Unnamed: 0
int64
0
16k
text_prompt
stringlengths
149
45.7k
code_prompt
stringlengths
67
140k
15,900
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: CSAL4243 Step1: Feature Scaling and Mean Normalization Step2: Initialize Hyper Parameters Step3: Model/Hypothesis Function Step5: Cost Function Step7: Gradient Descent Function Step8: Run Gradient Descent Step9: Plot Convergence Step10: Predict output using trained model
<ASSISTANT_TASK:> Python Code: %matplotlib inline import pandas as pd import numpy as np from sklearn import linear_model import matplotlib.pyplot as plt import matplotlib as mpl # read data in pandas frame dataframe = pd.read_csv('datasets/house_dataset2.csv', encoding='utf-8') # check data by printing first few rows dataframe.head() from mpl_toolkits.mplot3d import Axes3D fig = plt.figure() fig.set_size_inches(12.5, 7.5) ax = fig.add_subplot(111, projection='3d') ax.scatter(xs=dataframe['size'], ys=dataframe['bedrooms'], zs=dataframe['price']) ax.set_ylabel('bedrooms'); ax.set_xlabel('size'); ax.set_zlabel('price') # ax.view_init(10, -45) plt.show() Explanation: CSAL4243: Introduction to Machine Learning Muhammad Mudassir Khan (mudasssir.khan@ucp.edu.pk) Lecture 5: Multivariate Regression Overview Machine Learning pipeline Linear Regression with one variable Model Representation Vectorize Model Linear Regression with multiple variables Cost Function Gradient Descent Speed up gradient descent Feature Scaling Mean Normalization Combining Both Learning Rate $\alpha$ Automatic Covergence Test Linear Regression with Multiple Variables Example Read data Feature Scaling and Mean Normalization Initialize Hyper Parameters Model/Hypothesis Function Cost Function Gradient Descent Run Gradient Descent Plot Convergence Predict output using trained model Resources Credits <br> <br> Classification vs Regression <img style="float: left;" src="images/05_05.jpg" width=300> <img style="float: center;" src="images/05_04.png" width=400> <br> Machine Learning pipeline <img style="float: left;" src="images/model.png"> x is called input variables or input features. y is called output or target variable. Also sometimes known as label. h is called hypothesis or model. pair (x<sup>(i)</sup>,y<sup>(i)</sup>) is called a sample or training example dataset of all training examples is called training set. m is the number of samples in a dataset. n is the number of features in a dataset excluding label. <img style="float: left;" src="images/02_02.png", width=400> <br> <br> Linear Regression with one variable Model Representation Model is represented by h<sub>$\theta$</sub>(x) or simply h(x) For Linear regression with one input variable h(x) = $\theta$<sub>0</sub> + $\theta$<sub>1</sub>x <img style="float: left;" src="images/02_01.png"> $\theta$<sub>0</sub> and $\theta$<sub>1</sub> are called weights or parameters. Need to find $\theta$<sub>0</sub> and $\theta$<sub>1</sub> that maximizes the performance of model. <br> Vectorize Model <img style="float: right;" src="images/02_02.png" width=300> Write model in form of matrix multiplication $h(x)$ = $X \times \theta$ $X$ and $\theta$ are both matrices $X = \left[ \begin{array}{cc} x_1 \ x_2 \ x_3 \ ... \ x_{m} \end{array} \right]$ $h(x)$ = $\theta_0 + \theta_1 x$ = $X \times \theta$ = $\left[ \begin{array}{cc} 1 & x_i \end{array} \right] \times \left[ \begin{array}{cc} \theta_0 \ \theta_1 \end{array} \right]$ $h(x)$ = $\left[ \begin{array}{cc} \theta_0 + \theta_1 x_1 \ \theta_0 + \theta_1 x_2 \ \theta_0 + \theta_1 x_3 \ ... \ \theta_0 + \theta_1 x_{m} \end{array} \right] = \left[ \begin{array}{cc} 1 & x_1 \ 1 & x_2 \ 1 & x_3 \ ... \ 1 & x_{m} \end{array} \right] \times \left[ \begin{array}{cc} \theta_0 \ \theta_1 \end{array} \right]$ In given dataset $X$ has dimensions $m \times 1$ because of 1 variable $\theta$ has dimension $2\times 1$ Append a column vector of all 1's to X New X has dimensions $m\times 2$ $h(x) = X \times \theta$ has dimensions $m\times 1$ <br> Linear Regression with multiple variables <img style="float: right;" src="images/02_03.png" width=300> Model $h(x) = \theta_0 + \theta_1 x_1 + \theta_2 x_2 .... + \theta_n x_n$ Dimensions of $X$ is $m\times n$ $X = \left[ \begin{array}{cc} x_1^1 & x_1^2 & .. & x_1^{n} \ x_2^1 & x_2^2 & .. & x_2^{n} \ x_3^1 & x_3^2 & .. & x_3^{n} \ ... \ x_{m}^1 & x_{m}^2 & .. & x_{m}^{n} \end{array} \right]$ $\theta$ has dimension $(n+1)\times 1$ $\theta = \left[ \begin{array}{cc} \theta_0 \ \theta_1 \ \theta_2 \ ... \ \theta_{n} \ \end{array} \right]$ <br> - Append a column vector of all 1's to X - Now X has dimensions $m\times (n+1)$ $X = \left[ \begin{array}{cc} 1 & x_1^1 & x_1^2 & .. & x_1^{n} \ 1 & x_2^1 & x_2^2 & .. & x_2^{n} \ 1 & x_3^1 & x_3^2 & .. & x_3^{n} \ ... \ 1 & x_{m}^1 & x_{m}^2 & .. & x_{m}^{n} \end{array} \right]$ where $x_i$ is $i^{th}$ sample, e.g. $x_2 = [ \begin{array}{cc} 4.9 & 3.0 & 1.4 & 0.2 \end{array}]$ and $x_i^{j}$ is value of feature $j$ in the $i^{th}$ training example e.g. $x_2^3=1.4$ $h(x) = X \times \theta$ has dimensions $m\times 1$ <br> <br> <br> Cost Function Cost function = J($\theta$) = $\frac{1}{2m}\sum_{i=1}^{m} (h(x^i) - y^i)^2$ where $h(x) = \theta_0 + \theta_1 x_1 + \theta_2 x_2 .... + \theta_n x_n$ <img style="float: center;" src="images/03_02.png", width=300> <br> <br> Gradient Descent Cost function: J($\theta$) = $\frac{1}{2m}\sum_{i=1}^{m} (h(x^i) - y^i)^2$ Gradient descent equation: $\theta_j := \theta_j - \alpha \frac{\partial}{\partial \theta_j} J(\theta_0, \theta_1)$ <br> Replacing J($\theta$) for each j $\begin{align} & \text{repeat until convergence:} \; \lbrace \newline \; & \theta_0 := \theta_0 - \alpha \frac{1}{m} \sum\limits_{i=1}^{m} (h_\theta(x_{i}) - y_{i}) \cdot x^0_{i}\newline \; & \theta_1 := \theta_1 - \alpha \frac{1}{m} \sum\limits_{i=1}^{m} (h_\theta(x_{i}) - y_{i}) \cdot x^1_{i} \newline \; & \theta_2 := \theta_2 - \alpha \frac{1}{m} \sum\limits_{i=1}^{m} (h_\theta(x_{i}) - y_{i}) \cdot x^2_{i} \newline & \cdots \newline \rbrace \end{align}$ <br> or more generally $\begin{align}& \text{repeat until convergence:} \; \lbrace \newline \; & \theta_j := \theta_j - \alpha \frac{1}{m} \sum\limits_{i=1}^{m} (h_\theta(x_{i}) - y_{i}) \cdot x^j_{i} \; & \text{for j := 0...n}\newline \rbrace\end{align}$ <br> <img style="float: left;" src="images/03_04.gif"> Speed up gradient descent Tricks to make gradient descent converge faster to optimal value Each of our input values in roughly the same range. $\theta$ will descend quickly on small ranges and slowly on large ranges. $\theta$ Will oscillate inefficiently down to the optimum when the variables are very uneven. Aim is to have: $-1 \le x^i \le 1$ or $-0.5 \le x^i \le 0.5$ <br> Feature Scaling Divide the values of a feature by its range $x^i = \frac{x^i}{\max(x^i) - \min(x^i)}$ <img style="float: center;" src="images/05_06.png"> Mean Normalization Bring mean of each feature to zero $x^i = x^i - \mu^i$ where $\mu^i$ is the mean of feature $i$ Combining both $x^i = \frac{x^i - \mu^i}{\max(x^i) - \min(x^i)}$ or $x^i = \frac{x^i - \mu^i}{\rho^i}$ where $\rho^i$ is standard deviation of feature $i$ <br> Learning Rate $\alpha$ Appropriate $\alpha$ value will speed up gradient descent. If $\alpha$ is too small: slow convergence. If $\alpha$ is too large: may not decrease on every iteration and thus may not converge. <img src="images/05_01.png"> For implementation purpose, try out different values of $\alpha$ e.g. 0.001, 0.003, 0.01, 0.03, 0.1 and plot $J(\theta)$ with respect to iterations. Choose the one the makes gradient descent coverge quickly. <br> Automatic Covergence Test Plot $J(\theta)$ vs iterations. $J(\theta)$ should descrease on each iteration. If $J(\theta)$ decrease by a very small value in an iteration, you may have reached optimal value. <img style="float: left;" src="images/05_02.png"> <br> <br> Linear Regression with Multiple Variables Example Read data End of explanation dataframe.describe() #Quick visualize data plt.grid(True) plt.xlim([-1,5000]) dummy = plt.hist(dataframe["size"],label = 'Size') dummy = plt.hist(dataframe["bedrooms"],label = 'Bedrooms') plt.title('Clearly we need feature normalization.') plt.xlabel('Column Value') plt.ylabel('Counts') dummy = plt.legend() mean_size = dataframe["size"].mean() std_size = dataframe["size"].std() mean_bed = dataframe["bedrooms"].mean() std_bed = dataframe["bedrooms"].std() dataframe["size"] = (dataframe["size"] - mean_size)/std_size dataframe["bedrooms"] = (dataframe["bedrooms"] - mean_bed)/std_bed dataframe.describe() # reassign X # assign X X = np.array(dataframe[['size','bedrooms']]) X = np.insert(X,0,1,axis=1) #Quick visualize data plt.grid(True) plt.xlim([-5,5]) dummy = plt.hist(dataframe["size"],label = 'Size') dummy = plt.hist(dataframe["bedrooms"],label = 'Bedrooms') plt.title('Feature scaled and normalization.') plt.xlabel('Column Value') plt.ylabel('Counts') dummy = plt.legend() # assign X and y X = np.array(dataframe[['size','bedrooms']]) y = np.array(dataframe[['price']]) m = y.size # number of training examples # insert all 1's column for theta_0 X = np.insert(X,0,1,axis=1) # initialize theta # initial_theta = np.zeros((X.shape[1],1)) initial_theta = np.random.rand(X.shape[1],1) initial_theta X.shape initial_theta.shape Explanation: Feature Scaling and Mean Normalization End of explanation iterations = 1500 alpha = 0.1 Explanation: Initialize Hyper Parameters End of explanation def h(X, theta): #Linear hypothesis function hx = np.dot(X,theta) return hx Explanation: Model/Hypothesis Function End of explanation def computeCost(theta,X,y): #Cost function theta_start is an n- dimensional vector of initial theta guess X is matrix with n- columns and m- rows y is a matrix with m- rows and 1 column #note to self: *.shape is (rows, columns) return float((1./(2*m)) * np.dot((h(X,theta)-y).T,(h(X,theta)-y))) #Test that running computeCost with 0's as theta returns 65591548106.45744: initial_theta = np.zeros((X.shape[1],1)) #(theta is a vector with n rows and 1 columns (if X has n features) ) print (computeCost(initial_theta,X,y)) Explanation: Cost Function End of explanation #Actual gradient descent minimizing routine def gradientDescent(X, theta_start = np.zeros(2)): theta_start is an n- dimensional vector of initial theta guess X is matrix with n- columns and m- rows theta = theta_start j_history = [] #Used to plot cost as function of iteration theta_history = [] #Used to visualize the minimization path later on for meaninglessvariable in range(iterations): tmptheta = theta # append for plotting j_history.append(computeCost(theta,X,y)) theta_history.append(list(theta[:,0])) #Simultaneously updating theta values for j in range(len(tmptheta)): tmptheta[j] = theta[j] - (alpha/m)*np.sum((h(X,theta) - y)*np.array(X[:,j]).reshape(m,1)) theta = tmptheta return theta, theta_history, j_history Explanation: Gradient Descent Function End of explanation #Actually run gradient descent to get the best-fit theta values theta, thetahistory, j_history = gradientDescent(X,initial_theta) theta Explanation: Run Gradient Descent End of explanation plt.plot(j_history) plt.title("Convergence of Cost Function") plt.xlabel("Iteration number") plt.ylabel("Cost function") plt.show() Explanation: Plot Convergence End of explanation dataframe.head() x_test = np.array([1,0.130010,-0.22367]) print("$%0.2f" % float(h(x_test,theta))) hx = h(X, theta) from mpl_toolkits.mplot3d import Axes3D fig = plt.figure() fig.set_size_inches(12.5, 7.5) ax = fig.add_subplot(111, projection='3d') ax.scatter(xs=dataframe['size'], ys=dataframe['bedrooms'], zs=dataframe['price']) ax.set_ylabel('bedrooms'); ax.set_xlabel('size'); ax.set_zlabel('price') # ax.plot(xs=np.array(X[:,0],dtype=object).reshape(-1,1), ys=np.array(X[:,1],dtype=object).reshape(-1,1), zs=hx, color='green') ax.plot(X[:,0], X[:,1], np.array(hx[:,0]), label='fitted line', color='green') # ax.view_init(20, -165) plt.show() Explanation: Predict output using trained model End of explanation <END_TASK>
15,901
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: chart module This module relies on pygal library, so the returned charts are instances of pygal.chart. See options at pygal site I made a JavaScript 'equivalent' Step1: Time Series Step2: Chart series Step3: Chart seriesByRegion
<ASSISTANT_TASK:> Python Code: import ee from geetools import ui test_site = ee.Geometry.Point([-71, -42]) test_feat = ee.Feature(test_site, {'name': 'test feature'}) test_featcol = ee.FeatureCollection([ test_feat, test_feat.buffer(100).set('name', 'buffer 100'), test_feat.buffer(1000).set('name', 'buffer 1000') ]) Explanation: chart module This module relies on pygal library, so the returned charts are instances of pygal.chart. See options at pygal site I made a JavaScript 'equivalent': https://code.earthengine.google.com/b2922b860b85c1120250794fb82dfda8 End of explanation years = ee.List([2015, 2016, 2017, 2018]) col = ee.ImageCollection('COPERNICUS/S2').filterBounds(test_site) def make_time_series(year): ''' make a time series from year's list ''' eefilter = ee.Filter.calendarRange(year, field='year') filtered = col.filter(eefilter) return filtered.mean().set('system:time_start', ee.Date.fromYMD(year, 1, 1).millis()) time_series = ee.ImageCollection(years.map(make_time_series)) Explanation: Time Series End of explanation chart_ts = ui.chart.Image.series(**{ 'imageCollection': time_series, 'region': test_site, 'scale': 10, 'bands': ['B1', 'B2', 'B3'], # 'xProperty': 'B4', # You can use a band too! 'labels': ['band B1', 'B2 band', 'this is B3'] }) chart_ts.render_widget(width='50%') Explanation: Chart series End of explanation chart_ts_region = ui.chart.Image.seriesByRegion(**{ 'imageCollection': time_series, 'reducer': ee.Reducer.median(), 'regions': test_featcol, 'scale': 10, 'band': 'B11', 'seriesProperty': 'name' }) chart_ts_region.render_widget(height=500) Explanation: Chart seriesByRegion End of explanation <END_TASK>
15,902
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Truncated potentials for long-range electrostatics Björn Stenqvist and Mikael Lund, Div. of Theoretical Chemistry, Lund University, Sweden, 2016 In this notebook we test the validity of a variety of Wolf-inspired potentials and compare with a default Ewald summation on a sinple box of water. Install prerequisites using conda Step2: Create initial box of water First create a single water molecule pdb file, then use this to fill up a predefined box using the command line tool packmol. In this example, $N$ and $L$ are taken form the Gromacs spc216.gro file. Step4: Create OpenMM System classes for a variety of long-range correction schemes Here the idea is to create all the different setups and add then to the systemlist dictionary. Ewald summation (and PME) can be setup using the default NonbondedForce class, while CustomNonbondedForce must be used for custom pair potentials as used in Wolf, Stenqvist etc. For more information on how to control force objects and simulation, check, OpenMM python API. Lennard-Jones + Ewald at http Step5: Here we perform the actual MD simulation, incl. minimization of all defined systems Step6: Oxygen-oxygen radial distribution function Step7: Average potential energy Step8: Static dielectric constant The exact formula for $\epsilon_r$ is not given in the mdtraj documentation, but can be found in the source code on github, https Step9: Gromacs analysis We now generate a topology (top/tpr) file for Gromacs in order to use use their analysis tools, in particular for dipol correlations. These files are identitical for all electrostatic schemes. NOTE Step10: Convert trajectory to XTC format and calculate dipolar correlations See description of the -g option in gmx dipoles, relating to http
<ASSISTANT_TASK:> Python Code: %matplotlib inline from simtk.openmm.app import * from simtk.openmm import * from simtk.unit import * from sys import stdout import matplotlib, matplotlib.pyplot as plt import pandas as pd from io import StringIO import numpy as np, os matplotlib.rcParams.update({'font.size': 12}) # simulation parameters nsteps = 50000 # number of MD steps each of 2 fs nwater = 1000 # number of water molecules cutoff = 1.2*nanometers # pair potential cutoff platform = 'OpenCL' # 'OpenCL', 'CUDA', 'CPU' rho = 216/18.6206**3 # density from gromacs spc216.gro boxlen = (nwater/rho)**(1/3) print('box length = ', boxlen*angstrom) print('half box = ', boxlen*angstrom/2) Explanation: Truncated potentials for long-range electrostatics Björn Stenqvist and Mikael Lund, Div. of Theoretical Chemistry, Lund University, Sweden, 2016 In this notebook we test the validity of a variety of Wolf-inspired potentials and compare with a default Ewald summation on a sinple box of water. Install prerequisites using conda: bash $ conda config --add channels omnia $ conda install -c omnia openmm mdtraj packmol Also, Gromacs analysis tools (gmx) must be installed for parts of the analysis. End of explanation %%writefile hoh.pdb CRYST1 30.000 30.000 30.000 90.00 90.00 90.00 P 1 1 ATOM 1 OW HOH A 1 27.552 11.051 7.172 1.00 0.00 O ATOM 2 HW1 HOH A 1 27.900 10.721 8.050 1.00 0.00 H ATOM 3 HW2 HOH A 1 26.606 11.355 7.281 1.00 0.00 H END # write input file for packmol PACKMOL_INPUT = tolerance %f filetype pdb output %s # hoh will be put in a box # defined by the minimum coordinates x, y and z = 0. 0. 0. and maximum # coordinates box_size box_size box_size That is, they will be put in a cube of side # box_size (the keyword "inside cube 0. 0. 0. box_size") could be used as well. structure %s number %d inside box 0. 0. 0. %f %f %f add_box_sides 0.0 end structure % (2.,'water.pdb','hoh.pdb', nwater, boxlen, boxlen, boxlen) !echo '$PACKMOL_INPUT' > packmol_input.txt !packmol < packmol_input.txt > /dev/null Explanation: Create initial box of water First create a single water molecule pdb file, then use this to fill up a predefined box using the command line tool packmol. In this example, $N$ and $L$ are taken form the Gromacs spc216.gro file. End of explanation elec_to_kJmol = ( constants.elementary_charge**2 * AVOGADRO_CONSTANT_NA / (4*np.pi*1.0*8.854187817e-12 * constants.farad/constants.meter)).value_in_unit(kilojoule_per_mole*nanometer) def findForce(system, forcetype, add=True): Finds a specific force in the system force list - added if not found. for force in system.getForces(): if isinstance(force, forcetype): return force if add==True: system.addForce(forcetype()) return findForce(system, forcetype) return None def make_qpotential_system(topology, Rc=0.9*nanometers, moments=1000): ''' return a q-potential system (TO ''' ff = ForceField('spce-custom.xml') # this will create a CustomNonbondedForce system = ff.createSystem( topology, nonbondedMethod=CutoffPeriodic, nonbondedCutoff=Rc, constraints=HBonds, rigidWater=True) def qPochhammerSymbol( Rc, moments ): if isinstance(Rc, Quantity): Rc = Rc / nanometer # strip unit qP = 1.0 r = np.linspace(0, Rc, 5000) for i in range( moments ): qP *= (1 - (r/Rc)**(i+1) ) return qP qP = Continuous1DFunction( qPochhammerSymbol(Rc, moments), 0*nanometers, Rc) nonbonded = findForce(system, CustomNonbondedForce) nonbonded.addTabulatedFunction( 'qP', qP ) # 'qP(r)' can now be used in energy function nonbonded.addGlobalParameter( 'f', elec_to_kJmol ) # convert to kJ/mol nonbonded.setEnergyFunction( 'f * charge1 * charge2 * qP(r)/r' \ ' + 4 * epsilon * ( (sigma/r)^12 - (sigma/r)^6 )' \ ' ; sigma = 0.5 * ( sigma1+sigma2 ); epsilon = sqrt( epsilon1*epsilon2 )' ) print('qpot') print(' periodic boundaries: ', nonbonded.usesPeriodicBoundaryConditions()) print(' switching function: ', nonbonded.getUseSwitchingFunction()) print(' long-range correction:', nonbonded.getUseLongRangeCorrection()) print(' cutoff distance: ', nonbonded.getCutoffDistance()) print(' energy function: ', nonbonded.getEnergyFunction()) return system def make_grekiss_system(topology, Rc=0.9*nanometers): ''' return a grekis system ''' ff = ForceField('spce-custom.xml') # this will create a CustomNonbondedForce system = ff.createSystem( topology, nonbondedMethod=CutoffPeriodic, nonbondedCutoff=Rc, constraints=HBonds, rigidWater=True) nonbonded = findForce(system, CustomNonbondedForce) nonbonded.addGlobalParameter( 'Rc', Rc ) # 'Rc' can now be used in energy function nonbonded.addGlobalParameter( 'f', elec_to_kJmol ) # 'lB' bjerrum length in nm and kJ/mol nonbonded.setEnergyFunction( 'f * charge1 * charge2 * 1/r * ( 1 - 1.75*r/Rc + 5.25*(r/Rc)^5 - 7*(r/Rc)^6 + 2.5*(r/Rc)^7 )' \ ' + 4 * epsilon * ( (sigma/r)^12 - (sigma/r)^6 )' \ ' ; sigma = 0.5 * (sigma1+sigma2); epsilon = sqrt( epsilon1*epsilon2 )' ) print('fanourgakis') print(' periodic boundaries: ', nonbonded.usesPeriodicBoundaryConditions()) print(' switching function: ', nonbonded.getUseSwitchingFunction()) print(' long-range correction:', nonbonded.getUseLongRangeCorrection()) print(' cutoff distance: ', nonbonded.getCutoffDistance()) print(' energy function: ', nonbonded.getEnergyFunction()) return system def make_yonezawa_system(topology, Rc=0.9*nanometers): ''' return a yonezawa system ''' ff = ForceField('spce-custom.xml') # this will create a CustomNonbondedForce system = ff.createSystem( topology, nonbondedMethod=CutoffPeriodic, nonbondedCutoff=Rc, constraints=HBonds, rigidWater=True) nonbonded = findForce(system, CustomNonbondedForce) nonbonded.addGlobalParameter( 'Rc', Rc ) # 'Rc' can now be used in energy function nonbonded.addGlobalParameter( 'f', elec_to_kJmol ) # 'lB' bjerrum length in nm and kJ/mol nonbonded.setEnergyFunction( 'f * charge1 * charge2 * ( 1/r - 1/Rc + 1/Rc^2 * (r-Rc) )' \ ' + 4 * epsilon * ( (sigma/r)^12 - (sigma/r)^6 )' \ ' ; sigma = 0.5 * ( sigma1+sigma2 ); epsilon = sqrt( epsilon1 * epsilon2 )' ) print('yonezawa') print(' periodic boundaries: ', nonbonded.usesPeriodicBoundaryConditions()) print(' switching function: ', nonbonded.getUseSwitchingFunction()) print(' long-range correction:', nonbonded.getUseLongRangeCorrection()) print(' cutoff distance: ', nonbonded.getCutoffDistance()) print(' energy function: ', nonbonded.getEnergyFunction()) return system def make_ewald_system(topology, Rc=0.9*nanometers, method=Ewald, ljcorr=False): ''' returns an Ewald system ''' ff = ForceField('spce.xml') # this will create a NonbondedForce system = ff.createSystem( topology, nonbondedMethod=method, nonbondedCutoff=Rc, constraints=HBonds, rigidWater=True) nonbonded = findForce(system, NonbondedForce) nonbonded.setUseDispersionCorrection( ljcorr ) print('ewald') print(' err. tolerance: ', nonbonded.getEwaldErrorTolerance()) print(' LJ switching function:', nonbonded.getUseSwitchingFunction()) print(' LJ correction: ', nonbonded.getUseDispersionCorrection()) print(' cutoff distance: ', nonbonded.getCutoffDistance()) return system pdb = PDBFile('water.pdb') # here we define our list of system incl. plotting properties. All files from simulation # and analysis will be prefixed with 'qpot.', 'ewald.', etc. systemlist = { 'yonezawa': { 'system': make_yonezawa_system( pdb.topology, Rc=cutoff ), 'npt': False, 'run': False, 'color': 'blue', 'label': 'Yonezawa' }, 'grekiss': { 'system': make_grekiss_system( pdb.topology, Rc=cutoff ), 'npt': False, 'run': False, 'color': 'green', 'label': 'Fanourgakis' }, 'qpot' : { 'system': make_qpotential_system( pdb.topology, Rc=cutoff ), 'npt': False, 'run': False, 'color': 'red', 'label': '$q$ ($n=\infty$)' }, 'qpot2' : { 'system': make_qpotential_system( pdb.topology, Rc=cutoff, moments=2 ), 'npt': False, 'run': False, 'color': 'orange', 'label': '$q$ ($n=2$)' }, 'qpot3' : { 'system': make_qpotential_system( pdb.topology, Rc=cutoff, moments=3 ), 'npt': False, 'run': False, 'color': 'magenta', 'label': '$q$ ($n=3$)' }, 'ewald': { 'system': make_ewald_system( pdb.topology, Rc=cutoff ), 'npt': False, 'run': False, 'color': 'black', 'label': 'Ewald' } } Explanation: Create OpenMM System classes for a variety of long-range correction schemes Here the idea is to create all the different setups and add then to the systemlist dictionary. Ewald summation (and PME) can be setup using the default NonbondedForce class, while CustomNonbondedForce must be used for custom pair potentials as used in Wolf, Stenqvist etc. For more information on how to control force objects and simulation, check, OpenMM python API. Lennard-Jones + Ewald at http://docs.openmm.org/7.0.0/userguide/theory.html#nonbondedforce Long range correction beyond cut-off: http://docs.openmm.org/7.0.0/userguide/theory.html#customnonbondedforce TODO: Smoothen LJ part at $R_c$. Use switching function? End of explanation for name, prop in systemlist.items(): # loop over systems if prop['run']: print(name) if prop['npt']: # NVT -> NPT ensemble ? print(' adding barostat.') barostat = MonteCarloBarostat(1.0*bar, 298.15*kelvin, 25) system.addForce(barostat) integrator = LangevinIntegrator( 298.15*kelvin, 1.0/picoseconds, 2*femtoseconds ) integrator.setConstraintTolerance(0.00001) _platform = Platform.getPlatformByName( platform ) # CPU, OpenCL sim = Simulation(pdb.topology, prop['system'], integrator, _platform) sim.context.setPositions(pdb.positions) # set particle positions if os.path.isfile( name+'.chk' ): with open( name+'.chk', 'rb') as f: print(' loading restart file.') sim.context.loadCheckpoint( f.read() ) else: print(' minimizing energy...') sim.reporters.clear() %time sim.minimizeEnergy( tolerance=50*kilojoule/mole, maxIterations=1000 ) sim.context.setVelocitiesToTemperature( 298.15*kelvin ) # initial random velocities print(' running Production...') sim.reporters.clear() sim.reporters.append( DCDReporter( name+'.dcd', 1000) ) sim.reporters.append( StateDataReporter(name+'.energy', 1000, step=True, potentialEnergy=True, temperature=True, density=True) ) sim.reporters.append( StateDataReporter(stdout, 1000, step=True, potentialEnergy=True, temperature=True, density=True, separator='\t', progress=True, totalSteps = nsteps) ) %time sim.step( nsteps ) with open( name+'.chk', 'wb') as f: print(' saving restart file.') f.write( sim.context.createCheckpoint() ) # save final configuration to PDB file # todo: something is fishy here... water molecules are outside the box # in pdb although the trajectory looks fine. Å->nm problem? positions = sim.context.getState( getPositions=True ).getPositions() PDBFile.writeFile(sim.topology, positions, open( name+'.pdb', 'w')) print() Explanation: Here we perform the actual MD simulation, incl. minimization of all defined systems End of explanation import mdtraj as md for name, prop in systemlist.items(): print(name) traj = md.load(name+'.dcd', top=name+'.pdb') sel = traj.top.select('name O') OOpairs = traj.top.select_pairs('name O', 'name O') rdf = md.compute_rdf( traj, pairs=OOpairs, bin_width=0.005, r_range=[0.2, boxlen/2/10] ) plt.plot( rdf[0], rdf[1], label=prop['label'], color=prop['color'], lw=1) plt.xlim(0.2, 0.8) plt.xlabel(r'$r$/nm') plt.ylabel(r'$g(r)$') plt.legend(loc=0, frameon=False, fontsize='medium') Explanation: Oxygen-oxygen radial distribution function End of explanation for name, prop in systemlist.items(): step, Upot, temp, density = np.loadtxt(name+'.energy', skiprows=1, delimiter=',', unpack=True) Upot = Upot / pdb.topology.getNumResidues() # -> energy per water molecule plt.plot( step, Upot, label=prop['label'], color=prop['color'], lw=0.5 ) plt.plot( [0,step.max()], [Upot.mean(), Upot.mean()], color=prop['color'], ls='--', lw=0.5 ) prop['energy'] = Upot.mean() plt.xlabel(r'steps') plt.ylabel(r'$U_{pot}/N$ (kJ/mol)') plt.legend(loc=1, frameon=True, fontsize='small') Explanation: Average potential energy End of explanation def getChargeVector( force, stripunit=True ): ''' Extract particle charge array from force object if available Note that this function assumes that charge is located at parameter position 0 which is default in NonbondedForce while it can be anywhere in CustomNonbondedForce, following the order in which parameters were added in python/xml. ''' if isinstance(force, (NonbondedForce, CustomNonbondedForce) ): if 'getParticleParameters' in dir(force): chargelist = [] for i in range( force.getNumParticles() ): charge = force.getParticleParameters( i )[ 0 ] if stripunit: if isinstance(charge, Quantity): charge = charge / elementary_charge # strip unit chargelist.append( charge ) return chargelist return None for name, prop in systemlist.items(): for force in prop['system'].getForces(): if isinstance(force, (NonbondedForce, CustomNonbondedForce) ): charges = getChargeVector( force ) traj = md.load(name+'.dcd', top=name+'.pdb') l = len(traj) l2= int(l/2) eps12= md.static_dielectric( traj[1:l], charges, temperature=298.15) # all eps1 = md.static_dielectric( traj[1:l2], charges, temperature=298.15) # 1st half eps2 = md.static_dielectric( traj[l2:l], charges, temperature=298.15) # 2nd half prop['diel'] = eps12 print(name, ': diel. const.', eps12, eps1, eps2) pd.DataFrame(systemlist).T Explanation: Static dielectric constant The exact formula for $\epsilon_r$ is not given in the mdtraj documentation, but can be found in the source code on github, https://github.com/mdtraj/mdtraj/blob/master/mdtraj/geometry/thermodynamic_properties.py#L90 End of explanation %%bash rm -fR echo -n "6\n" | gmx -quiet -nobackup pdb2gmx -f water.pdb -o water.gro -p water.top -n water.ndx -water spce gmx -quiet -nobackup grompp -f water.mdp -c water.gro -p water.top -o water.tpr -maxwarn 10 Explanation: Gromacs analysis We now generate a topology (top/tpr) file for Gromacs in order to use use their analysis tools, in particular for dipol correlations. These files are identitical for all electrostatic schemes. NOTE: Depending on the size of the water box, grompp may complain that the cut-off is too short. Fix this by increasing the verlet tolerance. It will not matter for the analysis, but consider making a slightly larger system. End of explanation for name, prop in systemlist.items(): print(name) traj = md.load(name+'.dcd', top=name+'.pdb') traj.save_xtc(name+'.xtc') !echo -n "0\n" | gmx -quiet -nobackup dipoles -f $name'.xtc' -s water.tpr -temp 298.15 -g $name'.gkr.xvg' print() fig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(nrows=2, ncols=2, figsize=(8, 7)) for name, prop in systemlist.items(): r, G, cos, hOO, gOO, energy = np.loadtxt(name+'.gkr.xvg', skiprows=27, unpack=True) ax1.plot(r, G, label=prop['label'], color=prop['color'], lw=1 ) ax1.set_xlabel('$r$/nm') ax1.set_ylabel('$G_k(r)$') ax2.plot(r, cos, label=prop['label'], color=prop['color'], lw=1 ) ax2.set_xlabel('$r$/nm') ax2.set_ylabel('cosine something...') ax2.set_xlim(0.2, 0.8) ax2.set_ylim(-0.1, 0.55) ax3.plot(r, hOO, label=prop['label'], color=prop['color'], lw=1 ) ax3.set_xlabel('$r$/nm') ax3.set_ylabel('hOO') ax3.set_xlim(0.2, 0.8) ax4.plot(r, gOO, label=prop['label'], color=prop['color'], lw=1 ) ax4.set_xlabel('$r$/nm') ax4.set_ylabel('gOO') ax4.set_xlim(0.2, 0.8) ax2.legend(loc=0, frameon=False, fontsize='medium') plt.tight_layout() Explanation: Convert trajectory to XTC format and calculate dipolar correlations See description of the -g option in gmx dipoles, relating to http://manual.gromacs.org/programs/gmx-dipoles.html (Nymand/Linse). End of explanation <END_TASK>
15,903
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: A simple method for segmenting rectilinear crop plots The following outlines the method in https Step1: Read in a test image subset. Replace with your own if required parameters will need to be adjusted, needless to say complete segmentation is not guaranteed - will be dependent upon your image. Step2: The classical Canny edge detection. Experimental GUI below, we follow Canny's original rule of thumb with the low threshold being divided by 2. From this we can derive the parameters for the crop segmentation. Step3: Phase congruency edge detection See https Step4: Segment the plots The function houghseg segments the plots. In this case, phase congruency is 'cleaner' edge detector, so it is used. We are interested in the larger rectangular plots here so we use the min_area param to constrain this.
<ASSISTANT_TASK:> Python Code: %matplotlib inline import cv2 from geospatial_learn import raster from geospatial_learn.utilities import do_phasecong, houghseg from math import ceil import matplotlib.pyplot as plt from skimage.color import rgb2gray, label2rgb from skimage.feature import canny from skimage.exposure import rescale_intensity Explanation: A simple method for segmenting rectilinear crop plots The following outlines the method in https://www.frontiersin.org/articles/10.3389/fpls.2020.591886/full employing simple image processing techniques to extract crop plots. This is easily implemented using scikit-image or open cv, but available as a functionality within this library. The method employs edge detection, hough line detection then segment reduction to arrive at the plots of interest. The merit is the simplicity, though it is somewhat inflexible being limited to rectilinear plots. Needless to say complete segmentation is not guaranteed - will be dependent upon your image. End of explanation inRas = 'figures/weetestorig.tif' img = raster.raster2array(inRas, bands=[1,2,3]) # for testing below gray = rgb2gray(img) plt.imshow(img) plt.show() Explanation: Read in a test image subset. Replace with your own if required parameters will need to be adjusted, needless to say complete segmentation is not guaranteed - will be dependent upon your image. End of explanation def icanny(high_threshold, *args, **kwargs): #...do it inIm = gray#.astype(np.float) low_threshold = high_threshold / 2 edge = canny(inIm, low_threshold=low_threshold, high_threshold=high_threshold, *args, **kwargs) # Comment the first 2 lines if you want more space plt.figure(figsize=(15,15)) plt.subplot(121) plt.imshow(img) plt.subplot(122) plt.imshow(edge) plt.show() # return edge from ipywidgets import widgets cTester = widgets.interact(icanny, #k=widgets.IntSlider(min=3, max=100, step=2, continuous_update=False), sigma=widgets.IntSlider(min=0, max=100, step=1, continuous_update=False), #low_threshold=widgets.IntSlider(min=0, max=255, step=1, continuous_update=False), high_threshold=widgets.FloatSlider(min=0, max=1, step=0.01, continuous_update=False)) Explanation: The classical Canny edge detection. Experimental GUI below, we follow Canny's original rule of thumb with the low threshold being divided by 2. From this we can derive the parameters for the crop segmentation. End of explanation def iphase(*args, **kwargs): plt.figure(figsize=(15,15)) plt.subplot(121) plt.imshow(img) plt.subplot(122) edge = do_phasecong(gray, *args, **kwargs) plt.imshow(edge) plt.show() from ipywidgets import widgets cTester = widgets.interact(iphase, sigma=widgets.IntSlider(min=0, max=50, step=1, continuous_update=False), low_t=widgets.IntSlider(min=0, max=256, step=1, continuous_update=False), hi_t=widgets.IntSlider(min=0, max=256, step=1, continuous_update=False)) Explanation: Phase congruency edge detection See https://link.springer.com/article/10.1007/s004260000024 An alternative method for edges, which may or may not prove more useful albeit at a reduced speed with the python version ustilised here. End of explanation outShp = 'mytest.shp' segments = houghseg(inRas, outShp, edge='phase', sigma=4, min_area=4) plt.figure(figsize=(15,15)) plt.subplot(121) plt.imshow(img) plt.subplot(122) plt.imshow(segments, cmap='gray') Explanation: Segment the plots The function houghseg segments the plots. In this case, phase congruency is 'cleaner' edge detector, so it is used. We are interested in the larger rectangular plots here so we use the min_area param to constrain this. End of explanation <END_TASK>
15,904
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Basic notebook to look @ convergence of a 2D region in an FES. It will actually call sum hills with the stride you set in cell one , graph the FES and put the regions of convergence there Step1: Graph the final FES and plot the two squares on top of it Step2: The two functions below calculate the average free energy of a region by integrating over whichever boxes you defined above. Since the FES is discrete and points are equally spaced, this is trivially taken as a summation Step3: Below this is all testing of different read-in options Step4: Profiling speed of different read in options
<ASSISTANT_TASK:> Python Code: import numpy as np import matplotlib.pyplot as plt import glob import os from matplotlib.patches import Rectangle # define all variables for convergence script # these will pass to the bash magic below used to call plumed sum_hills dir="MetaD_converge" #where the intermediate fes will be stored hills="other/HILLS" #your HILLS file from the simulation finalfes='other/fes.dat' #the final fes.dat file stride=1000 kT=8.314e-3*300 #throughout we convert to kcal, but the HILLS are assumed to be in GROMACS units (kJ) ## here is where you set the boxes to define convergence regions C1=[-1.5,1.0] #center of box 1 C2=[1.0,-.5] edge1=1.0 #edge of box1 edge2=1.0 %%bash -s "$dir" "$hills" "$stride" "$kT" # calling sum hills and output to devnul HILLSFILE=HILLS rm -rf $1 mkdir $1 cp $2 $1 cd $1 plumed sum_hills --hills $HILLSFILE --kt $4 --stride $3 >& /dev/null Explanation: Basic notebook to look @ convergence of a 2D region in an FES. It will actually call sum hills with the stride you set in cell one , graph the FES and put the regions of convergence there End of explanation %matplotlib inline #read the data in from a text file fesdata = np.genfromtxt(finalfes,comments='#'); fesdata = fesdata[:,0:3] #what was your grid size? this calculates it dim=int(np.sqrt(np.size(fesdata)/3)) #some post-processing to be compatible with contourf X=np.reshape(fesdata[:,0],[dim,dim],order="F") #order F was 20% faster than A/C Y=np.reshape(fesdata[:,1],[dim,dim],order="F") Z=np.reshape((fesdata[:,2]-np.min(fesdata[:,2]))/4.184,[dim,dim],order="F") #convert to kcal/mol #what spacing do you want? assume units are in kJ/mol spacer=1 #this means 1kcal/mol spacing lines=20 levels=np.linspace(0,lines*spacer,num=(lines+1),endpoint=True) fig=plt.figure(figsize=(8,6)) axes = fig.add_subplot(111) xlabel='$\Phi$' ylabel='$\Psi$' plt.contourf(X, Y, Z, levels, cmap=plt.cm.bone,) plt.colorbar() axes.set_xlabel(xlabel, fontsize=20) axes.set_ylabel(ylabel, fontsize=20) currentAxis = plt.gca() currentAxis.add_patch(Rectangle((C1[0]-edge1/2, C1[1]-edge1/2), edge1, edge1,facecolor='none',edgecolor='yellow',linewidth='3')) currentAxis.add_patch(Rectangle((C2[0]-edge2/2, C2[1]-edge2/2), edge2, edge2,facecolor='none',edgecolor='yellow',linewidth='3')) plt.show() Explanation: Graph the final FES and plot the two squares on top of it End of explanation def diffNP(file): #read the data in from a text file # note - this is very slow fesdata = np.genfromtxt(file,comments='#'); A=0.0 B=0.0 dim=np.shape(fesdata)[0] for i in range(0, dim): x=fesdata[i][0] y=fesdata[i][1] z=fesdata[i][2] if x < C1[0]+edge1/2 and x > C1[0]-edge1/2 and y > C1[1]-edge1/2 and y < C1[1]+edge1/2: A+=np.exp(-z/kT) if x < C2[0]+edge2/2 and x > C2[0]-edge2/2 and y > C2[1]-edge2/2 and y < C2[1]+edge2/2: B+=np.exp(-z/kT) A=-kT*np.log(A) B=-kT*np.log(B) diff=(A-B)/4.184 #output in kcal return diff def diff(file): kT=8.314e-3*300 A=0.0 B=0.0 f = open(file, 'r') for line in f: if line[:1] != '#': line=line.strip() if line: columns = line.split() x=float(columns[0]) y=float(columns[1]) z=float(columns[2]) if x < C1[0]+edge1/2 and x > C1[0]-edge1/2 and y > C1[1]-edge1/2 and y < C1[1]+edge1/2: A+=np.exp(-z/kT) if x < C2[0]+edge2/2 and x > C2[0]-edge2/2 and y > C2[1]-edge2/2 and y < C2[1]+edge2/2: B+=np.exp(-z/kT) f.close A=-kT*np.log(A) B=-kT*np.log(B) diff=(A-B)/4.184 return diff diffvec=None rootdir = '/Users/jpfaendt/Learning/Python/ALA2_MetaD/MetaD_converge' i=0 diffvec=np.zeros((1,2)) #the variable func defines which function you are going to call to read in your data files fes_*.dat #func=diffNP uses the numpy read in (SLOW) #func=diff streams in data from a text file #to experience the differnece , uncomment out the print statements and run each way func=diff for infile in glob.glob( os.path.join(rootdir, 'fes_?.dat') ): if i >= i: diffvec.resize((i+1,2)) #print "current file is: " + infile diffvec[i][0]=i*1.0 diffvec[i][1]=func(infile) i+=1 for infile in glob.glob( os.path.join(rootdir, 'fes_??.dat') ): if i >= i: diffvec.resize((i+1,2)) #print "current file is: " + infile diffvec[i][0]=i*1.0 diffvec[i][1]=func(infile) i+=1 for infile in glob.glob( os.path.join(rootdir, 'fes_???.dat') ): if i >= i: diffvec.resize((i+1,2)) #print "current file is: " + infile diffvec[i][0]=i*1.0 diffvec[i][1]=func(infile) i+=1 fig = plt.figure(figsize=(6,6)) axes = fig.add_subplot(111) xlabel='time (generic)' ylabel='diff (A-B) (kcal/mol)' axes.plot(diffvec[:,0],diffvec[:,1]) axes.set_xlabel(xlabel, fontsize=20) axes.set_ylabel(ylabel, fontsize=20) plt.show() Explanation: The two functions below calculate the average free energy of a region by integrating over whichever boxes you defined above. Since the FES is discrete and points are equally spaced, this is trivially taken as a summation: $F_A = -k_BT * ln \sum_A exp\left(-F_{Ai}/k_BT\right) $ Don't forget that this is formally a free-energy plus some trivial constant but that the constant is equal for both regions $A$ and $B$ so that you will obtain the same free-energy difference irrespective of the reference point. On the other hand, it doesn't make much sense to just use the arbitrary nubmers coming from sum_hills, which are related only to the amount of aggregate bias produced in your simulation. This is why we reference the lowest point to zero on the contour plots. I left both functions in as a teaching tool to show how slow np.genfromtext is End of explanation ## #read the data in from a text file using genfrom txt fesdata = np.genfromtxt('MetaD_converge/fes_1.dat',comments='#'); kT=8.314e-3*300 A=0.0 B=0.0 dim=np.shape(fesdata)[0] for i in range(0, dim): x=fesdata[i][0] y=fesdata[i][1] z=fesdata[i][2] if x < C1[0]+edge1/2 and x > C1[0]-edge1/2 and y > C1[1]-edge1/2 and y < C1[1]+edge1/2: A+=np.exp(-z/kT) if x < C2[0]+edge2/2 and x > C2[0]-edge2/2 and y > C2[1]-edge2/2 and y < C2[1]+edge2/2: B+=np.exp(-z/kT) A=-kT*np.log(A) B=-kT*np.log(B) diff=(A-B)/4.184 diff ## #read the data in from a text file using read in commands kT=8.314e-3*300 A=0.0 B=0.0 f = open('MetaD_converge/fes_1.dat', 'r') for line in f: if line[:1] != '#': line=line.strip() if line: columns = line.split() x=float(columns[0]) y=float(columns[1]) z=float(columns[2]) if x < C1[0]+edge1/2 and x > C1[0]-edge1/2 and y > C1[1]-edge1/2 and y < C1[1]+edge1/2: A+=np.exp(-z/kT) if x < C2[0]+edge2/2 and x > C2[0]-edge2/2 and y > C2[1]-edge2/2 and y < C2[1]+edge2/2: B+=np.exp(-z/kT) f.close A=-kT*np.log(A) B=-kT*np.log(B) diff=(A-B)/4.184 diff Explanation: Below this is all testing of different read-in options: End of explanation file='MetaD/fes.dat' %timeit diffNP(file) %timeit diff(file) Explanation: Profiling speed of different read in options: End of explanation <END_TASK>
15,905
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Batch Normalization – Lesson What is it? What are it's benefits? How do we add it to a network? Let's see it work! What are you hiding? What is Batch Normalization?<a id='theory'></a> Batch normalization was introduced in Sergey Ioffe's and Christian Szegedy's 2015 paper Batch Normalization Step6: Neural network classes for testing The following class, NeuralNet, allows us to create identical neural networks with and without batch normalization. The code is heaviy documented, but there is also some additional discussion later. You do not need to read through it all before going through the rest of the notebook, but the comments within the code blocks may answer some of your questions. About the code Step9: There are quite a few comments in the code, so those should answer most of your questions. However, let's take a look at the most important lines. We add batch normalization to layers inside the fully_connected function. Here are some important points about that code Step10: Comparisons between identical networks, with and without batch normalization The next series of cells train networks with various settings to show the differences with and without batch normalization. They are meant to clearly demonstrate the effects of batch normalization. We include a deeper discussion of batch normalization later in the notebook. The following creates two networks using a ReLU activation function, a learning rate of 0.01, and reasonable starting weights. Step11: As expected, both networks train well and eventually reach similar test accuracies. However, notice that the model with batch normalization converges slightly faster than the other network, reaching accuracies over 90% almost immediately and nearing its max acuracy in 10 or 15 thousand iterations. The other network takes about 3 thousand iterations to reach 90% and doesn't near its best accuracy until 30 thousand or more iterations. If you look at the raw speed, you can see that without batch normalization we were computing over 1100 batches per second, whereas with batch normalization that goes down to just over 500. However, batch normalization allows us to perform fewer iterations and converge in less time over all. (We only trained for 50 thousand batches here so we could plot the comparison.) The following creates two networks with the same hyperparameters used in the previous example, but only trains for 2000 iterations. Step12: As you can see, using batch normalization produces a model with over 95% accuracy in only 2000 batches, and it was above 90% at somewhere around 500 batches. Without batch normalization, the model takes 1750 iterations just to hit 80% – the network with batch normalization hits that mark after around 200 iterations! (Note Step13: With the number of layers we're using and this small learning rate, using a sigmoid activation function takes a long time to start learning. It eventually starts making progress, but it took over 45 thousand batches just to get over 80% accuracy. Using batch normalization gets to 90% in around one thousand batches. The following creates two networks using a ReLU activation function, a learning rate of 1, and reasonable starting weights. Step14: Now we're using ReLUs again, but with a larger learning rate. The plot shows how training started out pretty normally, with the network with batch normalization starting out faster than the other. But the higher learning rate bounces the accuracy around a bit more, and at some point the accuracy in the network without batch normalization just completely crashes. It's likely that too many ReLUs died off at this point because of the high learning rate. The next cell shows the same test again. The network with batch normalization performs the same way, and the other suffers from the same problem again, but it manages to train longer before it happens. Step15: In both of the previous examples, the network with batch normalization manages to gets over 98% accuracy, and get near that result almost immediately. The higher learning rate allows the network to train extremely fast. The following creates two networks using a sigmoid activation function, a learning rate of 1, and reasonable starting weights. Step16: In this example, we switched to a sigmoid activation function. It appears to hande the higher learning rate well, with both networks achieving high accuracy. The cell below shows a similar pair of networks trained for only 2000 iterations. Step17: As you can see, even though these parameters work well for both networks, the one with batch normalization gets over 90% in 400 or so batches, whereas the other takes over 1700. When training larger networks, these sorts of differences become more pronounced. The following creates two networks using a ReLU activation function, a learning rate of 2, and reasonable starting weights. Step18: With this very large learning rate, the network with batch normalization trains fine and almost immediately manages 98% accuracy. However, the network without normalization doesn't learn at all. The following creates two networks using a sigmoid activation function, a learning rate of 2, and reasonable starting weights. Step19: Once again, using a sigmoid activation function with the larger learning rate works well both with and without batch normalization. However, look at the plot below where we train models with the same parameters but only 2000 iterations. As usual, batch normalization lets it train faster. Step20: In the rest of the examples, we use really bad starting weights. That is, normally we would use very small values close to zero. However, in these examples we choose randome values with a standard deviation of 5. If you were really training a neural network, you would not want to do this. But these examples demonstrate how batch normalization makes your network much more resilient. The following creates two networks using a ReLU activation function, a learning rate of 0.01, and bad starting weights. Step21: As the plot shows, without batch normalization the network never learns anything at all. But with batch normalization, it actually learns pretty well and gets to almost 80% accuracy. The starting weights obviously hurt the network, but you can see how well batch normalization does in overcoming them. The following creates two networks using a sigmoid activation function, a learning rate of 0.01, and bad starting weights. Step22: Using a sigmoid activation function works better than the ReLU in the previous example, but without batch normalization it would take a tremendously long time to train the network, if it ever trained at all. The following creates two networks using a ReLU activation function, a learning rate of 1, and bad starting weights.<a id="successful_example_lr_1"></a> Step23: The higher learning rate used here allows the network with batch normalization to surpass 90% in about 30 thousand batches. The network without it never gets anywhere. The following creates two networks using a sigmoid activation function, a learning rate of 1, and bad starting weights. Step24: Using sigmoid works better than ReLUs for this higher learning rate. However, you can see that without batch normalization, the network takes a long time tro train, bounces around a lot, and spends a long time stuck at 90%. The network with batch normalization trains much more quickly, seems to be more stable, and achieves a higher accuracy. The following creates two networks using a ReLU activation function, a learning rate of 2, and bad starting weights.<a id="successful_example_lr_2"></a> Step25: We've already seen that ReLUs do not do as well as sigmoids with higher learning rates, and here we are using an extremely high rate. As expected, without batch normalization the network doesn't learn at all. But with batch normalization, it eventually achieves 90% accuracy. Notice, though, how its accuracy bounces around wildly during training - that's because the learning rate is really much too high, so the fact that this worked at all is a bit of luck. The following creates two networks using a sigmoid activation function, a learning rate of 2, and bad starting weights. Step26: In this case, the network with batch normalization trained faster and reached a higher accuracy. Meanwhile, the high learning rate makes the network without normalization bounce around erratically and have trouble getting past 90%. Full Disclosure Step27: When we used these same parameters earlier, we saw the network with batch normalization reach 92% validation accuracy. This time we used different starting weights, initialized using the same standard deviation as before, and the network doesn't learn at all. (Remember, an accuracy around 10% is what the network gets if it just guesses the same value all the time.) The following creates two networks using a ReLU activation function, a learning rate of 2, and bad starting weights. Step29: When we trained with these parameters and batch normalization earlier, we reached 90% validation accuracy. However, this time the network almost starts to make some progress in the beginning, but it quickly breaks down and stops learning. Note Step31: This version of fully_connected is much longer than the original, but once again has extensive comments to help you understand it. Here are some important points Step32: In the following cell, we pass True for test_training_accuracy, which performs the same batch normalization that we normally perform during training. Step33: As you can see, the network guessed the same value every time! But why? Because during training, a network with batch normalization adjusts the values at each layer based on the mean and variance of that batch. The "batches" we are using for these predictions have a single input each time, so their values are the means, and their variances will always be 0. That means the network will normalize the values at any layer to zero. (Review the equations from before to see why a value that is equal to the mean would always normalize to zero.) So we end up with the same result for every input we give the network, because its the value the network produces when it applies its learned weights to zeros at every layer. Note
<ASSISTANT_TASK:> Python Code: # Import necessary packages import tensorflow as tf import tqdm import numpy as np import matplotlib.pyplot as plt %matplotlib inline # Import MNIST data so we have something for our experiments from tensorflow.examples.tutorials.mnist import input_data mnist = input_data.read_data_sets("MNIST_data/", one_hot=True) Explanation: Batch Normalization – Lesson What is it? What are it's benefits? How do we add it to a network? Let's see it work! What are you hiding? What is Batch Normalization?<a id='theory'></a> Batch normalization was introduced in Sergey Ioffe's and Christian Szegedy's 2015 paper Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift. The idea is that, instead of just normalizing the inputs to the network, we normalize the inputs to layers within the network. It's called "batch" normalization because during training, we normalize each layer's inputs by using the mean and variance of the values in the current mini-batch. Why might this help? Well, we know that normalizing the inputs to a network helps the network learn. But a network is a series of layers, where the output of one layer becomes the input to another. That means we can think of any layer in a neural network as the first layer of a smaller network. For example, imagine a 3 layer network. Instead of just thinking of it as a single network with inputs, layers, and outputs, think of the output of layer 1 as the input to a two layer network. This two layer network would consist of layers 2 and 3 in our original network. Likewise, the output of layer 2 can be thought of as the input to a single layer network, consistng only of layer 3. When you think of it like that - as a series of neural networks feeding into each other - then it's easy to imagine how normalizing the inputs to each layer would help. It's just like normalizing the inputs to any other neural network, but you're doing it at every layer (sub-network). Beyond the intuitive reasons, there are good mathematical reasons why it helps the network learn better, too. It helps combat what the authors call internal covariate shift. This discussion is best handled in the paper and in Deep Learning a book you can read online written by Ian Goodfellow, Yoshua Bengio, and Aaron Courville. Specifically, check out the batch normalization section of Chapter 8: Optimization for Training Deep Models. Benefits of Batch Normalization<a id="benefits"></a> Batch normalization optimizes network training. It has been shown to have several benefits: 1. Networks train faster – Each training iteration will actually be slower because of the extra calculations during the forward pass and the additional hyperparameters to train during back propagation. However, it should converge much more quickly, so training should be faster overall. 2. Allows higher learning rates – Gradient descent usually requires small learning rates for the network to converge. And as networks get deeper, their gradients get smaller during back propagation so they require even more iterations. Using batch normalization allows us to use much higher learning rates, which further increases the speed at which networks train. 3. Makes weights easier to initialize – Weight initialization can be difficult, and it's even more difficult when creating deeper networks. Batch normalization seems to allow us to be much less careful about choosing our initial starting weights. 4. Makes more activation functions viable – Some activation functions do not work well in some situations. Sigmoids lose their gradient pretty quickly, which means they can't be used in deep networks. And ReLUs often die out during training, where they stop learning completely, so we need to be careful about the range of values fed into them. Because batch normalization regulates the values going into each activation function, non-linearlities that don't seem to work well in deep networks actually become viable again. 5. Simplifies the creation of deeper networks – Because of the first 4 items listed above, it is easier to build and faster to train deeper neural networks when using batch normalization. And it's been shown that deeper networks generally produce better results, so that's great. 6. Provides a bit of regularlization – Batch normalization adds a little noise to your network. In some cases, such as in Inception modules, batch normalization has been shown to work as well as dropout. But in general, consider batch normalization as a bit of extra regularization, possibly allowing you to reduce some of the dropout you might add to a network. 7. May give better results overall – Some tests seem to show batch normalization actually improves the train.ing results. However, it's really an optimization to help train faster, so you shouldn't think of it as a way to make your network better. But since it lets you train networks faster, that means you can iterate over more designs more quickly. It also lets you build deeper networks, which are usually better. So when you factor in everything, you're probably going to end up with better results if you build your networks with batch normalization. Batch Normalization in TensorFlow<a id="implementation_1"></a> This section of the notebook shows you one way to add batch normalization to a neural network built in TensorFlow. The following cell imports the packages we need in the notebook and loads the MNIST dataset to use in our experiments. However, the tensorflow package contains all the code you'll actually need for batch normalization. End of explanation class NeuralNet: def __init__(self, initial_weights, activation_fn, use_batch_norm): Initializes this object, creating a TensorFlow graph using the given parameters. :param initial_weights: list of NumPy arrays or Tensors Initial values for the weights for every layer in the network. We pass these in so we can create multiple networks with the same starting weights to eliminate training differences caused by random initialization differences. The number of items in the list defines the number of layers in the network, and the shapes of the items in the list define the number of nodes in each layer. e.g. Passing in 3 matrices of shape (784, 256), (256, 100), and (100, 10) would create a network with 784 inputs going into a hidden layer with 256 nodes, followed by a hidden layer with 100 nodes, followed by an output layer with 10 nodes. :param activation_fn: Callable The function used for the output of each hidden layer. The network will use the same activation function on every hidden layer and no activate function on the output layer. e.g. Pass tf.nn.relu to use ReLU activations on your hidden layers. :param use_batch_norm: bool Pass True to create a network that uses batch normalization; False otherwise Note: this network will not use batch normalization on layers that do not have an activation function. # Keep track of whether or not this network uses batch normalization. self.use_batch_norm = use_batch_norm self.name = "With Batch Norm" if use_batch_norm else "Without Batch Norm" # Batch normalization needs to do different calculations during training and inference, # so we use this placeholder to tell the graph which behavior to use. self.is_training = tf.placeholder(tf.bool, name="is_training") # This list is just for keeping track of data we want to plot later. # It doesn't actually have anything to do with neural nets or batch normalization. self.training_accuracies = [] # Create the network graph, but it will not actually have any real values until after you # call train or test self.build_network(initial_weights, activation_fn) def build_network(self, initial_weights, activation_fn): Build the graph. The graph still needs to be trained via the `train` method. :param initial_weights: list of NumPy arrays or Tensors See __init__ for description. :param activation_fn: Callable See __init__ for description. self.input_layer = tf.placeholder(tf.float32, [None, initial_weights[0].shape[0]]) layer_in = self.input_layer for weights in initial_weights[:-1]: layer_in = self.fully_connected(layer_in, weights, activation_fn) self.output_layer = self.fully_connected(layer_in, initial_weights[-1]) def fully_connected(self, layer_in, initial_weights, activation_fn=None): Creates a standard, fully connected layer. Its number of inputs and outputs will be defined by the shape of `initial_weights`, and its starting weight values will be taken directly from that same parameter. If `self.use_batch_norm` is True, this layer will include batch normalization, otherwise it will not. :param layer_in: Tensor The Tensor that feeds into this layer. It's either the input to the network or the output of a previous layer. :param initial_weights: NumPy array or Tensor Initial values for this layer's weights. The shape defines the number of nodes in the layer. e.g. Passing in 3 matrix of shape (784, 256) would create a layer with 784 inputs and 256 outputs. :param activation_fn: Callable or None (default None) The non-linearity used for the output of the layer. If None, this layer will not include batch normalization, regardless of the value of `self.use_batch_norm`. e.g. Pass tf.nn.relu to use ReLU activations on your hidden layers. # Since this class supports both options, only use batch normalization when # requested. However, do not use it on the final layer, which we identify # by its lack of an activation function. if self.use_batch_norm and activation_fn: # Batch normalization uses weights as usual, but does NOT add a bias term. This is because # its calculations include gamma and beta variables that make the bias term unnecessary. # (See later in the notebook for more details.) weights = tf.Variable(initial_weights) linear_output = tf.matmul(layer_in, weights) # Apply batch normalization to the linear combination of the inputs and weights batch_normalized_output = tf.layers.batch_normalization(linear_output, training=self.is_training) # Now apply the activation function, *after* the normalization. return activation_fn(batch_normalized_output) else: # When not using batch normalization, create a standard layer that multiplies # the inputs and weights, adds a bias, and optionally passes the result # through an activation function. weights = tf.Variable(initial_weights) biases = tf.Variable(tf.zeros([initial_weights.shape[-1]])) linear_output = tf.add(tf.matmul(layer_in, weights), biases) return linear_output if not activation_fn else activation_fn(linear_output) def train(self, session, learning_rate, training_batches, batches_per_sample, save_model_as=None): Trains the model on the MNIST training dataset. :param session: Session Used to run training graph operations. :param learning_rate: float Learning rate used during gradient descent. :param training_batches: int Number of batches to train. :param batches_per_sample: int How many batches to train before sampling the validation accuracy. :param save_model_as: string or None (default None) Name to use if you want to save the trained model. # This placeholder will store the target labels for each mini batch labels = tf.placeholder(tf.float32, [None, 10]) # Define loss and optimizer cross_entropy = tf.reduce_mean( tf.nn.softmax_cross_entropy_with_logits(labels=labels, logits=self.output_layer)) # Define operations for testing correct_prediction = tf.equal(tf.argmax(self.output_layer, 1), tf.argmax(labels, 1)) accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32)) if self.use_batch_norm: # If we don't include the update ops as dependencies on the train step, the # tf.layers.batch_normalization layers won't update their population statistics, # which will cause the model to fail at inference time with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)): train_step = tf.train.GradientDescentOptimizer(learning_rate).minimize(cross_entropy) else: train_step = tf.train.GradientDescentOptimizer(learning_rate).minimize(cross_entropy) # Train for the appropriate number of batches. (tqdm is only for a nice timing display) for i in tqdm.tqdm(range(training_batches)): # We use batches of 60 just because the original paper did. You can use any size batch you like. batch_xs, batch_ys = mnist.train.next_batch(60) session.run(train_step, feed_dict={self.input_layer: batch_xs, labels: batch_ys, self.is_training: True}) # Periodically test accuracy against the 5k validation images and store it for plotting later. if i % batches_per_sample == 0: test_accuracy = session.run(accuracy, feed_dict={self.input_layer: mnist.validation.images, labels: mnist.validation.labels, self.is_training: False}) self.training_accuracies.append(test_accuracy) # After training, report accuracy against test data test_accuracy = session.run(accuracy, feed_dict={self.input_layer: mnist.validation.images, labels: mnist.validation.labels, self.is_training: False}) print('{}: After training, final accuracy on validation set = {}'.format(self.name, test_accuracy)) # If you want to use this model later for inference instead of having to retrain it, # just construct it with the same parameters and then pass this file to the 'test' function if save_model_as: tf.train.Saver().save(session, save_model_as) def test(self, session, test_training_accuracy=False, include_individual_predictions=False, restore_from=None): Trains a trained model on the MNIST testing dataset. :param session: Session Used to run the testing graph operations. :param test_training_accuracy: bool (default False) If True, perform inference with batch normalization using batch mean and variance; if False, perform inference with batch normalization using estimated population mean and variance. Note: in real life, *always* perform inference using the population mean and variance. This parameter exists just to support demonstrating what happens if you don't. :param include_individual_predictions: bool (default True) This function always performs an accuracy test against the entire test set. But if this parameter is True, it performs an extra test, doing 200 predictions one at a time, and displays the results and accuracy. :param restore_from: string or None (default None) Name of a saved model if you want to test with previously saved weights. # This placeholder will store the true labels for each mini batch labels = tf.placeholder(tf.float32, [None, 10]) # Define operations for testing correct_prediction = tf.equal(tf.argmax(self.output_layer, 1), tf.argmax(labels, 1)) accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32)) # If provided, restore from a previously saved model if restore_from: tf.train.Saver().restore(session, restore_from) # Test against all of the MNIST test data test_accuracy = session.run(accuracy, feed_dict={self.input_layer: mnist.test.images, labels: mnist.test.labels, self.is_training: test_training_accuracy}) print('-'*75) print('{}: Accuracy on full test set = {}'.format(self.name, test_accuracy)) # If requested, perform tests predicting individual values rather than batches if include_individual_predictions: predictions = [] correct = 0 # Do 200 predictions, 1 at a time for i in range(200): # This is a normal prediction using an individual test case. However, notice # we pass `test_training_accuracy` to `feed_dict` as the value for `self.is_training`. # Remember that will tell it whether it should use the batch mean & variance or # the population estimates that were calucated while training the model. pred, corr = session.run([tf.arg_max(self.output_layer,1), accuracy], feed_dict={self.input_layer: [mnist.test.images[i]], labels: [mnist.test.labels[i]], self.is_training: test_training_accuracy}) correct += corr predictions.append(pred[0]) print("200 Predictions:", predictions) print("Accuracy on 200 samples:", correct/200) Explanation: Neural network classes for testing The following class, NeuralNet, allows us to create identical neural networks with and without batch normalization. The code is heaviy documented, but there is also some additional discussion later. You do not need to read through it all before going through the rest of the notebook, but the comments within the code blocks may answer some of your questions. About the code: This class is not meant to represent TensorFlow best practices – the design choices made here are to support the discussion related to batch normalization. It's also important to note that we use the well-known MNIST data for these examples, but the networks we create are not meant to be good for performing handwritten character recognition. We chose this network architecture because it is similar to the one used in the original paper, which is complex enough to demonstrate some of the benefits of batch normalization while still being fast to train. End of explanation def plot_training_accuracies(*args, **kwargs): Displays a plot of the accuracies calculated during training to demonstrate how many iterations it took for the model(s) to converge. :param args: One or more NeuralNet objects You can supply any number of NeuralNet objects as unnamed arguments and this will display their training accuracies. Be sure to call `train` the NeuralNets before calling this function. :param kwargs: You can supply any named parameters here, but `batches_per_sample` is the only one we look for. It should match the `batches_per_sample` value you passed to the `train` function. fig, ax = plt.subplots() batches_per_sample = kwargs['batches_per_sample'] for nn in args: ax.plot(range(0,len(nn.training_accuracies)*batches_per_sample,batches_per_sample), nn.training_accuracies, label=nn.name) ax.set_xlabel('Training steps') ax.set_ylabel('Accuracy') ax.set_title('Validation Accuracy During Training') ax.legend(loc=4) ax.set_ylim([0,1]) plt.yticks(np.arange(0, 1.1, 0.1)) plt.grid(True) plt.show() def train_and_test(use_bad_weights, learning_rate, activation_fn, training_batches=50000, batches_per_sample=500): Creates two networks, one with and one without batch normalization, then trains them with identical starting weights, layers, batches, etc. Finally tests and plots their accuracies. :param use_bad_weights: bool If True, initialize the weights of both networks to wildly inappropriate weights; if False, use reasonable starting weights. :param learning_rate: float Learning rate used during gradient descent. :param activation_fn: Callable The function used for the output of each hidden layer. The network will use the same activation function on every hidden layer and no activate function on the output layer. e.g. Pass tf.nn.relu to use ReLU activations on your hidden layers. :param training_batches: (default 50000) Number of batches to train. :param batches_per_sample: (default 500) How many batches to train before sampling the validation accuracy. # Use identical starting weights for each network to eliminate differences in # weight initialization as a cause for differences seen in training performance # # Note: The networks will use these weights to define the number of and shapes of # its layers. The original batch normalization paper used 3 hidden layers # with 100 nodes in each, followed by a 10 node output layer. These values # build such a network, but feel free to experiment with different choices. # However, the input size should always be 784 and the final output should be 10. if use_bad_weights: # These weights should be horrible because they have such a large standard deviation weights = [np.random.normal(size=(784,100), scale=5.0).astype(np.float32), np.random.normal(size=(100,100), scale=5.0).astype(np.float32), np.random.normal(size=(100,100), scale=5.0).astype(np.float32), np.random.normal(size=(100,10), scale=5.0).astype(np.float32) ] else: # These weights should be good because they have such a small standard deviation weights = [np.random.normal(size=(784,100), scale=0.05).astype(np.float32), np.random.normal(size=(100,100), scale=0.05).astype(np.float32), np.random.normal(size=(100,100), scale=0.05).astype(np.float32), np.random.normal(size=(100,10), scale=0.05).astype(np.float32) ] # Just to make sure the TensorFlow's default graph is empty before we start another # test, because we don't bother using different graphs or scoping and naming # elements carefully in this sample code. tf.reset_default_graph() # build two versions of same network, 1 without and 1 with batch normalization nn = NeuralNet(weights, activation_fn, False) bn = NeuralNet(weights, activation_fn, True) # train and test the two models with tf.Session() as sess: tf.global_variables_initializer().run() nn.train(sess, learning_rate, training_batches, batches_per_sample) bn.train(sess, learning_rate, training_batches, batches_per_sample) nn.test(sess) bn.test(sess) # Display a graph of how validation accuracies changed during training # so we can compare how the models trained and when they converged plot_training_accuracies(nn, bn, batches_per_sample=batches_per_sample) Explanation: There are quite a few comments in the code, so those should answer most of your questions. However, let's take a look at the most important lines. We add batch normalization to layers inside the fully_connected function. Here are some important points about that code: 1. Layers with batch normalization do not include a bias term. 2. We use TensorFlow's tf.layers.batch_normalization function to handle the math. (We show lower-level ways to do this later in the notebook.) 3. We tell tf.layers.batch_normalization whether or not the network is training. This is an important step we'll talk about later. 4. We add the normalization before calling the activation function. In addition to that code, the training step is wrapped in the following with statement: python with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)): This line actually works in conjunction with the training parameter we pass to tf.layers.batch_normalization. Without it, TensorFlow's batch normalization layer will not operate correctly during inference. Finally, whenever we train the network or perform inference, we use the feed_dict to set self.is_training to True or False, respectively, like in the following line: python session.run(train_step, feed_dict={self.input_layer: batch_xs, labels: batch_ys, self.is_training: True}) We'll go into more details later, but next we want to show some experiments that use this code and test networks with and without batch normalization. Batch Normalization Demos<a id='demos'></a> This section of the notebook trains various networks with and without batch normalization to demonstrate some of the benefits mentioned earlier. We'd like to thank the author of this blog post Implementing Batch Normalization in TensorFlow. That post provided the idea of - and some of the code for - plotting the differences in accuracy during training, along with the idea for comparing multiple networks using the same initial weights. Code to support testing The following two functions support the demos we run in the notebook. The first function, plot_training_accuracies, simply plots the values found in the training_accuracies lists of the NeuralNet objects passed to it. If you look at the train function in NeuralNet, you'll see it that while it's training the network, it periodically measures validation accuracy and stores the results in that list. It does that just to support these plots. The second function, train_and_test, creates two neural nets - one with and one without batch normalization. It then trains them both and tests them, calling plot_training_accuracies to plot how their accuracies changed over the course of training. The really imporant thing about this function is that it initializes the starting weights for the networks outside of the networks and then passes them in. This lets it train both networks from the exact same starting weights, which eliminates performance differences that might result from (un)lucky initial weights. End of explanation train_and_test(False, 0.01, tf.nn.relu) Explanation: Comparisons between identical networks, with and without batch normalization The next series of cells train networks with various settings to show the differences with and without batch normalization. They are meant to clearly demonstrate the effects of batch normalization. We include a deeper discussion of batch normalization later in the notebook. The following creates two networks using a ReLU activation function, a learning rate of 0.01, and reasonable starting weights. End of explanation train_and_test(False, 0.01, tf.nn.relu, 2000, 50) Explanation: As expected, both networks train well and eventually reach similar test accuracies. However, notice that the model with batch normalization converges slightly faster than the other network, reaching accuracies over 90% almost immediately and nearing its max acuracy in 10 or 15 thousand iterations. The other network takes about 3 thousand iterations to reach 90% and doesn't near its best accuracy until 30 thousand or more iterations. If you look at the raw speed, you can see that without batch normalization we were computing over 1100 batches per second, whereas with batch normalization that goes down to just over 500. However, batch normalization allows us to perform fewer iterations and converge in less time over all. (We only trained for 50 thousand batches here so we could plot the comparison.) The following creates two networks with the same hyperparameters used in the previous example, but only trains for 2000 iterations. End of explanation train_and_test(False, 0.01, tf.nn.sigmoid) Explanation: As you can see, using batch normalization produces a model with over 95% accuracy in only 2000 batches, and it was above 90% at somewhere around 500 batches. Without batch normalization, the model takes 1750 iterations just to hit 80% – the network with batch normalization hits that mark after around 200 iterations! (Note: if you run the code yourself, you'll see slightly different results each time because the starting weights - while the same for each model - are different for each run.) In the above example, you should also notice that the networks trained fewer batches per second then what you saw in the previous example. That's because much of the time we're tracking is actually spent periodically performing inference to collect data for the plots. In this example we perform that inference every 50 batches instead of every 500, so generating the plot for this example requires 10 times the overhead for the same 2000 iterations. The following creates two networks using a sigmoid activation function, a learning rate of 0.01, and reasonable starting weights. End of explanation train_and_test(False, 1, tf.nn.relu) Explanation: With the number of layers we're using and this small learning rate, using a sigmoid activation function takes a long time to start learning. It eventually starts making progress, but it took over 45 thousand batches just to get over 80% accuracy. Using batch normalization gets to 90% in around one thousand batches. The following creates two networks using a ReLU activation function, a learning rate of 1, and reasonable starting weights. End of explanation train_and_test(False, 1, tf.nn.relu) Explanation: Now we're using ReLUs again, but with a larger learning rate. The plot shows how training started out pretty normally, with the network with batch normalization starting out faster than the other. But the higher learning rate bounces the accuracy around a bit more, and at some point the accuracy in the network without batch normalization just completely crashes. It's likely that too many ReLUs died off at this point because of the high learning rate. The next cell shows the same test again. The network with batch normalization performs the same way, and the other suffers from the same problem again, but it manages to train longer before it happens. End of explanation train_and_test(False, 1, tf.nn.sigmoid) Explanation: In both of the previous examples, the network with batch normalization manages to gets over 98% accuracy, and get near that result almost immediately. The higher learning rate allows the network to train extremely fast. The following creates two networks using a sigmoid activation function, a learning rate of 1, and reasonable starting weights. End of explanation train_and_test(False, 1, tf.nn.sigmoid, 2000, 50) Explanation: In this example, we switched to a sigmoid activation function. It appears to hande the higher learning rate well, with both networks achieving high accuracy. The cell below shows a similar pair of networks trained for only 2000 iterations. End of explanation train_and_test(False, 2, tf.nn.relu) Explanation: As you can see, even though these parameters work well for both networks, the one with batch normalization gets over 90% in 400 or so batches, whereas the other takes over 1700. When training larger networks, these sorts of differences become more pronounced. The following creates two networks using a ReLU activation function, a learning rate of 2, and reasonable starting weights. End of explanation train_and_test(False, 2, tf.nn.sigmoid) Explanation: With this very large learning rate, the network with batch normalization trains fine and almost immediately manages 98% accuracy. However, the network without normalization doesn't learn at all. The following creates two networks using a sigmoid activation function, a learning rate of 2, and reasonable starting weights. End of explanation train_and_test(False, 2, tf.nn.sigmoid, 2000, 50) Explanation: Once again, using a sigmoid activation function with the larger learning rate works well both with and without batch normalization. However, look at the plot below where we train models with the same parameters but only 2000 iterations. As usual, batch normalization lets it train faster. End of explanation train_and_test(True, 0.01, tf.nn.relu) Explanation: In the rest of the examples, we use really bad starting weights. That is, normally we would use very small values close to zero. However, in these examples we choose randome values with a standard deviation of 5. If you were really training a neural network, you would not want to do this. But these examples demonstrate how batch normalization makes your network much more resilient. The following creates two networks using a ReLU activation function, a learning rate of 0.01, and bad starting weights. End of explanation train_and_test(True, 0.01, tf.nn.sigmoid) Explanation: As the plot shows, without batch normalization the network never learns anything at all. But with batch normalization, it actually learns pretty well and gets to almost 80% accuracy. The starting weights obviously hurt the network, but you can see how well batch normalization does in overcoming them. The following creates two networks using a sigmoid activation function, a learning rate of 0.01, and bad starting weights. End of explanation train_and_test(True, 1, tf.nn.relu) Explanation: Using a sigmoid activation function works better than the ReLU in the previous example, but without batch normalization it would take a tremendously long time to train the network, if it ever trained at all. The following creates two networks using a ReLU activation function, a learning rate of 1, and bad starting weights.<a id="successful_example_lr_1"></a> End of explanation train_and_test(True, 1, tf.nn.sigmoid) Explanation: The higher learning rate used here allows the network with batch normalization to surpass 90% in about 30 thousand batches. The network without it never gets anywhere. The following creates two networks using a sigmoid activation function, a learning rate of 1, and bad starting weights. End of explanation train_and_test(True, 2, tf.nn.relu) Explanation: Using sigmoid works better than ReLUs for this higher learning rate. However, you can see that without batch normalization, the network takes a long time tro train, bounces around a lot, and spends a long time stuck at 90%. The network with batch normalization trains much more quickly, seems to be more stable, and achieves a higher accuracy. The following creates two networks using a ReLU activation function, a learning rate of 2, and bad starting weights.<a id="successful_example_lr_2"></a> End of explanation train_and_test(True, 2, tf.nn.sigmoid) Explanation: We've already seen that ReLUs do not do as well as sigmoids with higher learning rates, and here we are using an extremely high rate. As expected, without batch normalization the network doesn't learn at all. But with batch normalization, it eventually achieves 90% accuracy. Notice, though, how its accuracy bounces around wildly during training - that's because the learning rate is really much too high, so the fact that this worked at all is a bit of luck. The following creates two networks using a sigmoid activation function, a learning rate of 2, and bad starting weights. End of explanation train_and_test(True, 1, tf.nn.relu) Explanation: In this case, the network with batch normalization trained faster and reached a higher accuracy. Meanwhile, the high learning rate makes the network without normalization bounce around erratically and have trouble getting past 90%. Full Disclosure: Batch Normalization Doesn't Fix Everything Batch normalization isn't magic and it doesn't work every time. Weights are still randomly initialized and batches are chosen at random during training, so you never know exactly how training will go. Even for these tests, where we use the same initial weights for both networks, we still get different weights each time we run. This section includes two examples that show runs when batch normalization did not help at all. The following creates two networks using a ReLU activation function, a learning rate of 1, and bad starting weights. End of explanation train_and_test(True, 2, tf.nn.relu) Explanation: When we used these same parameters earlier, we saw the network with batch normalization reach 92% validation accuracy. This time we used different starting weights, initialized using the same standard deviation as before, and the network doesn't learn at all. (Remember, an accuracy around 10% is what the network gets if it just guesses the same value all the time.) The following creates two networks using a ReLU activation function, a learning rate of 2, and bad starting weights. End of explanation def fully_connected(self, layer_in, initial_weights, activation_fn=None): Creates a standard, fully connected layer. Its number of inputs and outputs will be defined by the shape of `initial_weights`, and its starting weight values will be taken directly from that same parameter. If `self.use_batch_norm` is True, this layer will include batch normalization, otherwise it will not. :param layer_in: Tensor The Tensor that feeds into this layer. It's either the input to the network or the output of a previous layer. :param initial_weights: NumPy array or Tensor Initial values for this layer's weights. The shape defines the number of nodes in the layer. e.g. Passing in 3 matrix of shape (784, 256) would create a layer with 784 inputs and 256 outputs. :param activation_fn: Callable or None (default None) The non-linearity used for the output of the layer. If None, this layer will not include batch normalization, regardless of the value of `self.use_batch_norm`. e.g. Pass tf.nn.relu to use ReLU activations on your hidden layers. if self.use_batch_norm and activation_fn: # Batch normalization uses weights as usual, but does NOT add a bias term. This is because # its calculations include gamma and beta variables that make the bias term unnecessary. weights = tf.Variable(initial_weights) linear_output = tf.matmul(layer_in, weights) num_out_nodes = initial_weights.shape[-1] # Batch normalization adds additional trainable variables: # gamma (for scaling) and beta (for shifting). gamma = tf.Variable(tf.ones([num_out_nodes])) beta = tf.Variable(tf.zeros([num_out_nodes])) # These variables will store the mean and variance for this layer over the entire training set, # which we assume represents the general population distribution. # By setting `trainable=False`, we tell TensorFlow not to modify these variables during # back propagation. Instead, we will assign values to these variables ourselves. pop_mean = tf.Variable(tf.zeros([num_out_nodes]), trainable=False) pop_variance = tf.Variable(tf.ones([num_out_nodes]), trainable=False) # Batch normalization requires a small constant epsilon, used to ensure we don't divide by zero. # This is the default value TensorFlow uses. epsilon = 1e-3 def batch_norm_training(): # Calculate the mean and variance for the data coming out of this layer's linear-combination step. # The [0] defines an array of axes to calculate over. batch_mean, batch_variance = tf.nn.moments(linear_output, [0]) # Calculate a moving average of the training data's mean and variance while training. # These will be used during inference. # Decay should be some number less than 1. tf.layers.batch_normalization uses the parameter # "momentum" to accomplish this and defaults it to 0.99 decay = 0.99 train_mean = tf.assign(pop_mean, pop_mean * decay + batch_mean * (1 - decay)) train_variance = tf.assign(pop_variance, pop_variance * decay + batch_variance * (1 - decay)) # The 'tf.control_dependencies' context tells TensorFlow it must calculate 'train_mean' # and 'train_variance' before it calculates the 'tf.nn.batch_normalization' layer. # This is necessary because the those two operations are not actually in the graph # connecting the linear_output and batch_normalization layers, # so TensorFlow would otherwise just skip them. with tf.control_dependencies([train_mean, train_variance]): return tf.nn.batch_normalization(linear_output, batch_mean, batch_variance, beta, gamma, epsilon) def batch_norm_inference(): # During inference, use the our estimated population mean and variance to normalize the layer return tf.nn.batch_normalization(linear_output, pop_mean, pop_variance, beta, gamma, epsilon) # Use `tf.cond` as a sort of if-check. When self.is_training is True, TensorFlow will execute # the operation returned from `batch_norm_training`; otherwise it will execute the graph # operation returned from `batch_norm_inference`. batch_normalized_output = tf.cond(self.is_training, batch_norm_training, batch_norm_inference) # Pass the batch-normalized layer output through the activation function. # The literature states there may be cases where you want to perform the batch normalization *after* # the activation function, but it is difficult to find any uses of that in practice. return activation_fn(batch_normalized_output) else: # When not using batch normalization, create a standard layer that multiplies # the inputs and weights, adds a bias, and optionally passes the result # through an activation function. weights = tf.Variable(initial_weights) biases = tf.Variable(tf.zeros([initial_weights.shape[-1]])) linear_output = tf.add(tf.matmul(layer_in, weights), biases) return linear_output if not activation_fn else activation_fn(linear_output) Explanation: When we trained with these parameters and batch normalization earlier, we reached 90% validation accuracy. However, this time the network almost starts to make some progress in the beginning, but it quickly breaks down and stops learning. Note: Both of the above examples use extremely bad starting weights, along with learning rates that are too high. While we've shown batch normalization can overcome bad values, we don't mean to encourage actually using them. The examples in this notebook are meant to show that batch normalization can help your networks train better. But these last two examples should remind you that you still want to try to use good network design choices and reasonable starting weights. It should also remind you that the results of each attempt to train a network are a bit random, even when using otherwise identical architectures. Batch Normalization: A Detailed Look<a id='implementation_2'></a> The layer created by tf.layers.batch_normalization handles all the details of implementing batch normalization. Many students will be fine just using that and won't care about what's happening at the lower levels. However, some students may want to explore the details, so here is a short explanation of what's really happening, starting with the equations you're likely to come across if you ever read about batch normalization. In order to normalize the values, we first need to find the average value for the batch. If you look at the code, you can see that this is not the average value of the batch inputs, but the average value coming out of any particular layer before we pass it through its non-linear activation function and then feed it as an input to the next layer. We represent the average as $\mu_B$, which is simply the sum of all of the values $x_i$ divided by the number of values, $m$ $$ \mu_B \leftarrow \frac{1}{m}\sum_{i=1}^m x_i $$ We then need to calculate the variance, or mean squared deviation, represented as $\sigma_{B}^{2}$. If you aren't familiar with statistics, that simply means for each value $x_i$, we subtract the average value (calculated earlier as $\mu_B$), which gives us what's called the "deviation" for that value. We square the result to get the squared deviation. Sum up the results of doing that for each of the values, then divide by the number of values, again $m$, to get the average, or mean, squared deviation. $$ \sigma_{B}^{2} \leftarrow \frac{1}{m}\sum_{i=1}^m (x_i - \mu_B)^2 $$ Once we have the mean and variance, we can use them to normalize the values with the following equation. For each value, it subtracts the mean and divides by the (almost) standard deviation. (You've probably heard of standard deviation many times, but if you have not studied statistics you might not know that the standard deviation is actually the square root of the mean squared deviation.) $$ \hat{x_i} \leftarrow \frac{x_i - \mu_B}{\sqrt{\sigma_{B}^{2} + \epsilon}} $$ Above, we said "(almost) standard deviation". That's because the real standard deviation for the batch is calculated by $\sqrt{\sigma_{B}^{2}}$, but the above formula adds the term epsilon, $\epsilon$, before taking the square root. The epsilon can be any small, positive constant - in our code we use the value 0.001. It is there partially to make sure we don't try to divide by zero, but it also acts to increase the variance slightly for each batch. Why increase the variance? Statistically, this makes sense because even though we are normalizing one batch at a time, we are also trying to estimate the population distribution – the total training set, which itself an estimate of the larger population of inputs your network wants to handle. The variance of a population is higher than the variance for any sample taken from that population, so increasing the variance a little bit for each batch helps take that into account. At this point, we have a normalized value, represented as $\hat{x_i}$. But rather than use it directly, we multiply it by a gamma value, $\gamma$, and then add a beta value, $\beta$. Both $\gamma$ and $\beta$ are learnable parameters of the network and serve to scale and shift the normalized value, respectively. Because they are learnable just like weights, they give your network some extra knobs to tweak during training to help it learn the function it is trying to approximate. $$ y_i \leftarrow \gamma \hat{x_i} + \beta $$ We now have the final batch-normalized output of our layer, which we would then pass to a non-linear activation function like sigmoid, tanh, ReLU, Leaky ReLU, etc. In the original batch normalization paper (linked in the beginning of this notebook), they mention that there might be cases when you'd want to perform the batch normalization after the non-linearity instead of before, but it is difficult to find any uses like that in practice. In NeuralNet's implementation of fully_connected, all of this math is hidden inside the following line, where linear_output serves as the $x_i$ from the equations: python batch_normalized_output = tf.layers.batch_normalization(linear_output, training=self.is_training) The next section shows you how to implement the math directly. Batch normalization without the tf.layers package Our implementation of batch normalization in NeuralNet uses the high-level abstraction tf.layers.batch_normalization, found in TensorFlow's tf.layers package. However, if you would like to implement batch normalization at a lower level, the following code shows you how. It uses tf.nn.batch_normalization from TensorFlow's neural net (nn) package. 1) You can replace the fully_connected function in the NeuralNet class with the below code and everything in NeuralNet will still work like it did before. End of explanation def batch_norm_test(test_training_accuracy): :param test_training_accuracy: bool If True, perform inference with batch normalization using batch mean and variance; if False, perform inference with batch normalization using estimated population mean and variance. weights = [np.random.normal(size=(784,100), scale=0.05).astype(np.float32), np.random.normal(size=(100,100), scale=0.05).astype(np.float32), np.random.normal(size=(100,100), scale=0.05).astype(np.float32), np.random.normal(size=(100,10), scale=0.05).astype(np.float32) ] tf.reset_default_graph() # Train the model bn = NeuralNet(weights, tf.nn.relu, True) # First train the network with tf.Session() as sess: tf.global_variables_initializer().run() bn.train(sess, 0.01, 2000, 2000) bn.test(sess, test_training_accuracy=test_training_accuracy, include_individual_predictions=True) Explanation: This version of fully_connected is much longer than the original, but once again has extensive comments to help you understand it. Here are some important points: It explicitly creates variables to store gamma, beta, and the population mean and variance. These were all handled for us in the previous version of the function. It initializes gamma to one and beta to zero, so they start out having no effect in this calculation: $y_i \leftarrow \gamma \hat{x_i} + \beta$. However, during training the network learns the best values for these variables using back propagation, just like networks normally do with weights. Unlike gamma and beta, the variables for population mean and variance are marked as untrainable. That tells TensorFlow not to modify them during back propagation. Instead, the lines that call tf.assign are used to update these variables directly. TensorFlow won't automatically run the tf.assign operations during training because it only evaluates operations that are required based on the connections it finds in the graph. To get around that, we add this line: with tf.control_dependencies([train_mean, train_variance]): before we run the normalization operation. This tells TensorFlow it needs to run those operations before running anything inside the with block. The actual normalization math is still mostly hidden from us, this time using tf.nn.batch_normalization. tf.nn.batch_normalization does not have a training parameter like tf.layers.batch_normalization did. However, we still need to handle training and inference differently, so we run different code in each case using the tf.cond operation. We use the tf.nn.moments function to calculate the batch mean and variance. 2) The current version of the train function in NeuralNet will work fine with this new version of fully_connected. However, it uses these lines to ensure population statistics are updated when using batch normalization: python if self.use_batch_norm: with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)): train_step = tf.train.GradientDescentOptimizer(learning_rate).minimize(cross_entropy) else: train_step = tf.train.GradientDescentOptimizer(learning_rate).minimize(cross_entropy) Our new version of fully_connected handles updating the population statistics directly. That means you can also simplify your code by replacing the above if/else condition with just this line: python train_step = tf.train.GradientDescentOptimizer(learning_rate).minimize(cross_entropy) 3) And just in case you want to implement every detail from scratch, you can replace this line in batch_norm_training: python return tf.nn.batch_normalization(linear_output, batch_mean, batch_variance, beta, gamma, epsilon) with these lines: python normalized_linear_output = (linear_output - batch_mean) / tf.sqrt(batch_variance + epsilon) return gamma * normalized_linear_output + beta And replace this line in batch_norm_inference: python return tf.nn.batch_normalization(linear_output, pop_mean, pop_variance, beta, gamma, epsilon) with these lines: python normalized_linear_output = (linear_output - pop_mean) / tf.sqrt(pop_variance + epsilon) return gamma * normalized_linear_output + beta As you can see in each of the above substitutions, the two lines of replacement code simply implement the following two equations directly. The first line calculates the following equation, with linear_output representing $x_i$ and normalized_linear_output representing $\hat{x_i}$: $$ \hat{x_i} \leftarrow \frac{x_i - \mu_B}{\sqrt{\sigma_{B}^{2} + \epsilon}} $$ And the second line is a direct translation of the following equation: $$ y_i \leftarrow \gamma \hat{x_i} + \beta $$ We still use the tf.nn.moments operation to implement the other two equations from earlier – the ones that calculate the batch mean and variance used in the normalization step. If you really wanted to do everything from scratch, you could replace that line, too, but we'll leave that to you. Why the difference between training and inference? In the original function that uses tf.layers.batch_normalization, we tell the layer whether or not the network is training by passing a value for its training parameter, like so: python batch_normalized_output = tf.layers.batch_normalization(linear_output, training=self.is_training) And that forces us to provide a value for self.is_training in our feed_dict, like we do in this example from NeuralNet's train function: python session.run(train_step, feed_dict={self.input_layer: batch_xs, labels: batch_ys, self.is_training: True}) If you looked at the low level implementation, you probably noticed that, just like with tf.layers.batch_normalization, we need to do slightly different things during training and inference. But why is that? First, let's look at what happens when we don't. The following function is similar to train_and_test from earlier, but this time we are only testing one network and instead of plotting its accuracy, we perform 200 predictions on test inputs, 1 input at at time. We can use the test_training_accuracy parameter to test the network in training or inference modes (the equivalent of passing True or False to the feed_dict for is_training). End of explanation batch_norm_test(True) Explanation: In the following cell, we pass True for test_training_accuracy, which performs the same batch normalization that we normally perform during training. End of explanation batch_norm_test(False) Explanation: As you can see, the network guessed the same value every time! But why? Because during training, a network with batch normalization adjusts the values at each layer based on the mean and variance of that batch. The "batches" we are using for these predictions have a single input each time, so their values are the means, and their variances will always be 0. That means the network will normalize the values at any layer to zero. (Review the equations from before to see why a value that is equal to the mean would always normalize to zero.) So we end up with the same result for every input we give the network, because its the value the network produces when it applies its learned weights to zeros at every layer. Note: If you re-run that cell, you might get a different value from what we showed. That's because the specific weights the network learns will be different every time. But whatever value it is, it should be the same for all 200 predictions. To overcome this problem, the network does not just normalize the batch at each layer. It also maintains an estimate of the mean and variance for the entire population. So when we perform inference, instead of letting it "normalize" all the values using their own means and variance, it uses the estimates of the population mean and variance that it calculated while training. So in the following example, we pass False for test_training_accuracy, which tells the network that we it want to perform inference with the population statistics it calculates during training. End of explanation <END_TASK>
15,906
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Training DeepMind's Atari DQN with Chimp Load Chimp modules Step1: Load Python packages Step2: Set training parameters Step3: You may want to set a smaller number of iterations (like 100000) - for illustration purposes. We set the GPU option to True, turn it off if your machine does not support it. Be sure to have the requested rom in the indicated directory. Step4: Now we initialize the simulator first, as we need to use some information it provides - e.g., number of actions. Step5: Here we define the convolutional network, in a format required by Chainer - the deep learning library we use. Step6: We then initialize the learner + chainer backend, replay memory, and agent modules. Step7: Now let the agent train. Step8: Visualizing results First, let's visualize the training and evaluation results. Step9: Evaluating the best policy Let's load the network that collected the highest reward per game episode
<ASSISTANT_TASK:> Python Code: from chimp.memories import ReplayMemoryHDF5 from chimp.learners.dqn_learner import DQNLearner from chimp.learners.chainer_backend import ChainerBackend from chimp.simulators.atari import AtariSimulator from chimp.agents import DQNAgent Explanation: Training DeepMind's Atari DQN with Chimp Load Chimp modules End of explanation %matplotlib inline import matplotlib.pyplot as plt import numpy as np import random import chainer import chainer.functions as F import chainer.links as L from chainer import Chain import os import pandas as ps Explanation: Load Python packages End of explanation settings = { # agent settings 'batch_size' : 32, 'print_every' : 5000, 'save_dir' : './results_atari', 'iterations' : 5000000, 'eval_iterations' : 5000, 'eval_every' : 50000, 'save_every' : 50000, 'initial_exploration' : 50000, 'epsilon_decay' : 0.000005, # subtract from epsilon every step 'eval_epsilon' : 0.05, # epsilon used in evaluation, 0 means no random actions 'epsilon' : 1.0, # Initial exploratoin rate 'learn_freq' : 4, 'history_sizes' : (4, 0, 0), # sizes of histories to use as nn inputs (o, a, r) 'model_dims' : (84,84), # Atari settings 'rom' : "Breakout.bin", 'rom_dir' : './roms', 'pad' : 15, # padding parameter - for image cropping - only along the length of the image, to obtain a square 'action_history' : True, # simulator settings 'viz' : True, 'viz_cropped' : False, # replay memory settings 'memory_size' : 1000000, # size of replay memory 'frame_skip' : 4, # number of frames to skip # learner settings 'learning_rate' : 0.00025, 'decay_rate' : 0.95, # decay rate for RMSprop, otherwise not used 'discount' : 0.99, # discount rate for RL 'clip_err' : False, # value to clip loss gradients to 'clip_reward' : 1, # value to clip reward values to 'target_net_update' : 10000, # update the update-generating target net every fixed number of iterations 'optim_name' : 'RMSprop', # currently supports "RMSprop", "ADADELTA", "ADAM" and "SGD"' 'gpu' : True, 'reward_rescale': False, # general 'seed_general' : 1723, 'seed_simulator' : 5632, 'seed_agent' : 9826, 'seed_memory' : 7563 } Explanation: Set training parameters End of explanation # set random seed np.random.seed(settings["seed_general"]) random.seed(settings["seed_general"]) Explanation: You may want to set a smaller number of iterations (like 100000) - for illustration purposes. We set the GPU option to True, turn it off if your machine does not support it. Be sure to have the requested rom in the indicated directory. End of explanation simulator = AtariSimulator(settings) Explanation: Now we initialize the simulator first, as we need to use some information it provides - e.g., number of actions. End of explanation #Define the network class Convolution(Chain): def __init__(self): super(Convolution, self).__init__( l1=F.Convolution2D(settings['history_sizes'][0], 32, ksize=8, stride=4, nobias=False, wscale=np.sqrt(2)), l2=F.Convolution2D(32, 64, ksize=4, stride=2, nobias=False, wscale=np.sqrt(2)), l3=F.Convolution2D(64, 64, ksize=3, stride=1, nobias=False, wscale=np.sqrt(2)), l4=F.Linear(3136, 512, wscale = np.sqrt(2)), l5=F.Linear(512, simulator.n_actions, wscale = np.sqrt(2)), ) def __call__(self, ohist, ahist): if len(ohist.data.shape) < 4: ohist = F.reshape(ohist,(1,4,84,84)) h1 = F.relu(self.l1(ohist/255.0)) h2 = F.relu(self.l2(h1)) h3 = F.relu(self.l3(h2)) h4 = F.relu(self.l4(h3)) output = self.l5(h4) return output net = Convolution() Explanation: Here we define the convolutional network, in a format required by Chainer - the deep learning library we use. End of explanation backend = ChainerBackend(settings) backend.set_net(net) learner = DQNLearner(settings, backend) memory = ReplayMemoryHDF5(settings) agent = DQNAgent(learner, memory, simulator, settings) Explanation: We then initialize the learner + chainer backend, replay memory, and agent modules. End of explanation agent.train() Explanation: Now let the agent train. End of explanation train_stats = ps.read_csv('%s/training_history.csv' % settings['save_dir'],delimiter=' ',header=None) train_stats.columns = ['Iteration','MSE Loss','Average Q-Value'] eval_stats = ps.read_csv('%s/evaluation_history.csv' % settings['save_dir'],delimiter=' ',header=None) eval_stats.columns = ['Iteration','Total Reward','Reward per Episode'] plt.plot(eval_stats['Iteration'], eval_stats['Reward per Episode']) plt.xlabel("Iteration") plt.ylabel("Avg. Reward per Episode") plt.grid(True) #plt.savefig(settings['save_dir'] + '_' + "evaluation_reward.svg", bbox_inches='tight') plt.show() plt.close() plt.plot(train_stats['Iteration'], train_stats['Average Q-Value']) plt.xlabel("Iteration") plt.ylabel("Avg. Q-Values") plt.grid(True) #plt.savefig(settings['save_dir'] + '_' + "training_q_values.svg", bbox_inches='tight') plt.show() plt.close() Explanation: Visualizing results First, let's visualize the training and evaluation results. End of explanation best_iteration_index = np.argmax(eval_stats['Reward per Episode']) best_iteration = str(int(eval_stats['Iteration'][best_iteration_index])) best_iteration agent.learner.load_net(settings['save_dir']+'/net_' + best_iteration + '.p') r_tot, r_per_episode, runtime = agent.simulate(10000, epsilon=0.05, viz=True) r_per_episode Explanation: Evaluating the best policy Let's load the network that collected the highest reward per game episode End of explanation <END_TASK>
15,907
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Copyright 2020 Google LLC. Licensed under the Apache License, Version 2.0 (the 'License'); you may not use this file except in compliance with the License. You may obtain a copy of the License at https Step1: Pairwise Regression Fairness We will be training a linear scoring function $f(x) = w^\top x$ where $x \in \mathbb{R}^d$ is the input feature vector. Our goal is to train the regression model subject to pairwise fairness constraints. Specifically, for the regression model $f$, we denote Step2: Evaluation Metrics We will need functions to convert labeled data into paired data. Step3: We will also need functions to evaluate the pairwise error rates for a linear model. Step4: Create Linear Model We then write a function to create the linear scoring model. Step5: Formulate Optimization Problem We are ready to formulate the constrained optimization problem using the TFCO library. Step6: Train Model The following function then trains the linear model by solving the above constrained optimization problem. We first provide a training function with minibatch gradient updates. There are three types of pairwise fairness criterion we handle (specified by 'constraint_type'), and assign the (pos_group, neg_group) pairs to compare accordingly. Step7: Summarize and Plot Results Having trained a model, we will need functions to summarize the various evaluation metrics. Step8: Experimental Results We now run experiments with two types of pairwise fairness criteria Step9: (1) Marginal Equal Opportunity For a scoring model $f Step10: (2) Pairwise Equal Opportunity Recall that we denote $err_{i,j}(f)$ as the pairwise error over example pairs where the higher label example is from group $i$, and the lower label example is from group $j$. $$ err_{i, j}(f) ~=~ \mathbf{E}\big[\mathbb{I}\big(f(x) < f(x')\big) \,\big|\, y > y',~ grp(x) = i, ~grp(x') = j\big] $$ We first constrain only the cross-group errors, highlighted below. <br> <table border='1' bordercolor='black'> <tr > <td bgcolor='white'> </td> <td bgcolor='white'> </td> <td bgcolor='white' colspan=2 align=center><b>Negative</b></td> </tr> <tr> <td bgcolor='white'></td> <td bgcolor='white'></td> <td>Group 0</td> <td>Group 1</td> </tr> <tr> <td bgcolor='white' rowspan=2><b>Positive</b></td> <td bgcolor='white'>Group 0</td> <td bgcolor='white'>$err_{0,0}$</td> <td bgcolor='white'>$\mathbf{err_{0,1}}$</td> </tr> <tr> <td>Group 1</td> <td bgcolor='white'>$\mathbf{err_{1,0}}$</td> <td bgcolor='white'>$err_{1,1}$</td> </tr> </table> <br> The optimization problem we solve constraints the cross-group pairwise errors to be similar
<ASSISTANT_TASK:> Python Code: import matplotlib.pyplot as plt import numpy as np import pandas as pd import random import sys from sklearn import model_selection import tensorflow as tf !pip install git+https://github.com/google-research/tensorflow_constrained_optimization import tensorflow_constrained_optimization as tfco Explanation: Copyright 2020 Google LLC. Licensed under the Apache License, Version 2.0 (the 'License'); you may not use this file except in compliance with the License. You may obtain a copy of the License at https://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an 'AS IS' BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. This colab contains TensorFlow code for implementing the constrained optimization methods presented in the paper: Harikrishna Narasimhan, Andrew Cotter, Maya Gupta, Serena Wang, 'Pairwise Fairness for Ranking and Regression', AAAI 2020. [<a href='https://arxiv.org/pdf/1906.05330.pdf'>link</a>] First, let's install and import the relevant libraries. End of explanation # We will divide the data into 25 minibatches and refer to them as 'queries'. num_queries = 25 # List of column names in the dataset. column_names = ["state", "county", "community", "communityname", "fold", "population", "householdsize", "racepctblack", "racePctWhite", "racePctAsian", "racePctHisp", "agePct12t21", "agePct12t29", "agePct16t24", "agePct65up", "numbUrban", "pctUrban", "medIncome", "pctWWage", "pctWFarmSelf", "pctWInvInc", "pctWSocSec", "pctWPubAsst", "pctWRetire", "medFamInc", "perCapInc", "whitePerCap", "blackPerCap", "indianPerCap", "AsianPerCap", "OtherPerCap", "HispPerCap", "NumUnderPov", "PctPopUnderPov", "PctLess9thGrade", "PctNotHSGrad", "PctBSorMore", "PctUnemployed", "PctEmploy", "PctEmplManu", "PctEmplProfServ", "PctOccupManu", "PctOccupMgmtProf", "MalePctDivorce", "MalePctNevMarr", "FemalePctDiv", "TotalPctDiv", "PersPerFam", "PctFam2Par", "PctKids2Par", "PctYoungKids2Par", "PctTeen2Par", "PctWorkMomYoungKids", "PctWorkMom", "NumIlleg", "PctIlleg", "NumImmig", "PctImmigRecent", "PctImmigRec5", "PctImmigRec8", "PctImmigRec10", "PctRecentImmig", "PctRecImmig5", "PctRecImmig8", "PctRecImmig10", "PctSpeakEnglOnly", "PctNotSpeakEnglWell", "PctLargHouseFam", "PctLargHouseOccup", "PersPerOccupHous", "PersPerOwnOccHous", "PersPerRentOccHous", "PctPersOwnOccup", "PctPersDenseHous", "PctHousLess3BR", "MedNumBR", "HousVacant", "PctHousOccup", "PctHousOwnOcc", "PctVacantBoarded", "PctVacMore6Mos", "MedYrHousBuilt", "PctHousNoPhone", "PctWOFullPlumb", "OwnOccLowQuart", "OwnOccMedVal", "OwnOccHiQuart", "RentLowQ", "RentMedian", "RentHighQ", "MedRent", "MedRentPctHousInc", "MedOwnCostPctInc", "MedOwnCostPctIncNoMtg", "NumInShelters", "NumStreet", "PctForeignBorn", "PctBornSameState", "PctSameHouse85", "PctSameCity85", "PctSameState85", "LemasSwornFT", "LemasSwFTPerPop", "LemasSwFTFieldOps", "LemasSwFTFieldPerPop", "LemasTotalReq", "LemasTotReqPerPop", "PolicReqPerOffic", "PolicPerPop", "RacialMatchCommPol", "PctPolicWhite", "PctPolicBlack", "PctPolicHisp", "PctPolicAsian", "PctPolicMinor", "OfficAssgnDrugUnits", "NumKindsDrugsSeiz", "PolicAveOTWorked", "LandArea", "PopDens", "PctUsePubTrans", "PolicCars", "PolicOperBudg", "LemasPctPolicOnPatr", "LemasGangUnitDeploy", "LemasPctOfficDrugUn", "PolicBudgPerPop", "ViolentCrimesPerPop"] dataset_url = "http://archive.ics.uci.edu/ml/machine-learning-databases/communities/communities.data" # Read dataset from the UCI web repository and assign column names. data_df = pd.read_csv(dataset_url, sep=",", names=column_names, na_values="?") # Make sure that there are no missing values in the "ViolentCrimesPerPop" column. assert(not data_df["ViolentCrimesPerPop"].isna().any()) # Real-valued label: "ViolentCrimesPerPop". labels_df = data_df["ViolentCrimesPerPop"] # Now that we have assigned binary labels, # we drop the "ViolentCrimesPerPop" column from the data frame. data_df.drop(columns="ViolentCrimesPerPop", inplace=True) # Group features. race_black_70_percentile = data_df["racepctblack"].quantile(q=0.7) groups_df = (data_df["racepctblack"] >= race_black_70_percentile) # Drop categorical features. data_df.drop(columns=["state", "county", "community", "communityname", "fold"], inplace=True) # Handle missing features. feature_names = data_df.columns for feature_name in feature_names: missing_rows = data_df[feature_name].isna() # Which rows have missing values? if missing_rows.any(): # Check if at least one row has a missing value. data_df[feature_name].fillna(0.0, inplace=True) # Fill NaN with 0. missing_rows.rename(feature_name + "_is_missing", inplace=True) data_df = data_df.join(missing_rows) # Append boolean "is_missing" feature. labels = labels_df.values.astype(np.float32) groups = groups_df.values.astype(np.float32) features = data_df.values.astype(np.float32) # Set random seed so that the results are reproducible. np.random.seed(123456) # We randomly divide the examples into 'num_queries' queries. queries = np.random.randint(0, num_queries, size=features.shape[0]) # Train and test indices. train_indices, test_indices = model_selection.train_test_split( range(features.shape[0]), test_size=0.4) # Train features, labels and protected groups. train_set = { 'features': features[train_indices, :], 'labels': labels[train_indices], 'groups': groups[train_indices], 'queries': queries[train_indices], 'dimension': features.shape[-1], 'num_queries': num_queries } # Test features, labels and protected groups. test_set = { 'features': features[test_indices, :], 'labels': labels[test_indices], 'groups': groups[test_indices], 'queries': queries[test_indices], 'dimension': features.shape[-1], 'num_queries': num_queries } Explanation: Pairwise Regression Fairness We will be training a linear scoring function $f(x) = w^\top x$ where $x \in \mathbb{R}^d$ is the input feature vector. Our goal is to train the regression model subject to pairwise fairness constraints. Specifically, for the regression model $f$, we denote: - $sqerr(f)$ as the squared error for model $f$. $$ sqerr(f) = \mathbf{E}\big[\big(f(x) - y\big)^2\big] $$ $err_{i,j}(f)$ as the pairwise error over example pairs where the higher label example is from group $i$, and the lower label example is from group $j$. $$ err_{i, j}(f) = \mathbf{E}\big[\mathbb{I}\big(f(x) < f(x')\big) \,\big|\, y > y',~ grp(x) = i, ~grp(x') = j\big] $$ <br> We then wish to solve the following constrained problem: $$min_f\; sqerr(f)$$ $$\text{ s.t. } |err_{i,j}(f) - err_{k,\ell}(f)| \leq \epsilon \;\;\; \forall ((i,j), (k,\ell)) \in \mathcal{G},$$ where $\mathcal{G}$ contains the pairs we are interested in constraining. Load Communities & Crime Data We will use the benchmark Communities and Crimes dataset from the UCI Machine Learning repository for our illustration. This dataset contains various demographic and racial distribution details (aggregated from census and law enforcement data sources) about different communities in the US, along with the per capita crime rate in each commmunity. Our goal is to predict the crime rate for a community, a regression problem. We consider communities where the percentage of black population is above the 70-th percentile as the protected group. End of explanation def pair_high_low_docs(data): # Returns a DataFrame of pairs of larger-smaller labeled regression examples # given in DataFrame. # For all pairs of docs, and remove rows that are not needed. pos_docs = data.copy() neg_docs = data.copy() # Include a merge key. pos_docs.insert(0, "merge_key", 0) neg_docs.insert(0, "merge_key", 0) # Merge docs and drop merge key and label column. pairs = pos_docs.merge(neg_docs, on="merge_key", how="outer", suffixes=("_pos", "_neg")) # Only retain rows where label_pos > label_neg. pairs = pairs[pairs.label_pos > pairs.label_neg] # Drop merge_key. pairs.drop(columns=["merge_key"], inplace=True) return pairs def convert_labeled_to_paired_data(data_dict, index=None): # Forms pairs of examples from each batch/query. # Converts data arrays to pandas DataFrame with required column names and # makes a call to convert_df_to_pairs and returns a dictionary. features = data_dict['features'] labels = data_dict['labels'] groups = data_dict['groups'] queries = data_dict['queries'] if index is not None: data_df = pd.DataFrame(features[queries == index, :]) data_df = data_df.assign(label=pd.DataFrame(labels[queries == index])) data_df = data_df.assign(group=pd.DataFrame(groups[queries == index])) data_df = data_df.assign(query_id=pd.DataFrame(queries[queries == index])) else: data_df = pd.DataFrame(features) data_df = data_df.assign(label=pd.DataFrame(labels)) data_df = data_df.assign(group=pd.DataFrame(groups)) data_df = data_df.assign(query_id=pd.DataFrame(queries)) # Forms pairs of positive-negative docs for each query in given DataFrame # if the DataFrame has a query_id column. Otherise forms pairs from all rows # of the DataFrame. data_pairs = data_df.groupby('query_id').apply(pair_high_low_docs) # Create groups ndarray. pos_groups = data_pairs['group_pos'].values.reshape(-1, 1) neg_groups = data_pairs['group_neg'].values.reshape(-1, 1) group_pairs = np.concatenate((pos_groups, neg_groups), axis=1) # Create queries ndarray. query_pairs = data_pairs['query_id_pos'].values.reshape(-1,) # Create features ndarray. feature_names = data_df.columns feature_names = feature_names.drop(['query_id', 'label']) feature_names = feature_names.drop(['group']) pos_features = data_pairs[[str(s) + '_pos' for s in feature_names]].values pos_features = pos_features.reshape(-1, 1, len(feature_names)) neg_features = data_pairs[[str(s) + '_neg' for s in feature_names]].values neg_features = neg_features.reshape(-1, 1, len(feature_names)) feature_pairs = np.concatenate((pos_features, neg_features), axis=1) # Paired data dict. paired_data = { 'feature_pairs': feature_pairs, 'group_pairs': group_pairs, 'query_pairs': query_pairs, 'features': features, 'labels': labels, 'queries': queries, 'dimension': data_dict['dimension'], 'num_queries': data_dict['num_queries'] } return paired_data Explanation: Evaluation Metrics We will need functions to convert labeled data into paired data. End of explanation def get_mask(groups, pos_group, neg_group=None): # Returns a boolean mask selecting positive-negative document pairs where # the protected group for the positive document is pos_group and # the protected group for the negative document (if specified) is neg_group. # Repeat group membership positive docs as many times as negative docs. mask_pos = groups[:, 0] == pos_group if neg_group is None: return mask_pos else: mask_neg = groups[:, 1] == neg_group return mask_pos & mask_neg def mean_squared_error(model, dataset): # Returns mean squared error for Keras model on dataset. scores = model.predict(dataset['features']) labels = dataset['labels'] return np.mean((scores - labels) ** 2) def group_error_rate(model, dataset, pos_group, neg_group=None): # Returns error rate for Keras model on data set, considering only document # pairs where the protected group for the positive document is pos_group, and # the protected group for the negative document (if specified) is neg_group. d = dataset['dimension'] scores0 = model.predict(dataset['feature_pairs'][:, 0, :].reshape(-1, d)) scores1 = model.predict(dataset['feature_pairs'][:, 1, :].reshape(-1, d)) mask = get_mask(dataset['group_pairs'], pos_group, neg_group) diff = scores0 - scores1 diff = diff[mask > 0].reshape((-1)) return np.mean(diff < 0) Explanation: We will also need functions to evaluate the pairwise error rates for a linear model. End of explanation def create_scoring_model(feature_pairs, features, dimension): # Returns a linear Keras scoring model, and returns a nullary function # returning predictions on the features. # Linear scoring model with no hidden layers. layers = [] # Input layer takes `dimension` inputs. layers.append(tf.keras.Input(shape=(dimension,))) layers.append(tf.keras.layers.Dense(1)) scoring_model = tf.keras.Sequential(layers) # Create a nullary function that returns applies the linear model to the # features and returns the tensor with the prediction differences on pairs. def prediction_diffs(): scores0 = scoring_model(feature_pairs()[:, 0, :].reshape(-1, dimension)) scores1 = scoring_model(feature_pairs()[:, 1, :].reshape(-1, dimension)) return scores0 - scores1 # Create a nullary function that returns the predictions on individual # examples. predictions = lambda: scoring_model(features()) return scoring_model, prediction_diffs, predictions Explanation: Create Linear Model We then write a function to create the linear scoring model. End of explanation def group_mask_fn(groups, pos_group, neg_group=None): # Returns a nullary function returning group mask. group_mask = lambda: np.reshape( get_mask(groups(), pos_group, neg_group), (-1)) return group_mask def formulate_problem( feature_pairs, group_pairs, features, labels, dimension, constraint_groups=[], constraint_slack=None): # Formulates a constrained problem that optimizes the squared error for a linear # model on the specified dataset, subject to pairwise fairness constraints # specified by the constraint_groups and the constraint_slack. # # Args: # feature_pairs: Nullary function returning paired features # group_pairs: Nullary function returning paired groups # features: Nullary function returning features # labels: Nullary function returning labels # dimension: Input dimension for scoring model # constraint_groups: List containing tuples of the form # ((pos_group0, neg_group0), (pos_group1, neg_group1)), specifying the # group memberships for the document pairs to compare in the constraints. # constraint_slack: slackness '\epsilon' allowed in the constraints. # Returns: # A RateMinimizationProblem object, and a Keras scoring model. # Create linear scoring model: we get back a Keras model and a nullary # function returning predictions on the features. scoring_model, prediction_diffs, predictions = create_scoring_model( feature_pairs, features, dimension) # Context for the optimization objective. context = tfco.rate_context(prediction_diffs) # Squared loss objective. squared_loss = lambda: tf.reduce_mean((predictions() - labels()) ** 2) # Constraint set. constraint_set = [] # Context for the constraints. for ((pos_group0, neg_group0), (pos_group1, neg_group1)) in constraint_groups: # Context for group 0. group_mask0 = group_mask_fn(group_pairs, pos_group0, neg_group0) context_group0 = context.subset(group_mask0) # Context for group 1. group_mask1 = group_mask_fn(group_pairs, pos_group1, neg_group1) context_group1 = context.subset(group_mask1) # Add constraints to constraint set. constraint_set.append( tfco.negative_prediction_rate(context_group0) <= ( tfco.negative_prediction_rate(context_group1) + constraint_slack)) constraint_set.append( tfco.negative_prediction_rate(context_group1) <= ( tfco.negative_prediction_rate(context_group0) + constraint_slack)) # Formulate constrained minimization problem. problem = tfco.RateMinimizationProblem( tfco.wrap_rate(squared_loss), constraint_set) return problem, scoring_model Explanation: Formulate Optimization Problem We are ready to formulate the constrained optimization problem using the TFCO library. End of explanation def train_model(train_set, params): # Trains the model with stochastic updates (one query per updates). # # Args: # train_set: Dictionary of "paired" training data. # params: Dictionary of hyper-paramters for training. # # Returns: # Trained model, list of objectives, list of group constraint violations. # Set random seed for reproducibility. random.seed(333333) np.random.seed(121212) tf.random.set_seed(212121) # Set up problem and model. if params['constrained']: # Constrained optimization. if params['constraint_type'] == 'marginal_equal_opportunity': constraint_groups = [((0, None), (1, None))] elif params['constraint_type'] == 'cross_group_equal_opportunity': constraint_groups = [((0, 1), (1, 0))] else: constraint_groups = [((0, 1), (1, 0)), ((0, 0), (1, 1))] else: # Unconstrained optimization. constraint_groups = [] # Dictionary that will hold batch features pairs, group pairs and labels for # current batch. We include one query per-batch. paired_batch = {} batch_index = 0 # Index of current query. # Data functions. feature_pairs = lambda: paired_batch['feature_pairs'] group_pairs = lambda: paired_batch['group_pairs'] features = lambda: paired_batch['features'] labels = lambda: paired_batch['labels'] # Create scoring model and constrained optimization problem. problem, scoring_model = formulate_problem( feature_pairs, group_pairs, features, labels, train_set['dimension'], constraint_groups, params['constraint_slack']) # Create a loss function for the problem. lagrangian_loss, update_ops, multipliers_variables = ( tfco.create_lagrangian_loss(problem, dual_scale=params['dual_scale'])) # Create optimizer optimizer = tf.keras.optimizers.Adagrad(learning_rate=params['learning_rate']) # List of trainable variables. var_list = ( scoring_model.trainable_weights + problem.trainable_variables + [multipliers_variables]) # List of objectives, group constraint violations. # violations, and snapshot of models during course of training. objectives = [] group_violations = [] models = [] feature_pair_batches = train_set['feature_pairs'] group_pair_batches = train_set['group_pairs'] query_pairs = train_set['query_pairs'] feature_batches = train_set['features'] label_batches = train_set['labels'] queries = train_set['queries'] print() # Run loops * iterations_per_loop full batch iterations. for ii in range(params['loops']): for jj in range(params['iterations_per_loop']): # Populate paired_batch dict with all pairs for current query. The batch # index is the same as the current query index. paired_batch = { 'feature_pairs': feature_pair_batches[query_pairs == batch_index], 'group_pairs': group_pair_batches[query_pairs == batch_index], 'features': feature_batches[queries == batch_index], 'labels': label_batches[queries == batch_index] } # Optimize loss. update_ops() optimizer.minimize(lagrangian_loss, var_list=var_list) # Update batch_index, and cycle back once last query is reached. batch_index = (batch_index + 1) % train_set['num_queries'] # Snap shot current model. model_copy = tf.keras.models.clone_model(scoring_model) model_copy.set_weights(scoring_model.get_weights()) models.append(model_copy) # Evaluate metrics for snapshotted model. error, gerr, group_viol = evaluate_results( scoring_model, train_set, params) objectives.append(error) group_violations.append( [x - params['constraint_slack'] for x in group_viol]) sys.stdout.write( '\r Loop %d: error = %.3f, max constraint violation = %.3f' % (ii, objectives[-1], max(group_violations[-1]))) print() if params['constrained']: # Find model iterate that trades-off between objective and group violations. best_index = tfco.find_best_candidate_index( np.array(objectives), np.array(group_violations), rank_objectives=False) else: # Find model iterate that achieves lowest objective. best_index = np.argmin(objectives) return models[best_index] Explanation: Train Model The following function then trains the linear model by solving the above constrained optimization problem. We first provide a training function with minibatch gradient updates. There are three types of pairwise fairness criterion we handle (specified by 'constraint_type'), and assign the (pos_group, neg_group) pairs to compare accordingly. End of explanation def evaluate_results(model, test_set, params): # Returns sqaured error, group error rates, group-level constraint violations. if params['constraint_type'] == 'marginal_equal_opportunity': g0_error = group_error_rate(model, test_set, 0) g1_error = group_error_rate(model, test_set, 1) group_violations = [g0_error - g1_error, g1_error - g0_error] return (mean_squared_error(model, test_set), [g0_error, g1_error], group_violations) else: g00_error = group_error_rate(model, test_set, 0, 0) g01_error = group_error_rate(model, test_set, 0, 1) g10_error = group_error_rate(model, test_set, 1, 1) g11_error = group_error_rate(model, test_set, 1, 1) group_violations_offdiag = [g01_error - g10_error, g10_error - g01_error] group_violations_diag = [g00_error - g11_error, g11_error - g00_error] if params['constraint_type'] == 'cross_group_equal_opportunity': return (mean_squared_error(model, test_set), [[g00_error, g01_error], [g10_error, g11_error]], group_violations_offdiag) else: return (mean_squared_error(model, test_set), [[g00_error, g01_error], [g10_error, g11_error]], group_violations_offdiag + group_violations_diag) def display_results( model, test_set, params, method, error_type, show_header=False): # Prints evaluation results for model on test data. error, group_error, diffs = evaluate_results(model, test_set, params) if params['constraint_type'] == 'marginal_equal_opportunity': if show_header: print('\nMethod\t\t\tError\t\tMSE\t\tGroup 0\t\tGroup 1\t\tDiff') print('%s\t%s\t\t%.3f\t\t%.3f\t\t%.3f\t\t%.3f' % ( method, error_type, error, group_error[0], group_error[1], np.max(diffs))) elif params['constraint_type'] == 'cross_group_equal_opportunity': if show_header: print('\nMethod\t\t\tError\t\tMSE\t\tGroup 0/1\tGroup 1/0\tDiff') print('%s\t%s\t\t%.3f\t\t%.3f\t\t%.3f\t\t%.3f' % ( method, error_type, error, group_error[0][1], group_error[1][0], np.max(diffs))) else: if show_header: print('\nMethod\t\t\tError\t\MSE\t\tGroup 0/1\tGroup 1/0\t' + 'Group 0/0\tGroup 1/1\tDiff') print('%s\t%s\t\t%.3f\t\t%.3f\t\t%.3f\t\t%.3f\t\t%.3f\t\t%.3f' % ( method, error_type, error, group_error[0][1], group_error[1][0], group_error[0][0], group_error[1][1], np.max(diffs))) Explanation: Summarize and Plot Results Having trained a model, we will need functions to summarize the various evaluation metrics. End of explanation # Convert train/test set to paired data for later evaluation. paired_train_set = convert_labeled_to_paired_data(train_set) paired_test_set = convert_labeled_to_paired_data(test_set) Explanation: Experimental Results We now run experiments with two types of pairwise fairness criteria: (1) marginal_equal_opportunity and (2) pairwise equal opportunity. In each case, we compare an unconstrained model trained to optimize just the squared error and a constrained model trained with pairwise fairness constraints. End of explanation # Model hyper-parameters. model_params = { 'loops': 10, 'iterations_per_loop': 250, 'learning_rate': 0.1, 'constraint_type': 'marginal_equal_opportunity', 'constraint_slack': 0.02, 'dual_scale': 1.0} # Unconstrained optimization. model_params['constrained'] = False model_unc = train_model(paired_train_set, model_params) display_results(model_unc, paired_train_set, model_params, 'Unconstrained ', 'Train', show_header=True) display_results(model_unc, paired_test_set, model_params, 'Unconstrained ', 'Test') # Constrained optimization with TFCO. model_params['constrained'] = True model_con = train_model(paired_train_set, model_params) display_results(model_con, paired_train_set, model_params, 'Constrained ', 'Train', show_header=True) display_results(model_con, paired_test_set, model_params, 'Constrained ', 'Test') Explanation: (1) Marginal Equal Opportunity For a scoring model $f: \mathbb{R}^d \rightarrow \mathbb{R}$, recall: - $sqerr(f)$ as the squared error for scoring function $f$. and we additionally define: $err_i(f)$ as the row-marginal pairwise error over example pairs where the higher label example is from group $i$, and the lower label is from either groups $$ err_i(f) = \mathbf{E}\big[\mathbb{I}\big(f(x) < f(x')\big) \,\big|\, y > y',~ grp(x) = i\big] $$ The constrained optimization problem we solve constraints the row-marginal pairwise errors to be similar: $$min_f\;sqerr(f)$$ $$\text{s.t. }\;|err_0(f) - err_1(f)| \leq 0.02$$ End of explanation # Model hyper-parameters. model_params = { 'loops': 10, 'iterations_per_loop': 250, 'learning_rate': 0.1, 'constraint_type': 'cross_group_equal_opportunity', 'constraint_slack': 0.02, 'dual_scale': 1.0} # Unconstrained optimization. model_params['constrained'] = False model_unc = train_model(paired_train_set, model_params) display_results(model_unc, paired_train_set, model_params, 'Unconstrained ', 'Train', show_header=True) display_results(model_unc, paired_test_set, model_params, 'Unconstrained ', 'Test') # Constrained optimization with TFCO. model_params['constrained'] = True model_con = train_model(paired_train_set, model_params) display_results(model_con, paired_train_set, model_params, 'Constrained ', 'Train', show_header=True) display_results(model_con, paired_test_set, model_params, 'Constrained ', 'Test') Explanation: (2) Pairwise Equal Opportunity Recall that we denote $err_{i,j}(f)$ as the pairwise error over example pairs where the higher label example is from group $i$, and the lower label example is from group $j$. $$ err_{i, j}(f) ~=~ \mathbf{E}\big[\mathbb{I}\big(f(x) < f(x')\big) \,\big|\, y > y',~ grp(x) = i, ~grp(x') = j\big] $$ We first constrain only the cross-group errors, highlighted below. <br> <table border='1' bordercolor='black'> <tr > <td bgcolor='white'> </td> <td bgcolor='white'> </td> <td bgcolor='white' colspan=2 align=center><b>Negative</b></td> </tr> <tr> <td bgcolor='white'></td> <td bgcolor='white'></td> <td>Group 0</td> <td>Group 1</td> </tr> <tr> <td bgcolor='white' rowspan=2><b>Positive</b></td> <td bgcolor='white'>Group 0</td> <td bgcolor='white'>$err_{0,0}$</td> <td bgcolor='white'>$\mathbf{err_{0,1}}$</td> </tr> <tr> <td>Group 1</td> <td bgcolor='white'>$\mathbf{err_{1,0}}$</td> <td bgcolor='white'>$err_{1,1}$</td> </tr> </table> <br> The optimization problem we solve constraints the cross-group pairwise errors to be similar: $$min_f\; sqerr(f)$$ $$\text{s.t. }\;\; |err_{0,1}(f) - err_{1,0}(f)| \leq 0.02$$ End of explanation <END_TASK>
15,908
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Whitening evoked data with a noise covariance Evoked data are loaded and then whitened using a given noise covariance matrix. It's an excellent quality check to see if baseline signals match the assumption of Gaussian white noise during the baseline period. Covariance estimation and diagnostic plots are based on [1]_. References .. [1] Engemann D. and Gramfort A. (2015) Automated model selection in covariance estimation and spatial whitening of MEG and EEG signals, vol. 108, 328-342, NeuroImage. Step1: Set parameters Step2: Compute covariance using automated regularization Step3: Show the evoked data Step4: We can then show whitening for our various noise covariance estimates. Here we should look to see if baseline signals match the assumption of Gaussian white noise. we expect values centered at 0 within 2 standard deviations for 95% of the time points. For the Global field power we expect a value of 1.
<ASSISTANT_TASK:> Python Code: # Authors: Alexandre Gramfort <alexandre.gramfort@telecom-paristech.fr> # Denis A. Engemann <denis.engemann@gmail.com> # # License: BSD (3-clause) import mne from mne import io from mne.datasets import sample from mne.cov import compute_covariance print(__doc__) Explanation: Whitening evoked data with a noise covariance Evoked data are loaded and then whitened using a given noise covariance matrix. It's an excellent quality check to see if baseline signals match the assumption of Gaussian white noise during the baseline period. Covariance estimation and diagnostic plots are based on [1]_. References .. [1] Engemann D. and Gramfort A. (2015) Automated model selection in covariance estimation and spatial whitening of MEG and EEG signals, vol. 108, 328-342, NeuroImage. End of explanation data_path = sample.data_path() raw_fname = data_path + '/MEG/sample/sample_audvis_filt-0-40_raw.fif' event_fname = data_path + '/MEG/sample/sample_audvis_filt-0-40_raw-eve.fif' raw = io.read_raw_fif(raw_fname, preload=True) raw.filter(1, 40, n_jobs=1, fir_design='firwin') raw.info['bads'] += ['MEG 2443'] # bads + 1 more events = mne.read_events(event_fname) # let's look at rare events, button presses event_id, tmin, tmax = 2, -0.2, 0.5 picks = mne.pick_types(raw.info, meg=True, eeg=True, eog=True, exclude='bads') reject = dict(mag=4e-12, grad=4000e-13, eeg=80e-6) epochs = mne.Epochs(raw, events, event_id, tmin, tmax, picks=picks, baseline=None, reject=reject, preload=True) # Uncomment next line to use fewer samples and study regularization effects # epochs = epochs[:20] # For your data, use as many samples as you can! Explanation: Set parameters End of explanation method_params = dict(diagonal_fixed=dict(mag=0.01, grad=0.01, eeg=0.01)) noise_covs = compute_covariance(epochs, tmin=None, tmax=0, method='auto', return_estimators=True, verbose=True, n_jobs=1, projs=None, rank=None, method_params=method_params) # With "return_estimator=True" all estimated covariances sorted # by log-likelihood are returned. print('Covariance estimates sorted from best to worst') for c in noise_covs: print("%s : %s" % (c['method'], c['loglik'])) Explanation: Compute covariance using automated regularization End of explanation evoked = epochs.average() evoked.plot(time_unit='s') # plot evoked response Explanation: Show the evoked data: End of explanation evoked.plot_white(noise_covs, time_unit='s') Explanation: We can then show whitening for our various noise covariance estimates. Here we should look to see if baseline signals match the assumption of Gaussian white noise. we expect values centered at 0 within 2 standard deviations for 95% of the time points. For the Global field power we expect a value of 1. End of explanation <END_TASK>
15,909
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: &larr; Back to Index NumPy and SciPy The quartet of NumPy, SciPy, Matplotlib, and IPython is a popular combination in the Python world. We will use each of these libraries in this workshop. Tutorial NumPy is one of the most popular libraries for numerical computing in the world. It is used in several disciplines including image processing, finance, bioinformatics, and more. This entire workshop is based upon NumPy and its derivatives. If you are new to NumPy, follow this NumPy Tutorial. SciPy is a Python library for scientific computing which builds on top of NumPy. If NumPy is like the Matlab core, then SciPy is like the Matlab toolboxes. It includes support for linear algebra, sparse matrices, spatial data structions, statistics, and more. While there is a SciPy Tutorial, it isn't critical that you follow it for this workshop. Special Arrays Step1: Slicing Arrays Step2: The optional third parameter indicates the increment value Step3: If you omit the start index, the slice implicitly starts from zero Step4: Array Arithmetic Step5: dot computes the dot product, or inner product, between arrays or matrices. Step6: Boolean Operations Step7: Distance Metrics Step8: The cosine distance measures the angle between two vectors Step9: Sorting NumPy arrays have a method, sort, which sorts the array in-place. Step10: numpy.argsort returns an array of indices, ind, such that x[ind] is a sorted version of x.
<ASSISTANT_TASK:> Python Code: print numpy.arange(5) print numpy.linspace(0, 5, 10, endpoint=False) print numpy.zeros(5) print numpy.ones(5) print numpy.ones((5,2)) print scipy.randn(5) # random Gaussian, zero-mean unit-variance print scipy.randn(5,2) Explanation: &larr; Back to Index NumPy and SciPy The quartet of NumPy, SciPy, Matplotlib, and IPython is a popular combination in the Python world. We will use each of these libraries in this workshop. Tutorial NumPy is one of the most popular libraries for numerical computing in the world. It is used in several disciplines including image processing, finance, bioinformatics, and more. This entire workshop is based upon NumPy and its derivatives. If you are new to NumPy, follow this NumPy Tutorial. SciPy is a Python library for scientific computing which builds on top of NumPy. If NumPy is like the Matlab core, then SciPy is like the Matlab toolboxes. It includes support for linear algebra, sparse matrices, spatial data structions, statistics, and more. While there is a SciPy Tutorial, it isn't critical that you follow it for this workshop. Special Arrays End of explanation x = numpy.arange(10) print x[2:4] print x[-1] Explanation: Slicing Arrays End of explanation print x[0:8:2] print x[4:2:-1] Explanation: The optional third parameter indicates the increment value: End of explanation print x[:4] print x[:999] print x[::-1] Explanation: If you omit the start index, the slice implicitly starts from zero: End of explanation x = numpy.arange(5) y = numpy.ones(5) print x+2*y Explanation: Array Arithmetic End of explanation x = scipy.randn(5) y = numpy.ones(5) print numpy.dot(x, y) x = scipy.randn(5,3) y = numpy.ones((3,2)) print numpy.dot(x, y) Explanation: dot computes the dot product, or inner product, between arrays or matrices. End of explanation x = numpy.arange(10) print x < 5 y = numpy.ones(10) print x < y Explanation: Boolean Operations End of explanation from scipy.spatial import distance print distance.euclidean([0, 0], [3, 4]) print distance.sqeuclidean([0, 0], [3, 4]) print distance.cityblock([0, 0], [3, 4]) print distance.chebyshev([0, 0], [3, 4]) Explanation: Distance Metrics End of explanation print distance.cosine([67, 0], [89, 0]) print distance.cosine([67, 0], [0, 89]) Explanation: The cosine distance measures the angle between two vectors: End of explanation x = scipy.randn(5) print x x.sort() print x Explanation: Sorting NumPy arrays have a method, sort, which sorts the array in-place. End of explanation x = scipy.randn(5) print x ind = numpy.argsort(x) print ind print x[ind] Explanation: numpy.argsort returns an array of indices, ind, such that x[ind] is a sorted version of x. End of explanation <END_TASK>
15,910
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Deep Convolutional Neural Network in TensorFlow In this notebook, we convert our LeNet-5-inspired, MNIST-classifying, deep convolutional network from Keras to TensorFlow (compare them side by side) following Aymeric Damien's style. Load dependencies Step1: Load data Step2: Set neural network hyperparameters Step3: Set parameters for each layer Step4: Define placeholder Tensors for inputs and labels Step5: Define types of layers Step6: Design neural network architecture Step7: Define dictionaries for storing weights and biases for each layer -- and initialize Step8: Build model Step9: Define model's loss and its optimizer Step10: Define evaluation metrics Step11: Create op for variable initialization Step12: Train the network in a session (identical to intermediate_net_in_tensorflow.ipynb except addition of display_progress)
<ASSISTANT_TASK:> Python Code: import numpy as np np.random.seed(42) import tensorflow as tf tf.set_random_seed(42) Explanation: Deep Convolutional Neural Network in TensorFlow In this notebook, we convert our LeNet-5-inspired, MNIST-classifying, deep convolutional network from Keras to TensorFlow (compare them side by side) following Aymeric Damien's style. Load dependencies End of explanation from tensorflow.examples.tutorials.mnist import input_data mnist = input_data.read_data_sets("MNIST_data/", one_hot=True) Explanation: Load data End of explanation epochs = 20 batch_size = 128 display_progress = 40 # after this many batches, output progress to screen wt_init = tf.contrib.layers.xavier_initializer() # weight initializer Explanation: Set neural network hyperparameters End of explanation # input layer: n_input = 784 # first convolutional layer: n_conv_1 = 32 k_conv_1 = 3 # k_size # second convolutional layer: n_conv_2 = 64 k_conv_2 = 3 # max pooling layer: pool_size = 2 mp_layer_dropout = 0.25 # dense layer: n_dense = 128 dense_layer_dropout = 0.5 # output layer: n_classes = 10 Explanation: Set parameters for each layer End of explanation x = tf.placeholder(tf.float32, [None, n_input]) y = tf.placeholder(tf.float32, [None, n_classes]) Explanation: Define placeholder Tensors for inputs and labels End of explanation # dense layer with ReLU activation: def dense(x, W, b): z = tf.add(tf.matmul(x, W), b) a = tf.nn.relu(z) return a # convolutional layer with ReLU activation: def conv2d(x, W, b, stride_length=1): xW = tf.nn.conv2d(x, W, strides=[1, stride_length, stride_length, 1], padding='SAME') z = tf.nn.bias_add(xW, b) a = tf.nn.relu(z) return a # max-pooling layer: def maxpooling2d(x, p_size): return tf.nn.max_pool(x, ksize=[1, p_size, p_size, 1], strides=[1, p_size, p_size, 1], padding='SAME') Explanation: Define types of layers End of explanation def network(x, weights, biases, n_in, mp_psize, mp_dropout, dense_dropout): # reshape linear MNIST pixel input into square image: square_dimensions = int(np.sqrt(n_in)) square_x = tf.reshape(x, shape=[-1, square_dimensions, square_dimensions, 1]) # convolutional and max-pooling layers: conv_1 = conv2d(square_x, weights['W_c1'], biases['b_c1']) conv_2 = conv2d(conv_1, weights['W_c2'], biases['b_c2']) pool_1 = maxpooling2d(conv_2, mp_psize) pool_1 = tf.nn.dropout(pool_1, 1-mp_dropout) # dense layer: flat = tf.reshape(pool_1, [-1, weights['W_d1'].get_shape().as_list()[0]]) dense_1 = dense(flat, weights['W_d1'], biases['b_d1']) dense_1 = tf.nn.dropout(dense_1, 1-dense_dropout) # output layer: out_layer_z = tf.add(tf.matmul(dense_1, weights['W_out']), biases['b_out']) return out_layer_z Explanation: Design neural network architecture End of explanation bias_dict = { 'b_c1': tf.Variable(tf.zeros([n_conv_1])), 'b_c2': tf.Variable(tf.zeros([n_conv_2])), 'b_d1': tf.Variable(tf.zeros([n_dense])), 'b_out': tf.Variable(tf.zeros([n_classes])) } # calculate number of inputs to dense layer: full_square_length = np.sqrt(n_input) pooled_square_length = int(full_square_length / pool_size) dense_inputs = pooled_square_length**2 * n_conv_2 weight_dict = { 'W_c1': tf.get_variable('W_c1', [k_conv_1, k_conv_1, 1, n_conv_1], initializer=wt_init), 'W_c2': tf.get_variable('W_c2', [k_conv_2, k_conv_2, n_conv_1, n_conv_2], initializer=wt_init), 'W_d1': tf.get_variable('W_d1', [dense_inputs, n_dense], initializer=wt_init), 'W_out': tf.get_variable('W_out', [n_dense, n_classes], initializer=wt_init) } Explanation: Define dictionaries for storing weights and biases for each layer -- and initialize End of explanation predictions = network(x, weight_dict, bias_dict, n_input, pool_size, mp_layer_dropout, dense_layer_dropout) Explanation: Build model End of explanation cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=predictions, labels=y)) optimizer = tf.train.AdamOptimizer().minimize(cost) Explanation: Define model's loss and its optimizer End of explanation correct_prediction = tf.equal(tf.argmax(predictions, 1), tf.argmax(y, 1)) accuracy_pct = tf.reduce_mean(tf.cast(correct_prediction, tf.float32)) * 100 Explanation: Define evaluation metrics End of explanation initializer_op = tf.global_variables_initializer() Explanation: Create op for variable initialization End of explanation with tf.Session() as session: session.run(initializer_op) print("Training for", epochs, "epochs.") # loop over epochs: for epoch in range(epochs): avg_cost = 0.0 # track cost to monitor performance during training avg_accuracy_pct = 0.0 # loop over all batches of the epoch: n_batches = int(mnist.train.num_examples / batch_size) for i in range(n_batches): # to reassure you something's happening! if i % display_progress == 0: print("Step ", i+1, " of ", n_batches, " in epoch ", epoch+1, ".", sep='') batch_x, batch_y = mnist.train.next_batch(batch_size) # feed batch data to run optimization and fetching cost and accuracy: _, batch_cost, batch_acc = session.run([optimizer, cost, accuracy_pct], feed_dict={x: batch_x, y: batch_y}) # accumulate mean loss and accuracy over epoch: avg_cost += batch_cost / n_batches avg_accuracy_pct += batch_acc / n_batches # output logs at end of each epoch of training: print("Epoch ", '%03d' % (epoch+1), ": cost = ", '{:.3f}'.format(avg_cost), ", accuracy = ", '{:.2f}'.format(avg_accuracy_pct), "%", sep='') print("Training Complete. Testing Model.\n") test_cost = cost.eval({x: mnist.test.images, y: mnist.test.labels}) test_accuracy_pct = accuracy_pct.eval({x: mnist.test.images, y: mnist.test.labels}) print("Test Cost:", '{:.3f}'.format(test_cost)) print("Test Accuracy: ", '{:.2f}'.format(test_accuracy_pct), "%", sep='') Explanation: Train the network in a session (identical to intermediate_net_in_tensorflow.ipynb except addition of display_progress) End of explanation <END_TASK>
15,911
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Sparse CNN The idea of the sparse CNN is to utilize only the important values of its filters and set those less impotant to zero. For terminology, we'll call the weights of a specific channel a "filter", and we'll call the values of each filter a "connection". Via Hebian inspired learning, those connections that see high absolute-values in both their inputs and outputs will be favored. This commparison of input and output will be refered to as the connection's strength. Connection Strength Calculations Initialize a typical conv layer to start. From here, three seperate methods will be attempted to calculate the strength of the connections. Step3: Vectorized Method The idea of this method is to utilize convolutional arithmetic to determine the input for a given output unit and a given connection. Suppose we initialize a weight matrix of exactly the same dimensions of our orignal conv layer, and set all of filters to 0 except for one connection. That is, new_conv.weight[ Step5: Vectorized Method + Grouping Same as the previous method, but utilizing the grouping argument of the conv layer so that only one is needed Step7: Vectorized Method with Less Redundancies Step9: Brute Force Method Computationally speaking, this is the same method as the preivous two. Only now, instead of using conv layers to assist in the computations, we use for loops the brute force our way through. This is more so for a sanity check on the first two to validate their outputs. Step10: Validation \ Time Trials Quick test to make sure they all give the same output. Let's see how long the take to run. Step13: Sparse Conv Layer Now to implement a conv layer that utilizes the second implementation above. Step14: Test Training a Network The following is a simple toy example copied mostly from Step15: Time Trials Quick test to compare runtime with and without updating the connections tensor. Step16: Network Setup Step17: Training Step18: Testing Accuracy
<ASSISTANT_TASK:> Python Code: c_out = 7 c_in = 8 conv = torch.nn.Conv2d(c_in, c_out, kernel_size=(2, 2), stride=(1, 1), padding=0, dilation=1, groups=1) input_tensor = torch.randn(2, c_in, 5, 3) output_tensor = conv(input_tensor) # Get params of conv layer. in_channels = conv.in_channels # Number of channels in the input image out_channels = conv.out_channels # Number of channels produced by the convolution kernel_size = conv.kernel_size # Size of the convolving kernel stride = conv.stride # Stride of the convolution. Default: 1 padding = conv.padding # Zero-padding added to both sides of the input. Default: 0 padding_mode = conv.padding_mode # zeros dilation = conv.dilation # Spacing between kernel elements. Default: 1 groups = conv.groups # Number of blocked connections from input channels to output channels. Default: 1 bias = conv.bias is not None Explanation: Sparse CNN The idea of the sparse CNN is to utilize only the important values of its filters and set those less impotant to zero. For terminology, we'll call the weights of a specific channel a "filter", and we'll call the values of each filter a "connection". Via Hebian inspired learning, those connections that see high absolute-values in both their inputs and outputs will be favored. This commparison of input and output will be refered to as the connection's strength. Connection Strength Calculations Initialize a typical conv layer to start. From here, three seperate methods will be attempted to calculate the strength of the connections. End of explanation def get_single_unit_conv(c, j, h, **kwargs): Constructs and returns conv layer with trainging disabled and all zero weights except along the output channels for unit specified as (c, j, h). # Construct conv. conv = torch.nn.Conv2d(**kwargs) # Turn off training. conv.train = False # Set weights to zero except those specified. with torch.no_grad(): conv.weight.set_(torch.zeros_like(conv.weight)) conv.weight[:, c, j, h] = 1 return conv # Get inidices that loop over all connections. single_unit_indxs = list(itertools.product(*[range(d) for d in conv.weight.shape[1:]])) single_unit_convs = [ get_single_unit_conv( c, j, h, in_channels=in_channels, out_channels=out_channels, kernel_size=kernel_size, stride=stride, padding=padding, padding_mode=padding_mode, dilation=dilation, groups=groups, bias=False, ) for c, j, h in single_unit_indxs ] def f1(): Calculate connection strengths. H = torch.zeros_like(conv.weight) s1 = torch.sigmoid(output_tensor).gt_(0.5) for idx, sconv in zip(single_unit_indxs, single_unit_convs): s2 = torch.sigmoid(sconv(input_tensor)).gt_(0.5) m = torch.sum(s1.mul(s2), (0, 2, 3,)) H[:, idx[0], idx[1], idx[2]] += m return H Explanation: Vectorized Method The idea of this method is to utilize convolutional arithmetic to determine the input for a given output unit and a given connection. Suppose we initialize a weight matrix of exactly the same dimensions of our orignal conv layer, and set all of filters to 0 except for one connection. That is, new_conv.weight[:, c, j, h] = 1. Now if we pass the input through new_conv, we'll be given an output tensor of the same size as the original, but with the input values arranged at the locations of their respective output through the connection. That is, ``` old_output = conv[input] new_output = new_conv[input] ==> for all b, j, and h (b being the batch), we have new_output[b, :, j, h] = input[<indices of input passed through connection conv.weight[:, c, j, h]>] examine_connections(old_output, new_output) # done in pointwise fashion `` With this vectorized calculation, we may then loop over all combinations ofc,j, andh`, compare the outputs to their respective inputs, and populate a matrix to record the strengths. End of explanation def get_single_unit_weights(shape, c, j, h, **kwargs): Constructs and returns conv layer with traingin diabled and all zero weights except along the output channels for unit specified as (c, j, h). # Construct weight. weight = torch.zeros(shape) # Set weights to zero except those specified. weight[:, c, j, h] = 1 return weight # Compute inidices that loop over all connections of a channel. filter_indxs = list(itertools.product(*[range(d) for d in conv.weight.shape[1:]])) # Compute indeces that loop over all channels and filters. # This will be used to unpack the pointwise comparisons of the output. connection_indxs = [] for idx in filter_indxs: i_ = list(idx) connection_indxs.extend([ [c_]+i_ for c_ in range(out_channels) ]) connection_indxs = list(zip(*connection_indxs)) # Create new conv layer that groups it's input and output. new_groups = len(filter_indxs) stacked_conv = torch.nn.Conv2d( in_channels=in_channels * new_groups, out_channels=out_channels * new_groups, kernel_size=kernel_size, stride=stride, padding=padding, padding_mode=padding_mode, dilation=dilation, groups=groups * new_groups, bias=False, ) # Populate the weight matrix with stacked tensors having only one non-zero unit. single_unit_weights = [ get_single_unit_weights( conv.weight.shape, c, j, h, ) for c, j, h in filter_indxs ] with torch.no_grad(): stacked_conv.weight.set_(torch.cat(single_unit_weights, dim=0)) def f2(): # print('------f2a--------') # print('input ', input_tensor.shape) stacked_input = input_tensor.repeat((1, new_groups, 1, 1)) stacked_output = stacked_conv(stacked_input) # print('stacked_input ', stacked_input.shape) # print('stacked_output', stacked_output.shape) H = torch.zeros_like(conv.weight) s1 = torch.sigmoid(stacked_output).gt_(0.5) s2 = torch.sigmoid(output_tensor).gt_(0.5).repeat((1, new_groups, 1, 1)) print('s1', s1.shape) print('s2', s2.shape) H_ = torch.sum(s2.mul(s1), (0, 2, 3,)) # print('H_', H_.shape) H[connection_indxs] = H_ # print('\n') return H Explanation: Vectorized Method + Grouping Same as the previous method, but utilizing the grouping argument of the conv layer so that only one is needed End of explanation def get_single_unit_weights_2b(shape, c, j, h, **kwargs): Constructs and returns conv layer with traingin diabled and all zero weights except along the output channels for unit specified as (c, j, h). # Construct weight. weight = torch.zeros(1, *shape[1:]) # Set weights to zero except those specified. weight[0, c, j, h] = 1 return weight # Compute inidices that loop over all connections of a channel. filter_indxs_2b = list(itertools.product(*[range(d) for d in conv.weight.shape[1:]])) # Compute indeces that loop over all channels and filters. # This will be used to unpack the pointwise comparisons of the output. connection_indxs_2b = [] for c_ in range(out_channels): for idx in filter_indxs_2b: i_ = list(idx) connection_indxs_2b.append([c_] + i_) connection_indxs_2b = list(zip(*connection_indxs_2b)) new_groups_2b = int(np.prod(conv.weight.shape[1:])) perm_indices_2b = [] for c_i in range(out_channels): perm_indices_2b.extend( [c_i] * new_groups_2b ) # Create new conv layer that groups it's input and output. stacked_conv_2b = torch.nn.Conv2d( in_channels=in_channels * new_groups_2b, out_channels=new_groups_2b, kernel_size=kernel_size, stride=stride, padding=padding, padding_mode=padding_mode, dilation=dilation, groups=groups * new_groups_2b, bias=False, ) # Populate the weight matrix with stacked tensors having only one non-zero unit. single_unit_weights_2b = [ get_single_unit_weights_2b( conv.weight.shape, c, j, h, ) for c, j, h in filter_indxs_2b ] with torch.no_grad(): stacked_conv_2b.weight.set_(torch.cat(single_unit_weights_2b, dim=0)) def f2b(): stacked_input = input_tensor.repeat((1, new_groups_2b, 1, 1)) stacked_output = stacked_conv_2b(stacked_input).repeat((1, out_channels, 1, 1)) H = torch.zeros_like(conv.weight) s1 = stacked_output s2 = output_tensor[:, perm_indices_2b, ...] mu_in = s1.mean(dim=0) mu_out = s2.mean(dim=0) std_in = s1.std(dim=0) std_out = s2.std(dim=0) corr = ((s1 - mu_in) * (s2 - mu_out)).mean(dim=0) / (std_in * std_out) corr[torch.where((std_in == 0 ) | (std_out == 0 ))] = 0 corr = corr.abs() H_ = torch.mean(corr, (1, 2)) H[connection_indxs_2b] = H_ return H Explanation: Vectorized Method with Less Redundancies End of explanation def coactivation(t1, t2): s = (torch.sigmoid(t1) > 0.5) * (torch.sigmoid(t2) > 0.5) return s def get_indeces_of_input_and_filter(n, m): Assumes dilation=1, i.e. typical conv. k1, k2 = kernel_size p1, p2 = padding s1, s2 = stride i1, i2 = (0, 0) i1 -= p1 i2 -= p2 i1 += n * s1 i2 += m * s2 if i2 == 2: import ipdb; ipdb.set_trace() indxs = [] for c_in in range(in_channels): for n_k1 in range(k1): for m_k2 in range(k2): filter_indx = (c_in, n_k1, m_k2) input_indx = (c_in, i1 + n_k1, i2 + m_k2) indxs.append((input_indx, filter_indx)) return indxs B = output_tensor.shape[0] N_out = output_tensor.shape[2] M_out = output_tensor.shape[3] C_in = conv.weight.shape[1] C_out = conv.weight.shape[0] def f3(): H = torch.zeros_like(conv.weight) for b in range(B): for c_out in range(C_out): for n_out in range(N_out): for m_out in range(M_out): unit_1 = output_tensor[b, c_out, n_out, m_out] indxs = get_indeces_of_input_and_filter(n_out, m_out) for input_indx, filter_indx in indxs: c_in, n_in, m_in = input_indx c_fl, n_fl, m_fl = filter_indx unit_2 = input_tensor[b, c_in, n_in, m_in] if coactivation(unit_1, unit_2): H[c_out, c_fl, n_fl, m_fl] += 1 return H Explanation: Brute Force Method Computationally speaking, this is the same method as the preivous two. Only now, instead of using conv layers to assist in the computations, we use for loops the brute force our way through. This is more so for a sanity check on the first two to validate their outputs. End of explanation assert f1().allclose(f2(), rtol=0, atol=0) and f2().allclose(f2b(), rtol=0, atol=0) and f2b().allclose(f3(), rtol=0, atol=0) %timeit f1() %timeit f2() %timeit f2b() # %timeit f3() Explanation: Validation \ Time Trials Quick test to make sure they all give the same output. Let's see how long the take to run. End of explanation class DSConv2d(torch.nn.Conv2d): def __init__(self, *args, **kwargs): super().__init__(*args, **kwargs) self.learning_iterations = 0 self.activity_threshold = 0.5 self.k1 = max(int(0.1 * np.prod(self.weight.shape[2:])), 1) self.k2 = max(int(0.15 * np.prod(self.weight.shape[2:])), 1) self.prune_dims = [0, 1] # Todo: sort self.connections_tensor = torch.zeros_like(self.weight) self.prune_mask = torch.ones_like(self.weight) # Compute inidices that loop over all connections of a channel. filter_indxs = list(itertools.product(*[range(d) for d in self.weight.shape[1:]])) # Compute indeces that loop over all channels and filters. # This will be used to unpack the pointwise comparisons of the output. self.connection_indxs = [] for idx in filter_indxs: i_ = list(idx) self.connection_indxs.extend([ [c]+i_ for c in range(self.weight.shape[0]) ]) self.connection_indxs = list(zip(*self.connection_indxs)) # Create new conv layer that groups it's input and output. self.new_groups = len(filter_indxs) self.stacked_conv = torch.nn.Conv2d( in_channels=self.in_channels * self.new_groups, out_channels=self.out_channels * self.new_groups, kernel_size=self.kernel_size, stride=self.stride, padding=self.padding, padding_mode=self.padding_mode, dilation=self.dilation, groups=self.groups * self.new_groups, bias=False, ) # Populate the weight matrix with stacked tensors having only one non-zero unit. single_unit_weights = [ self.get_single_unit_weights( self.weight.shape, c, j, h, ) for c, j, h in filter_indxs ] with torch.no_grad(): self.stacked_conv.weight.set_(torch.cat(single_unit_weights, dim=0)) def get_single_unit_weights(self, shape, c, j, h): Constructs and returns conv layer with traingin diabled and all zero weights except along the output channels for unit specified as (c, j, h). # Construct weight. weight = torch.zeros(self.weight.shape) # Set weights to zero except those specified. weight[:, c, j, h] = 1 return weight def update_connections_tensor(self, input_tensor, output_tensor): with torch.no_grad(): stacked_input = input_tensor.repeat((1, self.new_groups, 1, 1)) stacked_output = self.stacked_conv(stacked_input) s1 = torch.sigmoid(stacked_output).gt_(0.5) s2 = torch.sigmoid(output_tensor).gt_(0.5).repeat((1, self.new_groups, 1, 1)) H_ = torch.sum(s2.mul(s1), (0, 2, 3,)) self.connections_tensor[self.connection_indxs] = H_ def progress_connections(self): Prunes and add connections. with torch.no_grad(): # Get strengths of all connections. strengths = self.connections_tensor.numpy() shape = strengths.shape # Determine all combinations of prune dimensions all_dims = range(len(shape)) prune_indxs = [range(shape[d]) if d in self.prune_dims else [slice(None)] for d in all_dims] prune_indxs = itertools.product(*prune_indxs) # Along all combinations of prune dimensions: # - Keep strongest k1 connections # - Reinitilize trailing k2 - k1 connections. k1 = self.k1 k2 = self.k2 for idx in prune_indxs: # Get top k1'th strength. s = strengths[idx].flatten() v1 = np.partition(s, -k1)[-k1] # s.kthvalue(len(s) - k1).value # Keep top k1'th connection - prune those below c = self.weight[idx].flatten() prune_mask = (s < v1).astype(np.uint8) c[prune_mask] = 0 # Get trailing k2 - k1 connections. v2 = np.partition(s, -k2)[-k2] # s.kthvalue(len(s) - k2).value new_mask = (s > v2) & prune_mask # Reinitilized trailing k2 - k1 connections. # Note: [None, :] is added here as kaiming_uniform_ requires a 2d tensor. if len(c[new_mask]) > 0: torch.nn.init.kaiming_uniform_(c[new_mask][None, :]) # Reshape connections and update the weight. self.weight[idx] = c.reshape(self.weight[idx].shape) self.prune_mask = prune_mask # Reset connection strengths. self.connections_tensor = torch.zeros_like(self.weight) def prune_randomly(self): with torch.no_grad(): prune_mask = torch.rand(self.weight.shape) < 0.85 # prune 15% of weights self.weight[prune_mask] = 0 # Reinitialize those that are zero. keep_mask = ~prune_mask new_mask = (self.weight == 0) & keep_mask new_weights = self.weight[new_mask] if len(new_weights) > 0: torch.nn.init.kaiming_uniform_(new_weights[None, :]) self.weight[new_mask] = new_weights def __call__(self, input_tensor, *args, **kwargs): output_tensor = super().__call__(input_tensor, *args, **kwargs) if self.learning_iterations % 20 == 0: self.update_connections_tensor(input_tensor, output_tensor) self.learning_iterations += 1 return output_tensor Explanation: Sparse Conv Layer Now to implement a conv layer that utilizes the second implementation above. End of explanation root_path = '~/nta/datasets' transform = transforms.Compose( [transforms.ToTensor(), transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))]) trainset = torchvision.datasets.CIFAR10(root=root_path, train=True, download=True, transform=transform) trainloader = torch.utils.data.DataLoader(trainset, batch_size=4, shuffle=True, num_workers=2) testset = torchvision.datasets.CIFAR10(root=root_path, train=False, download=True, transform=transform) testloader = torch.utils.data.DataLoader(testset, batch_size=4, shuffle=False, num_workers=0) classes = ('plane', 'car', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck') Explanation: Test Training a Network The following is a simple toy example copied mostly from: https://pytorch.org/tutorials/beginner/blitz/cifar10_tutorial.html The main difference here is that the network utilizes the SparseCNN module. This exercise moslty servesto gain confidence in the implementation with respect to it's ability to run without errors - this is not concerned with verifying training improvements just yet. Load Data End of explanation cd = torch.nn.Conv2d(3, 6, 5) cs = DSConv2d(3, 6, 5) dataiter = iter(trainloader) images, labels = dataiter.next() print('Dense CNN foward pass:') %timeit cd(images) print('DSConv2d foward pass:') %timeit cs(images) Explanation: Time Trials Quick test to compare runtime with and without updating the connections tensor. End of explanation import torch.nn as nn import torch.nn.functional as F class Net(nn.Module): def __init__(self): super(Net, self).__init__() # self.conv1 = nn.Conv2d(3, 6, 5) self.conv1 = DSConv2d(3, 6, 5) self.pool = nn.MaxPool2d(2, 2) # self.conv2 = nn.Conv2d(6, 16, 5) self.conv2 = DSConv2d(6, 16, 5) self.fc1 = nn.Linear(16 * 5 * 5, 120) self.fc2 = nn.Linear(120, 84) self.fc3 = nn.Linear(84, 10) def forward(self, x): x = self.pool(F.relu(self.conv1(x))) x = self.pool(F.relu(self.conv2(x))) x = x.view(-1, 16 * 5 * 5) x = F.relu(self.fc1(x)) x = F.relu(self.fc2(x)) x = self.fc3(x) return x net = Net() Explanation: Network Setup End of explanation criterion = nn.CrossEntropyLoss() optimizer = optim.SGD(net.parameters(), lr=0.001, momentum=0.9) for epoch in range(2): # loop over the dataset multiple times running_loss = 0.0 for i, data in enumerate(trainloader, 0): # get the inputs; data is a list of [inputs, labels] inputs, labels = data # zero the parameter gradients optimizer.zero_grad() # forward + backward + optimize outputs = net(inputs) loss = criterion(outputs, labels) loss.backward() optimizer.step() # print statistics running_loss += loss.item() if i % 2000 == 1999: # print every 2000 mini-batches print('[%d, %5d] loss: %.3f' % (epoch + 1, i + 1, running_loss / 2000)) running_loss = 0.0 net.conv1.progress_connections() net.conv2.progress_connections() break break # # Compare with pruning random weights. # net.conv1.prune_randomly() # net.conv2.prune_randomly() print('Finished Training') Explanation: Training End of explanation class_correct = list(0. for i in range(10)) class_total = list(0. for i in range(10)) with torch.no_grad(): for data in testloader: images, labels = data outputs = net(images) _, predicted = torch.max(outputs, 1) c = (predicted == labels).squeeze() for i in range(4): label = labels[i] class_correct[label] += c[i].item() class_total[label] += 1 for i in range(10): print('Accuracy of %5s : %2d %%' % ( classes[i], 100 * class_correct[i] / class_total[i])) Explanation: Testing Accuracy End of explanation <END_TASK>
15,912
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Compute ICA components on epochs ICA is fit to MEG raw data. We assume that the non-stationary EOG artifacts have already been removed. The sources matching the ECG are automatically found and displayed. Subsequently, artefact detection and rejection quality are assessed. Finally, the impact on the evoked ERF is visualized. Note that this example does quite a bit of processing, so even on a fast machine it can take about a minute to complete. Step1: Fit ICA model using the FastICA algorithm, detect and inspect components
<ASSISTANT_TASK:> Python Code: # Authors: Denis Engemann <denis.engemann@gmail.com> # # License: BSD (3-clause) import mne from mne.preprocessing import ICA, create_ecg_epochs from mne.datasets import sample print(__doc__) Explanation: Compute ICA components on epochs ICA is fit to MEG raw data. We assume that the non-stationary EOG artifacts have already been removed. The sources matching the ECG are automatically found and displayed. Subsequently, artefact detection and rejection quality are assessed. Finally, the impact on the evoked ERF is visualized. Note that this example does quite a bit of processing, so even on a fast machine it can take about a minute to complete. End of explanation data_path = sample.data_path() raw_fname = data_path + '/MEG/sample/sample_audvis_filt-0-40_raw.fif' raw = mne.io.read_raw_fif(raw_fname, preload=True) raw.filter(1, 30, method='iir') raw.pick_types(meg=True, eeg=False, exclude='bads', stim=True) # longer + more epochs for more artifact exposure events = mne.find_events(raw, stim_channel='STI 014') epochs = mne.Epochs(raw, events, event_id=None, tmin=-0.2, tmax=0.5) ica = ICA(n_components=0.95, method='fastica').fit(epochs) ecg_epochs = create_ecg_epochs(raw, tmin=-.5, tmax=.5) ecg_inds, scores = ica.find_bads_ecg(ecg_epochs) ica.plot_components(ecg_inds) Explanation: Fit ICA model using the FastICA algorithm, detect and inspect components End of explanation <END_TASK>
15,913
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Class Session 4 Exercise Step1: Now, define a function that returns the index numbers of the neighbors of a vertex i, when the graph is stored in adjacency matrix format. So your function will accept as an input a NxN numpy matrix. Step2: Define a function that enumerates the neighbors of a vertex i, when the graph is stored in adjacency list format (a list of lists) Step3: Define a function that enumerates the neighbors of a vertex i, when the graph is stored in edge-list format (a numpy array of length-two-lists); use numpy.where and numpy.unique Step4: This next function is the simulation funtion. "n" is the number of vertices. It returns a length-three list containing the average running time for enumerating the neighbor vertices of a vertex in the graph. Step5: A simulation with 1000 vertices clearly shows that adjacency list is fastest Step6: We see the expected behavior, with the running time for the adjacency-matrix and edge-list formats going up when we increase "n", but there is hardly any change in the running time for the graph stored in adjacency list format
<ASSISTANT_TASK:> Python Code: import numpy as np import igraph import timeit import itertools Explanation: Class Session 4 Exercise: Comparing asymptotic running time for enumerating neighbors of all vertices in a graph We will measure the running time for enumerating the neighbor vertices for three different data structures for representing an undirected graph: adjacency matrix adjacency list edge list Let's assume that each vertex is labeled with a unique integer number. So if there are N vertices, the vertices are labeled 0, 2, 3, 4, ..., N-1. First, we will import all of the Python modules that we will need for this exercise: note how we assign a short name, "np" to the numpy module. This will save typing. End of explanation def enumerate_matrix(gmat, i): return np.nonzero(gmat[i,:])[1].tolist() Explanation: Now, define a function that returns the index numbers of the neighbors of a vertex i, when the graph is stored in adjacency matrix format. So your function will accept as an input a NxN numpy matrix. End of explanation def enumerate_adj_list(adj_list, i): return adj_list[i] Explanation: Define a function that enumerates the neighbors of a vertex i, when the graph is stored in adjacency list format (a list of lists): End of explanation def enumerate_edge_list(edge_list, i): inds1 = np.where(edge_list[:,0] == i)[0] elems1 = edge_list[inds1, 1].tolist() inds2 = np.where(edge_list[:,1] == i)[0] elems2 = edge_list[inds2, 0].tolist() return np.unique(elems1 + elems2).tolist() Explanation: Define a function that enumerates the neighbors of a vertex i, when the graph is stored in edge-list format (a numpy array of length-two-lists); use numpy.where and numpy.unique: End of explanation def do_sim(n): retlist = [] nrep = 10 nsubrep = 10 # this is (sort of) a Python way of doing the R function "replicate": for _ in itertools.repeat(None, nrep): # make a random undirected graph with fixed (average) vertex degree = 5 g = igraph.Graph.Barabasi(n, 5) # get the graph in three different representations g_matrix = np.matrix(g.get_adjacency().data) g_adj_list = g.get_adjlist() g_edge_list = np.array(g.get_edgelist()) start_time = timeit.default_timer() for _ in itertools.repeat(None, nsubrep): for i in range(0, n): enumerate_matrix(g_matrix, i) matrix_elapsed = timeit.default_timer() - start_time start_time = timeit.default_timer() for _ in itertools.repeat(None, nsubrep): for i in range(0, n): enumerate_adj_list(g_adj_list, i) adjlist_elapsed = timeit.default_timer() - start_time start_time = timeit.default_timer() for _ in itertools.repeat(None, nsubrep): for i in range(0, n): enumerate_edge_list(g_edge_list, i) edgelist_elapsed = timeit.default_timer() - start_time retlist.append([matrix_elapsed, adjlist_elapsed, edgelist_elapsed]) # average over replicates and then # divide by n so that the running time results are on a per-vertex basis return np.mean(np.array(retlist), axis=0)/n Explanation: This next function is the simulation funtion. "n" is the number of vertices. It returns a length-three list containing the average running time for enumerating the neighbor vertices of a vertex in the graph. End of explanation do_sim(1000)*1000 Explanation: A simulation with 1000 vertices clearly shows that adjacency list is fastest: (I multiply by 1000 just so the results are in ms.) End of explanation do_sim(2000)*1000 Explanation: We see the expected behavior, with the running time for the adjacency-matrix and edge-list formats going up when we increase "n", but there is hardly any change in the running time for the graph stored in adjacency list format: End of explanation <END_TASK>
15,914
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Recommender Engine Perhaps the most famous example of a recommender engine in the Data Science world was the Netflix competition started in 2006, in which teams from all around the world competed to improve on Netflix's reccomendation algorithm. The final prize of $1,000,000 was awarded to a team which developed a solution which had about a 10% increase in accuracy over Netflix's. In fact, this competition resulted in the development of some new techniques which are still in use. For more reading on this topic, see Simon Funk's blog post In this exercise, you will build a collaborative-filter recommendatin engine using both a cosine similarity approach and SVD (singular value decomposition). Before proceding download the MovieLens dataset. Importing and Pre-processing the Data First familiarize yourself with the data you downloaded, and then import the u.data file and take a look at the first few rows. Step1: Before building any recommendation engines, we'll have to get the data into a useful form. Do this by first splitting the data into testing and training sets, and then by constructing two new dataframes whose columns are each unique movie and rows are each unique user, filling in 0 for missing values. Step7: Now split the data into a training and test set, using a ratio 80/20 for train/test. Cosine Similarity Building a recommendation engine can be thought of as "filling in the holes" in the sparse matrix you made above. For example, take a look at the MovieLense data. You'll see that that matrix is mostly zeros. Our task here is to predict what a given user will rate a given movie depending on the users tastes. To determine a users taste, we can use cosine similarity which is given by $$s_u^{cos}(u_k,u_a) = \frac{ u_k \cdot u_a }{ \left \| u_k \right \| \left \| u_a \right \| } = \frac{ \sum x_{k,m}x_{a,m} }{ \sqrt{\sum x_{k,m}^2\sum x_{a,m}^2} }$$ for users $u_a$ and $u_k$ on ratings given by $x_{a,m}$ and $x_{b,m}$. This is just the cosine of the angle between the two vectors. Likewise, this can also be calculated for the similarity between two items, $i_a$ and $i_b$, given by $$s_u^{cos}(i_m,i_b) = \frac{ i_m \cdot i_b }{ \left \| i_m \right \| \left \| i_b \right \| } = \frac{ \sum x_{a,m} x_{a,b} }{ \sqrt{ \sum x_{a,m}^2 \sum x_{a,b}^2 } }$$ Then, the similarity between two users is given by $$\hat{x}{k,m} = \bar{x}{k} + \frac{\sum\limits_{u_a} s_u^{cos}(u_k,u_a) (x_{a,m})}{\sum\limits_{u_a}|s_u^{cos}(u_k,u_a)|}$$ and for items given by $$\hat{x}{k,m} = \frac{\sum\limits{i_b} s_u^{cos}(i_m,i_b) (x_{k,b}) }{\sum\limits_{i_b}|s_u^{cos}(i_m,i_b)|}$$ Use these ideas to construct a class cos_engine which can be used to recommend movies for a given user. Be sure to also test your algorithm, reporting its accuracy. Bonus Step13: SVD Above we used Cosine Similarity to fill the holes in our sparse matrix. Another, and much more popular, method for matrix completion is called a Singluar Value Decomposition. SVD factors our data matrix into three smaller matricies, given by $$\textbf{M} = \textbf{U} \bf{\Sigma} \textbf{V}^*$$ where $\textbf{M}$ is our data matrix, $\textbf{U}$ is a unitary matrix containg the latent variables in the user space, $\bf{\Sigma}$ is diagonal matrix containing the singular values of $\textbf{M}$, and $\textbf{V}$ is a unitary matrix containing the latent variables in the item space. For more information on the SVD see the Wikipedia article. Numpy contains a package to estimate the SVD of a sparse matrix. By making estimates of the matricies $\textbf{U}$, $\bf{\Sigma}$, and $\textbf{V}$, and then by multiplying them together, we can reconstruct an estimate for the matrix $\textbf{M}$ with all the holes filled in. Use these ideas to construct a class svd_engine which can be used to recommend movies for a given user. Be sure to also test your algorithm, reporting its accuracy. Bonus Step14: Overall RMSE of about 0.98 Step15: 7 is the optimal value of k in this case. Note that no cross-validation was performed! Now we'll build the best recommender and recommend 5 movies to each user.
<ASSISTANT_TASK:> Python Code: # Importing the data import pandas as pd import numpy as np header = ['user_id', 'item_id', 'rating', 'timestamp'] data_movie_raw = pd.read_csv('../data/ml-100k/u.data', sep='\t', names=header) data_movie_raw.head() Explanation: Recommender Engine Perhaps the most famous example of a recommender engine in the Data Science world was the Netflix competition started in 2006, in which teams from all around the world competed to improve on Netflix's reccomendation algorithm. The final prize of $1,000,000 was awarded to a team which developed a solution which had about a 10% increase in accuracy over Netflix's. In fact, this competition resulted in the development of some new techniques which are still in use. For more reading on this topic, see Simon Funk's blog post In this exercise, you will build a collaborative-filter recommendatin engine using both a cosine similarity approach and SVD (singular value decomposition). Before proceding download the MovieLens dataset. Importing and Pre-processing the Data First familiarize yourself with the data you downloaded, and then import the u.data file and take a look at the first few rows. End of explanation from sklearn.model_selection import train_test_split # First split into train and test sets data_train_raw, data_test_raw = train_test_split(data_movie_raw, train_size = 0.8) # Turning to pivot tables data_train = data_train_raw.pivot_table(index = 'user_id', columns = 'item_id', values = 'rating').fillna(0) data_test = data_test_raw.pivot_table(index = 'user_id', columns = 'item_id', values = 'rating').fillna(0) # Print the firest few rows data_train.head() Explanation: Before building any recommendation engines, we'll have to get the data into a useful form. Do this by first splitting the data into testing and training sets, and then by constructing two new dataframes whose columns are each unique movie and rows are each unique user, filling in 0 for missing values. End of explanation # Libraries import pandas as pd import numpy as np from sklearn.metrics.pairwise import cosine_similarity class cos_engine: def __init__(self, data_all): Constructor for cos_engine class Args: data_all: Raw dataset containing all movies to build a list of movies already seen by each user. # Create copy of data self.data_all = data_all.copy() # Now build a list of movies each of you has seen self.seen = [] for user in data_all.user_id.unique(): cur_seen = {} cur_seen["user"] = user cur_seen["seen"] = self.data_all[data_all.user_id == user].item_id self.seen.append(cur_seen) def fit(self, data_train): Performs cosine similarity on a sparse matrix data_train Args: data_train: A pandas data frame data to estimate cosine similarity # Create a copy of the dataframe self.data_train = data_train.copy() # Save the indices and column names self.users = self.data_train.index self.items = self.data_train.columns # Compute mean vectors self.user_means = self.data_train.replace(0, np.nan).mean(axis = 1) self.item_means = self.data_train.T.replace(0, np.nan).mean(axis = 1) # Get similarity matrices and compute sums for normalization # For non adjusted cosine similarity, neglect subtracting the means. self.data_train_adj = (self.data_train.replace(0, np.nan).T - self.user_means).fillna(0).T self.user_cos = cosine_similarity(self.data_train_adj) self.item_cos = cosine_similarity(self.data_train_adj.T) self.user_cos_sum = np.abs(self.user_cos).sum(axis = 1) self.item_cos_sum = np.abs(self.item_cos).sum(axis = 1) self.user_cos_sum = self.user_cos_sum.reshape(self.user_cos_sum.shape[0], 1) self.item_cos_sum = self.item_cos_sum.reshape(self.item_cos_sum.shape[0], 1) def predict(self, method = 'user'): Predicts using Cosine Similarity Args: method: A string indicating what method to use, user or item. Default user. Returns: A pandas dataframe containing the prediction values # Store prediction locally and turn to dataframe if method == 'user': self.pred = self.user_means[:, np.newaxis] + ((self.user_cos @ self.data_train_adj) / self.user_cos_sum) self.pred = pd.DataFrame(self.pred, index = data_train.index, columns = data_train.columns) elif method == 'item': self.pred = self.item_means[:, np.newaxis] + ((self.data_train @ self.item_cos) / self.item_cos_sum.T).T self.pred = pd.DataFrame(self.pred, columns = data_train.index.values, index = data_train.columns) return(self.pred) def test(self, data_test, root = False): Tests fit given test data in data_test Args: data_test: A pandas dataframe containing test data root: A boolean indicating whether to return the RMSE. Default False Returns: The Mean Squared Error of the fit if root = False, the Root Mean\ Squared Error otherwise. # Build a list of common indices (users) in the train and test set row_idx = list(set(self.pred.index) & set(data_test.index)) # Prime the variables for loop err = [] # To hold the Sum of Squared Errors N = 0 # To count preditions for MSE calculation for row in row_idx: # Get the rows test_row = data_test.loc[row, :] pred_row = self.pred.loc[row, :] # Get indices of nonzero elements in the test set idx = test_row.index[test_row.nonzero()[0]] # Get only common movies temp_test = test_row[idx] temp_pred = pred_row[idx] # Compute error and count temp_err = ((temp_pred - temp_test)**2).sum() N = N + len(idx) err.append(temp_err) mse = np.sum(err) / N # Switch for RMSE if root: err = np.sqrt(mse) else: err = mse return(err) def recommend(self, user, num_recs): Tests fit given test data in data_test Args: data_test: A pandas dataframe containing test data root: A boolean indicating whether to return the RMSE. Default False Returns: The Mean Squared Error of the fit if root = False, the Root Mean Squared Error otherwise. # Get list of already seen movies for this user idx_seen = next(item for item in self.seen if item["user"] == 2)["seen"] # Remove already seen movies and recommend rec = self.pred.loc[user, :].drop(idx_seen).nlargest(num_recs) return(rec.index) # Testing cos_en = cos_engine(data_movie_raw) cos_en.fit(data_train) # Predict using user similarity pred1 = cos_en.predict(method = 'user') err = cos_en.test(data_test, root = True) rec1 = cos_en.recommend(1, 5) print("RMSE:", err) print("Reccomendations for user 1:", rec1.values) # And now with item pred2 = cos_en.predict(method = 'item') err = cos_en.test(data_test, root = True) rec2 = cos_en.recommend(1, 5) print("RMSE:", err) print("Reccomendations for item 1:", rec2.values) Explanation: Now split the data into a training and test set, using a ratio 80/20 for train/test. Cosine Similarity Building a recommendation engine can be thought of as "filling in the holes" in the sparse matrix you made above. For example, take a look at the MovieLense data. You'll see that that matrix is mostly zeros. Our task here is to predict what a given user will rate a given movie depending on the users tastes. To determine a users taste, we can use cosine similarity which is given by $$s_u^{cos}(u_k,u_a) = \frac{ u_k \cdot u_a }{ \left \| u_k \right \| \left \| u_a \right \| } = \frac{ \sum x_{k,m}x_{a,m} }{ \sqrt{\sum x_{k,m}^2\sum x_{a,m}^2} }$$ for users $u_a$ and $u_k$ on ratings given by $x_{a,m}$ and $x_{b,m}$. This is just the cosine of the angle between the two vectors. Likewise, this can also be calculated for the similarity between two items, $i_a$ and $i_b$, given by $$s_u^{cos}(i_m,i_b) = \frac{ i_m \cdot i_b }{ \left \| i_m \right \| \left \| i_b \right \| } = \frac{ \sum x_{a,m} x_{a,b} }{ \sqrt{ \sum x_{a,m}^2 \sum x_{a,b}^2 } }$$ Then, the similarity between two users is given by $$\hat{x}{k,m} = \bar{x}{k} + \frac{\sum\limits_{u_a} s_u^{cos}(u_k,u_a) (x_{a,m})}{\sum\limits_{u_a}|s_u^{cos}(u_k,u_a)|}$$ and for items given by $$\hat{x}{k,m} = \frac{\sum\limits{i_b} s_u^{cos}(i_m,i_b) (x_{k,b}) }{\sum\limits_{i_b}|s_u^{cos}(i_m,i_b)|}$$ Use these ideas to construct a class cos_engine which can be used to recommend movies for a given user. Be sure to also test your algorithm, reporting its accuracy. Bonus: Use adjusted cosine similiarity. End of explanation # Libraries import pandas as pd import numpy as np import scipy.sparse as sp from scipy.sparse.linalg import svds class svd_engine: def __init__(self, data_all, k = 6): Constructor for svd_engine class Args: k: The number of latent variables to fit self.k = k # Create copy of data self.data_all = data_all.copy() # Now build a list of movies each you has seen self.seen = [] for user in data_all.user_id.unique(): cur_seen = {} cur_seen["user"] = user cur_seen["seen"] = self.data_all[data_all.user_id == user].item_id self.seen.append(cur_seen) def fit(self, data_train): Performs SVD on a sparse matrix data_train Args: data_train: A pandas data frame data to estimate SVD Returns: Matricies u, s, and vt of SVD # Save local copy of data self.data_train = data_train.copy() # Compute adjusted matrix self.user_means = self.data_train.replace(0, np.nan).mean(axis = 1) self.item_means = self.data_train.T.replace(0, np.nan).mean(axis = 1) self.data_train_adj = (self.data_train.replace(0, np.nan).T - self.user_means).fillna(0).T # Save the indices and column names self.users = data_train.index self.items = data_train.columns # Train the model self.u, self.s, self.vt = svds(self.data_train_adj, k = self.k) return(self.u, np.diag(self.s), self.vt) def predict(self): Predicts using SVD Returns: A pandas dataframe containing the prediction values # Store prediction locally and turn to dataframe, adding the mean back self.pred = pd.DataFrame(self.u @ np.diag(self.s) @ self.vt, index = self.users, columns = self.items) self.pred = self.user_means[:, np.newaxis] + self.pred return(self.pred) def test(self, data_test, root = False): Tests fit given test data in data_test Args: data_test: A pandas dataframe containing test data root: A boolean indicating whether to return the RMSE. Default False Returns: The Mean Squared Error of the fit if root = False, the Root Mean\ Squared Error otherwise. # Build a list of common indices (users) in the train and test set row_idx = list(set(self.pred.index) & set(data_test.index)) # Prime the variables for loop err = [] # To hold the Sum of Squared Errors N = 0 # To count predictions for MSE calculation for row in row_idx: # Get the rows test_row = data_test.loc[row, :] pred_row = self.pred.loc[row, :] # Get indices of nonzero elements in the test set idx = test_row.index[test_row.nonzero()[0]] # Get only common movies temp_test = test_row[idx] temp_pred = pred_row[idx] # Compute error and count temp_err = ((temp_pred - temp_test)**2).sum() N = N + len(idx) err.append(temp_err) mse = np.sum(err) / N # Switch for RMSE if root: err = np.sqrt(mse) else: err = mse return(err) def recommend(self, user, num_recs): Tests fit given test data in data_test Args: data_test: A pandas dataframe containing test data root: A boolean indicating whether to return the RMSE. Default False Returns: The Mean Squared Error of the fit if root = False, the Root Mean\ Squared Error otherwise. # Get list of already seen movies for this user idx_seen = next(item for item in self.seen if item["user"] == 2)["seen"] # Remove already seen movies and recommend rec = self.pred.loc[user, :].drop(idx_seen).nlargest(num_recs) return(rec.index) # Testing svd_en = svd_engine(data_movie_raw, k = 20) svd_en.fit(data_train) svd_en.predict() err = svd_en.test(data_test, root = True) rec = svd_en.recommend(1, 5) print("RMSE:", err) print("Reccomendations for user 1:", rec.values) Explanation: SVD Above we used Cosine Similarity to fill the holes in our sparse matrix. Another, and much more popular, method for matrix completion is called a Singluar Value Decomposition. SVD factors our data matrix into three smaller matricies, given by $$\textbf{M} = \textbf{U} \bf{\Sigma} \textbf{V}^*$$ where $\textbf{M}$ is our data matrix, $\textbf{U}$ is a unitary matrix containg the latent variables in the user space, $\bf{\Sigma}$ is diagonal matrix containing the singular values of $\textbf{M}$, and $\textbf{V}$ is a unitary matrix containing the latent variables in the item space. For more information on the SVD see the Wikipedia article. Numpy contains a package to estimate the SVD of a sparse matrix. By making estimates of the matricies $\textbf{U}$, $\bf{\Sigma}$, and $\textbf{V}$, and then by multiplying them together, we can reconstruct an estimate for the matrix $\textbf{M}$ with all the holes filled in. Use these ideas to construct a class svd_engine which can be used to recommend movies for a given user. Be sure to also test your algorithm, reporting its accuracy. Bonus: Tune any parameters. End of explanation # Parameter tuning import matplotlib.pyplot as plt err = [] for cur_k in range(1, 50): svd_en = svd_engine(data_movie_raw, k = cur_k) svd_en.fit(data_train) svd_en.predict() err.append(svd_en.test(data_test)) plt.plot(range(1, 50), err) plt.title('RMSE versus k') plt.xlabel('k') plt.ylabel('RMSE') plt.show() err.index(min(err)) Explanation: Overall RMSE of about 0.98 End of explanation # Build the engine svd_en = svd_engine(data_movie_raw, k = 7) svd_en.fit(data_train) svd_en.predict() # Now make recommendations recs = [] for user in data_movie_raw.user_id.unique(): temp_rec = svd_en.recommend(user, 5) recs.append(temp_rec) recs[0] Explanation: 7 is the optimal value of k in this case. Note that no cross-validation was performed! Now we'll build the best recommender and recommend 5 movies to each user. End of explanation <END_TASK>
15,915
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Unsupervised Analysis of Days of Week Treating crossing each day as features to learn about the relatinships between various days. Step1: Downloading Data We'll start by downloading the data (available on seattle.gov). Step2: Principal Component Analysis Step3: Comparing with Day of Week
<ASSISTANT_TASK:> Python Code: %matplotlib inline import matplotlib.pyplot as plt Explanation: Unsupervised Analysis of Days of Week Treating crossing each day as features to learn about the relatinships between various days. End of explanation from urllib import request FREMONT_URL = 'https://data.seattle.gov/api/views/65db-xm6k/rows.csv?accessType=DOWNLOAD' request.urlretrieve(FREMONT_URL, 'Fremont.csv') # magic function to show the content of the file %more Fremont.csv import pandas as pd df = pd.read_csv('Fremont.csv') # use read_csv to load the data into dataframe df.head() # Let's see the type of the data df.dtypes # change the Date column to datetime data type df['Date'] = pd.to_datetime(df['Date']) df.head() df.dtypes # Set the index to Date df.set_index('Date', inplace=True) df.head() df.apply(lambda x: sum(x.isnull())) # clear the data by delete the non-numeric df.dropna(inplace=True) df.apply(lambda x: sum(x.isnull())) df.columns df.plot() df.resample('W').sum().plot() df.columns=['West', 'East'] df.resample('w').sum().plot() # To see whether there is any annual trend of the number of rides df.resample('D').sum().rolling(365).sum().plot() # each point is the sum of the number of rides in the previuos 365 days # The y coordinate is not from 0 ax = df.resample('D').sum().rolling(365).sum().plot() ax.set_ylim(0, None) # DateimeIndex.time return numpy array of datetime.time, the time part of the Timestamps df.groupby(df.index.time).mean().plot() # plot the average of rides at each hours of the day # Create the pivoted table to investigate the pattern in each day df['Total'] = df['West'] + df['East'] pivoted = df.pivot_table(values='Total', index=df.index.time, columns=df.index.date) pivoted.head() pivoted.shape # delete the date with non-numeric pivoted.dropna(axis=1, inplace=True) pivoted.shape pivoted.plot(legend=False) # add transparent parameter alpha pivoted.plot(legend=False, alpha=0.01) Explanation: Downloading Data We'll start by downloading the data (available on seattle.gov). End of explanation # Get X with hours as mearsurement and date as observations X = pivoted.T.values X.shape X from sklearn.decomposition import PCA X2 = PCA(2, svd_solver='full').fit_transform(X) X2 X2.shape plt.scatter(X2[:, 0], X2[:, 1]) # use cluster algorithm Gaussian mixture model from sklearn.mixture import GaussianMixture gmm = GaussianMixture(2) gmm.fit(X) labels = gmm.predict(X) labels # plt.scatter(X2[:, 0], X2[:, 1], c=labels, cmap='rainbow') # plt.colorbar() plt.scatter(X2[:, 0], X2[:, 1], c=labels) plt.colorbar() labels # so labels == 1 represents the weekday pivoted.T[labels == 1].T.plot(legend=False, alpha=0.01) # labels == 0 represents the weekend or holiday pivoted.T[labels == 0].T.plot(legend=False, alpha=0.1) Explanation: Principal Component Analysis End of explanation pd.DatetimeIndex(pivoted.columns) # The DatetimeIndex.dayof week gives the day of the week dayofweek = pd.DatetimeIndex(pivoted.columns).dayofweek dayofweek # Then we plot the color of the weekday plt.scatter(X2[:, 0], X2[:, 1], c=dayofweek) plt.colorbar() # grab the day in label 0 which is not weekend dates = pd.DatetimeIndex(pivoted.columns) dates[(labels == 0) & (dayofweek < 5)] Explanation: Comparing with Day of Week End of explanation <END_TASK>
15,916
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: OCCT Bottle Tutorial Step1: VTK Viewer The following summarizes the mouse and keyboard commands for interacting with shapes rendered in the viewer. Keypress j / Keypress t
<ASSISTANT_TASK:> Python Code: height = 70. width = 50. thickness = 30. pnt1 = [-width/2., 0., 0.] pnt2 = [-width/2., -thickness/4., 0.] pnt3 = [0., -thickness/2., 0.] pnt4 = [width/2., -thickness/4., 0.] pnt5 = [width/2., 0., 0.] edge1 = Edge().createLine(start=pnt1, end=pnt2) edge2 = Edge().createArc3P(start=pnt2, end=pnt4, pnt=pnt3) edge3 = Edge().createLine(start=pnt4, end=pnt5) halfProfile = Wire([edge1, edge2, edge3]) mirrorPlane = Plane(origin=[0,0,0], xaxis=[1,0,0], yaxis=[0,0,1]) mirrorProfile = halfProfile.mirror(mirrorPlane, copy=True) allEdges = list(EdgeIterator(halfProfile)) + list(EdgeIterator(mirrorProfile)) fullProfile = Wire().createWire(allEdges) bottomFace = Face().createFace(fullProfile) body = Solid().extrude(bottomFace, (0, 0, 0), (0, 0, height)) body.fillet(thickness/12.) neckHeight = height/10 neckRadius = thickness/4 neck = Solid().createCylinder([0,0,0], [0,0,neckHeight], radius=neckRadius) neck.translate([0, 0, height]) body.fuse(neck) zMax = -1 neckTopFace = None for f in FaceIterator(body): [x, y , z] = f.centreOfMass() if z >= zMax: neckTopFace = f zMax = z body.shell(thickness/50., [neckTopFace], tolerance=1E-3) t_thick = neckHeight/5 t_height = neckHeight - t_thick t_radius = neckRadius + t_thick/4 t_pitch = t_height/2 t_angle = 0 # Note the following thread geometry is not correct. The profile # is wrong and there is a twist added to the profile. But it's # kind of close and good enough for this example. threadHelix = Edge().createHelix(pitch=t_pitch, height=t_height, radius=t_radius, angle = t_angle) threadFace = Face().createPolygonal([[0, 0, t_thick/2], [t_thick, .0, 0], [0, 0, -t_thick/2]]) threadFace.translate([t_radius, 0, 0]) thread = Solid().pipe(threadFace, threadHelix) thread.translate([0, 0, height]) body.fuse(thread) actor = body.toVtkActor() Explanation: OCCT Bottle Tutorial End of explanation try: a = get_QApplication([]) except: pass vtkWin = SimpleVtkViewer() vtkWin.add_actor(actor) # If the VTK window is blank/white, click on the window and hit 'r' to zoom to fit. Explanation: VTK Viewer The following summarizes the mouse and keyboard commands for interacting with shapes rendered in the viewer. Keypress j / Keypress t: toggle between joystick (position sensitive) and trackball (motion sensitive) styles. In joystick style, motion occurs continuously as long as a mouse button is pressed. In trackball style, motion occurs when the mouse button is pressed and the mouse pointer moves. Keypress c / Keypress a: toggle between camera and actor modes. In camera mode, mouse events affect the camera position and focal point. In actor mode, mouse events affect the actor that is under the mouse pointer. Button 1: rotate the camera around its focal point (if camera mode) or rotate the actor around its origin (if actor mode). The rotation is in the direction defined from the center of the renderer's viewport towards the mouse position. In joystick mode, the magnitude of the rotation is determined by the distance the mouse is from the center of the render window. Button 2: pan the camera (if camera mode) or translate the actor (if object mode). In joystick mode, the direction of pan or translation is from the center of the viewport towards the mouse position. In trackball mode, the direction of motion is the direction the mouse moves. (Note: with 2-button mice, pan is defined as &lt;Shift&gt;Button 1.) Button 3: zoom the camera (if camera mode) or scale the actor (if object mode). Zoom in/increase scale if the mouse position is in the top half of the viewport; zoom out/decrease scale if the mouse position is in the bottom half. In joystick mode, the amount of zoom is controlled by the distance of the mouse pointer from the horizontal centerline of the window. Keypress 3: toggle the render window into and out of stereo mode. By default, red-blue stereo pairs are created. Some systems support Crystal Eyes LCD stereo glasses; you have to invoke SetStereoTypeToCrystalEyes() on the rendering window object. Note: to use stereo you also need to pass a stereo=1 keyword argument to the render window object constructor. Keypress e: exit the application. Keypress f: fly to the picked point Keypress p: perform a pick operation. The render window interactor has an internal instance of vtkCellPicker that it uses to pick. Keypress r: reset the camera view along the current view direction. Centers the actors and moves the camera so that all actors are visible. Keypress s: modify the representation of all actors so that they are surfaces. Keypress u: invoke the user-defined function. Typically, this keypress will bring up an interactor that you can type commands in. Keypress w: modify the representation of all actors so that they are wireframe. End of explanation <END_TASK>
15,917
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Wind energy production forecast from Met.no weather forecast With this notebook we illustrate how one might improve weather forecast for wind energy production, considering that the height of wind turbines doesn’t match the height of wind speed commonly used in weather forecasts. Also, that wind energy production does not depend only on wind speed, but also air density. In this Notebook we will use three weather forecast from Planet OS Datahub Step1: <font color='red'>Please put your datahub API key into a file called APIKEY and place it to the notebook folder or assign your API key directly to the variable API_key!</font> Step2: We use a simple dh_py_access library for controlling REST requests. Step3: dh_py_access provides some easy to use functions to list dataset metadata Step4: Initialize coordinates and reftime. Please note that more recent reftime, than computed below, may be available for particular dataset. We just use conservative example for demo Step5: Fetch data and convert to Pandas dataframe Step6: Filter out necessary data from the DataFrames Step7: We easily see that differences between models can be larger than difference of 10m to 80m winds in the same model. What role does density play in energy production? From the theory we know that wind energy production is roughly $1/2 A \rho \mathbf{v}^3$, where $A$ is area, $\rho$ is air density and $\mathbf{v}$ is wind speed. We are not concerned about $A$, which is a turbine parameter, but we can analyse the linear relation of density and cube relation of wind speed itself. First, let's see how the density varies over time Step8: No let's energy production looks compared to the wind speed Step9: Finally, let's analyse how much density changes vary over the whole domain during one forecast. For this purpose, we download the whole density field with package api Step10: Maximum relative change of air density in single location is
<ASSISTANT_TASK:> Python Code: %matplotlib notebook import urllib.request import numpy as np import simplejson as json import matplotlib.pyplot as plt import matplotlib.dates as mdates import warnings import datetime import dateutil.parser import matplotlib.cbook warnings.filterwarnings("ignore",category=matplotlib.cbook.mplDeprecation) import requests from netCDF4 import Dataset from dh_py_access.lib.dataset import dataset as dataset import dh_py_access.package_api as package_api import dh_py_access.lib.datahub as datahub np.warnings.filterwarnings('ignore') Explanation: Wind energy production forecast from Met.no weather forecast With this notebook we illustrate how one might improve weather forecast for wind energy production, considering that the height of wind turbines doesn’t match the height of wind speed commonly used in weather forecasts. Also, that wind energy production does not depend only on wind speed, but also air density. In this Notebook we will use three weather forecast from Planet OS Datahub: GFS - A global model produced by NOAA that provides a 10-day forecast at a resolution of 0.25 degrees (15 days with lower resolution and fewer variables); FMI HIRLAM - A regional model limited to Northern-Europe and Northern-Atlantic with a forecast period of two days and resolution of 0.07 degrees; Met.no HARMONIE & Met.no HARMONIE dedicated wind forecast A regional model limited to Scandinavia with a resolution of 0.025 x 0.05 degrees. The dedicated wind forecast version is augmented by Planet OS to include the air density as explicit variable. API documentation is available at http://docs.planetos.com. If you have questions or comments, join the Planet OS Slack community to chat with our development team. For general information on usage of IPython/Jupyter and Matplotlib, please refer to their corresponding documentation. https://ipython.org/ and http://matplotlib.org/ End of explanation API_key = open("APIKEY").read().strip() dh=datahub.datahub_main(API_key) Explanation: <font color='red'>Please put your datahub API key into a file called APIKEY and place it to the notebook folder or assign your API key directly to the variable API_key!</font> End of explanation fmi_hirlam_surface = dataset('fmi_hirlam_surface',dh) metno_harmonie_metcoop = dataset('metno_harmonie_metcoop',dh) metno_harmonie_wind = dataset('metno_harmonie_wind_det',dh) gfs = dataset('noaa_gfs_pgrb2_global_forecast_recompute_0.25degree',dh) Explanation: We use a simple dh_py_access library for controlling REST requests. End of explanation ## Does not look good in github ##gfs.variable_names() Explanation: dh_py_access provides some easy to use functions to list dataset metadata End of explanation lon = 22 lat = 59+6./60 today = datetime.datetime.today() reft = datetime.datetime(today.year,today.month,today.day,int(today.hour/6)*6) - datetime.timedelta(hours=12) reft = reft.isoformat() ##reft = "2018-02-11T18:00:00" arg_dict = {'lon':lon,'lat':lat,'reftime_start':reft,'reftime_end':reft,'count':250} arg_dict_metno_wind_det = dict(arg_dict, **{'vars':'wind_u_z,wind_v_z,air_density_z'}) arg_dict_metno_harm_metcoop = dict(arg_dict, **{'vars':'u_wind_10m,v_wind_10m'}) arg_dict_hirlam = dict(arg_dict, **{'vars':'u-component_of_wind_height_above_ground,v-component_of_wind_height_above_ground'}) arg_dict_gfs = dict(arg_dict, **{'vars':'ugrd_m,vgrd_m','count':450}) Explanation: Initialize coordinates and reftime. Please note that more recent reftime, than computed below, may be available for particular dataset. We just use conservative example for demo End of explanation dmw = metno_harmonie_wind.get_json_data_in_pandas(**arg_dict_metno_wind_det) dmm = metno_harmonie_metcoop.get_json_data_in_pandas(**arg_dict_metno_harm_metcoop) dhs = fmi_hirlam_surface.get_json_data_in_pandas(**arg_dict_hirlam) dgfs = gfs.get_json_data_in_pandas(**arg_dict_gfs) ## show how to filter Pandas ## dgfs[dgfs['z']==80] Explanation: Fetch data and convert to Pandas dataframe End of explanation vel80_metno = np.array(np.sqrt(dmw[dmw['z']==80]['wind_u_z']**2 + dmw[dmw['z']==80]['wind_v_z']**2)) vel10_metno = np.array(np.sqrt(dmm['u_wind_10m']**2 + dmm['v_wind_10m']**2)) vel10_hirlam = np.array(np.sqrt(dhs['u-component_of_wind_height_above_ground']**2 + dhs['v-component_of_wind_height_above_ground']**2)) vel10_gfs = np.sqrt(dgfs[dgfs['z']==10]['ugrd_m']**2+dgfs[dgfs['z']==10]['vgrd_m']**2) vel80_gfs = np.sqrt(dgfs[dgfs['z']==80]['ugrd_m']**2+dgfs[dgfs['z']==80]['vgrd_m']**2) t_metno = [dateutil.parser.parse(i) for i in dmw[dmw['z']==80]['time']] t_metno_10 = [dateutil.parser.parse(i) for i in dmm['time']] t_hirlam = [dateutil.parser.parse(i) for i in dhs['time']] t_gfs_10 = [dateutil.parser.parse(i) for i in dgfs[dgfs['z']==10]['time']] t_gfs_80 = [dateutil.parser.parse(i) for i in dgfs[dgfs['z']==80]['time']] fig, ax = plt.subplots() days = mdates.DayLocator() daysFmt = mdates.DateFormatter('%Y-%m-%d') hours = mdates.HourLocator() ax.set_ylabel("wind speed") ax.plot(t_metno, vel80_metno, label='Metno 80m') ax.plot(t_metno_10, vel10_metno, label='Metno 10m') ax.plot(t_hirlam, vel10_hirlam, label='HIRLAM 10m') gfs_lim=67 ax.plot(t_gfs_10[:gfs_lim], vel10_gfs[:gfs_lim], label='GFS 10m') ax.plot(t_gfs_80[:gfs_lim], vel80_gfs[:gfs_lim], label='GFS 80m') ax.xaxis.set_major_locator(days) ax.xaxis.set_major_formatter(daysFmt) ax.xaxis.set_minor_locator(hours) #fig.autofmt_xdate() plt.legend() plt.grid() plt.savefig("model_comp") Explanation: Filter out necessary data from the DataFrames End of explanation fig, ax = plt.subplots() ax2 = ax.twinx() days = mdates.DayLocator() daysFmt = mdates.DateFormatter('%Y-%m-%d') hours = mdates.HourLocator() ax.plot(t_metno,vel80_metno) aird80 = dmw[dmw['z']==80]['air_density_z'] ax2.plot(t_metno,aird80,c='g') ax.xaxis.set_major_locator(days) ax.xaxis.set_major_formatter(daysFmt) ax.xaxis.set_minor_locator(hours) ax.set_ylabel("wind speed") ax2.set_ylabel("air density") fig.tight_layout() #fig.autofmt_xdate() plt.savefig("density_80m") Explanation: We easily see that differences between models can be larger than difference of 10m to 80m winds in the same model. What role does density play in energy production? From the theory we know that wind energy production is roughly $1/2 A \rho \mathbf{v}^3$, where $A$ is area, $\rho$ is air density and $\mathbf{v}$ is wind speed. We are not concerned about $A$, which is a turbine parameter, but we can analyse the linear relation of density and cube relation of wind speed itself. First, let's see how the density varies over time End of explanation fig, ax = plt.subplots() ax2 = ax.twinx() ax2.set_ylabel("energy production") ax.set_ylabel("wind speed") ax.plot(t_metno,vel80_metno, c='b', label='wind speed') ax2.plot(t_metno,aird80*vel80_metno**3, c='r', label='energy prod w.dens') ax.xaxis.set_major_locator(days) ax.xaxis.set_major_formatter(daysFmt) ax.xaxis.set_minor_locator(hours) fig.autofmt_xdate() lines, labels = ax.get_legend_handles_labels() lines2, labels2 = ax2.get_legend_handles_labels() ax2.legend(lines + lines2, labels + labels2, loc=0) Explanation: No let's energy production looks compared to the wind speed End of explanation density = package_api.package_api(dh,'metno_harmonie_wind_det','air_density_z',-20,60,10,80,'full_domain_harmonie') density.make_package() density.download_package() density_data = Dataset(density.get_local_file_name()) ## biggest change of density in one location during forecast period maxval = np.nanmax(density_data.variables['air_density_z'],axis=0) minval = np.nanmin(density_data.variables['air_density_z'],axis=0) Explanation: Finally, let's analyse how much density changes vary over the whole domain during one forecast. For this purpose, we download the whole density field with package api End of explanation print(np.nanmax(maxval-minval),np.nanmean(maxval-minval)) Explanation: Maximum relative change of air density in single location is: End of explanation <END_TASK>
15,918
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: 596 Step1: Discovering available alignments Alignments for a specified transcript This approach identifies available alignments and then selects the desired one to use with VariantMapper Step2: Alignments for a gene Step3: Alignments for a genomic region (new method) The hgvs data provider method get_tx_for_region used a supplied alignment method, which necessarily filtered alignments for a specific alignment method. A new method, get_alignments_for_region provides similar functionality with an optional alt_aln_method parameter. When alt_aln_method is None (or not provided), all alignments are returned. When alt_aln_method is provided, it behaves exactly like get_tx_for_region. Step4: Alternate method for transcript-to-genome projections
<ASSISTANT_TASK:> Python Code: from hgvs.easy import (__version__, parser, hdp, vm) from hgvs.exceptions import HGVSDataNotAvailableError __version__ # hgvs_g = "NC_000010.11:g.94762693G>A" # GRCh38 hgvs_g = "NC_000010.10:g.96522450G>A" # GRCh37 hgvs_c = "NM_000769.4:c.-13G>A" var_c = parser.parse(hgvs_c) var_g = parser.parse(hgvs_g) Explanation: 596: Support for custom transcripts See https://github.com/biocommons/hgvs/issues/596 UTA#220 added support for custom transcripts, which really means custom alignments associated with a new alt_aln_method, "splign-manual". The loading/data/splign-manual contains examples. Although hgvs supports multiple alignment methods already (e.g., one can specify "blat", "splign" in VariantMapper::g_to_c()), several areas of code assume one default alt_aln_method. This work was sponsored by Invitae, which uses the VariantMapper interface. Therefore, this new feature focuses on minimal changes to extend hgvs to better support using multiple alignment methods with VariantMapper. These are demonstrated below. Setup Initialize hgvs and create two variants from CYP2C19 to use as examples (VCV000634882.1) End of explanation hdp.get_tx_mapping_options(var_c.ac) # or, for a more complete example with many options: hdp.get_tx_mapping_options("NM_001807.4") Explanation: Discovering available alignments Alignments for a specified transcript This approach identifies available alignments and then selects the desired one to use with VariantMapper::c_to_g() as above. End of explanation alignments = hdp.get_tx_for_gene("CYP2C19") alignments.sort(key=lambda a: (a["tx_ac"], a["alt_ac"], a["alt_aln_method"])) alignments Explanation: Alignments for a gene End of explanation start = var_g.posedit.pos.start.base end = var_g.posedit.pos.end.base alignments = hdp.get_alignments_for_region(var_g.ac, start, end) alignments.sort(key=lambda a: (a["tx_ac"], a["alt_ac"], a["alt_aln_method"])) alignments Explanation: Alignments for a genomic region (new method) The hgvs data provider method get_tx_for_region used a supplied alignment method, which necessarily filtered alignments for a specific alignment method. A new method, get_alignments_for_region provides similar functionality with an optional alt_aln_method parameter. When alt_aln_method is None (or not provided), all alignments are returned. When alt_aln_method is provided, it behaves exactly like get_tx_for_region. End of explanation try: vm.c_to_g(var_c, "NC_000010.10") except HGVSDataNotAvailableError as e: print(f"Got {e!r}") vm.c_to_g(var_c, "NC_000010.10", alt_aln_method="splign-manual") Explanation: Alternate method for transcript-to-genome projections: Using try...except This approach follows the easier-to-ask-for-forgiveness-than-permission principle. Code tries "splign", which is expected to satisfy the majority of cases, and then falls back to "splign-manual" for failures. The advantage of this method is that there is only one database fetch for the most common case. End of explanation <END_TASK>
15,919
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: The decimal module implements fixed and floating point arithmetic using the model familiar to most people, rather than the IEEE floating point version implemented by most computer hardware and familiar to programmers. A Decimal instance can represent any number exactly, round up or down, and apply a limit to the number of significant digits. Decimal Step1: Decimals can also be created from tuples containing a sign flag Step2: Formatting Step3: Arithmetic Step4: Special Value Step5: Context Step6: Precision Step7: Local Context
<ASSISTANT_TASK:> Python Code: import decimal fmt = '{0:<25}{1:<25}' print(fmt.format('Input', 'Output')) print(fmt.format('-'*25, '-'*25)) #Integer print(fmt.format(5, decimal.Decimal(5))) #String print(fmt.format('3.14', decimal.Decimal('3.14'))) #Float f = 0.1 print(fmt.format(repr(f), decimal.Decimal(str(f)))) print('{:0.23g}{:<25}'.format(f, str(decimal.Decimal.from_float(f))[:25])) Explanation: The decimal module implements fixed and floating point arithmetic using the model familiar to most people, rather than the IEEE floating point version implemented by most computer hardware and familiar to programmers. A Decimal instance can represent any number exactly, round up or down, and apply a limit to the number of significant digits. Decimal End of explanation import decimal # Tuple t = (1, (1, 1), -2) print('Input :', t) print('Decimal:', decimal.Decimal(t)) Explanation: Decimals can also be created from tuples containing a sign flag End of explanation import decimal d = decimal.Decimal(1.1) print('Precision:') print('{:.1}'.format(d)) print('{:.2}'.format(d)) print('{:.3}'.format(d)) print('{:.18}'.format(d)) print('\nWidth and precision combined:') print('{:5.1f} {:5.1g}'.format(d, d)) print('{:5.2f} {:5.2g}'.format(d, d)) print('{:5.2f} {:5.2g}'.format(d, d)) print('\nZero padding:') print('{:05.1}'.format(d)) print('{:05.2}'.format(d)) print('{:05.3}'.format(d)) Explanation: Formatting End of explanation import decimal a = decimal.Decimal('5.1') b = decimal.Decimal('3.14') c = 4 d = 3.14 print('a =', repr(a)) print('b =', repr(b)) print('c =', repr(c)) print('d =', repr(d)) print() print('a + b =', a + b) print('a - b =', a - b) print('a * b =', a * b) print('a / b =', a / b) print() print('a + c =', a + c) print('a - c =', a - c) print('a * c =', a * c) print('a / c =', a / c) print() print('a + d =', end=' ') try: print(a + d) except TypeError as e: print(e) Explanation: Arithmetic End of explanation import decimal for value in ['Infinity', 'NaN', '0']: print(decimal.Decimal(value), decimal.Decimal('-' + value)) print() # Math with infinity print('Infinity + 1:', (decimal.Decimal('Infinity') + 1)) print('-Infinity + 1:', (decimal.Decimal('-Infinity') + 1)) # Print comparing NaN print(decimal.Decimal('NaN') == decimal.Decimal('Infinity')) print(decimal.Decimal('NaN') != decimal.Decimal(1)) Explanation: Special Value End of explanation import decimal context = decimal.getcontext() print('Emax =', context.Emax) print('Emin =', context.Emin) print('capitals =', context.capitals) print('prec =', context.prec) print('rounding =', context.rounding) print('flags =') for f, v in context.flags.items(): print(' {}: {}'.format(f, v)) print('traps =') for t, v in context.traps.items(): print(' {}: {}'.format(t, v)) Explanation: Context End of explanation import decimal d = decimal.Decimal('0.123456') for i in range(1, 5): decimal.getcontext().prec = i print(i, ':', d, d * 1) Explanation: Precision End of explanation import decimal with decimal.localcontext() as c: c.prec = 2 print('Local precision:', c.prec) print('3.14 / 3 =', (decimal.Decimal('3.14') / 3)) print() print('Default precision:', decimal.getcontext().prec) print('3.14 / 3 =', (decimal.Decimal('3.14') / 3)) import decimal # Set up a context with limited precision c = decimal.getcontext().copy() c.prec = 3 # Create our constant pi = c.create_decimal('3.1415') # The constant value is rounded off print('PI :', pi) # The result of using the constant uses the global context print('RESULT:', decimal.Decimal('2.01') * pi) Explanation: Local Context End of explanation <END_TASK>
15,920
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: ModelSelection.ipynb Choosing the number of states and a suitable timescale for hidden Markov models One of the challenges associated with using hidden Markov models is specifying the correct model. For example, how many hidden states should the model have? At what timescale should we bin our observations? How much data do we need in order to train an effective/useful/representative model? One possibility (which is conceptually very appealing) is to use a nonparametric Bayesian extension to the HMM, the HDP-HMM (hierarchical Dirichlet process hidden Markov model), in which the number of states can be directly inferred from the data, and moreover, where the number of states are allowed to grow as we obtain more and more data. Fortunately, even if we choose to use a simple HMM, model selection is perhaps not as important as one might at first think. More specifically, we will show that for a wide range of model states, and for a wide range of timescales, the HMM should return plausible and usable models, so that we can use them to learn something about the data even if we don't have a good idea of what the model parameters should be. Nevertheless, shifting over to the HDP-HMMs and especially to the HDP-HSMMs (semi-Markov models) where state durations are explicitly specified or learned, is certainly something that I would highly recommend. TODO Step1: Load data Here we consider lin2 data for gor01 on the first recording day (6-7-2006), since this session had the most units (91) of all the gor01 sessions, and lin2 has position data, whereas lin1 only has partial position data. Step2: Find most appropriate number of states using cross validation Here we split the data into training, validation, and test sets. We monitor the average log probability per sequence (normalized by length) for each of these sets, and we use the validation set to choose the number of model states $m$. Note to self Step3: Remarks Step4: Remarks Step5: Remarks Step6: Remarks Step7: Remarks Step8: Remarks Step9: then we stat to see the emergence of the S-shaped place field progressions again, indicating that the reward locations are overexpressed by several different states. This observation is even more pronounced if we increase the number of states further Step10: With enough expressiveness in the number of states, we see the S-shaped curve reappear, which suggests an overexpression of the reward locations, which is consistent with what we see with place cells in animals.
<ASSISTANT_TASK:> Python Code: import matplotlib.pyplot as plt import numpy as np import pandas as pd import seaborn as sns import sys from IPython.display import display, clear_output sys.path.insert(0, 'helpers') from efunctions import * # load my helper function(s) to save pdf figures, etc. from hc3 import load_data, get_sessions from hmmlearn import hmm # see https://github.com/ckemere/hmmlearn import klabtools as klab import seqtools as sq import importlib importlib.reload(sq) # reload module here only while prototyping... importlib.reload(klab) # reload module here only while prototyping... %matplotlib inline sns.set(rc={'figure.figsize': (12, 4),'lines.linewidth': 1.5}) sns.set_style("white") Explanation: ModelSelection.ipynb Choosing the number of states and a suitable timescale for hidden Markov models One of the challenges associated with using hidden Markov models is specifying the correct model. For example, how many hidden states should the model have? At what timescale should we bin our observations? How much data do we need in order to train an effective/useful/representative model? One possibility (which is conceptually very appealing) is to use a nonparametric Bayesian extension to the HMM, the HDP-HMM (hierarchical Dirichlet process hidden Markov model), in which the number of states can be directly inferred from the data, and moreover, where the number of states are allowed to grow as we obtain more and more data. Fortunately, even if we choose to use a simple HMM, model selection is perhaps not as important as one might at first think. More specifically, we will show that for a wide range of model states, and for a wide range of timescales, the HMM should return plausible and usable models, so that we can use them to learn something about the data even if we don't have a good idea of what the model parameters should be. Nevertheless, shifting over to the HDP-HMMs and especially to the HDP-HSMMs (semi-Markov models) where state durations are explicitly specified or learned, is certainly something that I would highly recommend. TODO: Take a look at e.g. https://www.cs.cmu.edu/~ggordon/siddiqi-gordon-moore.fast-hmm.pdf : fast HMM (order of magnitude faster than Baum-Welch) and better model fit: V-STACS. Import packages and initialization End of explanation datadirs = ['/home/etienne/Dropbox/neoReader/Data', 'C:/etienne/Dropbox/neoReader/Data', '/Users/etienne/Dropbox/neoReader/Data'] fileroot = next( (dir for dir in datadirs if os.path.isdir(dir)), None) animal = 'gor01'; month,day = (6,7); session = '16-40-19' # 91 units spikes = load_data(fileroot=fileroot, datatype='spikes',animal=animal, session=session, month=month, day=day, fs=32552, verbose=False) eeg = load_data(fileroot=fileroot, datatype='eeg', animal=animal, session=session, month=month, day=day,channels=[0,1,2], fs=1252, starttime=0, verbose=False) posdf = load_data(fileroot=fileroot, datatype='pos',animal=animal, session=session, month=month, day=day, verbose=False) speed = klab.get_smooth_speed(posdf,fs=60,th=8,cutoff=0.5,showfig=False,verbose=False) Explanation: Load data Here we consider lin2 data for gor01 on the first recording day (6-7-2006), since this session had the most units (91) of all the gor01 sessions, and lin2 has position data, whereas lin1 only has partial position data. End of explanation ## bin ALL spikes ds = 0.125 # bin spikes into 125 ms bins (theta-cycle inspired) binned_spikes_all = klab.bin_spikes(spikes.data, ds=ds, fs=spikes.samprate, verbose=True) ## identify boundaries for running (active) epochs and then bin those observations into separate sequences: runbdries = klab.get_boundaries_from_bins(eeg.samprate,bins=speed.active_bins,bins_fs=60) binned_spikes_bvr = klab.bin_spikes(spikes.data, fs=spikes.samprate, boundaries=runbdries, boundaries_fs=eeg.samprate, ds=ds) ## stack data for hmmlearn: seq_stk_bvr = sq.data_stack(binned_spikes_bvr, verbose=True) seq_stk_all = sq.data_stack(binned_spikes_all, verbose=True) ## split data into train, test, and validation sets: tr_b,vl_b,ts_b = sq.data_split(seq_stk_bvr, tr=60, vl=20, ts=20, randomseed = 0, verbose=False) Smax = 40 S = np.arange(start=5,step=1,stop=Smax+1) tr_ll = [] vl_ll = [] ts_ll = [] for num_states in S: clear_output(wait=True) print('Training and evaluating {}-state hmm'.format(num_states)) sys.stdout.flush() myhmm = sq.hmm_train(tr_b, num_states=num_states, n_iter=30, verbose=False) tr_ll.append( (np.array(list(sq.hmm_eval(myhmm, tr_b)))/tr_b.sequence_lengths ).mean()) vl_ll.append( (np.array(list(sq.hmm_eval(myhmm, vl_b)))/vl_b.sequence_lengths ).mean()) ts_ll.append( (np.array(list(sq.hmm_eval(myhmm, ts_b)))/ts_b.sequence_lengths ).mean()) clear_output(wait=True) print('Done!') sys.stdout.flush() num_states = 35 fig = plt.figure(1, figsize=(12, 4)) ax = fig.add_subplot(111) ax.annotate('plateau at approx ' + str(num_states), xy=(num_states, -38.5), xycoords='data', xytext=(-140, -30), textcoords='offset points', arrowprops=dict(arrowstyle="->", connectionstyle="angle3,angleA=0,angleB=-90"), ) ax.plot(S, tr_ll, lw=1.5, label='train') ax.plot(S, vl_ll, lw=1.5, label='validation') ax.plot(S, ts_ll, lw=1.5, label='test') ax.legend(loc=2) ax.set_xlabel('number of states') ax.set_ylabel('normalized (to single time bin) log likelihood') ax.axhspan(-38.5, -37.5, facecolor='0.75', alpha=0.25) ax.set_xlim([5, S[-1]]) Explanation: Find most appropriate number of states using cross validation Here we split the data into training, validation, and test sets. We monitor the average log probability per sequence (normalized by length) for each of these sets, and we use the validation set to choose the number of model states $m$. Note to self: I should re-write my data splitting routines to allow me to extract as many subsets as I want, so that I can do k-fold cross validation. End of explanation from placefieldviz import hmmplacefieldposviz num_states = 35 ds = 0.0625 # bin spikes into 62.5 ms bins (theta-cycle inspired) vth = 8 # units/sec velocity threshold for place fields #state_pos, peakorder = hmmplacefieldposviz(num_states=num_states, ds=ds, posdf=posdf, spikes=spikes, speed=speed, vth=vth) fig, axes = plt.subplots(4, 3, figsize=(17, 11)) axes = [item for sublist in axes for item in sublist] for ii, ax in enumerate(axes): vth = ii+1 state_pos, peakorder, stateorder = hmmplacefieldposviz(num_states=num_states, ds=ds, posdf=posdf, spikes=spikes, speed=speed, vth=vth, normalize=True, verbose=False) ax.matshow(state_pos[peakorder,:], interpolation='none', cmap='OrRd') #ax.set_xlabel('position bin') ax.set_ylabel('state') ax.set_xticklabels([]) ax.set_yticklabels([]) ax.set_title('learned place fields; RUN > ' + str(vth), y=1.02) ax.plot([13, 35], [0, num_states], color='k', linestyle='dashed', linewidth=1) ax.plot([7, 41], [0, num_states], color='k', linestyle='dashed', linewidth=1) ax.axis('tight') Explanation: Remarks: We see that the training error is decreasing (equivalently, the training log probability is increasing) over the entire range of states considered. Indeed, we have computed this for a much larger number of states, and the training error keeps on decreasing, whereas both the validation and test errors reach a plateau at around 30 or 35 states. As expected, the training set has the largest log probability (best agreement with model), but we might expect the test and validation sets to be about the same. For different subsets of our data this is indeed the case, but the more important thing in model selection is that the validation and test sets should have the same shape or behavior, so that we can choose an appropriate model parameter. However, if we wanted to predict what our log probability for any given sequence would be, then we probably need a little bit more data, for which the test and validation errors should agree more. Finally, we have also repeated the above analysis when we restricted ourselves to only using place cells in the model, and although the log probabilities were uniformly increased to around $-7$ or $-8$, the overall shape and characteristic behavior were left unchanged, so that model selection could be done either way. Place field visualization Previously we have only considered varying the number of model states for model selection, but of course choosing an appropriate timescale is perhaps just as important. We know, for example, that if our timescale is too short (or fast), then most of the bins will be empty, making it difficult for the model to learn appropriate representations and transitions. On the other hand, if our timescale is too coarse (or long or slow) then we will certainly miss SWR events, and we may even miss some behavioral events as well. Since theta is around 8 Hz for rodents, it might make sense to consider a timescale of 125 ms or even 62.5 ms for behaviorally relevant events, so that we can hope to capture half or full theta cycles in the observations. One might also reasonably ask: "even though the log probability has been optimized, how do we know that the learned model makes any sense? That is, that the model is plausible and useful?" One way to try to answer this question is to again consider the place fields that we learn from the data. Place field visualization is considered in more detail in StateClustering.ipynb, but here we simply want to see if we get plausible, behaviorally relevant state representations out when choosing different numbers of states, and different timescales, for example. Place fields for varying velocity thresholds We train our models on RUN data, so we might want to know how sensitive our model is to a specific velocity threshold. Using a smaller threshold will include more quiescent data, and using a larger threshold will exclude more data from being used to learn in the model. End of explanation from placefieldviz import hmmplacefieldposviz num_states = 35 ds = 0.0625 # bin spikes into 62.5 ms bins (theta-cycle inspired) vth = 8 # units/sec velocity threshold for place fields #state_pos, peakorder = hmmplacefieldposviz(num_states=num_states, ds=ds, posdf=posdf, spikes=spikes, speed=speed, vth=vth) fig, axes = plt.subplots(4, 3, figsize=(17, 11)) axes = [item for sublist in axes for item in sublist] for ii, ax in enumerate(axes): num_states = 5 + ii*5 state_pos, peakorder, stateorder = hmmplacefieldposviz(num_states=num_states, ds=ds, posdf=posdf, spikes=spikes, speed=speed, vth=vth, normalize=True) ax.matshow(state_pos[peakorder,:], interpolation='none', cmap='OrRd') #ax.set_xlabel('position bin') ax.set_ylabel('state') ax.set_xticklabels([]) ax.set_yticklabels([]) ax.set_title('learned place fields; RUN > ' + str(vth) + '; m = ' + str(num_states), y=1.02) ax.axis('tight') saveFigure('posterfigs/numstates.pdf') Explanation: Remarks: As can be expected, with low velocity thresholds, we see an overrepresentation of the reward locations, and only a relatively small number of states that are dedicated to encoding the position along the track. Recall that the track was shortened halfway through the recording session. Here, the reward locations for the longer track (first half of the experiment) and shorter track (second half of the experiment) are shown by the ends of the dashed lines. We notice that at some point, the movement velocity (for fixed state evolution) appears to be constant, and that at e.g. 8 units/sec we see a clear bifurcation in the place fields, so that states encode both positions before and after the track was shortened. Place fields for varying number of states Next, we take a look at how the place fields are affected by changing the number of states in the model. End of explanation from placefieldviz import hmmplacefieldposviz num_states = 35 ds = 0.0625 # bin spikes into 62.5 ms bins (theta-cycle inspired) vth = 8 # units/sec velocity threshold for place fields #state_pos, peakorder = hmmplacefieldposviz(num_states=num_states, ds=ds, posdf=posdf, spikes=spikes, speed=speed, vth=vth) fig, axes = plt.subplots(4, 3, figsize=(17, 11)) axes = [item for sublist in axes for item in sublist] for ii, ax in enumerate(axes): ds = (ii+1)*0.03125 state_pos, peakorder, stateorder = hmmplacefieldposviz(num_states=num_states, ds=ds, posdf=posdf, spikes=spikes, speed=speed, vth=vth, normalize=True) ax.matshow(state_pos[peakorder,:], interpolation='none', cmap='OrRd') #ax.set_xlabel('position bin') ax.set_ylabel('state') ax.set_xticklabels([]) ax.set_yticklabels([]) ax.set_title('learned place fields; RUN > ' + str(vth) + '; m = ' + str(num_states) + '; ds = ' + str(ds), y=1.02) ax.plot([13, 35], [0, num_states], color='k', linestyle='dashed', linewidth=1) ax.plot([7, 41], [0, num_states], color='k', linestyle='dashed', linewidth=1) ax.axis('tight') Explanation: Remarks: First, we see that independent of the number of states, the model captures the place field like nature of the underlying states very well. Furthermore, the bifurcation of some states to represent both the first and second halves of the experiment becomes clear with as few as 15 states, but interestingly this bifurcation fades as we add more states to the model, since there is enough flexibility to encode those shifting positions by their own states. Warning: However, in the case where we have many states so that the states are no longer bimodal, the strict linear ordering that we impose (ordering by peak firing location) can easily mask the underlying structural change in the environment. Place fields for varying timescales Next we investigate how the place fields are affected by changing the timescale of our observations. First, we consider timescales in the range of 31.25 ms to 375 ms, in increments of 31.25 ms. End of explanation from placefieldviz import hmmplacefieldposviz num_states = 35 ds = 0.0625 # bin spikes into 62.5 ms bins (theta-cycle inspired) vth = 8 # units/sec velocity threshold for place fields #state_pos, peakorder = hmmplacefieldposviz(num_states=num_states, ds=ds, posdf=posdf, spikes=spikes, speed=speed, vth=vth) fig, axes = plt.subplots(4, 3, figsize=(17, 11)) axes = [item for sublist in axes for item in sublist] for ii, ax in enumerate(axes): ds = (ii+1)*0.0625 state_pos, peakorder, stateorder = hmmplacefieldposviz(num_states=num_states, ds=ds, posdf=posdf, spikes=spikes, speed=speed, vth=vth, normalize=True) ax.matshow(state_pos[peakorder,:], interpolation='none', cmap='OrRd') #ax.set_xlabel('position bin') ax.set_ylabel('state') ax.set_xticklabels([]) ax.set_yticklabels([]) ax.set_title('learned place fields; RUN > ' + str(vth) + '; m = ' + str(num_states) + '; ds = ' + str(ds), y=1.02) ax.plot([13, 35], [0, num_states], color='k', linestyle='dashed', linewidth=1) ax.plot([7, 41], [0, num_states], color='k', linestyle='dashed', linewidth=1) ax.axis('tight') Explanation: Remarks: We notice that we clearly see the bimodal place fields when the timescales are sufficiently small, with a particularly clear example at 62.5 ms, for example. Larger timescales tend to focus on the longer track piece, with a single trajectory being skewed away towards the shorter track piece. Next we consider timescales in increments of 62.5 ms. End of explanation from placefieldviz import hmmplacefieldposviz num_states = 25 ds = 0.0625 # bin spikes into 62.5 ms bins (theta-cycle inspired) vth = 8 # units/sec velocity threshold for place fields state_pos_b, peakorder_b, stateorder_b = hmmplacefieldposviz(num_states=num_states, ds=ds, posdf=posdf, spikes=spikes, speed=speed, vth=vth, normalize=True, experiment='both') state_pos_1, peakorder_1, stateorder_1 = hmmplacefieldposviz(num_states=num_states, ds=ds, posdf=posdf, spikes=spikes, speed=speed, vth=vth, normalize=True, experiment='first') state_pos_2, peakorder_2, stateorder_2 = hmmplacefieldposviz(num_states=num_states, ds=ds, posdf=posdf, spikes=spikes, speed=speed, vth=vth, normalize=True, experiment='second') fig, (ax1, ax2, ax3) = plt.subplots(1, 3, figsize=(17, 3)) ax1.matshow(state_pos_b[peakorder_b,:], interpolation='none', cmap='OrRd') ax1.set_ylabel('state') ax1.set_xticklabels([]) ax1.set_yticklabels([]) ax1.set_title('learned place fields BOTH; RUN > ' + str(vth) + '; m = ' + str(num_states) + '; ds = ' + str(ds), y=1.02) ax1.plot([13, 35], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax1.plot([7, 41], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax1.axis('tight') ax2.matshow(state_pos_1[peakorder_1,:], interpolation='none', cmap='OrRd') ax2.set_ylabel('state') ax2.set_xticklabels([]) ax2.set_yticklabels([]) ax2.set_title('learned place fields FIRST; RUN > ' + str(vth) + '; m = ' + str(num_states) + '; ds = ' + str(ds), y=1.02) ax2.plot([13, 35], [0, num_states], color='gray', linestyle='dashed', linewidth=1) ax2.plot([7, 41], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax2.axis('tight') ax3.matshow(state_pos_2[peakorder_2,:], interpolation='none', cmap='OrRd') ax3.set_ylabel('state') ax3.set_xticklabels([]) ax3.set_yticklabels([]) ax3.set_title('learned place fields SECOND; RUN > ' + str(vth) + '; m = ' + str(num_states) + '; ds = ' + str(ds), y=1.02) ax3.plot([13, 35], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax3.plot([7, 41], [0, num_states], color='gray', linestyle='dashed', linewidth=1) ax3.axis('tight') saveFigure('posterfigs/expsplit.pdf') Explanation: Remarks: Again, we see that with larger timescales, the spatial resolution becomes more coarse, because we don't have that sufficiently many observations, and the modes of the place fields tend to lie close to those associated wit the longer track. Splitting the experimment in half Just as a confirmation of what we've seen so far, we next consider the place fields obtained when we split the experiment into its first and second halves, correponding to when the track was longer, and shorter, respectively. End of explanation from placefieldviz import hmmplacefieldposviz num_states = 45 ds = 0.0625 # bin spikes into 62.5 ms bins (theta-cycle inspired) vth = 8 # units/sec velocity threshold for place fields state_pos_b, peakorder_b, stateorder_b = hmmplacefieldposviz(num_states=num_states, ds=ds, posdf=posdf, spikes=spikes, speed=speed, vth=vth, normalize=True, experiment='both') state_pos_1, peakorder_1, stateorder_1 = hmmplacefieldposviz(num_states=num_states, ds=ds, posdf=posdf, spikes=spikes, speed=speed, vth=vth, normalize=True, experiment='first') state_pos_2, peakorder_2, stateorder_2 = hmmplacefieldposviz(num_states=num_states, ds=ds, posdf=posdf, spikes=spikes, speed=speed, vth=vth, normalize=True, experiment='second') fig, (ax1, ax2, ax3) = plt.subplots(1, 3, figsize=(17, 3)) ax1.matshow(state_pos_b[peakorder_b,:], interpolation='none', cmap='OrRd') ax1.set_ylabel('state') ax1.set_xticklabels([]) ax1.set_yticklabels([]) ax1.set_title('learned place fields BOTH; RUN > ' + str(vth) + '; m = ' + str(num_states) + '; ds = ' + str(ds), y=1.02) ax1.plot([13, 35], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax1.plot([7, 41], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax1.axis('tight') ax2.matshow(state_pos_1[peakorder_1,:], interpolation='none', cmap='OrRd') ax2.set_ylabel('state') ax2.set_xticklabels([]) ax2.set_yticklabels([]) ax2.set_title('learned place fields FIRST; RUN > ' + str(vth) + '; m = ' + str(num_states) + '; ds = ' + str(ds), y=1.02) ax2.plot([13, 35], [0, num_states], color='gray', linestyle='dashed', linewidth=1) ax2.plot([7, 41], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax2.axis('tight') ax3.matshow(state_pos_2[peakorder_2,:], interpolation='none', cmap='OrRd') ax3.set_ylabel('state') ax3.set_xticklabels([]) ax3.set_yticklabels([]) ax3.set_title('learned place fields SECOND; RUN > ' + str(vth) + '; m = ' + str(num_states) + '; ds = ' + str(ds), y=1.02) ax3.plot([13, 35], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax3.plot([7, 41], [0, num_states], color='gray', linestyle='dashed', linewidth=1) ax3.axis('tight') Explanation: Remarks: We clearly see the bimodal place fields when we use all of the data, and we see the unimodal place fields emerge as we focus on either the first, or the second half of the experiment. Notice that the reward locations are more concentrated, but that the velocity (with fixed state progression) is roughly constant. However, if we increase the number of states: End of explanation from placefieldviz import hmmplacefieldposviz num_states = 100 ds = 0.0625 # bin spikes into 62.5 ms bins (theta-cycle inspired) vth = 8 # units/sec velocity threshold for place fields state_pos_b, peakorder_b, stateorder_b = hmmplacefieldposviz(num_states=num_states, ds=ds, posdf=posdf, spikes=spikes, speed=speed, vth=vth, normalize=True, experiment='both') state_pos_1, peakorder_1, stateorder_1 = hmmplacefieldposviz(num_states=num_states, ds=ds, posdf=posdf, spikes=spikes, speed=speed, vth=vth, normalize=True, experiment='first') state_pos_2, peakorder_2, stateorder_2 = hmmplacefieldposviz(num_states=num_states, ds=ds, posdf=posdf, spikes=spikes, speed=speed, vth=vth, normalize=True, experiment='second') fig, (ax1, ax2, ax3) = plt.subplots(1, 3, figsize=(17, 3)) ax1.matshow(state_pos_b[peakorder_b,:], interpolation='none', cmap='OrRd') ax1.set_ylabel('state') ax1.set_xticklabels([]) ax1.set_yticklabels([]) ax1.set_title('learned place fields BOTH; RUN > ' + str(vth) + '; m = ' + str(num_states) + '; ds = ' + str(ds), y=1.02) ax1.plot([13, 35], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax1.plot([7, 41], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax1.axis('tight') ax2.matshow(state_pos_1[peakorder_1,:], interpolation='none', cmap='OrRd') ax2.set_ylabel('state') ax2.set_xticklabels([]) ax2.set_yticklabels([]) ax2.set_title('learned place fields FIRST; RUN > ' + str(vth) + '; m = ' + str(num_states) + '; ds = ' + str(ds), y=1.02) ax2.plot([13, 35], [0, num_states], color='gray', linestyle='dashed', linewidth=1) ax2.plot([7, 41], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax2.axis('tight') ax2.add_patch( patches.Rectangle( (-1, 0), # (x,y) 8, # width num_states, # height hatch='/', facecolor='w', alpha=0.5 ) ) ax2.add_patch( patches.Rectangle( (41, 0), # (x,y) 11, # width num_states, # height hatch='/', facecolor='w', alpha=0.5 ) ) ax3.matshow(state_pos_2[peakorder_2,:], interpolation='none', cmap='OrRd') ax3.set_ylabel('state') ax3.set_xticklabels([]) ax3.set_yticklabels([]) ax3.set_title('learned place fields SECOND; RUN > ' + str(vth) + '; m = ' + str(num_states) + '; ds = ' + str(ds), y=1.02) ax3.plot([13, 35], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax3.plot([7, 41], [0, num_states], color='gray', linestyle='dashed', linewidth=1) ax3.axis('tight') ax3.add_patch( patches.Rectangle( (-1, 0), # (x,y) 14, # width num_states, # height hatch='/', facecolor='w', alpha=0.5 ) ) ax3.add_patch( patches.Rectangle( (35, 0), # (x,y) 15, # width num_states, # height hatch='/', facecolor='w', alpha=0.5 ) ) Explanation: then we stat to see the emergence of the S-shaped place field progressions again, indicating that the reward locations are overexpressed by several different states. This observation is even more pronounced if we increase the number of states further: End of explanation import matplotlib.patches as patches fig, (ax1, ax2, ax3) = plt.subplots(1, 3, figsize=(17, 3)) ax1.matshow(state_pos_b[stateorder_b,:], interpolation='none', cmap='OrRd') ax1.set_ylabel('state') ax1.set_xticklabels([]) ax1.set_yticklabels([]) ax1.set_title('learned place fields BOTH; RUN > ' + str(vth) + '; m = ' + str(num_states) + '; ds = ' + str(ds), y=1.02) ax1.plot([13, 35], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax1.plot([7, 41], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax1.axis('tight') ax2.matshow(state_pos_1[stateorder_1,:], interpolation='none', cmap='OrRd') ax2.set_ylabel('state') ax2.set_xticklabels([]) ax2.set_yticklabels([]) ax2.set_title('learned place fields FIRST; RUN > ' + str(vth) + '; m = ' + str(num_states) + '; ds = ' + str(ds), y=1.02) ax2.plot([13, 13], [0, num_states], color='gray', linestyle='dashed', linewidth=1) ax2.plot([7, 7], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax2.plot([35, 35], [0, num_states], color='gray', linestyle='dashed', linewidth=1) ax2.plot([41, 41], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax2.axis('tight') ax2.add_patch( patches.Rectangle( (-1, 0), # (x,y) 8, # width num_states, # height hatch='/', facecolor='w', alpha=0.5 ) ) ax2.add_patch( patches.Rectangle( (41, 0), # (x,y) 11, # width num_states, # height hatch='/', facecolor='w', alpha=0.5 ) ) ax3.matshow(state_pos_2[stateorder_2,:], interpolation='none', cmap='OrRd') ax3.set_ylabel('state') ax3.set_xticklabels([]) ax3.set_yticklabels([]) ax3.set_title('learned place fields SECOND; RUN > ' + str(vth) + '; m = ' + str(num_states) + '; ds = ' + str(ds), y=1.02) ax3.plot([13, 13], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax3.plot([7, 7], [0, num_states], color='gray', linestyle='dashed', linewidth=1) ax3.plot([35, 35], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax3.plot([41, 41], [0, num_states], color='gray', linestyle='dashed', linewidth=1) ax3.axis('tight') ax3.add_patch( patches.Rectangle( (-1, 0), # (x,y) 14, # width num_states, # height hatch='/', facecolor='w', alpha=0.5 ) ) ax3.add_patch( patches.Rectangle( (35, 0), # (x,y) 15, # width num_states, # height hatch='/', facecolor='w', alpha=0.5 ) ) fig.suptitle('State ordering not by peak location, but by the state transition probability matrix', y=1.08, fontsize=14) saveFigure('posterfigs/zigzag.pdf') state_pos_b[state_pos_b < np.transpose(np.tile(state_pos_b.max(axis=1),[state_pos_b.shape[1],1]))] = 0 state_pos_b[state_pos_b == np.transpose(np.tile(state_pos_b.max(axis=1),[state_pos_b.shape[1],1]))] = 1 state_pos_1[state_pos_1 < np.transpose(np.tile(state_pos_1.max(axis=1),[state_pos_1.shape[1],1]))] = 0 state_pos_1[state_pos_1 == np.transpose(np.tile(state_pos_1.max(axis=1),[state_pos_1.shape[1],1]))] = 1 state_pos_2[state_pos_2 < np.transpose(np.tile(state_pos_2.max(axis=1),[state_pos_2.shape[1],1]))] = 0 state_pos_2[state_pos_2 == np.transpose(np.tile(state_pos_2.max(axis=1),[state_pos_2.shape[1],1]))] = 1 fig, (ax1, ax2, ax3) = plt.subplots(1, 3, figsize=(17, 3)) ax1.matshow(state_pos_b[peakorder_b,:], interpolation='none', cmap='OrRd') ax1.set_ylabel('state') ax1.set_xticklabels([]) ax1.set_yticklabels([]) ax1.set_title('learned place fields BOTH; RUN > ' + str(vth) + '; m = ' + str(num_states) + '; ds = ' + str(ds), y=1.02) ax1.plot([13, 35], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax1.plot([7, 41], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax1.axis('tight') ax2.matshow(state_pos_1[peakorder_1,:], interpolation='none', cmap='OrRd') ax2.set_ylabel('state') ax2.set_xticklabels([]) ax2.set_yticklabels([]) ax2.set_title('learned place fields FIRST; RUN > ' + str(vth) + '; m = ' + str(num_states) + '; ds = ' + str(ds), y=1.02) ax2.plot([13, 35], [0, num_states], color='gray', linestyle='dashed', linewidth=1) ax2.plot([7, 41], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax2.axis('tight') ax3.matshow(state_pos_2[peakorder_2,:], interpolation='none', cmap='OrRd') ax3.set_ylabel('state') ax3.set_xticklabels([]) ax3.set_yticklabels([]) ax3.set_title('learned place fields SECOND; RUN > ' + str(vth) + '; m = ' + str(num_states) + '; ds = ' + str(ds), y=1.02) ax3.plot([13, 35], [0, num_states], color='k', linestyle='dashed', linewidth=2) ax3.plot([7, 41], [0, num_states], color='gray', linestyle='dashed', linewidth=1) ax3.axis('tight') Explanation: With enough expressiveness in the number of states, we see the S-shaped curve reappear, which suggests an overexpression of the reward locations, which is consistent with what we see with place cells in animals. End of explanation <END_TASK>
15,921
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Basics of Marvin In this notebook, you will learn the common core functionality across many of the Marvin Tools. This includes the basics of accessing and handling MaNGA data from different locations, as well as a beginners guide of interacting with data via the core tools of Cubes and Maps. Much of this information can be found in the Getting Started section of the Marvin documentation. Table of Contents Data Access Modes <br> Interacting with Data Content <br> Object Datamodels <br> Extracting Data <br> Accessing Associated Objects<br> Quality and Target Flags<br> NSA and DAPall Metadata<br> Downloading Files<br> Interacting with Spaxels<br> <a id="common"></a> Common Core Functionality Marvin Tools provides the core functionality for accessing MaNGA data with Marvin. At the lowest level they are class wrappers around data products or elements (datacubes, maps, and spaxels). Their purpose is to provide a more natural way to interact with the data, unconstrained by specific data storage architectures such as files or databases, or by data location such as a laptop or remote server. All Marvin Tools are built off a set of common core functionalities and features, and thus behave in very similar ways. There are four main classes provided in Marvin Step1: You may see the info message No release version set. Setting default to DR15. The default data release Marvin uses is DR15. For this tutorial, we will be using DR15 data so there is no need to do anything further. In general, to change which release Marvin uses, we can use the Marvin config.setRelease method on the config object. Step2: <a id="modes"></a> Data Access Modes Let's instantiate our first cube, for galaxy with plate-IFU designation 8485-1901. Each marvin tool takes a string id input and can one of Step3: Once instantiated, each marvin tool indicates a mode and a data-origin. Note that mode="local" and data_origin="file". This indicates that we have accessed a file from our local filesystem. Our cube now has a filename attribute indicating its location on disk. Step4: What if we won't have the MaNGa data products for our target of interest? We can specify the plateifu or mangaid of the target and Marvin will access it remotely. Let's open a cube for 8485-1902. Step5: Now note that mode="remote" and data_origin="api". This indicates the cube has been remotely loaded using the built-in Marvin API. By specifying an input plateifu or mangaid, marvin will always first try to access the data locally, and then fall back to remote access. You can also explicitly load a file using the filename keyword argument. Step6: <a id="interaction"></a> Interacting with Data Content Regardless of how you load an object, the way we interact with all the tools, and the way they behave, is the same. They also all provide some quick access to basic metadata, like coordinates, header information, and WCS. <a id="metadata"></a> Step7: <a id="datamodel"></a> Object Datamodels Every Marvin object comes with a built-in datamodel, which shows us what extensions are available, how they are names, and what they contain. Step8: The datamodel for a cube contains 3 datacube extensions and 2 spectral extensions. Let's look at them more closely. Step9: This tells us that this cube has three associated 3D datacubes, flux, dispersion, and dispersion_prepixel, and two associated spectra, spectral_resolution and spectral_resolution_prepixel, as well as their associated units. The description attribute provides information on what the extension is. Step10: The names of each of the datacubes and spectral extensions are available as object attributes and can be used to access the associated data quantity. Let's access the flux for our current cube. Step11: The flux is represented as a 3D array with units. We can also access the inverse variance and the mask using flux.ivar and flux.mask, respectively. Marvin uses Astropy quantities to represent all internal multidimensional data, e.g. 3-d datacube and 2-d map arrays, or 1-d spectral data. <a id="extract"></a> Extracting data All marvin arrays can be sliced to extract subsets of data. Slices of datacubes produce new datacubes. Slices of spectra produce new spectra. A single slice from a datacube extracts a spectrum. Each spectrum can be easily displayed with its plot attribute. Step12: <a id="associates"></a> Accessing Associated Objects For a given target galaxy, there are many associated MaNGA Data Products from both the reduction and analysis pipelines. Marvin seamlessly connects these together so you don't have to access them individually. Let's access the DAP Maps objects associated with our cube. To do so we use the getMaps method on cube. All instances have getXXX methods that allow you to access associated data objects for a given target. Step13: When accessing associated objects, Marvin also smartly determines whether to open the file locally or remotely grab it over the API. In our case, the mode is local and data_origin is file. DAP Maps have different bintypes. The default binning is HYB10, a hybrid binning scheme to S/N~10. To load a different bintype, use the bintype keyword argument in the cube.getMaps method. A Maps behaves very similarly to a Cube and everything we have discussed above will still work. Instead of datacubes and spectra, a Maps object contains a set of 2D quantities called Map, each one of them representing a different property measured by the DAP. One can get a full list of all the properties available using the datamodel. Step14: Note that some properties such as spx_skycoo have multiple channels (in this case the on-sky x and y coordinates). To access the individual property, simply combine the property name and the channel name with an underscore, i.e. spx_skycoo_on_sky_x. As before, we can get more information about a property using the description attribute. Step15: More information about Marvin datamodels can be found here. See the DR15 Datamodel for a full description of the available datamodels for DR15. We can retrieve the map associated to a specific property directly from the Maps instance. For example, let’s get the Hα emission line flux (fitted by a Gaussian). As Map quantities are similar to DataCube, we can easily plot the 2d map. Step16: Note that the plot method returns the matplotlib Figure and Axes for the plot. We can use those to modify or save the plot. Marvin plotting routines try to select the best parameters, colour maps, and dynamic ranges. You can modify those by passing extra arguments to plot. You can learn more in the Map plotting section. We will talk about the Map class in detail in Working with Astropy Quantities and in Map. Remember that all associated objects are connected together. They can be accessed either from within an object instance or loaded independently from the class itself. From within an instance, the core tools, Maps, RSS, and ModelCube are available using the getXXX methods, as well as additional auxillary tools and object data, such as Image. Step17: <a id="flags"></a> Quality and Target Flags and Maskbits Each MaNGA target has associated quality assurance flags as well targeting flags. Quality flags indiciate the overall quality of the data and/or any problems with the data discovered during the reduction and analysis pipeline runs. Step18: In this case the MANGA_DRP3QUAL maskbit does not have any bit activated, which means the data is safe to use. See the Maskbits section for more information on how to interact with Marvin Maskbits. The targeting flags highlight the targeting bits set for this galaxy and tell you which sample the target belongs to. Step19: Note that in this case the galaxy belongs to the secondary sample from the final target selection (SECONDARY_v1_2_0) as well as to the secondary samples from several commissioning target selections. The galaxy does not have any ancillary bit (manga_target3). An manga_target3 bit set indicates the target belongs to one of the MaNGA Ancillary programs. <a id="mixin"></a> NSA, DAPall, and Value-Added catalog data For each target we can also access additional catalog data Step20: SDSS members produce Value-Added Catalogs (VACS) of MaNGA products. These products can be incorporated into Marvin. Any incorporated VACS in Marvin are contained in the vacs attribute attached to each tool. More information on VACs in Marvin can be found here. Currently the MaNGA-HI VAC for DR15 has been contributed into Marvin. Let's take a look at it. Step21: vacs is a container (list) of all names of the vacs contributed into Marvin. These names are dottable to access to vac data. These contributed custom products can be of any format chosen by the VAC owner, a Python class, a dictionary, an array, etc. Step22: <a id='download'></a> Downloading files While Marvin allows you to access data remotely, frequently you will find that you want to download the file associated to an object so that you can access it more quickly in the future. We can do that using the download method attached to all marvin core tools. Let’s try to load a cube that we know we do not have in out hard drive, and download it. Step23: Each tool contains a download method to download that individual file. If you want to download a list of targets in bulk, you can use the downloadList utility function. This function uses the built-in Python product sdss_access to download files from the Utah SAS. downloadList accepts as input a list of target ids, e.g. plateifus or mangaids. The dltype keyword indicates the type of file you wish to download. The default file is cube. Let's use the utility function to download a list of DAP MAPS files for three targets. Step24: <a id="spaxels"></a> Interacting with Spaxels All of the core Marvin Galaxy Tools have the ability to extract one ore more Spaxel objects. Spaxel objects are special Marvin objects to make it easier to deal with individual spaxels in galaxies and access their properties. See the Getting Started with Spaxels for more information on how to interact with Spaxels. We can either use the standard array slicing notation (0-indexed, origin of coordinates in the lower left corner of the array) Step25: or we can use getSpaxel method, which accepts multiple arguments. Note that by default, (x, y) coordinates passed to getSpaxel are measured from the center of the array.
<ASSISTANT_TASK:> Python Code: from marvin.tools import Cube Explanation: Basics of Marvin In this notebook, you will learn the common core functionality across many of the Marvin Tools. This includes the basics of accessing and handling MaNGA data from different locations, as well as a beginners guide of interacting with data via the core tools of Cubes and Maps. Much of this information can be found in the Getting Started section of the Marvin documentation. Table of Contents Data Access Modes <br> Interacting with Data Content <br> Object Datamodels <br> Extracting Data <br> Accessing Associated Objects<br> Quality and Target Flags<br> NSA and DAPall Metadata<br> Downloading Files<br> Interacting with Spaxels<br> <a id="common"></a> Common Core Functionality Marvin Tools provides the core functionality for accessing MaNGA data with Marvin. At the lowest level they are class wrappers around data products or elements (datacubes, maps, and spaxels). Their purpose is to provide a more natural way to interact with the data, unconstrained by specific data storage architectures such as files or databases, or by data location such as a laptop or remote server. All Marvin Tools are built off a set of common core functionalities and features, and thus behave in very similar ways. There are four main classes provided in Marvin: the Cube, the RSS, the Maps, and the ModelCube. We will start by focusing on a Marvin Cube to demonstrate the common features and handling of Marvin Tools. Everything discussed here also applies to any of the other available Marvin Tools listed above. Let's start by importing the Marvin Cube class. End of explanation from marvin import config # set the release to DR15 config.setRelease("DR15") Explanation: You may see the info message No release version set. Setting default to DR15. The default data release Marvin uses is DR15. For this tutorial, we will be using DR15 data so there is no need to do anything further. In general, to change which release Marvin uses, we can use the Marvin config.setRelease method on the config object. End of explanation # instantiate a cube for 8485-1901 cube = Cube('8485-1901') cube Explanation: <a id="modes"></a> Data Access Modes Let's instantiate our first cube, for galaxy with plate-IFU designation 8485-1901. Each marvin tool takes a string id input and can one of: a string filepath and name, a plate-IFU designation, or a manga-ID designation. Marvin will attempt to identify the input format and properly load the target data. End of explanation # print the local filename cube.filename Explanation: Once instantiated, each marvin tool indicates a mode and a data-origin. Note that mode="local" and data_origin="file". This indicates that we have accessed a file from our local filesystem. Our cube now has a filename attribute indicating its location on disk. End of explanation remote_cube = Cube('8485-1902') remote_cube Explanation: What if we won't have the MaNGa data products for our target of interest? We can specify the plateifu or mangaid of the target and Marvin will access it remotely. Let's open a cube for 8485-1902. End of explanation # Explicitly load a cube from a file on disk filename = '/Users/Brian/Work/sdss/sas/dr15/manga/spectro/redux/v2_4_3/8485/stack/manga-8485-1901-LOGCUBE.fits.gz' cube = Cube(filename=filename) cube Explanation: Now note that mode="remote" and data_origin="api". This indicates the cube has been remotely loaded using the built-in Marvin API. By specifying an input plateifu or mangaid, marvin will always first try to access the data locally, and then fall back to remote access. You can also explicitly load a file using the filename keyword argument. End of explanation print('Target Coordindates:', cube.ra, cube.dec) print('Header:') # access the PRIMARY header for the current object cube.header # show the WCS for the current object cube.wcs Explanation: <a id="interaction"></a> Interacting with Data Content Regardless of how you load an object, the way we interact with all the tools, and the way they behave, is the same. They also all provide some quick access to basic metadata, like coordinates, header information, and WCS. <a id="metadata"></a> End of explanation # look up the cube datmaodel datamodel = cube.datamodel datamodel Explanation: <a id="datamodel"></a> Object Datamodels Every Marvin object comes with a built-in datamodel, which shows us what extensions are available, how they are names, and what they contain. End of explanation # see what datacubes are available print('Datacubes:') datamodel.datacubes # see what additional spectral extensions are available print('Spectra:') datamodel.spectra Explanation: The datamodel for a cube contains 3 datacube extensions and 2 spectral extensions. Let's look at them more closely. End of explanation print('Flux description:', datamodel.datacubes.flux.description) print('Spectral Resolution description:', datamodel.spectra.spectral_resolution.description) Explanation: This tells us that this cube has three associated 3D datacubes, flux, dispersion, and dispersion_prepixel, and two associated spectra, spectral_resolution and spectral_resolution_prepixel, as well as their associated units. The description attribute provides information on what the extension is. End of explanation flux = cube.flux flux Explanation: The names of each of the datacubes and spectral extensions are available as object attributes and can be used to access the associated data quantity. Let's access the flux for our current cube. End of explanation # create a small datacube subset in a 10x10 pixel region around the center of the target central_pixel = int(cube.flux.shape[1]/2) lo = central_pixel - 5 hi = central_pixel + 5 subset = flux[:, lo:hi, lo:hi] subset # extract a single spectrum and display it spectrum = flux[:, 16, 17] spectrum spectrum.plot(show_std=True) Explanation: The flux is represented as a 3D array with units. We can also access the inverse variance and the mask using flux.ivar and flux.mask, respectively. Marvin uses Astropy quantities to represent all internal multidimensional data, e.g. 3-d datacube and 2-d map arrays, or 1-d spectral data. <a id="extract"></a> Extracting data All marvin arrays can be sliced to extract subsets of data. Slices of datacubes produce new datacubes. Slices of spectra produce new spectra. A single slice from a datacube extracts a spectrum. Each spectrum can be easily displayed with its plot attribute. End of explanation # access the Maps object from our cube maps = cube.getMaps() maps Explanation: <a id="associates"></a> Accessing Associated Objects For a given target galaxy, there are many associated MaNGA Data Products from both the reduction and analysis pipelines. Marvin seamlessly connects these together so you don't have to access them individually. Let's access the DAP Maps objects associated with our cube. To do so we use the getMaps method on cube. All instances have getXXX methods that allow you to access associated data objects for a given target. End of explanation # look at the datamodel for the maps maps.datamodel Explanation: When accessing associated objects, Marvin also smartly determines whether to open the file locally or remotely grab it over the API. In our case, the mode is local and data_origin is file. DAP Maps have different bintypes. The default binning is HYB10, a hybrid binning scheme to S/N~10. To load a different bintype, use the bintype keyword argument in the cube.getMaps method. A Maps behaves very similarly to a Cube and everything we have discussed above will still work. Instead of datacubes and spectra, a Maps object contains a set of 2D quantities called Map, each one of them representing a different property measured by the DAP. One can get a full list of all the properties available using the datamodel. End of explanation print('spx_skycoo_on_sky_x: ', maps.datamodel.spx_skycoo_on_sky_x.description) Explanation: Note that some properties such as spx_skycoo have multiple channels (in this case the on-sky x and y coordinates). To access the individual property, simply combine the property name and the channel name with an underscore, i.e. spx_skycoo_on_sky_x. As before, we can get more information about a property using the description attribute. End of explanation # grab the map of H-alpha emission line flux ha = maps.emline_gflux_ha_6564 ha # plot the H-alpha map fig, ax = ha.plot() Explanation: More information about Marvin datamodels can be found here. See the DR15 Datamodel for a full description of the available datamodels for DR15. We can retrieve the map associated to a specific property directly from the Maps instance. For example, let’s get the Hα emission line flux (fitted by a Gaussian). As Map quantities are similar to DataCube, we can easily plot the 2d map. End of explanation # load the same Maps object by itself from marvin.tools import Maps my_maps = Maps('8485-1901') my_maps # get the stellar velocity map stvel = my_maps.stellar_velocity # load an auxilliary tool, the gri color image of the target galaxy using the Marvin Image tool # get the Image of the galaxy and plot it image = cube.getImage() image.plot() Explanation: Note that the plot method returns the matplotlib Figure and Axes for the plot. We can use those to modify or save the plot. Marvin plotting routines try to select the best parameters, colour maps, and dynamic ranges. You can modify those by passing extra arguments to plot. You can learn more in the Map plotting section. We will talk about the Map class in detail in Working with Astropy Quantities and in Map. Remember that all associated objects are connected together. They can be accessed either from within an object instance or loaded independently from the class itself. From within an instance, the core tools, Maps, RSS, and ModelCube are available using the getXXX methods, as well as additional auxillary tools and object data, such as Image. End of explanation # look at the overall cube quality flags cube.quality_flag Explanation: <a id="flags"></a> Quality and Target Flags and Maskbits Each MaNGA target has associated quality assurance flags as well targeting flags. Quality flags indiciate the overall quality of the data and/or any problems with the data discovered during the reduction and analysis pipeline runs. End of explanation # look at the targeting flags for this galaxy cube.target_flags Explanation: In this case the MANGA_DRP3QUAL maskbit does not have any bit activated, which means the data is safe to use. See the Maskbits section for more information on how to interact with Marvin Maskbits. The targeting flags highlight the targeting bits set for this galaxy and tell you which sample the target belongs to. End of explanation # look at the NSA properties for this galaxy from the cube instance nsa = cube.nsa nsa # look at the DAP all properties from the maps instance maps.dapall Explanation: Note that in this case the galaxy belongs to the secondary sample from the final target selection (SECONDARY_v1_2_0) as well as to the secondary samples from several commissioning target selections. The galaxy does not have any ancillary bit (manga_target3). An manga_target3 bit set indicates the target belongs to one of the MaNGA Ancillary programs. <a id="mixin"></a> NSA, DAPall, and Value-Added catalog data For each target we can also access additional catalog data: the associated parameters from the NASA Sloan Atlas, and the DAPall file. The NSA catalog contains measured global galaxy photometric and shape profile information, while the DAPall catalog contains aggregrate statistics as measured by the MaNGA DAP pipeline. NSA properties are available on all Marvin objects under the nsa attribute. DAPall properties are available on all Marvin Maps and Modelcube objects. End of explanation # access the available VACs for this object vacs = cube.vacs vacs Explanation: SDSS members produce Value-Added Catalogs (VACS) of MaNGA products. These products can be incorporated into Marvin. Any incorporated VACS in Marvin are contained in the vacs attribute attached to each tool. More information on VACs in Marvin can be found here. Currently the MaNGA-HI VAC for DR15 has been contributed into Marvin. Let's take a look at it. End of explanation # access the mangahi vac hi = vacs.mangahi print(hi) print('data type:', type(hi)) # access any HI data available for 8485-1901 hi.data Explanation: vacs is a container (list) of all names of the vacs contributed into Marvin. These names are dottable to access to vac data. These contributed custom products can be of any format chosen by the VAC owner, a Python class, a dictionary, an array, etc. End of explanation # Check our remote cube we loaded earlier remote_cube # download the remote cube # remote_cube.download() # loading the cube a second time loads it from the new file new_cube = Cube('8485-1902') new_cube Explanation: <a id='download'></a> Downloading files While Marvin allows you to access data remotely, frequently you will find that you want to download the file associated to an object so that you can access it more quickly in the future. We can do that using the download method attached to all marvin core tools. Let’s try to load a cube that we know we do not have in out hard drive, and download it. End of explanation from marvin.utils.general import downloadList # create a list of target ids plateifus = ['8485-1901', '8485-1902', '7443-1901'] # download the DAP 2d MAPS files for these targets # dllist = downloadList(plateifus, dltype='maps') Explanation: Each tool contains a download method to download that individual file. If you want to download a list of targets in bulk, you can use the downloadList utility function. This function uses the built-in Python product sdss_access to download files from the Utah SAS. downloadList accepts as input a list of target ids, e.g. plateifus or mangaids. The dltype keyword indicates the type of file you wish to download. The default file is cube. Let's use the utility function to download a list of DAP MAPS files for three targets. End of explanation # extract the spaxel at array index x=10, y=15 spaxel = cube[15, 10] spaxel Explanation: <a id="spaxels"></a> Interacting with Spaxels All of the core Marvin Galaxy Tools have the ability to extract one ore more Spaxel objects. Spaxel objects are special Marvin objects to make it easier to deal with individual spaxels in galaxies and access their properties. See the Getting Started with Spaxels for more information on how to interact with Spaxels. We can either use the standard array slicing notation (0-indexed, origin of coordinates in the lower left corner of the array) End of explanation # get the spaxel at the center of the target galaxy central_spaxel = cube.getSpaxel(x=0, y=0) central_spaxel # we can plot the spectrum of the spaxel spaxel.flux.plot() Explanation: or we can use getSpaxel method, which accepts multiple arguments. Note that by default, (x, y) coordinates passed to getSpaxel are measured from the center of the array. End of explanation <END_TASK>
15,922
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: The big reset So I went ahead and cleared the memory. Step1: The mystery section remains the same. Step2: All the blocks are empty. Step3: The 'PresetStyle' settings are empty, too. Step4: Each of the registry settings are completely blank. Interesting things to note Step5: The only difference seems to be two bytes in the mystery section, at offsets 0x07 and 0x08. Perhaps this has to do with some kind of internal wear levelling or something. Registration extension Now that the memory has been cleared, we can hopefully figure out more about the registration settings. Recording Bank 3, Button 2 as the following settings Step6: I believe the only real way to get unrecorded settings is to reset the memory, which clears all the values to zero. This means that the first byte which has a value of 01 for all recorded settings can indeed be used as a flag... along with the FF byte at offset 24, and any other setting that cannot be set to a value of zero, such as the Pitch Bend range, Reverb type, Chorus type, and panel Sustain. Personally, I think it makes more sense for the first byte to act as the recorded flag, so I think I'll use that.
<ASSISTANT_TASK:> Python Code: import sys sys.path.append('..') import collections import mido from commons import dgxdump from commons.dumpdata import messages, songdata, regdata, regvalues old_syx_messages = mido.read_syx_file('../data/syxout5.syx') clear_syx_messages = mido.read_syx_file('../data/clear_bulk.txt') o_dump = dgxdump.DgxDump(old_syx_messages) c_dump = dgxdump.DgxDump(clear_syx_messages) # songs slices songslices = collections.OrderedDict([ ('songs', slice(0x00, 0x01)), ('mystery', slice(0x01, 0x15D)), ('tracks', slice(0x15D, 0x167)), ('durations', slice(0x167, 0x17B)), ('trackdurations', slice(0x17B, 0x1F3)), ('presetstyle', slice(0x1F3, 0x22F)), ('beginningblocks', slice(0x22F, 0x24D)), ('nextblocks', slice(0x24D, 0x2CF)), ('startmarker', slice(0x2CF, 0x2D5)), ('blockdata', slice(0x2D5, 0x106D5)), ('endmarker', slice(0x106D5, None)), ]) EXPECTED_SIZE = 0x106DB PRESETSTYLE = b'PresetStyle\0'*5 MARKER = b'PK0001' def hex_string(data): return " ".join("{:02X}".format(b) for b in data) def bin_string(data): return " ".join("{:08b}".format(b) for b in data) def line_hex(data, head=None, tail=0): if head is None: head = len(data) tailstart = len(data) - tail if tailstart <= head: return (hex_string(data)) else: return ("{} .. {}".format(hex_string(data[:head]), hex_string(data[tailstart:]))) def song_section(dump, section): return dump.song_data.data[songslices[section]] for sec in songslices: print(sec) print(line_hex(song_section(o_dump, sec), 32, 4)) print(line_hex(song_section(c_dump, sec), 32, 4)) song_section(o_dump, 'mystery') == song_section(c_dump, 'mystery') Explanation: The big reset So I went ahead and cleared the memory. End of explanation all(b==0 for b in song_section(c_dump, 'nextblocks')) all(b==0 for b in song_section(c_dump, 'blockdata')) Explanation: The mystery section remains the same. End of explanation bytes(song_section(c_dump, 'presetstyle')) Explanation: All the blocks are empty. End of explanation print(line_hex(o_dump.reg_data.data, 32, 4)) print(line_hex(c_dump.reg_data.data, 32, 4)) for bank in range(1, 8+1): for button in range(1, 2+1): print(bank, button) print(line_hex(o_dump.reg_data.settings.get_setting(bank, button).data)) print(line_hex(c_dump.reg_data.settings.get_setting(bank, button).data)) Explanation: The 'PresetStyle' settings are empty, too. End of explanation for x in range(2, 7): !diff -qs ../data/backup_experiment/cb1.txt ../data/backup_experiment/cb{x}.txt !diff -qs ../data/backup_experiment/cb1.txt ../data/clear_bulk.txt c2_syx_messages = mido.read_syx_file('../data/backup_experiment/cb1.txt') c2_dump = dgxdump.DgxDump(c2_syx_messages) c_dump.song_data.data == c2_dump.song_data.data c_dump.reg_data.data == c2_dump.reg_data.data for sec in songslices: c_sec = song_section(c_dump, sec) c2_sec = song_section(c2_dump, sec) if c_sec != c2_sec: print(sec) print(line_hex(c_sec, 32, 4)) print(line_hex(c2_sec, 32, 4)) for n, (a, b) in enumerate(zip(c_dump.song_data.data, c2_dump.song_data.data)): if a != b: print("{0:02X}: {1:02X} {2:02X} ({1:03d} {2:03d})".format(n, a, b)) Explanation: Each of the registry settings are completely blank. Interesting things to note: the first byte is 0 instead of 1, which probably indicates that the setting is unused. The bytes that were FF in each recorded setting are 00 here. Investigating FUNCTION backup According to the manual (page 49), the following settings can be saved to backup, i.e. persistent memory for startup bu holding the FUNCTION button: User songs (These are saved when recorded anyway) Style files (the ones loaded using SmartMedia) Touch response (ON/OFF) Registration memory These function settings: Tuning Split point Touch sensitivity Style volume Song volume Metronome volume Grade Demo cancel Language Media Select Panel Sustain. These backup settings are also cleared with the rest of the memory. The default values for these settings are as follows: | setting | default | |-------------------|--------------| | Touch response | ON | | Tuning | 000 | | Split point | 54 (F#2) | | Touch sensitivity | 2 (Medium) | | Style volume | 100 | | Song volume | 100 | | Metronome volume | 100 | | Grade | ON | | Demo cancel | OFF | | Language | English | | Media Select | Flash Memory | | Panel sustain | OFF | As an experiment, I changed the values of the function settings: | setting | new value | |-------------------|--------------| | Touch response | ON | | Tuning | 057 | | Split point | 112 (E7) | | Touch sensitivity | 3 (Hard) | | Style volume | 045 | | Song volume | 079 | | Metronome volume | 121 | | Grade | OFF | | Demo cancel | ON | | Language | Japanese | | Media Select | Smart Media | | Panel sustain | ON | and without making a backup: - took a bulk dump. (cb1.txt), - then made the backup, took another bulk dump, (cb2.txt), - restarted with the new settings, took another (cb3.txt), - reset everything to default without backup (cb4.txt), - made a backup again and took another dump (cb5.txt), - then restarted again (cb6.txt). All of these files were identical to each other, which suggests that these backup settings are not stored any part we can retrieve. However, there is one thing interesting about these files, in that they differ from the dump I got immediately after resetting the memory (clear_bulk.txt). End of explanation r1_dump = dgxdump.DgxDump(mido.read_syx_file('../data/post_clear/1reg.syx')) c2_dump.song_data.data == r1_dump.song_data.data c2_dump.reg_data.data == r1_dump.reg_data.data for bank in range(1, 8+1): for button in range(1, 2+1): if not all(x == 0 for x in r1_dump.reg_data.settings.get_setting(bank, button).data): print(bank, button) line_hex(r1_dump.reg_data.settings.get_setting(3, 2).data) for bb in [(3, 2), (1, 1)]: sets = r1_dump.reg_data.settings.get_setting(*bb) print(line_hex(sets.data)) sets.print_settings() sets.print_unusual() Explanation: The only difference seems to be two bytes in the mystery section, at offsets 0x07 and 0x08. Perhaps this has to do with some kind of internal wear levelling or something. Registration extension Now that the memory has been cleared, we can hopefully figure out more about the registration settings. Recording Bank 3, Button 2 as the following settings: | setting | value | |------------------|-------| | Style | 092 | | Accompaniment | ON | | Split point | 053 | | Main A/B | A | | Style vol | 050 | | Main voice | 060 | | Main Octave | -1 | | Main Volume | 054 | | Main Pan | 092 | | Main Reverb | 078 | | Main Chorus | 103 | | Split | ON | | Split voice | 003 | | Split Octave | 0 | | Split Volume | 108 | | Split Pan | 064 | | Split Reverb | 032 | | Split Chorus | 127 | | Dual | OFF | | Dual voice | 201 | | Dual Octave | +2 | | Dual Volume | 095 | | Dual Pan | 048 | | Dual Reverb | 017 | | Dual Chorus | 082 | | Pitch bend range | 05 | | Reverb type | --(Room) | | Chorus type | --(Celeste) | | Harmony | OFF | | Harmony type | 06(Trill1/4) | | Harmony volume | 085/---* | | Transpose | +03 | | Tempo | 080 | | Panel Sustain | ON | *This was set using a different Harmony type setting. End of explanation r2_dump = dgxdump.DgxDump(mido.read_syx_file('../data/post_clear/2reg.txt')) sets = r2_dump.reg_data.settings.get_setting(2,2) sets.print_settings() sets.print_unusual() Explanation: I believe the only real way to get unrecorded settings is to reset the memory, which clears all the values to zero. This means that the first byte which has a value of 01 for all recorded settings can indeed be used as a flag... along with the FF byte at offset 24, and any other setting that cannot be set to a value of zero, such as the Pitch Bend range, Reverb type, Chorus type, and panel Sustain. Personally, I think it makes more sense for the first byte to act as the recorded flag, so I think I'll use that. End of explanation <END_TASK>
15,923
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Numpy Exercise 4 Imports Step1: Complete graph Laplacian In discrete mathematics a Graph is a set of vertices or nodes that are connected to each other by edges or lines. If those edges don't have directionality, the graph is said to be undirected. Graphs are used to model social and communications networks (Twitter, Facebook, Internet) as well as natural systems such as molecules. A Complete Graph, $K_n$ on $n$ nodes has an edge that connects each node to every other node. Here is $K_5$ Step3: The Laplacian Matrix is a matrix that is extremely important in graph theory and numerical analysis. It is defined as $L=D-A$. Where $D$ is the degree matrix and $A$ is the adjecency matrix. For the purpose of this problem you don't need to understand the details of these matrices, although their definitions are relatively simple. The degree matrix for $K_n$ is an $n \times n$ diagonal matrix with the value $n-1$ along the diagonal and zeros everywhere else. Write a function to compute the degree matrix for $K_n$ using NumPy. Step5: The adjacency matrix for $K_n$ is an $n \times n$ matrix with zeros along the diagonal and ones everywhere else. Write a function to compute the adjacency matrix for $K_n$ using NumPy. Step6: Use NumPy to explore the eigenvalues or spectrum of the Laplacian L of $K_n$. What patterns do you notice as $n$ changes? Create a conjecture about the general Laplace spectrum of $K_n$.
<ASSISTANT_TASK:> Python Code: import numpy as np %matplotlib inline import matplotlib.pyplot as plt import seaborn as sns Explanation: Numpy Exercise 4 Imports End of explanation import networkx as nx K_5=nx.complete_graph(5) nx.draw(K_5) Explanation: Complete graph Laplacian In discrete mathematics a Graph is a set of vertices or nodes that are connected to each other by edges or lines. If those edges don't have directionality, the graph is said to be undirected. Graphs are used to model social and communications networks (Twitter, Facebook, Internet) as well as natural systems such as molecules. A Complete Graph, $K_n$ on $n$ nodes has an edge that connects each node to every other node. Here is $K_5$: End of explanation def complete_deg(n): Return the integer valued degree matrix D for the complete graph K_n. f=np.zeros((n,n), dtype=int) #creates integer matrix of size nxn t=np.diag(np.diag(f)+(n-1)) #makes the diagonal n-1 and zeros elsewhere return t D = complete_deg(5) assert D.shape==(5,5) assert D.dtype==np.dtype(int) assert np.all(D.diagonal()==4*np.ones(5)) assert np.all(D-np.diag(D.diagonal())==np.zeros((5,5),dtype=int)) Explanation: The Laplacian Matrix is a matrix that is extremely important in graph theory and numerical analysis. It is defined as $L=D-A$. Where $D$ is the degree matrix and $A$ is the adjecency matrix. For the purpose of this problem you don't need to understand the details of these matrices, although their definitions are relatively simple. The degree matrix for $K_n$ is an $n \times n$ diagonal matrix with the value $n-1$ along the diagonal and zeros everywhere else. Write a function to compute the degree matrix for $K_n$ using NumPy. End of explanation def complete_adj(n): Return the integer valued adjacency matrix A for the complete graph K_n. x=np.zeros((n,n), dtype=int) #creates integer matrix of size nxn o=np.ones((n,n), dtype=int) #makes the diagonal 0 and ones elsewhere return o+np.diag(np.diag(x)+-1) A = complete_adj(5) assert A.shape==(5,5) assert A.dtype==np.dtype(int) assert np.all(A+np.eye(5,dtype=int)==np.ones((5,5),dtype=int)) Explanation: The adjacency matrix for $K_n$ is an $n \times n$ matrix with zeros along the diagonal and ones everywhere else. Write a function to compute the adjacency matrix for $K_n$ using NumPy. End of explanation d=complete_deg(5) #calling function a=complete_adj(5) #calling function L=d-a #finding difference np.linalg.eigvals(L) #eigenvalue answer Explanation: Use NumPy to explore the eigenvalues or spectrum of the Laplacian L of $K_n$. What patterns do you notice as $n$ changes? Create a conjecture about the general Laplace spectrum of $K_n$. End of explanation <END_TASK>
15,924
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: ES-DOC CMIP6 Model Properties - Ocean MIP Era Step1: Document Authors Set document authors Step2: Document Contributors Specify document contributors Step3: Document Publication Specify document publication status Step4: Document Table of Contents 1. Key Properties 2. Key Properties --&gt; Seawater Properties 3. Key Properties --&gt; Bathymetry 4. Key Properties --&gt; Nonoceanic Waters 5. Key Properties --&gt; Software Properties 6. Key Properties --&gt; Resolution 7. Key Properties --&gt; Tuning Applied 8. Key Properties --&gt; Conservation 9. Grid 10. Grid --&gt; Discretisation --&gt; Vertical 11. Grid --&gt; Discretisation --&gt; Horizontal 12. Timestepping Framework 13. Timestepping Framework --&gt; Tracers 14. Timestepping Framework --&gt; Baroclinic Dynamics 15. Timestepping Framework --&gt; Barotropic 16. Timestepping Framework --&gt; Vertical Physics 17. Advection 18. Advection --&gt; Momentum 19. Advection --&gt; Lateral Tracers 20. Advection --&gt; Vertical Tracers 21. Lateral Physics 22. Lateral Physics --&gt; Momentum --&gt; Operator 23. Lateral Physics --&gt; Momentum --&gt; Eddy Viscosity Coeff 24. Lateral Physics --&gt; Tracers 25. Lateral Physics --&gt; Tracers --&gt; Operator 26. Lateral Physics --&gt; Tracers --&gt; Eddy Diffusity Coeff 27. Lateral Physics --&gt; Tracers --&gt; Eddy Induced Velocity 28. Vertical Physics 29. Vertical Physics --&gt; Boundary Layer Mixing --&gt; Details 30. Vertical Physics --&gt; Boundary Layer Mixing --&gt; Tracers 31. Vertical Physics --&gt; Boundary Layer Mixing --&gt; Momentum 32. Vertical Physics --&gt; Interior Mixing --&gt; Details 33. Vertical Physics --&gt; Interior Mixing --&gt; Tracers 34. Vertical Physics --&gt; Interior Mixing --&gt; Momentum 35. Uplow Boundaries --&gt; Free Surface 36. Uplow Boundaries --&gt; Bottom Boundary Layer 37. Boundary Forcing 38. Boundary Forcing --&gt; Momentum --&gt; Bottom Friction 39. Boundary Forcing --&gt; Momentum --&gt; Lateral Friction 40. Boundary Forcing --&gt; Tracers --&gt; Sunlight Penetration 41. Boundary Forcing --&gt; Tracers --&gt; Fresh Water Forcing 1. Key Properties Ocean key properties 1.1. Model Overview Is Required Step5: 1.2. Model Name Is Required Step6: 1.3. Model Family Is Required Step7: 1.4. Basic Approximations Is Required Step8: 1.5. Prognostic Variables Is Required Step9: 2. Key Properties --&gt; Seawater Properties Physical properties of seawater in ocean 2.1. Eos Type Is Required Step10: 2.2. Eos Functional Temp Is Required Step11: 2.3. Eos Functional Salt Is Required Step12: 2.4. Eos Functional Depth Is Required Step13: 2.5. Ocean Freezing Point Is Required Step14: 2.6. Ocean Specific Heat Is Required Step15: 2.7. Ocean Reference Density Is Required Step16: 3. Key Properties --&gt; Bathymetry Properties of bathymetry in ocean 3.1. Reference Dates Is Required Step17: 3.2. Type Is Required Step18: 3.3. Ocean Smoothing Is Required Step19: 3.4. Source Is Required Step20: 4. Key Properties --&gt; Nonoceanic Waters Non oceanic waters treatement in ocean 4.1. Isolated Seas Is Required Step21: 4.2. River Mouth Is Required Step22: 5. Key Properties --&gt; Software Properties Software properties of ocean code 5.1. Repository Is Required Step23: 5.2. Code Version Is Required Step24: 5.3. Code Languages Is Required Step25: 6. Key Properties --&gt; Resolution Resolution in the ocean grid 6.1. Name Is Required Step26: 6.2. Canonical Horizontal Resolution Is Required Step27: 6.3. Range Horizontal Resolution Is Required Step28: 6.4. Number Of Horizontal Gridpoints Is Required Step29: 6.5. Number Of Vertical Levels Is Required Step30: 6.6. Is Adaptive Grid Is Required Step31: 6.7. Thickness Level 1 Is Required Step32: 7. Key Properties --&gt; Tuning Applied Tuning methodology for ocean component 7.1. Description Is Required Step33: 7.2. Global Mean Metrics Used Is Required Step34: 7.3. Regional Metrics Used Is Required Step35: 7.4. Trend Metrics Used Is Required Step36: 8. Key Properties --&gt; Conservation Conservation in the ocean component 8.1. Description Is Required Step37: 8.2. Scheme Is Required Step38: 8.3. Consistency Properties Is Required Step39: 8.4. Corrected Conserved Prognostic Variables Is Required Step40: 8.5. Was Flux Correction Used Is Required Step41: 9. Grid Ocean grid 9.1. Overview Is Required Step42: 10. Grid --&gt; Discretisation --&gt; Vertical Properties of vertical discretisation in ocean 10.1. Coordinates Is Required Step43: 10.2. Partial Steps Is Required Step44: 11. Grid --&gt; Discretisation --&gt; Horizontal Type of horizontal discretisation scheme in ocean 11.1. Type Is Required Step45: 11.2. Staggering Is Required Step46: 11.3. Scheme Is Required Step47: 12. Timestepping Framework Ocean Timestepping Framework 12.1. Overview Is Required Step48: 12.2. Diurnal Cycle Is Required Step49: 13. Timestepping Framework --&gt; Tracers Properties of tracers time stepping in ocean 13.1. Scheme Is Required Step50: 13.2. Time Step Is Required Step51: 14. Timestepping Framework --&gt; Baroclinic Dynamics Baroclinic dynamics in ocean 14.1. Type Is Required Step52: 14.2. Scheme Is Required Step53: 14.3. Time Step Is Required Step54: 15. Timestepping Framework --&gt; Barotropic Barotropic time stepping in ocean 15.1. Splitting Is Required Step55: 15.2. Time Step Is Required Step56: 16. Timestepping Framework --&gt; Vertical Physics Vertical physics time stepping in ocean 16.1. Method Is Required Step57: 17. Advection Ocean advection 17.1. Overview Is Required Step58: 18. Advection --&gt; Momentum Properties of lateral momemtum advection scheme in ocean 18.1. Type Is Required Step59: 18.2. Scheme Name Is Required Step60: 18.3. ALE Is Required Step61: 19. Advection --&gt; Lateral Tracers Properties of lateral tracer advection scheme in ocean 19.1. Order Is Required Step62: 19.2. Flux Limiter Is Required Step63: 19.3. Effective Order Is Required Step64: 19.4. Name Is Required Step65: 19.5. Passive Tracers Is Required Step66: 19.6. Passive Tracers Advection Is Required Step67: 20. Advection --&gt; Vertical Tracers Properties of vertical tracer advection scheme in ocean 20.1. Name Is Required Step68: 20.2. Flux Limiter Is Required Step69: 21. Lateral Physics Ocean lateral physics 21.1. Overview Is Required Step70: 21.2. Scheme Is Required Step71: 22. Lateral Physics --&gt; Momentum --&gt; Operator Properties of lateral physics operator for momentum in ocean 22.1. Direction Is Required Step72: 22.2. Order Is Required Step73: 22.3. Discretisation Is Required Step74: 23. Lateral Physics --&gt; Momentum --&gt; Eddy Viscosity Coeff Properties of eddy viscosity coeff in lateral physics momemtum scheme in the ocean 23.1. Type Is Required Step75: 23.2. Constant Coefficient Is Required Step76: 23.3. Variable Coefficient Is Required Step77: 23.4. Coeff Background Is Required Step78: 23.5. Coeff Backscatter Is Required Step79: 24. Lateral Physics --&gt; Tracers Properties of lateral physics for tracers in ocean 24.1. Mesoscale Closure Is Required Step80: 24.2. Submesoscale Mixing Is Required Step81: 25. Lateral Physics --&gt; Tracers --&gt; Operator Properties of lateral physics operator for tracers in ocean 25.1. Direction Is Required Step82: 25.2. Order Is Required Step83: 25.3. Discretisation Is Required Step84: 26. Lateral Physics --&gt; Tracers --&gt; Eddy Diffusity Coeff Properties of eddy diffusity coeff in lateral physics tracers scheme in the ocean 26.1. Type Is Required Step85: 26.2. Constant Coefficient Is Required Step86: 26.3. Variable Coefficient Is Required Step87: 26.4. Coeff Background Is Required Step88: 26.5. Coeff Backscatter Is Required Step89: 27. Lateral Physics --&gt; Tracers --&gt; Eddy Induced Velocity Properties of eddy induced velocity (EIV) in lateral physics tracers scheme in the ocean 27.1. Type Is Required Step90: 27.2. Constant Val Is Required Step91: 27.3. Flux Type Is Required Step92: 27.4. Added Diffusivity Is Required Step93: 28. Vertical Physics Ocean Vertical Physics 28.1. Overview Is Required Step94: 29. Vertical Physics --&gt; Boundary Layer Mixing --&gt; Details Properties of vertical physics in ocean 29.1. Langmuir Cells Mixing Is Required Step95: 30. Vertical Physics --&gt; Boundary Layer Mixing --&gt; Tracers *Properties of boundary layer (BL) mixing on tracers in the ocean * 30.1. Type Is Required Step96: 30.2. Closure Order Is Required Step97: 30.3. Constant Is Required Step98: 30.4. Background Is Required Step99: 31. Vertical Physics --&gt; Boundary Layer Mixing --&gt; Momentum *Properties of boundary layer (BL) mixing on momentum in the ocean * 31.1. Type Is Required Step100: 31.2. Closure Order Is Required Step101: 31.3. Constant Is Required Step102: 31.4. Background Is Required Step103: 32. Vertical Physics --&gt; Interior Mixing --&gt; Details *Properties of interior mixing in the ocean * 32.1. Convection Type Is Required Step104: 32.2. Tide Induced Mixing Is Required Step105: 32.3. Double Diffusion Is Required Step106: 32.4. Shear Mixing Is Required Step107: 33. Vertical Physics --&gt; Interior Mixing --&gt; Tracers *Properties of interior mixing on tracers in the ocean * 33.1. Type Is Required Step108: 33.2. Constant Is Required Step109: 33.3. Profile Is Required Step110: 33.4. Background Is Required Step111: 34. Vertical Physics --&gt; Interior Mixing --&gt; Momentum *Properties of interior mixing on momentum in the ocean * 34.1. Type Is Required Step112: 34.2. Constant Is Required Step113: 34.3. Profile Is Required Step114: 34.4. Background Is Required Step115: 35. Uplow Boundaries --&gt; Free Surface Properties of free surface in ocean 35.1. Overview Is Required Step116: 35.2. Scheme Is Required Step117: 35.3. Embeded Seaice Is Required Step118: 36. Uplow Boundaries --&gt; Bottom Boundary Layer Properties of bottom boundary layer in ocean 36.1. Overview Is Required Step119: 36.2. Type Of Bbl Is Required Step120: 36.3. Lateral Mixing Coef Is Required Step121: 36.4. Sill Overflow Is Required Step122: 37. Boundary Forcing Ocean boundary forcing 37.1. Overview Is Required Step123: 37.2. Surface Pressure Is Required Step124: 37.3. Momentum Flux Correction Is Required Step125: 37.4. Tracers Flux Correction Is Required Step126: 37.5. Wave Effects Is Required Step127: 37.6. River Runoff Budget Is Required Step128: 37.7. Geothermal Heating Is Required Step129: 38. Boundary Forcing --&gt; Momentum --&gt; Bottom Friction Properties of momentum bottom friction in ocean 38.1. Type Is Required Step130: 39. Boundary Forcing --&gt; Momentum --&gt; Lateral Friction Properties of momentum lateral friction in ocean 39.1. Type Is Required Step131: 40. Boundary Forcing --&gt; Tracers --&gt; Sunlight Penetration Properties of sunlight penetration scheme in ocean 40.1. Scheme Is Required Step132: 40.2. Ocean Colour Is Required Step133: 40.3. Extinction Depth Is Required Step134: 41. Boundary Forcing --&gt; Tracers --&gt; Fresh Water Forcing Properties of surface fresh water forcing in ocean 41.1. From Atmopshere Is Required Step135: 41.2. From Sea Ice Is Required Step136: 41.3. Forced Mode Restoring Is Required
<ASSISTANT_TASK:> Python Code: # DO NOT EDIT ! from pyesdoc.ipython.model_topic import NotebookOutput # DO NOT EDIT ! DOC = NotebookOutput('cmip6', 'messy-consortium', 'sandbox-2', 'ocean') Explanation: ES-DOC CMIP6 Model Properties - Ocean MIP Era: CMIP6 Institute: MESSY-CONSORTIUM Source ID: SANDBOX-2 Topic: Ocean Sub-Topics: Timestepping Framework, Advection, Lateral Physics, Vertical Physics, Uplow Boundaries, Boundary Forcing. Properties: 133 (101 required) Model descriptions: Model description details Initialized From: -- Notebook Help: Goto notebook help page Notebook Initialised: 2018-02-15 16:54:10 Document Setup IMPORTANT: to be executed each time you run the notebook End of explanation # Set as follows: DOC.set_author("name", "email") # TODO - please enter value(s) Explanation: Document Authors Set document authors End of explanation # Set as follows: DOC.set_contributor("name", "email") # TODO - please enter value(s) Explanation: Document Contributors Specify document contributors End of explanation # Set publication status: # 0=do not publish, 1=publish. DOC.set_publication_status(0) Explanation: Document Publication Specify document publication status End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.model_overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: Document Table of Contents 1. Key Properties 2. Key Properties --&gt; Seawater Properties 3. Key Properties --&gt; Bathymetry 4. Key Properties --&gt; Nonoceanic Waters 5. Key Properties --&gt; Software Properties 6. Key Properties --&gt; Resolution 7. Key Properties --&gt; Tuning Applied 8. Key Properties --&gt; Conservation 9. Grid 10. Grid --&gt; Discretisation --&gt; Vertical 11. Grid --&gt; Discretisation --&gt; Horizontal 12. Timestepping Framework 13. Timestepping Framework --&gt; Tracers 14. Timestepping Framework --&gt; Baroclinic Dynamics 15. Timestepping Framework --&gt; Barotropic 16. Timestepping Framework --&gt; Vertical Physics 17. Advection 18. Advection --&gt; Momentum 19. Advection --&gt; Lateral Tracers 20. Advection --&gt; Vertical Tracers 21. Lateral Physics 22. Lateral Physics --&gt; Momentum --&gt; Operator 23. Lateral Physics --&gt; Momentum --&gt; Eddy Viscosity Coeff 24. Lateral Physics --&gt; Tracers 25. Lateral Physics --&gt; Tracers --&gt; Operator 26. Lateral Physics --&gt; Tracers --&gt; Eddy Diffusity Coeff 27. Lateral Physics --&gt; Tracers --&gt; Eddy Induced Velocity 28. Vertical Physics 29. Vertical Physics --&gt; Boundary Layer Mixing --&gt; Details 30. Vertical Physics --&gt; Boundary Layer Mixing --&gt; Tracers 31. Vertical Physics --&gt; Boundary Layer Mixing --&gt; Momentum 32. Vertical Physics --&gt; Interior Mixing --&gt; Details 33. Vertical Physics --&gt; Interior Mixing --&gt; Tracers 34. Vertical Physics --&gt; Interior Mixing --&gt; Momentum 35. Uplow Boundaries --&gt; Free Surface 36. Uplow Boundaries --&gt; Bottom Boundary Layer 37. Boundary Forcing 38. Boundary Forcing --&gt; Momentum --&gt; Bottom Friction 39. Boundary Forcing --&gt; Momentum --&gt; Lateral Friction 40. Boundary Forcing --&gt; Tracers --&gt; Sunlight Penetration 41. Boundary Forcing --&gt; Tracers --&gt; Fresh Water Forcing 1. Key Properties Ocean key properties 1.1. Model Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of ocean model. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.model_name') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 1.2. Model Name Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Name of ocean model code (NEMO 3.6, MOM 5.0,...) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.model_family') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "OGCM" # "slab ocean" # "mixed layer ocean" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 1.3. Model Family Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of ocean model. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.basic_approximations') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "Primitive equations" # "Non-hydrostatic" # "Boussinesq" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 1.4. Basic Approximations Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N Basic approximations made in the ocean. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.prognostic_variables') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "Potential temperature" # "Conservative temperature" # "Salinity" # "U-velocity" # "V-velocity" # "W-velocity" # "SSH" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 1.5. Prognostic Variables Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N List of prognostic variables in the ocean component. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.seawater_properties.eos_type') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Linear" # "Wright, 1997" # "Mc Dougall et al." # "Jackett et al. 2006" # "TEOS 2010" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 2. Key Properties --&gt; Seawater Properties Physical properties of seawater in ocean 2.1. Eos Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of EOS for sea water End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.seawater_properties.eos_functional_temp') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Potential temperature" # "Conservative temperature" # TODO - please enter value(s) Explanation: 2.2. Eos Functional Temp Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Temperature used in EOS for sea water End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.seawater_properties.eos_functional_salt') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Practical salinity Sp" # "Absolute salinity Sa" # TODO - please enter value(s) Explanation: 2.3. Eos Functional Salt Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Salinity used in EOS for sea water End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.seawater_properties.eos_functional_depth') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Pressure (dbars)" # "Depth (meters)" # TODO - please enter value(s) Explanation: 2.4. Eos Functional Depth Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Depth or pressure used in EOS for sea water ? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.seawater_properties.ocean_freezing_point') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "TEOS 2010" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 2.5. Ocean Freezing Point Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Equation used to compute the freezing point (in deg C) of seawater, as a function of salinity and pressure End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.seawater_properties.ocean_specific_heat') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 2.6. Ocean Specific Heat Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: FLOAT&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Specific heat in ocean (cpocean) in J/(kg K) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.seawater_properties.ocean_reference_density') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 2.7. Ocean Reference Density Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: FLOAT&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Boussinesq reference density (rhozero) in kg / m3 End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.bathymetry.reference_dates') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Present day" # "21000 years BP" # "6000 years BP" # "LGM" # "Pliocene" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 3. Key Properties --&gt; Bathymetry Properties of bathymetry in ocean 3.1. Reference Dates Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Reference date of bathymetry End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.bathymetry.type') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 3.2. Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is the bathymetry fixed in time in the ocean ? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.bathymetry.ocean_smoothing') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 3.3. Ocean Smoothing Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe any smoothing or hand editing of bathymetry in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.bathymetry.source') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 3.4. Source Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe source of bathymetry in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.nonoceanic_waters.isolated_seas') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 4. Key Properties --&gt; Nonoceanic Waters Non oceanic waters treatement in ocean 4.1. Isolated Seas Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe if/how isolated seas is performed End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.nonoceanic_waters.river_mouth') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 4.2. River Mouth Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe if/how river mouth mixing or estuaries specific treatment is performed End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.software_properties.repository') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 5. Key Properties --&gt; Software Properties Software properties of ocean code 5.1. Repository Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Location of code for this component. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.software_properties.code_version') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 5.2. Code Version Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Code version identifier. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.software_properties.code_languages') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 5.3. Code Languages Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N Code language(s). End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.resolution.name') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 6. Key Properties --&gt; Resolution Resolution in the ocean grid 6.1. Name Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 This is a string usually used by the modelling group to describe the resolution of this grid, e.g. ORCA025, N512L180, T512L70 etc. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.resolution.canonical_horizontal_resolution') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 6.2. Canonical Horizontal Resolution Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Expression quoted for gross comparisons of resolution, eg. 50km or 0.1 degrees etc. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.resolution.range_horizontal_resolution') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 6.3. Range Horizontal Resolution Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Range of horizontal resolution with spatial details, eg. 50(Equator)-100km or 0.1-0.5 degrees etc. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.resolution.number_of_horizontal_gridpoints') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 6.4. Number Of Horizontal Gridpoints Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Total number of horizontal (XY) points (or degrees of freedom) on computational grid. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.resolution.number_of_vertical_levels') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 6.5. Number Of Vertical Levels Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Number of vertical levels resolved on computational grid. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.resolution.is_adaptive_grid') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 6.6. Is Adaptive Grid Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Default is False. Set true if grid resolution changes during execution. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.resolution.thickness_level_1') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 6.7. Thickness Level 1 Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: FLOAT&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Thickness of first surface ocean level (in meters) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.tuning_applied.description') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 7. Key Properties --&gt; Tuning Applied Tuning methodology for ocean component 7.1. Description Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 General overview description of tuning: explain and motivate the main targets and metrics retained. &amp;Document the relative weight given to climate performance metrics versus process oriented metrics, &amp;and on the possible conflicts with parameterization level tuning. In particular describe any struggle &amp;with a parameter value that required pushing it to its limits to solve a particular model deficiency. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.tuning_applied.global_mean_metrics_used') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 7.2. Global Mean Metrics Used Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N List set of metrics of the global mean state used in tuning model/component End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.tuning_applied.regional_metrics_used') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 7.3. Regional Metrics Used Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N List of regional metrics of mean state (e.g THC, AABW, regional means etc) used in tuning model/component End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.tuning_applied.trend_metrics_used') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 7.4. Trend Metrics Used Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N List observed trend metrics used in tuning model/component End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.conservation.description') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 8. Key Properties --&gt; Conservation Conservation in the ocean component 8.1. Description Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Brief description of conservation methodology End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.conservation.scheme') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "Energy" # "Enstrophy" # "Salt" # "Volume of ocean" # "Momentum" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 8.2. Scheme Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N Properties conserved in the ocean by the numerical schemes End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.conservation.consistency_properties') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 8.3. Consistency Properties Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Any additional consistency properties (energy conversion, pressure gradient discretisation, ...)? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.conservation.corrected_conserved_prognostic_variables') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 8.4. Corrected Conserved Prognostic Variables Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Set of variables which are conserved by more than the numerical scheme alone. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.key_properties.conservation.was_flux_correction_used') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 8.5. Was Flux Correction Used Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Does conservation involve flux correction ? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.grid.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 9. Grid Ocean grid 9.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of grid in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.grid.discretisation.vertical.coordinates') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Z-coordinate" # "Z*-coordinate" # "S-coordinate" # "Isopycnic - sigma 0" # "Isopycnic - sigma 2" # "Isopycnic - sigma 4" # "Isopycnic - other" # "Hybrid / Z+S" # "Hybrid / Z+isopycnic" # "Hybrid / other" # "Pressure referenced (P)" # "P*" # "Z**" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 10. Grid --&gt; Discretisation --&gt; Vertical Properties of vertical discretisation in ocean 10.1. Coordinates Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of vertical coordinates in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.grid.discretisation.vertical.partial_steps') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 10.2. Partial Steps Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Using partial steps with Z or Z vertical coordinate in ocean ?* End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.grid.discretisation.horizontal.type') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Lat-lon" # "Rotated north pole" # "Two north poles (ORCA-style)" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 11. Grid --&gt; Discretisation --&gt; Horizontal Type of horizontal discretisation scheme in ocean 11.1. Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Horizontal grid type End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.grid.discretisation.horizontal.staggering') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Arakawa B-grid" # "Arakawa C-grid" # "Arakawa E-grid" # "N/a" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 11.2. Staggering Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Horizontal grid staggering type End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.grid.discretisation.horizontal.scheme') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Finite difference" # "Finite volumes" # "Finite elements" # "Unstructured grid" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 11.3. Scheme Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Horizontal discretisation scheme in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.timestepping_framework.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 12. Timestepping Framework Ocean Timestepping Framework 12.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of time stepping in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.timestepping_framework.diurnal_cycle') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "None" # "Via coupling" # "Specific treatment" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 12.2. Diurnal Cycle Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Diurnal cycle type End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.timestepping_framework.tracers.scheme') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Leap-frog + Asselin filter" # "Leap-frog + Periodic Euler" # "Predictor-corrector" # "Runge-Kutta 2" # "AM3-LF" # "Forward-backward" # "Forward operator" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 13. Timestepping Framework --&gt; Tracers Properties of tracers time stepping in ocean 13.1. Scheme Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Tracers time stepping scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.timestepping_framework.tracers.time_step') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 13.2. Time Step Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Tracers time step (in seconds) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.timestepping_framework.baroclinic_dynamics.type') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Preconditioned conjugate gradient" # "Sub cyling" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 14. Timestepping Framework --&gt; Baroclinic Dynamics Baroclinic dynamics in ocean 14.1. Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Baroclinic dynamics type End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.timestepping_framework.baroclinic_dynamics.scheme') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Leap-frog + Asselin filter" # "Leap-frog + Periodic Euler" # "Predictor-corrector" # "Runge-Kutta 2" # "AM3-LF" # "Forward-backward" # "Forward operator" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 14.2. Scheme Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Baroclinic dynamics scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.timestepping_framework.baroclinic_dynamics.time_step') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 14.3. Time Step Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Baroclinic time step (in seconds) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.timestepping_framework.barotropic.splitting') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "None" # "split explicit" # "implicit" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 15. Timestepping Framework --&gt; Barotropic Barotropic time stepping in ocean 15.1. Splitting Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Time splitting method End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.timestepping_framework.barotropic.time_step') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 15.2. Time Step Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Barotropic time step (in seconds) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.timestepping_framework.vertical_physics.method') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 16. Timestepping Framework --&gt; Vertical Physics Vertical physics time stepping in ocean 16.1. Method Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Details of vertical time stepping in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.advection.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 17. Advection Ocean advection 17.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of advection in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.advection.momentum.type') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Flux form" # "Vector form" # TODO - please enter value(s) Explanation: 18. Advection --&gt; Momentum Properties of lateral momemtum advection scheme in ocean 18.1. Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of lateral momemtum advection scheme in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.advection.momentum.scheme_name') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 18.2. Scheme Name Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Name of ocean momemtum advection scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.advection.momentum.ALE') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 18.3. ALE Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Using ALE for vertical advection ? (if vertical coordinates are sigma) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.advection.lateral_tracers.order') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 19. Advection --&gt; Lateral Tracers Properties of lateral tracer advection scheme in ocean 19.1. Order Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Order of lateral tracer advection scheme in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.advection.lateral_tracers.flux_limiter') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 19.2. Flux Limiter Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Monotonic flux limiter for lateral tracer advection scheme in ocean ? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.advection.lateral_tracers.effective_order') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 19.3. Effective Order Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: FLOAT&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Effective order of limited lateral tracer advection scheme in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.advection.lateral_tracers.name') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 19.4. Name Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Descriptive text for lateral tracer advection scheme in ocean (e.g. MUSCL, PPM-H5, PRATHER,...) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.advection.lateral_tracers.passive_tracers') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "Ideal age" # "CFC 11" # "CFC 12" # "SF6" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 19.5. Passive Tracers Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N Passive tracers advected End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.advection.lateral_tracers.passive_tracers_advection') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 19.6. Passive Tracers Advection Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Is advection of passive tracers different than active ? if so, describe. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.advection.vertical_tracers.name') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 20. Advection --&gt; Vertical Tracers Properties of vertical tracer advection scheme in ocean 20.1. Name Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Descriptive text for vertical tracer advection scheme in ocean (e.g. MUSCL, PPM-H5, PRATHER,...) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.advection.vertical_tracers.flux_limiter') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 20.2. Flux Limiter Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Monotonic flux limiter for vertical tracer advection scheme in ocean ? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 21. Lateral Physics Ocean lateral physics 21.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of lateral physics in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.scheme') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "None" # "Eddy active" # "Eddy admitting" # TODO - please enter value(s) Explanation: 21.2. Scheme Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of transient eddy representation in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.momentum.operator.direction') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Horizontal" # "Isopycnal" # "Isoneutral" # "Geopotential" # "Iso-level" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 22. Lateral Physics --&gt; Momentum --&gt; Operator Properties of lateral physics operator for momentum in ocean 22.1. Direction Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Direction of lateral physics momemtum scheme in the ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.momentum.operator.order') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Harmonic" # "Bi-harmonic" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 22.2. Order Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Order of lateral physics momemtum scheme in the ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.momentum.operator.discretisation') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Second order" # "Higher order" # "Flux limiter" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 22.3. Discretisation Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Discretisation of lateral physics momemtum scheme in the ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.momentum.eddy_viscosity_coeff.type') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Constant" # "Space varying" # "Time + space varying (Smagorinsky)" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 23. Lateral Physics --&gt; Momentum --&gt; Eddy Viscosity Coeff Properties of eddy viscosity coeff in lateral physics momemtum scheme in the ocean 23.1. Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Lateral physics momemtum eddy viscosity coeff type in the ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.momentum.eddy_viscosity_coeff.constant_coefficient') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 23.2. Constant Coefficient Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 If constant, value of eddy viscosity coeff in lateral physics momemtum scheme (in m2/s) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.momentum.eddy_viscosity_coeff.variable_coefficient') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 23.3. Variable Coefficient Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 If space-varying, describe variations of eddy viscosity coeff in lateral physics momemtum scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.momentum.eddy_viscosity_coeff.coeff_background') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 23.4. Coeff Background Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe background eddy viscosity coeff in lateral physics momemtum scheme (give values in m2/s) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.momentum.eddy_viscosity_coeff.coeff_backscatter') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 23.5. Coeff Backscatter Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is there backscatter in eddy viscosity coeff in lateral physics momemtum scheme ? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.tracers.mesoscale_closure') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 24. Lateral Physics --&gt; Tracers Properties of lateral physics for tracers in ocean 24.1. Mesoscale Closure Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is there a mesoscale closure in the lateral physics tracers scheme ? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.tracers.submesoscale_mixing') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 24.2. Submesoscale Mixing Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is there a submesoscale mixing parameterisation (i.e Fox-Kemper) in the lateral physics tracers scheme ? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.tracers.operator.direction') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Horizontal" # "Isopycnal" # "Isoneutral" # "Geopotential" # "Iso-level" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 25. Lateral Physics --&gt; Tracers --&gt; Operator Properties of lateral physics operator for tracers in ocean 25.1. Direction Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Direction of lateral physics tracers scheme in the ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.tracers.operator.order') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Harmonic" # "Bi-harmonic" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 25.2. Order Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Order of lateral physics tracers scheme in the ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.tracers.operator.discretisation') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Second order" # "Higher order" # "Flux limiter" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 25.3. Discretisation Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Discretisation of lateral physics tracers scheme in the ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.tracers.eddy_diffusity_coeff.type') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Constant" # "Space varying" # "Time + space varying (Smagorinsky)" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 26. Lateral Physics --&gt; Tracers --&gt; Eddy Diffusity Coeff Properties of eddy diffusity coeff in lateral physics tracers scheme in the ocean 26.1. Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Lateral physics tracers eddy diffusity coeff type in the ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.tracers.eddy_diffusity_coeff.constant_coefficient') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 26.2. Constant Coefficient Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 If constant, value of eddy diffusity coeff in lateral physics tracers scheme (in m2/s) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.tracers.eddy_diffusity_coeff.variable_coefficient') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 26.3. Variable Coefficient Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 If space-varying, describe variations of eddy diffusity coeff in lateral physics tracers scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.tracers.eddy_diffusity_coeff.coeff_background') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 26.4. Coeff Background Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe background eddy diffusity coeff in lateral physics tracers scheme (give values in m2/s) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.tracers.eddy_diffusity_coeff.coeff_backscatter') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 26.5. Coeff Backscatter Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is there backscatter in eddy diffusity coeff in lateral physics tracers scheme ? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.tracers.eddy_induced_velocity.type') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "GM" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 27. Lateral Physics --&gt; Tracers --&gt; Eddy Induced Velocity Properties of eddy induced velocity (EIV) in lateral physics tracers scheme in the ocean 27.1. Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of EIV in lateral physics tracers in the ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.tracers.eddy_induced_velocity.constant_val') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 27.2. Constant Val Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 If EIV scheme for tracers is constant, specify coefficient value (M2/s) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.tracers.eddy_induced_velocity.flux_type') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 27.3. Flux Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of EIV flux (advective or skew) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.lateral_physics.tracers.eddy_induced_velocity.added_diffusivity') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 27.4. Added Diffusivity Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of EIV added diffusivity (constant, flow dependent or none) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 28. Vertical Physics Ocean Vertical Physics 28.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of vertical physics in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.boundary_layer_mixing.details.langmuir_cells_mixing') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 29. Vertical Physics --&gt; Boundary Layer Mixing --&gt; Details Properties of vertical physics in ocean 29.1. Langmuir Cells Mixing Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is there Langmuir cells mixing in upper ocean ? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.boundary_layer_mixing.tracers.type') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Constant value" # "Turbulent closure - TKE" # "Turbulent closure - KPP" # "Turbulent closure - Mellor-Yamada" # "Turbulent closure - Bulk Mixed Layer" # "Richardson number dependent - PP" # "Richardson number dependent - KT" # "Imbeded as isopycnic vertical coordinate" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 30. Vertical Physics --&gt; Boundary Layer Mixing --&gt; Tracers *Properties of boundary layer (BL) mixing on tracers in the ocean * 30.1. Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of boundary layer mixing for tracers in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.boundary_layer_mixing.tracers.closure_order') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 30.2. Closure Order Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: FLOAT&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 If turbulent BL mixing of tracers, specific order of closure (0, 1, 2.5, 3) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.boundary_layer_mixing.tracers.constant') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 30.3. Constant Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 If constant BL mixing of tracers, specific coefficient (m2/s) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.boundary_layer_mixing.tracers.background') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 30.4. Background Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Background BL mixing of tracers coefficient, (schema and value in m2/s - may by none) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.boundary_layer_mixing.momentum.type') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Constant value" # "Turbulent closure - TKE" # "Turbulent closure - KPP" # "Turbulent closure - Mellor-Yamada" # "Turbulent closure - Bulk Mixed Layer" # "Richardson number dependent - PP" # "Richardson number dependent - KT" # "Imbeded as isopycnic vertical coordinate" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 31. Vertical Physics --&gt; Boundary Layer Mixing --&gt; Momentum *Properties of boundary layer (BL) mixing on momentum in the ocean * 31.1. Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of boundary layer mixing for momentum in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.boundary_layer_mixing.momentum.closure_order') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 31.2. Closure Order Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: FLOAT&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 If turbulent BL mixing of momentum, specific order of closure (0, 1, 2.5, 3) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.boundary_layer_mixing.momentum.constant') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 31.3. Constant Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 If constant BL mixing of momentum, specific coefficient (m2/s) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.boundary_layer_mixing.momentum.background') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 31.4. Background Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Background BL mixing of momentum coefficient, (schema and value in m2/s - may by none) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.interior_mixing.details.convection_type') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Non-penetrative convective adjustment" # "Enhanced vertical diffusion" # "Included in turbulence closure" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 32. Vertical Physics --&gt; Interior Mixing --&gt; Details *Properties of interior mixing in the ocean * 32.1. Convection Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of vertical convection in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.interior_mixing.details.tide_induced_mixing') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 32.2. Tide Induced Mixing Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe how tide induced mixing is modelled (barotropic, baroclinic, none) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.interior_mixing.details.double_diffusion') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 32.3. Double Diffusion Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is there double diffusion End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.interior_mixing.details.shear_mixing') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 32.4. Shear Mixing Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is there interior shear mixing End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.interior_mixing.tracers.type') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Constant value" # "Turbulent closure / TKE" # "Turbulent closure - Mellor-Yamada" # "Richardson number dependent - PP" # "Richardson number dependent - KT" # "Imbeded as isopycnic vertical coordinate" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 33. Vertical Physics --&gt; Interior Mixing --&gt; Tracers *Properties of interior mixing on tracers in the ocean * 33.1. Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of interior mixing for tracers in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.interior_mixing.tracers.constant') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 33.2. Constant Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 If constant interior mixing of tracers, specific coefficient (m2/s) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.interior_mixing.tracers.profile') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 33.3. Profile Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is the background interior mixing using a vertical profile for tracers (i.e is NOT constant) ? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.interior_mixing.tracers.background') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 33.4. Background Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Background interior mixing of tracers coefficient, (schema and value in m2/s - may by none) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.interior_mixing.momentum.type') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Constant value" # "Turbulent closure / TKE" # "Turbulent closure - Mellor-Yamada" # "Richardson number dependent - PP" # "Richardson number dependent - KT" # "Imbeded as isopycnic vertical coordinate" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 34. Vertical Physics --&gt; Interior Mixing --&gt; Momentum *Properties of interior mixing on momentum in the ocean * 34.1. Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of interior mixing for momentum in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.interior_mixing.momentum.constant') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 34.2. Constant Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 If constant interior mixing of momentum, specific coefficient (m2/s) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.interior_mixing.momentum.profile') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 34.3. Profile Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is the background interior mixing using a vertical profile for momentum (i.e is NOT constant) ? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.vertical_physics.interior_mixing.momentum.background') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 34.4. Background Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Background interior mixing of momentum coefficient, (schema and value in m2/s - may by none) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.uplow_boundaries.free_surface.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 35. Uplow Boundaries --&gt; Free Surface Properties of free surface in ocean 35.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of free surface in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.uplow_boundaries.free_surface.scheme') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Linear implicit" # "Linear filtered" # "Linear semi-explicit" # "Non-linear implicit" # "Non-linear filtered" # "Non-linear semi-explicit" # "Fully explicit" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 35.2. Scheme Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Free surface scheme in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.uplow_boundaries.free_surface.embeded_seaice') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 35.3. Embeded Seaice Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is the sea-ice embeded in the ocean model (instead of levitating) ? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.uplow_boundaries.bottom_boundary_layer.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 36. Uplow Boundaries --&gt; Bottom Boundary Layer Properties of bottom boundary layer in ocean 36.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of bottom boundary layer in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.uplow_boundaries.bottom_boundary_layer.type_of_bbl') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Diffusive" # "Acvective" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 36.2. Type Of Bbl Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of bottom boundary layer in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.uplow_boundaries.bottom_boundary_layer.lateral_mixing_coef') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 36.3. Lateral Mixing Coef Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 If bottom BL is diffusive, specify value of lateral mixing coefficient (in m2/s) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.uplow_boundaries.bottom_boundary_layer.sill_overflow') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 36.4. Sill Overflow Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe any specific treatment of sill overflows End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.boundary_forcing.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 37. Boundary Forcing Ocean boundary forcing 37.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of boundary forcing in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.boundary_forcing.surface_pressure') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 37.2. Surface Pressure Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe how surface pressure is transmitted to ocean (via sea-ice, nothing specific,...) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.boundary_forcing.momentum_flux_correction') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 37.3. Momentum Flux Correction Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe any type of ocean surface momentum flux correction and, if applicable, how it is applied and where. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.boundary_forcing.tracers_flux_correction') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 37.4. Tracers Flux Correction Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe any type of ocean surface tracers flux correction and, if applicable, how it is applied and where. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.boundary_forcing.wave_effects') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 37.5. Wave Effects Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe if/how wave effects are modelled at ocean surface. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.boundary_forcing.river_runoff_budget') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 37.6. River Runoff Budget Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe how river runoff from land surface is routed to ocean and any global adjustment done. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.boundary_forcing.geothermal_heating') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 37.7. Geothermal Heating Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe if/how geothermal heating is present at ocean bottom. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.boundary_forcing.momentum.bottom_friction.type') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Linear" # "Non-linear" # "Non-linear (drag function of speed of tides)" # "Constant drag coefficient" # "None" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 38. Boundary Forcing --&gt; Momentum --&gt; Bottom Friction Properties of momentum bottom friction in ocean 38.1. Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of momentum bottom friction in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.boundary_forcing.momentum.lateral_friction.type') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "None" # "Free-slip" # "No-slip" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 39. Boundary Forcing --&gt; Momentum --&gt; Lateral Friction Properties of momentum lateral friction in ocean 39.1. Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of momentum lateral friction in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.boundary_forcing.tracers.sunlight_penetration.scheme') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "1 extinction depth" # "2 extinction depth" # "3 extinction depth" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 40. Boundary Forcing --&gt; Tracers --&gt; Sunlight Penetration Properties of sunlight penetration scheme in ocean 40.1. Scheme Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of sunlight penetration scheme in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.boundary_forcing.tracers.sunlight_penetration.ocean_colour') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 40.2. Ocean Colour Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is the ocean sunlight penetration scheme ocean colour dependent ? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.boundary_forcing.tracers.sunlight_penetration.extinction_depth') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 40.3. Extinction Depth Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe and list extinctions depths for sunlight penetration scheme (if applicable). End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.boundary_forcing.tracers.fresh_water_forcing.from_atmopshere') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Freshwater flux" # "Virtual salt flux" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 41. Boundary Forcing --&gt; Tracers --&gt; Fresh Water Forcing Properties of surface fresh water forcing in ocean 41.1. From Atmopshere Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of surface fresh water forcing from atmos in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.boundary_forcing.tracers.fresh_water_forcing.from_sea_ice') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Freshwater flux" # "Virtual salt flux" # "Real salt flux" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 41.2. From Sea Ice Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of surface fresh water forcing from sea-ice in ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.ocean.boundary_forcing.tracers.fresh_water_forcing.forced_mode_restoring') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 41.3. Forced Mode Restoring Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Type of surface salinity restoring in forced mode (OMIP) End of explanation <END_TASK>
15,925
<SYSTEM_TASK:> Given the following text problem statement, write Python code to implement the functionality described below in problem statement <END_TASK> <USER_TASK:> Problem: Given a distance matrix, with similarity between various fruits :
<ASSISTANT_TASK:> Python Code: import numpy as np import pandas as pd import sklearn.cluster simM = load_data() model = sklearn.cluster.AgglomerativeClustering(affinity='precomputed', n_clusters=2, linkage='complete').fit(simM) cluster_labels = model.labels_ <END_TASK>
15,926
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Vertex SDK Step1: Restart the Kernel Once you've installed the Vertex SDK and Google cloud-storage, you need to restart the notebook kernel so it can find the packages. Step2: Before you begin GPU run-time Make sure you're running this notebook in a GPU runtime if you have that option. In Colab, select Runtime > Change Runtime Type > GPU Set up your GCP project The following steps are required, regardless of your notebook environment. Select or create a GCP project. When you first create an account, you get a $300 free credit towards your compute/storage costs. Make sure that billing is enabled for your project. Enable the AutoML APIs and Compute Engine APIs. Google Cloud SDK is already installed in AutoML Notebooks. Enter your project ID in the cell below. Then run the cell to make sure the Cloud SDK uses the right project for all the commands in this notebook. Note Step3: Region You can also change the REGION variable, which is used for operations throughout the rest of this notebook. Below are regions supported forAutoML. We recommend when possible, to choose the region closest to you. Currently project resources must be in the us-central1 region to use this API. Step4: Timestamp If you are in a live tutorial session, you might be using a shared test account or project. To avoid name collisions between users on resources created, you create a timestamp for each instance session, and append onto the name of resources which will be created in this tutorial. Step5: Authenticate your GCP account If you are using AutoML Notebooks, your environment is already authenticated. Skip this step. Note Step6: Create a Cloud Storage bucket The following steps are required, regardless of your notebook environment. This tutorial is designed to use training data that is in a public Cloud Storage bucket and a local Cloud Storage bucket for your batch predictions. You may alternatively use your own training data that you have stored in a local Cloud Storage bucket. Set the name of your Cloud Storage bucket below. It must be unique across all Cloud Storage buckets. Step7: Only if your bucket doesn't already exist Step8: Finally, validate access to your Cloud Storage bucket by examining its contents Step9: Set up variables Next, set up some variables used throughout the tutorial. Import libraries and define constants Step10: AutoML constants Setup up the following constants for AutoML Step11: Clients We use the Google APIs Client Library for Python to call the Vertex Training and Prediction API without manually constructing HTTP requests. Step12: Prepare a trainer script Package assembly Step13: Task.py contents Step14: Store training script on your Cloud Storage bucket Step15: Train a model projects.jobs.create Request Step16: Example output Step17: Response Step18: Example output Step19: projects.jobs.get Call Step20: Response Step21: Example output Step22: Serving function for trained model (image data) Step23: Example output Step24: Example output Step25: Example output Step26: Response Step27: Example output Step28: projects.jobs.get Call Step29: Response Step30: Example output Step31: Example output Step32: Example output Step33: Response Step34: Example output Step35: projects.models.versions.create Request Step36: Example output Step37: Response Step38: Example output Step39: Prepare input for online prediction Step40: projects.predict Request Step41: Example output Step42: Response Step43: Example output Step44: Response Step45: Example output
<ASSISTANT_TASK:> Python Code: ! pip3 install google-cloud-storage Explanation: Vertex SDK: Train & deploy a TensorFlow model with hosted runtimes (aka pre-built containers) Installation Install the Google cloud-storage library as well. End of explanation import os if not os.getenv("AUTORUN"): # Automatically restart kernel after installs import IPython app = IPython.Application.instance() app.kernel.do_shutdown(True) Explanation: Restart the Kernel Once you've installed the Vertex SDK and Google cloud-storage, you need to restart the notebook kernel so it can find the packages. End of explanation PROJECT_ID = "[your-project-id]" # @param {type:"string"} if PROJECT_ID == "" or PROJECT_ID is None or PROJECT_ID == "[your-project-id]": # Get your GCP project id from gcloud shell_output = !gcloud config list --format 'value(core.project)' 2>/dev/null PROJECT_ID = shell_output[0] print("Project ID:", PROJECT_ID) ! gcloud config set project $PROJECT_ID Explanation: Before you begin GPU run-time Make sure you're running this notebook in a GPU runtime if you have that option. In Colab, select Runtime > Change Runtime Type > GPU Set up your GCP project The following steps are required, regardless of your notebook environment. Select or create a GCP project. When you first create an account, you get a $300 free credit towards your compute/storage costs. Make sure that billing is enabled for your project. Enable the AutoML APIs and Compute Engine APIs. Google Cloud SDK is already installed in AutoML Notebooks. Enter your project ID in the cell below. Then run the cell to make sure the Cloud SDK uses the right project for all the commands in this notebook. Note: Jupyter runs lines prefixed with ! as shell commands, and it interpolates Python variables prefixed with $ into these commands. End of explanation REGION = "us-central1" # @param {type: "string"} Explanation: Region You can also change the REGION variable, which is used for operations throughout the rest of this notebook. Below are regions supported forAutoML. We recommend when possible, to choose the region closest to you. Currently project resources must be in the us-central1 region to use this API. End of explanation from datetime import datetime TIMESTAMP = datetime.now().strftime("%Y%m%d%H%M%S") Explanation: Timestamp If you are in a live tutorial session, you might be using a shared test account or project. To avoid name collisions between users on resources created, you create a timestamp for each instance session, and append onto the name of resources which will be created in this tutorial. End of explanation import os import sys # If you are running this notebook in Colab, run this cell and follow the # instructions to authenticate your Google Cloud account. This provides access # to your Cloud Storage bucket and lets you submit training jobs and prediction # requests. # If on AutoML, then don't execute this code if not os.path.exists("/opt/deeplearning/metadata/env_version"): if "google.colab" in sys.modules: from google.colab import auth as google_auth google_auth.authenticate_user() # If you are running this tutorial in a notebook locally, replace the string # below with the path to your service account key and run this cell to # authenticate your Google Cloud account. else: %env GOOGLE_APPLICATION_CREDENTIALS your_path_to_credentials.json # Log in to your account on Google Cloud ! gcloud auth login Explanation: Authenticate your GCP account If you are using AutoML Notebooks, your environment is already authenticated. Skip this step. Note: If you are on an AutoML notebook and run the cell, the cell knows to skip executing the authentication steps. End of explanation BUCKET_NAME = "[your-bucket-name]" # @param {type:"string"} if BUCKET_NAME == "" or BUCKET_NAME is None or BUCKET_NAME == "[your-bucket-name]": BUCKET_NAME = PROJECT_ID + "aip-" + TIMESTAMP Explanation: Create a Cloud Storage bucket The following steps are required, regardless of your notebook environment. This tutorial is designed to use training data that is in a public Cloud Storage bucket and a local Cloud Storage bucket for your batch predictions. You may alternatively use your own training data that you have stored in a local Cloud Storage bucket. Set the name of your Cloud Storage bucket below. It must be unique across all Cloud Storage buckets. End of explanation ! gsutil mb -l $REGION gs://$BUCKET_NAME Explanation: Only if your bucket doesn't already exist: Run the following cell to create your Cloud Storage bucket. End of explanation ! gsutil ls -al gs://$BUCKET_NAME Explanation: Finally, validate access to your Cloud Storage bucket by examining its contents: End of explanation import json import time from google.protobuf.json_format import MessageToJson from google.protobuf.struct_pb2 import Value from googleapiclient import discovery, errors Explanation: Set up variables Next, set up some variables used throughout the tutorial. Import libraries and define constants End of explanation # AutoM location root path for your dataset, model and endpoint resources PARENT = "projects/" + PROJECT_ID Explanation: AutoML constants Setup up the following constants for AutoML: PARENT: The AutoM location root path for dataset, model and endpoint resources. End of explanation cloudml = discovery.build("ml", "v1") Explanation: Clients We use the Google APIs Client Library for Python to call the Vertex Training and Prediction API without manually constructing HTTP requests. End of explanation ! rm -rf cifar ! mkdir cifar ! touch cifar/README.md setup_cfg = "[egg_info]\n\ tag_build =\n\ tag_date = 0" ! echo "$setup_cfg" > cifar/setup.cfg setup_py = "import setuptools\n\ # Requires TensorFlow Datasets\n\ setuptools.setup(\n\ install_requires=[\n\ 'tensorflow_datasets==1.3.0',\n\ ],\n\ packages=setuptools.find_packages())" ! echo "$setup_py" > cifar/setup.py pkg_info = "Metadata-Version: 1.0\n\ Name: Custom Training CIFAR-10\n\ Version: 0.0.0\n\ Summary: Demonstration training script\n\ Home-page: www.google.com\n\ Author: Google\n\ Author-email: aferlitsch@google.com\n\ License: Public\n\ Description: Demo\n\ Platform: Vertex AI" ! echo "$pkg_info" > cifar/PKG-INFO ! mkdir cifar/trainer ! touch cifar/trainer/__init__.py Explanation: Prepare a trainer script Package assembly End of explanation %%writefile cifar/trainer/task.py import tensorflow_datasets as tfds import tensorflow as tf from tensorflow.python.client import device_lib import argparse import os import sys tfds.disable_progress_bar() parser = argparse.ArgumentParser() parser.add_argument('--model-dir', dest='model_dir', default='/tmp/saved_model', type=str, help='Model dir.') parser.add_argument('--lr', dest='lr', default=0.01, type=float, help='Learning rate.') parser.add_argument('--epochs', dest='epochs', default=10, type=int, help='Number of epochs.') parser.add_argument('--steps', dest='steps', default=200, type=int, help='Number of steps per epoch.') parser.add_argument('--distribute', dest='distribute', type=str, default='single', help='distributed training strategy') args = parser.parse_args() print('Python Version = {}'.format(sys.version)) print('TensorFlow Version = {}'.format(tf.__version__)) print('TF_CONFIG = {}'.format(os.environ.get('TF_CONFIG', 'Not found'))) print('DEVICES', device_lib.list_local_devices()) if args.distribute == 'single': if tf.test.is_gpu_available(): strategy = tf.distribute.OneDeviceStrategy(device="/gpu:0") else: strategy = tf.distribute.OneDeviceStrategy(device="/cpu:0") elif args.distribute == 'mirror': strategy = tf.distribute.MirroredStrategy() elif args.distribute == 'multi': strategy = tf.distribute.experimental.MultiWorkerMirroredStrategy() print('num_replicas_in_sync = {}'.format(strategy.num_replicas_in_sync)) BUFFER_SIZE = 10000 BATCH_SIZE = 64 def make_datasets_unbatched(): def scale(image, label): image = tf.cast(image, tf.float32) image /= 255.0 return image, label datasets, info = tfds.load(name='cifar10', with_info=True, as_supervised=True) return datasets['train'].map(scale).cache().shuffle(BUFFER_SIZE).repeat() def build_and_compile_cnn_model(): model = tf.keras.Sequential([ tf.keras.layers.Conv2D(32, 3, activation='relu', input_shape=(32, 32, 3)), tf.keras.layers.MaxPooling2D(), tf.keras.layers.Conv2D(32, 3, activation='relu'), tf.keras.layers.MaxPooling2D(), tf.keras.layers.Flatten(), tf.keras.layers.Dense(10, activation='softmax') ]) model.compile( loss=tf.keras.losses.sparse_categorical_crossentropy, optimizer=tf.keras.optimizers.SGD(learning_rate=args.lr), metrics=['accuracy']) return model NUM_WORKERS = strategy.num_replicas_in_sync GLOBAL_BATCH_SIZE = BATCH_SIZE * NUM_WORKERS train_dataset = make_datasets_unbatched().batch(GLOBAL_BATCH_SIZE) with strategy.scope(): model = build_and_compile_cnn_model() model.fit(x=train_dataset, epochs=args.epochs, steps_per_epoch=args.steps) model.save(args.model_dir) Explanation: Task.py contents End of explanation ! rm -f cifar.tar cifar.tar.gz ! tar cvf cifar.tar cifar ! gzip cifar.tar ! gsutil cp cifar.tar.gz gs://$BUCKET_NAME/trainer_cifar.tar.gz Explanation: Store training script on your Cloud Storage bucket End of explanation JOB_NAME = "custom_job_TF_" + TIMESTAMP TRAINING_INPUTS = { "scaleTier": "CUSTOM", "masterType": "n1-standard-4", "masterConfig": {"acceleratorConfig": {"count": "1", "type": "NVIDIA_TESLA_K80"}}, "packageUris": ["gs://" + BUCKET_NAME + "/trainer_cifar.tar.gz"], "pythonModule": "trainer.task", "args": [ "--model-dir=" + "gs://{}/{}".format(BUCKET_NAME, JOB_NAME), "--epochs=" + str(20), "--steps=" + str(100), "--distribute=" + "single", ], "region": REGION, "runtimeVersion": "2.1", "pythonVersion": "3.7", } body = {"jobId": JOB_NAME, "trainingInput": TRAINING_INPUTS} request = cloudml.projects().jobs().create(parent=PARENT) request.body = json.loads(json.dumps(body, indent=2)) print(json.dumps(json.loads(request.to_json()), indent=2)) request = cloudml.projects().jobs().create(parent=PARENT, body=body) Explanation: Train a model projects.jobs.create Request End of explanation response = request.execute() Explanation: Example output: { "uri": "https://ml.googleapis.com/v1/projects/migration-ucaip-training/jobs?alt=json", "method": "POST", "body": { "jobId": "custom_job_TF_20210325211532", "trainingInput": { "scaleTier": "CUSTOM", "masterType": "n1-standard-4", "masterConfig": { "acceleratorConfig": { "count": "1", "type": "NVIDIA_TESLA_K80" } }, "packageUris": [ "gs://migration-ucaip-trainingaip-20210325211532/trainer_cifar.tar.gz" ], "pythonModule": "trainer.task", "args": [ "--model-dir=gs://migration-ucaip-trainingaip-20210325211532/custom_job_TF_20210325211532", "--epochs=20", "--steps=100", "--distribute=single" ], "region": "us-central1", "runtimeVersion": "2.1", "pythonVersion": "3.7" } }, "headers": { "accept": "application/json", "accept-encoding": "gzip, deflate", "user-agent": "(gzip)", "x-goog-api-client": "gdcl/1.12.8 gl-python/3.7.8" }, "methodId": "ml.projects.jobs.create", "resumable": null, "response_callbacks": [], "_in_error_state": false, "body_size": 0, "resumable_uri": null, "resumable_progress": 0 } Call End of explanation print(json.dumps(response, indent=2)) Explanation: Response End of explanation # The full unique ID for the custom training job custom_training_id = f'{PARENT}/jobs/{response["jobId"]}' # The short numeric ID for the custom training job custom_training_short_id = response["jobId"] print(custom_training_id) Explanation: Example output: { "jobId": "custom_job_TF_20210325211532", "trainingInput": { "scaleTier": "CUSTOM", "masterType": "n1-standard-4", "packageUris": [ "gs://migration-ucaip-trainingaip-20210325211532/trainer_cifar.tar.gz" ], "pythonModule": "trainer.task", "args": [ "--model-dir=gs://migration-ucaip-trainingaip-20210325211532/custom_job_TF_20210325211532", "--epochs=20", "--steps=100", "--distribute=single" ], "region": "us-central1", "runtimeVersion": "2.1", "pythonVersion": "3.7", "masterConfig": { "acceleratorConfig": { "count": "1", "type": "NVIDIA_TESLA_K80" } } }, "createTime": "2021-03-25T21:15:40Z", "state": "QUEUED", "trainingOutput": {}, "etag": "dH4whflp8Fg=" } End of explanation request = cloudml.projects().jobs().get(name=custom_training_id) response = request.execute() Explanation: projects.jobs.get Call End of explanation print(json.dumps(response, indent=2)) Explanation: Response End of explanation while True: response = cloudml.projects().jobs().get(name=custom_training_id).execute() if response["state"] != "SUCCEEDED": print("Training job has not completed:", response["state"]) if response["state"] == "FAILED": break else: break time.sleep(20) # model artifact output directory on Google Cloud Storage model_artifact_dir = response["trainingInput"]["args"][0].split("=")[-1] print("artifact location " + model_artifact_dir) Explanation: Example output: { "jobId": "custom_job_TF_20210325211532", "trainingInput": { "scaleTier": "CUSTOM", "masterType": "n1-standard-4", "packageUris": [ "gs://migration-ucaip-trainingaip-20210325211532/trainer_cifar.tar.gz" ], "pythonModule": "trainer.task", "args": [ "--model-dir=gs://migration-ucaip-trainingaip-20210325211532/custom_job_TF_20210325211532", "--epochs=20", "--steps=100", "--distribute=single" ], "region": "us-central1", "runtimeVersion": "2.1", "pythonVersion": "3.7", "masterConfig": { "acceleratorConfig": { "count": "1", "type": "NVIDIA_TESLA_K80" } } }, "createTime": "2021-03-25T21:15:40Z", "state": "PREPARING", "trainingOutput": {}, "etag": "eLnYfClHtKU=" } End of explanation import tensorflow as tf model = tf.keras.models.load_model(model_artifact_dir) CONCRETE_INPUT = "numpy_inputs" def _preprocess(bytes_input): decoded = tf.io.decode_jpeg(bytes_input, channels=3) decoded = tf.image.convert_image_dtype(decoded, tf.float32) resized = tf.image.resize(decoded, size=(32, 32)) rescale = tf.cast(resized / 255.0, tf.float32) return rescale @tf.function(input_signature=[tf.TensorSpec([None], tf.string)]) def preprocess_fn(bytes_inputs): decoded_images = tf.map_fn( _preprocess, bytes_inputs, dtype=tf.float32, back_prop=False ) return { CONCRETE_INPUT: decoded_images } # User needs to make sure the key matches model's input m_call = tf.function(model.call).get_concrete_function( [tf.TensorSpec(shape=[None, 32, 32, 3], dtype=tf.float32, name=CONCRETE_INPUT)] ) @tf.function( input_signature=[tf.TensorSpec([None], tf.string), tf.TensorSpec([None], tf.string)] ) def serving_fn(bytes_inputs, key): images = preprocess_fn(bytes_inputs) prob = m_call(**images) return {"prediction": prob, "key": key} tf.saved_model.save( model, model_artifact_dir, signatures={ "serving_default": serving_fn, }, ) loaded = tf.saved_model.load(model_artifact_dir) tensors_specs = list(loaded.signatures["serving_default"].structured_input_signature) print("Tensors specs:", tensors_specs) input_name = [v for k, v in tensors_specs[1].items() if k != "key"][0].name print("Bytes input tensor name:", input_name) Explanation: Serving function for trained model (image data) End of explanation import base64 import json import cv2 import numpy as np import tensorflow as tf (_, _), (x_test, y_test) = tf.keras.datasets.cifar10.load_data() test_image_1, test_label_1 = x_test[0], y_test[0] test_image_2, test_label_2 = x_test[1], y_test[1] cv2.imwrite("tmp1.jpg", (test_image_1 * 255).astype(np.uint8)) cv2.imwrite("tmp2.jpg", (test_image_2 * 255).astype(np.uint8)) gcs_input_uri = "gs://" + BUCKET_NAME + "/" + "test.json" with tf.io.gfile.GFile(gcs_input_uri, "w") as f: for img in ["tmp1.jpg", "tmp2.jpg"]: bytes = tf.io.read_file(img) b64str = base64.b64encode(bytes.numpy()).decode("utf-8") f.write(json.dumps({"key": img, input_name: {"b64": b64str}}) + "\n") ! gsutil cat $gcs_input_uri Explanation: Example output: Tensors specs: [(), {'bytes_inputs': TensorSpec(shape=(None,), dtype=tf.string, name='bytes_inputs'), 'key': TensorSpec(shape=(None,), dtype=tf.string, name='key')}] Bytes input tensor name: bytes_inputs Make batch predictions Prepare files for batch prediction End of explanation body = { "jobId": "custom_job_TF_pred_" + TIMESTAMP, "prediction_input": { "input_paths": gcs_input_uri, "output_path": "gs://" + f"{BUCKET_NAME}/batch_output/", "data_format": "JSON", "runtime_version": "2.1", "uri": model_artifact_dir, "region": "us-central1", }, } request = ( cloudml.projects() .jobs() .create( parent=PARENT, ) ) request.body = json.loads(json.dumps(body, indent=2)) print(json.dumps(json.loads(request.to_json()), indent=2)) request = cloudml.projects().jobs().create(parent=PARENT, body=body) Explanation: Example output: {"key": "tmp1.jpg", "bytes_inputs": {"b64": "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"}} {"key": "tmp2.jpg", "bytes_inputs": {"b64": "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"}} projects.jobs.create Request End of explanation response = request.execute() Explanation: Example output: { "uri": "https://ml.googleapis.com/v1/projects/migration-ucaip-training/jobs?alt=json", "method": "POST", "body": { "jobId": "custom_job_TF_pred_20210325211532", "prediction_input": { "input_paths": "gs://migration-ucaip-trainingaip-20210325211532/test.json", "output_path": "gs://migration-ucaip-trainingaip-20210325211532/batch_output/", "data_format": "JSON", "runtime_version": "2.1", "uri": "gs://migration-ucaip-trainingaip-20210325211532/custom_job_TF_20210325211532", "region": "us-central1" } }, "headers": { "accept": "application/json", "accept-encoding": "gzip, deflate", "user-agent": "(gzip)", "x-goog-api-client": "gdcl/1.12.8 gl-python/3.7.8" }, "methodId": "ml.projects.jobs.create", "resumable": null, "response_callbacks": [], "_in_error_state": false, "body_size": 0, "resumable_uri": null, "resumable_progress": 0 } Call End of explanation print(json.dumps(response, indent=2)) Explanation: Response End of explanation # The full unique ID for the batch prediction job batch_job_id = PARENT + "/jobs/" + response["jobId"] print(batch_job_id) Explanation: Example output: { "jobId": "custom_job_TF_pred_20210325211532", "predictionInput": { "dataFormat": "JSON", "inputPaths": [ "gs://migration-ucaip-trainingaip-20210325211532/test.json" ], "outputPath": "gs://migration-ucaip-trainingaip-20210325211532/batch_output/", "region": "us-central1", "runtimeVersion": "2.1", "uri": "gs://migration-ucaip-trainingaip-20210325211532/custom_job_TF_20210325211532", "framework": "TENSORFLOW" }, "createTime": "2021-03-25T21:34:56Z", "state": "QUEUED", "predictionOutput": { "outputPath": "gs://migration-ucaip-trainingaip-20210325211532/batch_output/" }, "etag": "QwNOFOfoKdY=" } End of explanation request = cloudml.projects().jobs().get(name=batch_job_id) response = request.execute() Explanation: projects.jobs.get Call End of explanation print(json.dumps(response, indent=2)) Explanation: Response End of explanation while True: response = request = cloudml.projects().jobs().get(name=batch_job_id).execute() if response["state"] != "SUCCEEDED": print("The job has not completed:", response["state"]) if response["state"] == "FAILED": break else: folder = response["predictionInput"]["outputPath"][:-1] ! gsutil ls $folder/prediction* ! gsutil cat $folder/prediction* break time.sleep(60) Explanation: Example output: { "jobId": "custom_job_TF_pred_20210325211532", "predictionInput": { "dataFormat": "JSON", "inputPaths": [ "gs://migration-ucaip-trainingaip-20210325211532/test.json" ], "outputPath": "gs://migration-ucaip-trainingaip-20210325211532/batch_output/", "region": "us-central1", "runtimeVersion": "2.1", "uri": "gs://migration-ucaip-trainingaip-20210325211532/custom_job_TF_20210325211532", "framework": "TENSORFLOW" }, "createTime": "2021-03-25T21:34:56Z", "state": "QUEUED", "predictionOutput": { "outputPath": "gs://migration-ucaip-trainingaip-20210325211532/batch_output/" }, "etag": "NSbtn4XnbbU=" } End of explanation request = cloudml.projects().models().create(parent=PARENT) request.body = json.loads(json.dumps({"name": "custom_job_TF_" + TIMESTAMP}, indent=2)) print(json.dumps(json.loads(request.to_json()), indent=2)) request = ( cloudml.projects() .models() .create(parent=PARENT, body={"name": "custom_job_TF_" + TIMESTAMP}) ) Explanation: Example output: gs://migration-ucaip-trainingaip-20210325211532/batch_output/prediction.errors_stats-00000-of-00001 gs://migration-ucaip-trainingaip-20210325211532/batch_output/prediction.results-00000-of-00001 {"prediction": [0.033321816474199295, 0.052459586411714554, 0.1548144668340683, 0.11401787400245667, 0.17382358014583588, 0.09015274047851562, 0.19865882396697998, 0.10446511209011078, 0.029874442145228386, 0.048411525785923004], "key": "tmp1.jpg"} {"prediction": [0.03346974775195122, 0.05255022272467613, 0.15449963510036469, 0.11388237029314041, 0.17408262193202972, 0.08989296853542328, 0.19814379513263702, 0.10520868003368378, 0.02989153563976288, 0.04837837815284729], "key": "tmp2.jpg"} Make online predictions Deploy the model projects.models.create Request End of explanation response = request.execute() Explanation: Example output: { "uri": "https://ml.googleapis.com/v1/projects/migration-ucaip-training/models?alt=json", "method": "POST", "body": { "name": "custom_job_TF_20210325211532" }, "headers": { "accept": "application/json", "accept-encoding": "gzip, deflate", "user-agent": "(gzip)", "x-goog-api-client": "gdcl/1.12.8 gl-python/3.7.8" }, "methodId": "ml.projects.models.create", "resumable": null, "response_callbacks": [], "_in_error_state": false, "body_size": 0, "resumable_uri": null, "resumable_progress": 0 } Call End of explanation print(json.dumps(response, indent=2)) Explanation: Response End of explanation # The full unique ID for the training pipeline model_id = response["name"] # The short numeric ID for the training pipeline model_short_name = model_id.split("/")[-1] print(model_id) Explanation: Example output: { "name": "projects/migration-ucaip-training/models/custom_job_TF_20210325211532", "regions": [ "us-central1" ], "etag": "fFH1QQbH3tA=" } End of explanation version = { "name": "custom_job_TF_" + TIMESTAMP, "deploymentUri": model_artifact_dir, "runtimeVersion": "2.1", "framework": "TENSORFLOW", "pythonVersion": "3.7", "machineType": "mls1-c1-m2", } request = ( cloudml.projects() .models() .versions() .create( parent=model_id, ) ) request.body = json.loads(json.dumps(version, indent=2)) print(json.dumps(json.loads(request.to_json()), indent=2)) request = cloudml.projects().models().versions().create(parent=model_id, body=version) Explanation: projects.models.versions.create Request End of explanation response = request.execute() Explanation: Example output: { "uri": "https://ml.googleapis.com/v1/projects/migration-ucaip-training/models/custom_job_TF_20210325211532/versions?alt=json", "method": "POST", "body": { "name": "custom_job_TF_20210325211532", "deploymentUri": "gs://migration-ucaip-trainingaip-20210325211532/custom_job_TF_20210325211532", "runtimeVersion": "2.1", "framework": "TENSORFLOW", "pythonVersion": "3.7", "machineType": "mls1-c1-m2" }, "headers": { "accept": "application/json", "accept-encoding": "gzip, deflate", "user-agent": "(gzip)", "x-goog-api-client": "gdcl/1.12.8 gl-python/3.7.8" }, "methodId": "ml.projects.models.versions.create", "resumable": null, "response_callbacks": [], "_in_error_state": false, "body_size": 0, "resumable_uri": null, "resumable_progress": 0 } Call End of explanation print(json.dumps(response, indent=2)) Explanation: Response End of explanation # The full unique ID for the model version model_version_name = response["metadata"]["version"]["name"] print(model_version_name) while True: response = ( cloudml.projects().models().versions().get(name=model_version_name).execute() ) if response["state"] == "READY": print("Model version created.") break time.sleep(60) Explanation: Example output: { "name": "projects/migration-ucaip-training/operations/create_custom_job_TF_20210325211532_custom_job_TF_20210325211532-1616708521927", "metadata": { "@type": "type.googleapis.com/google.cloud.ml.v1.OperationMetadata", "createTime": "2021-03-25T21:42:02Z", "operationType": "CREATE_VERSION", "modelName": "projects/migration-ucaip-training/models/custom_job_TF_20210325211532", "version": { "name": "projects/migration-ucaip-training/models/custom_job_TF_20210325211532/versions/custom_job_TF_20210325211532", "deploymentUri": "gs://migration-ucaip-trainingaip-20210325211532/custom_job_TF_20210325211532", "createTime": "2021-03-25T21:42:01Z", "runtimeVersion": "2.1", "etag": "3vf44xGDtdw=", "framework": "TENSORFLOW", "machineType": "mls1-c1-m2", "pythonVersion": "3.7" } } } End of explanation import base64 import json import cv2 import tensorflow as tf (_, _), (x_test, y_test) = tf.keras.datasets.cifar10.load_data() test_image_1, test_label_1 = x_test[0], y_test[0] test_image_2, test_label_2 = x_test[1], y_test[1] cv2.imwrite("tmp1.jpg", (test_image_1 * 255).astype(np.uint8)) cv2.imwrite("tmp2.jpg", (test_image_2 * 255).astype(np.uint8)) Explanation: Prepare input for online prediction End of explanation instances_list = [] for img in ["tmp1.jpg", "tmp2.jpg"]: bytes = tf.io.read_file(img) b64str = base64.b64encode(bytes.numpy()).decode("utf-8") instances_list.append({"key": img, input_name: {"b64": b64str}}) request = cloudml.projects().predict(name=model_version_name) request.body = json.loads(json.dumps({"instances": instances_list}, indent=2)) print(json.dumps(json.loads(request.to_json()), indent=2)) request = cloudml.projects().predict( name=model_version_name, body={"instances": instances_list} ) Explanation: projects.predict Request End of explanation response = request.execute() Explanation: Example output: { "uri": "https://ml.googleapis.com/v1/projects/migration-ucaip-training/models/custom_job_TF_20210325211532/versions/custom_job_TF_20210325211532:predict?alt=json", "method": "POST", "body": { "instances": [ { "key": "tmp1.jpg", "bytes_inputs": { "b64": "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" } }, { "key": "tmp2.jpg", "bytes_inputs": { "b64": "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" } } ] }, "headers": { "accept": "application/json", "accept-encoding": "gzip, deflate", "user-agent": "(gzip)", "x-goog-api-client": "gdcl/1.12.8 gl-python/3.7.8" }, "methodId": "ml.projects.predict", "resumable": null, "response_callbacks": [], "_in_error_state": false, "body_size": 0, "resumable_uri": null, "resumable_progress": 0 } Call End of explanation print(json.dumps(response, indent=2)) Explanation: Response End of explanation request = cloudml.projects().models().versions().delete(name=model_version_name) response = request.execute() Explanation: Example output: { "predictions": [ { "key": "tmp1.jpg", "prediction": [ 0.033321816474199295, 0.052459586411714554, 0.1548144668340683, 0.11401788890361786, 0.17382356524467468, 0.09015275537967682, 0.19865882396697998, 0.10446509718894958, 0.02987445704638958, 0.048411525785923004 ] }, { "key": "tmp2.jpg", "prediction": [ 0.03346974775195122, 0.052550218999385834, 0.15449965000152588, 0.11388237029314041, 0.17408263683319092, 0.08989296108484268, 0.19814379513263702, 0.10520866513252258, 0.02989153563976288, 0.04837837815284729 ] } ] } projects.models.versions.delete Call End of explanation print(json.dumps(response, indent=2)) Explanation: Response End of explanation delete_model = True delete_bucket = True # Delete the model using the Vertex AI fully qualified identifier for the model try: if delete_model: cloudml.projects().models().delete(name=model_id) except Exception as e: print(e) if delete_bucket and "BUCKET_NAME" in globals(): ! gsutil rm -r gs://$BUCKET_NAME Explanation: Example output: { "name": "projects/migration-ucaip-training/operations/delete_custom_job_TF_20210325211532_custom_job_TF_20210325211532-1616708584436", "metadata": { "@type": "type.googleapis.com/google.cloud.ml.v1.OperationMetadata", "createTime": "2021-03-25T21:43:04Z", "operationType": "DELETE_VERSION", "modelName": "projects/migration-ucaip-training/models/custom_job_TF_20210325211532", "version": { "name": "projects/migration-ucaip-training/models/custom_job_TF_20210325211532/versions/custom_job_TF_20210325211532", "deploymentUri": "gs://migration-ucaip-trainingaip-20210325211532/custom_job_TF_20210325211532", "createTime": "2021-03-25T21:42:01Z", "runtimeVersion": "2.1", "state": "READY", "etag": "Nu2QJaCl6vw=", "framework": "TENSORFLOW", "machineType": "mls1-c1-m2", "pythonVersion": "3.7" } } } Cleanup End of explanation <END_TASK>
15,927
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: <p> <img src="http Step1: Stirling array $\mathcal{S}$ Step2: power function Step3: inverse function Step4: sqrt function Step5: expt function Step6: log function Step7: sin function Step8: cos function
<ASSISTANT_TASK:> Python Code: from sympy import * from sympy.abc import n, i, N, x, lamda, phi, z, j, r, k, a, alpha from commons import * from matrix_functions import * from sequences import * import functions_catalog init_printing() Explanation: <p> <img src="http://www.cerm.unifi.it/chianti/images/logo%20unifi_positivo.jpg" alt="UniFI logo" style="float: left; width: 20%; height: 20%;"> <div align="right"> Massimo Nocentini<br> <small> <br>November 20, 2018: cleanup, `sin` and `cos` $g$ polys <br>November 16 and 18, 2016: classic $g$ polys </small> </div> </p> <br> <br> <div align="center"> <b>Abstract</b><br> Theory of matrix functions, with applications to Stirling array $\mathcal{S}$. </div> End of explanation from sympy.functions.combinatorial.numbers import stirling m=8 S2 = define(let=Symbol(r'\mathcal{{S}}_{{ {} }}'.format(m)), be=Matrix(m, m, lambda n,k: stirling(n,k, kind=2))) S2 # this version is the pure one S2 = define(let=Symbol(r'\mathcal{{S}}_{{ {} }}'.format(m)), be=Matrix(m, m, riordan_matrix_exponential( riordan_matrix_by_convolution(d=Eq(Function('d')(z), exp(z)), h=Eq(Function('h')(z), exp(z)-1), dim=m)))) S2 inspect(S2.rhs) production_matrix(S2.rhs) eigendata = spectrum(S2) eigendata data, eigenvals, multiplicities = eigendata.rhs Phi_poly = Phi_poly_ctor(deg=m-1) Phi_poly Phi_polynomials = component_polynomials(eigendata, early_eigenvals_subs=True) Phi_polynomials cmatrices = component_matrices(S2, Phi_polynomials) cmatrices Explanation: Stirling array $\mathcal{S}$ End of explanation f_power, g_power, G_power = functions_catalog.power(eigendata, Phi_polynomials) S2_power = G_power(S2) S2_power define(S2_power.lhs, S2_power.rhs.applyfunc(factor)) # factored S2_power.rhs[:,0] assert (S2.rhs**r).applyfunc(simplify) == S2_power.rhs inspect(S2_power.rhs) production_matrix(S2_power.rhs).applyfunc(factor) Explanation: power function End of explanation f_inverse, g_inverse, G_inverse = functions_catalog.inverse(eigendata, Phi_polynomials) S2_inverse = G_inverse(S2) S2_inverse, G_inverse(S2_inverse) inspect(S2_inverse.rhs) production_matrix(S2_inverse.rhs) assert S2_inverse.rhs*S2.rhs == Matrix(m, m, identity_matrix()) assert S2_inverse.rhs == S2_power.rhs.subs({r:-1}) Explanation: inverse function End of explanation f_sqrt, g_sqrt, G_sqrt = functions_catalog.square_root(eigendata, Phi_polynomials) S2_sqrt = G_sqrt(S2) S2_sqrt inspect(S2_sqrt.rhs) production_matrix(S2_sqrt.rhs) inspect(S2_sqrt.rhs) production_matrix(S2_sqrt.rhs, exp=False) assert S2_sqrt.rhs == S2.rhs**(S(1)/2) assert S2_sqrt.rhs == S2_power.rhs.subs({r:S(1)/2}) Explanation: sqrt function End of explanation f_exp, g_exp, G_exp = functions_catalog.exp(eigendata, Phi_polynomials) S2_exp = G_exp(S2) S2_exp define(S2_exp.lhs, S2_exp.rhs.applyfunc(factor)) S2_exp1 = define(let=Subs(S2_exp.lhs, alpha, 1), be=S2_exp.rhs.subs({alpha:1})) S2_exp1 inspect(S2_exp.rhs) inspect(S2_exp1.rhs) Explanation: expt function End of explanation f_log, g_log, G_log = functions_catalog.log(eigendata, Phi_polynomials) S2_log = G_log(S2) S2_log inspect(S2_log.rhs[1:,:-1]) production_matrix(S2_log.rhs[1:,:-1]) Explanation: log function End of explanation f_sin, g_sin, G_sin = functions_catalog.sin(eigendata, Phi_polynomials) S2_sin = G_sin(S2) S2_sin Explanation: sin function End of explanation f_cos, g_cos, G_cos = functions_catalog.cos(eigendata, Phi_polynomials) S2_cos = G_cos(S2) S2_cos assert (S2_sin.rhs**2 + S2_cos.rhs**2).applyfunc(trigsimp) == Matrix(m, m, identity_matrix()) Explanation: cos function End of explanation <END_TASK>
15,928
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Sets & Frozen Sets A set in python is an unordered sequence of unique and immutable objects, although the set itself is mutable. A frozen set is the same as a set except that the set itself is immutable. Since the set is a collection of unordered elements, indexing is not applicable. Constructing sets Step1: Adding, Updating, and Removing elements from a set Step2: Set Membership & Length Step3: Set Intersection, Disjoint, Union, and Difference Step4: Subsets and supersets A.issubset(B) returns True if A is a subset of B. A <= B is another way of writing it.<br> A.issuperset(B) returns True of A is a superset of B. A >= is another way of writing it. "<" is used to check if a set is a proper subset of another set. A proper subset B of the set A means that all the elements of B are in A but A has at least one more element that is not in B. Formally Step5: Symmetric difference The symmetric difference of two sets is the set of elements which are in one of either set, but not in both. Step6: Printing, Copying, & Clearing a set
<ASSISTANT_TASK:> Python Code: # from a list a = set([1,2,3,4]) a # using curly braces a = {1,2,3,4} a # using a tuple a = set((1,2,3,4)) a # start with and empty set and add elements to it a = set() a.add('hello') a # be careful in assigning a string as an element to a set. If assigned as below, it'll be broken up and only unique # characters are added to the set a = set('this is a string') a # this will do the same thing a = set(('string')) # if you want to assign a string as an element then do a = set() a.add('string') a # or a = set(('string',)) a # or a = set(['string']) a # or a = {'string'} a # this assignment will not work since we're one element of the set is a list which is a mutable object. # A set's elements should be immutable a = {"a", (1,2), [3,4]} # uniquness of the elements of a set a = {1,2,3,3,4,4,7,8,-1,3,4} # we contains only one copy of a repeated element a Explanation: Sets & Frozen Sets A set in python is an unordered sequence of unique and immutable objects, although the set itself is mutable. A frozen set is the same as a set except that the set itself is immutable. Since the set is a collection of unordered elements, indexing is not applicable. Constructing sets End of explanation A = {1,2,3,4} A.add('string') A # update a set A.update([26, 12, 9, 14]) print A # A.discard(x) will remove x from the set. If x is not in the set, nothing happens. A = {'a', 'b', 'c', 'd', 'e'} A.discard('b') print "A after discarding the element 'b'", A A.discard('f') print "A after discarding the element 'f' (which does not belong to A)", A # A.remove(x) works like A.discard(x) but if x is not in the set A, an KeyError exception is thrown A = {'a', 'b', 'c', 'd', 'e'} A.remove('b') print "A after A.remove('b')", A A.remove('f') # a KeyError is thrown print "A after A.remove('f') ('f' does not belong to A)", A Explanation: Adding, Updating, and Removing elements from a set End of explanation A = {'1', '2', '3', 1, 2, 3} print "Is '1' in A?", '1' in A print "Is '4' not in A?", '4' not in A print "Is 5 in A", 5 in A print "A is %s elements long"%len(A) Explanation: Set Membership & Length End of explanation # A.intersection(B) or A&B retrurns a set with the common elements between A and B. A = {"a","b","c","d","e"} B = {"c","d","e","f","g"} print "The intersection of A and B is ", A.intersection(B) print "The intersection of A and B using A&B notation is ", A&B # get the common elements of more than 2 sets A = set([1,2,3,4,5]) B = set([4,5,6,7,8]) C = set([4,5,1,8,9]) print "First way to get the intersection of 3 sets ", set.intersection(A, B, C) print "Second wat to get the intersection of 3 sets ", A&B&C # A.issdisjoint(B) returns True if the two sets have a null intersection, otherwise it returns False A = {'a', 'b', 'c', 'd', 'e', 'f'} B = {'c', 'd', 'e'} C = {'g', 'h', 'i', 'j'} print "A and B have no common elements? ", A.isdisjoint(B) print "A and C have no common elements? ", A.isdisjoint(C) # A.union(B) or A | B returns a set of the merged elements of A and B # the same element in both sets is counted once of course A = set([1,2,3]) B = set([2,3,4]) print "The merging of A and B is ", A.union(B) print "The merging of A and B using another notation is ", A | B # A.difference(B) or A-B returns the a set of the difference between A and B. # This can be chanied to include many sets A = {'a', 'b', 'c', 'd', 'e'} B = {'c', 'd', 'g', 'h'} C = {'d', 'b', 'f'} print "The elements that are in A but not in B", A.difference(B) print "The elemets that are in A but not in B or C", A.difference(B).difference(C) print "The elements that are in A but not in B", A-B print "The elemets that are in A but not in B or C", A-B-C # A.difference_update(B) or A = A - B find the difference between A and B and assigns it back to A. # In other words, it removes the elements from A that are in B and assigns it to A. A = {'a', 'b', 'c', 'd', 'e'} B = {'c', 'd', 'g', 'h'} A.difference_update(B) print "A after A.difference_update(B)", A A = {'a', 'b', 'c', 'd', 'e'} B = {'c', 'd', 'g', 'h'} A = A - B print "A after A=A-B", A Explanation: Set Intersection, Disjoint, Union, and Difference End of explanation A = {'a', 'b', 'c', 'd', 'e', 'f'} B = {'c', 'd', 'e'} print "Is A a subset of B? ", A.issubset(B) print "Is A a superset of B? ", A.issuperset(B) print "Is B a subset of A? ", B.issubset(A) print "Is B a superset of A? ", B.issuperset(A) # using the other notation print "Is A a subset of B? ", A <= B print "Is A a superset of B? ", A >= B print "Is B a subset of A? ", B <= A print "Is B a superset of A? ", B >= A print "Is B a superset of B? ", B >= B print "Is B a subset of B? ", B <= B print "Is A a proper subset of B? ", A < B print "Is B a proper subset of A? ", B < A print "Is A a proper superset of B? ", A > B print "Is B a proper superset of A? ", B > A print "Is B a proper subset of B? ", B < B print "Is B a proper superset of B? ", B > B # pop() removes and returns an arbitrary set element. The method raises a KeyError if the set is empty x = {"a","b","c","d","e"} x.pop() x.pop() Explanation: Subsets and supersets A.issubset(B) returns True if A is a subset of B. A <= B is another way of writing it.<br> A.issuperset(B) returns True of A is a superset of B. A >= is another way of writing it. "<" is used to check if a set is a proper subset of another set. A proper subset B of the set A means that all the elements of B are in A but A has at least one more element that is not in B. Formally: a proper subset of a set A is a subset of A that is not equal to A. ">" is used to check if a set is a proper superset of another set. A set A that is a proper superset of B means that all elements of B are in A but A contains at least one element that is not in B. Formally: A proper superset of a set A is a superset of A that is not equal to A. End of explanation A = set([1, 2, 3]) B = set([3, 4, 5]) print "Symmetric diffrence between A and B ", A.symmetric_difference(B) print "Symmetric difference between A and B using A ^ B is ", A ^ B A.symmetric_difference_update(B) # this assigns the results of symmetric_difference to A print "After symmetric_difference_update A is ", A Explanation: Symmetric difference The symmetric difference of two sets is the set of elements which are in one of either set, but not in both. End of explanation # print the elements of the set A = set("James") for n in A: print n, # print the elements of A on the same line A = {'string', 1,2,3,4,5,3,'a',(1,2,3)} for n in A: print n, # print the elements of A each on it's own line for n in A: print n # A.clear() clear the elemets of A A = {'string', 1,2,3,4,5,3,'a',(1,2,3)} print "A before A.clear()", A A.clear() print "A after A.clear()", A # A.copy() copies the contents of A into another set B so clearing A wont affect the contents of B A = {1,2,3,3,4,4,7,8,-1,3,4} # we contains only one copy of a repeated element B = A.copy() print "A is ", A print "B after B=A.copy() is", B A.clear() print "A after A.clear()", A print "B after A.clear()", B # B = A makes a "pointer" from A to B so modifying A will also modify B A = {1,2,3,3,4,4,7,8,-1,3,4} # we contains only one copy of a repeated element B = A print "A is ", A print "B after B=A.copy() is", B A.clear() print "A after A.clear()", A print "B after A.clear()", B Explanation: Printing, Copying, & Clearing a set End of explanation <END_TASK>
15,929
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Principal Component Analysis by Rene Zhang and Max Margenot Part of the Quantopian Lecture Series Step1: We will introduce PCA with an image processing example. A grayscale digital image can be represented by a matrix, whose $(i,j)^{th}$ entry corresponds to the measurement of gray scale at the $(i,j)^{th}$ pixel. The following gray-scale image has $200 \times 200$ pixels, though it can be changed on the fly. We store it in a matrix $\mathbf{X}$. The number of rows of the $\mathbf{X}$ is $200$, and the number of columns of $\mathbf{X}$ is $200$. Step2: We start with a simple checkboard pattern, add some random normal noise, and add a gradient. Step3: Set each row as a variable, with observations in the columns. Denote the covariance matrix of $\mathbf{X}$ as $\mathbf{C}$, where the size of $\mathbf{C}$ is $m \times m$. $\mathbf{C}$ is a matrix whose $(i,j)^{th}$ entry is the covariance between the $i^{th}$ row and $j^{th}$ row of the matrix $\mathbf{X}$. Step4: Performing principal component analysis decomposes the matrix $\mathbf{C}$ into Step5: The function LA.eigh lists the eigenvalues from small to large in $P$. Let us change the order first to list them from largest to smallest and make sure that $\mathbf{L}\mathbf{P}\mathbf{L}^{\top}==\mathbf{C}$. Step6: Here we plot all of the eigenvalues Step7: The $i^{th}$ principal component is given as $i^{th}$ row of $\mathbf{V}$, $$\mathbf{V} =\mathbf{L}^{\top} \mathbf{X}.$$ Step8: If we multiply both sides on the left by $\mathbf{L}$, we get the following Step9: The proportion of total variance due to the $i^{th}$ principal component is given by the ratio $\frac{\lambda_i}{\lambda_1 + \lambda_2 + \dots \lambda_m}.$ The sum of proportion of total variance should be $1$. As we defined, $\lambda_i$ is $i^{th}$ entry of $\mathbf{P}$, $$\sum_{i}\frac{P_i}{\text{trace}(P)} = 1$$ Where the trace$(P)$ is the sum of the diagonal of $P$. Step10: Recall the number of principal components is denoted as $k$. Let $k$ be $10, 20, 30, 60$ as examples and take a look at the corresponding approximated images. Step11: The number of variables in $X$ is $200$. When reducing the dimension to $k=60$, which uses half of the principal components, the approximated image is close to the original one. Moving forward, we do not have to do PCA by hand. Luckly, scikit-learn has an implementation that we can use. Next, let us show an example in quantitative finance using sklearn. PCA on a Portfolio Construct a portfolio with 10 stocks, IBM, MSFT, FB, T, INTC, ABX, NEM, AU, AEM, GFI. 5 of them are technology related and 5 of them are gold mining companies. In this case, there are 10 variables (companies), and each column is a variable. Step12: Notice that the grand bulk of the variance of the returns of these assets can be explained by the first two principal components. Now we collect the first two principal components and plot their contributions. Step13: From these principal components we can construct "statistical risk factors", similar to more conventional common risk factors. These should give us an idea of how much of the portfolio's returns comes from some unobservable statistical feature. Step14: The factor returns here are an analogue to the principal component matrix $\mathbf{V}$ in the image processing example. Step15: The factor exposures are an analogue to the eigenvector matrix $\mathbf{L}$ in the image processing example.
<ASSISTANT_TASK:> Python Code: from numpy import linalg as LA import numpy as np import pandas as pd import matplotlib.pyplot as plt Explanation: Principal Component Analysis by Rene Zhang and Max Margenot Part of the Quantopian Lecture Series: www.quantopian.com/lectures https://github.com/quantopian/research_public Applications in many fields, such as image processing, bioinformatics, and quantitative finance, involve large-scale data. Both the size and complexity of this data can make the computations required for analysis practically infeasible. Principal Component Analysis (PCA) is a classical method for dimension reduction. It uses the first several principal components, statistical features that explain most of the variation of a $m \times n$ data matrix $\mathbf{X}$, to describe the large-scale data matrix $\mathbf{X}$ economically. End of explanation def generate_test_image(m,n): X = np.zeros((m,n)) # generate a rectangle X[25:80,25:80] = 1 # generate a triangle for i in range(25, 80, 1): X[i+80:160, 100+i-1] = 2 # generate a circle for i in range(0,200,1): for j in range(0,200,1): if ((i - 135)*(i - 135) +(j - 53)*(j - 53) <= 900): X[i, j] = 3 return X X = generate_test_image(200,200) Explanation: We will introduce PCA with an image processing example. A grayscale digital image can be represented by a matrix, whose $(i,j)^{th}$ entry corresponds to the measurement of gray scale at the $(i,j)^{th}$ pixel. The following gray-scale image has $200 \times 200$ pixels, though it can be changed on the fly. We store it in a matrix $\mathbf{X}$. The number of rows of the $\mathbf{X}$ is $200$, and the number of columns of $\mathbf{X}$ is $200$. End of explanation imgplot = plt.imshow(X, cmap='gray') plt.title('Original Test Image'); m = X.shape[0] # num of rows n = X.shape[1] # num of columns Explanation: We start with a simple checkboard pattern, add some random normal noise, and add a gradient. End of explanation X = np.asarray(X, dtype=np.float64) C = np.cov(X) np.linalg.matrix_rank(C) Explanation: Set each row as a variable, with observations in the columns. Denote the covariance matrix of $\mathbf{X}$ as $\mathbf{C}$, where the size of $\mathbf{C}$ is $m \times m$. $\mathbf{C}$ is a matrix whose $(i,j)^{th}$ entry is the covariance between the $i^{th}$ row and $j^{th}$ row of the matrix $\mathbf{X}$. End of explanation P, L = LA.eigh(C) Explanation: Performing principal component analysis decomposes the matrix $\mathbf{C}$ into: $$\mathbf{C} = \mathbf{L}\mathbf{P}\mathbf{L}^{\top},$$ where $\mathbf{P}$ is a diagonal matrix $\mathbf{P}=\text{diag}(\lambda_1,\lambda_2,\dots,\lambda_m)$, with $\lambda_1 \geq \lambda_1 \geq \dots \lambda_m \geq 0$ being the eigenvalues of matrix $\mathbf{C}$. The matrix $\mathbf{L}$ is an orthogonal matrix, consisting the eigenvectors of matrix $\mathbf{C}$. End of explanation P = P[::-1] L = L[:,::-1] np.allclose(L.dot(np.diag(P)).dot(L.T), C) Explanation: The function LA.eigh lists the eigenvalues from small to large in $P$. Let us change the order first to list them from largest to smallest and make sure that $\mathbf{L}\mathbf{P}\mathbf{L}^{\top}==\mathbf{C}$. End of explanation plt.semilogy(P, '-o') plt.xlim([1, P.shape[0]]) plt.xlabel('eigenvalue index') plt.ylabel('eigenvalue in a log scale') plt.title('Eigenvalues of Covariance Matrix'); Explanation: Here we plot all of the eigenvalues: End of explanation V = L.T.dot(X) V.shape Explanation: The $i^{th}$ principal component is given as $i^{th}$ row of $\mathbf{V}$, $$\mathbf{V} =\mathbf{L}^{\top} \mathbf{X}.$$ End of explanation k = 200 X_tilde = L[:,0:k-1].dot(L[:,0:k-1].T).dot(X) np.allclose(X_tilde, X) plt.imshow(X_tilde, cmap='gray') plt.title('Approximated Image with full rank'); Explanation: If we multiply both sides on the left by $\mathbf{L}$, we get the following: $$\mathbf{L}\mathbf{L}^{\top} \mathbf{X}= \mathbf{L}\mathbf{V}.$$ The matrix $\mathbf{L}$ is the set of eigenvectors from a covariance matrix , so $\mathbf{L}\mathbf{L}^{\top} = \mathbf{I}$ and $\mathbf{L}\mathbf{L}^{\top}\mathbf{X} = \mathbf{X}$. The relationship among matrices of $\mathbf{X}$, $\mathbf{L}$, and $\mathbf{V}$ can be expressed as $$\mathbf{X} = \mathbf{L}\mathbf{V}.$$ To approximate $\mathbf{X}$, we use $k$ eigenvectors that have largest eigenvalues: $$\mathbf{X} \approx \mathbf{L[:, 1:k]}\mathbf{L[:, 1:k]}^{\top} \mathbf{X}.$$ Denote the approximated $\mathbf{X}$ as $\tilde{\mathbf{X}} = \mathbf{L[:, 1:k]}\mathbf{L[:, 1:k]}^{\top} \mathbf{X}$. When $k = m $, the $\tilde{\mathbf{X}}$ should be same as $\mathbf{X}$. End of explanation (P/P.sum()).sum() plt.plot((P/P.sum()).cumsum(), '-o') plt.title('Cumulative Sum of the Proportion of Total Variance') plt.xlabel('index') plt.ylabel('Proportion'); Explanation: The proportion of total variance due to the $i^{th}$ principal component is given by the ratio $\frac{\lambda_i}{\lambda_1 + \lambda_2 + \dots \lambda_m}.$ The sum of proportion of total variance should be $1$. As we defined, $\lambda_i$ is $i^{th}$ entry of $\mathbf{P}$, $$\sum_{i}\frac{P_i}{\text{trace}(P)} = 1$$ Where the trace$(P)$ is the sum of the diagonal of $P$. End of explanation X_tilde_10 = L[:,0:10-1].dot(V[0:10-1,:]) X_tilde_20 = L[:,0:20-1].dot(V[0:20-1,:]) X_tilde_30 = L[:,0:30-1].dot(V[0:30-1,:]) X_tilde_60 = L[:,0:60-1].dot(V[0:60-1,:]) fig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(2, 2, figsize=(12, 12)) ax1.imshow(X_tilde_10, cmap='gray') ax1.set(title='Approximated Image with k = 10') ax2.imshow(X_tilde_20, cmap='gray') ax2.set(title='Approximated Image with k = 20') ax3.imshow(X_tilde_30, cmap='gray') ax3.set(title='Approximated Image with k = 30') ax4.imshow(X_tilde_60, cmap='gray') ax4.set(title='Approximated Image with k = 60'); Explanation: Recall the number of principal components is denoted as $k$. Let $k$ be $10, 20, 30, 60$ as examples and take a look at the corresponding approximated images. End of explanation symbol = ['IBM','MSFT', 'FB', 'T', 'INTC', 'ABX','NEM', 'AU', 'AEM', 'GFI'] start = "2015-09-01" end = "2016-11-01" portfolio_returns = get_pricing(symbol, start_date=start, end_date=end, fields="price").pct_change()[1:] from sklearn.decomposition import PCA num_pc = 2 X = np.asarray(portfolio_returns) [n,m] = X.shape print 'The number of timestamps is {}.'.format(n) print 'The number of stocks is {}.'.format(m) pca = PCA(n_components=num_pc) # number of principal components pca.fit(X) percentage = pca.explained_variance_ratio_ percentage_cum = np.cumsum(percentage) print '{0:.2f}% of the variance is explained by the first 2 PCs'.format(percentage_cum[-1]*100) pca_components = pca.components_ Explanation: The number of variables in $X$ is $200$. When reducing the dimension to $k=60$, which uses half of the principal components, the approximated image is close to the original one. Moving forward, we do not have to do PCA by hand. Luckly, scikit-learn has an implementation that we can use. Next, let us show an example in quantitative finance using sklearn. PCA on a Portfolio Construct a portfolio with 10 stocks, IBM, MSFT, FB, T, INTC, ABX, NEM, AU, AEM, GFI. 5 of them are technology related and 5 of them are gold mining companies. In this case, there are 10 variables (companies), and each column is a variable. End of explanation x = np.arange(1,len(percentage)+1,1) plt.subplot(1, 2, 1) plt.bar(x, percentage*100, align = "center") plt.title('Contribution of principal components',fontsize = 16) plt.xlabel('principal components',fontsize = 16) plt.ylabel('percentage',fontsize = 16) plt.xticks(x,fontsize = 16) plt.yticks(fontsize = 16) plt.xlim([0, num_pc+1]) plt.subplot(1, 2, 2) plt.plot(x, percentage_cum*100,'ro-') plt.xlabel('principal components',fontsize = 16) plt.ylabel('percentage',fontsize = 16) plt.title('Cumulative contribution of principal components',fontsize = 16) plt.xticks(x,fontsize = 16) plt.yticks(fontsize = 16) plt.xlim([1, num_pc]) plt.ylim([50,100]); Explanation: Notice that the grand bulk of the variance of the returns of these assets can be explained by the first two principal components. Now we collect the first two principal components and plot their contributions. End of explanation factor_returns = X.dot(pca_components.T) factor_returns = pd.DataFrame(columns=["factor 1", "factor 2"], index=portfolio_returns.index, data=factor_returns) factor_returns.head() Explanation: From these principal components we can construct "statistical risk factors", similar to more conventional common risk factors. These should give us an idea of how much of the portfolio's returns comes from some unobservable statistical feature. End of explanation factor_exposures = pd.DataFrame(index=["factor 1", "factor 2"], columns=portfolio_returns.columns, data = pca.components_).T factor_exposures Explanation: The factor returns here are an analogue to the principal component matrix $\mathbf{V}$ in the image processing example. End of explanation labels = factor_exposures.index data = factor_exposures.values plt.subplots_adjust(bottom = 0.1) plt.scatter( data[:, 0], data[:, 1], marker='o', s=300, c='m', cmap=plt.get_cmap('Spectral')) plt.title('Scatter Plot of Coefficients of PC1 and PC2') plt.xlabel('factor exposure of PC1') plt.ylabel('factor exposure of PC2') for label, x, y in zip(labels, data[:, 0], data[:, 1]): plt.annotate( label, xy=(x, y), xytext=(-20, 20), textcoords='offset points', ha='right', va='bottom', bbox=dict(boxstyle='round,pad=0.5', fc='yellow', alpha=0.5), arrowprops=dict(arrowstyle = '->', connectionstyle='arc3,rad=0') ); Explanation: The factor exposures are an analogue to the eigenvector matrix $\mathbf{L}$ in the image processing example. End of explanation <END_TASK>
15,930
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Step1: Face Generation In this project, you'll use generative adversarial networks to generate new images of faces. Get the Data You'll be using two datasets in this project Step3: Explore the Data MNIST As you're aware, the MNIST dataset contains images of handwritten digits. You can view the first number of examples by changing show_n_images. Step5: CelebA The CelebFaces Attributes Dataset (CelebA) dataset contains over 200,000 celebrity images with annotations. Since you're going to be generating faces, you won't need the annotations. You can view the first number of examples by changing show_n_images. Step7: Preprocess the Data Since the project's main focus is on building the GANs, we'll preprocess the data for you. The values of the MNIST and CelebA dataset will be in the range of -0.5 to 0.5 of 28x28 dimensional images. The CelebA images will be cropped to remove parts of the image that don't include a face, then resized down to 28x28. The MNIST images are black and white images with a single [color channel](https Step10: Input Implement the model_inputs function to create TF Placeholders for the Neural Network. It should create the following placeholders Step13: Discriminator Implement discriminator to create a discriminator neural network that discriminates on images. This function should be able to reuse the variabes in the neural network. Use tf.variable_scope with a scope name of "discriminator" to allow the variables to be reused. The function should return a tuple of (tensor output of the generator, tensor logits of the generator). Step16: Generator Implement generator to generate an image using z. This function should be able to reuse the variabes in the neural network. Use tf.variable_scope with a scope name of "generator" to allow the variables to be reused. The function should return the generated 28 x 28 x out_channel_dim images. Step19: Loss Implement model_loss to build the GANs for training and calculate the loss. The function should return a tuple of (discriminator loss, generator loss). Use the following functions you implemented Step22: Optimization Implement model_opt to create the optimization operations for the GANs. Use tf.trainable_variables to get all the trainable variables. Filter the variables with names that are in the discriminator and generator scope names. The function should return a tuple of (discriminator training operation, generator training operation). Step25: Neural Network Training Show Output Use this function to show the current output of the generator during training. It will help you determine how well the GANs is training. Step27: Train Implement train to build and train the GANs. Use the following functions you implemented Step29: MNIST Test your GANs architecture on MNIST. After 2 epochs, the GANs should be able to generate images that look like handwritten digits. Make sure the loss of the generator is lower than the loss of the discriminator or close to 0. Step31: CelebA Run your GANs on CelebA. It will take around 20 minutes on the average GPU to run one epoch. You can run the whole epoch or stop when it starts to generate realistic faces.
<ASSISTANT_TASK:> Python Code: data_dir = './data' # FloydHub - Use with data ID "R5KrjnANiKVhLWAkpXhNBe" #data_dir = '/input' DON'T MODIFY ANYTHING IN THIS CELL import helper helper.download_extract('mnist', data_dir) helper.download_extract('celeba', data_dir) Explanation: Face Generation In this project, you'll use generative adversarial networks to generate new images of faces. Get the Data You'll be using two datasets in this project: - MNIST - CelebA Since the celebA dataset is complex and you're doing GANs in a project for the first time, we want you to test your neural network on MNIST before CelebA. Running the GANs on MNIST will allow you to see how well your model trains sooner. If you're using FloydHub, set data_dir to "/input" and use the FloydHub data ID "R5KrjnANiKVhLWAkpXhNBe". End of explanation show_n_images = 25 DON'T MODIFY ANYTHING IN THIS CELL %matplotlib inline import os from glob import glob from matplotlib import pyplot mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'mnist/*.jpg'))[:show_n_images], 28, 28, 'L') pyplot.imshow(helper.images_square_grid(mnist_images, 'L'), cmap='gray') Explanation: Explore the Data MNIST As you're aware, the MNIST dataset contains images of handwritten digits. You can view the first number of examples by changing show_n_images. End of explanation show_n_images = 25 DON'T MODIFY ANYTHING IN THIS CELL mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'img_align_celeba/*.jpg'))[:show_n_images], 28, 28, 'RGB') pyplot.imshow(helper.images_square_grid(mnist_images, 'RGB')) Explanation: CelebA The CelebFaces Attributes Dataset (CelebA) dataset contains over 200,000 celebrity images with annotations. Since you're going to be generating faces, you won't need the annotations. You can view the first number of examples by changing show_n_images. End of explanation DON'T MODIFY ANYTHING IN THIS CELL from distutils.version import LooseVersion import warnings import tensorflow as tf # Check TensorFlow Version assert LooseVersion(tf.__version__) >= LooseVersion('1.0'), 'Please use TensorFlow version 1.0 or newer. You are using {}'.format(tf.__version__) print('TensorFlow Version: {}'.format(tf.__version__)) # Check for a GPU if not tf.test.gpu_device_name(): warnings.warn('No GPU found. Please use a GPU to train your neural network.') else: print('Default GPU Device: {}'.format(tf.test.gpu_device_name())) Explanation: Preprocess the Data Since the project's main focus is on building the GANs, we'll preprocess the data for you. The values of the MNIST and CelebA dataset will be in the range of -0.5 to 0.5 of 28x28 dimensional images. The CelebA images will be cropped to remove parts of the image that don't include a face, then resized down to 28x28. The MNIST images are black and white images with a single [color channel](https://en.wikipedia.org/wiki/Channel_(digital_image%29) while the CelebA images have [3 color channels (RGB color channel)](https://en.wikipedia.org/wiki/Channel_(digital_image%29#RGB_Images). Build the Neural Network You'll build the components necessary to build a GANs by implementing the following functions below: - model_inputs - discriminator - generator - model_loss - model_opt - train Check the Version of TensorFlow and Access to GPU This will check to make sure you have the correct version of TensorFlow and access to a GPU End of explanation import problem_unittests as tests def model_inputs(image_width, image_height, image_channels, z_dim): Create the model inputs :param image_width: The input image width :param image_height: The input image height :param image_channels: The number of image channels :param z_dim: The dimension of Z :return: Tuple of (tensor of real input images, tensor of z data, learning rate) # TODO: Implement Function inputs_real = tf.placeholder(tf.float32, (None, image_width, image_height, image_channels), name='input_real') inputs_z = tf.placeholder(tf.float32, (None, z_dim), name='input_z') learning_rate = tf.placeholder(tf.float32, name='learning_rate') return inputs_real, inputs_z, learning_rate DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE tests.test_model_inputs(model_inputs) Explanation: Input Implement the model_inputs function to create TF Placeholders for the Neural Network. It should create the following placeholders: - Real input images placeholder with rank 4 using image_width, image_height, and image_channels. - Z input placeholder with rank 2 using z_dim. - Learning rate placeholder with rank 0. Return the placeholders in the following the tuple (tensor of real input images, tensor of z data) End of explanation def discriminator(images, reuse=False): Create the discriminator network :param image: Tensor of input image(s) :param reuse: Boolean if the weights should be reused :return: Tuple of (tensor output of the discriminator, tensor logits of the discriminator) # TODO: Implement Function alpha=0.2 # TODO: Implement Function with tf.variable_scope('discriminator', reuse=reuse): # input 28x28x3 x1 = tf.layers.conv2d(images, 64, 5, strides=1,padding='same') relu1 = tf.maximum(alpha*x1, x1) x2 = tf.layers.conv2d(relu1, 128, 5, strides=2, padding='same') bn2 = tf.layers.batch_normalization(x2, training=True) relu2 = tf.maximum(alpha * bn2, bn2) x3 = tf.layers.conv2d(relu2, 256, 5, strides=1, padding='same') bn3 = tf.layers.batch_normalization(x3, training=True) relu3 = tf.maximum(alpha*bn3, bn3) flat = tf.reshape(relu3, (-1, 7*7*512)) logits = tf.layers.dense(flat, 1) out = tf.sigmoid(logits) return out, logits DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE tests.test_discriminator(discriminator, tf) Explanation: Discriminator Implement discriminator to create a discriminator neural network that discriminates on images. This function should be able to reuse the variabes in the neural network. Use tf.variable_scope with a scope name of "discriminator" to allow the variables to be reused. The function should return a tuple of (tensor output of the generator, tensor logits of the generator). End of explanation def generator(z, out_channel_dim, is_train=True): Create the generator network :param z: Input z :param out_channel_dim: The number of channels in the output image :param is_train: Boolean if generator is being used for training :return: The tensor output of the generator # TODO: Implement Function alpha = 0.2 # TODO: Implement Function with tf.variable_scope('generator', reuse=False if is_train==True else True): x1 = tf.layers.dense(z, 7*7*512) x1 = tf.reshape(x1, (-1, 7, 7, 512)) x1 = tf.layers.batch_normalization(x1, training=is_train) x1 = tf.maximum(alpha*x1, x1) x2 = tf.layers.conv2d_transpose(x1, 256, 5, strides = 1, padding='same') x2 = tf.layers.batch_normalization(x2, training=is_train) x2 = tf.maximum(alpha*x2, x2) x3 = tf.layers.conv2d_transpose(x2, 128, 5, strides = 2, padding='same') x3 = tf.layers.batch_normalization(x3, training=is_train) x3 = tf.maximum(alpha*x3, x3) logits = tf.layers.conv2d_transpose(x3, out_channel_dim, 5, strides = 2, padding='same') out = tf.tanh(logits) return out DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE tests.test_generator(generator, tf) Explanation: Generator Implement generator to generate an image using z. This function should be able to reuse the variabes in the neural network. Use tf.variable_scope with a scope name of "generator" to allow the variables to be reused. The function should return the generated 28 x 28 x out_channel_dim images. End of explanation def model_loss(input_real, input_z, out_channel_dim): Get the loss for the discriminator and generator :param input_real: Images from the real dataset :param input_z: Z input :param out_channel_dim: The number of channels in the output image :return: A tuple of (discriminator loss, generator loss) # TODO: Implement Function g_model = generator(input_z, out_channel_dim, is_train=True) d_model_real, d_logits_real = discriminator(input_real, reuse=False) d_model_fake, d_logits_fake = discriminator(g_model, reuse=True) d_loss_real = tf.reduce_mean( tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_real, labels=tf.ones_like(d_model_real))) d_loss_fake = tf.reduce_mean( tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.zeros_like(d_model_fake))) g_loss = tf.reduce_mean( tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.ones_like(d_model_fake))) d_loss = d_loss_real + d_loss_fake return d_loss, g_loss DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE tests.test_model_loss(model_loss) Explanation: Loss Implement model_loss to build the GANs for training and calculate the loss. The function should return a tuple of (discriminator loss, generator loss). Use the following functions you implemented: - discriminator(images, reuse=False) - generator(z, out_channel_dim, is_train=True) End of explanation def model_opt(d_loss, g_loss, learning_rate, beta1): Get optimization operations :param d_loss: Discriminator loss Tensor :param g_loss: Generator loss Tensor :param learning_rate: Learning Rate Placeholder :param beta1: The exponential decay rate for the 1st moment in the optimizer :return: A tuple of (discriminator training operation, generator training operation) # TODO: Implement Function t_vars = tf.trainable_variables() d_vars = [var for var in t_vars if var.name.startswith('discriminator')] g_vars = [var for var in t_vars if var.name.startswith('generator')] # Optimize all_update_ops = tf.get_collection(tf.GraphKeys.UPDATE_OPS) g_update_ops = [var for var in all_update_ops if var.name.startswith('generator')] d_update_ops = [var for var in all_update_ops if var.name.startswith('discriminator')] with tf.control_dependencies(d_update_ops): d_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(d_loss, var_list=d_vars) with tf.control_dependencies(g_update_ops): g_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(g_loss, var_list=g_vars) return d_train_opt, g_train_opt DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE tests.test_model_opt(model_opt, tf) Explanation: Optimization Implement model_opt to create the optimization operations for the GANs. Use tf.trainable_variables to get all the trainable variables. Filter the variables with names that are in the discriminator and generator scope names. The function should return a tuple of (discriminator training operation, generator training operation). End of explanation DON'T MODIFY ANYTHING IN THIS CELL import numpy as np def show_generator_output(sess, n_images, input_z, out_channel_dim, image_mode): Show example output for the generator :param sess: TensorFlow session :param n_images: Number of Images to display :param input_z: Input Z Tensor :param out_channel_dim: The number of channels in the output image :param image_mode: The mode to use for images ("RGB" or "L") cmap = None if image_mode == 'RGB' else 'gray' z_dim = input_z.get_shape().as_list()[-1] example_z = np.random.uniform(-1, 1, size=[n_images, z_dim]) samples = sess.run( generator(input_z, out_channel_dim, False), feed_dict={input_z: example_z}) images_grid = helper.images_square_grid(samples, image_mode) pyplot.imshow(images_grid, cmap=cmap) pyplot.show() Explanation: Neural Network Training Show Output Use this function to show the current output of the generator during training. It will help you determine how well the GANs is training. End of explanation def train(epoch_count, batch_size, z_dim, learning_rate, beta1, get_batches, data_shape, data_image_mode): Train the GAN :param epoch_count: Number of epochs :param batch_size: Batch Size :param z_dim: Z dimension :param learning_rate: Learning Rate :param beta1: The exponential decay rate for the 1st moment in the optimizer :param get_batches: Function to get batches :param data_shape: Shape of the data :param data_image_mode: The image mode to use for images ("RGB" or "L") # TODO: Build Model z_size = z_dim steps = 0 input_real, input_z, testt_ = model_inputs(*data_shape[1:4], z_dim) d_loss, g_loss = model_loss(input_real, input_z, data_shape[3]) d_opt, g_opt = model_opt(d_loss, g_loss, learning_rate, beta1) with tf.Session() as sess: sess.run(tf.global_variables_initializer()) for epoch_i in range(epoch_count): for batch_images in get_batches(batch_size): # TODO: Train Model steps += 1 batch_images = 2*batch_images # Sample random noise for G batch_z = np.random.uniform(-1, 1, size=(batch_size, z_size)) # Run optimizers _ = sess.run(d_opt, feed_dict={input_real: batch_images, input_z: batch_z}) _ = sess.run(g_opt, feed_dict={input_z: batch_z}) if steps % 10 == 0: # At the end of each epoch, get the losses and print them out train_loss_d = d_loss.eval({input_z: batch_z, input_real: batch_images}) train_loss_g = g_loss.eval({input_z: batch_z}) print("Epoch {}/{}...".format(epoch_i+1, epochs), "Discriminator Loss: {:.4f}...".format(train_loss_d), "Generator Loss: {:.4f}".format(train_loss_g)) if steps % 100 == 0: show_generator_output(sess, 6, input_z, data_shape[3], data_image_mode) Explanation: Train Implement train to build and train the GANs. Use the following functions you implemented: - model_inputs(image_width, image_height, image_channels, z_dim) - model_loss(input_real, input_z, out_channel_dim) - model_opt(d_loss, g_loss, learning_rate, beta1) Use the show_generator_output to show generator output while you train. Running show_generator_output for every batch will drastically increase training time and increase the size of the notebook. It's recommended to print the generator output every 100 batches. End of explanation batch_size = 32 z_dim = 100 learning_rate = 0.002 beta1 = 0.5 DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE epochs = 2 mnist_dataset = helper.Dataset('mnist', glob(os.path.join(data_dir, 'mnist/*.jpg'))) with tf.Graph().as_default(): train(epochs, batch_size, z_dim, learning_rate, beta1, mnist_dataset.get_batches, mnist_dataset.shape, mnist_dataset.image_mode) Explanation: MNIST Test your GANs architecture on MNIST. After 2 epochs, the GANs should be able to generate images that look like handwritten digits. Make sure the loss of the generator is lower than the loss of the discriminator or close to 0. End of explanation batch_size = 128 z_dim = 100 learning_rate = 0.002 beta1 = 0.5 DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE epochs = 1 celeba_dataset = helper.Dataset('celeba', glob(os.path.join(data_dir, 'img_align_celeba/*.jpg'))) with tf.Graph().as_default(): train(epochs, batch_size, z_dim, learning_rate, beta1, celeba_dataset.get_batches, celeba_dataset.shape, celeba_dataset.image_mode) Explanation: CelebA Run your GANs on CelebA. It will take around 20 minutes on the average GPU to run one epoch. You can run the whole epoch or stop when it starts to generate realistic faces. End of explanation <END_TASK>
15,931
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Sentiment Analysis with an RNN In this notebook, you'll implement a recurrent neural network that performs sentiment analysis. Using an RNN rather than a feedfoward network is more accurate since we can include information about the sequence of words. Here we'll use a dataset of movie reviews, accompanied by labels. The architecture for this network is shown below. <img src="assets/network_diagram.png" width=400px> Here, we'll pass in words to an embedding layer. We need an embedding layer because we have tens of thousands of words, so we'll need a more efficient representation for our input data than one-hot encoded vectors. You should have seen this before from the word2vec lesson. You can actually train up an embedding with word2vec and use it here. But it's good enough to just have an embedding layer and let the network learn the embedding table on it's own. From the embedding layer, the new representations will be passed to LSTM cells. These will add recurrent connections to the network so we can include information about the sequence of words in the data. Finally, the LSTM cells will go to a sigmoid output layer here. We're using the sigmoid because we're trying to predict if this text has positive or negative sentiment. The output layer will just be a single unit then, with a sigmoid activation function. We don't care about the sigmoid outputs except for the very last one, we can ignore the rest. We'll calculate the cost from the output of the last step and the training label. Step1: Data preprocessing The first step when building a neural network model is getting your data into the proper form to feed into the network. Since we're using embedding layers, we'll need to encode each word with an integer. We'll also want to clean it up a bit. You can see an example of the reviews data above. We'll want to get rid of those periods. Also, you might notice that the reviews are delimited with newlines \n. To deal with those, I'm going to split the text into each review using \n as the delimiter. Then I can combined all the reviews back together into one big string. First, let's remove all punctuation. Then get all the text without the newlines and split it into individual words. Step2: Encoding the words The embedding lookup requires that we pass in integers to our network. The easiest way to do this is to create dictionaries that map the words in the vocabulary to integers. Then we can convert each of our reviews into integers so they can be passed into the network. Exercise Step3: Encoding the labels Our labels are "positive" or "negative". To use these labels in our network, we need to convert them to 0 and 1. Exercise Step4: If you built labels correctly, you should see the next output. Step5: Okay, a couple issues here. We seem to have one review with zero length. And, the maximum review length is way too many steps for our RNN. Let's truncate to 200 steps. For reviews shorter than 200, we'll pad with 0s. For reviews longer than 200, we can truncate them to the first 200 characters. Exercise Step6: Exercise Step7: If you build features correctly, it should look like that cell output below. Step8: Training, Validation, Test With our data in nice shape, we'll split it into training, validation, and test sets. Exercise Step9: With train, validation, and text fractions of 0.8, 0.1, 0.1, the final shapes should look like Step10: For the network itself, we'll be passing in our 200 element long review vectors. Each batch will be batch_size vectors. We'll also be using dropout on the LSTM layer, so we'll make a placeholder for the keep probability. Exercise Step11: Embedding Now we'll add an embedding layer. We need to do this because there are 74000 words in our vocabulary. It is massively inefficient to one-hot encode our classes here. You should remember dealing with this problem from the word2vec lesson. Instead of one-hot encoding, we can have an embedding layer and use that layer as a lookup table. You could train an embedding layer using word2vec, then load it here. But, it's fine to just make a new layer and let the network learn the weights. Exercise Step12: LSTM cell <img src="assets/network_diagram.png" width=400px> Next, we'll create our LSTM cells to use in the recurrent network (TensorFlow documentation). Here we are just defining what the cells look like. This isn't actually building the graph, just defining the type of cells we want in our graph. To create a basic LSTM cell for the graph, you'll want to use tf.contrib.rnn.BasicLSTMCell. Looking at the function documentation Step13: RNN forward pass <img src="assets/network_diagram.png" width=400px> Now we need to actually run the data through the RNN nodes. You can use tf.nn.dynamic_rnn to do this. You'd pass in the RNN cell you created (our multiple layered LSTM cell for instance), and the inputs to the network. outputs, final_state = tf.nn.dynamic_rnn(cell, inputs, initial_state=initial_state) Above I created an initial state, initial_state, to pass to the RNN. This is the cell state that is passed between the hidden layers in successive time steps. tf.nn.dynamic_rnn takes care of most of the work for us. We pass in our cell and the input to the cell, then it does the unrolling and everything else for us. It returns outputs for each time step and the final_state of the hidden layer. Exercise Step14: Output We only care about the final output, we'll be using that as our sentiment prediction. So we need to grab the last output with outputs[ Step15: Validation accuracy Here we can add a few nodes to calculate the accuracy which we'll use in the validation pass. Step16: Batching This is a simple function for returning batches from our data. First it removes data such that we only have full batches. Then it iterates through the x and y arrays and returns slices out of those arrays with size [batch_size]. Step17: Training Below is the typical training code. If you want to do this yourself, feel free to delete all this code and implement it yourself. Before you run this, make sure the checkpoints directory exists. Step18: Testing Step26: Test accuracy
<ASSISTANT_TASK:> Python Code: import numpy as np import tensorflow as tf with open('../sentiment-network/reviews.txt', 'r') as f: reviews = f.read() with open('../sentiment-network/labels.txt', 'r') as f: labels_orig = f.read() reviews[:2000] Explanation: Sentiment Analysis with an RNN In this notebook, you'll implement a recurrent neural network that performs sentiment analysis. Using an RNN rather than a feedfoward network is more accurate since we can include information about the sequence of words. Here we'll use a dataset of movie reviews, accompanied by labels. The architecture for this network is shown below. <img src="assets/network_diagram.png" width=400px> Here, we'll pass in words to an embedding layer. We need an embedding layer because we have tens of thousands of words, so we'll need a more efficient representation for our input data than one-hot encoded vectors. You should have seen this before from the word2vec lesson. You can actually train up an embedding with word2vec and use it here. But it's good enough to just have an embedding layer and let the network learn the embedding table on it's own. From the embedding layer, the new representations will be passed to LSTM cells. These will add recurrent connections to the network so we can include information about the sequence of words in the data. Finally, the LSTM cells will go to a sigmoid output layer here. We're using the sigmoid because we're trying to predict if this text has positive or negative sentiment. The output layer will just be a single unit then, with a sigmoid activation function. We don't care about the sigmoid outputs except for the very last one, we can ignore the rest. We'll calculate the cost from the output of the last step and the training label. End of explanation from string import punctuation all_text = ''.join([c for c in reviews if c not in punctuation]) reviews = all_text.split('\n') all_text = ' '.join(reviews) words = all_text.split() all_text[:2000] words[:100] Explanation: Data preprocessing The first step when building a neural network model is getting your data into the proper form to feed into the network. Since we're using embedding layers, we'll need to encode each word with an integer. We'll also want to clean it up a bit. You can see an example of the reviews data above. We'll want to get rid of those periods. Also, you might notice that the reviews are delimited with newlines \n. To deal with those, I'm going to split the text into each review using \n as the delimiter. Then I can combined all the reviews back together into one big string. First, let's remove all punctuation. Then get all the text without the newlines and split it into individual words. End of explanation import pickle with open('../embeddings/word_embeddings.pkl', 'rb') as f: [vocab_to_int, embed_mat] = pickle.load(f) embed_mat.shape bak = embed_mat[vocab_to_int['the'],:] # Reorganize so that 0 can be the empty string. embed_mat = np.concatenate((np.random.uniform(-1,1, (1,embed_mat.shape[1])), embed_mat), axis=0) vocab_to_int = {k:v+1 for k,v in vocab_to_int.items()} # embed_mat = embed_mat.copy() # embed_mat.resize((embed_mat.shape[0]+1, embed_mat.shape[1])) # embed_mat[-1,:] = embed_mat[0] # embed_mat[0,:] = np.random.uniform(-1,1, (1,embed_mat.shape[1])) # embed_mat.shape vocab_to_int[''] = 0 assert(all(bak == embed_mat[vocab_to_int['the'],:])) [k for k,v in vocab_to_int.items() if v == 0] embed_mat[vocab_to_int['stupid'],:] non_words = set(['','.','\n']) extra_words = set([w for w in set(words) if w not in vocab_to_int and w not in non_words]) new_vocab = [(word, index) for index,word in enumerate(extra_words, len(vocab_to_int))] embed_mat = np.concatenate( (embed_mat, np.random.uniform(-1,1, (len(extra_words), embed_mat.shape[1]))), axis=0) print("added {} extra words".format(len(extra_words))) vocab_to_int.update(new_vocab) del extra_words del new_vocab 37807/63641 reviews_ints = [[vocab_to_int[word] for word in review.split(' ') if word not in non_words] for review in reviews] set([word for word in set(words) if word not in vocab_to_int]) len(vocab_to_int) Explanation: Encoding the words The embedding lookup requires that we pass in integers to our network. The easiest way to do this is to create dictionaries that map the words in the vocabulary to integers. Then we can convert each of our reviews into integers so they can be passed into the network. Exercise: Now you're going to encode the words with integers. Build a dictionary that maps words to integers. Later we're going to pad our input vectors with zeros, so make sure the integers start at 1, not 0. Also, convert the reviews to integers and store the reviews in a new list called reviews_ints. End of explanation # Convert labels to 1s and 0s for 'positive' and 'negative' labels = np.array([(0 if l == 'negative' else 1) for l in labels_orig.split('\n')]) Explanation: Encoding the labels Our labels are "positive" or "negative". To use these labels in our network, we need to convert them to 0 and 1. Exercise: Convert labels from positive and negative to 1 and 0, respectively. End of explanation from collections import Counter review_lens = Counter([len(x) for x in reviews_ints]) print("Zero-length reviews: {}".format(review_lens[0])) print("Maximum review length: {}".format(max(review_lens))) Explanation: If you built labels correctly, you should see the next output. End of explanation x = [1,2,3] x[:10] # Filter out that review with 0 length new_values = [(review_ints[:200], label) for review_ints,label in zip(reviews_ints, labels) if len(review_ints) > 0] reviews_ints, labels = zip(*new_values) Explanation: Okay, a couple issues here. We seem to have one review with zero length. And, the maximum review length is way too many steps for our RNN. Let's truncate to 200 steps. For reviews shorter than 200, we'll pad with 0s. For reviews longer than 200, we can truncate them to the first 200 characters. Exercise: First, remove the review with zero length from the reviews_ints list. End of explanation seq_len = 200 features = np.array([([0] * (seq_len-len(review))) + review for review in reviews_ints]) labels = np.array(labels) Explanation: Exercise: Now, create an array features that contains the data we'll pass to the network. The data should come from review_ints, since we want to feed integers to the network. Each row should be 200 elements long. For reviews shorter than 200 words, left pad with 0s. That is, if the review is ['best', 'movie', 'ever'], [117, 18, 128] as integers, the row will look like [0, 0, 0, ..., 0, 117, 18, 128]. For reviews longer than 200, use on the first 200 words as the feature vector. This isn't trivial and there are a bunch of ways to do this. But, if you're going to be building your own deep learning networks, you're going to have to get used to preparing your data. End of explanation review = reviews_ints[0] len(review) features[:10,:] Explanation: If you build features correctly, it should look like that cell output below. End of explanation split_frac = 0.8 split_tv = int(features.shape[0] * split_frac) split_vt = int(round(features.shape[0] * (1-split_frac) / 2)) + split_tv train_x = features[:split_tv,:] train_y = labels[:split_vt ] val_x = features[split_tv:split_vt,:] val_y = labels[split_tv:split_vt] test_x = features[split_vt:,:] test_y = labels[split_vt: ] print("\t\t\tFeature Shapes:") print("Train set: \t\t{}".format(train_x.shape), "\nValidation set: \t{}".format(val_x.shape), "\nTest set: \t\t{}".format(test_x.shape)) Explanation: Training, Validation, Test With our data in nice shape, we'll split it into training, validation, and test sets. Exercise: Create the training, validation, and test sets here. You'll need to create sets for the features and the labels, train_x and train_y for example. Define a split fraction, split_frac as the fraction of data to keep in the training set. Usually this is set to 0.8 or 0.9. The rest of the data will be split in half to create the validation and testing data. End of explanation #run_number = 7 if 'run_number' in locals(): run_number += 1 else: run_number = 1 run_number lstm_size = 512 lstm_layers = 1 batch_size = 500 learning_rate = 0.001 Explanation: With train, validation, and text fractions of 0.8, 0.1, 0.1, the final shapes should look like: Feature Shapes: Train set: (20000, 200) Validation set: (2500, 200) Test set: (2500, 200) Build the graph Here, we'll build the graph. First up, defining the hyperparameters. lstm_size: Number of units in the hidden layers in the LSTM cells. Usually larger is better performance wise. Common values are 128, 256, 512, etc. lstm_layers: Number of LSTM layers in the network. I'd start with 1, then add more if I'm underfitting. batch_size: The number of reviews to feed the network in one training pass. Typically this should be set as high as you can go without running out of memory. learning_rate: Learning rate End of explanation n_words = len(vocab_to_int) # Create the graph object graph = tf.Graph() # Add nodes to the graph with graph.as_default(): inputs_ = tf.placeholder(tf.int32, shape=(None,seq_len), name='inputs') labels_ = tf.placeholder(tf.int32, shape=(None,1), name='labels') keep_prob = tf.placeholder(tf.float32, name='keep_prob') n_words Explanation: For the network itself, we'll be passing in our 200 element long review vectors. Each batch will be batch_size vectors. We'll also be using dropout on the LSTM layer, so we'll make a placeholder for the keep probability. Exercise: Create the inputs_, labels_, and drop out keep_prob placeholders using tf.placeholder. labels_ needs to be two-dimensional to work with some functions later. Since keep_prob is a scalar (a 0-dimensional tensor), you shouldn't provide a size to tf.placeholder. End of explanation import os os.makedirs('./logs/{}/val'.format(run_number), exist_ok=True) word_order = {v:k for k,v in vocab_to_int.items()} embedding_metadata_file = './logs/{}/val/metadata.tsv'.format(run_number) with open(embedding_metadata_file, 'w') as f: for i in range(len(word_order)): f.write(word_order[i]+'\n') projector_config = tf.contrib.tensorboard.plugins.projector.ProjectorConfig() embedding_config = projector_config.embeddings.add() embed_mat.dtype # Size of the embedding vectors (number of units in the embedding layer) if False: embed_size = 300 with graph.as_default(): with tf.name_scope('embedding'): embedding = tf.Variable( tf.random_uniform((n_words,embed_size), -1,1), name="word_embedding") embedding_config.tensor_name = embedding.name embedding_config.metadata_path = embedding_metadata_file embed = tf.nn.embedding_lookup(embedding, inputs_) tf.summary.histogram('embedding', embedding) else: embed_size = embed_mat.shape[1] with graph.as_default(): with tf.name_scope('embedding'): embedding = tf.Variable(embed_mat, name="word_embedding", dtype=tf.float32) embedding_config.tensor_name = embedding.name embedding_config.metadata_path = embedding_metadata_file embed = tf.nn.embedding_lookup(embedding, inputs_) tf.summary.histogram('embedding', embedding) Explanation: Embedding Now we'll add an embedding layer. We need to do this because there are 74000 words in our vocabulary. It is massively inefficient to one-hot encode our classes here. You should remember dealing with this problem from the word2vec lesson. Instead of one-hot encoding, we can have an embedding layer and use that layer as a lookup table. You could train an embedding layer using word2vec, then load it here. But, it's fine to just make a new layer and let the network learn the weights. Exercise: Create the embedding lookup matrix as a tf.Variable. Use that embedding matrix to get the embedded vectors to pass to the LSTM cell with tf.nn.embedding_lookup. This function takes the embedding matrix and an input tensor, such as the review vectors. Then, it'll return another tensor with the embedded vectors. So, if the embedding layer has 200 units, the function will return a tensor with size [batch_size, 200]. End of explanation with graph.as_default(): with tf.name_scope('LSTM'): # Your basic LSTM cell lstm = tf.contrib.rnn.BasicLSTMCell(lstm_size) # Add dropout to the cell drop = tf.contrib.rnn.DropoutWrapper(lstm, output_keep_prob=keep_prob) # Stack up multiple LSTM layers, for deep learning cell = tf.contrib.rnn.MultiRNNCell([drop] * lstm_layers) # Getting an initial state of all zeros initial_state = cell.zero_state(batch_size, tf.float32) Explanation: LSTM cell <img src="assets/network_diagram.png" width=400px> Next, we'll create our LSTM cells to use in the recurrent network (TensorFlow documentation). Here we are just defining what the cells look like. This isn't actually building the graph, just defining the type of cells we want in our graph. To create a basic LSTM cell for the graph, you'll want to use tf.contrib.rnn.BasicLSTMCell. Looking at the function documentation: tf.contrib.rnn.BasicLSTMCell(num_units, forget_bias=1.0, input_size=None, state_is_tuple=True, activation=&lt;function tanh at 0x109f1ef28&gt;) you can see it takes a parameter called num_units, the number of units in the cell, called lstm_size in this code. So then, you can write something like lstm = tf.contrib.rnn.BasicLSTMCell(num_units) to create an LSTM cell with num_units. Next, you can add dropout to the cell with tf.contrib.rnn.DropoutWrapper. This just wraps the cell in another cell, but with dropout added to the inputs and/or outputs. It's a really convenient way to make your network better with almost no effort! So you'd do something like drop = tf.contrib.rnn.DropoutWrapper(cell, output_keep_prob=keep_prob) Most of the time, your network will have better performance with more layers. That's sort of the magic of deep learning, adding more layers allows the network to learn really complex relationships. Again, there is a simple way to create multiple layers of LSTM cells with tf.contrib.rnn.MultiRNNCell: cell = tf.contrib.rnn.MultiRNNCell([drop] * lstm_layers) Here, [drop] * lstm_layers creates a list of cells (drop) that is lstm_layers long. The MultiRNNCell wrapper builds this into multiple layers of RNN cells, one for each cell in the list. So the final cell you're using in the network is actually multiple (or just one) LSTM cells with dropout. But it all works the same from an achitectural viewpoint, just a more complicated graph in the cell. Exercise: Below, use tf.contrib.rnn.BasicLSTMCell to create an LSTM cell. Then, add drop out to it with tf.contrib.rnn.DropoutWrapper. Finally, create multiple LSTM layers with tf.contrib.rnn.MultiRNNCell. Here is a tutorial on building RNNs that will help you out. End of explanation with graph.as_default(): with tf.name_scope('LSTM'): outputs, final_state = tf.nn.dynamic_rnn(cell, embed, initial_state=initial_state) Explanation: RNN forward pass <img src="assets/network_diagram.png" width=400px> Now we need to actually run the data through the RNN nodes. You can use tf.nn.dynamic_rnn to do this. You'd pass in the RNN cell you created (our multiple layered LSTM cell for instance), and the inputs to the network. outputs, final_state = tf.nn.dynamic_rnn(cell, inputs, initial_state=initial_state) Above I created an initial state, initial_state, to pass to the RNN. This is the cell state that is passed between the hidden layers in successive time steps. tf.nn.dynamic_rnn takes care of most of the work for us. We pass in our cell and the input to the cell, then it does the unrolling and everything else for us. It returns outputs for each time step and the final_state of the hidden layer. Exercise: Use tf.nn.dynamic_rnn to add the forward pass through the RNN. Remember that we're actually passing in vectors from the embedding layer, embed. End of explanation with graph.as_default(): with tf.name_scope('Prediction'): predictions = tf.contrib.layers.fully_connected(outputs[:, -1], 1, activation_fn=tf.sigmoid) with tf.name_scope('Loss'): cost = tf.losses.mean_squared_error(labels_, predictions) tf.summary.scalar('cost', cost) optimizer = tf.train.AdamOptimizer(learning_rate).minimize(cost) Explanation: Output We only care about the final output, we'll be using that as our sentiment prediction. So we need to grab the last output with outputs[:, -1], the calculate the cost from that and labels_. End of explanation with graph.as_default(): with tf.name_scope('Accuracy'): correct_pred = tf.equal(tf.cast(tf.round(predictions), tf.int32), labels_) accuracy = tf.reduce_mean(tf.cast(correct_pred, tf.float32)) tf.summary.scalar('accuracy',accuracy) Explanation: Validation accuracy Here we can add a few nodes to calculate the accuracy which we'll use in the validation pass. End of explanation def get_batches(x, y, batch_size=100): n_batches = len(x)//batch_size x, y = x[:n_batches*batch_size], y[:n_batches*batch_size] for ii in range(0, len(x), batch_size): yield x[ii:ii+batch_size], y[ii:ii+batch_size] Explanation: Batching This is a simple function for returning batches from our data. First it removes data such that we only have full batches. Then it iterates through the x and y arrays and returns slices out of those arrays with size [batch_size]. End of explanation train_y.mean() epochs = 20 with graph.as_default(): merged = tf.summary.merge_all() saver = tf.train.Saver() with tf.Session(graph=graph) as sess: sess.run(tf.global_variables_initializer()) train_writer = tf.summary.FileWriter('./logs/{}/train'.format(run_number), sess.graph) val_writer = tf.summary.FileWriter('./logs/{}/val'.format(run_number)) tf.contrib.tensorboard.plugins.projector.visualize_embeddings(val_writer, embedding_config) iteration = 1 for e in range(epochs): state = sess.run(initial_state) for ii, (x, y) in enumerate(get_batches(train_x, train_y, batch_size), 1): feed = {inputs_: x, labels_: y[:, None], keep_prob: 0.5, initial_state: state} summary, loss, state, _ = sess.run([merged, cost, final_state, optimizer], feed_dict=feed) train_writer.add_summary(summary, iteration) if iteration%5==0: print("Epoch: {}/{}".format(e, epochs), "Iteration: {}".format(iteration), "Train loss: {:.3f}".format(loss)) if iteration%25==0: val_acc = [] val_state = sess.run(cell.zero_state(batch_size, tf.float32)) for x, y in get_batches(val_x, val_y, batch_size): feed = {inputs_: x, labels_: y[:, None], keep_prob: 1, initial_state: val_state} summary, batch_acc, val_state = sess.run([merged, accuracy, final_state], feed_dict=feed) val_acc.append(batch_acc) val_writer.add_summary(summary, iteration) saver.save(sess, './logs/{}/model.ckpt'.format(run_number), iteration) print("Val acc: {:.3f}".format(np.mean(val_acc))) iteration +=1 saver.save(sess, "checkpoints/sentiment.ckpt") train_writer.flush() val_writer.flush() Explanation: Training Below is the typical training code. If you want to do this yourself, feel free to delete all this code and implement it yourself. Before you run this, make sure the checkpoints directory exists. End of explanation test_acc = [] with tf.Session(graph=graph) as sess: saver.restore(sess, tf.train.latest_checkpoint('checkpoints')) test_state = sess.run(cell.zero_state(batch_size, tf.float32)) for ii, (x, y) in enumerate(get_batches(test_x, test_y, batch_size), 1): feed = {inputs_: x, labels_: y[:, None], keep_prob: 1, initial_state: test_state} batch_acc, test_state = sess.run([accuracy, final_state], feed_dict=feed) test_acc.append(batch_acc) print("Test accuracy: {:.3f}".format(np.mean(test_acc))) Explanation: Testing End of explanation print(features.shape) def TestSomeText(text): text = text.lower() delete = ['.','!',',','"',"'",'\n'] for d in delete: text = text.replace(d," ") text_ints = [vocab_to_int[word] for word in text.split(' ') if word in vocab_to_int] print(len(text_ints)) text_ints = text_ints[:seq_len] #print(text_ints) #text_features = np.zeros((batch_size,seq_len)) text_features = np.array([([0] * (seq_len-len(text_ints))) + text_ints] * batch_size) #print(text_features) #print(text_features.shape) with tf.Session(graph=graph) as sess: saver.restore(sess, tf.train.latest_checkpoint('checkpoints')) test_state = sess.run(cell.zero_state(batch_size, tf.float32)) for i in range(1): feed = {inputs_: text_features, labels_: [[0]]*batch_size, keep_prob: 1, initial_state: test_state} pred, mycost, test_state = sess.run([predictions, accuracy, final_state], feed_dict=feed) return pred[0,0] TestSomeText("highly underrated movie") #pred[0,0] TestSomeText('overrated movie') TestSomeText(I ve been looking forward to a viking film or TV series for many years and when my wishes were finally granted, I was very worried that this production was going to be total crap. After viewing the first two episodes I do not worry about that anymore. Thank you, Odin As a person of some historical knowledge of the viking era, I can point out numerous flaws - but they don't ruin the story for me, so I will let them slip. Historical accounts about those days are, after all, not entirely reliable. Happy to see Travis Fimmel in a role that totally suits him. A physical and intense character, with that spice of humor that is the viking trademark from the sagas. Gabriel Byrne plays a stern leader, that made me think of him in "Prince of Jutland", and Clive Standen seems like he's going to surprise us. Been pondering the Game of Thrones comparison, since I love that show too, but in my opinion Vikings has its own thing going on. Way fewer lead characters to begin with, and also a more straight forward approach. Plenty of room for more series with this high class! Can I wish for more than the planned nine episodes, PLEASE!!!) TestSomeText(vikings) delete = ['.','!',',','"',"'",'\n'] for d in delete: text = text.replace(d," ") TestSomeText(Pirates of the Caribbean has always been a franchise that makes no attempt for Oscar worthy material but in its own way is massively enjoyable. Pirates of the Caribbean: Dead Men Tell No Tales certainly embraces the aspects of the original movie while also incorporating new plot lines that fit in well with plots from the original story. With the introduction of Henry and Karina there is a new love interest that is provided to the audience that rivals that of Will and Elizabeth Turner's. Henry Turner is portrayed as an almost exact copy of his father except just a teensy bit worse at sword fighting while Karina differs from the usual women as she remains just as important, if not more, as Henry as she guides the course towards Posiedon's trident. Jack Sparrow is entertaining as always with his usual drunk characteristics. For those of you who are tired of Sparrow acting this way Don't SEE THE MOVIE Jack sparrow isn't going to change because it doesn't make sense for his character to suddenly change. All together the movie was expertly written and expertly performed by the entire cast even Kiera Knightely who didn't manage to get one word throughout the whole movie. I know as a major fan of the Pirates of the Caribbean I can't wait to see what happens for the future of the franchise. ) pred,mycost = TestSomeText(text) pred[0,0] TestSomeText(If your child is a fan of the Wimpy Kid series, they'll no doubt enjoy this one, it's entertaining and lowbrow enough to also appease the moodiest of teens and grumpiest adults.) TestSomeText(At first I thought the film was going to be just a normal thriller but it turned out to be a thousand times better than I expected. The film is truly original and was so dark & sinister that gives the tensive mood also it is emotionally & psychologically thrilling, the whole movie is charged with pulse pounding suspense and seems like it's really happening. It's amazing that how they managed to make an 80 minute movie with just a guy in a phone booth but the full credit goes to Colin Farrell and Larry Cohen the writer not Joel Schumacher because he is a crappy director. Joel Schumacher's films are rubbish especially The Number 23, Phone Booth was shot in 10 days with a budget of $10 million so it wasn't a hard job to make it, that's why Joel doesn't get any credit but the cast & crew did a fantastic job. I also really liked the raspberry coloured shirt Colin was wearing and it was an excellent choice of clothing because the viewers are going to watch him throughout the whole film. When I first saw the movie I fell in love with it and I bought it on DVD the next day and I've seen it about 20 times and I'm still not fed up with it. Phone Booth is and always will be Colin Farrell's best film! Overall it is simply one of my favourite films and I even argued over my friend because he didn't like it. ) delete = ['.','!',',','"',"'",'\n'] for d in delete: text = text.replace(d," ") text pred,mycost = TestSomeText(text) pred[0,0] TestSomeText(There are few quality movies or series about the Vikings, and this series is outstanding and well worth waiting for. Not only is Vikings a series that is a joy to watch, it is also a series that is easy to recommend. I personally feel that the creator and producers did a fine job of giving the viewer quality material. Now, there are a few inconsistencies with the series, most notably would be the idea that Vikings had very little knowledge of other European countries and were amazed by these people across the big waters. In reality Vikings engaged in somewhat normal commercial activities with other Anglo-Saxons, so the idea that Vikings were as amazed as they seemed when they realize that other people were out there is not that realistic. However, it is this small inconsistency that goes a long way in holding the premise together. I simply love the series and would recommend it to anyone wanting to watch a quality show.) delete = ['.','!',',','"',"'",'\n'] for d in delete: text = text.replace(d," ") pred,mycost = TestSomeText(text) pred[0,0] TestSomeText(This movie didn't feel any different from the other anime movies out there. Sure, the sibling dynamics were good, as well as the family values, the childhood memories and older brother anxiety. The main idea was interesting, with the new baby seeming rather like a boss sent into the family to spy on the parents and solve a big problem for his company. You can't help but identify with the older kid, especially if you have younger siblings. But eventually, the action was a bit main stream. The action scenes were not original and kind of boring. Other than that, the story became a little complicated when you start to think about what's real and what's not. The narration was good and the animation was nice, with the cute babies and puppies. So, 4 out of 10. ) delete = ['.','!',',','"',"'",'\n'] for d in delete: text = text.replace(d," ") text pred,mycost = TestSomeText(text) pred[0,0] TestSomeText('seriously awesome movie') Explanation: Test accuracy: 0.748 Test accuracy: 0.784 End of explanation <END_TASK>
15,932
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Workshop – Introduction to Neural Networks using TensorFlow and Keras http Step1: How does Tensorflow Low Level API look like? Step2: Interactive usage of Low Level API Step3: Calling a TensorFlow Model deployed on Google Cloud ML Service Step4: What can Keras do for us? Abstracting from defining operations on tensors Just define the network architecture and you are done Keras compile the description of the model down to TensorFlow operations Many other backends are supported https
<ASSISTANT_TASK:> Python Code: import warnings warnings.filterwarnings('ignore') %matplotlib inline %pylab inline from distutils.version import StrictVersion import sklearn print(sklearn.__version__) assert StrictVersion(sklearn.__version__ ) >= StrictVersion('0.18.1') import tensorflow as tf tf.logging.set_verbosity(tf.logging.ERROR) print(tf.__version__) assert StrictVersion(tf.__version__) >= StrictVersion('1.1.0') import keras print(keras.__version__) assert StrictVersion(keras.__version__) >= StrictVersion('2.0.0') import pandas as pd print(pd.__version__) assert StrictVersion(pd.__version__) >= StrictVersion('0.20.0') Explanation: Workshop – Introduction to Neural Networks using TensorFlow and Keras http://www.mcubed.london/sessions/workshop-introduction-to-neural-networks-using-tensorflow-and-keras/ <h1 style="color:red">Executing these notebooks on Azure is fine, however, for improved performance and reliability it is best to run it on a local machine or, ideally, on more powerful GPU hardware</h1> <h2 style="color:red">You can download a complete library from Azure Notebooks as a zip</h2> (local) installation goes like this: Install Anaconda 3 for Python 3.6: https://www.anaconda.com/download/ Install TensorFlow for Anaconda: https://www.tensorflow.org/install/install_linux#InstallingAnaconda e.g. like pip install --ignore-installed --upgrade https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-1.3.0-cp36-cp36m-linux_x86_64.whl Install Keras using pip: pip install keras --upgrade <h2 style="color:red">If all checks below pass and you can download the training material you are good to go, no matter what platform you are on</h2> Notebook Basics This is a Jupyter Notebook You type commands which will be executed on a server (possibly remote) A notebook saves all output of the commands executed, thus you can view it including results without executing it Before using a notebook to type or edit commands make sure you execute all code Do this by selecting Kernel, Restart & Run All in the menu Tips This is your personal notebook, add Python comments or markdown to take notes Executing Cells Everthing can be down using the Menu and Toolbar Using the keyboard you can be faster, click on the Keyboard Symbol in the toolbar to learn about more short cuts The most important shortcut is Shit+Enter which executes a cell and navigates to the next one Preparations Import and check required dependencies End of explanation # graph definition matrix1 = tf.constant([[3., 3.]]) matrix2 = tf.constant([[2.],[2.]]) product = tf.matmul(matrix1, matrix2) # launching the graph in a session with tf.Session() as sess: result = sess.run([product]) print(result) Explanation: How does Tensorflow Low Level API look like? End of explanation sess = tf.InteractiveSession() x = tf.Variable([1.0, 2.0]) a = tf.constant([3.0, 3.0]) # Initialize 'x' using the run() method of its initializer op. x.initializer.run() # Add an op to subtract 'a' from 'x'. Run it and print the result sub = tf.subtract(x, a) print(sub.eval()) # ==> [-2. -1.] # Close the Session when we're done. sess.close() Explanation: Interactive usage of Low Level API End of explanation !cat sample_iris.json # Example for iris, model exported as Tensorflow # gsutil cp -R 1 gs://irisnn # create model and version at https://console.cloud.google.com/mlengine # in a DOS shell on local machine in this folder # gcloud ml-engine predict --model=irisnn --json-instances=./sample_iris.json # SCORES # [0.9954029321670532, 0.004596732556819916, 3.3544753819114703e-07] Explanation: Calling a TensorFlow Model deployed on Google Cloud ML Service End of explanation !curl -O https://raw.githubusercontent.com/DJCordhose/speed-limit-signs/master/data/speed-limit-signs.zip from zipfile import ZipFile zip = ZipFile(r'speed-limit-signs.zip') zip.extractall('.') # Calling a Keras based model to classify speed limit signs !curl -H "Content-Type: application/json" -X GET -d '{"url": "https://github.com/DJCordhose/speed-limit-signs/raw/master/data/real-world/4/100-sky-cutoff-detail.jpg", "model": "default" }' http://ec2-52-43-39-37.us-west-2.compute.amazonaws.com:8888 Explanation: What can Keras do for us? Abstracting from defining operations on tensors Just define the network architecture and you are done Keras compile the description of the model down to TensorFlow operations Many other backends are supported https://github.com/fchollet/keras#getting-started-30-seconds-to-keras Our task for today Train a system to classify speed limit signs: * We have a pretty realistic example * https://twitter.com/art_sobolev/status/907857395757481985?s=03: I don't think it's fine to only list experiments on toy datasets, and hide failures on more complicated cases. * we need to find ways to keep our models simple to let them train in a few minutes Getting the data End of explanation <END_TASK>
15,933
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Qualité de l'air mesurée dans la station Châtelet Source de données utilisée (fichiers CSV en open data) Step1: Infos diverses sur le DataFrame Step2: Analyse de la concentration en particules fines (PM10)
<ASSISTANT_TASK:> Python Code: %matplotlib inline #%matplotlib notebook import matplotlib matplotlib.rcParams['figure.figsize'] = (9, 9) import pandas as pd def conv_func(s): s = s.replace('<', '') if s == 'ND': return np.nan elif s.strip() == '': return np.nan else: return float(s) url = "https://data.iledefrance.fr/explore/dataset/qualite-de-lair-mesuree-dans-la-station-chatelet/download/?format=csv&timezone=Europe/Berlin&use_labels_for_header=true" #dtype_dict = {'NO': np.float64, # 'NO2': np.float64, # 'PM10': np.float64, # 'CO2': np.float64, # 'TEMP': np.float64, # 'HUMI': np.float64} converter_dict = {'NO': conv_func, 'NO2': conv_func, 'PM10': conv_func, 'CO2': conv_func, 'TEMP': conv_func, 'HUMI': conv_func} df = pd.read_csv(url, #encoding='iso-8859-1', index_col=0, sep=';', decimal=',', parse_dates=["DATE/HEURE"], #dtype=dtype_dict, #na_values='ND', converters=converter_dict) df = df.sort_index() Explanation: Qualité de l'air mesurée dans la station Châtelet Source de données utilisée (fichiers CSV en open data): https://data.iledefrance.fr/explore/dataset/qualite-de-lair-mesuree-dans-la-station-chatelet/ Récupération des données End of explanation df.head() df.columns df.dtypes df.index Explanation: Infos diverses sur le DataFrame End of explanation df.PM10.plot(figsize=(18,6)); df.PM10.resample('7D').mean().plot(figsize=(18,6)); df.PM10.rolling('7D').mean().plot(figsize=(18,6)); df.PM10.resample('1M').mean().plot(figsize=(18,6)); ts = df.PM10 # https://jakevdp.github.io/PythonDataScienceHandbook/03.11-working-with-time-series.html#Digging-into-the-data ts_mean = ts.groupby(ts.index.time).mean() ts_median = ts.groupby(ts.index.time).median() ts_quartile_1 = ts.groupby(ts.index.time).quantile(0.25) ts_quartile_3 = ts.groupby(ts.index.time).quantile(0.75) ts_percentile_5 = ts.groupby(ts.index.time).quantile(0.05) ts_percentile_95 = ts.groupby(ts.index.time).quantile(0.95) ts_min = ts.groupby(ts.index.time).min() ts_max = ts.groupby(ts.index.time).max() color = "blue" ax = ts_mean.plot(y='duration', figsize=(18, 12), color=color, label="mean", alpha=0.75) ts_median.plot(ax=ax, color=color, label="median", style="--", alpha=0.75) ts_quartile_1.plot(ax=ax, color=color, alpha=0.5, style="-.", label="1st quartile") ts_quartile_3.plot(ax=ax, color=color, alpha=0.5, style="-.", label="3rd quartile") ts_percentile_5.plot(ax=ax, color=color, alpha=0.25, style=":", label="5th percentile") ts_percentile_95.plot(ax=ax, color=color, alpha=0.25, style=":", label="95th percentile") ts_min.plot(ax=ax, color=color, alpha=0.2, style=":", label="min") ts_max.plot(ax=ax, color=color, alpha=0.2, style=":", label="max") plt.fill_between(ts_percentile_5.index, ts_percentile_5.values, ts_percentile_95.values, facecolor=color, alpha=0.1) plt.fill_between(ts_quartile_1.index, ts_quartile_1.values, ts_quartile_3.values, facecolor=color, alpha=0.1) ts = df.TEMP ax2 = ax.twinx() # https://jakevdp.github.io/PythonDataScienceHandbook/03.11-working-with-time-series.html#Digging-into-the-data ts_mean = ts.groupby(ts.index.time).mean() ts_median = ts.groupby(ts.index.time).median() ts_quartile_1 = ts.groupby(ts.index.time).quantile(0.25) ts_quartile_3 = ts.groupby(ts.index.time).quantile(0.75) ts_percentile_5 = ts.groupby(ts.index.time).quantile(0.05) ts_percentile_95 = ts.groupby(ts.index.time).quantile(0.95) ts_min = ts.groupby(ts.index.time).min() ts_max = ts.groupby(ts.index.time).max() color = "red" ax2 = ts_mean.plot(y='duration', figsize=(18, 12), color=color, label="mean", alpha=0.75) ts_median.plot(ax=ax2, color=color, label="median", style="--", alpha=0.75) ts_quartile_1.plot(ax=ax2, color=color, alpha=0.5, style="-.", label="1st quartile") ts_quartile_3.plot(ax=ax2, color=color, alpha=0.5, style="-.", label="3rd quartile") ts_percentile_5.plot(ax=ax2, color=color, alpha=0.25, style=":", label="5th percentile") ts_percentile_95.plot(ax=ax2, color=color, alpha=0.25, style=":", label="95th percentile") ts_min.plot(ax=ax2, color=color, alpha=0.2, style=":", label="min") ts_max.plot(ax=ax2, color=color, alpha=0.2, style=":", label="max") plt.fill_between(ts_percentile_5.index, ts_percentile_5.values, ts_percentile_95.values, facecolor=color, alpha=0.1) plt.fill_between(ts_quartile_1.index, ts_quartile_1.values, ts_quartile_3.values, facecolor=color, alpha=0.1) ax.legend(loc='upper left') ax2.legend(loc='upper right'); ax.set_xlabel('Time') ax.set_ylabel('PM10'); ax2.set_ylabel('Temperature'); ax = df.PM10.groupby(df.index.time).mean().plot(figsize=(18,6), color="blue") ax.set_xlabel("Time") ax2 = ax.twinx() df.TEMP.groupby(df.index.time).mean().plot(ax=ax2, color="red") ax.legend(loc='upper left') ax2.legend(loc='upper right'); ax = df.PM10.groupby(df.index.weekday).mean().plot(figsize=(18,6), color="blue") ax.set_xlabel("Weekday") ax2 = ax.twinx() df.TEMP.groupby(df.index.weekday).mean().plot(ax=ax2, color="red") ax.legend(loc='upper left') ax2.legend(loc='upper right'); ax = df.PM10.groupby(df.index.month).mean().plot(figsize=(18,6), color="blue") ax.set_xlabel("Month") ax2 = ax.twinx() df.TEMP.groupby(df.index.month).mean().plot(ax=ax2, color="red") ax.legend(loc='upper left') ax2.legend(loc='upper right'); Explanation: Analyse de la concentration en particules fines (PM10) End of explanation <END_TASK>
15,934
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Factorial HMM Example synthetic data Step1: Test out learned distribution inside of SMC We'll compare it against a baseline of "bootstrap" SMC, which proposes from the transition dynamics of the individual HMMs. Step2: Look at rate of path coalescence
<ASSISTANT_TASK:> Python Code: devices = factorial_hmm.gen_devices() T = 50 np.random.seed(20) X, Y = factorial_hmm.gen_dataset(devices, T) plt.figure(figsize=(15,3.5)) plt.plot(Y) plt.figure(figsize=(15,10)) plt.imshow((X*devices).T, interpolation='None', aspect=1); plt.yticks(np.arange(len(devices)), devices); print len(devices), 2**len(devices) trace_train = [] trace_validation = [] dist_est = cde.ConditionalBinaryMADE(len(devices)+1, len(devices), H=300, num_layers=4) if USE_GPU: dist_est.cuda() dist_est.load_state_dict(torch.load('../saved/trained_hmm_params.rar')) Explanation: Factorial HMM Example synthetic data: 20 different "devices", each with different power consumptions, turning on and off following separate Markov models End of explanation X_hat_bootstrap, ancestry_bootstrap, ESS_bootstrap = \ factorial_hmm.run_smc(devices, Y, 500, factorial_hmm.baseline_proposal, verbose=False) Y_hat_bootstrap = np.dot(X_hat_bootstrap, devices) nn_proposal = factorial_hmm.make_nn_proposal(dist_est) X_hat_nn, ancestry_nn, ESS_nn = \ factorial_hmm.run_smc(devices, Y, 500, nn_proposal, verbose=False) Y_hat_nn = np.dot(X_hat_nn, devices) plt.hist(ESS_bootstrap, histtype='stepfilled', linewidth=2, alpha=0.5, bins=20,edgeColor='k') plt.hist(ESS_nn, histtype='stepfilled', linewidth=2, alpha=0.5, bins=20,edgeColor='k') plt.xlim([0,plt.xlim()[1]]) plt.legend(['bootstrap', 'nnsmc']) plt.title('Histogram of effective sample size of SMC filtering distribution'); plt.figure(figsize=(16,4)) plt.title('Ancestral paths for bootstrap proposals (blue) and nn (green)') plt.plot(ancestry_bootstrap.T, color=sns.color_palette()[0]); plt.plot(ancestry_nn.T, color=sns.color_palette()[1]); plt.ylim(0,ancestry_nn.shape[0]) plt.xlim(0,T-1); plt.figure(figsize=(14,3.25)) plt.plot(np.dot(X_hat_nn, devices).T, color=sns.color_palette()[1], alpha=0.1) plt.plot(np.arange(len(Y)), Y,'k--') plt.xlim([0,T-1]) plt.xlabel('Time step') plt.ylabel('Total energy usage') Explanation: Test out learned distribution inside of SMC We'll compare it against a baseline of "bootstrap" SMC, which proposes from the transition dynamics of the individual HMMs. End of explanation ANC_PRIOR = [] ANC_NN = [] def count_uniques(ancestry): K, T = ancestry.shape counts = np.empty((T,), dtype=int) for t in xrange(T): counts[t] = len(np.unique(ancestry[:,t])) return counts def run_iter(): X,Y = factorial_hmm.gen_dataset(devices, T=30) X_particles_baseline, ancestry_baseline, _ = \ factorial_hmm.run_smc(devices, Y, 100, factorial_hmm.baseline_proposal, verbose=False) print "smc complete" X_particles, ancestry_nnsmc, _ = \ factorial_hmm.run_smc(devices, Y, 500, nn_proposal, verbose=False) print "nn complete" ANC_PRIOR.append(count_uniques(ancestry_baseline)) ANC_NN.append(count_uniques(ancestry_nnsmc)) return X,Y for i in xrange(10): print "iteration", i+1 X_tmp, Y_tmp = run_iter() plt.figure(figsize=(8,3.5)) plt.plot(np.arange(len(X_tmp)), np.mean(ANC_PRIOR, 0)); plt.plot(np.arange(len(X_tmp)), np.mean(ANC_NN, 0)); plt.legend(['Bootstrap SMC', 'NN-SMC'], loc='upper left') pm = np.mean(ANC_PRIOR, 0) psd = np.std(ANC_PRIOR, 0) safe_lb = (pm - psd) * (pm - psd > 1.0) + (pm - psd <= 1.0) plt.fill_between(np.arange(len(X_tmp)), safe_lb, pm+psd, alpha=0.25, color=sns.color_palette()[0]); pm = np.mean(ANC_NN, 0) psd = np.std(ANC_NN, 0) plt.fill_between(np.arange(len(X_tmp)), pm-psd, pm+psd, alpha=0.25, color=sns.color_palette()[1]); plt.semilogy(); plt.xlabel('Time step') plt.ylabel('Surviving paths') plt.ylim(1, 100) plt.xlim(0, len(X_tmp)-1) plt.tight_layout() Explanation: Look at rate of path coalescence End of explanation <END_TASK>
15,935
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Gaussian Process Regression in Pytorch Thomas Viehmann, &#116;&#118;&#64;&#108;&#101;&#114;&#110;&#97;&#112;&#112;&#97;&#114;&#97;&#116;&#46;&#100;&#101; Modelled after GPFlow Regression notebook by James Hensman Step1: Let's have a regression example Step2: Creating the model Not adapted to the data yet... Step3: Maximum-A-Posteriori One commonly used approach to model selection is to maximize the marginal log likelihood. This is the "gp" equivalent of a maximum-likelihood estimate. Step4: Hamiltonian Monte Carlo We can go more Bayesian by putting a prior on the parameters and do Hamiltonian Monte Carlo to draw parameters. Step5: Plotting simulated functions (Note that the simulations are for the de-noised functions - i.e. without the noise contribution of the likelihood.) Step6: Sparse Regression
<ASSISTANT_TASK:> Python Code: from matplotlib import pyplot %matplotlib inline import IPython import torch import numpy import sys, os sys.path.append(os.path.join(os.getcwd(),'..')) pyplot.style.use('ggplot') import candlegp import candlegp.training.hmc Explanation: Gaussian Process Regression in Pytorch Thomas Viehmann, &#116;&#118;&#64;&#108;&#101;&#114;&#110;&#97;&#112;&#112;&#97;&#114;&#97;&#116;&#46;&#100;&#101; Modelled after GPFlow Regression notebook by James Hensman End of explanation N = 12 X = torch.rand(N,1).double() Y = (torch.sin(12*X) + 0.6*torch.cos(25*X) + torch.randn(N,1).double()*0.1+3.0).squeeze(1) pyplot.figure() pyplot.plot(X.numpy(), Y.numpy(), 'kx', mew=2) Explanation: Let's have a regression example End of explanation k = candlegp.kernels.Matern52(1, lengthscales=torch.tensor([0.3], dtype=torch.double), variance=torch.tensor([1.0], dtype=torch.double)) mean = candlegp.mean_functions.Linear(torch.tensor([1], dtype=torch.double), torch.tensor([0], dtype=torch.double)) m = candlegp.models.GPR(X, Y.unsqueeze(1), kern=k, mean_function=mean) m.likelihood.variance.set(torch.tensor([0.01], dtype=torch.double)) m xstar = torch.linspace(0,1,100).double() mu, var = m.predict_y(xstar.unsqueeze(1)) cred_size = (var**0.5*2).squeeze(1) mu = mu.squeeze(1) pyplot.plot(xstar.numpy(),mu.data.numpy(),'b') pyplot.fill_between(xstar.numpy(),mu.data.numpy()+cred_size.data.numpy(), mu.data.numpy()-cred_size.data.numpy(),facecolor='0.75') pyplot.plot(X.numpy(), Y.numpy(), 'kx', mew=2) Explanation: Creating the model Not adapted to the data yet... End of explanation opt = torch.optim.LBFGS(m.parameters(), lr=1e-2, max_iter=40) def eval_model(): obj = m() opt.zero_grad() obj.backward() return obj for i in range(50): obj = m() opt.zero_grad() obj.backward() opt.step(eval_model) if i%5==0: print(i,':',obj.item()) m xstar = torch.linspace(0,1,100).double() mu, var = m.predict_y(xstar.unsqueeze(1)) cred_size = (var**0.5*2).squeeze(1) mu = mu.squeeze(1) pyplot.plot(xstar.numpy(),mu.data.numpy(),'b') pyplot.fill_between(xstar.numpy(),mu.data.numpy()+cred_size.data.numpy(), mu.data.numpy()-cred_size.data.numpy(),facecolor='0.75') pyplot.plot(X.numpy(), Y.numpy(), 'kx', mew=2) Explanation: Maximum-A-Posteriori One commonly used approach to model selection is to maximize the marginal log likelihood. This is the "gp" equivalent of a maximum-likelihood estimate. End of explanation k2 = candlegp.kernels.RBF(1, lengthscales=torch.tensor([0.3], dtype=torch.double), variance=torch.tensor([1.0], dtype=torch.double)) mean2 = candlegp.mean_functions.Linear(torch.tensor([1], dtype=torch.double), torch.tensor([0], dtype=torch.double)) m2 = candlegp.models.GPR(X, Y.unsqueeze(1), kern=k2, mean_function=mean2) m2.load_state_dict(m.state_dict()) dt = torch.double m2.likelihood.variance.prior = candlegp.priors.Gamma(1.0,1.0, dtype=dt) m2.kern.variance.prior = candlegp.priors.Gamma(1.0,1.0, dtype=dt) m2.kern.lengthscales.prior = candlegp.priors.Gamma(1.0,1.0,dtype=dt) m2.mean_function.A.prior = candlegp.priors.Gaussian(0.0,10.0, dtype=dt) m2.mean_function.b.prior = candlegp.priors.Gaussian(0.0,10.0, dtype=dt) print("likelihood with priors",m2().item()) m2 # res = candlegp.training.hmc.hmc_sample(m2,500,0.2,burn=50, thin=10) res = candlegp.training.hmc.hmc_sample(m2,50,0.2,burn=50, thin=10) pyplot.plot(res[0]); pyplot.title("likelihood"); for (n,p0),p,c in zip(m.named_parameters(),res[1:],['r','g','b','y','b']): pyplot.plot(torch.stack(p).squeeze().numpy(), c=c, label=n) pyplot.plot((0,len(p)),(p0.data.view(-1)[0],p0.data.view(-1)[0]), c=c) pyplot.legend(); Explanation: Hamiltonian Monte Carlo We can go more Bayesian by putting a prior on the parameters and do Hamiltonian Monte Carlo to draw parameters. End of explanation xstar = torch.linspace(0,1,100).double() mc_params = torch.stack([torch.cat(p, dim=0).view(-1) for p in res[1:]], dim=1) allsims = [] for ps in mc_params[:50]: for mp, p in zip(m2.parameters(), ps): with torch.no_grad(): mp.set(p) allsims.append(m2.predict_f_samples(xstar.unsqueeze(1), 1).squeeze(0).t()) allsims = torch.cat(allsims, dim=0) pyplot.plot(xstar.numpy(),allsims.data.numpy().T, 'b', lw=2, alpha=0.1) mu, var = m.predict_y(xstar.unsqueeze(1)) cred_size = (var**0.5*2).squeeze(1) mu = mu.squeeze(1) pyplot.plot(xstar.numpy(),mu.data.numpy(),'b') pyplot.fill_between(xstar.numpy(),mu.data.numpy()+cred_size.data.numpy(), mu.data.numpy()-cred_size.data.numpy(),facecolor='0.75') pyplot.plot(X.numpy(), Y.numpy(), 'kx', mew=2) Explanation: Plotting simulated functions (Note that the simulations are for the de-noised functions - i.e. without the noise contribution of the likelihood.) End of explanation k3 = candlegp.kernels.RBF(1, lengthscales=torch.tensor([0.3], dtype=torch.double), variance=torch.tensor([1.0], dtype=torch.double)) mean3 = candlegp.mean_functions.Linear(torch.tensor([1], dtype=torch.double), torch.tensor([0], dtype=torch.double)) m3 = candlegp.models.SGPR(X, Y.unsqueeze(1), k3, X[:7].clone(), mean_function=mean3) m3.likelihood.variance.set(torch.tensor([0.01], dtype=torch.double)) m3 opt = torch.optim.LBFGS(m3.parameters(), lr=1e-2, max_iter=40) def eval_model(): obj = m3() opt.zero_grad() obj.backward() return obj for i in range(50): obj = m3() opt.zero_grad() obj.backward() opt.step(eval_model) if i%5==0: print(i,':',obj.item()) m3 xstar = torch.linspace(0,1,100).double() mu, var = m3.predict_y(xstar.unsqueeze(1)) cred_size = (var**0.5*2).squeeze(1) mu = mu.squeeze(1) pyplot.plot(xstar.numpy(),mu.data.numpy(),'b') pyplot.fill_between(xstar.numpy(),mu.data.numpy()+cred_size.data.numpy(), mu.data.numpy()-cred_size.data.numpy(),facecolor='0.75') pyplot.plot(X.numpy(), Y.numpy(), 'kx', mew=2) pyplot.plot(m3.Z.data.numpy(), torch.zeros(m3.Z.size(0)).numpy(),'o') Explanation: Sparse Regression End of explanation <END_TASK>
15,936
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Basic Principles of Machine Learning Here we'll dive into the basic principles of machine learning, and how to utilize them via the Scikit-Learn API. After briefly introducing scikit-learn's Estimator object, we'll cover supervised learning, including classification and regression problems, and unsupervised learning, including dimensionality reduction and clustering problems. Step1: The Scikit-learn Estimator Object Every algorithm is exposed in scikit-learn via an ''Estimator'' object. For instance a linear regression is implemented as so Step2: Estimator parameters Step3: Estimated Model parameters Step4: The model found a line with a slope 2 and intercept 1, as we'd expect. Supervised Learning Step5: You can also do probabilistic predictions Step6: Exercise Use a different estimator on the same problem Step7: As above, we can plot a line of best fit Step8: Scikit-learn also has some more sophisticated models, which can respond to finer features in the data Step9: Whether either of these is a "good" fit or not depends on a number of things; we'll discuss details of how to choose a model later in the tutorial. Exercise Explore the RandomForestRegressor object using IPython's help features (i.e. put a question mark after the object). What arguments are available to RandomForestRegressor? How does the above plot change if you change these arguments? These class-level arguments are known as hyperparameters, and we will discuss later how you to select hyperparameters in the model validation section. Unsupervised Learning Step10: Clustering Step11: Recap Step12: A more useful way to look at the results is to view the confusion matrix, or the matrix showing the frequency of inputs and outputs Step13: For each class, all 50 training samples are correctly identified. But this does not mean that our model is perfect! In particular, such a model generalizes extremely poorly to new data. We can simulate this by splitting our data into a training set and a testing set. Scikit-learn contains some convenient routines to do this Step14: This paints a better picture of the true performance of our classifier Step15: Original source on the scikit-learn website Quick Application Step16: Let's plot a few of these Step17: Here the data is simply each pixel value within an 8x8 grid Step18: So our data have 1797 samples in 64 dimensions. Unsupervised Learning Step19: We see here that the digits are fairly well-separated in the parameter space; this tells us that a supervised classification algorithm should perform fairly well. Let's give it a try. Classification on Digits Let's try a classification task on the digits. The first thing we'll want to do is split the digits into a training and testing sample Step20: Let's use a simple logistic regression which (despite its confusing name) is a classification algorithm Step21: We can check our classification accuracy by comparing the true values of the test set to the predictions Step22: This single number doesn't tell us where we've gone wrong Step23: We might also take a look at some of the outputs along with their predicted labels. We'll make the bad labels red
<ASSISTANT_TASK:> Python Code: %matplotlib inline import numpy as np import matplotlib.pyplot as plt Explanation: Basic Principles of Machine Learning Here we'll dive into the basic principles of machine learning, and how to utilize them via the Scikit-Learn API. After briefly introducing scikit-learn's Estimator object, we'll cover supervised learning, including classification and regression problems, and unsupervised learning, including dimensionality reduction and clustering problems. End of explanation from sklearn.linear_model import LinearRegression Explanation: The Scikit-learn Estimator Object Every algorithm is exposed in scikit-learn via an ''Estimator'' object. For instance a linear regression is implemented as so: End of explanation model = LinearRegression(normalize=True) print(model.normalize) print(model) model2 = LinearRegression() model? Explanation: Estimator parameters: All the parameters of an estimator can be set when it is instantiated, and have suitable default values: End of explanation x = np.arange(10) y = 2 * x + 1 print(x) print(y) plt.plot(x, y, 'o'); # The input data for sklearn is 2D: (samples == 10 x features == 1) X = x[:, np.newaxis] print(X) print(y) # fit the model on our data model.fit(X, y) # underscore at the end indicates a fit parameter print(model.coef_) print(model.intercept_) # residual error around fit model.residues_ model.score(X, y) Explanation: Estimated Model parameters: When data is fit with an estimator, parameters are estimated from the data at hand. All the estimated parameters are attributes of the estimator object ending by an underscore: End of explanation from sklearn import neighbors, datasets iris = datasets.load_iris() X, y = iris.data, iris.target # create the model knn = neighbors.KNeighborsClassifier(n_neighbors=5) # fit the model knn.fit(X, y) # What kind of iris has 3cm x 5cm sepal and 4cm x 2cm petal? # call the "predict" method: result = knn.predict([[5, 3, 4, 2],]) print(iris.target_names[result]) Explanation: The model found a line with a slope 2 and intercept 1, as we'd expect. Supervised Learning: Classification and Regression In Supervised Learning, we have a dataset consisting of both features and labels. The task is to construct an estimator which is able to predict the label of an object given the set of features. A relatively simple example is predicting the species of iris given a set of measurements of its flower. This is a relatively simple task. Some more complicated examples are: given a multicolor image of an object through a telescope, determine whether that object is a star, a quasar, or a galaxy. given a photograph of a person, identify the person in the photo. given a list of movies a person has watched and their personal rating of the movie, recommend a list of movies they would like (So-called recommender systems: a famous example is the Netflix Prize). What these tasks have in common is that there is one or more unknown quantities associated with the object which needs to be determined from other observed quantities. Supervised learning is further broken down into two categories, classification and regression. In classification, the label is discrete, while in regression, the label is continuous. For example, in astronomy, the task of determining whether an object is a star, a galaxy, or a quasar is a classification problem: the label is from three distinct categories. On the other hand, we might wish to estimate the age of an object based on such observations: this would be a regression problem, because the label (age) is a continuous quantity. Classification Example K nearest neighbors (kNN) is one of the simplest learning strategies: given a new, unknown observation, look up in your reference database which ones have the closest features and assign the predominant class. Let's try it out on our iris classification problem: End of explanation iris.target_names knn.predict_proba([[5, 3, 4, 2],]) from fig_code import plot_iris_knn plot_iris_knn() Explanation: You can also do probabilistic predictions: End of explanation # Create some simple data import numpy as np np.random.seed(0) X = np.random.random(size=(20, 1)) y = 3 * X.squeeze() + 2 + np.random.randn(20) plt.plot(X.squeeze(), y, 'o'); Explanation: Exercise Use a different estimator on the same problem: sklearn.svm.SVC. Note that you don't have to know what it is to use it. We're simply trying out the interface here If you finish early, try to create a similar plot as above with the SVC estimator. Regression Example One of the simplest regression problems is fitting a line to data, which we saw above. Scikit-learn also contains more sophisticated regression algorithms End of explanation model = LinearRegression() model.fit(X, y) # Plot the data and the model prediction X_fit = np.linspace(0, 1, 100)[:, np.newaxis] y_fit = model.predict(X_fit) plt.plot(X.squeeze(), y, 'o') plt.plot(X_fit.squeeze(), y_fit); Explanation: As above, we can plot a line of best fit: End of explanation # Fit a Random Forest from sklearn.ensemble import RandomForestRegressor model = RandomForestRegressor(n_estimators=10, max_depth=5) model.fit(X, y) # Plot the data and the model prediction X_fit = np.linspace(0, 1, 100)[:, np.newaxis] y_fit = model.predict(X_fit) plt.plot(X.squeeze(), y, 'o') plt.plot(X_fit.squeeze(), y_fit); Explanation: Scikit-learn also has some more sophisticated models, which can respond to finer features in the data: End of explanation X, y = iris.data, iris.target from sklearn.decomposition import PCA pca = PCA(n_components=0.95) pca.fit(X) X_reduced = pca.transform(X) print("Reduced dataset shape:", X_reduced.shape) import pylab as plt plt.scatter(X_reduced[:, 0], X_reduced[:, 1], #c=y, cmap='RdYlBu') print("Meaning of the 2 components:") for component in pca.components_: print(" + ".join("%.3f x %s" % (value, name) for value, name in zip(component, iris.feature_names))) Explanation: Whether either of these is a "good" fit or not depends on a number of things; we'll discuss details of how to choose a model later in the tutorial. Exercise Explore the RandomForestRegressor object using IPython's help features (i.e. put a question mark after the object). What arguments are available to RandomForestRegressor? How does the above plot change if you change these arguments? These class-level arguments are known as hyperparameters, and we will discuss later how you to select hyperparameters in the model validation section. Unsupervised Learning: Dimensionality Reduction and Clustering Unsupervised Learning addresses a different sort of problem. Here the data has no labels, and we are interested in finding similarities between the objects in question. In a sense, you can think of unsupervised learning as a means of discovering labels from the data itself. Unsupervised learning comprises tasks such as dimensionality reduction, clustering, and density estimation. For example, in the iris data discussed above, we can used unsupervised methods to determine combinations of the measurements which best display the structure of the data. As we'll see below, such a projection of the data can be used to visualize the four-dimensional dataset in two dimensions. Some more involved unsupervised learning problems are: given detailed observations of distant galaxies, determine which features or combinations of features best summarize the information. given a mixture of two sound sources (for example, a person talking over some music), separate the two (this is called the blind source separation problem). given a video, isolate a moving object and categorize in relation to other moving objects which have been seen. Sometimes the two may even be combined: e.g. Unsupervised learning can be used to find useful features in heterogeneous data, and then these features can be used within a supervised framework. Dimensionality Reduction: PCA Principle Component Analysis (PCA) is a dimension reduction technique that can find the combinations of variables that explain the most variance. Consider the iris dataset. It cannot be visualized in a single 2D plot, as it has 4 features. We are going to extract 2 combinations of sepal and petal dimensions to visualize it: End of explanation from sklearn.cluster import KMeans k_means = KMeans(n_clusters=3, random_state=0) # Fixing the RNG in kmeans k_means.fit(X) y_pred = k_means.predict(X) plt.scatter(X_reduced[:, 0], X_reduced[:, 1], c=y_pred, cmap='RdYlBu'); Explanation: Clustering: K-means Clustering groups together observations that are homogeneous with respect to a given criterion, finding ''clusters'' in the data. Note that these clusters will uncover relevent hidden structure of the data only if the criterion used highlights it. End of explanation from sklearn.neighbors import KNeighborsClassifier X, y = iris.data, iris.target clf = KNeighborsClassifier(n_neighbors=1) clf.fit(X, y) y_pred = clf.predict(X) print(np.all(y == y_pred)) Explanation: Recap: Scikit-learn's estimator interface Scikit-learn strives to have a uniform interface across all methods, and we'll see examples of these below. Given a scikit-learn estimator object named model, the following methods are available: Available in all Estimators model.fit() : fit training data. For supervised learning applications, this accepts two arguments: the data X and the labels y (e.g. model.fit(X, y)). For unsupervised learning applications, this accepts only a single argument, the data X (e.g. model.fit(X)). Available in supervised estimators model.predict() : given a trained model, predict the label of a new set of data. This method accepts one argument, the new data X_new (e.g. model.predict(X_new)), and returns the learned label for each object in the array. model.predict_proba() : For classification problems, some estimators also provide this method, which returns the probability that a new observation has each categorical label. In this case, the label with the highest probability is returned by model.predict(). model.score() : for classification or regression problems, most (all?) estimators implement a score method. Scores are between 0 and 1, with a larger score indicating a better fit. Available in unsupervised estimators model.predict() : predict labels in clustering algorithms. model.transform() : given an unsupervised model, transform new data into the new basis. This also accepts one argument X_new, and returns the new representation of the data based on the unsupervised model. model.fit_transform() : some estimators implement this method, which more efficiently performs a fit and a transform on the same input data. Model Validation An important piece of machine learning is model validation: that is, determining how well your model will generalize from the training data to future unlabeled data. Let's look at an example using the nearest neighbor classifier. This is a very simple classifier: it simply stores all training data, and for any unknown quantity, simply returns the label of the closest training point. With the iris data, it very easily returns the correct prediction for each of the input points: End of explanation from sklearn.metrics import confusion_matrix print(confusion_matrix(y, y_pred)) Explanation: A more useful way to look at the results is to view the confusion matrix, or the matrix showing the frequency of inputs and outputs: End of explanation from sklearn.model_selection import train_test_split Xtrain, Xtest, ytrain, ytest = train_test_split(X, y, random_state=23) clf.fit(Xtrain, ytrain) ypred = clf.predict(Xtest) print(confusion_matrix(ytest, ypred)) Xtest.shape[0] / len(X) Explanation: For each class, all 50 training samples are correctly identified. But this does not mean that our model is perfect! In particular, such a model generalizes extremely poorly to new data. We can simulate this by splitting our data into a training set and a testing set. Scikit-learn contains some convenient routines to do this: End of explanation from IPython.display import Image Image("http://scikit-learn.org/dev/_static/ml_map.png") Explanation: This paints a better picture of the true performance of our classifier: apparently there is some confusion between the second and third species, which we might anticipate given what we've seen of the data above. This is why it's extremely important to use a train/test split when evaluating your models. We'll go into more depth on model evaluation later in this tutorial. Flow Chart: How to Choose your Estimator This is a flow chart created by scikit-learn super-contributor Andreas Mueller which gives a nice summary of which algorithms to choose in various situations. Keep it around as a handy reference! End of explanation from sklearn import datasets digits = datasets.load_digits() digits.images.shape Explanation: Original source on the scikit-learn website Quick Application: Optical Character Recognition To demonstrate the above principles on a more interesting problem, let's consider OCR (Optical Character Recognition) – that is, recognizing hand-written digits. In the wild, this problem involves both locating and identifying characters in an image. Here we'll take a shortcut and use scikit-learn's set of pre-formatted digits, which is built-in to the library. Loading and visualizing the digits data We'll use scikit-learn's data access interface and take a look at this data: End of explanation fig, axes = plt.subplots(10, 10, figsize=(8, 8)) fig.subplots_adjust(hspace=0.1, wspace=0.1) for i, ax in enumerate(axes.flat): ax.imshow(digits.images[i], cmap='binary', interpolation='nearest') ax.text(0.05, 0.05, str(digits.target[i]), transform=ax.transAxes, color='green') ax.set_xticks([]) ax.set_yticks([]) Explanation: Let's plot a few of these: End of explanation # The images themselves print(digits.images.shape) print(digits.images[0]) # The data for use in our algorithms print(digits.data.shape) print(digits.data[0]) # The target label print(digits.target) Explanation: Here the data is simply each pixel value within an 8x8 grid: End of explanation from sklearn.manifold import Isomap iso = Isomap(n_components=2) data_projected = iso.fit_transform(digits.data) data_projected.shape plt.scatter(data_projected[:, 0], data_projected[:, 1], c=digits.target, edgecolor='none', alpha=0.5, cmap=plt.cm.get_cmap('nipy_spectral', 10)); plt.colorbar(label='digit label', ticks=range(10)) plt.clim(-0.5, 9.5) Explanation: So our data have 1797 samples in 64 dimensions. Unsupervised Learning: Dimensionality Reduction We'd like to visualize our points within the 64-dimensional parameter space, but it's difficult to plot points in 64 dimensions! Instead we'll reduce the dimensions to 2, using an unsupervised method. Here, we'll make use of a manifold learning algorithm called Isomap, and transform the data to two dimensions. End of explanation from sklearn.model_selection import train_test_split Xtrain, Xtest, ytrain, ytest = train_test_split(digits.data, digits.target, random_state=2) print(Xtrain.shape, Xtest.shape) Explanation: We see here that the digits are fairly well-separated in the parameter space; this tells us that a supervised classification algorithm should perform fairly well. Let's give it a try. Classification on Digits Let's try a classification task on the digits. The first thing we'll want to do is split the digits into a training and testing sample: End of explanation from sklearn.linear_model import LogisticRegression clf = LogisticRegression(penalty='l2') clf.fit(Xtrain, ytrain) ypred = clf.predict(Xtest) Explanation: Let's use a simple logistic regression which (despite its confusing name) is a classification algorithm: End of explanation from sklearn.metrics import accuracy_score accuracy_score(ytest, ypred) Explanation: We can check our classification accuracy by comparing the true values of the test set to the predictions: End of explanation from sklearn.metrics import confusion_matrix print(confusion_matrix(ytest, ypred)) plt.imshow(np.log(confusion_matrix(ytest, ypred)), cmap='Blues', interpolation='nearest') plt.grid(False) plt.ylabel('true') plt.xlabel('predicted'); Explanation: This single number doesn't tell us where we've gone wrong: one nice way to do this is to use the confusion matrix End of explanation fig, axes = plt.subplots(10, 10, figsize=(8, 8)) fig.subplots_adjust(hspace=0.1, wspace=0.1) for i, ax in enumerate(axes.flat): ax.imshow(Xtest[i].reshape(8, 8), cmap='binary') ax.text(0.05, 0.05, str(ypred[i]), transform=ax.transAxes, color='green' if (ytest[i] == ypred[i]) else 'red') ax.set_xticks([]) ax.set_yticks([]) Explanation: We might also take a look at some of the outputs along with their predicted labels. We'll make the bad labels red: End of explanation <END_TASK>
15,937
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: This is functionally similar to the the other notebook. All the operations here have been vectorized. This results in much much faster code, but is also much unreadable. The vectorization also necessitated the replacement of the Gauss-Seidel smoother with under-relaxed Jacobi. That change has had some effect since GS is "twice as better" as Jacobi. The Making of a Preconditioner ---Vectorized Version This is a demonstration of a multigrid preconditioned krylov solver in python3. The code and more examples are present on github here. The problem solved is a Poisson equation on a rectangular domain with homogenous dirichlet boundary conditions. Finite difference with cell-centered discretization is used to get a second order accurate solution, that is further improved to 4th order using deferred correction. The first step is a multigrid algorithm. This is the simplest 2D geometric multigrid solver. 1. Multigrid algorithm We need some terminology before going further. - Approximation Step1: 1.1 Smoothing operator This can be a certain number of Jacobi or a Gauss-Seidel iterations. Below is defined smoother that does under-relaxed Jacobi sweeps and returns the result along with the residual. Step2: 1.2 Interpolation Operator This operator takes values on a coarse grid and transfers them onto a fine grid. It is also called prolongation. The function below uses bilinear interpolation for this purpose. 'v' is on a coarse grid and we want to interpolate it on a fine grid and store it in v_f. Step3: 1.3 Restriction This is exactly the opposite of the interpolation. It takes values from the find grid and transfers them onto the coarse grid. It is kind of an averaging process. This is fundamentally different from interpolation. Each coarse grid point is surrounded by four fine grid points. So quite simply we take the value of the coarse point to be the average of 4 fine points. Here 'v' is the fine grid quantity and 'v_c' is the coarse grid quantity Step4: 1.4 Bottom Solver Note that we have looped over the coarse grid in both the cases above. It is easier to access the variables this way. The last part is the Bottom Solver. This must be something that gives us the exact/converged solution to what ever we feed it. What we feed to the bottom solver is the problem at the coarsest level. This has generally has very few points (e.g 2x2=4 in our case) and can be solved exactly by the smoother itself with few iterations. That is what we do here but, any other direct method can also be used. 50 Iterations are used here. If we coarsify to just one point, then just one iteration will solve it exactly. 1.5 V-cycle Now that we have all the parts, we are ready to build our multigrid algorithm. First we will look at a V-cycle. It is self explanatory. It is a recursive function ,i.e., it calls itself. It takes as input an initial guess 'u', the rhs 'f', the number of multigrid levels 'num_levels' among other things. At each level the V cycle calls another V-cycle. At the lowest level the solving is exact. Step5: Thats it! Now we can see it in action. We can use a problem with a known solution to test our code. The following functions set up a rhs for a problem with homogenous dirichlet BC on the unit square. Step6: Let us set up the problem, discretization and solver details. The number of divisions along each dimension is given as a power of two function of the number of levels. In principle this is not required, but having it makes the inter-grid transfers easy. The coarsest problem is going to have a 2-by-2 grid. Step7: Now we can call the solver Step8: True error is the difference of the approximation with the analytical solution. It is largely the discretization error. This what would be present when we solve the discrete equation with a direct/exact method like gaussian elimination. We see that true error stops reducing at the 5th cycle. The approximation is not getting any better after this point. So we can stop after 5 cycles. But, in general we dont know the true error. In practice we use the norm of the (relative) residual as a stopping criterion. As the cycles progress the floating point round-off error limit is reached and the residual also stops decreasing. This was the multigrid V cycle. We can use this as preconditioner to a Krylov solver. But before we get to that let's complete the multigrid introduction by looking at the Full Multi-Grid algorithm. You can skip this section safely. 1.6 Full Multi-Grid We started with a zero initial guess for the V-cycle. Presumably, if we had a better initial guess we would get better results. So we solve a coarse problem exactly and interpolate it onto the fine grid and use that as the initial guess for the V-cycle. The result of doing this recursively is the Full Multi-Grid(FMG) Algorithm. Unlike the V-cycle which was an iterative procedure, FMG is a direct solver. There is no successive improvement of the approximation. It straight away gives us an approximation that is within the discretization error. The FMG algorithm is given below. Step9: Lets call the FMG solver for the same problem Step10: It works wonderfully. The residual is large but the true error is within the discretization level. FMG is said to be scalable because the amount of work needed is linearly proportional to the the size of the problem. In big-O notation, FMG is $\mathcal{O}(N)$. Where N is the number of unknowns. Exact methods (Gaussian Elimination, LU decomposition ) are typically $\mathcal{O}(N^3)$ 2. Stationary iterative methods as preconditioners A preconditioner reduces the condition number of the coefficient matrix, thereby making it easier to solve. We dont explicitly need a matrix because we dont access the elements by index, coefficient matrix or preconditioner. What we do need is the action of the matrix on a vector. That is, we need only the matrix-vector product. The coefficient matrix can be defined as a function that takes in a vector and returns the matrix vector product. Any stationary method has an iteration matrix associated with it. This is easily seen for Jacobi or GS methods. This iteration matrix can be used as a preconditioner. But we dont explicitly need it. The stationary iterative method for solving an equation can be written as a Richardson iteration. When the initial guess is set to zero and one iteration is performed, what you get is the action of the preconditioner on the RHS vector. That is, we get a preconditioner-vector product, which is what we want. This allows us to use any blackbox stationary iterative method as a preconditioner To repeat, if there is a stationary iterative method that you want to use as a preconditioner, set the initial guess to zero, set the RHS to the vector you want to multiply the preconditioner with and perform one iteration of the stationary method. We can use the multigrid V-cycle as a preconditioner this way. We cant use FMG because it is not an iterative method. The matrix as a function can be defined using LinearOperator from scipy.sparse.linalg. It gives us an object which works like a matrix in-so-far as the product with a vector is concerned. It can be used as a regular 2D numpy array in multiplication with a vector. This can be passed to CG(), GMRES() or BiCGStab() as a preconditioner. Having a symmetric preconditioner would be nice because it will retain the symmetry if the original problem is symmetric and we can still use CG. If the preconditioner is not symmetric CG will not converge, and we would have to use a more general solver. Below is the code for defining a V-Cycle preconditioner. The default is one V-cycle. In the V-cycle, the defaults are one pre-sweep, one post-sweep. Step11: Let us define the Poisson matrix also as a LinearOperator Step12: The nested function is required because "matvec" in LinearOperator takes only one argument-- the vector. But we require the grid details and boundary condition information to create the Poisson matrix. Now will use these to solve a problem. Unlike earlier where we used an analytical solution and RHS, we will start with a random vector which will be our exact solution, and multiply it with the Poisson matrix to get the Rhs vector for the problem. There is no analytical equation associated with the matrix equation. The scipy sparse solve routines do not return the number of iterations performed. We can use this wrapper to get the number of iterations Step13: Lets look at what happens with and without the preconditioner. Step14: Without the preconditioner ~150 iterations were needed, where as with the V-cycle preconditioner the solution was obtained in far fewer iterations. Let's try with CG
<ASSISTANT_TASK:> Python Code: import numpy as np Explanation: This is functionally similar to the the other notebook. All the operations here have been vectorized. This results in much much faster code, but is also much unreadable. The vectorization also necessitated the replacement of the Gauss-Seidel smoother with under-relaxed Jacobi. That change has had some effect since GS is "twice as better" as Jacobi. The Making of a Preconditioner ---Vectorized Version This is a demonstration of a multigrid preconditioned krylov solver in python3. The code and more examples are present on github here. The problem solved is a Poisson equation on a rectangular domain with homogenous dirichlet boundary conditions. Finite difference with cell-centered discretization is used to get a second order accurate solution, that is further improved to 4th order using deferred correction. The first step is a multigrid algorithm. This is the simplest 2D geometric multigrid solver. 1. Multigrid algorithm We need some terminology before going further. - Approximation: - Residual: - Exact solution (of the discrete problem) - Correction This is a geometric multigrid algorithm, where a series of nested grids are used. There are four parts to a multigrid algorithm - Smoothing Operator (a.k.a Relaxation) - Restriction Operator - Interpolation Operator (a.k.a Prolongation Operator) - Bottom solver We will define each of these in sequence. These operators act of different quantities that are stored at the cell center. We will get to exactly what later on. To begin import numpy. End of explanation def Jacrelax(nx,ny,u,f,iters=1): ''' under-relaxed Jacobi iteration ''' dx=1.0/nx; dy=1.0/ny Ax=1.0/dx**2; Ay=1.0/dy**2 Ap=1.0/(2.0*(Ax+Ay)) #Dirichlet BC u[ 0,:] = -u[ 1,:] u[-1,:] = -u[-2,:] u[:, 0] = -u[:, 1] u[:,-1] = -u[:,-2] for it in range(iters): u[1:nx+1,1:ny+1] = 0.8*Ap*(Ax*(u[2:nx+2,1:ny+1] + u[0:nx,1:ny+1]) + Ay*(u[1:nx+1,2:ny+2] + u[1:nx+1,0:ny]) - f[1:nx+1,1:ny+1])+0.2*u[1:nx+1,1:ny+1] #Dirichlet BC u[ 0,:] = -u[ 1,:] u[-1,:] = -u[-2,:] u[:, 0] = -u[:, 1] u[:,-1] = -u[:,-2] res=np.zeros([nx+2,ny+2]) res[1:nx+1,1:ny+1]=f[1:nx+1,1:ny+1]-(( Ax*(u[2:nx+2,1:ny+1]+u[0:nx,1:ny+1]) + Ay*(u[1:nx+1,2:ny+2]+u[1:nx+1,0:ny]) - 2.0*(Ax+Ay)*u[1:nx+1,1:ny+1])) return u,res Explanation: 1.1 Smoothing operator This can be a certain number of Jacobi or a Gauss-Seidel iterations. Below is defined smoother that does under-relaxed Jacobi sweeps and returns the result along with the residual. End of explanation def prolong(nx,ny,v): ''' interpolate 'v' to the fine grid ''' v_f=np.zeros([2*nx+2,2*ny+2]) v_f[1:2*nx:2 ,1:2*ny:2 ] = 0.5625*v[1:nx+1,1:ny+1]+0.1875*(v[0:nx ,1:ny+1]+v[1:nx+1,0:ny] )+0.0625*v[0:nx ,0:ny ] v_f[2:2*nx+1:2,1:2*ny:2 ] = 0.5625*v[1:nx+1,1:ny+1]+0.1875*(v[2:nx+2,1:ny+1]+v[1:nx+1,0:ny] )+0.0625*v[2:nx+2,0:ny ] v_f[1:2*nx:2 ,2:2*ny+1:2] = 0.5625*v[1:nx+1,1:ny+1]+0.1875*(v[0:nx ,1:ny+1]+v[1:nx+1,2:ny+2])+0.0625*v[0:nx ,2:ny+2] v_f[2:2*nx+1:2,2:2*ny+1:2] = 0.5625*v[1:nx+1,1:ny+1]+0.1875*(v[2:nx+2,1:ny+1]+v[1:nx+1,2:ny+2])+0.0625*v[2:nx+2,2:ny+2] return v_f Explanation: 1.2 Interpolation Operator This operator takes values on a coarse grid and transfers them onto a fine grid. It is also called prolongation. The function below uses bilinear interpolation for this purpose. 'v' is on a coarse grid and we want to interpolate it on a fine grid and store it in v_f. End of explanation def restrict(nx,ny,v): ''' restrict 'v' to the coarser grid ''' v_c=np.zeros([nx+2,ny+2]) v_c[1:nx+1,1:ny+1]=0.25*(v[1:2*nx:2,1:2*ny:2]+v[1:2*nx:2,2:2*ny+1:2]+v[2:2*nx+1:2,1:2*ny:2]+v[2:2*nx+1:2,2:2*ny+1:2]) return v_c Explanation: 1.3 Restriction This is exactly the opposite of the interpolation. It takes values from the find grid and transfers them onto the coarse grid. It is kind of an averaging process. This is fundamentally different from interpolation. Each coarse grid point is surrounded by four fine grid points. So quite simply we take the value of the coarse point to be the average of 4 fine points. Here 'v' is the fine grid quantity and 'v_c' is the coarse grid quantity End of explanation def V_cycle(nx,ny,num_levels,u,f,level=1): if(level==num_levels):#bottom solve u,res=Jacrelax(nx,ny,u,f,iters=50) return u,res #Step 1: Relax Au=f on this grid u,res=Jacrelax(nx,ny,u,f,iters=1) #Step 2: Restrict residual to coarse grid res_c=restrict(nx//2,ny//2,res) #Step 3:Solve A e_c=res_c on the coarse grid. (Recursively) e_c=np.zeros_like(res_c) e_c,res_c=V_cycle(nx//2,ny//2,num_levels,e_c,res_c,level+1) #Step 4: Interpolate(prolong) e_c to fine grid and add to u u+=prolong(nx//2,ny//2,e_c) #Step 5: Relax Au=f on this grid u,res=Jacrelax(nx,ny,u,f,iters=1) return u,res Explanation: 1.4 Bottom Solver Note that we have looped over the coarse grid in both the cases above. It is easier to access the variables this way. The last part is the Bottom Solver. This must be something that gives us the exact/converged solution to what ever we feed it. What we feed to the bottom solver is the problem at the coarsest level. This has generally has very few points (e.g 2x2=4 in our case) and can be solved exactly by the smoother itself with few iterations. That is what we do here but, any other direct method can also be used. 50 Iterations are used here. If we coarsify to just one point, then just one iteration will solve it exactly. 1.5 V-cycle Now that we have all the parts, we are ready to build our multigrid algorithm. First we will look at a V-cycle. It is self explanatory. It is a recursive function ,i.e., it calls itself. It takes as input an initial guess 'u', the rhs 'f', the number of multigrid levels 'num_levels' among other things. At each level the V cycle calls another V-cycle. At the lowest level the solving is exact. End of explanation #analytical solution def Uann(x,y): return (x**3-x)*(y**3-y) #RHS corresponding to above def source(x,y): return 6*x*y*(x**2+ y**2 - 2) Explanation: Thats it! Now we can see it in action. We can use a problem with a known solution to test our code. The following functions set up a rhs for a problem with homogenous dirichlet BC on the unit square. End of explanation #input max_cycles = 30 nlevels = 6 NX = 2*2**(nlevels-1) NY = 2*2**(nlevels-1) tol = 1e-15 #the grid has one layer of ghost cellss uann=np.zeros([NX+2,NY+2])#analytical solution u =np.zeros([NX+2,NY+2])#approximation f =np.zeros([NX+2,NY+2])#RHS #calcualte the RHS and exact solution DX=1.0/NX DY=1.0/NY xc=np.linspace(0.5*DX,1-0.5*DX,NX) yc=np.linspace(0.5*DY,1-0.5*DY,NY) XX,YY=np.meshgrid(xc,yc,indexing='ij') uann[1:NX+1,1:NY+1]=Uann(XX,YY) f[1:NX+1,1:NY+1] =source(XX,YY) Explanation: Let us set up the problem, discretization and solver details. The number of divisions along each dimension is given as a power of two function of the number of levels. In principle this is not required, but having it makes the inter-grid transfers easy. The coarsest problem is going to have a 2-by-2 grid. End of explanation print('mgd2d.py solver:') print('NX:',NX,', NY:',NY,', tol:',tol,'levels: ',nlevels) for it in range(1,max_cycles+1): u,res=V_cycle(NX,NY,nlevels,u,f) rtol=np.max(np.max(np.abs(res))) if(rtol<tol): break error=uann[1:NX+1,1:NY+1]-u[1:NX+1,1:NY+1] print(' cycle: ',it,', L_inf(res.)= ',rtol,',L_inf(true error): ',np.max(np.max(np.abs(error)))) error=uann[1:NX+1,1:NY+1]-u[1:NX+1,1:NY+1] print('L_inf (true error): ',np.max(np.max(np.abs(error)))) Explanation: Now we can call the solver End of explanation def FMG(nx,ny,num_levels,f,nv=1,level=1): if(level==num_levels):#bottom solve u=np.zeros([nx+2,ny+2]) u,res=Jacrelax(nx,ny,u,f,iters=50) return u,res #Step 1: Restrict the rhs to a coarse grid f_c=restrict(nx//2,ny//2,f) #Step 2: Solve the coarse grid problem using FMG u_c,_=FMG(nx//2,ny//2,num_levels,f_c,nv,level+1) #Step 3: Interpolate u_c to the fine grid u=prolong(nx//2,ny//2,u_c) #step 4: Execute 'nv' V-cycles for _ in range(nv): u,res=V_cycle(nx,ny,num_levels-level,u,f) return u,res Explanation: True error is the difference of the approximation with the analytical solution. It is largely the discretization error. This what would be present when we solve the discrete equation with a direct/exact method like gaussian elimination. We see that true error stops reducing at the 5th cycle. The approximation is not getting any better after this point. So we can stop after 5 cycles. But, in general we dont know the true error. In practice we use the norm of the (relative) residual as a stopping criterion. As the cycles progress the floating point round-off error limit is reached and the residual also stops decreasing. This was the multigrid V cycle. We can use this as preconditioner to a Krylov solver. But before we get to that let's complete the multigrid introduction by looking at the Full Multi-Grid algorithm. You can skip this section safely. 1.6 Full Multi-Grid We started with a zero initial guess for the V-cycle. Presumably, if we had a better initial guess we would get better results. So we solve a coarse problem exactly and interpolate it onto the fine grid and use that as the initial guess for the V-cycle. The result of doing this recursively is the Full Multi-Grid(FMG) Algorithm. Unlike the V-cycle which was an iterative procedure, FMG is a direct solver. There is no successive improvement of the approximation. It straight away gives us an approximation that is within the discretization error. The FMG algorithm is given below. End of explanation print('mgd2d.py FMG solver:') print('NX:',NX,', NY:',NY,', levels: ',nlevels) u,res=FMG(NX,NY,nlevels,f,nv=1) rtol=np.max(np.max(np.abs(res))) print(' FMG L_inf(res.)= ',rtol) error=uann[1:NX+1,1:NY+1]-u[1:NX+1,1:NY+1] print('L_inf (true error): ',np.max(np.max(np.abs(error)))) Explanation: Lets call the FMG solver for the same problem End of explanation from scipy.sparse.linalg import LinearOperator,bicgstab,cg def MGVP(nx,ny,num_levels): ''' Multigrid Preconditioner. Returns a (scipy.sparse.linalg.) LinearOperator that can be passed to Krylov solvers as a preconditioner. ''' def pc_fn(v): u =np.zeros([nx+2,ny+2]) f =np.zeros([nx+2,ny+2]) f[1:nx+1,1:ny+1] =v.reshape([nx,ny]) #in practice this copying can be avoived #perform one V cycle u,res=V_cycle(nx,ny,num_levels,u,f) return u[1:nx+1,1:ny+1].reshape(v.shape) M=LinearOperator((nx*ny,nx*ny), matvec=pc_fn) return M Explanation: It works wonderfully. The residual is large but the true error is within the discretization level. FMG is said to be scalable because the amount of work needed is linearly proportional to the the size of the problem. In big-O notation, FMG is $\mathcal{O}(N)$. Where N is the number of unknowns. Exact methods (Gaussian Elimination, LU decomposition ) are typically $\mathcal{O}(N^3)$ 2. Stationary iterative methods as preconditioners A preconditioner reduces the condition number of the coefficient matrix, thereby making it easier to solve. We dont explicitly need a matrix because we dont access the elements by index, coefficient matrix or preconditioner. What we do need is the action of the matrix on a vector. That is, we need only the matrix-vector product. The coefficient matrix can be defined as a function that takes in a vector and returns the matrix vector product. Any stationary method has an iteration matrix associated with it. This is easily seen for Jacobi or GS methods. This iteration matrix can be used as a preconditioner. But we dont explicitly need it. The stationary iterative method for solving an equation can be written as a Richardson iteration. When the initial guess is set to zero and one iteration is performed, what you get is the action of the preconditioner on the RHS vector. That is, we get a preconditioner-vector product, which is what we want. This allows us to use any blackbox stationary iterative method as a preconditioner To repeat, if there is a stationary iterative method that you want to use as a preconditioner, set the initial guess to zero, set the RHS to the vector you want to multiply the preconditioner with and perform one iteration of the stationary method. We can use the multigrid V-cycle as a preconditioner this way. We cant use FMG because it is not an iterative method. The matrix as a function can be defined using LinearOperator from scipy.sparse.linalg. It gives us an object which works like a matrix in-so-far as the product with a vector is concerned. It can be used as a regular 2D numpy array in multiplication with a vector. This can be passed to CG(), GMRES() or BiCGStab() as a preconditioner. Having a symmetric preconditioner would be nice because it will retain the symmetry if the original problem is symmetric and we can still use CG. If the preconditioner is not symmetric CG will not converge, and we would have to use a more general solver. Below is the code for defining a V-Cycle preconditioner. The default is one V-cycle. In the V-cycle, the defaults are one pre-sweep, one post-sweep. End of explanation def Laplace(nx,ny): ''' Action of the Laplace matrix on a vector v ''' def mv(v): u =np.zeros([nx+2,ny+2]) u[1:nx+1,1:ny+1]=v.reshape([nx,ny]) dx=1.0/nx; dy=1.0/ny Ax=1.0/dx**2; Ay=1.0/dy**2 #BCs. Needs to be generalized! u[ 0,:] = -u[ 1,:] u[-1,:] = -u[-2,:] u[:, 0] = -u[:, 1] u[:,-1] = -u[:,-2] ut = (Ax*(u[2:nx+2,1:ny+1]+u[0:nx,1:ny+1]) + Ay*(u[1:nx+1,2:ny+2]+u[1:nx+1,0:ny]) - 2.0*(Ax+Ay)*u[1:nx+1,1:ny+1]) return ut.reshape(v.shape) A = LinearOperator((nx*ny,nx*ny), matvec=mv) return A Explanation: Let us define the Poisson matrix also as a LinearOperator End of explanation def solve_sparse(solver,A, b,tol=1e-10,maxiter=500,M=None): num_iters = 0 def callback(xk): nonlocal num_iters num_iters+=1 x,status=solver(A, b,tol=tol,maxiter=maxiter,callback=callback,M=M) return x,status,num_iters Explanation: The nested function is required because "matvec" in LinearOperator takes only one argument-- the vector. But we require the grid details and boundary condition information to create the Poisson matrix. Now will use these to solve a problem. Unlike earlier where we used an analytical solution and RHS, we will start with a random vector which will be our exact solution, and multiply it with the Poisson matrix to get the Rhs vector for the problem. There is no analytical equation associated with the matrix equation. The scipy sparse solve routines do not return the number of iterations performed. We can use this wrapper to get the number of iterations End of explanation A = Laplace(NX,NY) #Exact solution and RHS uex=np.random.rand(NX*NY,1) b=A*uex #Multigrid Preconditioner M=MGVP(NX,NY,nlevels) u,info,iters=solve_sparse(bicgstab,A,b,tol=1e-10,maxiter=500) print('Without preconditioning. status:',info,', Iters: ',iters) error=uex-u print('error :',np.max(np.abs(error))) u,info,iters=solve_sparse(bicgstab,A,b,tol=1e-10,maxiter=500,M=M) print('With preconditioning. status:',info,', Iters: ',iters) error=uex-u print('error :',np.max(np.abs(error))) Explanation: Lets look at what happens with and without the preconditioner. End of explanation u,info,iters=solve_sparse(cg,A,b,tol=1e-10,maxiter=500) print('Without preconditioning. status:',info,', Iters: ',iters) error=uex-u print('error :',np.max(np.abs(error))) u,info,iters=solve_sparse(cg,A,b,tol=1e-10,maxiter=500,M=M) print('With preconditioning. status:',info,', Iters: ',iters) error=uex-u print('error :',np.max(np.abs(error))) Explanation: Without the preconditioner ~150 iterations were needed, where as with the V-cycle preconditioner the solution was obtained in far fewer iterations. Let's try with CG: End of explanation <END_TASK>
15,938
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Lab 1 Step1: The core tables in the data warehouse are derived from 5 separate core operational systems (each with many tables) Step2: Question Step3: Question - How many columns of data does each table have (sorted by most to least?) - Which table has the most columns of data? Step4: Previewing sample rows of data values In the BigQuery UI, find the Resources panel and search for catalog_sales. You may need to add the qwiklabs-resources project to your UI by clicking + Add Data -> Pin a project and entering qwiklabs-resources Click on the catalog_sales table name for the tpcds_2t_baseline dataset under qwiklabs-resources Question - How many rows are in the table? - How large is the table in TB? Hint Step5: A note on our data Step6: Running the first benchmark test Now let's run the first query against our dataset and note the execution time. Tip Step7: It should execute in just a few seconds. Then try running it again and see if you get the same performance. BigQuery will automatically cache the results from the first time you ran the query and then serve those same results to you when you can the query again. We can confirm this by analyzing the query job statistics. Viewing BigQuery job statistics Let's list our five most recent query jobs run on BigQuery using the bq command line interface. Then we will get even more detail on our most recent job with the bq show command. Be sure to replace the job id with your own. Step9: Looking at the job statistics we can see our most recent query hit cache - cacheHit Step11: 132 GB will be processed. At the time of writing, BigQuery pricing is \$5 per 1 TB (or 1000 GB) of data after the first free 1 TB each month. Assuming we've exhausted our 1 TB free this month, this would be \$0.66 to run. Now let's run it an ensure we're not pulling from cache so we get an accurate time-to-completion benchmark. Step12: If you're an experienced BigQuery user, you likely have seen these same metrics in the Web UI as well as highlighted in the red box below Step13: Use the BigQuery Data Transfer Service to copy an existing dataset Enable the BigQuery Data Transfer Service API Navigate to the BigQuery console and the existing qwiklabs-resources dataset Click Copy Dataset In the pop-up, choose your project name and the newly created dataset name from the previous step Click Copy Wait for the transfer to complete Verify you now have the baseline data in your project Run the below query and confirm you see data. Note that if you omit the project-id ahead of the dataset name in the FROM clause, BigQuery will assume your default project. Step14: Setup an automated test Running each of the 99 queries manually via the Console UI would be a tedious effort. We'll show you how you can run all 99 programmatically and automatically log the output (time and GB processed) to a log file for analysis. Below is a shell script that Step15: Viewing the benchmark results As part of the benchmark test, we stored the processing time of each query into a new perf BigQuery table. We can query that table and get some performance stats for our test. First are each of the tests we ran Step16: And finally, the overall statistics for the entire test Step17: Benchmarking all 99 queries As we mentioned before, we already ran all 99 queries and recorded the results and made them available for you to query in a public table Step18: And the results of the complete test
<ASSISTANT_TASK:> Python Code: %%bigquery SELECT dataset_id, table_id, -- Convert bytes to GB. ROUND(size_bytes/pow(10,9),2) as size_gb, -- Convert UNIX EPOCH to a timestamp. TIMESTAMP_MILLIS(creation_time) AS creation_time, TIMESTAMP_MILLIS(last_modified_time) as last_modified_time, row_count, CASE WHEN type = 1 THEN 'table' WHEN type = 2 THEN 'view' ELSE NULL END AS type FROM `qwiklabs-resources.tpcds_2t_baseline.__TABLES__` ORDER BY size_gb DESC Explanation: Lab 1: Explore and Benchmark a BigQuery Dataset for Performance Overview In this lab you will take an existing 2TB+ TPC-DS benchmark dataset and learn the data warehouse optimization methods you can apply to the dataset in BigQuery to improve performance. What you'll do In this lab, you will learn how to: Use BigQuery to access and query the TPC-DS benchmark dataset Run pre-defined queries to establish baseline performance benchmarks Prerequisites This is an advanced level SQL lab. Before taking it, you should have experience with SQL. Familiarity with BigQuery is also highly recommended. If you need to get up to speed in these areas, you should take this Data Analyst series of labs first: Quest: BigQuery for Data Analysts Once you're ready, scroll down to learn about the services you will be using and how to properly set up your lab environment. BigQuery BigQuery is Google's fully managed, NoOps, low cost analytics database. With BigQuery you can query terabytes and terabytes of data without managing infrastructure or needing a database administrator. BigQuery uses SQL and takes advantage of the pay-as-you-go model. BigQuery allows you to focus on analyzing data to find meaningful insights. TPC-DS Background In order to benchmark the performance of a data warehouse we first must get tables and data to run queries against. There is a public organization, TPC, that provides large benchmarking datasets to companies explicitly for this purpose. The purpose of TPC benchmarks is to provide relevant, objective performance data to industry users. The TPC-DS Dataset we will be using comprises of 25 tables and 99 queries that simulate common data analysis tasks. View the full documentation here. Exploring TPC-DS in BigQuery The TPC-DS tables have been loaded into BigQuery and you will explore ways to optimize the performance of common queries by using BigQuery data warehousing best practices. We have limited the size to 2TB for the timing of this lab but the dataset itself can be expanded as needed. Note: The TPC Benchmark and TPC-DS are trademarks of the Transaction Processing Performance Council (http://www.tpc.org). The Cloud DW benchmark is derived from the TPC-DS Benchmark and as such is not comparable to published TPC-DS results. Exploring the Schema with SQL Question: - How many tables are in the dataset? - What is the name of the largest table (in GB)? How many rows does it have? End of explanation %%bigquery SELECT * FROM `qwiklabs-resources.tpcds_2t_baseline.INFORMATION_SCHEMA.COLUMNS` Explanation: The core tables in the data warehouse are derived from 5 separate core operational systems (each with many tables): These systems are driven by the core functions of our retail business. As you can see, our store accepts sales from online (web), mail-order (catalog), and in-store. The business must keep track of inventory and can offer promotional discounts on items sold. Exploring all available columns of data Question: - How many columns of data are in the entire dataset (all tables)? End of explanation %%bigquery SELECT * FROM `qwiklabs-resources.tpcds_2t_baseline.INFORMATION_SCHEMA.COLUMNS` WHERE is_partitioning_column = 'YES' OR clustering_ordinal_position IS NOT NULL Explanation: Question: - Are any of the columns of data in this baseline dataset partitioned or clustered? End of explanation %%bigquery SELECT COUNT(column_name) AS column_count, table_name FROM `qwiklabs-resources.tpcds_2t_baseline.INFORMATION_SCHEMA.COLUMNS` GROUP BY table_name ORDER BY column_count DESC, table_name Explanation: Question - How many columns of data does each table have (sorted by most to least?) - Which table has the most columns of data? End of explanation %%bigquery --verbose SELECT cs_item_sk, COUNT(cs_order_number) AS total_orders, SUM(cs_quantity) AS total_quantity, SUM(cs_ext_sales_price) AS total_revenue, SUM(cs_net_profit) AS total_profit FROM `qwiklabs-resources.tpcds_2t_baseline.catalog_sales` GROUP BY cs_item_sk ORDER BY total_orders DESC LIMIT 100 Explanation: Previewing sample rows of data values In the BigQuery UI, find the Resources panel and search for catalog_sales. You may need to add the qwiklabs-resources project to your UI by clicking + Add Data -> Pin a project and entering qwiklabs-resources Click on the catalog_sales table name for the tpcds_2t_baseline dataset under qwiklabs-resources Question - How many rows are in the table? - How large is the table in TB? Hint: Use the Details button in the web UI to quickly access table metadata Question: - Preview the data and find the Catalog Sales Extended Sales Price cs_ext_sales_price field (which is calculated based on product quantity * sales price) - Are there any missing data values for Catalog Sales Quantity (cs_quantity)? - Are there any missing values for cs_ext_ship_cost? For what type of product could this be expected? (Digital products) Create an example sales report Write a query that shows key sales stats for each item sold from the Catalog and execute it in the BigQuery UI: - total orders - total unit quantity - total revenue - total profit - sorted by total orders highest to lowest, limit 100 End of explanation !head --lines=50 'sql/example_baseline_queries.sql' Explanation: A note on our data: The TPC-DS benchmark allows data warehouse practicioners to generate any volume of data programmatically. Since the rows of data are system generated, they may not make the most sense in a business context (like why are we selling our top product at such a huge profit loss!). The good news is that to benchmark our performance we care most about the volume of rows and columns to run our benchmark against. Analyzing query performance Click on Execution details Refer to the chart below (which should be similar to your results) and answer the following questions. Question - How long did it take the query to run? 5.1s - How much data in GB was processed? 150GB - How much slot time was consumed? 1hr 24min - How many rows were input? 2,881,495,086 - How many rows were output as the end result (before the limit)? 23,300 - What does the output rows mean in the context of our query? (23,300 unique cs_item_sk) Side note: Slot Time We know the query took 5.1 seconds to run so what does the 1hr 24 min slot time metric mean? Inside of the BigQuery service are lots of virtual machines that massively process your data and query logic in parallel. These workers, or "slots", work together to process a single query job really quickly. For accounts with on-demand pricing, you can have up to 2,000 slots. So say we had 30 minutes of slot time or 1800 seconds. If the query took 20 seconds in total to run, but it was 1800 seconds worth of work, how many workers at minimum worked on it? 1800/20 = 90 And that's assuming each worker instantly had all the data it needed (no shuffling of data between workers) and was at full capacity for all 20 seconds! In reality, workers have a variety of tasks (waiting for data, reading it, performing computations, and writing data) and also need to compare notes with eachother on what work was already done on the job. The good news for you is that you don't need to worry about optimizing these workers or the underlying data to run perfectly in parallel. That's why BigQuery is a managed service -- there's an entire team dedicated to hardware and data storage optimization. In case you were wondering, the worker limit for your project is 2,000 slots at once. Running a performance benchmark To performance benchmark our data warehouse in BigQuery we need to create more than just a single SQL report. The good news is the TPC-DS dataset ships with 99 standard benchmark queries that we can run and log the performance outcomes. In this lab, we are doing no adjustments to the existing data warehouse tables (no partitioning, no clustering, no nesting) so we can establish a performance benchmark to beat in future labs. Viewing the 99 pre-made SQL queries We have a long SQL file with 99 standard queries against this dataset stored in our /sql/ directory. Let's view the first 50 lines of those baseline queries to get familiar with how we will be performance benchmarking our dataset. End of explanation %%bigquery --verbose # start query 1 in stream 0 using template query96.tpl select count(*) from `qwiklabs-resources.tpcds_2t_baseline.store_sales` as store_sales ,`qwiklabs-resources.tpcds_2t_baseline.household_demographics` as household_demographics ,`qwiklabs-resources.tpcds_2t_baseline.time_dim` as time_dim, `qwiklabs-resources.tpcds_2t_baseline.store` as store where ss_sold_time_sk = time_dim.t_time_sk and ss_hdemo_sk = household_demographics.hd_demo_sk and ss_store_sk = s_store_sk and time_dim.t_hour = 8 and time_dim.t_minute >= 30 and household_demographics.hd_dep_count = 5 and store.s_store_name = 'ese' order by count(*) limit 100; Explanation: Running the first benchmark test Now let's run the first query against our dataset and note the execution time. Tip: You can use the --verbose flag in %%bigquery magics to return the job and completion time. End of explanation !bq ls -j -a -n 5 !bq show --format=prettyjson -j 612a4b28-cb5c-4e0b-ad5b-ebd51c3b2439 Explanation: It should execute in just a few seconds. Then try running it again and see if you get the same performance. BigQuery will automatically cache the results from the first time you ran the query and then serve those same results to you when you can the query again. We can confirm this by analyzing the query job statistics. Viewing BigQuery job statistics Let's list our five most recent query jobs run on BigQuery using the bq command line interface. Then we will get even more detail on our most recent job with the bq show command. Be sure to replace the job id with your own. End of explanation %%bash bq query \ --dry_run \ --nouse_cache \ --use_legacy_sql=false \ \ select count(*) from \`qwiklabs-resources.tpcds_2t_baseline.store_sales\` as store_sales ,\`qwiklabs-resources.tpcds_2t_baseline.household_demographics\` as household_demographics ,\`qwiklabs-resources.tpcds_2t_baseline.time_dim\` as time_dim, \`qwiklabs-resources.tpcds_2t_baseline.store\` as store where ss_sold_time_sk = time_dim.t_time_sk and ss_hdemo_sk = household_demographics.hd_demo_sk and ss_store_sk = s_store_sk and time_dim.t_hour = 8 and time_dim.t_minute >= 30 and household_demographics.hd_dep_count = 5 and store.s_store_name = 'ese' order by count(*) limit 100; # Convert bytes to GB 132086388641 / 1e+9 Explanation: Looking at the job statistics we can see our most recent query hit cache - cacheHit: true and therefore - totalBytesProcessed: 0. While this is great in normal uses for BigQuery (you aren't charged for queries that hit cache) it kind of ruins our performance test. While cache is super useful we want to disable it for testing purposes. Disabling Cache and Dry Running Queries As of the time this lab was created, you can't pass a flag to %%bigquery iPython notebook magics to disable cache or to quickly see the amount of data processed. So we will use the traditional bq command line interface in bash. First we will do a dry run of the query without processing any data just to see how many bytes of data would be processed. Then we will remove that flag and ensure nouse_cache is set to avoid hitting cache as well. End of explanation %%bash bq query \ --nouse_cache \ --use_legacy_sql=false \ \ select count(*) from \`qwiklabs-resources.tpcds_2t_baseline.store_sales\` as store_sales ,\`qwiklabs-resources.tpcds_2t_baseline.household_demographics\` as household_demographics ,\`qwiklabs-resources.tpcds_2t_baseline.time_dim\` as time_dim, \`qwiklabs-resources.tpcds_2t_baseline.store\` as store where ss_sold_time_sk = time_dim.t_time_sk and ss_hdemo_sk = household_demographics.hd_demo_sk and ss_store_sk = s_store_sk and time_dim.t_hour = 8 and time_dim.t_minute >= 30 and household_demographics.hd_dep_count = 5 and store.s_store_name = 'ese' order by count(*) limit 100; Explanation: 132 GB will be processed. At the time of writing, BigQuery pricing is \$5 per 1 TB (or 1000 GB) of data after the first free 1 TB each month. Assuming we've exhausted our 1 TB free this month, this would be \$0.66 to run. Now let's run it an ensure we're not pulling from cache so we get an accurate time-to-completion benchmark. End of explanation %%bash export PROJECT_ID=$(gcloud config list --format 'value(core.project)') export BENCHMARK_DATASET_NAME=tpcds_2t_baseline # Name of the dataset you want to create ## Create a BigQuery dataset for tpcds_2t_flat_part_clust if it doesn't exist datasetexists=$(bq ls -d | grep -w $BENCHMARK_DATASET_NAME) if [ -n "$datasetexists" ]; then echo -e "BigQuery dataset $BENCHMARK_DATASET_NAME already exists, let's not recreate it." else echo "Creating BigQuery dataset titled: $BENCHMARK_DATASET_NAME" bq --location=US mk --dataset \ --description 'Benchmark Dataset' \ $PROJECT:$BENCHMARK_DATASET_NAME echo "\nHere are your current datasets:" bq ls fi Explanation: If you're an experienced BigQuery user, you likely have seen these same metrics in the Web UI as well as highlighted in the red box below: It's a matter of preference whether you do your work in the Web UI or the command line -- each has it's advantages. One major advantage of using the bq command line interface is the ability to create a script that will run the remaining 98 benchmark queries for us and log the results. Copy the qwiklabs-resources dataset into your own GCP project We will use the new BigQuery Transfer Service to quickly copy our large dataset from the qwiklabs-resources GCP project into your own so you can perform the benchmarking. Create a new baseline dataset in your project End of explanation %%bigquery SELECT COUNT(*) AS store_transaction_count FROM tpcds_2t_baseline.store_sales Explanation: Use the BigQuery Data Transfer Service to copy an existing dataset Enable the BigQuery Data Transfer Service API Navigate to the BigQuery console and the existing qwiklabs-resources dataset Click Copy Dataset In the pop-up, choose your project name and the newly created dataset name from the previous step Click Copy Wait for the transfer to complete Verify you now have the baseline data in your project Run the below query and confirm you see data. Note that if you omit the project-id ahead of the dataset name in the FROM clause, BigQuery will assume your default project. End of explanation %%bash # runs the SQL queries from the TPCDS benchmark # Pull the current Google Cloud Platform project name BQ_DATASET="tpcds_2t_baseline" # let's start by benchmarking our baseline dataset QUERY_FILE_PATH="./sql/example_baseline_queries.sql" # the full test is on 99_baseline_queries but that will take 80+ mins to run IFS=";" # create perf table to keep track of run times for all 99 queries printf "\033[32;1m Housekeeping tasks... \033[0m\n\n"; printf "Creating a reporting table perf to track how fast each query runs..."; perf_table_ddl="CREATE TABLE IF NOT EXISTS $BQ_DATASET.perf(performance_test_num int64, query_num int64, elapsed_time_sec int64, ran_on int64)" bq rm -f $BQ_DATASET.perf bq query --nouse_legacy_sql $perf_table_ddl start=$(date +%s) index=0 for select_stmt in $(<$QUERY_FILE_PATH)  do # run the test until you hit a line with the string 'END OF BENCHMARK' in the file if [[ "$select_stmt" == *'END OF BENCHMARK'* ]]; then break fi printf "\n\033[32;1m Let's benchmark this query... \033[0m\n"; printf "$select_stmt"; SECONDS=0; bq query --use_cache=false --nouse_legacy_sql $select_stmt # critical to turn cache off for this test duration=$SECONDS # get current timestamp in milliseconds ran_on=$(date +%s) index=$((index+1)) printf "\n\033[32;1m Here's how long it took... \033[0m\n\n"; echo "Query $index ran in $(($duration / 60)) minutes and $(($duration % 60)) seconds." printf "\n\033[32;1m Writing to our benchmark table... \033[0m\n\n"; insert_stmt="insert into $BQ_DATASET.perf(performance_test_num, query_num, elapsed_time_sec, ran_on) values($start, $index, $duration, $ran_on)" printf "$insert_stmt" bq query --nouse_legacy_sql $insert_stmt done end=$(date +%s) printf "Benchmark test complete" Explanation: Setup an automated test Running each of the 99 queries manually via the Console UI would be a tedious effort. We'll show you how you can run all 99 programmatically and automatically log the output (time and GB processed) to a log file for analysis. Below is a shell script that: 1. Accepts a BigQuery dataset to benchmark 2. Accepts a list of semi-colon separated queries to run 3. Loops through each query and calls the bq query command 4. Records the execution time into a separate BigQuery performance table perf Execute the below statement and follow along with the results as you benchmark a few example queries (don't worry, we've already ran the full 99 recently so you won't have to). After executing, wait 1-2 minutes for the benchmark test to complete End of explanation %%bigquery SELECT * FROM tpcds_2t_baseline.perf WHERE # Let's only pull the results from our most recent test performance_test_num = (SELECT MAX(performance_test_num) FROM tpcds_2t_baseline.perf) ORDER BY ran_on Explanation: Viewing the benchmark results As part of the benchmark test, we stored the processing time of each query into a new perf BigQuery table. We can query that table and get some performance stats for our test. First are each of the tests we ran: End of explanation %%bigquery SELECT TIMESTAMP_SECONDS(MAX(performance_test_num)) AS test_date, MAX(performance_test_num) AS latest_performance_test_num, COUNT(DISTINCT query_num) AS count_queries_benchmarked, SUM(elapsed_time_sec) AS total_time_sec, MIN(elapsed_time_sec) AS fastest_query_time_sec, MAX(elapsed_time_sec) AS slowest_query_time_sec FROM tpcds_2t_baseline.perf WHERE performance_test_num = (SELECT MAX(performance_test_num) FROM tpcds_2t_baseline.perf) Explanation: And finally, the overall statistics for the entire test: End of explanation %%bigquery SELECT TIMESTAMP_SECONDS(performance_test_num) AS test_date, query_num, TIMESTAMP_SECONDS(ran_on) AS query_ran_on, TIMESTAMP_SECONDS(ran_on + elapsed_time_sec) AS query_completed_on, elapsed_time_sec FROM `qwiklabs-resources.tpcds_2t_baseline.perf` # public table WHERE # Let's only pull the results from our most recent test performance_test_num = (SELECT MAX(performance_test_num) FROM `qwiklabs-resources.tpcds_2t_baseline.perf`) ORDER BY ran_on Explanation: Benchmarking all 99 queries As we mentioned before, we already ran all 99 queries and recorded the results and made them available for you to query in a public table: End of explanation %%bigquery SELECT TIMESTAMP_SECONDS(MAX(performance_test_num)) AS test_date, COUNT(DISTINCT query_num) AS count_queries_benchmarked, SUM(elapsed_time_sec) AS total_time_sec, ROUND(SUM(elapsed_time_sec)/60,2) AS total_time_min, MIN(elapsed_time_sec) AS fastest_query_time_sec, MAX(elapsed_time_sec) AS slowest_query_time_sec, ROUND(AVG(elapsed_time_sec),2) AS avg_query_time_sec FROM `qwiklabs-resources.tpcds_2t_baseline.perf` WHERE performance_test_num = (SELECT MAX(performance_test_num) FROM `qwiklabs-resources.tpcds_2t_baseline.perf`) Explanation: And the results of the complete test: End of explanation <END_TASK>
15,939
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Задание 1 Написать функцию на языке Python, формирующую список всех исполнителей композиций жанра Рок. Список должен быть упорядочен в порядке убывания. Step1: Задание 2 Написать функцию на языке Python, создающую таблицу Студентов Students(id, name, gpa). Ключ - id. Step2: Проверим, что таблица создана Step3: Задание 3 Для созданной выше функции реализовть возможность добавления списка студентов вида [['Ivanov', 1.2], ['Petrov', 2.3]]. ID новых студентов должно начинаться с максимального ID в таблице + 1. (Например, если в таблице максимальный ID - 10, то у Петрова должно быть - 11, у Иванова - 12). Функция должна предполагать вставки списка любой ограниченной длины. Получаем max(id) + 1 Step4: Задание 4 Добавить таблицу Факультетов Faculties(fac_id, name). Для таблицы Students добавить новое поле fac_id с внещним ключом на таблицу факультетов. Step5: Задание 5 Написать функцию, осуществляющую обновления всех факультетов у каждого студента. Функция должна выводить информацию о студенту, приглашение на вход для обновления факультета и обновление факультета. При возникновение вставки функция должна обрабатывать исключение, и продолжать работу. Для начала добавим в таблицу факультетов пару записей Step6: Задание 6 Написать функцию, осущетсвляющую перевод части учеников на новый факультет. На входе
<ASSISTANT_TASK:> Python Code: def task1(): cursor = db.cursor() cursor.execute(''' select distinct ar.Name from tracks t inner join albums al on t.albumid = al.albumid inner join artists ar on al.artistid = ar.artistid inner join genres g on t.genreid = g.genreid where g.name = 'Rock' ''') ar = cursor.fetchall() return [x[0] for x in ar] task1() Explanation: Задание 1 Написать функцию на языке Python, формирующую список всех исполнителей композиций жанра Рок. Список должен быть упорядочен в порядке убывания. End of explanation def task2(): cursor=db.cursor() cursor.execute(''' DROP TABLE IF EXISTS students''') cursor.execute(''' CREATE TABLE Students(id INTEGERE PRIMARY KEY, name TEXT, gpa NUMBER(10,2))''') db.commit() task2() Explanation: Задание 2 Написать функцию на языке Python, создающую таблицу Студентов Students(id, name, gpa). Ключ - id. End of explanation %%sql select * from students Explanation: Проверим, что таблица создана End of explanation %%sql select coalesce(max(id)+1, 1) as new_id from students def task3(l_students): cursor = db.cursor() cursor.execute( ''' SELECT COALESCE(MAX(ID)+1, 1) AS new_id FROM students''') new_id = cursor.fetchone()[0] for i, student in enumerate(l_students): cursor.execute(''' INSERT INTO Students(id, name, gpa) VALUES(?,?,?)''', (new_id + i, student[0], student[1])) db.commit() task3([['Ivanov', 3.2], ['Petrov', 4.2]]) %%sql SELECT * FROM Students Explanation: Задание 3 Для созданной выше функции реализовть возможность добавления списка студентов вида [['Ivanov', 1.2], ['Petrov', 2.3]]. ID новых студентов должно начинаться с максимального ID в таблице + 1. (Например, если в таблице максимальный ID - 10, то у Петрова должно быть - 11, у Иванова - 12). Функция должна предполагать вставки списка любой ограниченной длины. Получаем max(id) + 1 End of explanation def task4(): cursor = db.cursor() cursor.execute('''DROP TABLE IF EXISTS faculties''') cursor.execute('''CREATE TABLE faculties(fac_id INTEGER PRIMARY KEY, name TEXT)''') cursor.execute('''ALTER TABLE students ADD fac_id INTEGER REFERENCES faculties(fac_id)''') db.commit() task4() %%sql select * from faculties %%sql select * from Students Explanation: Задание 4 Добавить таблицу Факультетов Faculties(fac_id, name). Для таблицы Students добавить новое поле fac_id с внещним ключом на таблицу факультетов. End of explanation %%sql INSERT INTO faculties(fac_id, name) VALUES (1, 'IT'), (2, 'KIB'), (3, 'Math') %%sql select * from faculties a = input('1 {}', '2') def task5(): cursor = db.cursor() cursor.execute('Select id, name, gpa from Students') a = cursor.fetchall() for x in a: print("Введите факультет для студента {} с id = {} и gpa = {}".format(x[1], x[0], x[2])) fac_name = input() cursor.execute("SELECT fac_id from faculties where name = ?", (fac_name, )) # Проверяем есть ли такая запись try: fac_id = cursor.fetchone()[0] except TypeError: continue cursor.execute("Update students set fac_id = ? where id = ?", (fac_id, x[0],)) db.commit() task5() %%sql SELECT * FROM students task5() %%sql SELECT * FROM Students Explanation: Задание 5 Написать функцию, осуществляющую обновления всех факультетов у каждого студента. Функция должна выводить информацию о студенту, приглашение на вход для обновления факультета и обновление факультета. При возникновение вставки функция должна обрабатывать исключение, и продолжать работу. Для начала добавим в таблицу факультетов пару записей End of explanation def task6(fac_name, l_id): cursor = db.cursor() cursor.execute( ''' SELECT COALESCE(MAX(fac_id)+1, 1) AS new_fac_id FROM faculties''') new_id = cursor.fetchone()[0] cursor.execute(''' INSERT INTO faculties(fac_id, name) VALUES(?,?)''', (new_id, fac_name,)) for x in l_id: cursor.execute(''' Update students set fac_id = ? where id = ?''', (new_id, x, )) db.commit() task6('Hist', [1]) %%sql select * from students Explanation: Задание 6 Написать функцию, осущетсвляющую перевод части учеников на новый факультет. На входе: Имя факультета, Список студентов для перехода на новый факультет. На выходе Добавление новой записи в таблицу факультетов, Обновление записей в таблице студентов. End of explanation <END_TASK>
15,940
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Bayesian Hierarchical Stacking Step1: 1. Exploratory Data Analysis <a class="anchor" id="1"></a> The data we have to work with looks at households in Bangladesh, some of which were affected by high levels of arsenic in their water. Would affected households want to switch to a neighbour's well? We'll split the data into a train and test set, and then we'll train six different models to try to predict whether households would switch wells. Then, we'll see how we can stack them when predicting on the test set! But first, let's load it in and visualise it! Each row represents a household, and the features we have available to us are Step2: Next, we'll choose 200 observations to be part of our train set, and 1500 to be part of our test set. Step4: 2. Prepare 6 different candidate models <a class="anchor" id="2"></a> 2.1 Feature Engineering <a class="anchor" id="2.1"></a> First, let's add a few new columns Step5: 2.2 Training <a class="anchor" id="2.2"></a> Each model will be trained in the same way - with a Bernoulli likelihood and a logit link function. Step6: 2.3 Estimate leave-one-out cross-validated score for each training point <a class="anchor" id="2.3"></a> Rather than refitting each model 100 times, we will estimate the leave-one-out cross-validated score using LOO. Step7: 3. Bayesian Hierarchical Stacking <a class="anchor" id="3"></a> 3.1 Prepare stacking datasets <a class="anchor" id="3.1"></a> To determine how the stacking weights should vary across training and test sets, we will need to create "stacking datasets" which include all the features which we want the stacking weights to depend on. How should such features be included? For discrete features, this is easy, we just one-hot-encode them. But for continuous features, we need a trick. In Equation (16), the authors recommend the following Step9: 3.2 Define stacking model <a class="anchor" id="3.2"></a> What we seek to find is a matrix of weights $W$ with which to multiply the models' predictions. Let's define a matrix $Pred$ such that $Pred_{i,k}$ represents the prediction made for point $i$ by model $k$. Then the final prediction for point $i$ will then be Step10: We can now extract the weights with which to weight the different models from the posterior, and then visualise how they vary across the training set. Let's compare them with what the weights would've been if we'd just used fixed stacking weights (computed using ArviZ - see their docs for details). Step11: 4. Evaluate on test set <a class="anchor" id="4"></a> 4.1 Stack predictions <a class="anchor" id="4.1"></a> Now, for each model, let's evaluate the log predictive density for each point in the test set. Once we have predictions for each model, we need to think about how to combine them, such that for each test point, we get a single prediction. We decided we'd do this in three ways Step12: 4.2 Compare methods <a class="anchor" id="4.2"></a> Let's compare the negative log predictive density scores on the test set (note - lower is better)
<ASSISTANT_TASK:> Python Code: !pip install -q numpyro@git+https://github.com/pyro-ppl/numpyro import os from IPython.display import set_matplotlib_formats import arviz as az import matplotlib.pyplot as plt import numpy as np import pandas as pd from scipy.interpolate import BSpline import seaborn as sns import jax import jax.numpy as jnp import numpyro import numpyro.distributions as dist plt.style.use("seaborn") if "NUMPYRO_SPHINXBUILD" in os.environ: set_matplotlib_formats("svg") numpyro.set_host_device_count(4) assert numpyro.__version__.startswith("0.9.2") %matplotlib inline Explanation: Bayesian Hierarchical Stacking: Well Switching Case Study <figure> <img src="https://i.imgur.com/CiUmZKx.jpeg" width="700px"/> <figcaption>Photo by Belinda Fewings, https://unsplash.com/photos/6p-KtXCBGNw.</figcaption> </figure> Table of Contents Intro 1. Exploratory Data Analysis 2. Prepare 6 Different Models 2.1 Feature Engineering 2.2 Training 3. Bayesian Hierarchical Stacking 3.1 Prepare stacking datasets 3.2 Define stacking model 4. Evaluate on test set 4.1 Stack predictions 4.2 Compare methods Conclusion References Intro <a class="anchor" id="1"></a> Suppose you have just fit 6 models to a dataset, and need to choose which one to use to make predictions on your test set. How do you choose which one to use? A couple of common tactics are: - choose the best model based on cross-validation; - average the models, using weights based on cross-validation scores. In the paper Bayesian hierarchical stacking: Some models are (somewhere) useful, a new technique is introduced: average models based on weights which are allowed to vary across according to the input data, based on a hierarchical structure. Here, we'll implement the first case study from that paper - readers are nonetheless encouraged to look at the original paper to find other cases studies, as well as theoretical results. Code from the article (in R / Stan) can be found here. End of explanation wells = pd.read_csv( "http://stat.columbia.edu/~gelman/arm/examples/arsenic/wells.dat", sep=" " ) wells.head() fig, ax = plt.subplots(2, 2, figsize=(12, 6)) fig.suptitle("Target variable plotted against various predictors") sns.scatterplot(data=wells, x="arsenic", y="switch", ax=ax[0][0]) sns.scatterplot(data=wells, x="dist", y="switch", ax=ax[0][1]) sns.barplot( data=wells.groupby("assoc")["switch"].mean().reset_index(), x="assoc", y="switch", ax=ax[1][0], ) ax[1][0].set_ylabel("Proportion switch") sns.barplot( data=wells.groupby("educ")["switch"].mean().reset_index(), x="educ", y="switch", ax=ax[1][1], ) ax[1][1].set_ylabel("Proportion switch"); Explanation: 1. Exploratory Data Analysis <a class="anchor" id="1"></a> The data we have to work with looks at households in Bangladesh, some of which were affected by high levels of arsenic in their water. Would affected households want to switch to a neighbour's well? We'll split the data into a train and test set, and then we'll train six different models to try to predict whether households would switch wells. Then, we'll see how we can stack them when predicting on the test set! But first, let's load it in and visualise it! Each row represents a household, and the features we have available to us are: switch: whether a household switched to another well; arsenic: level of arsenic in drinking water; educ: level of education of "head of household"; dist100: distance to nearest safe-drinking well; assoc: whether the household participates in any community activities. End of explanation np.random.seed(1) train_id = wells.sample(n=200).index test_id = wells.loc[~wells.index.isin(train_id)].sample(n=1500).index y_train = wells.loc[train_id, "switch"].to_numpy() y_test = wells.loc[test_id, "switch"].to_numpy() Explanation: Next, we'll choose 200 observations to be part of our train set, and 1500 to be part of our test set. End of explanation wells["edu0"] = wells["educ"].isin(np.arange(0, 1)).astype(int) wells["edu1"] = wells["educ"].isin(np.arange(1, 6)).astype(int) wells["edu2"] = wells["educ"].isin(np.arange(6, 12)).astype(int) wells["edu3"] = wells["educ"].isin(np.arange(12, 18)).astype(int) wells["logarsenic"] = np.log(wells["arsenic"]) wells["assoc_half"] = wells["assoc"] / 2.0 wells["as_square"] = wells["logarsenic"] ** 2 wells["as_third"] = wells["logarsenic"] ** 3 wells["dist100"] = wells["dist"] / 100.0 wells["intercept"] = 1 def bs(x, knots, degree): Generate the B-spline basis matrix for a polynomial spline. Parameters ---------- x predictor variable. knots locations of internal breakpoints (not padded). degree degree of the piecewise polynomial. Returns ------- pd.DataFrame Spline basis matrix. Notes ----- This mirrors ``bs`` from splines package in R. padded_knots = np.hstack( [[x.min()] * (degree + 1), knots, [x.max()] * (degree + 1)] ) return pd.DataFrame( BSpline(padded_knots, np.eye(len(padded_knots) - degree - 1), degree)(x)[:, 1:], index=x.index, ) knots = np.quantile(wells.loc[train_id, "logarsenic"], np.linspace(0.1, 0.9, num=10)) spline_arsenic = bs(wells["logarsenic"], knots=knots, degree=3) knots = np.quantile(wells.loc[train_id, "dist100"], np.linspace(0.1, 0.9, num=10)) spline_dist = bs(wells["dist100"], knots=knots, degree=3) features_0 = ["intercept", "dist100", "arsenic", "assoc", "edu1", "edu2", "edu3"] features_1 = ["intercept", "dist100", "logarsenic", "assoc", "edu1", "edu2", "edu3"] features_2 = [ "intercept", "dist100", "arsenic", "as_third", "as_square", "assoc", "edu1", "edu2", "edu3", ] features_3 = ["intercept", "dist100", "assoc", "edu1", "edu2", "edu3"] features_4 = ["intercept", "logarsenic", "assoc", "edu1", "edu2", "edu3"] features_5 = ["intercept", "dist100", "logarsenic", "assoc", "educ"] X0 = wells.loc[train_id, features_0].to_numpy() X1 = wells.loc[train_id, features_1].to_numpy() X2 = wells.loc[train_id, features_2].to_numpy() X3 = ( pd.concat([wells.loc[:, features_3], spline_arsenic], axis=1) .loc[train_id] .to_numpy() ) X4 = pd.concat([wells.loc[:, features_4], spline_dist], axis=1).loc[train_id].to_numpy() X5 = wells.loc[train_id, features_5].to_numpy() X0_test = wells.loc[test_id, features_0].to_numpy() X1_test = wells.loc[test_id, features_1].to_numpy() X2_test = wells.loc[test_id, features_2].to_numpy() X3_test = ( pd.concat([wells.loc[:, features_3], spline_arsenic], axis=1) .loc[test_id] .to_numpy() ) X4_test = ( pd.concat([wells.loc[:, features_4], spline_dist], axis=1).loc[test_id].to_numpy() ) X5_test = wells.loc[test_id, features_5].to_numpy() train_x_list = [X0, X1, X2, X3, X4, X5] test_x_list = [X0_test, X1_test, X2_test, X3_test, X4_test, X5_test] K = len(train_x_list) Explanation: 2. Prepare 6 different candidate models <a class="anchor" id="2"></a> 2.1 Feature Engineering <a class="anchor" id="2.1"></a> First, let's add a few new columns: - edu0: whether educ is 0, - edu1: whether educ is between 1 and 5, - edu2: whether educ is between 6 and 11, - edu3: whether educ is between 12 and 17, - logarsenic: natural logarithm of arsenic, - assoc_half: half of assoc, - as_square: natural logarithm of arsenic, squared, - as_third: natural logarithm of arsenic, cubed, - dist100: dist divided by 100, - intercept: just a columns of 1s. We're going to start by fitting 6 different models to our train set: logistic regression using intercept, arsenic, assoc, edu1, edu2, and edu3; same as above, but with logarsenic instead of arsenic; same as the first one, but with square and cubic features as well; same as the first one, but with spline features derived from logarsenic as well; same as the first one, but with spline features derived from dist100 as well; same as the first one, but with educ instead of the binary edu variables. End of explanation def logistic(x, y=None): beta = numpyro.sample("beta", dist.Normal(0, 3).expand([x.shape[1]])) logits = numpyro.deterministic("logits", jnp.matmul(x, beta)) numpyro.sample( "obs", dist.Bernoulli(logits=logits), obs=y, ) fit_list = [] for k in range(K): sampler = numpyro.infer.NUTS(logistic) mcmc = numpyro.infer.MCMC( sampler, num_chains=4, num_samples=1000, num_warmup=1000, progress_bar=False ) rng_key = jax.random.fold_in(jax.random.PRNGKey(13), k) mcmc.run(rng_key, x=train_x_list[k], y=y_train) fit_list.append(mcmc) Explanation: 2.2 Training <a class="anchor" id="2.2"></a> Each model will be trained in the same way - with a Bernoulli likelihood and a logit link function. End of explanation def find_point_wise_loo_score(fit): return az.loo(az.from_numpyro(fit), pointwise=True, scale="log").loo_i.values lpd_point = np.vstack([find_point_wise_loo_score(fit) for fit in fit_list]).T exp_lpd_point = np.exp(lpd_point) Explanation: 2.3 Estimate leave-one-out cross-validated score for each training point <a class="anchor" id="2.3"></a> Rather than refitting each model 100 times, we will estimate the leave-one-out cross-validated score using LOO. End of explanation dist100_median = wells.loc[wells.index[train_id], "dist100"].median() logarsenic_median = wells.loc[wells.index[train_id], "logarsenic"].median() wells["dist100_l"] = (wells["dist100"] - dist100_median).clip(upper=0) wells["dist100_r"] = (wells["dist100"] - dist100_median).clip(lower=0) wells["logarsenic_l"] = (wells["logarsenic"] - logarsenic_median).clip(upper=0) wells["logarsenic_r"] = (wells["logarsenic"] - logarsenic_median).clip(lower=0) stacking_features = [ "edu0", "edu1", "edu2", "edu3", "assoc_half", "dist100_l", "dist100_r", "logarsenic_l", "logarsenic_r", ] X_stacking_train = wells.loc[train_id, stacking_features].to_numpy() X_stacking_test = wells.loc[test_id, stacking_features].to_numpy() Explanation: 3. Bayesian Hierarchical Stacking <a class="anchor" id="3"></a> 3.1 Prepare stacking datasets <a class="anchor" id="3.1"></a> To determine how the stacking weights should vary across training and test sets, we will need to create "stacking datasets" which include all the features which we want the stacking weights to depend on. How should such features be included? For discrete features, this is easy, we just one-hot-encode them. But for continuous features, we need a trick. In Equation (16), the authors recommend the following: if you have a continuous feature f, then replace it with the following two features: f_l: f minus the median of f, clipped above at 0; f_r: f minus the median of f, clipped below at 0; End of explanation def stacking( X, d_discrete, X_test, exp_lpd_point, tau_mu, tau_sigma, *, test, ): Get weights with which to stack candidate models' predictions. Parameters ---------- X Training stacking matrix: features on which stacking weights should depend, for the training set. d_discrete Number of discrete features in `X` and `X_test`. The first `d_discrete` features from these matrices should be the discrete ones, with the continuous ones coming after them. X_test Test stacking matrix: features on which stacking weights should depend, for the testing set. exp_lpd_point LOO score evaluated at each point in the training set, for each candidate model. tau_mu Hyperprior for mean of `beta`, for discrete features. tau_sigma Hyperprior for standard deviation of `beta`, for continuous features. test Whether to calculate stacking weights for test set. Notes ----- Naming of variables mirrors what's used in the original paper. N = X.shape[0] d = X.shape[1] N_test = X_test.shape[0] K = lpd_point.shape[1] # number of candidate models with numpyro.plate("Candidate models", K - 1, dim=-2): # mean effect of discrete features on stacking weights mu = numpyro.sample("mu", dist.Normal(0, tau_mu)) # standard deviation effect of discrete features on stacking weights sigma = numpyro.sample("sigma", dist.HalfNormal(scale=tau_sigma)) with numpyro.plate("Discrete features", d_discrete, dim=-1): # effect of discrete features on stacking weights tau = numpyro.sample("tau", dist.Normal(0, 1)) with numpyro.plate("Continuous features", d - d_discrete, dim=-1): # effect of continuous features on stacking weights beta_con = numpyro.sample("beta_con", dist.Normal(0, 1)) # effects of features on stacking weights beta = numpyro.deterministic( "beta", jnp.hstack([(sigma.squeeze() * tau.T + mu.squeeze()).T, beta_con]) ) assert beta.shape == (K - 1, d) # stacking weights (in unconstrained space) f = jnp.hstack([X @ beta.T, jnp.zeros((N, 1))]) assert f.shape == (N, K) # log probability of LOO training scores weighted by stacking weights. log_w = jax.nn.log_softmax(f, axis=1) # stacking weights (constrained to sum to 1) numpyro.deterministic("w", jnp.exp(log_w)) logp = jax.nn.logsumexp(lpd_point + log_w, axis=1) numpyro.factor("logp", jnp.sum(logp)) if test: # test set stacking weights (in unconstrained space) f_test = jnp.hstack([X_test @ beta.T, jnp.zeros((N_test, 1))]) # test set stacking weights (constrained to sum to 1) w_test = numpyro.deterministic("w_test", jax.nn.softmax(f_test, axis=1)) sampler = numpyro.infer.NUTS(stacking) mcmc = numpyro.infer.MCMC( sampler, num_chains=4, num_samples=1000, num_warmup=1000, progress_bar=False ) mcmc.run( jax.random.PRNGKey(17), X=X_stacking_train, d_discrete=4, X_test=X_stacking_test, exp_lpd_point=exp_lpd_point, tau_mu=1.0, tau_sigma=0.5, test=True, ) trace = mcmc.get_samples() Explanation: 3.2 Define stacking model <a class="anchor" id="3.2"></a> What we seek to find is a matrix of weights $W$ with which to multiply the models' predictions. Let's define a matrix $Pred$ such that $Pred_{i,k}$ represents the prediction made for point $i$ by model $k$. Then the final prediction for point $i$ will then be: $$ \sum_k W_{i, k}Pred_{i,k} $$ Such a matrix $W$ would be required to have each column sum to $1$. Hence, we calculate each row $W_i$ of $W$ as: $$ W_i = \text{softmax}(X_\text{stacking}_i \cdot \beta), $$ where $\beta$ is a matrix whose values we seek to determine. For the discrete features, $\beta$ is given a hierarchical structure over the possible inputs. Continuous features, on the other hand, get no hierarchical structure in this case study and just vary according to the input values. Notice how, for the discrete features, a non-centered parametrisation is used. Also note that we only need to estimate K-1 columns of $\beta$, because the weights W_{i, k} will have to sum to 1 for each i. End of explanation fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(16, 6), sharey=True) training_stacking_weights = trace["w"].mean(axis=0) sns.scatterplot(data=pd.DataFrame(training_stacking_weights), ax=ax[0]) fixed_weights = ( az.compare({idx: fit for idx, fit in enumerate(fit_list)}, method="stacking") .sort_index()["weight"] .to_numpy() ) fixed_weights_df = pd.DataFrame( np.repeat( fixed_weights[jnp.newaxis, :], len(X_stacking_train), axis=0, ) ) sns.scatterplot(data=fixed_weights_df, ax=ax[1]) ax[0].set_title("Training weights from Bayesian Hierarchical stacking") ax[1].set_title("Fixed weights stacking") ax[0].set_xlabel("Index") ax[1].set_xlabel("Index") fig.suptitle( "Bayesian Hierarchical Stacking weights can vary according to the input", fontsize=18, ) fig.tight_layout(); Explanation: We can now extract the weights with which to weight the different models from the posterior, and then visualise how they vary across the training set. Let's compare them with what the weights would've been if we'd just used fixed stacking weights (computed using ArviZ - see their docs for details). End of explanation # for each candidate model, extract the posterior predictive logits train_preds = [] for k in range(K): predictive = numpyro.infer.Predictive(logistic, fit_list[k].get_samples()) rng_key = jax.random.fold_in(jax.random.PRNGKey(19), k) train_pred = predictive(rng_key, x=train_x_list[k])["logits"] train_preds.append(train_pred.mean(axis=0)) # reshape, so we have (N, K) train_preds = np.vstack(train_preds).T # same as previous cell, but for test set test_preds = [] for k in range(K): predictive = numpyro.infer.Predictive(logistic, fit_list[k].get_samples()) rng_key = jax.random.fold_in(jax.random.PRNGKey(20), k) test_pred = predictive(rng_key, x=test_x_list[k])["logits"] test_preds.append(test_pred.mean(axis=0)) test_preds = np.vstack(test_preds).T # get the stacking weights for the test set test_stacking_weights = trace["w_test"].mean(axis=0) # get predictions using the stacking weights bhs_predictions = (test_stacking_weights * test_preds).sum(axis=1) # get predictions using only the model with the best LOO score model_selection_preds = test_preds[:, lpd_point.sum(axis=0).argmax()] # get predictions using fixed stacking weights, dependent on the LOO score fixed_weights_preds = (fixed_weights * test_preds).sum(axis=1) Explanation: 4. Evaluate on test set <a class="anchor" id="4"></a> 4.1 Stack predictions <a class="anchor" id="4.1"></a> Now, for each model, let's evaluate the log predictive density for each point in the test set. Once we have predictions for each model, we need to think about how to combine them, such that for each test point, we get a single prediction. We decided we'd do this in three ways: - Bayesian Hierarchical Stacking (bhs_pred); - choosing the model with the best training set LOO score (model_selection_preds); - fixed-weights stacking (fixed_weights_preds). End of explanation fig, ax = plt.subplots(figsize=(12, 6)) neg_log_pred_densities = np.vstack( [ -dist.Bernoulli(logits=bhs_predictions).log_prob(y_test), -dist.Bernoulli(logits=model_selection_preds).log_prob(y_test), -dist.Bernoulli(logits=fixed_weights_preds).log_prob(y_test), ] ).T neg_log_pred_density = pd.DataFrame( neg_log_pred_densities, columns=[ "Bayesian Hierarchical Stacking", "Model selection", "Fixed stacking weights", ], ) sns.barplot( data=neg_log_pred_density.reindex( columns=neg_log_pred_density.mean(axis=0).sort_values(ascending=False).index ), orient="h", ax=ax, ) ax.set_title( "Bayesian Hierarchical Stacking performs best here", fontdict={"fontsize": 18} ) ax.set_xlabel("Negative mean log predictive density (lower is better)"); Explanation: 4.2 Compare methods <a class="anchor" id="4.2"></a> Let's compare the negative log predictive density scores on the test set (note - lower is better): End of explanation <END_TASK>
15,941
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Paris Saclay Center for Data Science Titanic RAMP Step1: Exploratory data analysis Loading the data Step2: The original training data frame has 891 rows. In the starting kit, we give you a subset of 445 rows. Some passengers have missing information Step3: About two thirds of the passengers perished in the event. A dummy classifier that systematically returns "0" would have an accuracy of 62%, higher than that of a random model. Some plots Features densities and co-evolution A scatterplot matrix allows us to visualize Step4: Non-linearly transformed data The Fare variable has a very heavy tail. We can log-transform it. Step5: Plot the bivariate distributions and marginals of two variables Another way of visualizing relationships between variables is to plot their bivariate distributions. Step6: Making predictions A basic prediction workflow, using scikit-learn, will be presented below. First, we will perform some simple preprocessing of our data Step7: The preprocessor object created with make_column_transformer can be used in a scikit-learn pipeline. A pipeline assembles several steps together and can be used to cross validate an entire workflow. Generally, transformation steps are combined with a final estimator. We will create a pipeline consisting of the preprocessor created above and a final estimator, LogisticRegression. Step8: We can cross-validate our pipeline using cross_val_score. Below we will have specified cv=8 meaning KFold cross-valdiation splitting will be used, with 8 folds. The Area Under the Receiver Operating Characteristic Curve (ROC AUC) score is calculated for each split. The output score will be an array of 8 scores from each KFold. The score mean and standard of the 8 scores is printed at the end. Step9: Testing Once you have created a model with cross-valdiation scores you are happy with, you can test how well your model performs on the independent test data. First we will read in our test data Step10: Next we need to fit our pipeline on our training data Step11: Now we can predict on our test data Step12: Finally, we can calculate how well our model performed on the test data Step13: RAMP submissions For submitting to the RAMP site, you will need to write a submission.py file that defines a get_estimator function that returns a scikit-learn pipeline. For example, to submit our basic example above, we would define our pipeline within the function and return the pipeline at the end. Remember to include all the necessary imports at the beginning of the file. Step14: If you take a look at the sample submission in the directory submissions/starting_kit, you will find a file named submission.py, which has the above code in it. You can test that the sample submission works by running ramp_test_submission in your terminal (ensure that ramp-workflow has been installed and you are in the titanic ramp kit directory). Alternatively, within this notebook you can run
<ASSISTANT_TASK:> Python Code: %matplotlib inline import os import glob import numpy as np from scipy import io import matplotlib.pyplot as plt import pandas as pd Explanation: Paris Saclay Center for Data Science Titanic RAMP: survival prediction of Titanic passengers Benoit Playe (Institut Curie/Mines ParisTech), Chloé-Agathe Azencott (Institut Curie/Mines ParisTech), Alex Gramfort (LTCI/Télécom ParisTech), Balázs Kégl (LAL/CNRS) Introduction This is an initiation project to introduce RAMP and get you to know how it works. The goal is to develop prediction models able to identify people who survived from the sinking of the Titanic, based on gender, age, and ticketing information. The data we will manipulate is from the Titanic kaggle challenge. Requirements numpy>=1.10.0 matplotlib>=1.5.0 pandas>=0.19.0 scikit-learn>=0.17 (different syntaxes for v0.17 and v0.18) seaborn>=0.7.1 End of explanation train_filename = 'data/train.csv' data = pd.read_csv(train_filename) y_train = data['Survived'].values X_train = data.drop(['Survived', 'PassengerId'], axis=1) X_train.head(5) data.describe() data.count() Explanation: Exploratory data analysis Loading the data End of explanation data.groupby('Survived').count() Explanation: The original training data frame has 891 rows. In the starting kit, we give you a subset of 445 rows. Some passengers have missing information: in particular Age and Cabin info can be missing. The meaning of the columns is explained on the challenge website: Predicting survival The goal is to predict whether a passenger has survived from other known attributes. Let us group the data according to the Survived columns: End of explanation from pandas.plotting import scatter_matrix scatter_matrix(data.get(['Fare', 'Pclass', 'Age']), alpha=0.2, figsize=(8, 8), diagonal='kde'); Explanation: About two thirds of the passengers perished in the event. A dummy classifier that systematically returns "0" would have an accuracy of 62%, higher than that of a random model. Some plots Features densities and co-evolution A scatterplot matrix allows us to visualize: * on the diagonal, the density estimation for each feature * on each of the off-diagonal plots, a scatterplot between two features. Each dot represents an instance. End of explanation data_plot = data.get(['Age', 'Survived']) data_plot = data.assign(LogFare=lambda x : np.log(x.Fare + 10.)) scatter_matrix(data_plot.get(['Age', 'LogFare']), alpha=0.2, figsize=(8, 8), diagonal='kde'); data_plot.plot(kind='scatter', x='Age', y='LogFare', c='Survived', s=50, cmap=plt.cm.Paired); Explanation: Non-linearly transformed data The Fare variable has a very heavy tail. We can log-transform it. End of explanation import seaborn as sns sns.set() sns.set_style("whitegrid") sns.jointplot(data_plot.Age[data_plot.Survived == 1], data_plot.LogFare[data_plot.Survived == 1], kind="kde", size=7, space=0, color="b"); sns.jointplot(data_plot.Age[data_plot.Survived == 0], data_plot.LogFare[data_plot.Survived == 0], kind="kde", size=7, space=0, color="y"); Explanation: Plot the bivariate distributions and marginals of two variables Another way of visualizing relationships between variables is to plot their bivariate distributions. End of explanation from sklearn.compose import make_column_transformer from sklearn.preprocessing import OneHotEncoder from sklearn.impute import SimpleImputer categorical_cols = ['Sex', 'Pclass', 'Embarked'] numerical_cols = ['Age', 'SibSp', 'Parch', 'Fare'] preprocessor = make_column_transformer( (OneHotEncoder(handle_unknown='ignore'), categorical_cols), (SimpleImputer(strategy='constant', fill_value=-1), numerical_cols), ) Explanation: Making predictions A basic prediction workflow, using scikit-learn, will be presented below. First, we will perform some simple preprocessing of our data: one-hot encode the categorical features: Sex, Pclass, Embarked for the numerical columns Age, SibSp, Parch, Fare, fill in missing values with a default value (-1) all remaining columns will be dropped This can be done succintly with make_column_transformer which performs specific transformations on specific features. End of explanation from sklearn.pipeline import Pipeline from sklearn.linear_model import LogisticRegression pipeline = Pipeline([ ('transformer', preprocessor), ('classifier', LogisticRegression()), ]) Explanation: The preprocessor object created with make_column_transformer can be used in a scikit-learn pipeline. A pipeline assembles several steps together and can be used to cross validate an entire workflow. Generally, transformation steps are combined with a final estimator. We will create a pipeline consisting of the preprocessor created above and a final estimator, LogisticRegression. End of explanation from sklearn.model_selection import cross_val_score scores = cross_val_score(pipeline, X_train, y_train, cv=8, scoring='roc_auc') print("mean: %e (+/- %e)" % (scores.mean(), scores.std())) Explanation: We can cross-validate our pipeline using cross_val_score. Below we will have specified cv=8 meaning KFold cross-valdiation splitting will be used, with 8 folds. The Area Under the Receiver Operating Characteristic Curve (ROC AUC) score is calculated for each split. The output score will be an array of 8 scores from each KFold. The score mean and standard of the 8 scores is printed at the end. End of explanation # test_filename = 'data/test.csv' # data = pd.read_csv(test_filename) # y_test = data['Survived'].values # X_test = data.drop(['Survived', 'PassengerId'], axis=1) # X_test.head(5) Explanation: Testing Once you have created a model with cross-valdiation scores you are happy with, you can test how well your model performs on the independent test data. First we will read in our test data: End of explanation # clf = pipeline.fit(X_train, y_train) Explanation: Next we need to fit our pipeline on our training data: End of explanation # y_pred = pipeline.predict(X_test) Explanation: Now we can predict on our test data: End of explanation # from sklearn.metrics import roc_auc_score # score = roc_auc_score(y_test, y_pred) # score Explanation: Finally, we can calculate how well our model performed on the test data: End of explanation from sklearn.compose import make_column_transformer from sklearn.preprocessing import OneHotEncoder from sklearn.impute import SimpleImputer from sklearn.pipeline import Pipeline from sklearn.linear_model import LogisticRegression def get_estimator(): categorical_cols = ['Sex', 'Pclass', 'Embarked'] numerical_cols = ['Age', 'SibSp', 'Parch', 'Fare'] preprocessor = make_column_transformer( (OneHotEncoder(handle_unknown='ignore'), categorical_cols), (SimpleImputer(strategy='constant', fill_value=-1), numerical_cols), ) pipeline = Pipeline([ ('transformer', preprocessor), ('classifier', LogisticRegression()), ]) return pipeline Explanation: RAMP submissions For submitting to the RAMP site, you will need to write a submission.py file that defines a get_estimator function that returns a scikit-learn pipeline. For example, to submit our basic example above, we would define our pipeline within the function and return the pipeline at the end. Remember to include all the necessary imports at the beginning of the file. End of explanation # !ramp_test_submission Explanation: If you take a look at the sample submission in the directory submissions/starting_kit, you will find a file named submission.py, which has the above code in it. You can test that the sample submission works by running ramp_test_submission in your terminal (ensure that ramp-workflow has been installed and you are in the titanic ramp kit directory). Alternatively, within this notebook you can run: End of explanation <END_TASK>
15,942
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: QuTiP Lecture Step1: Introduction The quantum two-level system (TLS) is the simplest possible model for quantum light-matter interaction. In the version we simulate here, the system is driven by a continuous-mode coherent state, whose dipolar interaction with the system is represented by the following Hamiltonain $$ H =\hbar \omega_0 \sigma^\dagger \sigma + \frac{\hbar\Omega(t)}{2}\left( \sigma\textrm{e}^{-i\omega_dt} + \sigma^\dagger \textrm{e}^{i\omega_dt}\right),$$ where $\omega_0$ is the system's transition frequency, $\sigma$ is the system's atomic lowering operator, $\omega_d$ is the coherent state's center frequency, and $\Omega(t)$ is the coherent state's driving strength. The time-dependence can be removed to simplify the simulation by a rotating frame transformation, and is particularly simple when the driving field is resonant with the transition frequency ($\omega_d=\omega_0$). Then, $$ H_r =\frac{\hbar\Omega(t)}{2}\left( \sigma+ \sigma^\dagger \right).$$ Problem parameters We will explore emission from the two-level system under two different regimes Step2: Setup the operators, Hamiltonian, and initial state Step3: Calculate the emission flux We evolve the system with the Lindblad master equation solver, and we request that the expectation values of the number operator $\hat{n}=\sigma^{\dagger} \sigma$ are returned by the solver. If the probability of two photodetections were negligible over the course of the pulse, then $\langle \hat{n}(t) \rangle$ would be the probability density of a detection occuring on an ideal detector at time $t$. Step4: Visualize the emission flux We plot the emission flux from two different three-level systems. The system labelled 'exponential wavepacket' was excited with a short pulse, while the system labelled 'Gaussian wavepacket' was excited with a long pulse. Step5: Calculate the correlators involved in two-photon interference When two indistinguishable photons impinge on two ports of a beamsplitter at the same time, they exit together. Such a configuration, with intensity detectors at each output, is known as a Hong-Ou-Mandel interferometer. This interferometer measures two-photon interference in the pulse-wise cross-correlation of the detection records of its photon counters. If each of the two interfering wavepackets has unity average photon number, then the formula for this normalized cross-correlation is $$ g^{(2)}\textrm{HOM}[0] = \frac{1}{4}\left(g^{(2)}{aa}[0] + g^{(2)}_{bb}[0])\right) + \frac{1}{2}\left(1-\textrm{Re}\iint \mathop{\textrm{d} t} \mathop{\textrm{d} t'} \left[G^{(1)}_a(t,t')\right]^*G^{(1)}_b(t,t')\right) ,$$ where $a$ and $b$ label the wavepackets, $g^{(2)}{aa(bb)}[0]$ are the sources' measured degrees of second-order coherence, and $G^{(1)}{a(b)}(t,t')$ are the sources' first-order optical coherences. Specifically $$ G^{(1)}{a(b)}(t,t') = \gamma \langle \sigma{a(b)}^\dagger(t) \sigma_{a(b)}(t') \rangle,$$ which we will calculate in this notebook with the master equation solver and quantum regression theorem. Importantly, if $g^{(2)}\textrm{HOM}[0]=0$ then the photons always exit an output port of the beamsplitter together. Note Step6: Interpolate these functions, in preparation for time delays Step7: Calculate measured degrees of HOM cross-correlation Step8: Visualize the two-photon interference visibilities Here, the exponential wavepackets show good HOM interference visibility at zero delay, dipping down to almost zero (red). Meanwhile, in the case of the Gaussian wavepackets, the significant probability of re-excitation (due to $g^{(2)}_{aa}[0]=0.44$) partially destroys the HOM interference (blue); the HOM dip at zero delay still is below the nonclassical threshold (dashed line). However, if the Gaussian and exponential wavepackets, each generated by two-level systems, are interfered with one-another then the HOM interference is not below the nonclassical threshold (purple). This is a result of the re-excitation action that scrambles the phase of the first-order coherence. Step9: Versions
<ASSISTANT_TASK:> Python Code: %matplotlib inline %config InlineBackend.figure_format = 'retina' import matplotlib.pyplot as plt import numpy as np from scipy.interpolate import interp2d from qutip import * Explanation: QuTiP Lecture: Pulse-wise two-photon interference of emission from a two-level system K.A. Fischer, Stanford University This Jupyter notebook demonstrates how to simulate the two-photon interference between two potentially identical two-level systems, using QuTiP: The Quantum Toolbox in Python. The purpose is to help characterize the quality of a two-level system as a single-photon source and the capability of two such sources to interfere; ideal pulsed single-photon sources have both zero second-order coherence and complete first-order coherence, and the sources must match in spatio-temporal overlap to be indistinguishable. In this notebook, we will assume the spatial modes are comparable so we can focus on temporal effects. This notebook closely follows an example from my simulation paper, <a href="http://dx.doi.org/10.1088/1367-2630/18/11/113053">Dynamical modeling of pulsed two-photon interference</a>, published as New J. Phys. 18 113053 (2016). For more information about QuTiP see the project web page: http://qutip.org/ End of explanation # shared parameters gamma = 1 # decay rate tlist = np.linspace(0, 13, 300) taulist = tlist # parameters for TLS with exponential shape wavepacket (short pulse) tp_e = 0.060 # Gaussian pulse parameter Om_e = 19.40 # driving strength t_offset_e = 0.405 pulse_shape_e = Om_e/2 * np.exp(-(tlist - t_offset_e) ** 2 / (2 * tp_e ** 2)) # parameters for TLS with Gaussian shape wavepacket (long pulse) tp_G = 2.000 # Gaussian pulse parameter Om_G = 0.702 # driving strength t_offset_G = 5 pulse_shape_G = Om_G/2 * np.exp(-(tlist - t_offset_G) ** 2 / (2 * tp_G ** 2)) Explanation: Introduction The quantum two-level system (TLS) is the simplest possible model for quantum light-matter interaction. In the version we simulate here, the system is driven by a continuous-mode coherent state, whose dipolar interaction with the system is represented by the following Hamiltonain $$ H =\hbar \omega_0 \sigma^\dagger \sigma + \frac{\hbar\Omega(t)}{2}\left( \sigma\textrm{e}^{-i\omega_dt} + \sigma^\dagger \textrm{e}^{i\omega_dt}\right),$$ where $\omega_0$ is the system's transition frequency, $\sigma$ is the system's atomic lowering operator, $\omega_d$ is the coherent state's center frequency, and $\Omega(t)$ is the coherent state's driving strength. The time-dependence can be removed to simplify the simulation by a rotating frame transformation, and is particularly simple when the driving field is resonant with the transition frequency ($\omega_d=\omega_0$). Then, $$ H_r =\frac{\hbar\Omega(t)}{2}\left( \sigma+ \sigma^\dagger \right).$$ Problem parameters We will explore emission from the two-level system under two different regimes: under excitation by a short pulse which gives rise to an exponential wavepacket and under excitation by a long pulse which gives rise to a Gaussian wavepacket. (Short and long are relative to the spontaneous emission time of the atomic transition.) In both cases, the driving strengths are chosen such that the expected number of photodetections is unity, i.e. $$ \gamma\int \langle \sigma^\dagger (t) \sigma(t)\rangle=1 .$$ As a result, we can compare the statistics of the emission directly and the normalizations become trivial. Note, we use units where $\hbar=1$. End of explanation # initial state psi0 = fock(2, 0) # ground state # operators sm = destroy(2) # atomic lowering operator n = [sm.dag()*sm] # number operator # Hamiltonian H_I = sm + sm.dag() H_e = [[H_I, pulse_shape_e]] H_G = [[H_I, pulse_shape_G]] # collapse operator that describes dissipation c_ops = [np.sqrt(gamma) * sm] # represents spontaneous emission Explanation: Setup the operators, Hamiltonian, and initial state End of explanation n_e = mesolve(H_e, psi0, tlist, c_ops, n).expect[0] n_G = mesolve(H_G, psi0, tlist, c_ops, n).expect[0] Explanation: Calculate the emission flux We evolve the system with the Lindblad master equation solver, and we request that the expectation values of the number operator $\hat{n}=\sigma^{\dagger} \sigma$ are returned by the solver. If the probability of two photodetections were negligible over the course of the pulse, then $\langle \hat{n}(t) \rangle$ would be the probability density of a detection occuring on an ideal detector at time $t$. End of explanation fig, ax = plt.subplots(figsize=(8,5)) ax.plot(tlist, n_e, 'r', label="exponential wavepacket") ax.plot(tlist, n_G, 'b', label="Gaussian wavepacket") ax.legend() ax.set_xlim(0, 13) ax.set_ylim(0, 1) ax.set_xlabel('Time, $t$ [$1/\gamma$]') ax.set_ylabel('Emission flux [$\gamma$]') ax.set_title('TLS emission shapes'); Explanation: Visualize the emission flux We plot the emission flux from two different three-level systems. The system labelled 'exponential wavepacket' was excited with a short pulse, while the system labelled 'Gaussian wavepacket' was excited with a long pulse. End of explanation # specify relevant operators to calculate the correlation # <A(t+tau)B(t)> a_op = sm.dag() b_op = sm # calculate two-time correlations G1_t_tau_e = correlation_2op_2t(H_e, psi0, tlist, taulist, c_ops, a_op, b_op, reverse=True) G1_t_tau_e_r = correlation_2op_2t(H_e, psi0, tlist, taulist, c_ops, a_op, b_op) G1_t_tau_G = correlation_2op_2t(H_G, psi0, tlist, taulist, c_ops, a_op, b_op, reverse=True) G1_t_tau_G_r = correlation_2op_2t(H_G, psi0, tlist, taulist, c_ops, a_op, b_op) # g^(2)[0] values calculated for the sources in question in the # notebook 'Pulse-wise second-order optical coherences of emission # from a two-level system' g20_e = 0.03 g20_G = 0.44 Explanation: Calculate the correlators involved in two-photon interference When two indistinguishable photons impinge on two ports of a beamsplitter at the same time, they exit together. Such a configuration, with intensity detectors at each output, is known as a Hong-Ou-Mandel interferometer. This interferometer measures two-photon interference in the pulse-wise cross-correlation of the detection records of its photon counters. If each of the two interfering wavepackets has unity average photon number, then the formula for this normalized cross-correlation is $$ g^{(2)}\textrm{HOM}[0] = \frac{1}{4}\left(g^{(2)}{aa}[0] + g^{(2)}_{bb}[0])\right) + \frac{1}{2}\left(1-\textrm{Re}\iint \mathop{\textrm{d} t} \mathop{\textrm{d} t'} \left[G^{(1)}_a(t,t')\right]^*G^{(1)}_b(t,t')\right) ,$$ where $a$ and $b$ label the wavepackets, $g^{(2)}{aa(bb)}[0]$ are the sources' measured degrees of second-order coherence, and $G^{(1)}{a(b)}(t,t')$ are the sources' first-order optical coherences. Specifically $$ G^{(1)}{a(b)}(t,t') = \gamma \langle \sigma{a(b)}^\dagger(t) \sigma_{a(b)}(t') \rangle,$$ which we will calculate in this notebook with the master equation solver and quantum regression theorem. Importantly, if $g^{(2)}\textrm{HOM}[0]=0$ then the photons always exit an output port of the beamsplitter together. Note: for the sources in question $g^{(2)}{aa(bb)}[0]$ was calculated in the QuTiP example, <i>Pulse-wise second-order optical coherences of emission from a two-level system</i>. End of explanation t_delay_list = np.linspace(-5, 0, 50) TAULIST, TLIST = np.meshgrid(taulist, tlist) c1_e = interp2d(taulist, tlist, np.real(G1_t_tau_e)) c1_e_f = lambda tau, t: c1_e(tau, t) c2_e = interp2d(taulist, tlist, np.real(G1_t_tau_e_r)) c2_e_f = lambda tau, t: c2_e(tau, t) c1_G = interp2d(taulist, tlist, np.real(G1_t_tau_G)) c1_G_f = lambda tau, t: c1_G(tau, t) c2_G = interp2d(taulist, tlist, np.real(G1_t_tau_G_r)) c2_G_f = lambda tau, t: c2_G(tau, t) Explanation: Interpolate these functions, in preparation for time delays End of explanation # two delayed exponential wavepackets interfere def g2HOM_exponential(t_delay): corr_e = np.array( [[c1_e_f(tau, t)[0] * c2_e_f(tau, t - t_delay)[0] for tau in taulist] for t in tlist] ) return g20_e/2 + 1/2*abs(1 - 2 * np.trapz(np.trapz(corr_e, TLIST, axis=0), taulist) ) g2HOM_e = parallel_map(g2HOM_exponential, t_delay_list) # two delayed Gaussian wavepackets interfere def g2HOM_Gaussian(t_delay): corr_G = np.array( [[c1_G_f(tau, t)[0] * c2_G_f(tau, t - t_delay)[0] for tau in taulist] for t in tlist] ) return g20_G/2 + 1/2*abs(1 - 2 * np.trapz(np.trapz(corr_G, TLIST, axis=0), taulist) ) g2HOM_G = parallel_map(g2HOM_Gaussian, t_delay_list) # a delayed Gaussian wavepacket interferes with an exponential # wavepacket def g2HOM_Gaussian_exp(t_delay): corr_Ge = np.array( [[c1_e_f(tau, t)[0] * c2_G_f(tau, t - t_delay)[0] for tau in taulist] for t in tlist] ) return (g20_e + g20_G)/4 + 1/2*abs(1 - 2 * np.trapz(np.trapz(corr_Ge, TLIST, axis=0), taulist) ) g2HOM_Ge = parallel_map(g2HOM_Gaussian_exp, t_delay_list + 5.45) Explanation: Calculate measured degrees of HOM cross-correlation End of explanation fig, ax = plt.subplots(figsize=(8,5)) ax.plot(np.concatenate([t_delay_list, -t_delay_list[-2::-1]]), np.concatenate([g2HOM_e, g2HOM_e[-2::-1]]), 'r', label="Two exponential") ax.plot(np.concatenate([t_delay_list, -t_delay_list[-2::-1]]), np.concatenate([g2HOM_G, g2HOM_G[-2::-1]]), 'b', label="Two Gaussian") ax.plot(np.concatenate([t_delay_list, -t_delay_list[-2::-1]]), np.concatenate([g2HOM_Ge, g2HOM_Ge[-2::-1]]), 'purple', label="One exponential &\n one Gaussian") ax.legend(loc=3) ax.set_xlim(-5, 5) ax.axhline(y=0.5, color='k', linestyle='dashed') ax.set_xlabel('Wavepacket delay [$1/\gamma$]') ax.set_ylabel('$g^{(2)}_{HOM}[0]$') ax.set_title('Two-photon interference visibilities for '+ 'delayed wavepackets'); Explanation: Visualize the two-photon interference visibilities Here, the exponential wavepackets show good HOM interference visibility at zero delay, dipping down to almost zero (red). Meanwhile, in the case of the Gaussian wavepackets, the significant probability of re-excitation (due to $g^{(2)}_{aa}[0]=0.44$) partially destroys the HOM interference (blue); the HOM dip at zero delay still is below the nonclassical threshold (dashed line). However, if the Gaussian and exponential wavepackets, each generated by two-level systems, are interfered with one-another then the HOM interference is not below the nonclassical threshold (purple). This is a result of the re-excitation action that scrambles the phase of the first-order coherence. End of explanation from qutip.ipynbtools import version_table version_table() Explanation: Versions End of explanation <END_TASK>
15,943
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Comparing surrogate models Tim Head, July 2016. Step1: Bayesian optimization or sequential model-based optimization uses a surrogate model to model the expensive to evaluate function func. There are several choices for what kind of surrogate model to use. This example compares the performance of Step2: This shows the value of the two-dimensional branin function and the three minima. Objective The objective of this example is to find one of these minima in as few iterations as possible. One iteration is defined as one call to the branin function. We will evaluate each model several times using a different seed for the random number generator. Then compare the average performance of these models. This makes the comparison more robust against models that get "lucky". Step3: Note that this can take a few minutes.
<ASSISTANT_TASK:> Python Code: import numpy as np np.random.seed(123) %matplotlib inline import matplotlib.pyplot as plt plt.rcParams["figure.figsize"] = (10, 6) plt.set_cmap("viridis") Explanation: Comparing surrogate models Tim Head, July 2016. End of explanation from skopt.benchmarks import branin as _branin def branin(x, noise_level=0.): return _branin(x) + noise_level * np.random.randn() from matplotlib.colors import LogNorm def plot_branin(): fig, ax = plt.subplots() x1_values = np.linspace(-5, 10, 100) x2_values = np.linspace(0, 15, 100) x_ax, y_ax = np.meshgrid(x1_values, x2_values) vals = np.c_[x_ax.ravel(), y_ax.ravel()] fx = np.reshape([branin(val) for val in vals], (100, 100)) cm = ax.pcolormesh(x_ax, y_ax, fx, norm=LogNorm(vmin=fx.min(), vmax=fx.max())) minima = np.array([[-np.pi, 12.275], [+np.pi, 2.275], [9.42478, 2.475]]) ax.plot(minima[:, 0], minima[:, 1], "r.", markersize=14, lw=0, label="Minima") cb = fig.colorbar(cm) cb.set_label("f(x)") ax.legend(loc="best", numpoints=1) ax.set_xlabel("X1") ax.set_xlim([-5, 10]) ax.set_ylabel("X2") ax.set_ylim([0, 15]) plot_branin() Explanation: Bayesian optimization or sequential model-based optimization uses a surrogate model to model the expensive to evaluate function func. There are several choices for what kind of surrogate model to use. This example compares the performance of: gaussian processes, extra trees, and random forests as surrogate models. A purely random optimization strategy is used as a baseline. Toy model We will use the branin function as toy model for the expensive function. In a real world application this function would be unknown and expensive to evaluate. End of explanation from functools import partial from skopt import gp_minimize, forest_minimize, dummy_minimize func = partial(branin, noise_level=2.0) bounds = [(-5.0, 10.0), (0.0, 15.0)] x0 = [2.5, 7.5] n_calls = 80 def run(minimizer, n_iter=20): return [minimizer(func, bounds, x0=x0, n_calls=n_calls, random_state=n) for n in range(n_iter)] # Random search dummy_res = run(dummy_minimize) # Gaussian processes gp_res = run(gp_minimize) # Random forest rf_res = run(partial(forest_minimize, base_estimator="rf")) # Extra trees et_res = run(partial(forest_minimize, base_estimator="et")) Explanation: This shows the value of the two-dimensional branin function and the three minima. Objective The objective of this example is to find one of these minima in as few iterations as possible. One iteration is defined as one call to the branin function. We will evaluate each model several times using a different seed for the random number generator. Then compare the average performance of these models. This makes the comparison more robust against models that get "lucky". End of explanation from skopt.plots import plot_convergence plot_convergence(("dummy_minimize", dummy_res), ("gp_minimize", gp_res), ("forest_minimize('rf')", rf_res), ("forest_minimize('et)", et_res), true_minimum=0.397887, yscale="log") Explanation: Note that this can take a few minutes. End of explanation <END_TASK>
15,944
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: ES-DOC CMIP6 Model Properties - Land MIP Era Step1: Document Authors Set document authors Step2: Document Contributors Specify document contributors Step3: Document Publication Specify document publication status Step4: Document Table of Contents 1. Key Properties 2. Key Properties --&gt; Conservation Properties 3. Key Properties --&gt; Timestepping Framework 4. Key Properties --&gt; Software Properties 5. Grid 6. Grid --&gt; Horizontal 7. Grid --&gt; Vertical 8. Soil 9. Soil --&gt; Soil Map 10. Soil --&gt; Snow Free Albedo 11. Soil --&gt; Hydrology 12. Soil --&gt; Hydrology --&gt; Freezing 13. Soil --&gt; Hydrology --&gt; Drainage 14. Soil --&gt; Heat Treatment 15. Snow 16. Snow --&gt; Snow Albedo 17. Vegetation 18. Energy Balance 19. Carbon Cycle 20. Carbon Cycle --&gt; Vegetation 21. Carbon Cycle --&gt; Vegetation --&gt; Photosynthesis 22. Carbon Cycle --&gt; Vegetation --&gt; Autotrophic Respiration 23. Carbon Cycle --&gt; Vegetation --&gt; Allocation 24. Carbon Cycle --&gt; Vegetation --&gt; Phenology 25. Carbon Cycle --&gt; Vegetation --&gt; Mortality 26. Carbon Cycle --&gt; Litter 27. Carbon Cycle --&gt; Soil 28. Carbon Cycle --&gt; Permafrost Carbon 29. Nitrogen Cycle 30. River Routing 31. River Routing --&gt; Oceanic Discharge 32. Lakes 33. Lakes --&gt; Method 34. Lakes --&gt; Wetlands 1. Key Properties Land surface key properties 1.1. Model Overview Is Required Step5: 1.2. Model Name Is Required Step6: 1.3. Description Is Required Step7: 1.4. Land Atmosphere Flux Exchanges Is Required Step8: 1.5. Atmospheric Coupling Treatment Is Required Step9: 1.6. Land Cover Is Required Step10: 1.7. Land Cover Change Is Required Step11: 1.8. Tiling Is Required Step12: 2. Key Properties --&gt; Conservation Properties TODO 2.1. Energy Is Required Step13: 2.2. Water Is Required Step14: 2.3. Carbon Is Required Step15: 3. Key Properties --&gt; Timestepping Framework TODO 3.1. Timestep Dependent On Atmosphere Is Required Step16: 3.2. Time Step Is Required Step17: 3.3. Timestepping Method Is Required Step18: 4. Key Properties --&gt; Software Properties Software properties of land surface code 4.1. Repository Is Required Step19: 4.2. Code Version Is Required Step20: 4.3. Code Languages Is Required Step21: 5. Grid Land surface grid 5.1. Overview Is Required Step22: 6. Grid --&gt; Horizontal The horizontal grid in the land surface 6.1. Description Is Required Step23: 6.2. Matches Atmosphere Grid Is Required Step24: 7. Grid --&gt; Vertical The vertical grid in the soil 7.1. Description Is Required Step25: 7.2. Total Depth Is Required Step26: 8. Soil Land surface soil 8.1. Overview Is Required Step27: 8.2. Heat Water Coupling Is Required Step28: 8.3. Number Of Soil layers Is Required Step29: 8.4. Prognostic Variables Is Required Step30: 9. Soil --&gt; Soil Map Key properties of the land surface soil map 9.1. Description Is Required Step31: 9.2. Structure Is Required Step32: 9.3. Texture Is Required Step33: 9.4. Organic Matter Is Required Step34: 9.5. Albedo Is Required Step35: 9.6. Water Table Is Required Step36: 9.7. Continuously Varying Soil Depth Is Required Step37: 9.8. Soil Depth Is Required Step38: 10. Soil --&gt; Snow Free Albedo TODO 10.1. Prognostic Is Required Step39: 10.2. Functions Is Required Step40: 10.3. Direct Diffuse Is Required Step41: 10.4. Number Of Wavelength Bands Is Required Step42: 11. Soil --&gt; Hydrology Key properties of the land surface soil hydrology 11.1. Description Is Required Step43: 11.2. Time Step Is Required Step44: 11.3. Tiling Is Required Step45: 11.4. Vertical Discretisation Is Required Step46: 11.5. Number Of Ground Water Layers Is Required Step47: 11.6. Lateral Connectivity Is Required Step48: 11.7. Method Is Required Step49: 12. Soil --&gt; Hydrology --&gt; Freezing TODO 12.1. Number Of Ground Ice Layers Is Required Step50: 12.2. Ice Storage Method Is Required Step51: 12.3. Permafrost Is Required Step52: 13. Soil --&gt; Hydrology --&gt; Drainage TODO 13.1. Description Is Required Step53: 13.2. Types Is Required Step54: 14. Soil --&gt; Heat Treatment TODO 14.1. Description Is Required Step55: 14.2. Time Step Is Required Step56: 14.3. Tiling Is Required Step57: 14.4. Vertical Discretisation Is Required Step58: 14.5. Heat Storage Is Required Step59: 14.6. Processes Is Required Step60: 15. Snow Land surface snow 15.1. Overview Is Required Step61: 15.2. Tiling Is Required Step62: 15.3. Number Of Snow Layers Is Required Step63: 15.4. Density Is Required Step64: 15.5. Water Equivalent Is Required Step65: 15.6. Heat Content Is Required Step66: 15.7. Temperature Is Required Step67: 15.8. Liquid Water Content Is Required Step68: 15.9. Snow Cover Fractions Is Required Step69: 15.10. Processes Is Required Step70: 15.11. Prognostic Variables Is Required Step71: 16. Snow --&gt; Snow Albedo TODO 16.1. Type Is Required Step72: 16.2. Functions Is Required Step73: 17. Vegetation Land surface vegetation 17.1. Overview Is Required Step74: 17.2. Time Step Is Required Step75: 17.3. Dynamic Vegetation Is Required Step76: 17.4. Tiling Is Required Step77: 17.5. Vegetation Representation Is Required Step78: 17.6. Vegetation Types Is Required Step79: 17.7. Biome Types Is Required Step80: 17.8. Vegetation Time Variation Is Required Step81: 17.9. Vegetation Map Is Required Step82: 17.10. Interception Is Required Step83: 17.11. Phenology Is Required Step84: 17.12. Phenology Description Is Required Step85: 17.13. Leaf Area Index Is Required Step86: 17.14. Leaf Area Index Description Is Required Step87: 17.15. Biomass Is Required Step88: 17.16. Biomass Description Is Required Step89: 17.17. Biogeography Is Required Step90: 17.18. Biogeography Description Is Required Step91: 17.19. Stomatal Resistance Is Required Step92: 17.20. Stomatal Resistance Description Is Required Step93: 17.21. Prognostic Variables Is Required Step94: 18. Energy Balance Land surface energy balance 18.1. Overview Is Required Step95: 18.2. Tiling Is Required Step96: 18.3. Number Of Surface Temperatures Is Required Step97: 18.4. Evaporation Is Required Step98: 18.5. Processes Is Required Step99: 19. Carbon Cycle Land surface carbon cycle 19.1. Overview Is Required Step100: 19.2. Tiling Is Required Step101: 19.3. Time Step Is Required Step102: 19.4. Anthropogenic Carbon Is Required Step103: 19.5. Prognostic Variables Is Required Step104: 20. Carbon Cycle --&gt; Vegetation TODO 20.1. Number Of Carbon Pools Is Required Step105: 20.2. Carbon Pools Is Required Step106: 20.3. Forest Stand Dynamics Is Required Step107: 21. Carbon Cycle --&gt; Vegetation --&gt; Photosynthesis TODO 21.1. Method Is Required Step108: 22. Carbon Cycle --&gt; Vegetation --&gt; Autotrophic Respiration TODO 22.1. Maintainance Respiration Is Required Step109: 22.2. Growth Respiration Is Required Step110: 23. Carbon Cycle --&gt; Vegetation --&gt; Allocation TODO 23.1. Method Is Required Step111: 23.2. Allocation Bins Is Required Step112: 23.3. Allocation Fractions Is Required Step113: 24. Carbon Cycle --&gt; Vegetation --&gt; Phenology TODO 24.1. Method Is Required Step114: 25. Carbon Cycle --&gt; Vegetation --&gt; Mortality TODO 25.1. Method Is Required Step115: 26. Carbon Cycle --&gt; Litter TODO 26.1. Number Of Carbon Pools Is Required Step116: 26.2. Carbon Pools Is Required Step117: 26.3. Decomposition Is Required Step118: 26.4. Method Is Required Step119: 27. Carbon Cycle --&gt; Soil TODO 27.1. Number Of Carbon Pools Is Required Step120: 27.2. Carbon Pools Is Required Step121: 27.3. Decomposition Is Required Step122: 27.4. Method Is Required Step123: 28. Carbon Cycle --&gt; Permafrost Carbon TODO 28.1. Is Permafrost Included Is Required Step124: 28.2. Emitted Greenhouse Gases Is Required Step125: 28.3. Decomposition Is Required Step126: 28.4. Impact On Soil Properties Is Required Step127: 29. Nitrogen Cycle Land surface nitrogen cycle 29.1. Overview Is Required Step128: 29.2. Tiling Is Required Step129: 29.3. Time Step Is Required Step130: 29.4. Prognostic Variables Is Required Step131: 30. River Routing Land surface river routing 30.1. Overview Is Required Step132: 30.2. Tiling Is Required Step133: 30.3. Time Step Is Required Step134: 30.4. Grid Inherited From Land Surface Is Required Step135: 30.5. Grid Description Is Required Step136: 30.6. Number Of Reservoirs Is Required Step137: 30.7. Water Re Evaporation Is Required Step138: 30.8. Coupled To Atmosphere Is Required Step139: 30.9. Coupled To Land Is Required Step140: 30.10. Quantities Exchanged With Atmosphere Is Required Step141: 30.11. Basin Flow Direction Map Is Required Step142: 30.12. Flooding Is Required Step143: 30.13. Prognostic Variables Is Required Step144: 31. River Routing --&gt; Oceanic Discharge TODO 31.1. Discharge Type Is Required Step145: 31.2. Quantities Transported Is Required Step146: 32. Lakes Land surface lakes 32.1. Overview Is Required Step147: 32.2. Coupling With Rivers Is Required Step148: 32.3. Time Step Is Required Step149: 32.4. Quantities Exchanged With Rivers Is Required Step150: 32.5. Vertical Grid Is Required Step151: 32.6. Prognostic Variables Is Required Step152: 33. Lakes --&gt; Method TODO 33.1. Ice Treatment Is Required Step153: 33.2. Albedo Is Required Step154: 33.3. Dynamics Is Required Step155: 33.4. Dynamic Lake Extent Is Required Step156: 33.5. Endorheic Basins Is Required Step157: 34. Lakes --&gt; Wetlands TODO 34.1. Description Is Required
<ASSISTANT_TASK:> Python Code: # DO NOT EDIT ! from pyesdoc.ipython.model_topic import NotebookOutput # DO NOT EDIT ! DOC = NotebookOutput('cmip6', 'test-institute-3', 'sandbox-1', 'land') Explanation: ES-DOC CMIP6 Model Properties - Land MIP Era: CMIP6 Institute: TEST-INSTITUTE-3 Source ID: SANDBOX-1 Topic: Land Sub-Topics: Soil, Snow, Vegetation, Energy Balance, Carbon Cycle, Nitrogen Cycle, River Routing, Lakes. Properties: 154 (96 required) Model descriptions: Model description details Initialized From: -- Notebook Help: Goto notebook help page Notebook Initialised: 2018-02-15 16:54:46 Document Setup IMPORTANT: to be executed each time you run the notebook End of explanation # Set as follows: DOC.set_author("name", "email") # TODO - please enter value(s) Explanation: Document Authors Set document authors End of explanation # Set as follows: DOC.set_contributor("name", "email") # TODO - please enter value(s) Explanation: Document Contributors Specify document contributors End of explanation # Set publication status: # 0=do not publish, 1=publish. DOC.set_publication_status(0) Explanation: Document Publication Specify document publication status End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.key_properties.model_overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: Document Table of Contents 1. Key Properties 2. Key Properties --&gt; Conservation Properties 3. Key Properties --&gt; Timestepping Framework 4. Key Properties --&gt; Software Properties 5. Grid 6. Grid --&gt; Horizontal 7. Grid --&gt; Vertical 8. Soil 9. Soil --&gt; Soil Map 10. Soil --&gt; Snow Free Albedo 11. Soil --&gt; Hydrology 12. Soil --&gt; Hydrology --&gt; Freezing 13. Soil --&gt; Hydrology --&gt; Drainage 14. Soil --&gt; Heat Treatment 15. Snow 16. Snow --&gt; Snow Albedo 17. Vegetation 18. Energy Balance 19. Carbon Cycle 20. Carbon Cycle --&gt; Vegetation 21. Carbon Cycle --&gt; Vegetation --&gt; Photosynthesis 22. Carbon Cycle --&gt; Vegetation --&gt; Autotrophic Respiration 23. Carbon Cycle --&gt; Vegetation --&gt; Allocation 24. Carbon Cycle --&gt; Vegetation --&gt; Phenology 25. Carbon Cycle --&gt; Vegetation --&gt; Mortality 26. Carbon Cycle --&gt; Litter 27. Carbon Cycle --&gt; Soil 28. Carbon Cycle --&gt; Permafrost Carbon 29. Nitrogen Cycle 30. River Routing 31. River Routing --&gt; Oceanic Discharge 32. Lakes 33. Lakes --&gt; Method 34. Lakes --&gt; Wetlands 1. Key Properties Land surface key properties 1.1. Model Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of land surface model. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.key_properties.model_name') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 1.2. Model Name Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Name of land surface model code (e.g. MOSES2.2) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.key_properties.description') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 1.3. Description Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 General description of the processes modelled (e.g. dymanic vegation, prognostic albedo, etc.) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.key_properties.land_atmosphere_flux_exchanges') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "water" # "energy" # "carbon" # "nitrogen" # "phospherous" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 1.4. Land Atmosphere Flux Exchanges Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N Fluxes exchanged with the atmopshere. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.key_properties.atmospheric_coupling_treatment') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 1.5. Atmospheric Coupling Treatment Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe the treatment of land surface coupling with the Atmosphere model component, which may be different for different quantities (e.g. dust: semi-implicit, water vapour: explicit) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.key_properties.land_cover') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "bare soil" # "urban" # "lake" # "land ice" # "lake ice" # "vegetated" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 1.6. Land Cover Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N Types of land cover defined in the land surface model End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.key_properties.land_cover_change') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 1.7. Land Cover Change Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe how land cover change is managed (e.g. the use of net or gross transitions) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.key_properties.tiling') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 1.8. Tiling Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe the general tiling procedure used in the land surface (if any). Include treatment of physiography, land/sea, (dynamic) vegetation coverage and orography/roughness End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.key_properties.conservation_properties.energy') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 2. Key Properties --&gt; Conservation Properties TODO 2.1. Energy Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe if/how energy is conserved globally and to what level (e.g. within X [units]/year) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.key_properties.conservation_properties.water') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 2.2. Water Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe if/how water is conserved globally and to what level (e.g. within X [units]/year) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.key_properties.conservation_properties.carbon') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 2.3. Carbon Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe if/how carbon is conserved globally and to what level (e.g. within X [units]/year) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.key_properties.timestepping_framework.timestep_dependent_on_atmosphere') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 3. Key Properties --&gt; Timestepping Framework TODO 3.1. Timestep Dependent On Atmosphere Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is a time step dependent on the frequency of atmosphere coupling? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.key_properties.timestepping_framework.time_step') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 3.2. Time Step Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overall timestep of land surface model (i.e. time between calls) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.key_properties.timestepping_framework.timestepping_method') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 3.3. Timestepping Method Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 General description of time stepping method and associated time step(s) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.key_properties.software_properties.repository') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 4. Key Properties --&gt; Software Properties Software properties of land surface code 4.1. Repository Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Location of code for this component. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.key_properties.software_properties.code_version') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 4.2. Code Version Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Code version identifier. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.key_properties.software_properties.code_languages') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 4.3. Code Languages Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N Code language(s). End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.grid.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 5. Grid Land surface grid 5.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of the grid in the land surface End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.grid.horizontal.description') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 6. Grid --&gt; Horizontal The horizontal grid in the land surface 6.1. Description Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe the general structure of the horizontal grid (not including any tiling) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.grid.horizontal.matches_atmosphere_grid') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 6.2. Matches Atmosphere Grid Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Does the horizontal grid match the atmosphere? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.grid.vertical.description') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 7. Grid --&gt; Vertical The vertical grid in the soil 7.1. Description Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe the general structure of the vertical grid in the soil (not including any tiling) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.grid.vertical.total_depth') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 7.2. Total Depth Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 The total depth of the soil (in metres) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 8. Soil Land surface soil 8.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of soil in the land surface End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.heat_water_coupling') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 8.2. Heat Water Coupling Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe the coupling between heat and water in the soil End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.number_of_soil layers') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 8.3. Number Of Soil layers Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 The number of soil layers End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.prognostic_variables') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 8.4. Prognostic Variables Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 List the prognostic variables of the soil scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.soil_map.description') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 9. Soil --&gt; Soil Map Key properties of the land surface soil map 9.1. Description Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 General description of soil map End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.soil_map.structure') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 9.2. Structure Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the soil structure map End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.soil_map.texture') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 9.3. Texture Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the soil texture map End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.soil_map.organic_matter') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 9.4. Organic Matter Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the soil organic matter map End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.soil_map.albedo') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 9.5. Albedo Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the soil albedo map End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.soil_map.water_table') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 9.6. Water Table Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the soil water table map, if any End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.soil_map.continuously_varying_soil_depth') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 9.7. Continuously Varying Soil Depth Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Does the soil properties vary continuously with depth? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.soil_map.soil_depth') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 9.8. Soil Depth Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the soil depth map End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.snow_free_albedo.prognostic') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 10. Soil --&gt; Snow Free Albedo TODO 10.1. Prognostic Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is snow free albedo prognostic? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.snow_free_albedo.functions') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "vegetation type" # "soil humidity" # "vegetation state" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 10.2. Functions Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N If prognostic, describe the dependancies on snow free albedo calculations End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.snow_free_albedo.direct_diffuse') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "distinction between direct and diffuse albedo" # "no distinction between direct and diffuse albedo" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 10.3. Direct Diffuse Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 If prognostic, describe the distinction between direct and diffuse albedo End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.snow_free_albedo.number_of_wavelength_bands') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 10.4. Number Of Wavelength Bands Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 If prognostic, enter the number of wavelength bands used End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.hydrology.description') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 11. Soil --&gt; Hydrology Key properties of the land surface soil hydrology 11.1. Description Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 General description of the soil hydrological model End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.hydrology.time_step') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 11.2. Time Step Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Time step of river soil hydrology in seconds End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.hydrology.tiling') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 11.3. Tiling Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the soil hydrology tiling, if any. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.hydrology.vertical_discretisation') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 11.4. Vertical Discretisation Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe the typical vertical discretisation End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.hydrology.number_of_ground_water_layers') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 11.5. Number Of Ground Water Layers Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 The number of soil layers that may contain water End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.hydrology.lateral_connectivity') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "perfect connectivity" # "Darcian flow" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 11.6. Lateral Connectivity Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N Describe the lateral connectivity between tiles End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.hydrology.method') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Bucket" # "Force-restore" # "Choisnel" # "Explicit diffusion" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 11.7. Method Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 The hydrological dynamics scheme in the land surface model End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.hydrology.freezing.number_of_ground_ice_layers') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 12. Soil --&gt; Hydrology --&gt; Freezing TODO 12.1. Number Of Ground Ice Layers Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 How many soil layers may contain ground ice End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.hydrology.freezing.ice_storage_method') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 12.2. Ice Storage Method Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe the method of ice storage End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.hydrology.freezing.permafrost') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 12.3. Permafrost Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe the treatment of permafrost, if any, within the land surface scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.hydrology.drainage.description') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 13. Soil --&gt; Hydrology --&gt; Drainage TODO 13.1. Description Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 General describe how drainage is included in the land surface scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.hydrology.drainage.types') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "Gravity drainage" # "Horton mechanism" # "topmodel-based" # "Dunne mechanism" # "Lateral subsurface flow" # "Baseflow from groundwater" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 13.2. Types Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N Different types of runoff represented by the land surface model End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.heat_treatment.description') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 14. Soil --&gt; Heat Treatment TODO 14.1. Description Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 General description of how heat treatment properties are defined End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.heat_treatment.time_step') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 14.2. Time Step Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Time step of soil heat scheme in seconds End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.heat_treatment.tiling') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 14.3. Tiling Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the soil heat treatment tiling, if any. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.heat_treatment.vertical_discretisation') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 14.4. Vertical Discretisation Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe the typical vertical discretisation End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.heat_treatment.heat_storage') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Force-restore" # "Explicit diffusion" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 14.5. Heat Storage Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Specify the method of heat storage End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.soil.heat_treatment.processes') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "soil moisture freeze-thaw" # "coupling with snow temperature" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 14.6. Processes Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N Describe processes included in the treatment of soil heat End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.snow.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 15. Snow Land surface snow 15.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of snow in the land surface End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.snow.tiling') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 15.2. Tiling Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the snow tiling, if any. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.snow.number_of_snow_layers') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 15.3. Number Of Snow Layers Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 The number of snow levels used in the land surface scheme/model End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.snow.density') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "prognostic" # "constant" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 15.4. Density Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Description of the treatment of snow density End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.snow.water_equivalent') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "prognostic" # "diagnostic" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 15.5. Water Equivalent Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Description of the treatment of the snow water equivalent End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.snow.heat_content') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "prognostic" # "diagnostic" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 15.6. Heat Content Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Description of the treatment of the heat content of snow End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.snow.temperature') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "prognostic" # "diagnostic" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 15.7. Temperature Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Description of the treatment of snow temperature End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.snow.liquid_water_content') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "prognostic" # "diagnostic" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 15.8. Liquid Water Content Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Description of the treatment of snow liquid water End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.snow.snow_cover_fractions') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "ground snow fraction" # "vegetation snow fraction" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 15.9. Snow Cover Fractions Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N Specify cover fractions used in the surface snow scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.snow.processes') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "snow interception" # "snow melting" # "snow freezing" # "blowing snow" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 15.10. Processes Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N Snow related processes in the land surface scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.snow.prognostic_variables') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 15.11. Prognostic Variables Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 List the prognostic variables of the snow scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.snow.snow_albedo.type') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "prognostic" # "prescribed" # "constant" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 16. Snow --&gt; Snow Albedo TODO 16.1. Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe the treatment of snow-covered land albedo End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.snow.snow_albedo.functions') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "vegetation type" # "snow age" # "snow density" # "snow grain type" # "aerosol deposition" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 16.2. Functions Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N *If prognostic, * End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 17. Vegetation Land surface vegetation 17.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of vegetation in the land surface End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.time_step') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 17.2. Time Step Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Time step of vegetation scheme in seconds End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.dynamic_vegetation') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 17.3. Dynamic Vegetation Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is there dynamic evolution of vegetation? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.tiling') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 17.4. Tiling Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the vegetation tiling, if any. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.vegetation_representation') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "vegetation types" # "biome types" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 17.5. Vegetation Representation Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Vegetation classification used End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.vegetation_types') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "broadleaf tree" # "needleleaf tree" # "C3 grass" # "C4 grass" # "vegetated" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 17.6. Vegetation Types Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N List of vegetation types in the classification, if any End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.biome_types') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "evergreen needleleaf forest" # "evergreen broadleaf forest" # "deciduous needleleaf forest" # "deciduous broadleaf forest" # "mixed forest" # "woodland" # "wooded grassland" # "closed shrubland" # "opne shrubland" # "grassland" # "cropland" # "wetlands" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 17.7. Biome Types Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N List of biome types in the classification, if any End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.vegetation_time_variation') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "fixed (not varying)" # "prescribed (varying from files)" # "dynamical (varying from simulation)" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 17.8. Vegetation Time Variation Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 How the vegetation fractions in each tile are varying with time End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.vegetation_map') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 17.9. Vegetation Map Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 If vegetation fractions are not dynamically updated , describe the vegetation map used (common name and reference, if possible) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.interception') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 17.10. Interception Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is vegetation interception of rainwater represented? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.phenology') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "prognostic" # "diagnostic (vegetation map)" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 17.11. Phenology Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Treatment of vegetation phenology End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.phenology_description') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 17.12. Phenology Description Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 General description of the treatment of vegetation phenology End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.leaf_area_index') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "prescribed" # "prognostic" # "diagnostic" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 17.13. Leaf Area Index Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Treatment of vegetation leaf area index End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.leaf_area_index_description') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 17.14. Leaf Area Index Description Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 General description of the treatment of leaf area index End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.biomass') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "prognostic" # "diagnostic" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 17.15. Biomass Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 *Treatment of vegetation biomass * End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.biomass_description') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 17.16. Biomass Description Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 General description of the treatment of vegetation biomass End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.biogeography') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "prognostic" # "diagnostic" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 17.17. Biogeography Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Treatment of vegetation biogeography End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.biogeography_description') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 17.18. Biogeography Description Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 General description of the treatment of vegetation biogeography End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.stomatal_resistance') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "light" # "temperature" # "water availability" # "CO2" # "O3" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 17.19. Stomatal Resistance Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N Specify what the vegetation stomatal resistance depends on End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.stomatal_resistance_description') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 17.20. Stomatal Resistance Description Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 General description of the treatment of vegetation stomatal resistance End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.vegetation.prognostic_variables') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 17.21. Prognostic Variables Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 List the prognostic variables of the vegetation scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.energy_balance.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 18. Energy Balance Land surface energy balance 18.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of energy balance in land surface End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.energy_balance.tiling') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 18.2. Tiling Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the energy balance tiling, if any. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.energy_balance.number_of_surface_temperatures') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 18.3. Number Of Surface Temperatures Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 The maximum number of distinct surface temperatures in a grid cell (for example, each subgrid tile may have its own temperature) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.energy_balance.evaporation') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "alpha" # "beta" # "combined" # "Monteith potential evaporation" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 18.4. Evaporation Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N Specify the formulation method for land surface evaporation, from soil and vegetation End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.energy_balance.processes') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "transpiration" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 18.5. Processes Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N Describe which processes are included in the energy balance scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 19. Carbon Cycle Land surface carbon cycle 19.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of carbon cycle in land surface End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.tiling') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 19.2. Tiling Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the carbon cycle tiling, if any. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.time_step') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 19.3. Time Step Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Time step of carbon cycle in seconds End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.anthropogenic_carbon') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "grand slam protocol" # "residence time" # "decay time" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 19.4. Anthropogenic Carbon Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N Describe the treament of the anthropogenic carbon pool End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.prognostic_variables') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 19.5. Prognostic Variables Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 List the prognostic variables of the carbon scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.vegetation.number_of_carbon_pools') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 20. Carbon Cycle --&gt; Vegetation TODO 20.1. Number Of Carbon Pools Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Enter the number of carbon pools used End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.vegetation.carbon_pools') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 20.2. Carbon Pools Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 List the carbon pools used End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.vegetation.forest_stand_dynamics') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 20.3. Forest Stand Dynamics Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the treatment of forest stand dyanmics End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.vegetation.photosynthesis.method') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 21. Carbon Cycle --&gt; Vegetation --&gt; Photosynthesis TODO 21.1. Method Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the general method used for photosynthesis (e.g. type of photosynthesis, distinction between C3 and C4 grasses, Nitrogen depencence, etc.) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.vegetation.autotrophic_respiration.maintainance_respiration') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 22. Carbon Cycle --&gt; Vegetation --&gt; Autotrophic Respiration TODO 22.1. Maintainance Respiration Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the general method used for maintainence respiration End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.vegetation.autotrophic_respiration.growth_respiration') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 22.2. Growth Respiration Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the general method used for growth respiration End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.vegetation.allocation.method') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 23. Carbon Cycle --&gt; Vegetation --&gt; Allocation TODO 23.1. Method Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe the general principle behind the allocation scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.vegetation.allocation.allocation_bins') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "leaves + stems + roots" # "leaves + stems + roots (leafy + woody)" # "leaves + fine roots + coarse roots + stems" # "whole plant (no distinction)" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 23.2. Allocation Bins Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Specify distinct carbon bins used in allocation End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.vegetation.allocation.allocation_fractions') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "fixed" # "function of vegetation type" # "function of plant allometry" # "explicitly calculated" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 23.3. Allocation Fractions Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe how the fractions of allocation are calculated End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.vegetation.phenology.method') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 24. Carbon Cycle --&gt; Vegetation --&gt; Phenology TODO 24.1. Method Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe the general principle behind the phenology scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.vegetation.mortality.method') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 25. Carbon Cycle --&gt; Vegetation --&gt; Mortality TODO 25.1. Method Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe the general principle behind the mortality scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.litter.number_of_carbon_pools') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 26. Carbon Cycle --&gt; Litter TODO 26.1. Number Of Carbon Pools Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Enter the number of carbon pools used End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.litter.carbon_pools') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 26.2. Carbon Pools Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 List the carbon pools used End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.litter.decomposition') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 26.3. Decomposition Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 List the decomposition methods used End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.litter.method') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 26.4. Method Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 List the general method used End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.soil.number_of_carbon_pools') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 27. Carbon Cycle --&gt; Soil TODO 27.1. Number Of Carbon Pools Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Enter the number of carbon pools used End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.soil.carbon_pools') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 27.2. Carbon Pools Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 List the carbon pools used End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.soil.decomposition') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 27.3. Decomposition Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 List the decomposition methods used End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.soil.method') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 27.4. Method Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 List the general method used End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.permafrost_carbon.is_permafrost_included') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 28. Carbon Cycle --&gt; Permafrost Carbon TODO 28.1. Is Permafrost Included Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is permafrost included? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.permafrost_carbon.emitted_greenhouse_gases') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 28.2. Emitted Greenhouse Gases Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 List the GHGs emitted End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.permafrost_carbon.decomposition') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 28.3. Decomposition Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 List the decomposition methods used End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.carbon_cycle.permafrost_carbon.impact_on_soil_properties') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 28.4. Impact On Soil Properties Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the impact of permafrost on soil properties End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.nitrogen_cycle.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 29. Nitrogen Cycle Land surface nitrogen cycle 29.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of the nitrogen cycle in the land surface End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.nitrogen_cycle.tiling') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 29.2. Tiling Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the notrogen cycle tiling, if any. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.nitrogen_cycle.time_step') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 29.3. Time Step Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Time step of nitrogen cycle in seconds End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.nitrogen_cycle.prognostic_variables') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 29.4. Prognostic Variables Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 List the prognostic variables of the nitrogen scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.river_routing.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 30. River Routing Land surface river routing 30.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of river routing in the land surface End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.river_routing.tiling') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 30.2. Tiling Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the river routing, if any. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.river_routing.time_step') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 30.3. Time Step Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Time step of river routing scheme in seconds End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.river_routing.grid_inherited_from_land_surface') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 30.4. Grid Inherited From Land Surface Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is the grid inherited from land surface? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.river_routing.grid_description') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 30.5. Grid Description Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 General description of grid, if not inherited from land surface End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.river_routing.number_of_reservoirs') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 30.6. Number Of Reservoirs Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Enter the number of reservoirs End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.river_routing.water_re_evaporation') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "flood plains" # "irrigation" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 30.7. Water Re Evaporation Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N TODO End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.river_routing.coupled_to_atmosphere') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 30.8. Coupled To Atmosphere Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Is river routing coupled to the atmosphere model component? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.river_routing.coupled_to_land') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 30.9. Coupled To Land Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the coupling between land and rivers End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.river_routing.quantities_exchanged_with_atmosphere') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "heat" # "water" # "tracers" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 30.10. Quantities Exchanged With Atmosphere Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N If couple to atmosphere, which quantities are exchanged between river routing and the atmosphere model components? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.river_routing.basin_flow_direction_map') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "present day" # "adapted for other periods" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 30.11. Basin Flow Direction Map Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 What type of basin flow direction map is being used? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.river_routing.flooding') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 30.12. Flooding Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the representation of flooding, if any End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.river_routing.prognostic_variables') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 30.13. Prognostic Variables Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 List the prognostic variables of the river routing End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.river_routing.oceanic_discharge.discharge_type') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "direct (large rivers)" # "diffuse" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 31. River Routing --&gt; Oceanic Discharge TODO 31.1. Discharge Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Specify how rivers are discharged to the ocean End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.river_routing.oceanic_discharge.quantities_transported') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "heat" # "water" # "tracers" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 31.2. Quantities Transported Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N Quantities that are exchanged from river-routing to the ocean model component End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.lakes.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 32. Lakes Land surface lakes 32.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of lakes in the land surface End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.lakes.coupling_with_rivers') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 32.2. Coupling With Rivers Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Are lakes coupled to the river routing model component? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.lakes.time_step') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 32.3. Time Step Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Time step of lake scheme in seconds End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.lakes.quantities_exchanged_with_rivers') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "heat" # "water" # "tracers" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 32.4. Quantities Exchanged With Rivers Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N If coupling with rivers, which quantities are exchanged between the lakes and rivers End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.lakes.vertical_grid') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 32.5. Vertical Grid Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the vertical grid of lakes End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.lakes.prognostic_variables') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 32.6. Prognostic Variables Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 List the prognostic variables of the lake scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.lakes.method.ice_treatment') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 33. Lakes --&gt; Method TODO 33.1. Ice Treatment Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is lake ice included? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.lakes.method.albedo') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "prognostic" # "diagnostic" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 33.2. Albedo Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Describe the treatment of lake albedo End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.lakes.method.dynamics') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "No lake dynamics" # "vertical" # "horizontal" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 33.3. Dynamics Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N Which dynamics of lakes are treated? horizontal, vertical, etc. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.lakes.method.dynamic_lake_extent') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 33.4. Dynamic Lake Extent Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is a dynamic lake extent scheme included? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.lakes.method.endorheic_basins') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 33.5. Endorheic Basins Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Basins not flowing to ocean included? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.land.lakes.wetlands.description') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 34. Lakes --&gt; Wetlands TODO 34.1. Description Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Describe the treatment of wetlands, if any End of explanation <END_TASK>
15,945
<SYSTEM_TASK:> Given the following text problem statement, write Python code to implement the functionality described below in problem statement <END_TASK> <USER_TASK:> Problem: I'm using tensorflow 2.10.0.
<ASSISTANT_TASK:> Python Code: import tensorflow as tf import numpy as np np.random.seed(10) a = tf.constant(np.random.rand(50, 100, 512)) def g(a): return tf.expand_dims(a, 2) result = g(a.__copy__()) <END_TASK>
15,946
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Collation outputs Introduction In practice Table Step1: Table Step2: Before moving to the other outputs, try to produce the simple HTML output by changing the code above. The value required in the output keyword should be html. Table Step3: Table Step4: And, finally, you can test the XML/TEI output that produces XML following the TEI parallel segmentation encoding guidelines. Step5: Graph
<ASSISTANT_TASK:> Python Code: #import the collatex library from collatex import * #create an instance of the collateX engine collation = Collation() #add witnesses to the collateX instance collation.add_plain_witness( "A", "The quick brown fox jumped over the lazy dog.") collation.add_plain_witness( "B", "The brown fox jumped over the dog." ) collation.add_plain_witness( "C", "The bad fox jumped over the lazy dog." ) #collate the witnesses and store the result in a vaiable called 'table' #as we have not specified an output this will be sored in plain text table = collate(collation) #print the collation result print(table) Explanation: Collation outputs Introduction In practice Table: HTML Table: JSON Table: XML and XML/TEI Graph: SVG Exercise What's next Introduction In this tutorial we will be trying different outputs for our collation, meaning different graphical representations, formats and visualizations of the result. The visualization of the collation result is an open discussion: several possibilities have been suggested and used and new ones are always being proposed. When the output of the collation is a printed format, such as a book, it is rare to see anything different from the traditional critical apparatus. Now that output formats are more frequently digital (or at least have a digital component), collation tools tend to offer more than one visualization option. This is the case for both Juxta and CollateX. The different visualizations are not incompatible; on the contrary, they can be complementary, highlighting different aspects of the result and suitable for different users or different stages of the workflow. In the previous tutorials we used the alignment table and the graph. The alignment table, in use since the 1960's, is the equivalent of the matrix of bioinformatic for sequence alignment (for example, strings of DNA). In contrast, the graph is meant to represent the fluidity of the text and its variation. The idea of a graph-oriented model for expressing textual variance has been originally developed by Desmond Schmidt (2008). You can refer to this video, for a presentation on Apparatus vs. Graph – an Interface as Scholarly Argument by Tara Andrews and Joris van Zundert. Other outputs, such as the histogram and the side-by-side visualization offered by Juxta, allow users to visualize the result of the comparison between two witnesses only. This reflects the way the algorithm is built and shows that the graphical representation is connected with the approach to collation that informs the software. CollateX has two main ways to conceive of the collation result: as a table (with many different formatting options) and as a graph: - table formats - plain text table (no need to specify the output) - HTML table (output='html') - HTML vertical table with colors (output='html2') - JSON (output='json') - XML (output='xml') - XML/TEI (output='tei') - graph format - SVG (output='svg') In practice Even though we have already encountered some of these outputs, it is worth going through them one more time focussing on part of the code that needs to change to produce the different formats. Table: plain text In this tutorial we will use some simple texts already used in the previous tutorial: the fox and dog example. Let's start with the most simple output, for which we don't need to specify any output format (note that you can name the variable containing the output anything you like, but in this tutorial we call it alignment_table, table or graph) In the code cell below the lines starting with a hash (#) are comments and are not executed. They are there in this instance to help you remember what the different parts of the code do. You do not need to use them in your notebook (although sometimes it is helpful to add comments to your code so you remember what things do). End of explanation table = collate(collation, output='html2') Explanation: Table: HTML Now let's try a different output. This time we still want a table format but instead of it being in plain text we would like it exported in HTML (the markup language used for web pages), and we would like it to be displayed vertically with nice colors to highlight the comparison. To achieve this all you need to do is add the keyword output to the collate command and give it that value html2. End of explanation table = collate(collation, output='json') print(table) Explanation: Before moving to the other outputs, try to produce the simple HTML output by changing the code above. The value required in the output keyword should be html. Table: JSON The same alignment table can be exported in a variety of formats, as we have seen, including JSON (Javascript Object Notation), a format widely used for storing and interchanging data nowadays. In order to produce JSON as output, we need to specify json as the output format. End of explanation table = collate(collation, output='xml') print(table) Explanation: Table: XML and XML/TEI We can use the same procedure in order to export the table in XML or XML/TEI (the latter produces a condensed version of the table only listing witnesses at points of divergence - also called a negative apparatus). To do this you just specify a different output format. Let's start with the XML output (that you can later post-process using XSLT or other tools). End of explanation table = collate(collation, output='tei') print(table) Explanation: And, finally, you can test the XML/TEI output that produces XML following the TEI parallel segmentation encoding guidelines. End of explanation graph = collate(collation, output='svg') Explanation: Graph: SVG And now for something different: try with the graph, exported in the SVG format End of explanation <END_TASK>
15,947
<SYSTEM_TASK:> Given the following text problem statement, write Python code to implement the functionality described below in problem statement <END_TASK> <USER_TASK:> Problem: How do I find all rows in a pandas DataFrame which have the max value for count column, after grouping by ['Sp','Value'] columns?
<ASSISTANT_TASK:> Python Code: import pandas as pd df = pd.DataFrame({'Sp':['MM1','MM1','MM1','MM2','MM2','MM2','MM4','MM4','MM4'], 'Value':['S1','S1','S3','S3','S4','S4','S2','S2','S2'], 'Mt':['a','n','cb','mk','bg','dgd','rd','cb','uyi'], 'count':[3,2,5,8,10,1,2,2,7]}) def g(df): return df[df.groupby(['Sp', 'Value'])['count'].transform(max) == df['count']] result = g(df.copy()) <END_TASK>
15,948
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Study of Correlation Between Building Demolition and Associated Features Capstone Project for Data Science at Scale on Coursera Repo is located here Chen Yang yangcnju@gmail.com Step1: Objective Build a model to make predictions on blighted buildings based on real data from data.detroitmi.gov as given by coursera. Building demolition is very important for the city to turn around and revive its economy. However, it's no easy task. Accurate predictions can provide guidance on potential blighted buildings and help avoid complications at early stages. Building List The buildings were defined as described below Step2: Features Three kinds (311-calls, blight-violations, and crimes) of incident counts and coordinates (normalized) was used in the end. I also tried to generate more features by differentiating each kind of crimes or each kind of violations in this notebook. However, these differentiated features lead to smaller AUC scores. Data The buildings were down-sampled to contain same number of blighted buildings and non-blighted ones. The ratio between train and test was set at a ratio of 80 Step3: This model resulted in an AUC score of 0.858 on test data. Feature importances are shown below Step4: Locations were most important features in this model. Although I tried using more features generated by differentiating different kind of crimes or violations, the AUC scores did not improve. Feature importance can also be viewed using tree representation Step5: To reduce variance of the model, since overfitting was observed during training. I also tried to reduce variance by including in more nonblighted buildings by sampling again multiple times with replacement (bagging). A final AUC score of 0.8625 was achieved. The resulted ROC Curve on test data is shown below
<ASSISTANT_TASK:> Python Code: import numpy as np import pandas as pd import matplotlib.pyplot as plt from IPython.display import Image %matplotlib inline Explanation: Study of Correlation Between Building Demolition and Associated Features Capstone Project for Data Science at Scale on Coursera Repo is located here Chen Yang yangcnju@gmail.com End of explanation # The resulted buildings: Image("./data/buildings_distribution.png") Explanation: Objective Build a model to make predictions on blighted buildings based on real data from data.detroitmi.gov as given by coursera. Building demolition is very important for the city to turn around and revive its economy. However, it's no easy task. Accurate predictions can provide guidance on potential blighted buildings and help avoid complications at early stages. Building List The buildings were defined as described below: Building sizes were estimated using parcel info downloaded here at data.detroitmi.gov. Details can be found in this notebook. A event table was constructed from the 4 files (detroit-311.csv, detroit-blight-violations.csv, detroit-crime.csv, and detroit-demolition-permits.tsv) using their coordinates, as shown here. Buildings were defined using these coordinates with an estimated building size (median of all parcels). Each building was represented as a same sized rectangle. End of explanation Image('./data/train_process.png') Explanation: Features Three kinds (311-calls, blight-violations, and crimes) of incident counts and coordinates (normalized) was used in the end. I also tried to generate more features by differentiating each kind of crimes or each kind of violations in this notebook. However, these differentiated features lead to smaller AUC scores. Data The buildings were down-sampled to contain same number of blighted buildings and non-blighted ones. The ratio between train and test was set at a ratio of 80:20. During training using xgboost, the train data was further separated into train and evaluation with a ratio of 80:20 for monitoring. Model A Gradient Boosted Tree model using Xgboost achieved AUC score of 0.85 on evaluation data set: End of explanation Image('./data/feature_f_scores.png') Explanation: This model resulted in an AUC score of 0.858 on test data. Feature importances are shown below: End of explanation Image('./data/bst_tree.png') Explanation: Locations were most important features in this model. Although I tried using more features generated by differentiating different kind of crimes or violations, the AUC scores did not improve. Feature importance can also be viewed using tree representation: End of explanation Image('./data/ROC_Curve_combined.png') Explanation: To reduce variance of the model, since overfitting was observed during training. I also tried to reduce variance by including in more nonblighted buildings by sampling again multiple times with replacement (bagging). A final AUC score of 0.8625 was achieved. The resulted ROC Curve on test data is shown below: End of explanation <END_TASK>
15,949
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Histograma e Estatística de uma imagem Histograma de uma imagem A imagem normalmente é composta de um grande número de pixels; hoje em dia, celulares tiram fotografias com a resolução espacial que produz alguns milhões de pixels. Uma das caracterizações ou "assinaturas" mais eficientes da imagem é seu histograma. Estude o pequeno tutorial disponível em Step1: Uso do matplotlib Veja aqui neste breve tutorial como utilizar o matplotlib Step2: Procure utilizar a função iah2stats operando no histograma para calcular estes mesmos valores. utilize ia.iah2stats para calcular média e mediana da imagem da retina.tif Step3: Histograma de imagens aleatórias Calcule e plote os histogramas das imagens f1, f2 e f3 a seguir, dando uma justificativa para cada caso em relação ao formato do histograma. Aproveite para estudar as funções de geração de dados aleatórios "randint", "shuffle" e "random.normal"
<ASSISTANT_TASK:> Python Code: %matplotlib inline import matplotlib.pyplot as plt import matplotlib.image as mpimg import numpy as np import sys,os ia898path = os.path.abspath('../../') if ia898path not in sys.path: sys.path.append(ia898path) import ia898.src as ia !ls -l ../../ia898/data f = mpimg.imread('../data/retina.tif') plt.imshow(f,cmap='gray'); h = ia.histogram(f) plt.plot(h); Explanation: Histograma e Estatística de uma imagem Histograma de uma imagem A imagem normalmente é composta de um grande número de pixels; hoje em dia, celulares tiram fotografias com a resolução espacial que produz alguns milhões de pixels. Uma das caracterizações ou "assinaturas" mais eficientes da imagem é seu histograma. Estude o pequeno tutorial disponível em: Tutorial - Histograma e Estatísticas e após isto, plote com adshow(iaplot()) o histograma da imagem a seguir: End of explanation %%time print('f.mean()=', f.mean()) print('np.median(f)=', np.median(f)) %%time h = ia.histogram(f) v = ia.h2stats(h) print('média:',v[0]) print('mediano:',v[8]) Explanation: Uso do matplotlib Veja aqui neste breve tutorial como utilizar o matplotlib: tutorial_matplotlib Estatísticas de uma imagem Estude a função ia636:iah2stats que calcula diversas estatísticas da imagem a partir de seu histograma. Observe que estas estatísticas podem ser obtidas tanto diretamente da imagem como do histograma. A grande vantagem de calcular a estatística a partir do histograma é que o histograma é tipicamente um vetor de 256 elementos enquanto que a imagem pode ser de milhões de pixels. Assim, quando se precisa calcular diversos valores estatísticos de uma imagem, calcula-se seu histograma e posteriormente seus valores estatísticos, calculados a partir do histogram. Como ilustração, iremos medir o tempo para calcular a média (np.mean) e o valor mediano (np.median) fazendo o cálculo diretamente na imagem e comparar com o tempo de se calcular estas estatísticas a partir do histograma: End of explanation plt.imshow(f,cmap='gray') h = ia.histogram(f) v = ia.h2stats(h) print('mean =',v[0]) print('variance =',v[1]) print('skewness =',v[2]) print('kurtosis = ',v[3]) print('entropy = ',v[4]) print('mode = ',v[5]) print('percentil 1% = ',v[6]) print('percentil 10% = ',v[7]) print('percentil 50% = ',v[8]) print('percentil 90% = ',v[9]) print('percentil 99% = ',v[10]) Explanation: Procure utilizar a função iah2stats operando no histograma para calcular estes mesmos valores. utilize ia.iah2stats para calcular média e mediana da imagem da retina.tif End of explanation f1 = np.random.randint(0,256, (128,256)).astype(np.uint8) plt.imshow( f1, 'f1') h1 = ia.histogram(f1) plt.bar( np.arange(h1.size), h1) plt.imshow( ia.iafig2img(fig2), 'histograma de f1') f_2 = np.resize(np.arange(256, dtype=np.uint8), (128,256)) # imagem rampa f2 = np.ravel(f_2) np.random.shuffle(f2) adshow(f2.reshape(f_2.shape), 'f2') h2 = ia.iahistogram(f2) fig3 = plt.figure() plt.bar( np.arange(h2.size), h2) adshow( ia.iafig2img(fig3), 'histograma de f2') f_3 = np.random.randn(128,256) f3 = ia.ianormalize(f_3).astype(np.uint8) adshow(f3, 'f3') h3 = ia.iahistogram(f3) fig4 = plt.figure() plt.bar( np.arange(h3.size), h3) adshow( ia.iafig2img(fig4), 'histograma de f3') Explanation: Histograma de imagens aleatórias Calcule e plote os histogramas das imagens f1, f2 e f3 a seguir, dando uma justificativa para cada caso em relação ao formato do histograma. Aproveite para estudar as funções de geração de dados aleatórios "randint", "shuffle" e "random.normal": End of explanation <END_TASK>
15,950
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: The Basic Tools of the Deep Life Sciences Welcome to DeepChem's introductory tutorial for the deep life sciences. This series of notebooks is a step-by-step guide for you to get to know the new tools and techniques needed to do deep learning for the life sciences. We'll start from the basics, assuming that you're new to machine learning and the life sciences, and build up a repertoire of tools and techniques that you can use to do meaningful work in the life sciences. Scope Step1: You can of course run this tutorial locally if you prefer. In this case, don't run the above cell since it will download and install Anaconda on your local machine. In either case, we can now import the deepchem package to play with. Step2: Training a Model with DeepChem Step3: I won't say too much about this code right now. We will see many similar examples in later tutorials. There are two details I do want to draw your attention to. First, notice the featurizer argument passed to the load_delaney() function. Molecules can be represented in many ways. We therefore tell it which representation we want to use, or in more technical language, how to "featurize" the data. Second, notice that we actually get three different data sets Step4: Here again I will not say much about the code. Later tutorials will give lots more information about GraphConvModel, as well as other types of models provided by DeepChem. We now need to train the model on the data set. We simply give it the data set and tell it how many epochs of training to perform (that is, how many complete passes through the data to make). Step5: If everything has gone well, we should now have a fully trained model! But do we? To find out, we must evaluate the model on the test set. We do that by selecting an evaluation metric and calling evaluate() on the model. For this example, let's use the Pearson correlation, also known as r<sup>2</sup>, as our metric. We can evaluate it on both the training set and test set. Step6: Notice that it has a higher score on the training set than the test set. Models usually perform better on the particular data they were trained on than they do on similar but independent data. This is called "overfitting", and it is the reason it is essential to evaluate your model on an independent test set. Our model still has quite respectable performance on the test set. For comparison, a model that produced totally random outputs would have a correlation of 0, while one that made perfect predictions would have a correlation of 1. Our model does quite well, so now we can use it to make predictions about other molecules we care about. Since this is just a tutorial and we don't have any other molecules we specifically want to predict, let's just use the first ten molecules from the test set. For each one we print out the chemical structure (represented as a SMILES string) and the predicted log(solubility). To put these predictions in context, we print out the log(solubility) values from the test set as well.
<ASSISTANT_TASK:> Python Code: !pip install --pre deepchem Explanation: The Basic Tools of the Deep Life Sciences Welcome to DeepChem's introductory tutorial for the deep life sciences. This series of notebooks is a step-by-step guide for you to get to know the new tools and techniques needed to do deep learning for the life sciences. We'll start from the basics, assuming that you're new to machine learning and the life sciences, and build up a repertoire of tools and techniques that you can use to do meaningful work in the life sciences. Scope: This tutorial will encompass both the machine learning and data handling needed to build systems for the deep life sciences. Colab This tutorial and the rest in the sequences are designed to be done in Google colab. If you'd like to open this notebook in colab, you can use the following link. Why do the DeepChem Tutorial? 1) Career Advancement: Applying AI in the life sciences is a booming industry at present. There are a host of newly funded startups and initiatives at large pharmaceutical and biotech companies centered around AI. Learning and mastering DeepChem will bring you to the forefront of this field and will prepare you to enter a career in this field. 2) Humanitarian Considerations: Disease is the oldest cause of human suffering. From the dawn of human civilization, humans have suffered from pathogens, cancers, and neurological conditions. One of the greatest achievements of the last few centuries has been the development of effective treatments for many diseases. By mastering the skills in this tutorial, you will be able to stand on the shoulders of the giants of the past to help develop new medicine. 3) Lowering the Cost of Medicine: The art of developing new medicine is currently an elite skill that can only be practiced by a small core of expert practitioners. By enabling the growth of open source tools for drug discovery, you can help democratize these skills and open up drug discovery to more competition. Increased competition can help drive down the cost of medicine. Getting Extra Credit If you're excited about DeepChem and want to get more involved, there are some things that you can do right now: Star DeepChem on GitHub! - https://github.com/deepchem/deepchem Join the DeepChem forums and introduce yourself! - https://forum.deepchem.io Say hi on the DeepChem gitter - https://gitter.im/deepchem/Lobby Make a YouTube video teaching the contents of this notebook. Prerequisites This tutorial sequence will assume some basic familiarity with the Python data science ecosystem. We will assume that you have familiarity with libraries such as Numpy, Pandas, and TensorFlow. We'll provide some brief refreshers on basics through the tutorial so don't worry if you're not an expert. Setup The first step is to get DeepChem up and running. We recommend using Google Colab to work through this tutorial series. You'll also need to run the following commands to get DeepChem installed on your colab notebook. End of explanation import deepchem as dc dc.__version__ Explanation: You can of course run this tutorial locally if you prefer. In this case, don't run the above cell since it will download and install Anaconda on your local machine. In either case, we can now import the deepchem package to play with. End of explanation tasks, datasets, transformers = dc.molnet.load_delaney(featurizer='GraphConv') train_dataset, valid_dataset, test_dataset = datasets Explanation: Training a Model with DeepChem: A First Example Deep learning can be used to solve many sorts of problems, but the basic workflow is usually the same. Here are the typical steps you follow. Select the data set you will train your model on (or create a new data set if there isn't an existing suitable one). Create the model. Train the model on the data. Evaluate the model on an independent test set to see how well it works. Use the model to make predictions about new data. With DeepChem, each of these steps can be as little as one or two lines of Python code. In this tutorial we will walk through a basic example showing the complete workflow to solve a real world scientific problem. The problem we will solve is predicting the solubility of small molecules given their chemical formulas. This is a very important property in drug development: if a proposed drug isn't soluble enough, you probably won't be able to get enough into the patient's bloodstream to have a therapeutic effect. The first thing we need is a data set of measured solubilities for real molecules. One of the core components of DeepChem is MoleculeNet, a diverse collection of chemical and molecular data sets. For this tutorial, we can use the Delaney solubility data set. The property of solubility in this data set is reported in log(solubility) where solubility is measured in moles/liter. End of explanation model = dc.models.GraphConvModel(n_tasks=1, mode='regression', dropout=0.2) Explanation: I won't say too much about this code right now. We will see many similar examples in later tutorials. There are two details I do want to draw your attention to. First, notice the featurizer argument passed to the load_delaney() function. Molecules can be represented in many ways. We therefore tell it which representation we want to use, or in more technical language, how to "featurize" the data. Second, notice that we actually get three different data sets: a training set, a validation set, and a test set. Each of these serves a different function in the standard deep learning workflow. Now that we have our data, the next step is to create a model. We will use a particular kind of model called a "graph convolutional network", or "graphconv" for short. End of explanation model.fit(train_dataset, nb_epoch=100) Explanation: Here again I will not say much about the code. Later tutorials will give lots more information about GraphConvModel, as well as other types of models provided by DeepChem. We now need to train the model on the data set. We simply give it the data set and tell it how many epochs of training to perform (that is, how many complete passes through the data to make). End of explanation metric = dc.metrics.Metric(dc.metrics.pearson_r2_score) print("Training set score:", model.evaluate(train_dataset, [metric], transformers)) print("Test set score:", model.evaluate(test_dataset, [metric], transformers)) Explanation: If everything has gone well, we should now have a fully trained model! But do we? To find out, we must evaluate the model on the test set. We do that by selecting an evaluation metric and calling evaluate() on the model. For this example, let's use the Pearson correlation, also known as r<sup>2</sup>, as our metric. We can evaluate it on both the training set and test set. End of explanation solubilities = model.predict_on_batch(test_dataset.X[:10]) for molecule, solubility, test_solubility in zip(test_dataset.ids, solubilities, test_dataset.y): print(solubility, test_solubility, molecule) Explanation: Notice that it has a higher score on the training set than the test set. Models usually perform better on the particular data they were trained on than they do on similar but independent data. This is called "overfitting", and it is the reason it is essential to evaluate your model on an independent test set. Our model still has quite respectable performance on the test set. For comparison, a model that produced totally random outputs would have a correlation of 0, while one that made perfect predictions would have a correlation of 1. Our model does quite well, so now we can use it to make predictions about other molecules we care about. Since this is just a tutorial and we don't have any other molecules we specifically want to predict, let's just use the first ten molecules from the test set. For each one we print out the chemical structure (represented as a SMILES string) and the predicted log(solubility). To put these predictions in context, we print out the log(solubility) values from the test set as well. End of explanation <END_TASK>
15,951
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: DV360 Automation Step1: 0.2 Setup your GCP project To utilise the DV360 API, you need a Google Cloud project. For the purpose of this workshop, we've done this for you, but normally you'd have to complete the following steps, before you can make requests using the DV360 API Step2: 0.4 Set DV360 account settings Next, we need to set our DV360 parameters, and generate a sandbox (test) campaign. Note, if you'd prefer to use an existing campaign, update CAMPAIGN_ID below. Step4: Create a new 'sandbox' campaign to use with the rest of the exercises Executing the following code block will overwrite any CAMPAIGN_ID used above. Step5: 1A) SDF using DBM API (sunset) Important Step6: Define a boilerplate targeting template that all Line Items should adhere too Step7: Modify latest SDF LineItems file and update the columns according to the targeting template Step8: Upload the output .csv file in the DV360 UI Once the changes have been applied successfully, check the 'Targeting' controls within 'Line Item details' 1.3 SDF + Entity Read Files What are Entity Read Files (ERFs)? ERFs are flat files (.JSON) in Google Cloud Storage that contain lookup values for DV360 entities like geographies, creatives, etc. Each DV360 entity (Advertiser, Campaign, LineItem, etc) has a corresponding .JSON file in Cloud Storage retained free-of-charge for 60 days from their processing date. ERFs consist of 1 file per entity type, written x1 per day to two seperate Cloud buckets Step9: Retrieve a list of country codes / IDs from GeoLocation.json for each of our store locations Step10: Download the latest SDF LineItems (because we've made changes since our last download) Step11: Modify the contents of the latest SDF output, then save a new CSV with updated Geo Targeting IDs Step12: Upload the output .csv file in the DV360 UI Once the changes have been applied successfully, check the 'Targeting' controls within 'Line Item details' 1.4 SDF + Cloud Vision API Next, let's look at how we you can utilise external APIs. Download the 'product_feed' tab from Google Store as CSV (File >> Download >> CSV) Execute the following code block and upload 'product_feed.csv' This will create a new Python dictionary (key Step14: Define a function to send images to the Cloud Vision API Step15: Run our images through the function, and return a lookup table Step16: Now we have our new labels from the Vision API, we need to write these into the keywords targeting field Step18: Upload the output .csv file in the DV360 UI Once the changes have been applied successfully, check the 'Targeting' controls within 'Line Item details' 1.5 Optimisation using Reports Next, we'll look at how you could combine reporting data, with operations such as optimising bid multipliers or deactivating activity. Note Step19: Download an updated SDF LineItems file, and if the LineItem ID is in the poor performers list, add a Geo bid multiplier to half the bids (0.5) Step20: Note the only rows included in the output, are those that we want to modify. Upload the output .csv file in the DV360 UI Once the changes have been applied successfully, check the 'Targeting' controls within 'Line Item details' 1.6 Challenge Challenge Step21: Solution Step23: Upload the output .csv file in the DV360 UI 1B) SDF using DV360 API Reference Step24: Define a boilerplate targeting template that all Line Items should adhere too Step25: Modify latest SDF LineItems file and update the columns according to the targeting template Step26: Upload the output .csv file in the DV360 UI Once the changes have been applied successfully, check the 'Targeting' controls within 'Line Item details' 1.3 SDF + Entity Read Files What are Entity Read Files (ERFs)? ERFs are flat files (.JSON) in Google Cloud Storage that contain lookup values for DV360 entities like geographies, creatives, etc. Each DV360 entity (Advertiser, Campaign, LineItem, etc) has a corresponding .JSON file in Cloud Storage retained free-of-charge for 60 days from their processing date. ERFs consist of 1 file per entity type, written x1 per day to two seperate Cloud buckets Step27: Retrieve a list of country codes / IDs from GeoLocation.json for each of our store locations Step28: Download the latest SDF LineItems (because we've made changes since our last download) Step29: Modify the contents of the latest SDF output, then save a new CSV with updated Geo Targeting IDs Step30: Upload the output .csv file in the DV360 UI Once the changes have been applied successfully, check the 'Targeting' controls within 'Line Item details' 1.4 SDF + Cloud Vision API Next, let's look at how we you can utilise external APIs. Download the 'product_feed' tab from Google Store as CSV (File >> Download >> CSV) Execute the following code block and upload 'product_feed.csv' This will create a new Python dictionary (key Step32: Define a function to send images to the Cloud Vision API Step33: Run our images through the function, and return a lookup table (reference) Step34: View the results of our Vision analysis Step35: Download the latest SDF LineItems (because we've made changes since our last download) Step36: Now we have our new labels from the Vision API, we need to write these into the keywords targeting field Step37: Upload the output .csv file in the DV360 UI Once the changes have been applied successfully, check the 'Targeting' controls within 'Line Item details' 1.5 Optimisation using Reports Next, we'll look at how you could combine reporting data, with operations such as optimising bid multipliers or deactivating activity. Note Step38: Download an updated SDF LineItems file, and if the LineItem ID is in the poor performers list, add a Geo bid multiplier to half the bids (0.5) Step39: Upload the output .csv file in the DV360 UI Once the changes have been applied successfully, check the 'Targeting' controls within 'Line Item details' 1.6 Challenge Challenge Step40: Solution Step42: Upload the output .csv file in the DV360 UI 2) Display & Video 360 API What is the Display & Video 360 API? The Display & Video 360 API (formly known as the DV360 Write API) is the programmatic interface for the Display & Video 360 platform. It allows developers to easily and efficiently automate complex Display & Video 360 workflows, such as creating insertion orders and setting targeting options for individual line items. We'll use it now to build upon the campaign we created earlier using SDF. Reference Step43: Upload the extended feed for Google Store's new territories Step45: Create Insertion Order template Here we're defining a new a function called 'create_insertion_order'. Note Step47: Create LineItem template Here we define a new function called 'create_lineitem', based on a template we specified. Note Step48: Build our new campaign First, we'll loop through the list of countries generated at the beginning, and for each country, create a new Insertion Order by calling our function 'create_insertion_order'. Within that loop, we find every product that is sold in the corresponding country-code, and create a new Line Item for every matching product using our function 'create_lineitem'. Sit tight, this one can take a while (~10 mins)... Link to DV360 UI Step49: If successful, the result should look similar to the below in DV360 Step50: Apply individual targeting criteria to single entity Step51: Applying individual targeting criteria to multiple entities Step53: 2.3 Bulk targeting Bulk updates using templated targeting controls Step54: Retrieve list of active LineItems, and Apply bulk targeting Step56: 2.4 Optimisation (external trigger) The following optimisations will be completed on your campaign, created earlier. Create functions to 'deactivate' or 'optimise' Lineitems Step57: Creat list of out of stock products Step58: Process optimisation Step59: 2.5 Optimisation (reporting data) As your new campaign has no performance data, the following optimisations will be completed on an existing campaign with historical data. Create new performance report and fetch results Step60: Load report to Pandas DataFrame Step61: Create two lists of poorly performing LineItems 1. LineItems that should be paused 2. Lineitems to reduce bids Step62: Process optimisation Step64: 2.6 Creative upload Uploading Display creatives from remote storage (http) The following demonstrates how to upload image assets from remote storage, but it's also possible to upload from local storage. Reference Step65: Upload image creatives Note, all of the following assets are the same dimension (300x250) and type 'CREATIVE_TYPE_STANDARD'. When uploading assets of multiple sizes, the creatives.create body must reflect this. Step66: 2.7 Challenge Challenge Step67: Solution Step73: Link to DV360 UI Resources Getting started with SDF in DV360 guide Structured Data Files (SDF) developer guide Getting started with the Display & Video 360 API developer guide Getting started with the DoubleClick Bid Manager API developer guide How to access Entity Read Files Quickstart
<ASSISTANT_TASK:> Python Code: !pip install google-api-python-client !pip install google-cloud-vision import csv import datetime import io import json import pprint from google.api_core import retry from google.cloud import vision from google.colab import files from google_auth_oauthlib.flow import InstalledAppFlow from googleapiclient import discovery from googleapiclient import http import pandas as pd import requests print('Successfully imported Python libraries!') Explanation: DV360 Automation: codelab Author: Matt Lynam Objective Enable Display & Video 360 (DV360) advertisers to increase workflow efficiency by utilising the right automation solution according to their needs, resources and technical capability. Goals * Provide an overview of the current automation suite available in DV360 * Demonstrate the capabilities and limitations of DV360's UI and APIs * Explore common advertiser use cases and pitfalls * Acquire hands-on experience by applying key concepts using a fictional case study 0) Setup and authentication Google Colab primer Google Colaboratory, or "Colab" for short, allows you to write and execute Python in your browser, with: - Zero configuration required - Free access to GPUs - Easy sharing & colaboration A notebook is a list of cells, containing either explanatory text or executable code and its output. This is a text cell. Useful Colab tips * Double-click within the cell to edit * Code cells can be executed by clicking the Play icon in the left gutter of the cell; or with Cmd/Ctrl + Enter to run the cell in place; * Use Cmd/Ctrl + / to comment out a line of code 0.1 Install Python client libraries Run the following block to install the latest Google Python Client Library and import additional libraries used for this workshop. End of explanation API_SCOPES = ['https://www.googleapis.com/auth/doubleclickbidmanager', 'https://www.googleapis.com/auth/display-video', 'https://www.googleapis.com/auth/devstorage.read_only', 'https://www.googleapis.com/auth/cloud-vision'] # Authenticate using user credentials stored in client_secrets.json client_secrets_file = files.upload() client_secrets_json = json.loads(next(iter(client_secrets_file.values()))) flow = InstalledAppFlow.from_client_config(client_secrets_json, API_SCOPES) credentials = flow.run_console() print('Success!') # Build DBM Read API service object dbm_service = discovery.build( 'doubleclickbidmanager', 'v1.1', credentials=credentials) print('DBM API service object created') # Build Google Cloud Storage Read API service object gcs_service = discovery.build('storage', 'v1', credentials=credentials) print('GCS service object created') # Create Display Video API service object display_video_service = discovery.build( 'displayvideo', 'v1', credentials=credentials) print('Display Video API service object created') Explanation: 0.2 Setup your GCP project To utilise the DV360 API, you need a Google Cloud project. For the purpose of this workshop, we've done this for you, but normally you'd have to complete the following steps, before you can make requests using the DV360 API: Select or create a Google Cloud Platform project. Enable billing on your project. Enable the 'Display & Video 360' and 'DoubleClick Bid Manager' API from the API library Create GCP credentials We've also generated credentials for you, but if you needed to generate new credentials, this would be the process: Go to the API credentials page in the Cloud Platform Console. Fill out the required fields on the OAuth consent screen. On the credentials page, click Create credentials >> OAuth client ID. Select Other as the application type, and then click Create. Download the credentials by clicking the Download JSON button Reference: https://developers.google.com/display-video/api/guides/how-tos/authorizing 0.3 Authentication Next, we'll permission the application to submit authorised API requests on our behalf using OAuth authentication. The following scopes are specified in an array: * DBM API * Display Video API * GCP Storage Read * Cloud Vision API Reference: * Example OAuth2 Python Library * Google scopes End of explanation PARTNER_ID = '234340' #@param {type:"string"} ADVERTISER_ID = '2436036' #@param {type:"string"} CAMPAIGN_ID = '4258803' #@param {type:"string"} # For use with legacy DBM API SDF_VERSION = '5.3' #@param {type:"string"} # For use with DV360 API SDF_VERSION_DV360 = 'SDF_VERSION_5_3' #@param {type:"string"} print('DV360 settings saved!') Explanation: 0.4 Set DV360 account settings Next, we need to set our DV360 parameters, and generate a sandbox (test) campaign. Note, if you'd prefer to use an existing campaign, update CAMPAIGN_ID below. End of explanation YOUR_NAME = 'Matt' #@param {type:"string"} # Set dates for new campaign month = datetime.datetime.today().strftime('%m') day = datetime.datetime.today().strftime('%d') year = datetime.datetime.today().strftime('%Y') month_plus30 = (datetime.datetime.today() + datetime.timedelta(days=30)).strftime('%m') day_plus30 = (datetime.datetime.today() + datetime.timedelta(days=30)).strftime('%d') year_plus30 = (datetime.datetime.today() + datetime.timedelta(days=30)).strftime('%Y') def create_campaign(YOUR_NAME): Creates a new DV360 Campaign object. campaign_name = f'{year}-{month}-{day} | {YOUR_NAME}' campaign_obj = { 'displayName': campaign_name, 'entityStatus': 'ENTITY_STATUS_ACTIVE', 'campaignGoal': { 'campaignGoalType': 'CAMPAIGN_GOAL_TYPE_ONLINE_ACTION', 'performanceGoal': { 'performanceGoalType': 'PERFORMANCE_GOAL_TYPE_CPC', 'performanceGoalAmountMicros': 1000000 } }, 'campaignFlight': { 'plannedSpendAmountMicros': 1000000, 'plannedDates': { 'startDate': { 'year': year, 'month': month, 'day': day }, 'endDate': { 'year': year_plus30, 'month': month_plus30, 'day': day_plus30 } } }, 'frequencyCap': { 'maxImpressions': 10, 'timeUnit': 'TIME_UNIT_DAYS', 'timeUnitCount': 1 } } # Create the campaign. campaign = display_video_service.advertisers().campaigns().create( advertiserId=ADVERTISER_ID, body=campaign_obj ).execute() return campaign new_campaign = create_campaign(YOUR_NAME) # Display the new campaign. CAMPAIGN_ID = new_campaign['campaignId'] print(f"\nCampaign '{new_campaign['name']}' was created." f"\nCampaign id: '{new_campaign['campaignId']}'" f"\nCampaign name: '{new_campaign['displayName']}'" f"\nCampaign status: '{new_campaign['entityStatus']}'") Explanation: Create a new 'sandbox' campaign to use with the rest of the exercises Executing the following code block will overwrite any CAMPAIGN_ID used above. End of explanation # Configure the sdf.download request request_body = { 'fileTypes': ['LINE_ITEM'], 'filterType': 'CAMPAIGN_ID', 'filterIds': [CAMPAIGN_ID], 'version': SDF_VERSION } # Make the request to download all SDF LineItems for your new campaign request = dbm_service.sdf().download(body=request_body) response = request.execute() # Load SDF response to Pandas DataFrame sdf_df = pd.read_csv(io.StringIO(response['lineItems'])) # Show sample (5 rows) of DataFrame sdf_df.head() Explanation: 1A) SDF using DBM API (sunset) Important: the SDF resource (sdf.download) for the DBM API has migrated to a new endpoint (displayvideo.googleapis.com). SDF methods using this (doubleclickbidmanager.googleapis.com) endpoint have been sunset, and will not be updated moving forward. Please follow track 1B, for code samples using the DV360 API. Reference: https://developers.google.com/bid-manager/v1.1/sdf Structured Data Files (SDF) are a way of using spreadsheets to make bulk changes to DV360 entities, including Campaigns, Insertion Orders, Line Items, TrueView Ad Groups, TrueView Ads and deals. SDF are the first step on the path to full automation in DV360, but only allow you to automate so far, as we'll explore now... 1.1 Manually create SDF Create a copy of the Google Store product feed Update the highlighted cells (B2:B3) on the tab called "sdf_insertionorders" Save the updated "sdf_insertionorders" tab and "sdf_lineitems" tab to .CSV (File >> Download >> CSV) Upload the two .CSV files together in the DV360 UI This will create a very basic campaign, with 2 insertion orders, and 10 lineitems per insertion order. 1.2 Editing SDF programmatically Our new LineItems are missing some important targeting and inventory controls: * Channels (e.g. groups of publisher URLs) * Inventory source * Brand safety * Geo targeting Let’s use software to make these changes for us... End of explanation targeting_template = { 'Channel Targeting - Include': '2580510;', 'Channel Targeting - Exclude': '2580509;', 'Inventory Source Targeting - Include': '1;', 'Inventory Source Targeting - Exclude': '6; 8; 9; 10; 2; 11; 12; 13; 16; 20; 23; 27; 29; 30; 31; 34; 35; 36; ' '38; 43; 46; 50; 51; 56; 60; 63; 67; 74;', 'Digital Content Labels - Exclude': 'G; PG; T;', 'Brand Safety Sensitivity Setting': 'Use custom', 'Brand Safety Custom Settings': 'Adult; Alcohol; Derogatory; Downloads & Sharing; Drugs; Gambling; ' 'Politics; Profanity; Religion; Sensitive social issues; Suggestive; ' 'Tobacco; Tragedy; Transportation Accidents; Violence; Weapons;' } Explanation: Define a boilerplate targeting template that all Line Items should adhere too End of explanation # Overwrite targeting columns using 'targeting_template' sdf_df['Channel Targeting - Include'] = targeting_template[ 'Channel Targeting - Include'] sdf_df['Channel Targeting - Exclude'] = targeting_template[ 'Channel Targeting - Exclude'] sdf_df['Inventory Source Targeting - Include'] = targeting_template[ 'Inventory Source Targeting - Include'] sdf_df['Inventory Source Targeting - Exclude'] = targeting_template[ 'Inventory Source Targeting - Exclude'] sdf_df['Digital Content Labels - Exclude'] = targeting_template[ 'Digital Content Labels - Exclude'] sdf_df['Brand Safety Sensitivity Setting'] = targeting_template[ 'Brand Safety Sensitivity Setting'] sdf_df['Brand Safety Custom Settings'] = targeting_template[ 'Brand Safety Custom Settings'] # Save modified dataframe to remote storage in Colab sdf_df.to_csv('sdf_update1_controls.csv', index=False) # Show sample (5 rows) of DataFrame sdf_df.head() # Download modified csv to local storage files.download('sdf_update1_controls.csv') print( "Success, check your downloads for a file called 'sdf_update1_controls.csv'" ) Explanation: Modify latest SDF LineItems file and update the columns according to the targeting template End of explanation # Actually today-7 to avoid issues with collection yesterday = datetime.date.today() - datetime.timedelta(7) # Download public ERF for geolocation info request = gcs_service.objects().get_media( bucket='gdbm-public', object='entity/' + yesterday.strftime('%Y%m%d') + '.0.GeoLocation.json') response = request.execute() geolocations = json.loads(response) print('GeoLocation.json successfully downloaded \n') print("Here's a random sample of 5 entries:\n") pprint.pprint(geolocations[0:5]) Explanation: Upload the output .csv file in the DV360 UI Once the changes have been applied successfully, check the 'Targeting' controls within 'Line Item details' 1.3 SDF + Entity Read Files What are Entity Read Files (ERFs)? ERFs are flat files (.JSON) in Google Cloud Storage that contain lookup values for DV360 entities like geographies, creatives, etc. Each DV360 entity (Advertiser, Campaign, LineItem, etc) has a corresponding .JSON file in Cloud Storage retained free-of-charge for 60 days from their processing date. ERFs consist of 1 file per entity type, written x1 per day to two seperate Cloud buckets: Public (10 .JSON files) - contain common public data such as GeoLocation and Language which are stored in the gdbm-public bucket (the same bucket for every DV360 user). Private (13 .JSON files) - contain information about the DV360 Partner's campaigns, creatives, budgets and other private data and are stored in Partner-specific buckets (restricted to specific users) Reference: https://developers.google.com/bid-manager/guides/entity-read/overview ERFs can be used to speed up, and automate, the creation of SDF files. Let's explore this now... Download yesterday's GeoLocation.json from public ERF bucket using Google Cloud Storage API End of explanation # Provide a list of store locations store_locations = ['United Kingdom', 'France', 'Spain', 'Germany', 'Portugal'] # Create a new dictionary to save the country code and ID later on geo_targeting_ids = {} # Note: GeoLocation.json is over 800,000 lines for location in geolocations: if location['canonical_name'] in store_locations: geo_targeting_ids[location['country_code']] = location['id'] print(location) print(geo_targeting_ids) Explanation: Retrieve a list of country codes / IDs from GeoLocation.json for each of our store locations End of explanation # Configure the sdf.download request request_body = { 'fileTypes': ['LINE_ITEM'], 'filterType': 'CAMPAIGN_ID', 'filterIds': [CAMPAIGN_ID], 'version': SDF_VERSION } # Make the request to download all SDF LineItems for your new campaign request = dbm_service.sdf().download(body=request_body) response = request.execute() # Load SDF response to Pandas DataFrame sdf_df = pd.read_csv(io.StringIO(response['lineItems'])) # Show sample (5 rows) of DataFrame sdf_df.head() Explanation: Download the latest SDF LineItems (because we've made changes since our last download) End of explanation for country in geo_targeting_ids: target_country = geo_targeting_ids[country] sdf_df.loc[sdf_df.Name.str.contains(country), 'Geography Targeting - Include'] = f'{target_country};' # Save modified dataframe to remote storage in Colab sdf_df.to_csv('sdf_update2_geo.csv', index=False) # Display updated DataFrame sdf_df.head() # Download modified csv to local storage files.download('sdf_update2_geo.csv') print("Success, look for a file called 'sdf_update2_geo.csv' in your downloads folder") Explanation: Modify the contents of the latest SDF output, then save a new CSV with updated Geo Targeting IDs End of explanation # Upload product feed using Colab's upload utility product_feed_csv = files.upload() contents = next(iter(product_feed_csv.values())).decode('utf-8') products = csv.DictReader(io.StringIO(contents)) image_url_list = {} # Iterate through each row and update dict() with sku:link for row in products: image_url_list[row['sku']] = row['image_link'] pprint.pprint(image_url_list) Explanation: Upload the output .csv file in the DV360 UI Once the changes have been applied successfully, check the 'Targeting' controls within 'Line Item details' 1.4 SDF + Cloud Vision API Next, let's look at how we you can utilise external APIs. Download the 'product_feed' tab from Google Store as CSV (File >> Download >> CSV) Execute the following code block and upload 'product_feed.csv' This will create a new Python dictionary (key:value pairing), mapping SKUs with their image link Warning: Cloud Vision API is paid product, utilising the following example in your own Cloud project will incur costs. Try out the Cloud Vision API for free at cloud.google.com/vision End of explanation def vision_analysis(image_url): Process images using the Cloud Vision API. # Assign image URL image = vision.Image() image.source.image_uri = image_url # Instantiates a Vision client client = vision.ImageAnnotatorClient(credentials=credentials) # Performs label detection on the image file vision_response = client.label_detection(image=image) dv360_targeting_keywords = [] labels = [] for label in vision_response.label_annotations: dv360_targeting_keywords.append(label.description) label = f'{label.description} ({label.score:.2%})' labels.append(label) return dv360_targeting_keywords, labels Explanation: Define a function to send images to the Cloud Vision API End of explanation imageslookup = {} for sku, url in image_url_list.items(): imageslookup[sku], vision_labels = vision_analysis(url) print(f'Analysis completed for: {url}') print('Labels (confidence score):') pprint.pprint(vision_labels, indent=4) print('=' * 30) print('\n\nLookup table:') pprint.pprint(imageslookup, indent=4) Explanation: Run our images through the function, and return a lookup table End of explanation # Configure the sdf.download request request_body = { 'fileTypes': ['LINE_ITEM'], 'filterType': 'CAMPAIGN_ID', 'filterIds': [CAMPAIGN_ID], 'version': SDF_VERSION } request = dbm_service.sdf().download(body=request_body) response = request.execute() # Load SDF response to Pandas DataFrame sdf_df = pd.read_csv(io.StringIO(response['lineItems'])) for product in imageslookup: sdf_df.loc[sdf_df.Name.str.contains(product), 'Keyword Targeting - Include'] = ';'.join( imageslookup[product]).lower() # Save modified dataframe to remote storage in Colab sdf_df.to_csv('sdf_update3_keywords.csv', index=False) # Show sample (5 rows) of DataFrame sdf_df.head() # Download modified csv to local storage files.download('sdf_update3_keywords.csv') print("Success, look for the file called 'sdf_update3_keywords.csv' in your downloads folder") Explanation: Now we have our new labels from the Vision API, we need to write these into the keywords targeting field End of explanation # Define DV360 report definition (i.e. metrics and filters) report_definition = { 'params': { 'type': 'TYPE_GENERAL', 'metrics': [ 'METRIC_IMPRESSIONS', 'METRIC_CLICKS', 'METRIC_CTR', 'METRIC_REVENUE_ADVERTISER' ], 'groupBys': [ 'FILTER_ADVERTISER', 'FILTER_INSERTION_ORDER', 'FILTER_LINE_ITEM', 'FILTER_ADVERTISER_CURRENCY' ], 'filters': [{ 'type': 'FILTER_ADVERTISER', 'value': ADVERTISER_ID }], }, 'metadata': { 'title': 'DV360 Automation API-generated report', 'dataRange': 'LAST_90_DAYS', 'format': 'csv' }, 'schedule': { 'frequency': 'ONE_TIME' } } # Create new query using report definition operation = dbm_service.queries().createquery(body=report_definition).execute() pprint.pprint(operation) # Runs the given Queries.getquery request, retrying with an exponential # backoff. Returns completed operation. Will raise an exception if the # operation takes more than five hours to complete. @retry.Retry( predicate=retry.if_exception_type(Exception), initial=5, maximum=60, deadline=18000) def check_get_query_completion(getquery_request): Queries metadata to check for completion. completion_response = getquery_request.execute() pprint.pprint(completion_response) if completion_response['metadata']['running']: raise Exception('The operation has not completed.') return completion_response getquery_request = dbm_service.queries().getquery(queryId=operation['queryId']) getquery_response = check_get_query_completion(getquery_request) report_url = getquery_response['metadata'][ 'googleCloudStoragePathForLatestReport'] # Use skipfooter to remove report footer from data report_df = pd.read_csv(report_url, skipfooter=16, engine='python') report_df.head(10) # Define our 'KPIs' ctr_target = 0.15 imp_threshold = 10000 # Convert IDs to remove decimal point, then string report_df['Line Item ID'] = report_df['Line Item ID'].apply(int) poor_performers = report_df.query( 'Impressions > @imp_threshold & (Clicks / Impressions)*100 < @ctr_target') # Convert results to Python list poor_performers = list(poor_performers['Line Item ID']) print(f'There are {len(poor_performers)} LineItems with a CTR' f' < {ctr_target}% and over {imp_threshold} impressions:' f'\n{poor_performers}') Explanation: Upload the output .csv file in the DV360 UI Once the changes have been applied successfully, check the 'Targeting' controls within 'Line Item details' 1.5 Optimisation using Reports Next, we'll look at how you could combine reporting data, with operations such as optimising bid multipliers or deactivating activity. Note: your new campaign has no performance history, so we'll use an existing campaign for this exercise. End of explanation # Configure the sdf.download request request_body = { 'fileTypes': ['LINE_ITEM'], 'filterType': 'CAMPAIGN_ID', 'filterIds': ['1914007'], 'version': SDF_VERSION } request = dbm_service.sdf().download(body=request_body) response = request.execute() # Load SDF response to Pandas DataFrame sdf_df = pd.read_csv(io.StringIO(response['lineItems'])) for li in poor_performers: geo = sdf_df.loc[sdf_df['Line Item Id'] == li, 'Geography Targeting - Include'].iloc[0] sdf_df.loc[sdf_df['Line Item Id'] == li, 'Bid Multipliers'] = f'(geo; {geo} 0.5;);' # Save modified dataframe to remote storage in Colab sdf_df.to_csv('sdf_update4_bidmultipliers.csv', index=False) # Display updated DataFrame sdf_df.head() files.download('sdf_update4_bidmultipliers.csv') print('Success, your new SDF file has been downloaded') Explanation: Download an updated SDF LineItems file, and if the LineItem ID is in the poor performers list, add a Geo bid multiplier to half the bids (0.5) End of explanation #TODO Explanation: Note the only rows included in the output, are those that we want to modify. Upload the output .csv file in the DV360 UI Once the changes have been applied successfully, check the 'Targeting' controls within 'Line Item details' 1.6 Challenge Challenge: update your campaign with both language and audience targeting. All Lineitems should target the following Google audiences Affinity Categories » Technology » Mobile Enthusiasts Affinity Categories » Technology » Technophiles » High-End Computer Aficionado In-Market Categories » Consumer Electronics LineItems for France, should be targeted at French speakers LineItems for Great Britain, should be targeted at English speakers Tips Google Audience IDs can be found in the DV360 UI or by downloading an SDF with an existing audience applied Language IDs can be found in the Language.json ERF file or by downloading an SDF with the language already applied End of explanation # Format today-2 in required date format yesterday = (datetime.date.today() - datetime.timedelta(2)).strftime('%Y%m%d') # Download ERF for Language.json from public GCS bucket request = gcs_service.objects().get_media( bucket='gdbm-public', object='entity/' + yesterday + '.0.Language.json') response = request.execute() languages = json.loads(response) language_targets = ['en', 'fr'] lang_targeting_ids = {} # Search language.json for language targets 'en' and 'fr' for lang in languages: if lang['code'] in language_targets: lang_targeting_ids[lang['code']] = lang['id'] print(lang) print(lang_targeting_ids) # Define targeting template targeting_template = { 'Affinity & In Market Targeting - Include': '4569529;4586809;4497529;', } # Configure the sdf.download request request_body = { 'fileTypes': ['LINE_ITEM'], 'filterType': 'CAMPAIGN_ID', 'filterIds': [CAMPAIGN_ID], 'version': SDF_VERSION } request = dbm_service.sdf().download(body=request_body) response = request.execute() # Load SDF response to Pandas DataFrame sdf_df = pd.read_csv(io.StringIO(response['lineItems'])) # Update DataFrame with Language and Audience targeting sdf_df.loc[sdf_df.Name.str.contains('GB'), 'Language Targeting - Include'] = f"{lang_targeting_ids['en']};" sdf_df.loc[sdf_df.Name.str.contains('FR'), 'Language Targeting - Include'] = f"{lang_targeting_ids['fr']};" sdf_df['Affinity & In Market Targeting - Include'] = targeting_template[ 'Affinity & In Market Targeting - Include'] # Save modified dataframe to remote storage in Colab sdf_df.to_csv('sdf_update5_challenge.csv', index=False) # Display updated DataFrame sdf_df.head() # Download file to disk using Colab syntax files.download('sdf_update5_challenge.csv') print("Success, check your downloads for a file called 'sdf_update5_challenge.csv'") Explanation: Solution End of explanation def download_sdf(request_body): Download sdf .zip, extract .csv files, load 'SDF-LineItems.csv' to Pandas DataFrame. # Create the sdfdownloadtask sdf_operation = display_video_service.sdfdownloadtasks().create( body=sdf_body).execute() print(f'Operation {sdf_operation["name"]} was created.') # Configure the operations.get request get_request = display_video_service.sdfdownloadtasks().operations().get( name=sdf_operation['name']) # Runs the given operations.get request, retrying with an exponential # backoff. Returns completed operation. Will raise an exception if the # operation takes more than five hours to complete. @retry.Retry(predicate=retry.if_exception_type(Exception), initial=5, maximum=60, deadline=18000) def check_sdf_downloadtask_completion(get_request): operation = get_request.execute() if 'done' not in operation: raise Exception('The operation has not completed.') return operation # Get current status of operation with exponential backoff retry logic operation = check_sdf_downloadtask_completion(get_request) # Check if the operation finished with an error and return if 'error' in operation: raise Exception(f'The operation finished in error with code {operation["error"]["code"]} {operation["error"]["message"]}') print('The operation completed successfully.') print('Resource {operation["response"]["resourceName"]} was created.') # Extract download file resource name to use in download request resource_name = operation['response']['resourceName'] # Configure the Media.download request dowload_request = display_video_service.media().download_media( resourceName=resource_name) output_file = f"{resource_name.replace('/','-')}.zip" # Create output stream for downloaded file outstream = io.FileIO(output_file, mode='wb') # Make downloader object downloader = http.MediaIoBaseDownload(outstream, dowload_request) # Download media file in chunks until finished download_finished = False while download_finished is False: _, download_finished = downloader.next_chunk() print(f'File downloaded to {output_file}') # Load output into a Pandas dataframe df = pd.read_csv(output_file, compression='zip') return df print('Download SDF function created') Explanation: Upload the output .csv file in the DV360 UI 1B) SDF using DV360 API Reference: https://developers.google.com/display-video/api/reference/rest/v1/sdfdownloadtasks/create Structured Data Files (SDF) are a way of using spreadsheets to make bulk changes to DV360 entities, including Campaigns, Insertion Orders, Line Items, TrueView Ad Groups, TrueView Ads and deals. SDF are the first step on the path to full automation in DV360, but only allow you to automate so far, as we'll explore now... 1.1 Manually create SDF Create a copy of the Google Store product feed Update the highlighted cells (B2:B3) on the tab called "sdf_insertionorders" Save the updated "sdf_insertionorders" tab and "sdf_lineitems" tab to .CSV (File >> Download >> CSV) Upload the two .CSV files together in the DV360 UI This will create a very basic campaign, with 2 insertion orders, and 10 lineitems per insertion order. 1.2 Editing SDF programmatically Our new LineItems are missing some important targeting and inventory controls: * Channels (e.g. groups of publisher URLs) * Inventory source * Brand safety * Geo targeting Let’s use software to make these changes for us... Create a function to download SDFs As we'll be downloading multiple SDF files in the next exercises, we've created a function to handle to the download process for us. End of explanation targeting_template = { 'Channel Targeting - Include': '2580510;', 'Channel Targeting - Exclude': '2580509;', 'Inventory Source Targeting - Include': '1;', 'Inventory Source Targeting - Exclude': '6; 8; 9; 10; 2; 11; 12; 13; 16; 20; 23; 27; 29; 30; 31; 34; 35; 36; ' '38; 43; 46; 50; 51; 56; 60; 63; 67; 74;', 'Digital Content Labels - Exclude': 'G; PG; T;', 'Brand Safety Sensitivity Setting': 'Use custom', 'Brand Safety Custom Settings': 'Adult; Alcohol; Derogatory; Downloads & Sharing; Drugs; Gambling; ' 'Politics; Profanity; Religion; Sensitive social issues; Suggestive; ' 'Tobacco; Tragedy; Transportation Accidents; Violence; Weapons;' } Explanation: Define a boilerplate targeting template that all Line Items should adhere too End of explanation # Configure the sdfdownloadtasks.create request sdf_body = { 'version': SDF_VERSION_DV360, 'advertiserId': ADVERTISER_ID, 'parentEntityFilter': { 'fileType': ['FILE_TYPE_LINE_ITEM'], 'filterType': 'FILTER_TYPE_CAMPAIGN_ID', 'filterIds': [CAMPAIGN_ID] } } # Fetch updated SDF lineitem sdf_df = download_sdf(sdf_body) # Overwrite targeting columns using 'targeting_template' sdf_df['Channel Targeting - Include'] = targeting_template[ 'Channel Targeting - Include'] sdf_df['Channel Targeting - Exclude'] = targeting_template[ 'Channel Targeting - Exclude'] sdf_df['Inventory Source Targeting - Include'] = targeting_template[ 'Inventory Source Targeting - Include'] sdf_df['Inventory Source Targeting - Exclude'] = targeting_template[ 'Inventory Source Targeting - Exclude'] sdf_df['Digital Content Labels - Exclude'] = targeting_template[ 'Digital Content Labels - Exclude'] sdf_df['Brand Safety Sensitivity Setting'] = targeting_template[ 'Brand Safety Sensitivity Setting'] sdf_df['Brand Safety Custom Settings'] = targeting_template[ 'Brand Safety Custom Settings'] # Save modified dataframe to remote storage in Colab sdf_df.to_csv('sdf_update1_controls.csv', index=False) # Show sample (5 rows) of DataFrame sdf_df.head() # Download modified csv to local storage files.download('sdf_update1_controls.csv') print( "Success, check your downloads for a file called 'sdf_update1_controls.csv'" ) Explanation: Modify latest SDF LineItems file and update the columns according to the targeting template End of explanation # Actually today-7 to avoid issues with collection yesterday = datetime.date.today() - datetime.timedelta(7) # Download public ERF for geolocation info request = gcs_service.objects().get_media( bucket='gdbm-public', object='entity/' + yesterday.strftime('%Y%m%d') + '.0.GeoLocation.json') response = request.execute() geolocations = json.loads(response) print('GeoLocation.json successfully downloaded \n') print("Here's a random sample of 5 entries:\n") pprint.pprint(geolocations[0:5]) Explanation: Upload the output .csv file in the DV360 UI Once the changes have been applied successfully, check the 'Targeting' controls within 'Line Item details' 1.3 SDF + Entity Read Files What are Entity Read Files (ERFs)? ERFs are flat files (.JSON) in Google Cloud Storage that contain lookup values for DV360 entities like geographies, creatives, etc. Each DV360 entity (Advertiser, Campaign, LineItem, etc) has a corresponding .JSON file in Cloud Storage retained free-of-charge for 60 days from their processing date. ERFs consist of 1 file per entity type, written x1 per day to two seperate Cloud buckets: Public (10 .JSON files) - contain common public data such as GeoLocation and Language which are stored in the gdbm-public bucket (the same bucket for every DV360 user). Private (13 .JSON files) - contain information about the DV360 Partner's campaigns, creatives, budgets and other private data and are stored in Partner-specific buckets (restricted to specific users) Reference: https://developers.google.com/bid-manager/guides/entity-read/overview ERFs can be used to speed up, and automate, the creation of SDF files. Let's explore this now... Download yesterday's GeoLocation.json from public ERF bucket using Google Cloud Storage API End of explanation # Provide a list of store locations store_locations = ['United Kingdom', 'France', 'Spain', 'Germany', 'Portugal'] # Create a new dictionary to save the country code and ID later on geo_targeting_ids = {} # Note: GeoLocation.json is over 800,000 lines for location in geolocations: if location['canonical_name'] in store_locations: geo_targeting_ids[location['country_code']] = location['id'] print(location) print(geo_targeting_ids) Explanation: Retrieve a list of country codes / IDs from GeoLocation.json for each of our store locations End of explanation # Configure the sdfdownloadtasks.create request sdf_body = { 'version': SDF_VERSION_DV360, 'advertiserId': ADVERTISER_ID, 'parentEntityFilter': { 'fileType': ['FILE_TYPE_LINE_ITEM'], 'filterType': 'FILTER_TYPE_CAMPAIGN_ID', 'filterIds': [CAMPAIGN_ID] } } sdf_df = download_sdf(sdf_body) sdf_df.head() Explanation: Download the latest SDF LineItems (because we've made changes since our last download) End of explanation for country in geo_targeting_ids: target_country = geo_targeting_ids[country] sdf_df.loc[sdf_df.Name.str.contains(country), 'Geography Targeting - Include'] = f'{target_country};' # Save modified dataframe to remote storage in Colab sdf_df.to_csv('sdf_update2_geo.csv', index=False) # Display updated DataFrame sdf_df.head() # Download modified csv to local storage files.download('sdf_update2_geo.csv') print("Success, see file 'sdf_update2_geo.csv' in your downloads folder") Explanation: Modify the contents of the latest SDF output, then save a new CSV with updated Geo Targeting IDs End of explanation # Upload product feed using Colab's upload utility product_feed_csv = files.upload() contents = next(iter(product_feed_csv.values())).decode('utf-8') products = csv.DictReader(io.StringIO(contents)) image_url_list = {} # Iterate through each row and update dict() with sku:link for row in products: image_url_list[row['sku']] = row['image_link'] pprint.pprint(image_url_list) Explanation: Upload the output .csv file in the DV360 UI Once the changes have been applied successfully, check the 'Targeting' controls within 'Line Item details' 1.4 SDF + Cloud Vision API Next, let's look at how we you can utilise external APIs. Download the 'product_feed' tab from Google Store as CSV (File >> Download >> CSV) Execute the following code block and upload 'product_feed.csv' This will create a new Python dictionary (key:value pairing), mapping SKUs with their image link Warning: Cloud Vision API is paid product, utilising the following example in your own Cloud project will incur costs. Try out the Cloud Vision API for free at cloud.google.com/vision End of explanation def vision_analysis(image_url): Process images using the Cloud Vision API. # Assign image URL image = vision.Image() image.source.image_uri = image_url # Instantiates a Vision client client = vision.ImageAnnotatorClient(credentials=credentials) # Performs label detection on the image file response = client.label_detection(image=image) dv360_targeting_keywords = [] vision_labels = [] for label in response.label_annotations: dv360_targeting_keywords.append(label.description) label = f'{label.description} ({label.score:.2%})' vision_labels.append(label) return dv360_targeting_keywords, vision_labels print("Vision function created") Explanation: Define a function to send images to the Cloud Vision API End of explanation imageslookup = {} for sku, url in image_url_list.items(): imageslookup[sku], vision_labels = vision_analysis(url) print(f'Analysis completed for: {url}') print('Labels (confidence score):') pprint.pprint(vision_labels, indent=4) print('=' * 30) Explanation: Run our images through the function, and return a lookup table (reference) End of explanation print('\n\nLookup table:') pprint.pprint(imageslookup, indent=4) Explanation: View the results of our Vision analysis: End of explanation # Configure the sdfdownloadtasks.create request sdf_body = { 'version': SDF_VERSION_DV360, 'advertiserId': ADVERTISER_ID, 'parentEntityFilter': { 'fileType': ['FILE_TYPE_LINE_ITEM'], 'filterType': 'FILTER_TYPE_CAMPAIGN_ID', 'filterIds': [CAMPAIGN_ID] } } sdf_df = download_sdf(sdf_body) sdf_df.head() Explanation: Download the latest SDF LineItems (because we've made changes since our last download) End of explanation for product in imageslookup: sdf_df.loc[sdf_df.Name.str.contains(product), 'Keyword Targeting - Include'] = ';'.join( imageslookup[product]).lower() # Save modified dataframe to remote storage in Colab sdf_df.to_csv('sdf_update3_keywords.csv', index=False) sdf_df.head() # Download modified csv to local storage files.download('sdf_update3_keywords.csv') print("Success, see 'sdf_update3_keywords.csv' in your downloads folder") Explanation: Now we have our new labels from the Vision API, we need to write these into the keywords targeting field End of explanation # Define DV360 report definition (i.e. metrics and filters) report_definition = { 'params': { 'type': 'TYPE_GENERAL', 'metrics': [ 'METRIC_IMPRESSIONS', 'METRIC_CLICKS', 'METRIC_CTR', 'METRIC_REVENUE_ADVERTISER' ], 'groupBys': [ 'FILTER_ADVERTISER', 'FILTER_INSERTION_ORDER', 'FILTER_LINE_ITEM', 'FILTER_ADVERTISER_CURRENCY' ], 'filters': [{ 'type': 'FILTER_ADVERTISER', 'value': ADVERTISER_ID }], }, 'metadata': { 'title': 'DV360 Automation API-generated report', 'dataRange': 'LAST_90_DAYS', 'format': 'csv' }, 'schedule': { 'frequency': 'ONE_TIME' } } # Create new query using report definition operation = dbm_service.queries().createquery(body=report_definition).execute() pprint.pprint(operation) # Runs the given Queries.getquery request, retrying with an exponential # backoff. Returns completed operation. Will raise an exception if the # operation takes more than five hours to complete. @retry.Retry(predicate=retry.if_exception_type(Exception), initial=5, maximum=60, deadline=18000) def check_get_query_completion(getquery_request): response = getquery_request.execute() pprint.pprint(response) if response['metadata']['running']: raise Exception('The operation has not completed.') return response getquery_request = dbm_service.queries().getquery(queryId=operation['queryId']) response = check_get_query_completion(getquery_request) report_url = response['metadata']['googleCloudStoragePathForLatestReport'] # Use skipfooter to remove report footer from data report_df = pd.read_csv(report_url, skipfooter=16, engine='python') report_df.head(10) # Define our 'KPIs' ctr_target = 0.15 imp_threshold = 1000 # Convert IDs to remove decimal point, then string report_df['Line Item ID'] = report_df['Line Item ID'].apply(int) poor_performers = report_df.query( 'Impressions > @imp_threshold & (Clicks / Impressions)*100 < @ctr_target') # Convert results to Python list poor_performers = list(poor_performers['Line Item ID']) print(f'There are {len(poor_performers)} LineItems with a CTR' f' < {ctr_target}% and over {imp_threshold} impressions:' f'\n{poor_performers}') Explanation: Upload the output .csv file in the DV360 UI Once the changes have been applied successfully, check the 'Targeting' controls within 'Line Item details' 1.5 Optimisation using Reports Next, we'll look at how you could combine reporting data, with operations such as optimising bid multipliers or deactivating activity. Note: your new campaign has no performance history, so we'll use an existing campaign for this exercise. End of explanation # Configure the sdfdownloadtasks.create request sdf_body = { 'version': SDF_VERSION_DV360, 'advertiserId': ADVERTISER_ID, 'parentEntityFilter': { 'fileType': ['FILE_TYPE_LINE_ITEM'], 'filterType': 'FILTER_TYPE_CAMPAIGN_ID', 'filterIds': ['1914007'] } } sdf_df = download_sdf(sdf_body) sdf_df.head() for li in poor_performers: geo = sdf_df.loc[sdf_df['Line Item Id'] == li, 'Geography Targeting - Include'].iloc[0] sdf_df.loc[sdf_df['Line Item Id'] == li, 'Bid Multipliers'] = f'(geo; {geo} 0.5;);' # Save modified dataframe to remote storage in Colab sdf_df.to_csv('sdf_update4_bidmultipliers.csv', index=False) # Display updated DataFrame sdf_df.head() files.download('sdf_update4_bidmultipliers.csv') print('Success, your new SDF file has been downloaded') Explanation: Download an updated SDF LineItems file, and if the LineItem ID is in the poor performers list, add a Geo bid multiplier to half the bids (0.5) End of explanation #TODO Explanation: Upload the output .csv file in the DV360 UI Once the changes have been applied successfully, check the 'Targeting' controls within 'Line Item details' 1.6 Challenge Challenge: update your campaign with both language and audience targeting. All Lineitems should target the following Google audiences Affinity Categories » Technology » Mobile Enthusiasts Affinity Categories » Technology » Technophiles » High-End Computer Aficionado In-Market Categories » Consumer Electronics LineItems for France, should be targeted at French speakers LineItems for Great Britain, should be targeted at English speakers Tips Google Audience IDs can be found in the DV360 UI or by downloading an SDF with an existing audience applied Language IDs can be found in the Language.json ERF file or by downloading an SDF with the language already applied End of explanation # Format today-7 in required date format yesterday = (datetime.date.today() - datetime.timedelta(7)).strftime('%Y%m%d') # Download ERF for Language.json from public GCS bucket request = gcs_service.objects().get_media( bucket='gdbm-public', object='entity/' + yesterday + '.0.Language.json') response = request.execute() languages = json.loads(response) language_targets = ['en', 'fr'] lang_targeting_ids = {} # Search language.json for language targets 'en' and 'fr' for lang in languages: if lang['code'] in language_targets: lang_targeting_ids[lang['code']] = lang['id'] print(lang) print(lang_targeting_ids) # Define targeting template targeting_template = { 'Affinity & In Market Targeting - Include': '4569529;4586809;4497529;', } # Configure the sdfdownloadtasks.create request sdf_body = { 'version': SDF_VERSION_DV360, 'advertiserId': ADVERTISER_ID, 'parentEntityFilter': { 'fileType': ['FILE_TYPE_LINE_ITEM'], 'filterType': 'FILTER_TYPE_CAMPAIGN_ID', 'filterIds': [CAMPAIGN_ID] } } sdf_df = download_sdf(sdf_body) # Update DataFrame with Language and Audience targeting sdf_df.loc[sdf_df.Name.str.contains('GB'), 'Language Targeting - Include'] = f"{lang_targeting_ids['en']};" sdf_df.loc[sdf_df.Name.str.contains('FR'), 'Language Targeting - Include'] = f"{lang_targeting_ids['fr']};" sdf_df['Affinity & In Market Targeting - Include'] = targeting_template[ 'Affinity & In Market Targeting - Include'] # Save modified dataframe to remote storage in Colab sdf_df.to_csv('sdf_update5_challenge.csv', index=False) # Display updated DataFrame sdf_df.head() # Download file to disk using Colab syntax files.download('sdf_update5_challenge.csv') print("Success, see downloads folder for file 'sdf_update5_challenge.csv'") Explanation: Solution End of explanation request = display_video_service.advertisers().lineItems().list( advertiserId=ADVERTISER_ID, filter='entityStatus="ENTITY_STATUS_ACTIVE"', pageSize=1 ) response = request.execute() # Check if response is empty. if not response: print('Advertiser has no active Line Items') else: pprint.pprint(response['lineItems']) def get_active_lineitems(ADVERTISER_ID, CAMPAIGN_ID): Returns list of Lineitems with active status. list_lineitems = display_video_service.advertisers().lineItems().list( advertiserId=ADVERTISER_ID, filter=f'entityStatus="ENTITY_STATUS_ACTIVE" AND campaignId="{CAMPAIGN_ID}"', fields='lineItems(lineItemId,displayName)' # Return only two fields ).execute() active_lineitems = [li['lineItemId'] for li in list_lineitems['lineItems']] return active_lineitems Explanation: Upload the output .csv file in the DV360 UI 2) Display & Video 360 API What is the Display & Video 360 API? The Display & Video 360 API (formly known as the DV360 Write API) is the programmatic interface for the Display & Video 360 platform. It allows developers to easily and efficiently automate complex Display & Video 360 workflows, such as creating insertion orders and setting targeting options for individual line items. We'll use it now to build upon the campaign we created earlier using SDF. Reference: https://developers.google.com/display-video/api/reference/rest 2.1 Campaign builds Check Advertiser (ADVERTISER_ID) has active Lineitems End of explanation # Upload product feed using Colab's upload utility product_feed_csv = files.upload() contents = next(iter(product_feed_csv.values())).decode('utf-8') products = list(csv.DictReader(io.StringIO(contents))) # Create unique list of country-codes -- set() automatically de dupes unique_country_codes = set([row['country code'] for row in products]) print(unique_country_codes) Explanation: Upload the extended feed for Google Store's new territories: Spain, Germany and Portugal. End of explanation def create_insertion_order(parent_campaign_id, new_io_name): Creates a new DV360 insertion order object. # Define our new Insertion Order boilerplate new_insertion_order = { 'campaignId': parent_campaign_id, 'displayName': new_io_name, # Define naming convention 'entityStatus': 'ENTITY_STATUS_DRAFT', 'pacing': { 'pacingPeriod': 'PACING_PERIOD_DAILY', 'pacingType': 'PACING_TYPE_EVEN', 'dailyMaxMicros': '1000000' # Equiv to $1 or local currency }, 'frequencyCap': { 'unlimited': False, 'timeUnit': 'TIME_UNIT_MONTHS', 'timeUnitCount': 1, 'maxImpressions': 5 }, 'performanceGoal': { 'performanceGoalType': 'PERFORMANCE_GOAL_TYPE_CPC', 'performanceGoalAmountMicros': '1000000', # $1 CPM/CPC target }, 'bidStrategy': { 'fixedBid': { 'bidAmountMicros': '0' }, }, 'budget': { 'automationType': 'INSERTION_ORDER_AUTOMATION_TYPE_NONE', 'budgetUnit': 'BUDGET_UNIT_CURRENCY', 'budgetSegments': [{ 'budgetAmountMicros': '30000000', # Equiv to $30 or local currency 'description': 'My first segment', 'dateRange': { 'startDate': { 'year': year, 'month': month, 'day': day }, 'endDate': { 'year': year_plus30, 'month': month_plus30, 'day': day_plus30 } } }] } } # API create() request to generate new Insertion Order newinsertionorder_request = display_video_service.advertisers( ).insertionOrders().create( advertiserId=ADVERTISER_ID, body=new_insertion_order).execute() # Define patch to activate new Insertion Order afer creation patch = { 'entityStatus': 'ENTITY_STATUS_ACTIVE', } # API patch() request display_video_service.advertisers().insertionOrders().patch( advertiserId=ADVERTISER_ID, insertionOrderId=newinsertionorder_request['insertionOrderId'], updateMask='entityStatus', body=patch).execute() print(newinsertionorder_request) return newinsertionorder_request print('Insertion Order function created') Explanation: Create Insertion Order template Here we're defining a new a function called 'create_insertion_order'. Note: all new Insertion Orders and Line Items created using the DV360 API are created in 'Draft' mode (as a safety mechanism), and must be activated with a second API call, or via the UI (e.g. manually by a trader). End of explanation def create_lineitem(parent_io_id, new_li_name): Creates a new DV360 lineitem object. # Define our new LineItem boilerplate new_lineitem = { 'advertiserId': ADVERTISER_ID, 'insertionOrderId': parent_io_id, 'displayName': new_li_name, # Define naming convention 'lineItemType': 'LINE_ITEM_TYPE_DISPLAY_DEFAULT', 'entityStatus': 'ENTITY_STATUS_DRAFT', 'flight': { 'flightDateType': 'LINE_ITEM_FLIGHT_DATE_TYPE_INHERITED', }, 'pacing': { 'pacingPeriod': 'PACING_PERIOD_DAILY', 'pacingType': 'PACING_TYPE_EVEN', 'dailyMaxMicros': '1000000' }, 'frequencyCap': { 'timeUnit': 'TIME_UNIT_MONTHS', 'timeUnitCount': 1, 'maxImpressions': 5 }, 'partnerRevenueModel': { 'markupType': 'PARTNER_REVENUE_MODEL_MARKUP_TYPE_TOTAL_MEDIA_COST_MARKUP' }, 'budget': { 'budgetAllocationType': 'LINE_ITEM_BUDGET_ALLOCATION_TYPE_UNLIMITED', 'budgetUnit': 'BUDGET_UNIT_CURRENCY' }, 'bidStrategy': { 'fixedBid': { 'bidAmountMicros': '1000000' } } } # API create() request to generate new Lineitem newlineitem_request = display_video_service.advertisers().lineItems().create( advertiserId=ADVERTISER_ID, body=new_lineitem).execute() # Define patch to activate new Line Item afer creation patch = { 'entityStatus': 'ENTITY_STATUS_ACTIVE', } # API patch() request display_video_service.advertisers().lineItems().patch( advertiserId=ADVERTISER_ID, lineItemId=newlineitem_request['lineItemId'], updateMask='entityStatus', body=patch).execute() print(newlineitem_request) return newlineitem_request print('LineItem function created') Explanation: Create LineItem template Here we define a new function called 'create_lineitem', based on a template we specified. Note: the following template does not include any targeting controls by default. Normally, we strongly encourage the addition of targeting before activating a line item. End of explanation %%time for country_code in unique_country_codes: # Create() and patch() new Insertion Order io_name = f'Google Store | {country_code} | Display | Prospecting' insertionorder = create_insertion_order(CAMPAIGN_ID, io_name) for row in products: if country_code in row['country code']: # Create() and patch() new LineItem li_name = f"{row['country code']} | {row['title']} | {row['sku']}" lineitem = create_lineitem(insertionorder['insertionOrderId'], li_name) print('Process completed') Explanation: Build our new campaign First, we'll loop through the list of countries generated at the beginning, and for each country, create a new Insertion Order by calling our function 'create_insertion_order'. Within that loop, we find every product that is sold in the corresponding country-code, and create a new Line Item for every matching product using our function 'create_lineitem'. Sit tight, this one can take a while (~10 mins)... Link to DV360 UI End of explanation # Create the page token variable. next_page_token = '' while True: # Request the targeting options list. response = display_video_service.targetingTypes().targetingOptions().list( advertiserId=ADVERTISER_ID, targetingType='TARGETING_TYPE_BROWSER', pageToken=next_page_token).execute() # Check if response is empty. if not response: print('List request returned no Targeting Options') break # Iterate over retrieved targeting options. options_dict = {} for option in response['targetingOptions']: options_dict[ option['targetingOptionId']] = option['browserDetails']['displayName'] # Break out of loop if there is no next page. if 'nextPageToken' not in response: break # Update the next page token. next_page_token = response['nextPageToken'] pprint.pprint(options_dict) Explanation: If successful, the result should look similar to the below in DV360: 2.2 Individual targeting Reference: https://developers.google.com/display-video/api/guides/managing-line-items/targeting Retrieve a list of available targeting options using targetingTypes().targetingOptions() The following example demonstrates retrieving of Browser targeting options only. The "BrowserDetails" field is only applicable with "TARGETING_TYPE_BROWSER". End of explanation # Return list of Lineitems with active status active_lineitems = get_active_lineitems(ADVERTISER_ID, CAMPAIGN_ID) # Fetch first Lineitem ID lineitem_id = active_lineitems[0] # Create a assigned targeting option object. assigned_targeting_option_obj = { 'browserDetails': { 'targetingOptionId': '500072' } } # Create the assigned targeting option. assigned_targeting_option = display_video_service.advertisers().lineItems( ).targetingTypes().assignedTargetingOptions().create( advertiserId=ADVERTISER_ID, lineItemId=f'{lineitem_id}', targetingType='TARGETING_TYPE_BROWSER', body=assigned_targeting_option_obj ).execute() # Display the new assigned targeting option. print(f"Assigned Targeting Option {assigned_targeting_option['name']} created.") Explanation: Apply individual targeting criteria to single entity End of explanation # Create the page token variable. next_page_token = '' while True: # Request the targeting options list. response = display_video_service.googleAudiences().list( advertiserId=ADVERTISER_ID, filter='displayName : "Technology"', pageToken=next_page_token).execute() # Check if response is empty. if not response: print('List request returned no Targeting Options') break # Iterate over retrieved targeting options. options_dict = {} for option in response['googleAudiences']: options_dict[option['googleAudienceId']] = [ option['displayName'], option['googleAudienceType'] ] # Break out of loop if there is no next page. if 'nextPageToken' not in response: break # Update the next page token. next_page_token = response['nextPageToken'] pprint.pprint(response) google_audience_id = '92948' # Return list of Lineitems with active status active_lineitems = get_active_lineitems(ADVERTISER_ID, CAMPAIGN_ID) # Create a assigned targeting option object. assigned_targeting_option_obj = { 'audienceGroupDetails': { 'includedGoogleAudienceGroup': { 'settings': [{ 'googleAudienceId': f'{google_audience_id}' }] } } } pprint.pprint(assigned_targeting_option_obj) # Update bulk targeting for li in active_lineitems: # Create the assigned targeting option. assigned_targeting_option = display_video_service.advertisers().lineItems( ).targetingTypes().assignedTargetingOptions().create( advertiserId=ADVERTISER_ID, lineItemId=f'{li}', targetingType='TARGETING_TYPE_AUDIENCE_GROUP', body=assigned_targeting_option_obj).execute() # Display the new assigned targeting option. print(f"Targeting Option {assigned_targeting_option['name']} created.") Explanation: Applying individual targeting criteria to multiple entities End of explanation def set_default_li_targeting(lineitem_id): Sets default LineItem targeting according to standard template. # Define 'Channels' create_channel_assigned_targetingoptions = [] for targeting_id in ['1777746835', '1778039430']: create_channel_assigned_targetingoptions.append( {'channelDetails': { 'channelId': targeting_id, 'negative': False }}) # Define 'Inventory' create_inventory_assigned_targetingoptions = [] for targeting_id in ['1']: create_inventory_assigned_targetingoptions.append( {'inventorySourceDetails': {'inventorySourceId': targeting_id}} ) # Define 'Sensitive categories' create_sensitive_cat_assigned_targetingoptions = [] sensitive_category = [ '1163177997', '1163178297', '118521027123', '118521027843', '118521028083', '118521028563', '118521028803', '1596254697' ] for targeting_id in sensitive_category: create_sensitive_cat_assigned_targetingoptions.append({ 'sensitiveCategoryExclusionDetails': { 'excludedTargetingOptionId': targeting_id } }) # Define 'Digital content labels' create_digital_content_assigned_targetingoptions = [] content_rating_tier = ['19875634320', '19875634200', '19875634080'] for targeting_id in content_rating_tier: create_digital_content_assigned_targetingoptions.append({ 'digitalContentLabelExclusionDetails': { 'excludedTargetingOptionId': targeting_id } }) # Contruct request bulk_edit_line_item_request = { 'createRequests': [ { 'targetingType': 'TARGETING_TYPE_CHANNEL', 'assignedTargetingOptions': [ create_channel_assigned_targetingoptions ] }, { 'targetingType': 'TARGETING_TYPE_INVENTORY_SOURCE', 'assignedTargetingOptions': [ create_inventory_assigned_targetingoptions ] }, { 'targetingType': 'TARGETING_TYPE_SENSITIVE_CATEGORY_EXCLUSION', 'assignedTargetingOptions': [ create_sensitive_cat_assigned_targetingoptions ] }, { 'targetingType': 'TARGETING_TYPE_DIGITAL_CONTENT_LABEL_EXCLUSION', 'assignedTargetingOptions': [ create_digital_content_assigned_targetingoptions ] }, ] } # Edit the line item targeting. bulk_request = display_video_service.advertisers().lineItems( ).bulkEditLineItemAssignedTargetingOptions( advertiserId=ADVERTISER_ID, lineItemId=lineitem_id, body=bulk_edit_line_item_request ) bulk_response = bulk_request.execute() # Check if response is empty. # If not, iterate over and display new assigned targeting options. if not bulk_response: print('Bulk edit request created no new AssignedTargetingOptions') else: for assigned_targeting_option in bulk_response[ 'createdAssignedTargetingOptions']: print(f"Targeting Option {assigned_targeting_option['name']} created.") print('Lineitem targeting function created') Explanation: 2.3 Bulk targeting Bulk updates using templated targeting controls End of explanation # Return list of Lineitems with active status active_lineitems = get_active_lineitems(ADVERTISER_ID, CAMPAIGN_ID) # Update bulk targeting for li in active_lineitems: set_default_li_targeting(li) Explanation: Retrieve list of active LineItems, and Apply bulk targeting End of explanation def optimise_lineitem(lineitem_id, action): Optimises lineitem according to given parameter. lineitem_object = display_video_service.advertisers().lineItems().get( advertiserId=ADVERTISER_ID, lineItemId=lineitem_id).execute() if lineitem_object['entityStatus'] == 'ENTITY_STATUS_ACTIVE': if action == 'pause': patch = { 'entityStatus': 'ENTITY_STATUS_PAUSED', } lineitem_patched = display_video_service.advertisers().lineItems().patch( advertiserId=ADVERTISER_ID, lineItemId=lineitem_id, updateMask='entityStatus', body=patch).execute() print(f"LineItemID {lineitem_patched['name']} was paused") elif action == 'optimise': patch = {'bidStrategy': {'fixedBid': {'bidAmountMicros': '500000'},}} lineitem_patched = display_video_service.advertisers().lineItems().patch( advertiserId=ADVERTISER_ID, lineItemId=lineitem_id, updateMask='bidStrategy', body=patch).execute() print(f"{lineitem_patched['name']} was optimised") else: print("Not a valid action, must be either 'pause' or 'optimise'") else: print( f"{lineitem_object['name']} already paused/archived - no action taken") print('Optimisation function created') Explanation: 2.4 Optimisation (external trigger) The following optimisations will be completed on your campaign, created earlier. Create functions to 'deactivate' or 'optimise' Lineitems End of explanation out_of_stock_list = [] products = csv.DictReader(io.StringIO(contents)) # Iterate through each row, checking for products where availability = 0 for row in products: if row['availability'] == '0': out_of_stock_list.append(row['sku']) # This should generate a list of 9 SKUs that are no-longer in stock print( f'Found {len(out_of_stock_list)} out-of-stock products {out_of_stock_list}') Explanation: Creat list of out of stock products End of explanation # Return list of Lineitems with active status active_lineitems = get_active_lineitems(ADVERTISER_ID, CAMPAIGN_ID) # Iterate through out-of-stock list. If sku is found in lineitem's name, perform optimisation. for product in out_of_stock_list: for key, value in active_lineitems.items(): if product in key: optimise_lineitem(value, 'pause') Explanation: Process optimisation End of explanation # Define DV360 report definition (i.e. metrics and filters) report_definition = { 'params': { 'type': 'TYPE_GENERAL', 'metrics': [ 'METRIC_IMPRESSIONS', 'METRIC_CLICKS', 'METRIC_CTR', 'METRIC_REVENUE_ADVERTISER' ], 'groupBys': [ 'FILTER_ADVERTISER', 'FILTER_INSERTION_ORDER', 'FILTER_LINE_ITEM', 'FILTER_ADVERTISER_CURRENCY' ], 'filters': [{ 'type': 'FILTER_ADVERTISER', 'value': ADVERTISER_ID }], }, 'metadata': { 'title': 'DV360 Automation API-generated report', 'dataRange': 'LAST_90_DAYS', 'format': 'csv' }, 'schedule': { 'frequency': 'ONE_TIME' } } # Create new query using report definition operation = dbm_service.queries().createquery(body=report_definition).execute() pprint.pprint(operation) # Runs the given Queries.getquery request, retrying with an exponential # backoff. Returns completed operation. Will raise an exception if the # operation takes more than five hours to complete. @retry.Retry( predicate=retry.if_exception_type(Exception), initial=5, maximum=60, deadline=18000) def check_get_query_completion(getquery_request): response = getquery_request.execute() pprint.pprint(response) if response['metadata']['running']: raise Exception('The operation has not completed.') return response getquery_request = dbm_service.queries().getquery(queryId=operation['queryId']) response = check_get_query_completion(getquery_request) Explanation: 2.5 Optimisation (reporting data) As your new campaign has no performance data, the following optimisations will be completed on an existing campaign with historical data. Create new performance report and fetch results End of explanation # Capture report URL from response report_url = response['metadata']['googleCloudStoragePathForLatestReport'] # Use skipfooter to remove report footer from data report_df = pd.read_csv(report_url, skipfooter=16, engine='python') report_df.head(10) Explanation: Load report to Pandas DataFrame End of explanation # Define our 'KPIs' ctr_to_pause = 0.1 ctr_to_optimise = 0.3 imp_threshold = 5000 # Convert IDs to remove decimal point, then string report_df['Line Item ID'] = report_df['Line Item ID'].apply(int) lineitems_to_pause = report_df.query('Impressions > @imp_threshold and (Clicks / Impressions)*100 < @ctr_to_pause') lineitems_to_reducebid = report_df.query('Impressions > @imp_threshold and (Clicks / Impressions)*100 > @ctr_to_pause < @ctr_to_optimise') # Convert results to Python list lineitems_to_pause = list(lineitems_to_pause['Line Item ID']) lineitems_to_reducebid = list(lineitems_to_reducebid['Line Item ID']) print(f'Found {len(lineitems_to_pause)} LineItems with a CTR' f'< {ctr_to_pause}% and > {imp_threshold} impressions:' f'{lineitems_to_pause}') print(f'Found {len(lineitems_to_reducebid)} LineItems with a CTR' f' between {ctr_to_pause}%-{ctr_to_optimise}%, and > {imp_threshold}' f'\n impressions: {lineitems_to_reducebid}') Explanation: Create two lists of poorly performing LineItems 1. LineItems that should be paused 2. Lineitems to reduce bids End of explanation %%time if lineitems_to_pause: for lineitem in lineitems_to_pause: optimise_lineitem(str(lineitem), 'pause') if lineitems_to_reducebid: for lineitem in lineitems_to_reducebid: optimise_lineitem(str(lineitem), 'optimise') print('Optimisation completed') Explanation: Process optimisation End of explanation def upload_creative_image_asset(asset_url, click_url): Creates a new DV360 creative object. # Fetch asset from cloud storage using requests library asset = requests.get(asset_url) # Create upload object from http image url fh = io.BytesIO(asset.content) media_body = http.MediaIoBaseUpload(fh, mimetype='image/png', chunksize=1024*1024, resumable=True) # Extract filename from url path filename = str(asset_url.rsplit(sep='/', maxsplit=1)[1]) # Create the request body body = {'filename': filename} # Upload the asset asset_request = display_video_service.advertisers().assets().upload( advertiserId=ADVERTISER_ID, body=body, media_body=media_body).execute() # Display the new asset media ID print(f"Asset was created with media ID {asset_request['asset']['mediaId']}") display_name = f'{filename}'.split(sep='.')[0].lower() + ' 300x250' # Create a creative object. creative_obj = { 'displayName': f'{display_name}', 'entityStatus': 'ENTITY_STATUS_ACTIVE', 'creativeType': 'CREATIVE_TYPE_STANDARD', 'hostingSource': 'HOSTING_SOURCE_HOSTED', 'dimensions': { 'widthPixels': 300, 'heightPixels': 250 }, 'assets': [{ 'asset': { 'mediaId': asset_request['asset']['mediaId'] }, 'role': 'ASSET_ROLE_MAIN' }], 'exitEvents': [{ 'type': 'EXIT_EVENT_TYPE_DEFAULT', 'url': f'{click_url}', }] } creative_request = display_video_service.advertisers().creatives().create( advertiserId=ADVERTISER_ID, body=creative_obj ).execute() # Display the new creative ID print(f"Creative was created with ID {creative_request['creativeId']}" f" and DisplayName '{creative_request['displayName']}'") pprint.pprint(creative_request) print('Creative upload function defined') Explanation: 2.6 Creative upload Uploading Display creatives from remote storage (http) The following demonstrates how to upload image assets from remote storage, but it's also possible to upload from local storage. Reference: https://developers.google.com/display-video/api/guides/creating-creatives/overview End of explanation image_assets = { 'https://github.com/google/dv360-automation/blob/master/docs/images/googlestore/pixelbook.png?raw=true': 'https://store.google.com/product/google_pixelbook', 'https://github.com/google/dv360-automation/blob/master/docs/images/googlestore/googlehome.png?raw=true': 'https://store.google.com/product/google_home_hub', 'https://github.com/google/dv360-automation/blob/master/docs/images/googlestore/googlehomemini.png?raw=true': 'https://store.google.com/product/google_home_mini', 'https://github.com/google/dv360-automation/blob/master/docs/images/googlestore/pixel2.png?raw=true': 'https://store.google.com/product/pixel_2', 'https://github.com/google/dv360-automation/blob/master/docs/images/googlestore/chromecastultra.png?raw=true': 'https://store.google.com/product/chromecast_ultra' } for asset, click_url in image_assets.items(): upload_creative_image_asset(asset, click_url) Explanation: Upload image creatives Note, all of the following assets are the same dimension (300x250) and type 'CREATIVE_TYPE_STANDARD'. When uploading assets of multiple sizes, the creatives.create body must reflect this. End of explanation #TODO Explanation: 2.7 Challenge Challenge: build a new campaign for 'Google Airways' using the flights feed provided here. Tips You don't need to rewrite any functions, reuse the existing ones Don't forget to use print() statements to see progress within a for loop Your final campaign should look similar to the below: End of explanation %%time # Load flight information from CSV file googleairways_routes = files.upload() contents = next(iter(googleairways_routes.values())).decode('utf-8') routes = list(csv.DictReader(io.StringIO(contents))) # Create a unique set (de-duped) of cities from the routes provided unique_cities = set() for row in routes: unique_cities.add(row['airport-city']) print(unique_cities) # Create Campaign and Patch() new_campaign = create_campaign('Google Airways') print(new_campaign) # Step through each city within our unique set of cities for city in unique_cities: # Create Insertion Order and Patch() io_name = f'Flights | {city}' create_io = create_insertion_order(new_campaign['campaignId'], io_name) # Step through each route(row) of the CSV upload for row in routes: if city == row['airport-city']: # Create LineItems and Patch() li_name = f"Flight {row['flightno']} | {row['depairport-city']} to {row['arrairport-city']}" create_lis = create_lineitem(create_io['insertionOrderId'], li_name) print('Process completed') Explanation: Solution End of explanation # Exclude following campaigns in the reset process protected_campaigns = ['1914007','985747'] def reset_demo_account(): Reset DV360 account to earlier state. print('Resetting DV360 account...') # Reactivate Campaigns list_campaigns = display_video_service.advertisers().campaigns().list( advertiserId=ADVERTISER_ID, filter='entityStatus="ENTITY_STATUS_ACTIVE"').execute() results = list_campaigns['campaigns'] print(f'Found {len(results)} active campaigns') for index, campaign in enumerate(results, start=1): print(f'Campaign {index} of {len(results)}') pause_campaign(campaign['campaignId']) # Reactivate LineItems list_lineitems = display_video_service.advertisers().lineItems().list( advertiserId=ADVERTISER_ID, filter='entityStatus="ENTITY_STATUS_PAUSED" AND campaignId="1914007"' ).execute() if not list_lineitems: print('No paused lineitems found') else: for index, li in enumerate(list_lineitems['lineItems'], start=1): print(f"Lineitem {index} of {len(list_lineitems['lineItems'])}") lineitem_id = li['lineItemId'] activate_lineitem(lineitem_id) print('Account reset completed') def delete_campaign(campaign_id): Updates DV360 campaign object status to deleted. if campaign_id in protected_campaigns: print(f'Campaign ID {campaign_id} not deleted (protected campaign)') else: try: display_video_service.advertisers().campaigns().delete( advertiserId=ADVERTISER_ID, campaignId=campaign_id).execute() print(f'{campaign_id} successfully deleted') except Exception: print('Could not delete campaign') def archive_campaign(campaign_id): Updates DV360 campaign object status to archived. patch = {'entityStatus': 'ENTITY_STATUS_ARCHIVED'} if campaign_id in protected_campaigns: print(f'Campaign ID {campaign_id} not archived (protected campaign)') else: archive_campaign = display_video_service.advertisers().campaigns().patch( advertiserId=ADVERTISER_ID, campaignId=campaign_id, updateMask='entityStatus', body=patch).execute() print(f'Campaign ID {campaign_id} successfully archived') def pause_campaign(campaign_id): Updates DV360 campaign object status to paused. patch = {'entityStatus': 'ENTITY_STATUS_PAUSED'} if campaign_id in protected_campaigns: print(f'Campaign ID {campaign_id} not paused (protected campaign)') else: display_video_service.advertisers().campaigns().patch( advertiserId=ADVERTISER_ID, campaignId=campaign_id, updateMask='entityStatus', body=patch).execute() print(f'Campaign ID {campaign_id} successfully paused') def activate_lineitem(lineitem_id): Updates DV360 lineitem object status to active. patch = {'entityStatus': 'ENTITY_STATUS_ACTIVE'} display_video_service.advertisers().lineItems().patch( lineItemId=lineitem_id, advertiserId=ADVERTISER_ID, updateMask='entityStatus', body=patch).execute() print(f'Lineitem ID {lineitem_id} reactivated') # @title { display-mode: "form" } #@markdown Reset DV360 account # Call main function to intialise reset procedure reset_demo_account() Explanation: Link to DV360 UI Resources Getting started with SDF in DV360 guide Structured Data Files (SDF) developer guide Getting started with the Display & Video 360 API developer guide Getting started with the DoubleClick Bid Manager API developer guide How to access Entity Read Files Quickstart: Setup the Vision API Please help us improve this workshop by completing the satisfaction survey Thank you! Clean up To clean up all of the DV360 resources used during these exercises, you can run the following script. Warning: this will remove all Campaigns from the DV360 advertiser specified in ADVERTISER_ID, unless they are explicitly defined as a 'protected_campaign' End of explanation <END_TASK>
15,952
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Graphical User Interfaces Object oriented programming and particularly inheritance is commonly used for creating GUIs. There are a large number of different frameworks supporting building GUIs. The following are particularly relevant Step1: Although this works, it is visually unappealing. We can improve on this using styles and themes. Step2: As our applications get more complicated we must give greater thought to the layout. The following example comes from the TkDocs site. Step4: Matplotlib For simple programs, displaying data and taking basic input, often a command line application will be much faster to implement than a GUI. The times when I have moved away from the command line it has been to interact with image data and plots. Here, matplotlib often works very well. Either it can be embedded in a larger application or it can be used directly. There are a number of examples on the matplotlib site. Here is one stripped down example of one recent GUI I have used.
<ASSISTANT_TASK:> Python Code: import tkinter as tk class Application(tk.Frame): def __init__(self, master=None): tk.Frame.__init__(self, master) self.pack() self.createWidgets() def createWidgets(self): self.hi_there = tk.Button(self) self.hi_there["text"] = "Hello World\n(click me)" self.hi_there["command"] = self.say_hi self.hi_there.pack(side="top") self.QUIT = tk.Button(self, text="QUIT", fg="red", command=root.destroy) self.QUIT.pack(side="bottom") def say_hi(self): print("hi there, everyone!") root = tk.Tk() app = Application(master=root) app.mainloop() Explanation: Graphical User Interfaces Object oriented programming and particularly inheritance is commonly used for creating GUIs. There are a large number of different frameworks supporting building GUIs. The following are particularly relevant: TkInter - This is the official/default GUI framework guidata - A GUI framework for dataset display and editing VTK - A GUI framework for data visualization pyqtgraph - A GUI framework for data visualization, easily installed with conda install pyqtgraph matplotlib - As well as creating plots matplotlib can support interaction TkInter TkInter is widely used with plenty of documentation available but may prove somewhat limited for more data intensive applications. Documentation from the standard library Further documentation from python.org TkDocs Let's look at a simple example from the documentation End of explanation import tkinter as tk from tkinter import ttk class Application(ttk.Frame): def __init__(self, master=None): super().__init__(master, padding="3 3 12 12") self.grid(column=0, row=0, ) self.createWidgets() self.master.title('Test') def createWidgets(self): self.hi_there = ttk.Button(self) self.hi_there["text"] = "Hello World\n(click me)" self.hi_there["command"] = self.say_hi self.QUIT = ttk.Button(self, text="QUIT", style='Alert.TButton', command=root.destroy) for child in self.winfo_children(): child.grid_configure(padx=10, pady=10) def say_hi(self): print("hi there, everyone!") root = tk.Tk() app = Application(master=root) s = ttk.Style() s.configure('TButton', font='helvetica 24') s.configure('Alert.TButton', foreground='red') root.mainloop() Explanation: Although this works, it is visually unappealing. We can improve on this using styles and themes. End of explanation from tkinter import * from tkinter import ttk def calculate(*args): try: value = float(feet.get()) meters.set((0.3048 * value * 10000.0 + 0.5)/10000.0) except ValueError: pass root = Tk() root.title("Feet to Meters") mainframe = ttk.Frame(root, padding="3 3 12 12") mainframe.grid(column=0, row=0, sticky=(N, W, E, S)) mainframe.columnconfigure(0, weight=1) mainframe.rowconfigure(0, weight=1) feet = StringVar() meters = StringVar() feet_entry = ttk.Entry(mainframe, width=7, textvariable=feet) feet_entry.grid(column=2, row=1, sticky=(W, E)) ttk.Label(mainframe, textvariable=meters).grid(column=2, row=2, sticky=(W, E)) ttk.Button(mainframe, text="Calculate", command=calculate).grid(column=3, row=3, sticky=W) ttk.Label(mainframe, text="feet").grid(column=3, row=1, sticky=W) ttk.Label(mainframe, text="is equivalent to").grid(column=1, row=2, sticky=E) ttk.Label(mainframe, text="meters").grid(column=3, row=2, sticky=W) for child in mainframe.winfo_children(): child.grid_configure(padx=5, pady=5) feet_entry.focus() root.bind('<Return>', calculate) root.mainloop() Explanation: As our applications get more complicated we must give greater thought to the layout. The following example comes from the TkDocs site. End of explanation Do a mouseclick somewhere, move the mouse to some destination, release the button. This class gives click- and release-events and also draws a line or a box from the click-point to the actual mouseposition (within the same axes) until the button is released. Within the method 'self.ignore()' it is checked wether the button from eventpress and eventrelease are the same. from matplotlib.widgets import RectangleSelector import matplotlib.pyplot as plt import matplotlib.cbook as cbook def line_select_callback(eclick, erelease): 'eclick and erelease are the press and release events' x1, y1 = eclick.xdata, eclick.ydata x2, y2 = erelease.xdata, erelease.ydata print ("(%3.2f, %3.2f) --> (%3.2f, %3.2f)" % (x1, y1, x2, y2)) print (" The button you used were: %s %s" % (eclick.button, erelease.button)) def toggle_selector(event): print (' Key pressed.') if event.key in ['Q', 'q'] and toggle_selector.RS.active: print (' RectangleSelector deactivated.') toggle_selector.RS.set_active(False) if event.key in ['A', 'a'] and not toggle_selector.RS.active: print (' RectangleSelector activated.') toggle_selector.RS.set_active(True) image_file = cbook.get_sample_data('grace_hopper.png') image = plt.imread(image_file) fig, current_ax = plt.subplots() plt.imshow(image) toggle_selector.RS = RectangleSelector(current_ax, line_select_callback, drawtype='box', useblit=True, button=[1,3], # don't use middle button minspanx=5, minspany=5, spancoords='pixels') plt.connect('key_press_event', toggle_selector) plt.show() Explanation: Matplotlib For simple programs, displaying data and taking basic input, often a command line application will be much faster to implement than a GUI. The times when I have moved away from the command line it has been to interact with image data and plots. Here, matplotlib often works very well. Either it can be embedded in a larger application or it can be used directly. There are a number of examples on the matplotlib site. Here is one stripped down example of one recent GUI I have used. End of explanation <END_TASK>
15,953
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Basics of NUmpy & Pandas Numpy Numpy uses array whereas pandas used scaler <br /> Step1: Array are similar to python list , but it all element must be of same data type, and it faster than list Step2: Lets see some of functionality Step3: Vector Operation Step4: vector [+-*/] Scaler Step5: vector & boolean vector Step6: num[bl],, what it will return ?? <h4>It return array of values corresponding to which elemnt in bl is True Step7: find all elemnt greter than 100 from num Step8: <h5> All element less than 50 ?? Step9: In-place operation in numpay (Diff between += and +) Step10: <h5>this happen becouse a and b both point to same array and + operation create a new array and then a point to that so b remain unaffected" </h5> Step11: Pandas Series <h4> Basics are same as numpy array but pandas series also contain lots of functionality and speciality Step12: <h6>See All basic results using describe() function
<ASSISTANT_TASK:> Python Code: import numpy as np Explanation: Basics of NUmpy & Pandas Numpy Numpy uses array whereas pandas used scaler <br /> End of explanation num = np.array([3,4,2,5,7,23,56,23,7,23,89,43,676,43]) num Explanation: Array are similar to python list , but it all element must be of same data type, and it faster than list End of explanation print "Mean :",num.mean() print "sum :",num.sum() print "max :",num.max() print "std :",num.std() #slicing num[:5] #find index of any element let say max print "index of max :",num.argmax() print "data Type of array :",num.dtype Explanation: Lets see some of functionality End of explanation a=np.array([5,6,15]) b=np.array([5,4,-5]) # Addition print "{} + {} = {}".format(a,b,a+b) print "{} * {} = {}".format(a,b,a*b) print "{} / {} = {}".format(a,b,a/b) # If size mismatch then error occure b=np.array([5,4,-5,5]) print "{} + {} = {}".format(a,b,a+b) Explanation: Vector Operation End of explanation print "{} + {} = {}".format(a,3,a+3) print "{} * {} = {}".format(a,3,a*3) print "{} / {} = {}".format(a,3,a/3) Explanation: vector [+-*/] Scaler End of explanation num=np.array([5,6,15,65,32,656,23,435,2,45,21]) bl=np.array([False,True,True,False,True,False,True,False,True,True,False]) num[6] Explanation: vector & boolean vector End of explanation num[bl] Explanation: num[bl],, what it will return ?? <h4>It return array of values corresponding to which elemnt in bl is True End of explanation num[num>100] Explanation: find all elemnt greter than 100 from num End of explanation num[num<50] Explanation: <h5> All element less than 50 ?? End of explanation a=np.array([5,6,15]) b=a a += 2 print b print "this happen becouse a and b both point to same array and += is In-place operation so it maintain that" a=np.array([5,6,15]) b=a a = a + 2 print b Explanation: In-place operation in numpay (Diff between += and +) End of explanation a=np.array([5,6,15]) b=a[:3] b[0]=1000 print a,"Reason is similar as +=" Explanation: <h5>this happen becouse a and b both point to same array and + operation create a new array and then a point to that so b remain unaffected" </h5> End of explanation import pandas as pd num = pd.Series([3,4,2,5,7,23,56,23,7,23,89,43,676,43]) num Explanation: Pandas Series <h4> Basics are same as numpy array but pandas series also contain lots of functionality and speciality End of explanation num.describe() Explanation: <h6>See All basic results using describe() function End of explanation <END_TASK>
15,954
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: NoSQL (MongoDB) (Ejemplo de Replica Set y Sharding) Esta hoja muestra cómo crear un clúster (simulado con docker-compose donde se mantienen dos replica sets (conjuntos de réplicas) y una colección (posts) que se divide en cada una de esas dos replica sets. Al final se indica cómo escalar y añadir más hosts a las replica sets. Step1: Usaremos la librería pymongo para python. La cargamos a continuación. Step2: La conexión se inicia con MongoClient en el host descrito en el fichero docker-compose.yml (mongo). Step3: Importación de los ficheros CSV. Por ahora creamos una colección diferente para cada uno. Después estudiaremos cómo poder optimizar el acceso usando agregación. Step4: Creamos el índice por el que se creará el shard. Por defecto es de tipo ASCENDING. Step5: Añadimos al shard ambos ordenadores de ambos replica sets, y activamos el sharding en la tabla posts. Step6:
<ASSISTANT_TASK:> Python Code: !pip install --upgrade pymongo from pprint import pprint as pp import pandas as pd import matplotlib.pyplot as plt import matplotlib %matplotlib inline matplotlib.style.use('ggplot') %%bash sudo apt-get update sudo apt-get install -y mongodb-clients Explanation: NoSQL (MongoDB) (Ejemplo de Replica Set y Sharding) Esta hoja muestra cómo crear un clúster (simulado con docker-compose donde se mantienen dos replica sets (conjuntos de réplicas) y una colección (posts) que se divide en cada una de esas dos replica sets. Al final se indica cómo escalar y añadir más hosts a las replica sets. End of explanation %%bash cat << END | mongo --host mongo_mongocfg1_1 rs.initiate(); // Iniciar el replica set de los servidores de configuración sleep(1000); rs.status() END %%bash cat <<END | mongo --host mongo_mongors1_1 rs.initiate(); // Inicio del RS sleep(1000); cfg = rs.conf(); cfg.members[0].host = "mongo_mongors1_1"; // Cambiar el host porque docker coge el hostname del contenedor rs.reconfig(cfg); // Reconfigurar //rs.add(host2/3) // Añadir posibles nuevos hosts añadidos al RS rs.status() END %%bash cat <<END | mongo --host mongo_mongors2_1 rs.initiate(); sleep(1000); cfg = rs.conf(); cfg.members[0].host = "mongo_mongors2_1"; rs.reconfig(cfg); //rs.add(host2/3) rs.status() END import pymongo from pymongo import MongoClient Explanation: Usaremos la librería pymongo para python. La cargamos a continuación. End of explanation client = MongoClient("mongo_mongors1_1",27017) client client.database_names() %%bash file=../Posts.csv test -e $file || wget http://neuromancer.inf.um.es:8080/es.stackoverflow/`basename ${file}`.gz -O - 2>/dev/null | gunzip > $file db = client.stackoverflow db = client['stackoverflow'] db posts = db.posts posts Explanation: La conexión se inicia con MongoClient en el host descrito en el fichero docker-compose.yml (mongo). End of explanation %%bash mongoimport --db stackoverflow --collection posts --drop --type csv \ --headerline --host=mongo_mongors1_1 --file ../Posts.csv posts.count() Explanation: Importación de los ficheros CSV. Por ahora creamos una colección diferente para cada uno. Después estudiaremos cómo poder optimizar el acceso usando agregación. End of explanation posts.create_index('Id') Explanation: Creamos el índice por el que se creará el shard. Por defecto es de tipo ASCENDING. End of explanation %%bash cat <<END | mongo --host mongo_mongos_1 sh.addShard("mongors1/mongo_mongors1_1:27017"); sh.addShard("mongors2/mongo_mongors2_1:27017"); sh.enableSharding("stackoverflow"); sh.shardCollection('stackoverflow.posts', {'Id': 1}) END %%bash cat <<END | mongo --host mongo_mongos_1 sh.status() END Explanation: Añadimos al shard ambos ordenadores de ambos replica sets, y activamos el sharding en la tabla posts. End of explanation post = posts.find_one() post posts.find({"PostTypeId": 2}).explain() Explanation: End of explanation <END_TASK>
15,955
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Using a Pre-trained PyTorch Model for Inference In this demo, we will use a pre-trained model to perform inference on a single image. There are 3 components to this demo Step1: Model Step2: Input Step3: Output Step4: Human
<ASSISTANT_TASK:> Python Code: import torch import torchvision import torchvision.transforms as transforms import timm from einops import rearrange from PIL import Image Explanation: Using a Pre-trained PyTorch Model for Inference In this demo, we will use a pre-trained model to perform inference on a single image. There are 3 components to this demo: 1. Input 2. Model 3. Output We will cover these components in detail below. Let us first import the required packages. End of explanation use_timm = False # Download and load the pretrained ResNet-18. if use_timm: resnet = timm.create_model('resnet18', pretrained=True) else: resnet = torchvision.models.resnet18(pretrained=True) resnet.eval() Explanation: Model: Loading a pre-trained ResNet18 model We use a pre-trained ResNet18 model for inference. The model is available from torchvision or from timm. When we use a model for inference, we need to specify the eval mode. This is because the model in train mode by default, and we need to disable all the dropout layers. End of explanation filename = input() # Load a PIL Image given a file name from the current directory. img = Image.open(filename) # Display the loaded image on notebook. display(img) # Resize the image to 256x256. # Then crop the center square of the image. # Next, convert the image to a PyTorch Tensor. # Lastly, normalize the image so that it has mean and standard deviation as shown below. # Reference for image transforms: https://github.com/pytorch/examples/blob/42e5b996718797e45c46a25c55b031e6768f8440/imagenet/main.py#L89-L101 normalize = transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]) transform = transforms.Compose([ transforms.Resize(256), transforms.CenterCrop(224), transforms.ToTensor(), normalize, ]) # PIL image undergoes transforms. img = transform(img) # A simplified version is to simply transform the image to a tensor #img = transforms.ToTensor()(img) # Check img type and shape print("Type:", img.dtype) print("Shape:", img.shape) Explanation: Input: Loading an input image We can use matplotlib image to load an image into a numpy array. However, PyTorch transforms expects a PIL image. While we can convert numpy array to PIL, we can load an image directly into a PIL image. End of explanation # We need the tensor to have a batch dimension of 1. img = rearrange(img, 'c h w -> 1 c h w') print("New shape:", img.shape) with torch.no_grad(): pred = resnet(img) print("Prediction shape:", pred.shape) pred = torch.argmax(pred, dim=1) print("Predicted index", pred) Explanation: Output: Making a prediction We will now use img tensor as input to the pre-trained resnet18 model. Before running the model for prediction, there are 2 things that we should do: Include a batch dimension. In this case, we are using a single image, so we need to add a batch size of 1. We use rearrange for this. Execute inference within torch.no_grad() context manager. This is because we do not want to track the gradients. The expected output is a torch.Tensor of shape (1, 1000). resnet18 was pre-trained on ImageNet1k. We can use torch.argmax to get the index of the maximum value. End of explanation import urllib filename = "imagenet1000_labels.txt" url = "https://gist.githubusercontent.com/yrevar/942d3a0ac09ec9e5eb3a/raw/238f720ff059c1f82f368259d1ca4ffa5dd8f9f5/imagenet1000_clsidx_to_labels.txt" # Download the file if it does not exist if not os.path.isfile(filename): urllib.request.urlretrieve(url, filename) with open(filename) as f: idx2label = eval(f.read()) print("Predicted label:", idx2label[pred.cpu().numpy()[0]]) Explanation: Human: Convert class index to label To make sense of the predicted index, we need to convert it to a label. We can use https://gist.github.com/yrevar/942d3a0ac09ec9e5eb3a to get the mapping from index to label. End of explanation <END_TASK>
15,956
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Sweden Step1: 2. Join data to identify common water samples The recent data from Salar includes more water samples than are currently in RESA2. In addition, it looks as though the sample depths in RESA2 might be incorrect. As a first step, let's see if we can match water samples based on sample date and station codes. Step2: Next. check how much of the data downloaded by Salar is already in the database. Step3: So, the recent data downloaded by Salar includes 2412 records for these 25 stations, of which 1910 are already in RESA2. Let's check if the values for Mg are comparable. Step4: Excellent - the values in the RESA2 database for Mg match the values in the data Salar has recently downloaded. However, the sample depths do not match Step5: Next, get the sample IDs associated with these 25 sites and join them to the data above.
<ASSISTANT_TASK:> Python Code: # Read data in_xls = (r'C:\Data\James_Work\Staff\Heleen_d_W\ICP_Waters\TOC_Trends_Analysis_2015' r'\Swedish_Ca_Data\Missing_Data_25_Swedish_Sites.xlsx') smhi_df = pd.read_excel(in_xls, sheetname='salar_data') resa_df = pd.read_excel(in_xls, sheetname='from_resa_10-02-2017') # Get just the cols of interest smhi_df = smhi_df[['stn_code', 'stn_name', 'date', 'depth_m', 'ca_mekv/l', 'mg_mekv/l']] resa_df = resa_df[['stn_code', 'stn_name', 'date', 'depth1', 'depth2', 'ca_uekv/l', 'mg_uekv/l']] smhi_df.head(10) resa_df.head(10) Explanation: Sweden: missing Ca data Heleen has discovered that Ca data are missing from RESA2 for 25 Swedish sites in the TOC trends analysis - see e-mail received 07/02/2017 at 16.12. The relevant background data is gathered together in this spreadsheet: C:\Data\James_Work\Staff\Heleen_d_W\ICP_Waters\TOC_Trends_Analysis_2015\Swedish_Ca_Data\Missing_Data_25_Swedish_Sites.xlsx The 25 sites (with their RESA2 site codes) are listed in the resa2_stns_missing_data sheet. Salar has extracted the data for these sites from the SMHI database and I have copied this to the salar_data sheet and simplified the column headings (see e-mail from Heleen received 09/02/2017 at 10.35 for the original). The sheet from_resa_10-02-2017 shows all the Mg and Ca data from the Swedish sites currently associated with the ICPW_TOCTRENDS_2015_SE project (as of 10/02/2017, before any errors were corrected). This should make it possible to identify which water samples are already in the database. 1. Read data End of explanation # Join datasets df = pd.merge(resa_df, smhi_df, how='inner', left_on=['stn_code', 'date'], right_on=['stn_code', 'date']) df.head() Explanation: 2. Join data to identify common water samples The recent data from Salar includes more water samples than are currently in RESA2. In addition, it looks as though the sample depths in RESA2 might be incorrect. As a first step, let's see if we can match water samples based on sample date and station codes. End of explanation print len(smhi_df) print len(df) Explanation: Next. check how much of the data downloaded by Salar is already in the database. End of explanation # Convert Mg to ueq/l and round to nearest whole number df['mg_old_ueq/l'] = (df['mg_uekv/l'] + 0.5).astype(int) df['mg_new_ueq/l'] = (df['mg_mekv/l']*1000 + 0.5).astype(int) print 'All Mg values match:', (df['mg_old_ueq/l'] == df['mg_new_ueq/l']).all() df.head() Explanation: So, the recent data downloaded by Salar includes 2412 records for these 25 stations, of which 1910 are already in RESA2. Let's check if the values for Mg are comparable. End of explanation # Connect to db r2_func_path = r'C:\Data\James_Work\Staff\Heleen_d_W\ICP_Waters\Upload_Template\useful_resa2_code.py' resa2 = imp.load_source('useful_resa2_code', r2_func_path) engine, conn = resa2.connect_to_resa2() Explanation: Excellent - the values in the RESA2 database for Mg match the values in the data Salar has recently downloaded. However, the sample depths do not match: all the samples in RESA2 are assumed to have come from the surface (depth1=0 and depth2=0), whereas Salar's data indicates they actually come from a variety of depths between 0.5 and 2 m. In the trend analysis so far, we've focused on samples taken within the upper 0.5 m of the water column. How much does this matter? If it is important, we'll need to correct the sample depths for all the Swedish sites, which will involve Salar downloading some more data followed by a bit of additonal work from me. 3. Add Ca data Heleen has replied to my e-mail to say not to worry about the water depths - just add the new Ca data to the existing water sample IDs (see e-mail received 10/02/2017 at 16.19). First, connect to the database. End of explanation # Get all water samples for these sites stns = tuple(df['stn_code'].unique()) sql = ('SELECT station_id, sample_date, water_sample_id ' 'FROM resa2.water_samples ' 'WHERE station_id IN %s' % str(stns)) samp_df = pd.read_sql_query(sql, engine) # Join df = pd.merge(df, samp_df, how='left', left_on=['date', 'stn_code'], right_on=['sample_date', 'station_id']) df.head() # Extract columns of interest df = df[['water_sample_id', 'ca_mekv/l']] # Rename columns df.columns = ['sample_id', 'value'] # Add new columns df['method_id'] = 10551 df['approved'] = 'YES' df['entered_by'] = 'JES' df['entered_date'] = pd.datetime(2017, 2, 10) # Reorder columns df = df[['sample_id', 'method_id', 'value', 'approved', 'entered_by', 'entered_date']] df.head(10) # Write to database df.to_sql(name='WATER_CHEMISTRY_VALUES2', schema='RESA2', con=engine, index=False, if_exists='append') Explanation: Next, get the sample IDs associated with these 25 sites and join them to the data above. End of explanation <END_TASK>
15,957
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Step1: Language Translation In this project, you’re going to take a peek into the realm of neural network machine translation. You’ll be training a sequence to sequence model on a dataset of English and French sentences that can translate new sentences from English to French. Get the Data Since translating the whole language of English to French will take lots of time to train, we have provided you with a small portion of the English corpus. Step3: Explore the Data Play around with view_sentence_range to view different parts of the data. Step6: Implement Preprocessing Function Text to Word Ids As you did with other RNNs, you must turn the text into a number so the computer can understand it. In the function text_to_ids(), you'll turn source_text and target_text from words to ids. However, you need to add the &lt;EOS&gt; word id at the end of each sentence from target_text. This will help the neural network predict when the sentence should end. You can get the &lt;EOS&gt; word id by doing Step8: Preprocess all the data and save it Running the code cell below will preprocess all the data and save it to file. Step10: Check Point This is your first checkpoint. If you ever decide to come back to this notebook or have to restart the notebook, you can start from here. The preprocessed data has been saved to disk. Step12: Check the Version of TensorFlow and Access to GPU This will check to make sure you have the correct version of TensorFlow and access to a GPU Step15: Build the Neural Network You'll build the components necessary to build a Sequence-to-Sequence model by implementing the following functions below Step18: Process Decoding Input Implement process_decoding_input using TensorFlow to remove the last word id from each batch in target_data and concat the GO ID to the beginning of each batch. Step21: Encoding Implement encoding_layer() to create a Encoder RNN layer using tf.nn.dynamic_rnn(). Step24: Decoding - Training Create training logits using tf.contrib.seq2seq.simple_decoder_fn_train() and tf.contrib.seq2seq.dynamic_rnn_decoder(). Apply the output_fn to the tf.contrib.seq2seq.dynamic_rnn_decoder() outputs. Step27: Decoding - Inference Create inference logits using tf.contrib.seq2seq.simple_decoder_fn_inference() and tf.contrib.seq2seq.dynamic_rnn_decoder(). Step30: Build the Decoding Layer Implement decoding_layer() to create a Decoder RNN layer. Create RNN cell for decoding using rnn_size and num_layers. Create the output fuction using lambda to transform it's input, logits, to class logits. Use the your decoding_layer_train(encoder_state, dec_cell, dec_embed_input, sequence_length, decoding_scope, output_fn, keep_prob) function to get the training logits. Use your decoding_layer_infer(encoder_state, dec_cell, dec_embeddings, start_of_sequence_id, end_of_sequence_id, maximum_length, vocab_size, decoding_scope, output_fn, keep_prob) function to get the inference logits. Note Step33: Build the Neural Network Apply the functions you implemented above to Step34: Neural Network Training Hyperparameters Tune the following parameters Step36: Build the Graph Build the graph using the neural network you implemented. Step39: Train Train the neural network on the preprocessed data. If you have a hard time getting a good loss, check the forms to see if anyone is having the same problem. Step41: Save Parameters Save the batch_size and save_path parameters for inference. Step43: Checkpoint Step46: Sentence to Sequence To feed a sentence into the model for translation, you first need to preprocess it. Implement the function sentence_to_seq() to preprocess new sentences. Convert the sentence to lowercase Convert words into ids using vocab_to_int Convert words not in the vocabulary, to the &lt;UNK&gt; word id. Step48: Translate This will translate translate_sentence from English to French.
<ASSISTANT_TASK:> Python Code: DON'T MODIFY ANYTHING IN THIS CELL import helper import problem_unittests as tests source_path = 'data/small_vocab_en' target_path = 'data/small_vocab_fr' source_text = helper.load_data(source_path) target_text = helper.load_data(target_path) Explanation: Language Translation In this project, you’re going to take a peek into the realm of neural network machine translation. You’ll be training a sequence to sequence model on a dataset of English and French sentences that can translate new sentences from English to French. Get the Data Since translating the whole language of English to French will take lots of time to train, we have provided you with a small portion of the English corpus. End of explanation view_sentence_range = (0, 10) DON'T MODIFY ANYTHING IN THIS CELL import numpy as np print('Dataset Stats') print('Roughly the number of unique words: {}'.format(len({word: None for word in source_text.split()}))) sentences = source_text.split('\n') word_counts = [len(sentence.split()) for sentence in sentences] print('Number of sentences: {}'.format(len(sentences))) print('Average number of words in a sentence: {}'.format(np.average(word_counts))) print() print('English sentences {} to {}:'.format(*view_sentence_range)) print('\n'.join(source_text.split('\n')[view_sentence_range[0]:view_sentence_range[1]])) print() print('French sentences {} to {}:'.format(*view_sentence_range)) print('\n'.join(target_text.split('\n')[view_sentence_range[0]:view_sentence_range[1]])) Explanation: Explore the Data Play around with view_sentence_range to view different parts of the data. End of explanation def text_to_ids(source_text, target_text, source_vocab_to_int, target_vocab_to_int): Convert source and target text to proper word ids :param source_text: String that contains all the source text. :param target_text: String that contains all the target text. :param source_vocab_to_int: Dictionary to go from the source words to an id :param target_vocab_to_int: Dictionary to go from the target words to an id :return: A tuple of lists (source_id_text, target_id_text) # TODO: Implement Function source_id_text =[] target_id_text =[] for sentences in source_text.split('\n'): sentence_out = [source_vocab_to_int[word] for word in sentences.split()] source_id_text.append(sentence_out) for sentences in target_text.split('\n'): sentence_out = [target_vocab_to_int[word] for word in sentences.split()] sentence_out.append(target_vocab_to_int['<EOS>']) target_id_text.append(sentence_out) return source_id_text, target_id_text DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE tests.test_text_to_ids(text_to_ids) Explanation: Implement Preprocessing Function Text to Word Ids As you did with other RNNs, you must turn the text into a number so the computer can understand it. In the function text_to_ids(), you'll turn source_text and target_text from words to ids. However, you need to add the &lt;EOS&gt; word id at the end of each sentence from target_text. This will help the neural network predict when the sentence should end. You can get the &lt;EOS&gt; word id by doing: python target_vocab_to_int['&lt;EOS&gt;'] You can get other word ids using source_vocab_to_int and target_vocab_to_int. End of explanation DON'T MODIFY ANYTHING IN THIS CELL helper.preprocess_and_save_data(source_path, target_path, text_to_ids) Explanation: Preprocess all the data and save it Running the code cell below will preprocess all the data and save it to file. End of explanation DON'T MODIFY ANYTHING IN THIS CELL import numpy as np import helper (source_int_text, target_int_text), (source_vocab_to_int, target_vocab_to_int), _ = helper.load_preprocess() Explanation: Check Point This is your first checkpoint. If you ever decide to come back to this notebook or have to restart the notebook, you can start from here. The preprocessed data has been saved to disk. End of explanation DON'T MODIFY ANYTHING IN THIS CELL from distutils.version import LooseVersion import warnings import tensorflow as tf # Check TensorFlow Version assert LooseVersion(tf.__version__) in [LooseVersion('1.0.0'), LooseVersion('1.0.1')], 'This project requires TensorFlow version 1.0 You are using {}'.format(tf.__version__) print('TensorFlow Version: {}'.format(tf.__version__)) # Check for a GPU if not tf.test.gpu_device_name(): warnings.warn('No GPU found. Please use a GPU to train your neural network.') else: print('Default GPU Device: {}'.format(tf.test.gpu_device_name())) Explanation: Check the Version of TensorFlow and Access to GPU This will check to make sure you have the correct version of TensorFlow and access to a GPU End of explanation def model_inputs(): Create TF Placeholders for input, targets, and learning rate. :return: Tuple (input, targets, learning rate, keep probability) inputs = tf.placeholder(tf.int32, [None, None], name='input') targets = tf.placeholder(tf.int32, [None, None], name='target') learning_rate = tf.placeholder(tf.float32, name='learning_rate') keep_prob = tf.placeholder(tf.float32, name='keep_prob') return inputs, targets, learning_rate, keep_prob DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE tests.test_model_inputs(model_inputs) Explanation: Build the Neural Network You'll build the components necessary to build a Sequence-to-Sequence model by implementing the following functions below: - model_inputs - process_decoding_input - encoding_layer - decoding_layer_train - decoding_layer_infer - decoding_layer - seq2seq_model Input Implement the model_inputs() function to create TF Placeholders for the Neural Network. It should create the following placeholders: Input text placeholder named "input" using the TF Placeholder name parameter with rank 2. Targets placeholder with rank 2. Learning rate placeholder with rank 0. Keep probability placeholder named "keep_prob" using the TF Placeholder name parameter with rank 0. Return the placeholders in the following the tuple (Input, Targets, Learing Rate, Keep Probability) End of explanation def process_decoding_input(target_data, target_vocab_to_int, batch_size): Preprocess target data for dencoding :param target_data: Target Placehoder :param target_vocab_to_int: Dictionary to go from the target words to an id :param batch_size: Batch Size :return: Preprocessed target data go_batch = tf.concat([tf.fill([batch_size, 1], target_vocab_to_int['<GO>']), tf.strided_slice(target_data, [0,0], [batch_size, -1], [1,1])], 1) return go_batch DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE tests.test_process_decoding_input(process_decoding_input) Explanation: Process Decoding Input Implement process_decoding_input using TensorFlow to remove the last word id from each batch in target_data and concat the GO ID to the beginning of each batch. End of explanation def encoding_layer(rnn_inputs, rnn_size, num_layers, keep_prob): Create encoding layer :param rnn_inputs: Inputs for the RNN :param rnn_size: RNN Size :param num_layers: Number of layers :param keep_prob: Dropout keep probability :return: RNN state # TODO: Implement Function basic_cell = tf.contrib.rnn.BasicLSTMCell(rnn_size) multi_cell = tf.contrib.rnn.MultiRNNCell([basic_cell] * num_layers) dropout = tf.contrib.rnn.DropoutWrapper(multi_cell, keep_prob) output, state = tf.nn.dynamic_rnn(dropout, rnn_inputs, dtype=tf.float32) return state DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE tests.test_encoding_layer(encoding_layer) Explanation: Encoding Implement encoding_layer() to create a Encoder RNN layer using tf.nn.dynamic_rnn(). End of explanation def decoding_layer_train(encoder_state, dec_cell, dec_embed_input, sequence_length, decoding_scope, output_fn, keep_prob): Create a decoding layer for training :param encoder_state: Encoder State :param dec_cell: Decoder RNN Cell :param dec_embed_input: Decoder embedded input :param sequence_length: Sequence Length :param decoding_scope: TenorFlow Variable Scope for decoding :param output_fn: Function to apply the output layer :param keep_prob: Dropout keep probability :return: Train Logits # TODO: Implement Function simple_decoder_fn_train = tf.contrib.seq2seq.simple_decoder_fn_train(encoder_state) output, _, _ = tf.contrib.seq2seq.dynamic_rnn_decoder(dec_cell, simple_decoder_fn_train, dec_embed_input, sequence_length, scope=decoding_scope) logits = output_fn(tf.nn.dropout(output, keep_prob)) return logits DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE tests.test_decoding_layer_train(decoding_layer_train) Explanation: Decoding - Training Create training logits using tf.contrib.seq2seq.simple_decoder_fn_train() and tf.contrib.seq2seq.dynamic_rnn_decoder(). Apply the output_fn to the tf.contrib.seq2seq.dynamic_rnn_decoder() outputs. End of explanation def decoding_layer_infer(encoder_state, dec_cell, dec_embeddings, start_of_sequence_id, end_of_sequence_id, maximum_length, vocab_size, decoding_scope, output_fn, keep_prob): Create a decoding layer for inference :param encoder_state: Encoder state :param dec_cell: Decoder RNN Cell :param dec_embeddings: Decoder embeddings :param start_of_sequence_id: GO ID :param end_of_sequence_id: EOS Id :param maximum_length: The maximum allowed time steps to decode :param vocab_size: Size of vocabulary :param decoding_scope: TensorFlow Variable Scope for decoding :param output_fn: Function to apply the output layer :param keep_prob: Dropout keep probability :return: Inference Logits # TODO: Implement Function simple_decoder_fn_inference = tf.contrib.seq2seq.simple_decoder_fn_inference(output_fn, encoder_state, dec_embeddings, start_of_sequence_id, end_of_sequence_id, maximum_length, vocab_size) logits, _, _ = tf.contrib.seq2seq.dynamic_rnn_decoder(dec_cell, simple_decoder_fn_inference, scope=decoding_scope) return logits DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE tests.test_decoding_layer_infer(decoding_layer_infer) Explanation: Decoding - Inference Create inference logits using tf.contrib.seq2seq.simple_decoder_fn_inference() and tf.contrib.seq2seq.dynamic_rnn_decoder(). End of explanation def decoding_layer(dec_embed_input, dec_embeddings, encoder_state, vocab_size, sequence_length, rnn_size, num_layers, target_vocab_to_int, keep_prob): Create decoding layer :param dec_embed_input: Decoder embedded input :param dec_embeddings: Decoder embeddings :param encoder_state: The encoded state :param vocab_size: Size of vocabulary :param sequence_length: Sequence Length :param rnn_size: RNN Size :param num_layers: Number of layers :param target_vocab_to_int: Dictionary to go from the target words to an id :param keep_prob: Dropout keep probability :return: Tuple of (Training Logits, Inference Logits) # TODO: Implement Function basic_cell = tf.contrib.rnn.BasicLSTMCell(rnn_size) dropout = tf.contrib.rnn.DropoutWrapper(basic_cell, keep_prob) multi_cell = tf.contrib.rnn.MultiRNNCell([dropout] * num_layers) with tf.variable_scope("decoding") as decoding_scope: output_fn = lambda x: tf.contrib.layers.fully_connected(x, vocab_size, None, scope=decoding_scope) decoding_logits_train = decoding_layer_train(encoder_state, multi_cell, dec_embed_input, sequence_length, decoding_scope, output_fn, keep_prob) with tf.variable_scope("decoding", reuse=True) as decoding_scope: decoding_logits_infer = decoding_layer_infer(encoder_state, multi_cell, dec_embeddings, target_vocab_to_int['<GO>'], target_vocab_to_int['<EOS>'], sequence_length, vocab_size, decoding_scope, output_fn, keep_prob) return decoding_logits_train, decoding_logits_infer DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE tests.test_decoding_layer(decoding_layer) Explanation: Build the Decoding Layer Implement decoding_layer() to create a Decoder RNN layer. Create RNN cell for decoding using rnn_size and num_layers. Create the output fuction using lambda to transform it's input, logits, to class logits. Use the your decoding_layer_train(encoder_state, dec_cell, dec_embed_input, sequence_length, decoding_scope, output_fn, keep_prob) function to get the training logits. Use your decoding_layer_infer(encoder_state, dec_cell, dec_embeddings, start_of_sequence_id, end_of_sequence_id, maximum_length, vocab_size, decoding_scope, output_fn, keep_prob) function to get the inference logits. Note: You'll need to use tf.variable_scope to share variables between training and inference. End of explanation def seq2seq_model(input_data, target_data, keep_prob, batch_size, sequence_length, source_vocab_size, target_vocab_size, enc_embedding_size, dec_embedding_size, rnn_size, num_layers, target_vocab_to_int): Build the Sequence-to-Sequence part of the neural network :param input_data: Input placeholder :param target_data: Target placeholder :param keep_prob: Dropout keep probability placeholder :param batch_size: Batch Size :param sequence_length: Sequence Length :param source_vocab_size: Source vocabulary size :param target_vocab_size: Target vocabulary size :param enc_embedding_size: Decoder embedding size :param dec_embedding_size: Encoder embedding size :param rnn_size: RNN Size :param num_layers: Number of layers :param target_vocab_to_int: Dictionary to go from the target words to an id :return: Tuple of (Training Logits, Inference Logits) # TODO: Implement Function encoder_embed_input = tf.contrib.layers.embed_sequence(input_data, source_vocab_size, enc_embedding_size) encoder_output = encoding_layer(encoder_embed_input, rnn_size, num_layers, keep_prob) decoder_input = process_decoding_input(target_data, target_vocab_to_int, batch_size) decoder_embed = tf.Variable(tf.random_uniform([target_vocab_size, dec_embedding_size])) decoder_embed_input = tf.nn.embedding_lookup(decoder_embed, decoder_input) t_logits, i_logits = decoding_layer(decoder_embed_input, decoder_embed, encoder_output, target_vocab_size, sequence_length, rnn_size, num_layers, target_vocab_to_int, keep_prob) return t_logits, i_logits DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE tests.test_seq2seq_model(seq2seq_model) Explanation: Build the Neural Network Apply the functions you implemented above to: Apply embedding to the input data for the encoder. Encode the input using your encoding_layer(rnn_inputs, rnn_size, num_layers, keep_prob). Process target data using your process_decoding_input(target_data, target_vocab_to_int, batch_size) function. Apply embedding to the target data for the decoder. Decode the encoded input using your decoding_layer(dec_embed_input, dec_embeddings, encoder_state, vocab_size, sequence_length, rnn_size, num_layers, target_vocab_to_int, keep_prob). End of explanation # Number of Epochs epochs = 5 # Batch Size batch_size = 256 # RNN Size rnn_size = 512 # Number of Layers num_layers = 2 # Embedding Size encoding_embedding_size = 128 decoding_embedding_size = 128 # Learning Rate learning_rate = 0.001 # Dropout Keep Probability keep_probability = 0.5 Explanation: Neural Network Training Hyperparameters Tune the following parameters: Set epochs to the number of epochs. Set batch_size to the batch size. Set rnn_size to the size of the RNNs. Set num_layers to the number of layers. Set encoding_embedding_size to the size of the embedding for the encoder. Set decoding_embedding_size to the size of the embedding for the decoder. Set learning_rate to the learning rate. Set keep_probability to the Dropout keep probability End of explanation DON'T MODIFY ANYTHING IN THIS CELL save_path = 'checkpoints/dev' (source_int_text, target_int_text), (source_vocab_to_int, target_vocab_to_int), _ = helper.load_preprocess() max_source_sentence_length = max([len(sentence) for sentence in source_int_text]) train_graph = tf.Graph() with train_graph.as_default(): input_data, targets, lr, keep_prob = model_inputs() sequence_length = tf.placeholder_with_default(max_source_sentence_length, None, name='sequence_length') input_shape = tf.shape(input_data) train_logits, inference_logits = seq2seq_model( tf.reverse(input_data, [-1]), targets, keep_prob, batch_size, sequence_length, len(source_vocab_to_int), len(target_vocab_to_int), encoding_embedding_size, decoding_embedding_size, rnn_size, num_layers, target_vocab_to_int) tf.identity(inference_logits, 'logits') with tf.name_scope("optimization"): # Loss function cost = tf.contrib.seq2seq.sequence_loss( train_logits, targets, tf.ones([input_shape[0], sequence_length])) # Optimizer optimizer = tf.train.AdamOptimizer(lr) # Gradient Clipping gradients = optimizer.compute_gradients(cost) capped_gradients = [(tf.clip_by_value(grad, -1., 1.), var) for grad, var in gradients if grad is not None] train_op = optimizer.apply_gradients(capped_gradients) Explanation: Build the Graph Build the graph using the neural network you implemented. End of explanation DON'T MODIFY ANYTHING IN THIS CELL import time def get_accuracy(target, logits): Calculate accuracy max_seq = max(target.shape[1], logits.shape[1]) if max_seq - target.shape[1]: target = np.pad( target, [(0,0),(0,max_seq - target.shape[1])], 'constant') if max_seq - logits.shape[1]: logits = np.pad( logits, [(0,0),(0,max_seq - logits.shape[1]), (0,0)], 'constant') return np.mean(np.equal(target, np.argmax(logits, 2))) train_source = source_int_text[batch_size:] train_target = target_int_text[batch_size:] valid_source = helper.pad_sentence_batch(source_int_text[:batch_size]) valid_target = helper.pad_sentence_batch(target_int_text[:batch_size]) with tf.Session(graph=train_graph) as sess: sess.run(tf.global_variables_initializer()) for epoch_i in range(epochs): for batch_i, (source_batch, target_batch) in enumerate( helper.batch_data(train_source, train_target, batch_size)): start_time = time.time() _, loss = sess.run( [train_op, cost], {input_data: source_batch, targets: target_batch, lr: learning_rate, sequence_length: target_batch.shape[1], keep_prob: keep_probability}) batch_train_logits = sess.run( inference_logits, {input_data: source_batch, keep_prob: 1.0}) batch_valid_logits = sess.run( inference_logits, {input_data: valid_source, keep_prob: 1.0}) train_acc = get_accuracy(target_batch, batch_train_logits) valid_acc = get_accuracy(np.array(valid_target), batch_valid_logits) end_time = time.time() print('Epoch {:>3} Batch {:>4}/{} - Train Accuracy: {:>6.3f}, Validation Accuracy: {:>6.3f}, Loss: {:>6.3f}' .format(epoch_i, batch_i, len(source_int_text) // batch_size, train_acc, valid_acc, loss)) # Save Model saver = tf.train.Saver() saver.save(sess, save_path) print('Model Trained and Saved') Explanation: Train Train the neural network on the preprocessed data. If you have a hard time getting a good loss, check the forms to see if anyone is having the same problem. End of explanation DON'T MODIFY ANYTHING IN THIS CELL # Save parameters for checkpoint helper.save_params(save_path) Explanation: Save Parameters Save the batch_size and save_path parameters for inference. End of explanation DON'T MODIFY ANYTHING IN THIS CELL import tensorflow as tf import numpy as np import helper import problem_unittests as tests _, (source_vocab_to_int, target_vocab_to_int), (source_int_to_vocab, target_int_to_vocab) = helper.load_preprocess() load_path = helper.load_params() Explanation: Checkpoint End of explanation def sentence_to_seq(sentence, vocab_to_int): Convert a sentence to a sequence of ids :param sentence: String :param vocab_to_int: Dictionary to go from the words to an id :return: List of word ids # TODO: Implement Function unknown_word = vocab_to_int['<UNK>'] sentence_lowercase = sentence.lower() word_ids = [vocab_to_int.get(word, unknown_word) for word in sentence_lowercase.split()] return word_ids DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE tests.test_sentence_to_seq(sentence_to_seq) Explanation: Sentence to Sequence To feed a sentence into the model for translation, you first need to preprocess it. Implement the function sentence_to_seq() to preprocess new sentences. Convert the sentence to lowercase Convert words into ids using vocab_to_int Convert words not in the vocabulary, to the &lt;UNK&gt; word id. End of explanation translate_sentence = 'he saw a old yellow truck .' DON'T MODIFY ANYTHING IN THIS CELL translate_sentence = sentence_to_seq(translate_sentence, source_vocab_to_int) loaded_graph = tf.Graph() with tf.Session(graph=loaded_graph) as sess: # Load saved model loader = tf.train.import_meta_graph(load_path + '.meta') loader.restore(sess, load_path) input_data = loaded_graph.get_tensor_by_name('input:0') logits = loaded_graph.get_tensor_by_name('logits:0') keep_prob = loaded_graph.get_tensor_by_name('keep_prob:0') translate_logits = sess.run(logits, {input_data: [translate_sentence], keep_prob: 1.0})[0] print('Input') print(' Word Ids: {}'.format([i for i in translate_sentence])) print(' English Words: {}'.format([source_int_to_vocab[i] for i in translate_sentence])) print('\nPrediction') print(' Word Ids: {}'.format([i for i in np.argmax(translate_logits, 1)])) print(' French Words: {}'.format([target_int_to_vocab[i] for i in np.argmax(translate_logits, 1)])) Explanation: Translate This will translate translate_sentence from English to French. End of explanation <END_TASK>
15,958
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: ES-DOC CMIP6 Model Properties - Aerosol MIP Era Step1: Document Authors Set document authors Step2: Document Contributors Specify document contributors Step3: Document Publication Specify document publication status Step4: Document Table of Contents 1. Key Properties 2. Key Properties --&gt; Software Properties 3. Key Properties --&gt; Timestep Framework 4. Key Properties --&gt; Meteorological Forcings 5. Key Properties --&gt; Resolution 6. Key Properties --&gt; Tuning Applied 7. Transport 8. Emissions 9. Concentrations 10. Optical Radiative Properties 11. Optical Radiative Properties --&gt; Absorption 12. Optical Radiative Properties --&gt; Mixtures 13. Optical Radiative Properties --&gt; Impact Of H2o 14. Optical Radiative Properties --&gt; Radiative Scheme 15. Optical Radiative Properties --&gt; Cloud Interactions 16. Model 1. Key Properties Key properties of the aerosol model 1.1. Model Overview Is Required Step5: 1.2. Model Name Is Required Step6: 1.3. Scheme Scope Is Required Step7: 1.4. Basic Approximations Is Required Step8: 1.5. Prognostic Variables Form Is Required Step9: 1.6. Number Of Tracers Is Required Step10: 1.7. Family Approach Is Required Step11: 2. Key Properties --&gt; Software Properties Software properties of aerosol code 2.1. Repository Is Required Step12: 2.2. Code Version Is Required Step13: 2.3. Code Languages Is Required Step14: 3. Key Properties --&gt; Timestep Framework Physical properties of seawater in ocean 3.1. Method Is Required Step15: 3.2. Split Operator Advection Timestep Is Required Step16: 3.3. Split Operator Physical Timestep Is Required Step17: 3.4. Integrated Timestep Is Required Step18: 3.5. Integrated Scheme Type Is Required Step19: 4. Key Properties --&gt; Meteorological Forcings ** 4.1. Variables 3D Is Required Step20: 4.2. Variables 2D Is Required Step21: 4.3. Frequency Is Required Step22: 5. Key Properties --&gt; Resolution Resolution in the aersosol model grid 5.1. Name Is Required Step23: 5.2. Canonical Horizontal Resolution Is Required Step24: 5.3. Number Of Horizontal Gridpoints Is Required Step25: 5.4. Number Of Vertical Levels Is Required Step26: 5.5. Is Adaptive Grid Is Required Step27: 6. Key Properties --&gt; Tuning Applied Tuning methodology for aerosol model 6.1. Description Is Required Step28: 6.2. Global Mean Metrics Used Is Required Step29: 6.3. Regional Metrics Used Is Required Step30: 6.4. Trend Metrics Used Is Required Step31: 7. Transport Aerosol transport 7.1. Overview Is Required Step32: 7.2. Scheme Is Required Step33: 7.3. Mass Conservation Scheme Is Required Step34: 7.4. Convention Is Required Step35: 8. Emissions Atmospheric aerosol emissions 8.1. Overview Is Required Step36: 8.2. Method Is Required Step37: 8.3. Sources Is Required Step38: 8.4. Prescribed Climatology Is Required Step39: 8.5. Prescribed Climatology Emitted Species Is Required Step40: 8.6. Prescribed Spatially Uniform Emitted Species Is Required Step41: 8.7. Interactive Emitted Species Is Required Step42: 8.8. Other Emitted Species Is Required Step43: 8.9. Other Method Characteristics Is Required Step44: 9. Concentrations Atmospheric aerosol concentrations 9.1. Overview Is Required Step45: 9.2. Prescribed Lower Boundary Is Required Step46: 9.3. Prescribed Upper Boundary Is Required Step47: 9.4. Prescribed Fields Mmr Is Required Step48: 9.5. Prescribed Fields Mmr Is Required Step49: 10. Optical Radiative Properties Aerosol optical and radiative properties 10.1. Overview Is Required Step50: 11. Optical Radiative Properties --&gt; Absorption Absortion properties in aerosol scheme 11.1. Black Carbon Is Required Step51: 11.2. Dust Is Required Step52: 11.3. Organics Is Required Step53: 12. Optical Radiative Properties --&gt; Mixtures ** 12.1. External Is Required Step54: 12.2. Internal Is Required Step55: 12.3. Mixing Rule Is Required Step56: 13. Optical Radiative Properties --&gt; Impact Of H2o ** 13.1. Size Is Required Step57: 13.2. Internal Mixture Is Required Step58: 14. Optical Radiative Properties --&gt; Radiative Scheme Radiative scheme for aerosol 14.1. Overview Is Required Step59: 14.2. Shortwave Bands Is Required Step60: 14.3. Longwave Bands Is Required Step61: 15. Optical Radiative Properties --&gt; Cloud Interactions Aerosol-cloud interactions 15.1. Overview Is Required Step62: 15.2. Twomey Is Required Step63: 15.3. Twomey Minimum Ccn Is Required Step64: 15.4. Drizzle Is Required Step65: 15.5. Cloud Lifetime Is Required Step66: 15.6. Longwave Bands Is Required Step67: 16. Model Aerosol model 16.1. Overview Is Required Step68: 16.2. Processes Is Required Step69: 16.3. Coupling Is Required Step70: 16.4. Gas Phase Precursors Is Required Step71: 16.5. Scheme Type Is Required Step72: 16.6. Bulk Scheme Species Is Required
<ASSISTANT_TASK:> Python Code: # DO NOT EDIT ! from pyesdoc.ipython.model_topic import NotebookOutput # DO NOT EDIT ! DOC = NotebookOutput('cmip6', 'cmcc', 'sandbox-3', 'aerosol') Explanation: ES-DOC CMIP6 Model Properties - Aerosol MIP Era: CMIP6 Institute: CMCC Source ID: SANDBOX-3 Topic: Aerosol Sub-Topics: Transport, Emissions, Concentrations, Optical Radiative Properties, Model. Properties: 69 (37 required) Model descriptions: Model description details Initialized From: -- Notebook Help: Goto notebook help page Notebook Initialised: 2018-02-15 16:53:50 Document Setup IMPORTANT: to be executed each time you run the notebook End of explanation # Set as follows: DOC.set_author("name", "email") # TODO - please enter value(s) Explanation: Document Authors Set document authors End of explanation # Set as follows: DOC.set_contributor("name", "email") # TODO - please enter value(s) Explanation: Document Contributors Specify document contributors End of explanation # Set publication status: # 0=do not publish, 1=publish. DOC.set_publication_status(0) Explanation: Document Publication Specify document publication status End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.model_overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: Document Table of Contents 1. Key Properties 2. Key Properties --&gt; Software Properties 3. Key Properties --&gt; Timestep Framework 4. Key Properties --&gt; Meteorological Forcings 5. Key Properties --&gt; Resolution 6. Key Properties --&gt; Tuning Applied 7. Transport 8. Emissions 9. Concentrations 10. Optical Radiative Properties 11. Optical Radiative Properties --&gt; Absorption 12. Optical Radiative Properties --&gt; Mixtures 13. Optical Radiative Properties --&gt; Impact Of H2o 14. Optical Radiative Properties --&gt; Radiative Scheme 15. Optical Radiative Properties --&gt; Cloud Interactions 16. Model 1. Key Properties Key properties of the aerosol model 1.1. Model Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of aerosol model. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.model_name') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 1.2. Model Name Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Name of aerosol model code End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.scheme_scope') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "troposhere" # "stratosphere" # "mesosphere" # "mesosphere" # "whole atmosphere" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 1.3. Scheme Scope Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N Atmospheric domains covered by the aerosol model End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.basic_approximations') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 1.4. Basic Approximations Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Basic approximations made in the aerosol model End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.prognostic_variables_form') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "3D mass/volume ratio for aerosols" # "3D number concenttration for aerosols" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 1.5. Prognostic Variables Form Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N Prognostic variables in the aerosol model End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.number_of_tracers') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 1.6. Number Of Tracers Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Number of tracers in the aerosol model End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.family_approach') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 1.7. Family Approach Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Are aerosol calculations generalized into families of species? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.software_properties.repository') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 2. Key Properties --&gt; Software Properties Software properties of aerosol code 2.1. Repository Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Location of code for this component. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.software_properties.code_version') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 2.2. Code Version Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Code version identifier. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.software_properties.code_languages') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 2.3. Code Languages Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N Code language(s). End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.timestep_framework.method') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Uses atmospheric chemistry time stepping" # "Specific timestepping (operator splitting)" # "Specific timestepping (integrated)" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 3. Key Properties --&gt; Timestep Framework Physical properties of seawater in ocean 3.1. Method Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Mathematical method deployed to solve the time evolution of the prognostic variables End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.timestep_framework.split_operator_advection_timestep') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 3.2. Split Operator Advection Timestep Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Timestep for aerosol advection (in seconds) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.timestep_framework.split_operator_physical_timestep') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 3.3. Split Operator Physical Timestep Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Timestep for aerosol physics (in seconds). End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.timestep_framework.integrated_timestep') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 3.4. Integrated Timestep Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Timestep for the aerosol model (in seconds) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.timestep_framework.integrated_scheme_type') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Explicit" # "Implicit" # "Semi-implicit" # "Semi-analytic" # "Impact solver" # "Back Euler" # "Newton Raphson" # "Rosenbrock" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 3.5. Integrated Scheme Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Specify the type of timestep scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.meteorological_forcings.variables_3D') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 4. Key Properties --&gt; Meteorological Forcings ** 4.1. Variables 3D Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Three dimensionsal forcing variables, e.g. U, V, W, T, Q, P, conventive mass flux End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.meteorological_forcings.variables_2D') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 4.2. Variables 2D Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Two dimensionsal forcing variables, e.g. land-sea mask definition End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.meteorological_forcings.frequency') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 4.3. Frequency Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Frequency with which meteological forcings are applied (in seconds). End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.resolution.name') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 5. Key Properties --&gt; Resolution Resolution in the aersosol model grid 5.1. Name Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 This is a string usually used by the modelling group to describe the resolution of this grid, e.g. ORCA025, N512L180, T512L70 etc. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.resolution.canonical_horizontal_resolution') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 5.2. Canonical Horizontal Resolution Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Expression quoted for gross comparisons of resolution, eg. 50km or 0.1 degrees etc. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.resolution.number_of_horizontal_gridpoints') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 5.3. Number Of Horizontal Gridpoints Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Total number of horizontal (XY) points (or degrees of freedom) on computational grid. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.resolution.number_of_vertical_levels') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 5.4. Number Of Vertical Levels Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Number of vertical levels resolved on computational grid. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.resolution.is_adaptive_grid') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 5.5. Is Adaptive Grid Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Default is False. Set true if grid resolution changes during execution. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.tuning_applied.description') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 6. Key Properties --&gt; Tuning Applied Tuning methodology for aerosol model 6.1. Description Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 General overview description of tuning: explain and motivate the main targets and metrics retained. &amp;Document the relative weight given to climate performance metrics versus process oriented metrics, &amp;and on the possible conflicts with parameterization level tuning. In particular describe any struggle &amp;with a parameter value that required pushing it to its limits to solve a particular model deficiency. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.tuning_applied.global_mean_metrics_used') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 6.2. Global Mean Metrics Used Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N List set of metrics of the global mean state used in tuning model/component End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.tuning_applied.regional_metrics_used') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 6.3. Regional Metrics Used Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N List of regional metrics of mean state used in tuning model/component End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.key_properties.tuning_applied.trend_metrics_used') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 6.4. Trend Metrics Used Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N List observed trend metrics used in tuning model/component End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.transport.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 7. Transport Aerosol transport 7.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of transport in atmosperic aerosol model End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.transport.scheme') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Uses Atmospheric chemistry transport scheme" # "Specific transport scheme (eulerian)" # "Specific transport scheme (semi-lagrangian)" # "Specific transport scheme (eulerian and semi-lagrangian)" # "Specific transport scheme (lagrangian)" # TODO - please enter value(s) Explanation: 7.2. Scheme Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Method for aerosol transport modeling End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.transport.mass_conservation_scheme') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "Uses Atmospheric chemistry transport scheme" # "Mass adjustment" # "Concentrations positivity" # "Gradients monotonicity" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 7.3. Mass Conservation Scheme Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N Method used to ensure mass conservation. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.transport.convention') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "Uses Atmospheric chemistry transport scheme" # "Convective fluxes connected to tracers" # "Vertical velocities connected to tracers" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 7.4. Convention Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N Transport by convention End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.emissions.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 8. Emissions Atmospheric aerosol emissions 8.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of emissions in atmosperic aerosol model End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.emissions.method') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "None" # "Prescribed (climatology)" # "Prescribed CMIP6" # "Prescribed above surface" # "Interactive" # "Interactive above surface" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 8.2. Method Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N Method used to define aerosol species (several methods allowed because the different species may not use the same method). End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.emissions.sources') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "Vegetation" # "Volcanos" # "Bare ground" # "Sea surface" # "Lightning" # "Fires" # "Aircraft" # "Anthropogenic" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 8.3. Sources Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N Sources of the aerosol species are taken into account in the emissions scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.emissions.prescribed_climatology') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # Valid Choices: # "Constant" # "Interannual" # "Annual" # "Monthly" # "Daily" # TODO - please enter value(s) Explanation: 8.4. Prescribed Climatology Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Specify the climatology type for aerosol emissions End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.emissions.prescribed_climatology_emitted_species') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 8.5. Prescribed Climatology Emitted Species Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 List of aerosol species emitted and prescribed via a climatology End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.emissions.prescribed_spatially_uniform_emitted_species') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 8.6. Prescribed Spatially Uniform Emitted Species Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 List of aerosol species emitted and prescribed as spatially uniform End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.emissions.interactive_emitted_species') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 8.7. Interactive Emitted Species Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 List of aerosol species emitted and specified via an interactive method End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.emissions.other_emitted_species') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 8.8. Other Emitted Species Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 List of aerosol species emitted and specified via an &quot;other method&quot; End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.emissions.other_method_characteristics') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 8.9. Other Method Characteristics Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Characteristics of the &quot;other method&quot; used for aerosol emissions End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.concentrations.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 9. Concentrations Atmospheric aerosol concentrations 9.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of concentrations in atmosperic aerosol model End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.concentrations.prescribed_lower_boundary') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 9.2. Prescribed Lower Boundary Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 List of species prescribed at the lower boundary. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.concentrations.prescribed_upper_boundary') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 9.3. Prescribed Upper Boundary Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 List of species prescribed at the upper boundary. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.concentrations.prescribed_fields_mmr') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 9.4. Prescribed Fields Mmr Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 List of species prescribed as mass mixing ratios. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.concentrations.prescribed_fields_mmr') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 9.5. Prescribed Fields Mmr Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 List of species prescribed as AOD plus CCNs. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.optical_radiative_properties.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 10. Optical Radiative Properties Aerosol optical and radiative properties 10.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of optical and radiative properties End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.optical_radiative_properties.absorption.black_carbon') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 11. Optical Radiative Properties --&gt; Absorption Absortion properties in aerosol scheme 11.1. Black Carbon Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: FLOAT&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Absorption mass coefficient of black carbon at 550nm (if non-absorbing enter 0) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.optical_radiative_properties.absorption.dust') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 11.2. Dust Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: FLOAT&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Absorption mass coefficient of dust at 550nm (if non-absorbing enter 0) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.optical_radiative_properties.absorption.organics') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 11.3. Organics Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: FLOAT&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 Absorption mass coefficient of organics at 550nm (if non-absorbing enter 0) End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.optical_radiative_properties.mixtures.external') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 12. Optical Radiative Properties --&gt; Mixtures ** 12.1. External Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is there external mixing with respect to chemical composition? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.optical_radiative_properties.mixtures.internal') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 12.2. Internal Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is there internal mixing with respect to chemical composition? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.optical_radiative_properties.mixtures.mixing_rule') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 12.3. Mixing Rule Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 If there is internal mixing with respect to chemical composition then indicate the mixinrg rule End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.optical_radiative_properties.impact_of_h2o.size') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 13. Optical Radiative Properties --&gt; Impact Of H2o ** 13.1. Size Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Does H2O impact size? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.optical_radiative_properties.impact_of_h2o.internal_mixture') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 13.2. Internal Mixture Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Does H2O impact internal mixture? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.optical_radiative_properties.radiative_scheme.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 14. Optical Radiative Properties --&gt; Radiative Scheme Radiative scheme for aerosol 14.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of radiative scheme End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.optical_radiative_properties.radiative_scheme.shortwave_bands') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 14.2. Shortwave Bands Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Number of shortwave bands End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.optical_radiative_properties.radiative_scheme.longwave_bands') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 14.3. Longwave Bands Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Number of longwave bands End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.optical_radiative_properties.cloud_interactions.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 15. Optical Radiative Properties --&gt; Cloud Interactions Aerosol-cloud interactions 15.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of aerosol-cloud interactions End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.optical_radiative_properties.cloud_interactions.twomey') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 15.2. Twomey Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Is the Twomey effect included? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.optical_radiative_properties.cloud_interactions.twomey_minimum_ccn') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 15.3. Twomey Minimum Ccn Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.1 If the Twomey effect is included, then what is the minimum CCN number? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.optical_radiative_properties.cloud_interactions.drizzle') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 15.4. Drizzle Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Does the scheme affect drizzle? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.optical_radiative_properties.cloud_interactions.cloud_lifetime') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # Valid Choices: # True # False # TODO - please enter value(s) Explanation: 15.5. Cloud Lifetime Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: BOOLEAN&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Does the scheme affect cloud lifetime? End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.optical_radiative_properties.cloud_interactions.longwave_bands') # PROPERTY VALUE: # Set as follows: DOC.set_value(value) # TODO - please enter value(s) Explanation: 15.6. Longwave Bands Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: INTEGER&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Number of longwave bands End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.model.overview') # PROPERTY VALUE: # Set as follows: DOC.set_value("value") # TODO - please enter value(s) Explanation: 16. Model Aerosol model 16.1. Overview Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: STRING&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.1 Overview of atmosperic aerosol model End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.model.processes') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "Dry deposition" # "Sedimentation" # "Wet deposition (impaction scavenging)" # "Wet deposition (nucleation scavenging)" # "Coagulation" # "Oxidation (gas phase)" # "Oxidation (in cloud)" # "Condensation" # "Ageing" # "Advection (horizontal)" # "Advection (vertical)" # "Heterogeneous chemistry" # "Nucleation" # TODO - please enter value(s) Explanation: 16.2. Processes Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N Processes included in the Aerosol model. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.model.coupling') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "Radiation" # "Land surface" # "Heterogeneous chemistry" # "Clouds" # "Ocean" # "Cryosphere" # "Gas phase chemistry" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 16.3. Coupling Is Required: FALSE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 0.N Other model components coupled to the Aerosol model End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.model.gas_phase_precursors') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "DMS" # "SO2" # "Ammonia" # "Iodine" # "Terpene" # "Isoprene" # "VOC" # "NOx" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 16.4. Gas Phase Precursors Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N List of gas phase aerosol precursors. End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.model.scheme_type') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "Bulk" # "Modal" # "Bin" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 16.5. Scheme Type Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N Type(s) of aerosol scheme used by the aerosols model (potentially multiple: some species may be covered by one type of aerosol scheme and other species covered by another type). End of explanation # PROPERTY ID - DO NOT EDIT ! DOC.set_id('cmip6.aerosol.model.bulk_scheme_species') # PROPERTY VALUE(S): # Set as follows: DOC.set_value("value") # Valid Choices: # "Sulphate" # "Nitrate" # "Sea salt" # "Dust" # "Ice" # "Organic" # "Black carbon / soot" # "SOA (secondary organic aerosols)" # "POM (particulate organic matter)" # "Polar stratospheric ice" # "NAT (Nitric acid trihydrate)" # "NAD (Nitric acid dihydrate)" # "STS (supercooled ternary solution aerosol particule)" # "Other: [Please specify]" # TODO - please enter value(s) Explanation: 16.6. Bulk Scheme Species Is Required: TRUE&nbsp;&nbsp;&nbsp;&nbsp;Type: ENUM&nbsp;&nbsp;&nbsp;&nbsp;Cardinality: 1.N List of species covered by the bulk scheme. End of explanation <END_TASK>
15,959
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Use Word2Vec in gensim to train a word embedding model using the content from NIPS papers. Step1: Gensim word2vec https Step2: Train a word2vec model Step3: Create a representation of each paper The representation is simply a set of embedded words taken from the abstract and the title. Step4: Load the saved pickle and check Step5: filter words by DF
<ASSISTANT_TASK:> Python Code: %load_ext autoreload %autoreload 2 %matplotlib inline #%config InlineBackend.figure_format = 'svg' #config InlineBackend.figure_format = 'pdf' from IPython.core.display import HTML import gensim as gen import gensim.models.word2vec as w2v import matplotlib.pyplot as plt from nltk.tokenize import WhitespaceTokenizer import numpy as np import os import pandas as pd try: import cPickle as pickle except: import pickle import re import scipy.stats as stats import scipy.sparse as sp import string import sys import csv # load the pickle containing the document-term matrix, # put the abstracts in, and dump it to a file. fyear = 1988 tyear = 2015 dt_fpath = 'DT_%d_%d_wabs.p'%(fyear, tyear) with open(dt_fpath, 'r') as f: info = pickle.load(f) info.keys() list_abs = info['abstracts'] list_abs[:2] # make each abstract a list of words list_list_abs = [ab.split(' ') for ab in list_abs if ab is not None] print list_list_abs[20] Explanation: Use Word2Vec in gensim to train a word embedding model using the content from NIPS papers. End of explanation def paper_dataframe(fpath): rows = [] with open(fpath, 'r') as csvfile: reader = csv.reader(csvfile, delimiter=',', quotechar='"') # Each read gives ['Id', 'Title', 'EventType', 'PdfName', 'Abstract', 'PaperText'] reader.next() for row in reader: rows.append(tuple(row)) data = pd.DataFrame(rows, columns=['Id', 'Title', 'EventType', 'PdfName', 'Abstract', 'PaperText']) return data text = ',sdf,.-23\][](s)' re.sub(r'([^\w])+', ' ', text, flags=re.DOTALL) def tokenize_simple(text): # replace spaces with one space text = re.sub(r'\s+', ' ', text, flags=re.DOTALL) # remove non-English words text = re.sub(r'[^\w]+', ' ', text, flags=re.DOTALL) # naive tokenization tokens = [w.lower().strip() for w in text.split(' ') if len(w) > 1] return tokens dframe = paper_dataframe('Papers1988_2015.csv') n_docs = dframe.shape[0] tok_papers = [] tok_abstracts = [] for i in xrange(n_docs): paper = dframe['PaperText'][i] paper_tokens = tokenize_simple(paper) tok_papers.append(paper_tokens) ab = list_abs[i] if ab is None: ab_tokens = [] else: ab_tokens = tokenize_simple(ab) tok_abstracts.append(ab_tokens) Explanation: Gensim word2vec https://radimrehurek.com/gensim/models/word2vec.html#id6 End of explanation # size means the latent dimension # sentences = an iterable where each item is a list of words size = 50 window = 5 dest_fname = 'w2v_size%d_win%d.p'%(size, window) model = w2v.Word2Vec(tok_papers, size=size, window=window, min_count=5, workers=4) model.save(dest_fname) model.wv.similarity('neural', 'deep') model.wv.similarity('neural', 'kernel') model.wv.doesnt_match('supervised unsupervised neuron reinforcement'.split()) model.wv.doesnt_match('kernel gretton hsic mmd'.split()) model.wv['kernel'] 'kernel' in model.wv Explanation: Train a word2vec model End of explanation titles = info['titles'] # each element is the representation of the paper. # This is a matrix with each row corresponding to the embedding # of a word in the abstract and the title. paper_reps = [] for i in xrange(n_docs): title_tokens = tokenize_simple(titles[i]) rep_words = tok_abstracts[i] + title_tokens # embed each word in rep_words (if in the vocabulary) rep = [] for w in rep_words: # only embed words that are in the vocabulary if w in model.wv: embed = model.wv[w] rep.append(embed) mat = np.vstack(rep) paper_reps.append(mat) len(paper_reps) # save the pickle with the paper representations dt_dest = 'DT_%d_%d_wembed.p'%(fyear, tyear) info['paper_reps'] = paper_reps with open(dt_dest, 'w') as f: pickle.dump(info, f) Explanation: Create a representation of each paper The representation is simply a set of embedded words taken from the abstract and the title. End of explanation with open('DT_%d_%d_wembed.p'%(fyear, tyear), 'r') as f: info = pickle.load(f) info.keys() DT = info['DT'] abstracts = info['abstracts'] paper_reps = info['paper_reps'] titles = info['titles'] words = info['words'] Explanation: Load the saved pickle and check End of explanation # document frequency of each word n_docs = DT.shape[0] DF = np.array( (DT > 0).sum(0) )[0] df_lb = 7 df_ub = int(0.15*n_docs) print('n = #docs: %d'%n_docs) print('original #words: %d'%len(words)) print('#words with %d <= df: %d'% (df_lb, np.sum(DF>=df_lb) ) ) print('#words with df <= %d: %d'% (df_ub, np.sum(DF<=df_ub) ) ) df_I = np.logical_and(DF>=df_lb, DF<=df_ub) print('#words with %d <= df <= %d: %d'% (df_lb, df_ub, np.sum( df_I) ) ) df_words = np.array(words)[df_I] print df_words.tolist() # filter out words fDT = DT[:, df_I] fwords = np.array(words)[df_I].tolist() info['DT'] = fDT info['words'] = fwords dffiltered_fname = 'DT_%d_%d_wem_df%d_%d.p'%(fyear, tyear, df_lb, df_ub) with open(dffiltered_fname, 'w') as f: pickle.dump(info, f) Explanation: filter words by DF End of explanation <END_TASK>
15,960
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Using ICsound csoundmagics includes an ICsound class which is adapted from Andrés Cabrera's icsound module. ICsound is bound to the %%csound and %csound magics command. This notebook is an adaptation of Andrés' icsound test notebook. Starting the Csound engine To use ICsound create an ICsound instance Step1: Creating an ICsound object automatically starts the engine Step2: You can set the properties of the Csound engine with parameters to the startEngine() function. Step3: The engine runs in a separate thread, so it doesn't block execution of python. Step4: Use the %%csound magic command to directly type csound language code in the cell and send it to the engine. The number after the magic command is optional; it references the slot where the engine is running. If omitted, slot#1 is assumed. Step5: So where did it print? Step6: By default, messages from Csound are not shown, but they are stored in an internal buffer. You can view them with the printLog() function. If the log is getting too long and confusing, use the clearLog() function. Function tables You can create csound f-tables directly from python lists or numpy arrays Step7: Tables can be plotted in the usual matplotlib way, but ICsound provides a plotTable function which styles the graphs. Step8: You can get the function table values from the csound instance Step9: Tables can also be passed by their variable name in Csound Step10: The following will create 320 tables with 720 points each Step11: Sending instruments You can send instruments to a running csound engine with the %%csound magic. Any syntax errors will be displayed inline. Step12: Channels Csound channels can be used to send values to Csound. They can affect running instances of instruments by using the invalue/chnget opcodes Step13: You can also read the channels from Csound. These channels can be set from ICsound or within instruments with the outvalue/chnset opcodes Step14: Recording the output You can record the realtime output from csound Step15: Remote engines You can also interact with engines through UDP. Note that not all operations are available, notably reading f-tables, but you can send instruments and note events to the remote engine. Step16: Now send notes and instruments from the client Step17: And show the log in the server Step18: Stopping the engine Step19: If we don't need cs_client anymore, we can delete its slot with the %csound line magic (note the single % sign and the negative slot#). The python instance cs_client can then be deleted Step20: Audification Reading Earthquake data through a web API (might take a few minutes) Step21: Instrument to play back the earthquake data stored in a table Step22: Listen Step23: Slower Step24: Quicker Step25: Other tests Another engine
<ASSISTANT_TASK:> Python Code: import numpy as np import matplotlib.pyplot as plt %load_ext csoundmagics Explanation: Using ICsound csoundmagics includes an ICsound class which is adapted from Andrés Cabrera's icsound module. ICsound is bound to the %%csound and %csound magics command. This notebook is an adaptation of Andrés' icsound test notebook. Starting the Csound engine To use ICsound create an ICsound instance: End of explanation cs = ICsound(port=12894) Explanation: Creating an ICsound object automatically starts the engine: End of explanation help(cs.startEngine) Explanation: You can set the properties of the Csound engine with parameters to the startEngine() function. End of explanation cs.startEngine() Explanation: The engine runs in a separate thread, so it doesn't block execution of python. End of explanation %%csound 1 gkinstr init 1 %%csound print i(gkinstr) Explanation: Use the %%csound magic command to directly type csound language code in the cell and send it to the engine. The number after the magic command is optional; it references the slot where the engine is running. If omitted, slot#1 is assumed. End of explanation cs.printLog() Explanation: So where did it print? End of explanation cs.fillTable(1, np.array([8, 7, 9, 1, 1, 1])) cs.fillTable(2, [4, 5, 7, 0, 8, 7, 9, 6]) Explanation: By default, messages from Csound are not shown, but they are stored in an internal buffer. You can view them with the printLog() function. If the log is getting too long and confusing, use the clearLog() function. Function tables You can create csound f-tables directly from python lists or numpy arrays: End of explanation cs.plotTable(1) cs.plotTable(2, reuse=True) plt.grid() Explanation: Tables can be plotted in the usual matplotlib way, but ICsound provides a plotTable function which styles the graphs. End of explanation cs.table(2) cs.makeTable(2, 1024, 10, 1) cs.makeTable(3, 1024, -10, 0.5, 1) cs.plotTable(2) cs.plotTable(3, reuse=True) #ylim((-1.1,1.1)) cs.table(2)[100: 105] Explanation: You can get the function table values from the csound instance: End of explanation %%csound 1 giHalfSine ftgen 0, 0, 1024, 9, .5, 1, 0 cs.plotTable('giHalfSine') Explanation: Tables can also be passed by their variable name in Csound: End of explanation randsig = np.random.random((320, 720)) i = 0 for i, row in enumerate(randsig): cs.fillTable(50 + i, row) print(i, '..', end=' ') cs.plotTable(104) Explanation: The following will create 320 tables with 720 points each: End of explanation %%csound 1 instr 1 asig asds %%csound 1 instr 1 asig oscil 0.5, 440 outs asig, asig %%csound 1 instr 1 asig oscil 0.5, 440 outs asig, asig endin Explanation: Sending instruments You can send instruments to a running csound engine with the %%csound magic. Any syntax errors will be displayed inline. End of explanation cs.setChannel("val", 20) Explanation: Channels Csound channels can be used to send values to Csound. They can affect running instances of instruments by using the invalue/chnget opcodes: End of explanation cs.channel("val") Explanation: You can also read the channels from Csound. These channels can be set from ICsound or within instruments with the outvalue/chnset opcodes: End of explanation cs.startRecord("out.wav") cs.sendScore("i 1 0 1") import time time.sleep(1) cs.stopRecord() !aplay out.wav Explanation: Recording the output You can record the realtime output from csound: End of explanation cs_client = ICsound() cs_client.startClient() cs.clearLog() Explanation: Remote engines You can also interact with engines through UDP. Note that not all operations are available, notably reading f-tables, but you can send instruments and note events to the remote engine. End of explanation cs_client.sendScore("i 1 0 1") cs_client.sendCode("print i(gkinstr)") Explanation: Now send notes and instruments from the client: End of explanation cs.printLog() Explanation: And show the log in the server: End of explanation cs.stopEngine() cs Explanation: Stopping the engine End of explanation %csound -2 del cs_client Explanation: If we don't need cs_client anymore, we can delete its slot with the %csound line magic (note the single % sign and the negative slot#). The python instance cs_client can then be deleted: End of explanation prefix = 'http://service.iris.edu/irisws/timeseries/1/query?' SCNL_parameters = 'net=IU&sta=ANMO&loc=00&cha=BHZ&' times = 'starttime=2005-01-01T00:00:00&endtime=2005-01-02T00:00:00&' output = 'output=ascii' import urllib f = urllib.request.urlopen(prefix + SCNL_parameters + times + output) timeseries = f.read() import ctcsound data = ctcsound.pstring(timeseries).split('\n') dates = [] values = [] for line in data[1:-1]: date, val = line.split() dates.append(date) values.append(float(val)) plt.plot(values) cs.startEngine() cs.fillTable(1, values) Explanation: Audification Reading Earthquake data through a web API (might take a few minutes): End of explanation %%csound 1 instr 1 idur = p3 itable = p4 asig poscil 1/8000, 1/p3, p4 outs asig, asig endin Explanation: Instrument to play back the earthquake data stored in a table: End of explanation cs.sendScore('i 1 0 3 1') Explanation: Listen: End of explanation cs.sendScore('i 1 0 7 1') Explanation: Slower: End of explanation cs.sendScore('i 1 0 1 1') Explanation: Quicker: End of explanation ics = ICsound(bufferSize=64) ics.listInterfaces() %%csound 2 instr 1 asig oscil 0.5, 440 outs asig, asig endin ics.sendScore("i 1 0 0.5") %csound -2 del ics cs.stopEngine() Explanation: Other tests Another engine: End of explanation <END_TASK>
15,961
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Udacity MLND Capstone Project "Determination of students’ interaction patterns with an intelligent tutoring system and study of their correlation with successful learning" Step 1 (sensitivity check, run 1) Step1: Do some preprocessing to group the data by 'Anon Stud Id' and extract features for further analysis Step2: Note to reviewers Step3: Reading from the scratch instead Step4: Making backup for stud_data in HDF5 format Step5: Choosing a student subset for a sensitivity check (note that this step updates stud_data) Step6: Clustering Write a new clustering algorithm that Step7: Choose the pair of columns with best score Step8: Visualising the KMeans clusters Step9: Then, consider adding one more column to further increase the score Step10: As expected, the pair ['frac_1s_hints', 'max_probl_views'] still gives the best score. Now, trying with different clusterers. MeanShift Step11: GaussianMixture Step12: AgglomerativeClustering Step13: Visualising the AgglomerativeClustering clusters Step14: Further clustering of obtained KMeans groups Step15: Visualise obtained clusters Step16: As we see, group 01 contains more students with "gaming" behaviour, so I proceed with group 00 Step17: So, there is a subgroup 001 of 1001 students that do not use many hints. What about the rest (000, 5482 students)? Step18: Splitting group 0000 (students with large 'num_sess' and 'num_probs') Step19: As we see, these two groups represent students with "intermediate experience" (00000) and "largest experience" (00001). During this sensitivity check, I splitted 8082 students (90% of ASSISTments students) into 6 different groups
<ASSISTANT_TASK:> Python Code: # Select test_size and random_state for splitting a subset test_size=0.1 random_state=0 import pandas as pd import numpy as np %matplotlib inline import matplotlib.pyplot as plt import matplotlib.cm as cm import time import gzip import shutil import seaborn as sns from collections import Counter from sklearn.mixture import GaussianMixture from sklearn.cluster import KMeans, MeanShift, estimate_bandwidth, AgglomerativeClustering from sklearn.metrics import silhouette_score #, make_scorer from sklearn.preprocessing import StandardScaler, MinMaxScaler from sklearn.neighbors import kneighbors_graph from sklearn.model_selection import train_test_split Explanation: Udacity MLND Capstone Project "Determination of students’ interaction patterns with an intelligent tutoring system and study of their correlation with successful learning" Step 1 (sensitivity check, run 1) End of explanation def hdf_fixed_write_compress(df): df.to_hdf('data1-step1.hdf','test',mode='w',complib='blosc') return def hdf_fixed_read_compress(): df = pd.read_hdf('data.hdf','test') return df with gzip.open('data1.hdf.gz', 'rb') as f_in, open('data.hdf', 'wb') as f_out: shutil.copyfileobj(f_in, f_out) !ls -lh data.hdf data = hdf_fixed_read_compress() data.head() Explanation: Do some preprocessing to group the data by 'Anon Stud Id' and extract features for further analysis End of explanation def prepare_stud_data_new(df): start_time = time.time() stud_list = df['Anon Student Id'].unique() cols=['num_sess', \ 'num_days', \ 'num_probs', \ 'num_atts', \ 'num_hints', \ 'frac_corr_atts', \ 'frac_3s_atts', \ 'frac_1s_hints', \ 'time_atts', \ 'time_hints', \ 'max_probl_views', \ 'max_atts'] numbers = [] #stud_data = pd.DataFrame(columns=cols) stud_info_df = pd.DataFrame() i = 0 for stud_name in stud_list: stud_info_df = df[df['Anon Student Id'] == stud_name].copy() # total number of days loading the system num_days = len(set(stud_info_df['Day'])) # total number of sessions opened num_sessions = len(set(stud_info_df['Session Id'])) # total number of problems entered num_problems = len(set(stud_info_df['Problem Name'])) # total number of attempts made by the student num_attempts = stud_info_df[stud_info_df['Student Response Type'] == 0].shape[0] # total number of hints made by the student num_hints = stud_info_df[stud_info_df['Student Response Type'] == 1].shape[0] # fraction of short attemps (with time <= 3 sec) if (num_attempts > 0): frac_3s_atts = stud_info_df[(stud_info_df['Student Response Type'] == 0) & (stud_info_df['Duration (sec)'] <= 3.0)].shape[0] / num_attempts else: frac_3s_atts = 0 # fraction of short hints (with time <= 1 sec) if (num_hints > 0): frac_1s_hints = stud_info_df[(stud_info_df['Student Response Type'] == 1) & (stud_info_df['Duration (sec)'] <= 1.0)].shape[0] / num_hints else: frac_1s_hints = 0 # fraction of correct attempts if (num_attempts > 0): fraction_correct_attempts = stud_info_df[(stud_info_df['Student Response Type'] == 0) & (stud_info_df['Outcome'] == 0)].shape[0] / num_attempts else: fraction_correct_attempts = 0 # total number of time spent for attempts (in seconds) total_time_attempts = stud_info_df[stud_info_df['Student Response Type'] == 0]['Duration (sec)'].sum() # total number of time spent for hints (in seconds) total_time_hints = stud_info_df[stud_info_df['Student Response Type'] == 1]['Duration (sec)'].sum() # averaged maximal numbers of 'Problem View' avg_max_problem_views = stud_info_df[['Problem Name', 'Problem View']].groupby(['Problem Name']).agg(np.max).mean()[0] # averaged maximal number of attempts ('x') avg_max_attempts = stud_info_df[['Problem Name', 'x']].groupby(['Problem Name']).agg(np.max).mean()[0] stud_name = i # assign unique numerical ID to each student if num_attempts != 0: avd_time_att = total_time_attempts / num_attempts else: avg_time_att = 0 if num_hints != 0: avg_time_hint = total_time_hints / num_hints else: avg_time_hint = 0 numbers.append([num_sessions, \ num_days, \ num_problems, \ num_attempts, \ num_hints, \ fraction_correct_attempts, \ frac_3s_atts, \ frac_1s_hints, \ total_time_attempts, \ total_time_hints, \ avg_max_problem_views, \ avg_max_attempts]) print("\r\t>>> Progress\t:{:.4%}".format((i + 1)/len(stud_list)), end='') i += 1 stud_data = pd.DataFrame(data=numbers, columns=cols) end_time = time.time() print("\n\t>>> Exec. time\t:{}s".format(end_time-start_time)) return stud_data Explanation: Note to reviewers: this algorithm is quite slow (~45 minutes), so you may consider processing a substantial subset of data (e.g. processing 500,000 rows takes only ~1 minute). End of explanation #stud_data = prepare_stud_data_new(data.head(500000).copy()) #stud_data = prepare_stud_data_new(data.copy()) stud_data = pd.read_hdf('stud_data.hdf','test') Explanation: Reading from the scratch instead: End of explanation #stud_data.to_hdf('stud_data.hdf','test',mode='w',complib='blosc') stud_data.shape stud_data.describe() Explanation: Making backup for stud_data in HDF5 format: End of explanation print(test_size, random_state) stud_data_1, stud_data_2 = train_test_split(stud_data, test_size=test_size, random_state=random_state) stud_data_1.shape[0]/stud_data.shape[0] stud_data = stud_data_1 Explanation: Choosing a student subset for a sensitivity check (note that this step updates stud_data): End of explanation # old name: process_data def transform_data(selected_columns, data): ''' Apply log-transform and MinMaxScaler() to the selected data columns which are not fractions (frac_*) Parameters ========== selected_columns : list list of columns to leave in processed data data : pandas.DataFrame data to process (note that data should contain all selected_columns) Returns ======= log_scaled_data : pandas.DataFrame log-transformed and scaled data selected by selected_columns ''' data.reset_index(drop=True, inplace=True) log_data = data[selected_columns].copy() skewed = log_data.columns.tolist() skewed = [item for item in skewed if not item.startswith('frac_')] log_data[skewed] = log_data[skewed].apply(lambda x: np.log10(x + 1)) scaler = MinMaxScaler().fit(log_data) log_scaled_data = scaler.transform(log_data) log_scaled_data = pd.DataFrame(log_scaled_data, columns=log_data.columns) return log_scaled_data def replace_group_numbers(best_preds): ''' Replace group numbers in best_preds with sorting by group size (so that the largest group is 0, the second largest is 1 etc.) Parameters ========== best_preds : numpy array unsorted array of predictions Returns ======= best_preds_sorted : numpy array sorted array of predictions ''' pp = pd.DataFrame(best_preds, columns = ["old_group"]) dict_pp = {item[0]: i for i, item in enumerate(Counter(best_preds).most_common())} pp['new_group'] = pp['old_group'].replace(dict_pp) best_preds_sorted = np.array(pp['new_group']) return best_preds_sorted def kmeans(log_scaled_data): ''' Apply KMeans clustering algorithm with 2 <= cluster_number <= 6 to log_scaled_data (transformed and scaled by transform_data() function) Parameters ========== log_scaled_data : pandas.DataFrame data log-transormed and MinMaxScaler()-ed for KMeans clustering Returns ======= best_clusterer : sklearn Model clustering algorithm with the largest Silhouette Coefficient best_score : float the largest value of the Silhouette Coefficient best_preds_sorted : numpy.array array with clustering predictions for log_scaled_data (0 is the largest cluster, 1 is the second largest etc.) ''' best_score = 0 for n_clusters in range(2,6): clusterer = KMeans(n_clusters=n_clusters, n_init=10, random_state=0) clusterer.fit(log_scaled_data) preds = clusterer.predict(log_scaled_data) # Calculate the mean silhouette coefficient for the number of clusters chosen score = silhouette_score(log_scaled_data, preds) if best_score < score: best_clusterer = clusterer # Predict the cluster for each data point best_preds = best_clusterer.predict(log_scaled_data) best_score = score best_clusters = n_clusters best_preds_sorted = replace_group_numbers(best_preds) return best_clusterer, best_score, best_preds_sorted Explanation: Clustering Write a new clustering algorithm that: - starts from stud_data or its subset (with monotonic index) - finds a 2-column set with the largest score (using KMeans) and renames it that 0 is the largest group, 1 is the second largest etc. - returns index file (with indices 0, 1, ) that could be used for further analysis End of explanation all_columns = ['num_sess', 'num_days', 'num_probs', 'num_atts', 'num_hints', 'frac_corr_atts', \ 'frac_3s_atts', 'frac_1s_hints', 'time_atts', 'time_hints', 'max_probl_views', 'max_atts'] def choose_pair_columns_kmeans(all_columns, log_scaled_all_data): ''' Selects pair of columns in data that produces clusters with the largest score. In this function, only KMeans clustering algorithm is used Parameters ========== all_columns : list list of columns to look for the pair with the largest score log_scaled_data : pandas DataFrame properly scaled DataFrame with all columns Returns ======= best_columns : list pair of data columns with the largest score best_score : float the largest value of the score best_clusterer : sklearn Model clustering algorithm with the largest score best_preds : numpy.array array with clustering predictions for log_scaled_data (0 is the largest cluster, 1 is the second largest etc.) ''' best_score = 0 best_columns = [] j = 0 l = len(all_columns) num_pairs = (l-1)*l/2 for column in all_columns: selected_columns = [column] columns_to_add = [a for a in all_columns if (a not in selected_columns)] for column1 in columns_to_add: if all_columns.index(column) < all_columns.index(column1): selected_columns = [column, column1] print("\r\t>>> Progress\t:{:.4%}".format((j+1)/num_pairs), end='') j += 1 #log_scaled_data = transform_data(selected_columns, stud_data) clusterer, score, preds = kmeans(log_scaled_all_data[selected_columns]) if score > best_score: best_score = score best_clusterer = clusterer best_preds = preds best_columns = selected_columns.copy() return best_columns, best_score, best_clusterer, best_preds start_time = time.time() log_scaled_all_data = transform_data(all_columns, stud_data) # consider skipping the step below because it takes some time (~5 minutes) best_columns, best_kmeans_score, best_kmeans_clusterer, best_kmeans_preds = choose_pair_columns_kmeans(all_columns, log_scaled_all_data) # Instead run it single time (6 seconds only) #best_columns = ['frac_1s_hints', 'max_probl_views'] #best_kmeans_clusterer, best_kmeans_score, best_kmeans_preds = kmeans(log_scaled_all_data[best_columns]) end_time = time.time() print("\n\t>>> Exec. time\t:{}s".format(end_time-start_time)) print("\t>>> Best pair of cols:", best_columns) print("\t>>> Best score:", best_kmeans_score) print("\t>>> Best clusterer:", best_kmeans_clusterer) print("\t>>> Best preds:", best_kmeans_preds) def preds_to_indices(preds): # gives array and returns array of indices with 1s new_list = [] for i, val in enumerate(preds): if val == 1: new_list.append(i) return np.array(new_list) Explanation: Choose the pair of columns with best score: End of explanation log_scaled_all_data.describe() best_kmeans_preds_mask = preds_to_indices(best_kmeans_preds) log_scaled_all_data_kmeans_0 = log_scaled_all_data.copy()[~log_scaled_all_data.index.isin(best_kmeans_preds_mask)] log_scaled_all_data_kmeans_1 = log_scaled_all_data.copy()[log_scaled_all_data.index.isin(best_kmeans_preds_mask)] plt.scatter(log_scaled_all_data_kmeans_0['frac_1s_hints'], \ log_scaled_all_data_kmeans_0['max_probl_views'], \ alpha=0.6, s=15, c='lightgreen') plt.scatter(log_scaled_all_data_kmeans_1['frac_1s_hints'], \ log_scaled_all_data_kmeans_1['max_probl_views'], \ alpha=0.6, s=15, c='grey') plt.xlim([0.0, 0.6]) plt.ylim([0.0, 0.4]) plt.figtext(x=0.64, y=0.56, s='Group 1', ha='center', size=14, color='black') plt.figtext(x=0.20, y=0.19, s='Group 0', ha='center', size=14, color='darkgreen') ax = plt.gca() ax.set_xlabel('frac_1s_hints', size=14) ax.set_ylabel('max_probl_views', size=14) plt.plot((0.14, 0.14), (0.001, 0.399), 'k--', c='blue') plt.show() print(log_scaled_all_data_kmeans_0.shape, log_scaled_all_data_kmeans_1.shape) Explanation: Visualising the KMeans clusters: End of explanation def cols_iterate_kmeans(selected_columns, best_score, best_clusterer, best_preds): all_columns = ['num_sess', 'num_days', 'num_probs', 'num_atts', \ 'num_hints', 'frac_corr_atts', 'frac_3s_atts', 'frac_1s_hints', \ 'time_atts', 'time_hints', 'max_probl_views', 'max_atts'] columns_to_add = [a for a in all_columns if (a not in selected_columns)] #print(columns_to_add) for column in columns_to_add: print("*"*40) print("*** Trying to add column", column) print("*"*40) selected_columns.append(column) log_scaled_data = transform_data(selected_columns, stud_data) clusterer, score, preds = kmeans(log_scaled_data) if score > best_score: print("!!! Success !!!") best_score = score best_clusterer = clusterer best_preds = preds print("!!! New score is", best_score) print("!!! New best clusterer is", best_clusterer) print("!!! New best selected_columns are", selected_columns) columns_to_add.remove(column) else: print("!!! Last score is equal or worse then our best one") print("!!! According to Occam's razor, remove the column", column) selected_columns.remove(column) print("!!! Still the best selected columns are", selected_columns) return selected_columns, best_score, best_clusterer, best_preds # Just skip this step, it does not give new results: kmeans_clusterer = best_kmeans_clusterer kmeans_score = best_kmeans_score kmeans_preds = best_kmeans_preds selected_columns = best_columns # ['frac_1s_hints', 'max_probl_views'] new_columns, new_kmeans_score, new_kmeans_clusterer, new_kmeans_preds = cols_iterate_kmeans(selected_columns, kmeans_score, kmeans_clusterer, kmeans_preds) if new_kmeans_score > kmeans_score: print("+++ SUCCESS") selected_columns = new_columns best_kmeans_score = new_kmeans_score best_kmeans_clusterer = new_kmeans_clusterer best_kmeans_preds = new_kmeans_preds else: print("--- GIVE UP") Explanation: Then, consider adding one more column to further increase the score: End of explanation def largest_cluster_fraction(preds): ''' calculates the fraction of students that are in the largest group Parameters ========== preds : list list of predictions Returns ======= fraction : float largest fraction of students best_i : integer number of the largest group ''' fraction = 0 ll = len(preds) for i in np.unique(preds): frac = len(preds[preds == i])/ll if frac > fraction: fraction = frac best_i = i return fraction, best_i # Rewrite similar to kmeans procedure !!! def meanshift(log_scaled_data): ''' Apply MeanShift clustering algorithm to log_scaled_data (transformed and scaled by transform_data() function) Number of clusters is selected according to estimate_badwidth procedure with quantiles in np.linspace(0.01, 0.99, 99) Parameters ========== log_scaled_data : pandas.DataFrame data log-transormed and MinMaxScaler()-ed for KMeans clustering Returns ======= best_clusterer : sklearn Model clustering algorithm with the largest Silhouette Coefficient best_score : float the largest value of the Silhouette Coefficient best_preds_sorted : numpy.array array with clustering predictions for log_scaled_data (0 is the largest cluster, 1 is the second largest etc.) cluster_frac : float fraction of students inside the largest group ''' start_time = time.time() best_score = 0 best_cluster_frac = 0 for alpha in np.linspace(0.01, 0.99, 99): bandwidth = estimate_bandwidth(log_scaled_data, quantile=alpha, n_samples=None, random_state=0) clusterer = MeanShift(bandwidth=bandwidth, bin_seeding=True) clusterer.fit(log_scaled_data) preds = clusterer.fit_predict(log_scaled_data) cluster_frac = largest_cluster_fraction(preds)[0] # Calculate the mean silhouette coefficient for the number of clusters chosen try: score = silhouette_score(log_scaled_data, preds) except ValueError: score = 0 print(alpha, clusterer.cluster_centers_.shape[0], score, cluster_frac) # setting cluster_frac > 0.85, the value obtained in KMeans algorithm for ['frac_1s_hints', 'max_probl_views'] if (best_score < score) and (cluster_frac < 0.85): best_clusterer = clusterer best_preds = preds best_score = score best_clusters = clusterer.cluster_centers_.shape[0] best_cluster_frac = cluster_frac print('*'*68) print("Our best model has", best_clusters, "clusters and sihlouette is", best_score) end_time = time.time() print("Running time is {}s".format(end_time-start_time)) print('>'*68) best_preds_sorted = replace_group_numbers(best_preds) cluster_frac = best_cluster_frac return best_clusterer, best_score, best_preds_sorted, cluster_frac # Rinning MeanShift is too slow: runs about 9 min for 1 pair, # and produces too bad results (largest score = 0.56 for reasonable max_fractions < 0.85) start_time = time.time() log_scaled_data = transform_data(best_columns, stud_data) best_meanshift_clusterer, best_meanshift_score, best_meanshift_preds, _ = meanshift(log_scaled_data) print(best_meanshift_clusterer, best_meanshift_score, best_meanshift_preds) end_time = time.time() print("Running time is {}s".format(end_time-start_time)) Explanation: As expected, the pair ['frac_1s_hints', 'max_probl_views'] still gives the best score. Now, trying with different clusterers. MeanShift: End of explanation def gaussmix(log_scaled_data): # GaussianMixture start_time = time.time() max_score = 0 for n_clusters in range(2,6): clusterer = GaussianMixture(random_state=0, n_init=50, n_components=n_clusters).fit(log_scaled_data) preds = clusterer.predict(log_scaled_data) # Calculate the mean silhouette coefficient for the number of clusters chosen score = silhouette_score(log_scaled_data, preds) print("For our model with", clusterer.n_components, "clusters, the sihlouette score is", score) if max_score < score: best_clusterer = clusterer # Predict the cluster for each data point best_preds = best_clusterer.predict(log_scaled_data) max_score = score best_clusters = n_clusters print('*'*68) print("Our best model has", best_clusters, "clusters and sihlouette is", max_score) end_time = time.time() print("Running time is {}s".format(end_time-start_time)) print('>'*68) best_preds_sorted = replace_group_numbers(best_preds) return best_clusterer, max_score, best_preds_sorted def run_clustering_gaussmix(log_scaled_data): best_score = 0 print(">>> GaussianMixture:") clusterer, score, preds = gaussmix(log_scaled_data) if score > best_score: best_clusterer = clusterer best_score = score best_preds = preds print("Best clusterer is", best_clusterer) print("Max score is", best_score) print("Best preds is", best_preds) return best_clusterer, best_score, best_preds # ~0.6 min running time but very small score (~0.39) start_time = time.time() log_scaled_data = transform_data(best_columns, stud_data) gaussmix_best_clusterer, gaussmix_best_score, gaussmix_best_preds = run_clustering_gaussmix(log_scaled_data) print(gaussmix_best_clusterer, gaussmix_best_score, gaussmix_best_preds) end_time = time.time() print("Running time is {}s".format(end_time-start_time)) Explanation: GaussianMixture: End of explanation def agglom(log_scaled_data): # AgglomerativeClustering with 'ward' connectivity start_time = time.time() max_score = 0 for n_clusters in range(2,3): # use only 2 clusters connectivity = kneighbors_graph(log_scaled_data, n_neighbors=100, include_self=False) # make connectivity symmetric connectivity = 0.5 * (connectivity + connectivity.T) clusterer = AgglomerativeClustering(n_clusters=n_clusters, \ linkage='ward', \ connectivity=connectivity) preds = clusterer.fit_predict(log_scaled_data) # Calculate the mean silhouette coefficient for the number of clusters chosen score = silhouette_score(log_scaled_data, preds) print("For our model with", clusterer.n_clusters, "clusters, and the sihlouette score is", score) if max_score < score: best_clusterer = clusterer # Predict the cluster for each data point best_preds = preds max_score = score best_clusters = n_clusters print('*'*68) print("Our best model has", best_clusters, "clusters and sihlouette is", max_score) end_time = time.time() print("Running time is {}s".format(end_time-start_time)) print('>'*68) best_preds_sorted = replace_group_numbers(best_preds) return best_clusterer, max_score, best_preds_sorted def run_clustering_agglom(log_scaled_data): best_score = 0 print(">>> AgglomerativeClustering:") clusterer, score, preds = agglom(log_scaled_data) if score > best_score: best_clusterer = clusterer best_score = score best_preds = preds print("Best clusterer is", best_clusterer) print("Max score is", best_score) print("Best preds is", best_preds) return best_clusterer, best_score, best_preds # Gives results very similar to KMeans but takes ~4 times more running time start_time = time.time() log_scaled_data = transform_data(best_columns, stud_data) best_agglom_clusterer, best_agglom_score, best_agglom_preds = run_clustering_agglom(log_scaled_data) print(best_agglom_clusterer, best_agglom_score, best_agglom_preds) end_time = time.time() print("Running time is {}s".format(end_time-start_time)) Explanation: AgglomerativeClustering: End of explanation best_agglom_preds_mask = preds_to_indices(best_agglom_preds) log_scaled_data_agglom_0 = log_scaled_data.copy()[~log_scaled_data.index.isin(best_agglom_preds_mask)] log_scaled_data_agglom_1 = log_scaled_data.copy()[log_scaled_data.index.isin(best_agglom_preds_mask)] plt.scatter(log_scaled_data_agglom_0['frac_1s_hints'], \ log_scaled_data_agglom_0['max_probl_views'], \ alpha=0.6, s=15, c='lightgreen') plt.scatter(log_scaled_data_agglom_1['frac_1s_hints'], \ log_scaled_data_agglom_1['max_probl_views'], \ alpha=0.6, s=15, c='grey') plt.xlim([0.0, 0.6]) plt.ylim([0.0, 0.4]) plt.figtext(x=0.64, y=0.56, s='Group 1', ha='center', size=14, color='black') plt.figtext(x=0.20, y=0.19, s='Group 0', ha='center', size=14, color='darkgreen') ax = plt.gca() ax.set_xlabel('frac_1s_hints', size=14) ax.set_ylabel('max_probl_views', size=14) #plt.plot((0.145, 0.145), (0.001, 0.399), 'k--', c='blue') plt.show() Explanation: Visualising the AgglomerativeClustering clusters: End of explanation best_kmeans_preds_mask = preds_to_indices(best_kmeans_preds) log_scaled_all_data_kmeans_0 = log_scaled_all_data.copy()[~log_scaled_all_data.index.isin(best_kmeans_preds_mask)] # In this particular splitting, take drop=False to save the initial index # (simplifying students recovery for step 2) log_scaled_all_data_kmeans_0.reset_index(inplace=True, drop=False) log_scaled_all_data_kmeans_0.index start_time = time.time() best_kmeans_columns_0, \ best_kmeans_score_0, \ best_kmeans_clusterer_0, \ best_kmeans_preds_0 = choose_pair_columns_kmeans(all_columns, log_scaled_all_data_kmeans_0) # best_kmeans_columns_0 = ['frac_3s_atts', 'max_probl_views'] # best_kmeans_clusterer_0, best_kmeans_score_0, best_kmeans_preds_0 = kmeans(log_scaled_all_data_kmeans_0[best_kmeans_columns_0]) end_time = time.time() print("\n\t>>> Exec. time\t:{}s".format(end_time-start_time)) print("\t>>> Best pair of cols:", best_kmeans_columns_0) print("\t>>> Best score:", best_kmeans_score_0) print("\t>>> Best clusterer:", best_kmeans_clusterer_0) print("\t>>> Best preds:", best_kmeans_preds_0) print(sum(best_kmeans_preds_0), len(best_kmeans_preds_0), len(best_kmeans_preds_0[best_kmeans_preds_0 == 0])) log_scaled_all_data_kmeans_0.reset_index(inplace=True, drop=True) Explanation: Further clustering of obtained KMeans groups: I start from group 0 that contains 6934 students: End of explanation best_kmeans_preds_mask_0 = preds_to_indices(best_kmeans_preds_0) log_scaled_all_data_kmeans_00 = log_scaled_all_data_kmeans_0.copy()[~log_scaled_all_data_kmeans_0.index.isin(best_kmeans_preds_mask_0)] log_scaled_all_data_kmeans_01 = log_scaled_all_data_kmeans_0.copy()[log_scaled_all_data_kmeans_0.index.isin(best_kmeans_preds_mask_0)] plt.scatter(log_scaled_all_data_kmeans_00[best_kmeans_columns_0[0]], \ log_scaled_all_data_kmeans_00[best_kmeans_columns_0[1]], \ alpha=0.6, s=15, c='lightgreen') plt.scatter(log_scaled_all_data_kmeans_01[best_kmeans_columns_0[0]], \ log_scaled_all_data_kmeans_01[best_kmeans_columns_0[1]], \ alpha=0.6, s=15, c='grey') # plt.xlim([0.0, 0.6]) # plt.ylim([0.0, 0.4]) # plt.figtext(x=0.64, y=0.56, s='Group 01', ha='center', size=14, color='black') # plt.figtext(x=0.20, y=0.69, s='Group 00', ha='center', size=14, color='darkgreen') ax = plt.gca() ax.set_xlabel(best_kmeans_columns_0[0], size=14) ax.set_ylabel(best_kmeans_columns_0[1], size=14) #plt.plot((0.13, 0.13), (0.001, 0.499), 'k--', c='blue') plt.show() Explanation: Visualise obtained clusters: End of explanation len(best_kmeans_preds_0) #best_kmeans_preds_mask_0 = preds_to_indices(best_kmeans_preds_0) # already implemented during group0 visualisation log_scaled_all_data_kmeans_00 = log_scaled_all_data_kmeans_0.copy()[~log_scaled_all_data_kmeans_0.index.isin(best_kmeans_preds_mask_0)] log_scaled_all_data_kmeans_00.reset_index(inplace=True, drop=True) log_scaled_all_data_kmeans_00.index start_time = time.time() best_kmeans_columns_00, \ best_kmeans_score_00, \ best_kmeans_clusterer_00, \ best_kmeans_preds_00 = choose_pair_columns_kmeans(all_columns, log_scaled_all_data_kmeans_00) # best_kmeans_columns_00 = ['frac_3s_atts', 'time_hints'] # best_kmeans_clusterer_00, \ # best_kmeans_score_00, \ # best_kmeans_preds_00 = kmeans(log_scaled_all_data_kmeans_00[best_kmeans_columns_00]) end_time = time.time() print("\n\t>>> Exec. time\t:{}s".format(end_time-start_time)) print("\t>>> Best pair of cols:", best_kmeans_columns_00) print("\t>>> Best score:", best_kmeans_score_00) print("\t>>> Best clusterer:", best_kmeans_clusterer_00) print("\t>>> Best preds:", best_kmeans_preds_00) print(sum(best_kmeans_preds_00), len(best_kmeans_preds_00), len(best_kmeans_preds_00[best_kmeans_preds_00 == 0])) best_kmeans_preds_mask_00 = preds_to_indices(best_kmeans_preds_00) log_scaled_all_data_kmeans_000 = log_scaled_all_data_kmeans_00.copy()[~log_scaled_all_data_kmeans_00.index.isin(best_kmeans_preds_mask_00)] log_scaled_all_data_kmeans_001 = log_scaled_all_data_kmeans_00.copy()[log_scaled_all_data_kmeans_00.index.isin(best_kmeans_preds_mask_00)] plt.scatter(log_scaled_all_data_kmeans_000[best_kmeans_columns_00[0]], \ log_scaled_all_data_kmeans_000[best_kmeans_columns_00[1]], \ alpha=0.6, s=15, c='lightgreen') plt.scatter(log_scaled_all_data_kmeans_001[best_kmeans_columns_00[0]], \ log_scaled_all_data_kmeans_001[best_kmeans_columns_00[1]], \ alpha=0.6, s=15, c='grey') # plt.xlim([0.0, 0.6]) # plt.ylim([0.0, 0.4]) # plt.figtext(x=0.64, y=0.56, s='Group 01', ha='center', size=14, color='black') # plt.figtext(x=0.20, y=0.69, s='Group 00', ha='center', size=14, color='darkgreen') ax = plt.gca() ax.set_xlabel(best_kmeans_columns_00[0], size=14) ax.set_ylabel(best_kmeans_columns_00[1], size=14) #plt.plot((0.13, 0.13), (0.001, 0.499), 'k--', c='blue') plt.show() Explanation: As we see, group 01 contains more students with "gaming" behaviour, so I proceed with group 00: End of explanation log_scaled_all_data_kmeans_000 = log_scaled_all_data_kmeans_00.copy()[~log_scaled_all_data_kmeans_00.index.isin(best_kmeans_preds_mask_00)] log_scaled_all_data_kmeans_000.reset_index(inplace=True, drop=True) log_scaled_all_data_kmeans_000.index start_time = time.time() best_kmeans_columns_000, \ best_kmeans_score_000, \ best_kmeans_clusterer_000, \ best_kmeans_preds_000 = choose_pair_columns_kmeans(all_columns, log_scaled_all_data_kmeans_000) # best_kmeans_columns_000 = ['num_sess', 'num_probs'] # best_kmeans_clusterer_000, \ # best_kmeans_score_000, \ # best_kmeans_preds_000 = kmeans(log_scaled_all_data_kmeans_000[best_kmeans_columns_000]) end_time = time.time() print("\n\t>>> Exec. time\t:{}s".format(end_time-start_time)) print("\t>>> Best pair of cols:", best_kmeans_columns_000) print("\t>>> Best score:", best_kmeans_score_000) print("\t>>> Best clusterer:", best_kmeans_clusterer_000) print("\t>>> Best preds:", best_kmeans_preds_000) print(sum(best_kmeans_preds_000), len(best_kmeans_preds_000), len(best_kmeans_preds_000[best_kmeans_preds_000 == 0])) best_kmeans_preds_mask_000 = preds_to_indices(best_kmeans_preds_000) log_scaled_all_data_kmeans_0000 = log_scaled_all_data_kmeans_000.copy()[~log_scaled_all_data_kmeans_000.index.isin(best_kmeans_preds_mask_000)] log_scaled_all_data_kmeans_0001 = log_scaled_all_data_kmeans_000.copy()[log_scaled_all_data_kmeans_000.index.isin(best_kmeans_preds_mask_000)] plt.scatter(log_scaled_all_data_kmeans_0000[best_kmeans_columns_000[0]], \ log_scaled_all_data_kmeans_0000[best_kmeans_columns_000[1]], \ alpha=0.6, s=15, c='lightgreen') plt.scatter(log_scaled_all_data_kmeans_0001[best_kmeans_columns_000[0]], \ log_scaled_all_data_kmeans_0001[best_kmeans_columns_000[1]], \ alpha=0.6, s=15, c='grey') # plt.figtext(x=0.64, y=0.56, s='Group 01', ha='center', size=14, color='black') # plt.figtext(x=0.20, y=0.69, s='Group 00', ha='center', size=14, color='darkgreen') ax = plt.gca() ax.set_xlabel(best_kmeans_columns_000[0], size=14) ax.set_ylabel(best_kmeans_columns_000[1], size=14) #plt.plot((0.13, 0.13), (0.001, 0.499), 'k--', c='blue') plt.show() Explanation: So, there is a subgroup 001 of 1001 students that do not use many hints. What about the rest (000, 5482 students)? End of explanation log_scaled_all_data_kmeans_0000 = log_scaled_all_data_kmeans_000.copy()[~log_scaled_all_data_kmeans_000.index.isin(best_kmeans_preds_mask_000)] log_scaled_all_data_kmeans_0000.reset_index(inplace=True, drop=True) log_scaled_all_data_kmeans_0000.index start_time = time.time() best_kmeans_columns_0000, \ best_kmeans_score_0000, \ best_kmeans_clusterer_0000, \ best_kmeans_preds_0000 = choose_pair_columns_kmeans(all_columns, log_scaled_all_data_kmeans_0000) # best_kmeans_columns_0000 = ['num_sess', 'num_probs'] # best_kmeans_clusterer_0000, \ # best_kmeans_score_0000, \ # best_kmeans_preds_0000 = kmeans(log_scaled_all_data_kmeans_0000[best_kmeans_columns_0000]) end_time = time.time() print("\n\t>>> Exec. time\t:{}s".format(end_time-start_time)) print("\t>>> Best pair of cols:", best_kmeans_columns_0000) print("\t>>> Best score:", best_kmeans_score_0000) print("\t>>> Best clusterer:", best_kmeans_clusterer_0000) print("\t>>> Best preds:", best_kmeans_preds_0000) print(sum(best_kmeans_preds_0000), \ len(best_kmeans_preds_0000), \ len(best_kmeans_preds_0000[best_kmeans_preds_0000 == 0])) best_kmeans_preds_mask_0000 = preds_to_indices(best_kmeans_preds_0000) log_scaled_all_data_kmeans_00000 = log_scaled_all_data_kmeans_0000.copy()[~log_scaled_all_data_kmeans_0000.index.isin(best_kmeans_preds_mask_0000)] log_scaled_all_data_kmeans_00001 = log_scaled_all_data_kmeans_0000.copy()[log_scaled_all_data_kmeans_0000.index.isin(best_kmeans_preds_mask_0000)] plt.scatter(log_scaled_all_data_kmeans_00000[best_kmeans_columns_0000[0]], \ log_scaled_all_data_kmeans_00000[best_kmeans_columns_0000[1]], \ alpha=0.6, s=15, c='lightgreen') plt.scatter(log_scaled_all_data_kmeans_00001[best_kmeans_columns_0000[0]], \ log_scaled_all_data_kmeans_00001[best_kmeans_columns_0000[1]], \ alpha=0.6, s=15, c='grey') # plt.xlim([0.0, 0.6]) # plt.ylim([0.0, 0.4]) # plt.figtext(x=0.64, y=0.56, s='Group 01', ha='center', size=14, color='black') # plt.figtext(x=0.20, y=0.69, s='Group 00', ha='center', size=14, color='darkgreen') ax = plt.gca() ax.set_xlabel(best_kmeans_columns_0000[0], size=14) ax.set_ylabel(best_kmeans_columns_0000[1], size=14) #plt.plot((0.13, 0.13), (0.001, 0.499), 'k--', c='blue') plt.show() Explanation: Splitting group 0000 (students with large 'num_sess' and 'num_probs') End of explanation group1_index = np.array(log_scaled_all_data_kmeans_1.index) len(group1_index) group2_index = np.array(log_scaled_all_data_kmeans_01['index']) len(group2_index) group3_index = np.array(log_scaled_all_data_kmeans_001['index']) len(group3_index) group4_index = np.array(log_scaled_all_data_kmeans_0001['index']) len(group4_index) group5_index = np.array(log_scaled_all_data_kmeans_00000['index']) len(group5_index) group6_index = np.array(log_scaled_all_data_kmeans_00001['index']) len(group6_index) def create_joint_cluster_index(): ''' Saves group index files into cluster_index.csv for further analysis ''' cluster_index_lst = [] for i in range(len(stud_data)+1): if i in group1_index: cluster_index_lst.append(1) elif i in group2_index: cluster_index_lst.append(2) elif i in group3_index: cluster_index_lst.append(3) elif i in group4_index: cluster_index_lst.append(4) elif i in group5_index: cluster_index_lst.append(5) elif i in group6_index: cluster_index_lst.append(6) print(Counter(cluster_index_lst)) cluster_index = pd.Series(cluster_index_lst, dtype=int) cluster_index.to_csv('cluster_index_run1.csv') return create_joint_cluster_index() ! ls -lh cluster_index_run1.csv Explanation: As we see, these two groups represent students with "intermediate experience" (00000) and "largest experience" (00001). During this sensitivity check, I splitted 8082 students (90% of ASSISTments students) into 6 different groups: - group 1, 1148 students with large 'frac_1s_hints' ("gaming" behaviour); - group 2, 451 students with small 'frac_1s_hints' and large 'frac_3s_atts' ("gaming" behaviour); - group 3, 1001 students with small 'time_hints' ("non-gaming" behaviour, small usage of hints); - group 4, 2151 students with small 'num_sess' and 'num_probs' ("non-gaming" behaviour, large usage of hints, small experience); - group 5, 1734 students with medium 'num_sess' and 'num_probs' ("non-gaming" behaviour, large usage of hints, medium experience); - group 6, 1597 students with large 'num_sess' and 'num_probs' ("non-gaming" behaviour, large usage of hints, large experience). The final result of this step is the joint cluster index that contains numbers 1-6 for each student: End of explanation <END_TASK>
15,962
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Introduction to NLTK We have seen how to do some basic text processing in Python, now we introduce an open source framework for natural language processing that can further help to work with human languages Step1: Tokens The basic atomic part of each text are the tokens. A token is the NLP name for a sequence of characters that we want to treat as a group. We have seen how we can extract tokens by splitting the text at the blank spaces. NTLK has a function word_tokenize() for it Step2: 21 tokens extracted, which include words and punctuation. Note that the tokens are different than what a split by blank spaces would obtained, e.g. "can't" is by NTLK considered TWO tokens Step3: And we can apply it to an entire book, "The Prince" by Machiavelli that we used last time Step4: As mentioned above, the NTLK tokeniser works in a more sophisticated way than just splitting by spaces, therefore we got this time more tokens. Sentences NTLK has a function to tokenise a text not in words but in sentences. Step5: As you see, it is not splitting just after each full stop but check if it's part of an acronym (U.S.) or a number (0.99). It also splits correctly sentences after question or exclamation marks but not after commas. Step6: Most common tokens What are the 20 most frequently occurring (unique) tokens in the text? What is their frequency? The NTLK FreqDist class is used to encode “frequency distributions”, which count the number of times that something occurs, for example a token. Its most_common() method then returns a list of tuples where each tuple is of the form (token, frequency). The list is sorted in descending order of frequency. Step7: Comma is the most common Step8: We can also remove any capital letters before tokenising Step9: Now we removed the punctuation and the capital letters but the most common token is "the", not a significative word ... As we have seen last time, these are so-called stop words that are very common and are normally stripped from a text when doing these kind of analysis. Meaningful most common tokens A simple approach could be to filter the tokens that have a length greater than 5 and frequency of more than 150. Step10: This would work but would leave out also tokens such as I and you which are actually significative. The better approach - that we have seen earlier how - is to remove stopwords using external files containing the stop words. NLTK has a corpus of stop words in several languages Step11: Now we excluded words such as the but we can improve further the list by looking at semantically similar words, such as plural and singular versions. Step12: Stemming Above, in the list of words we have both prince and princes which are respectively the singular and plural version of the same word (the stem). The same would happen with verb conjugation (love and loving are considered different words but are actually inflections of the same verb). Stemmer is the tool that reduces such inflectional forms into their stem, base or root form and NLTK has several of them (each with a different heuristic algorithm). Step13: And now we apply one of the NLTK stemmer, the Porter stemmer Step14: As you see, all 5 different words have been reduced to the same stem and would be now the same lexical token. Step15: Now the word princ is counted 281 times, exactly like the sum of prince and princes. A note here Step16: Lemma Lemmatization usually refers to doing things properly with the use of a vocabulary and morphological analysis of words, normally aiming to remove inflectional endings only and to return the base or dictionary form of a word, which is known as the lemma. While a stemmer operates on a single word without knowledge of the context, a lemmatiser can take the context in consideration. NLTK has also a built-in lemmatiser, so let's see it in action Step17: We tell the lemmatise that the words are nouns. In this case it considers the same lemma words such as list (singular noun) and lists (plural noun) but leave as they are the other words. Step18: We get a different result if we say that the words are verbs. They have all the same lemma, in fact they could be all different inflections or conjugation of a verb. The type of words that can be used are Step19: It works with different adjectives, it doesn't look only at prefixes and suffixes. You would wonder why stemmers are used, instead of always using lemmatisers Step20: Yes, the lemma now is prince. But note that we consider all words in the book as nouns, while actually a proper way would be to apply the correct type to each single word. Part of speech (PoS) In traditional grammar, a part of speech (abbreviated form Step21: The NLTK function pos_tag() will tag each token with the estimated PoS. NLTK has 13 categories of PoS. You can check the acronym using the NLTK help function Step22: Which are the most common PoS in The Prince book? Step24: It's not nouns (NN) but interections (IN) such as preposition or conjunction. Extra note
<ASSISTANT_TASK:> Python Code: sampleText1 = "The Elephant's 4 legs: THE Pub! You can't believe it or can you, the believer?" sampleText2 = "Pierre Vinken, 61 years old, will join the board as a nonexecutive director Nov. 29." Explanation: Introduction to NLTK We have seen how to do some basic text processing in Python, now we introduce an open source framework for natural language processing that can further help to work with human languages: NLTK (Natural Language ToolKit). Tokenise a text Let's start with a simple text in a Python string: End of explanation import nltk s1Tokens = nltk.word_tokenize(sampleText1) s1Tokens len(s1Tokens) Explanation: Tokens The basic atomic part of each text are the tokens. A token is the NLP name for a sequence of characters that we want to treat as a group. We have seen how we can extract tokens by splitting the text at the blank spaces. NTLK has a function word_tokenize() for it: End of explanation s2Tokens = nltk.word_tokenize(sampleText2) s2Tokens Explanation: 21 tokens extracted, which include words and punctuation. Note that the tokens are different than what a split by blank spaces would obtained, e.g. "can't" is by NTLK considered TWO tokens: "can" and "n't" (= "not") while a tokeniser that splits text by spaces would consider it a single token: "can't". Let's see another example: End of explanation # If you would like to work with the raw text you can use 'bookRaw' with open('../datasets/ThePrince.txt', 'r') as f: bookRaw = f.read() bookTokens = nltk.word_tokenize(bookRaw) bookText = nltk.Text(bookTokens) # special format nBookTokens= len(bookTokens) # or alternatively len(bookText) print ("*** Analysing book ***") print ("The book is {} chars long".format (len(bookRaw))) print ("The book has {} tokens".format (nBookTokens)) Explanation: And we can apply it to an entire book, "The Prince" by Machiavelli that we used last time: End of explanation text1 = "This is the first sentence. A liter of milk in the U.S. costs $0.99. Is this the third sentence? Yes, it is!" sentences = nltk.sent_tokenize(text1) len(sentences) sentences Explanation: As mentioned above, the NTLK tokeniser works in a more sophisticated way than just splitting by spaces, therefore we got this time more tokens. Sentences NTLK has a function to tokenise a text not in words but in sentences. End of explanation sentences = nltk.sent_tokenize(bookRaw) # extract sentences nSent = len(sentences) print ("The book has {} sentences".format (nSent)) print ("and each sentence has in average {} tokens".format (nBookTokens / nSent)) Explanation: As you see, it is not splitting just after each full stop but check if it's part of an acronym (U.S.) or a number (0.99). It also splits correctly sentences after question or exclamation marks but not after commas. End of explanation def get_top_words(tokens): # Calculate frequency distribution fdist = nltk.FreqDist(tokens) return fdist.most_common() topBook = get_top_words(bookTokens) # Output top 20 words topBook[:20] Explanation: Most common tokens What are the 20 most frequently occurring (unique) tokens in the text? What is their frequency? The NTLK FreqDist class is used to encode “frequency distributions”, which count the number of times that something occurs, for example a token. Its most_common() method then returns a list of tuples where each tuple is of the form (token, frequency). The list is sorted in descending order of frequency. End of explanation topWords = [(freq, word) for (word,freq) in topBook if word.isalpha() and freq > 400] topWords Explanation: Comma is the most common: we need to remove the punctuation. Most common alphanumeric tokens We can use isalpha() to check if the token is a word and not punctuation. End of explanation def preprocessText(text, lowercase=True): if lowercase: tokens = nltk.word_tokenize(text.lower()) else: tokens = nltk.word_tokenize(text) return [word for word in tokens if word.isalpha()] bookWords = preprocessText(bookRaw) topBook = get_top_words(bookWords) # Output top 20 words topBook[:20] print ("*** Analysing book ***") print ("The text has now {} words (tokens)".format (len(bookWords))) Explanation: We can also remove any capital letters before tokenising: End of explanation meaningfulWords = [word for (word,freq) in topBook if len(word) > 5 and freq > 80] sorted(meaningfulWords) Explanation: Now we removed the punctuation and the capital letters but the most common token is "the", not a significative word ... As we have seen last time, these are so-called stop words that are very common and are normally stripped from a text when doing these kind of analysis. Meaningful most common tokens A simple approach could be to filter the tokens that have a length greater than 5 and frequency of more than 150. End of explanation from nltk.corpus import stopwords stopwordsEN = set(stopwords.words('english')) # english language betterWords = [w for w in bookWords if w not in stopwordsEN] topBook = get_top_words(betterWords) # Output top 20 words topBook[:20] Explanation: This would work but would leave out also tokens such as I and you which are actually significative. The better approach - that we have seen earlier how - is to remove stopwords using external files containing the stop words. NLTK has a corpus of stop words in several languages: End of explanation 'princes' in betterWords betterWords.count("prince") + betterWords.count("princes") Explanation: Now we excluded words such as the but we can improve further the list by looking at semantically similar words, such as plural and singular versions. End of explanation input1 = "List listed lists listing listings" words1 = input1.lower().split(' ') words1 Explanation: Stemming Above, in the list of words we have both prince and princes which are respectively the singular and plural version of the same word (the stem). The same would happen with verb conjugation (love and loving are considered different words but are actually inflections of the same verb). Stemmer is the tool that reduces such inflectional forms into their stem, base or root form and NLTK has several of them (each with a different heuristic algorithm). End of explanation porter = nltk.PorterStemmer() [porter.stem(t) for t in words1] Explanation: And now we apply one of the NLTK stemmer, the Porter stemmer: End of explanation stemmedWords = [porter.stem(w) for w in betterWords] topBook = get_top_words(stemmedWords) topBook[:20] # Output top 20 words Explanation: As you see, all 5 different words have been reduced to the same stem and would be now the same lexical token. End of explanation from nltk.stem.snowball import SnowballStemmer stemmerIT = SnowballStemmer("italian") inputIT = "Io ho tre mele gialle, tu hai una mela gialla e due pere verdi" wordsIT = inputIT.split(' ') [stemmerIT.stem(w) for w in wordsIT] Explanation: Now the word princ is counted 281 times, exactly like the sum of prince and princes. A note here: Stemming usually refers to a crude heuristic process that chops off the ends of words in the hope of achieving this goal correctly most of the time, and often includes the removal of derivational affixes. Prince and princes become princ. A different flavour is the lemmatisation that we will see in one second, but first a note about stemming in other languages than English. Stemming in other languages Snowball is an improvement created by Porter: a language to create stemmers and have rules for many more languages than English. For example Italian: End of explanation from nltk.stem import WordNetLemmatizer lemmatizer = WordNetLemmatizer() words1 [lemmatizer.lemmatize(w, 'n') for w in words1] # n = nouns Explanation: Lemma Lemmatization usually refers to doing things properly with the use of a vocabulary and morphological analysis of words, normally aiming to remove inflectional endings only and to return the base or dictionary form of a word, which is known as the lemma. While a stemmer operates on a single word without knowledge of the context, a lemmatiser can take the context in consideration. NLTK has also a built-in lemmatiser, so let's see it in action: End of explanation [lemmatizer.lemmatize(w, 'v') for w in words1] # v = verbs Explanation: We tell the lemmatise that the words are nouns. In this case it considers the same lemma words such as list (singular noun) and lists (plural noun) but leave as they are the other words. End of explanation words2 = ['good', 'better'] [porter.stem(w) for w in words2] [lemmatizer.lemmatize(w, 'a') for w in words2] Explanation: We get a different result if we say that the words are verbs. They have all the same lemma, in fact they could be all different inflections or conjugation of a verb. The type of words that can be used are: 'n' = noun, 'v'=verb, 'a'=adjective, 'r'=adverb End of explanation lemmatisedWords = [lemmatizer.lemmatize(w, 'n') for w in betterWords] topBook = get_top_words(lemmatisedWords) topBook[:20] # Output top 20 words Explanation: It works with different adjectives, it doesn't look only at prefixes and suffixes. You would wonder why stemmers are used, instead of always using lemmatisers: stemmers are much simpler, smaller and faster and for many applications good enough. Now we lemmatise the book: End of explanation text1 = "Children shouldn't drink a sugary drink before bed." tokensT1 = nltk.word_tokenize(text1) nltk.pos_tag(tokensT1) Explanation: Yes, the lemma now is prince. But note that we consider all words in the book as nouns, while actually a proper way would be to apply the correct type to each single word. Part of speech (PoS) In traditional grammar, a part of speech (abbreviated form: PoS or POS) is a category of words which have similar grammatical properties.  For example, an adjective (red, big, quiet, ...) describe properties while a verb (throw, walk, have) describe actions or states. Commonly listed parts of speech are noun, verb, adjective, adverb, pronoun, preposition, conjunction, interjection. End of explanation nltk.help.upenn_tagset('RB') Explanation: The NLTK function pos_tag() will tag each token with the estimated PoS. NLTK has 13 categories of PoS. You can check the acronym using the NLTK help function: End of explanation tokensAndPos = nltk.pos_tag(bookTokens) posList = [thePOS for (word, thePOS) in tokensAndPos] fdistPos = nltk.FreqDist(posList) fdistPos.most_common(5) nltk.help.upenn_tagset('IN') Explanation: Which are the most common PoS in The Prince book? End of explanation # Parsing sentence structure text2 = nltk.word_tokenize("Alice loves Bob") grammar = nltk.CFG.fromstring( S -> NP VP VP -> V NP NP -> 'Alice' | 'Bob' V -> 'loves' ) parser = nltk.ChartParser(grammar) trees = parser.parse_all(text2) for tree in trees: print(tree) Explanation: It's not nouns (NN) but interections (IN) such as preposition or conjunction. Extra note: Parsing the grammar structure Words can be ambiguous and sometimes is not easy to understand which kind of POS is a word, for example in the sentence "visiting aunts can be a nuisance", is visiting a verb or an adjective? Tagging a PoS depends on the context, which can be ambiguous. Making sense of a sentence is easier if it follows a well-defined grammatical structure, such as : subject + verb + object NLTK allows to define a formal grammar which can then be used to parse a text. The NLTK ChartParser is a procedure for finding one or more trees (sentences have internal organisation that can be represented using a tree) corresponding to a grammatically well-formed sentence. End of explanation <END_TASK>
15,963
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: SciPy for Economists Scipy provides many user-friendly and efficient numerical routines, e.g. numerical integration and optimization. The full documentation is available at http Step1: Let us have a look at the relationship. Step2: Estimation using Linear Algebra Tools We can determine the $\hat{\beta} = (X^T X)^{-1}X^T Y$ using basic linear algebra tools from NumPy. Step3: Estimation using Optimization Tools Let us now determine $\hat{\beta}$ using Maximum Likelihood (ML) estimation. So, we need to maximize the following log-likelihood function Step4: Formatting
<ASSISTANT_TASK:> Python Code: # standard library import numpy as np # Parametrization num_agents = 1000 num_covars = 3 betas_true = np.array([0.22, 0.30, -0.1]).T sd_true = 0.01 # Sampling of observables np.random.seed(123) X = np.random.rand(num_agents, num_covars) X[:,0] = 1 # Sampling disturbances eps = np.random.normal(loc=0.0, scale=sd_true, size=num_agents) # Create endogenous outcome idx_true = np.dot(X, betas_true) Y = idx_true + eps # Checks assert (X.dtype == 'float') assert (Y.dtype == 'float') assert (np.all(np.isfinite(X))) assert (np.all(np.isfinite(Y))) assert (X.shape == (num_agents, num_covars)) assert (Y.shape == (num_agents, )) assert (np.all(X[:,0] == 1.0)) Explanation: SciPy for Economists Scipy provides many user-friendly and efficient numerical routines, e.g. numerical integration and optimization. The full documentation is available at http://docs.scipy.org/doc/scipy/reference/. We will use the provided tools to simulate and estimate and Ordinary Least Squares (OLS) regression: $Y = X\beta + \epsilon$ We will proceed in three steps: Simulated Sample Estimate Model usig Linear Algebra Tools (NumPy) Estimate Model usig Optimization Tools (SciPy) Of course, OLS and other statistical models are readily available in the StatsModels Library http://statsmodels.sourceforge.net/. Simulate Sample End of explanation %pylab inline import matplotlib.pyplot as plt fig = plt.figure() ax = fig.add_subplot(111) ax.set_ylabel(r'$Y$'), ax.set_xlabel(r'$ X\beta $') ax.plot(idx_true, Y, 'o') Explanation: Let us have a look at the relationship. End of explanation # Let us get the estimates. beta_hat = np.dot(np.dot(np.linalg.inv(np.dot(X.T,X)), X.T), Y) sd_hat = np.sqrt(np.var(Y - np.dot(X, beta_hat))) # Let us have a look now. print('Results for beta', beta_hat, ' Results for sd', sd_hat) Explanation: Estimation using Linear Algebra Tools We can determine the $\hat{\beta} = (X^T X)^{-1}X^T Y$ using basic linear algebra tools from NumPy. End of explanation # standard library from scipy.optimize import minimize from scipy.stats import norm # Auxiliary functions. def sample_likelihood(paras, X, Y): ''' Construct sample likelihood. ''' # Antibugging. assert (isinstance(paras, np.ndarray)) assert (paras.dtype == 'float') assert (X.ndim == 2), (Y.ndim == 2) # Auxiliary objects. num_agents = Y.shape[0] # Summing over the sample. contribs = 0.0 for i in range(num_agents): contrib = individual_likelihood(paras, X[i,:], Y[i]) contribs += contrib # Modifications. contribs = np.mean(contribs) # Finishing. return contribs def individual_likelihood(paras, x, y): ''' This function determines the an individual's contribution to the sample likelihood. ''' # Antibugging. assert (isinstance(paras, np.ndarray)) assert (paras.dtype == 'float') assert (x.ndim == 1), (y.ndim == 1) # Distribute parameters. betas, sd = paras[:-1], paras[-1] # Calculate likelihood contribution. resid = (y - np.dot(x, betas))/sd contrib = (1.0/sd)*norm.pdf(resid) # Modifications. contrib = np.clip(contrib, 1e-20, np.inf) contrib = -np.log(contrib) # Finishing. return contrib ''' Main calculations. ''' # Construct parameters. paras = np.concatenate((betas_true, [sd_true])) # Single individual. individual_likelihood(paras, X[1,:], Y[1]) # Full sample. sample_likelihood(paras, X, Y) # Optimization. x0 = paras #x0 = [0.0, 0.0, 0.0, 1.0] for optimizer in ['BFGS', 'Nelder-Mead']: rslt = minimize(sample_likelihood, x0, args=(X, Y), method=optimizer) Explanation: Estimation using Optimization Tools Let us now determine $\hat{\beta}$ using Maximum Likelihood (ML) estimation. So, we need to maximize the following log-likelihood function: $$L(\beta, \sigma) = \sum_{i = 1, ..., N} \log\left(\frac{1}{\sigma}\phi\left(\frac{Y_i - X_i\beta}{\sigma}\right)\right)$$ SciPy offers a convenient interface to alternative optimization algorithms. Let us check it out online. End of explanation import urllib; from IPython.core.display import HTML HTML(urllib.urlopen('http://bit.ly/1Ki3iXw').read()) Explanation: Formatting End of explanation <END_TASK>
15,964
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Graded = 7/8 Step1: & for multiple parameters Step2: 2) What genres are most represented in the search results? Edit your previous printout to also display a list of their genres in the format GENRE_1, GENRE_2, GENRE_3". If there are no genres, print "No genres listed". Tip Step3: ANSWER Step4: 3) Use a for loop to determine who BESIDES Lil Wayne has the highest popularity rating. Is it the same artist who has the largest number of followers? Step5: ANSWER Step6: 5) Pick two of your favorite Lils to fight it out, and use their IDs to print out their top tracks. Step7: Will the world explode if a musician swears? Get an average popularity for their explicit songs vs. their non-explicit songs. How many minutes of explicit songs do they have? Non-explicit? Step8: QUESTION Step9: 7) Since we're talking about Lils, what about Biggies? How many total "Biggie" artists are there? How many total "Lil"s? If you made 1 request every 5 seconds, how long would it take to download information on all the Lils vs the Biggies? Step10: 8) Out of the top 50 "Lil"s and the top 50 "Biggie"s, who is more popular on average?
<ASSISTANT_TASK:> Python Code: import requests response = requests.get('https://api.spotify.com/v1/search?query=lil&type=artist&market=US&limit=50') Explanation: Graded = 7/8 End of explanation data = response.json() data.keys() artist_data = data['artists'] artist_data.keys() lil_names = artist_data['items'] #lil_names = list of dictionaries = list of artist name, popularity, type, genres etc Explanation: & for multiple parameters End of explanation for names in lil_names: if not names['genres']: print(names['name'], names['popularity'], "there are no genres listed") else: print(names['name'], names['popularity'], names["genres"]) #Join all the lists of genres in the dictionary and then count the number of elements in it Explanation: 2) What genres are most represented in the search results? Edit your previous printout to also display a list of their genres in the format GENRE_1, GENRE_2, GENRE_3". If there are no genres, print "No genres listed". Tip: "how to join a list Python" might be a helpful search End of explanation #ANSWER: all_genres = [] for artist in lil_names: print("All genres we've heard of:", all_genres) #The conditional: None print("Current artist has:", artist['genres']) all_genres = all_genres + artist['genres'] #genre_list = ", ".join(artist['genres']) #print(artist['name'], ":", genre_list) all_genres.count('dirty soup rap') all_genres.count('crunk') #This is bad because dirty south rap shows up four times. We need a unique list of genres for genre in all_genres: genre_count = all_genres.count(genre) print(genre, "shows up", genre_count, "times") #To remove duplicates. You need to turn a list into a set. unique_genres = set(list_with_duplicates) for genre in unique_genres: genre_count = all_genres.count(genre) print(genre, "shows up", genre_count, "times") #There is a library that comes with Python called Collections #Inside of it is a magic thing called Counter. import collections # will import the whole collections #you can also type from collections import Counter #all_genres is a list of strings of genrs with duplicates #counter will count all te genres for us counts = collections.Counter(all_genres) counts.most_common(4) #will give you the four most common genres #HOW TO AUTOMATE GETTING ALL THE RESULTS response = requests.get("https://api.spotify.com/v1/search?query=lil&type=artist&market=US&limit=10") small_data = response.json() small_data['artists'] print(len(small_data['artists']['items'])) #We only get 10 artists print(data['artists']['total']) import math page_count = math.ceil(4502/50) #math.ceil rounds up #math.ceil(page_count) page_count #First page artists 1-50: #https://api.spotify.com/v1/search?query=lil&type=artist&market=US&limit=50 #Second page artists 51-100: #https://api.spotify.com/v1/search?query=lil&type=artist&market=US&limit=50&offset=50 #Third page artists 101-150: #https://api.spotify.com/v1/search?query=lil&type=artist&market=US&limit=50&offset=100 #Fourth page artists 151-200: #https://api.spotify.com/v1/search?query=lil&type=artist&market=US&limit=50&offset=150 for page in [0, 1, 2, 3, 4]: offset = (page) * 50 #because page 2 is 50 and 2-1 = 1 x 50 = 50 print("We are on page", page, "with an offset of", offset) for page in range(91): #Get a page offset = page * 50 print("We are on page", page, "with an offset of", offset) #Make the request with a changed offset ?offset [offset] #data = response.json() #add all our new artists to our list of existing artists #all_artists = all_artists + data['artists]['items] print("Successfully retrieved", len(all_artists), "artists") Explanation: ANSWER: for artist in artists: print(artist['name'], artist['popularity']) if len(artist['genres']) > 0: genres = ", ".join(artist['genres']) print("Genre list: ", genres else: print("No genres listed") OR if len(artist['name']) == 0: OR if not len(artist['genres']) == 0: "-".join(your_list) to join lists End of explanation #TA-Stephan:can't just print the names yourself. The code must do it. for popularity in lil_names: print(popularity['name'], popularity['popularity'], popularity['followers']) print("Lil Yachty, Lil Uzi Vert, Lil Jon have the highest popularity ratings besides Lil Wayne, and they do not have the largest number of followers.") Explanation: 3) Use a for loop to determine who BESIDES Lil Wayne has the highest popularity rating. Is it the same artist who has the largest number of followers? End of explanation for kim in lil_names: print(kim['name'], kim['id']) response = requests.get("https://api.spotify.com/v1/artists/5tth2a3v0sWwV1C7bApBdX/") kim_data = response.json() #print(kim_data) kim_followers = kim_data['followers'] total_kim_followers = kim_followers['total'] #print(total_kim_followers) for artists in lil_names: if artists["followers"]["total"] > total_kim_followers: print(artists['name'], artists['popularity']) #ANSWER: for artist in artists: #print("Looking at", artist['name']) if artist['name'] == "Lil' Kim": print("Found Lil Kim") print(artist['popularity']) else: pass #print("Not Lil Kim") lil_kim_popularity = 62 for artist in artists: if artist['popularity'] > lil_kim_popularity: print(artist['name'], "is more popular with a score of", artist['popularity']) more_popular_than_lil_kim.append(artist['name']) else: print(artist['name'], "is less popular with a score of", artist['popularity']) print("#### More popular than Lil Kim ####"): print(artist_name) more_popular_string = ", ".join(more_popular_than_lil_kim) print("Artists mroe popular than Lil Kim are:", more_popular_string) Explanation: ANSWER: #jonathansoma.com/site/lede/foundations/python-patterns/looping-problems/ second_most_popular_name = "" second_most_popular_score = 0 for artist in artists: print("Looking at", artist['name]', "who has a popularity score of", artist['popularity']) #THE CONDITIONAL aka what you are testing print("Comparing", artist['popularity'], "to", most_popular_score) if artist['popularity'] > most_popular_score and artist['name'] != 'Lil Wayne': OR if artist['popularity'] &gt; most_popular_score: if artist['name'] == "Lil Wayne": print("Nice try, Lil Wayne, we don't care") else: print("Not Lil Wayne, updating our notebook") #The change, aka what you are keeping track of most_popular_name = artist['name'] most_popular_score = artist['popularity'] print(most_popular_name, most_popular_score) But what if more than one person has the highest score? Aggregation Problem. When you're looping through a series of objects and sometimes you want to add one of those objects to a different list. target score = 72 initial condition second_best_artist = [] for artist in artists: print("Looking at", artist['name'], "who has a popularity of", ) #2: Conditional. When we want to add someone to our list if artist['popularity'] == 72: print("!!! The artist's popularity is 72") #The Change #Add that artist to our list #.append(newthing) is how we do that in Python second_best_artists.append(artist['name']) print("Our second best artists are:") for artist in second_best_artist: print(artist) Print a list of Lil's that are more popular than Lil' Kim. End of explanation #Let's pick Lil Wayne and Lil Mama because I don't know who most of these people are wayne_id = "55Aa2cqylxrFIXC767Z865" response = requests.get("https://api.spotify.com/v1/artists/" + wayne_id + "/top-tracks?country=US") wayne_data = response.json() top_wayne_tracks = wayne_data['tracks'] for track in top_wayne_tracks: print(track["name"]) mama_id = "5qK5bOC6wLtuLhG5KvU17c" response = requests.get("https://api.spotify.com/v1/artists/" + mama_id + "/top-tracks?country=US") mama_data = response.json() top_mama_tracks = mama_data['tracks'] for track in top_mama_tracks: print(track["name"]) Explanation: 5) Pick two of your favorite Lils to fight it out, and use their IDs to print out their top tracks. End of explanation wayne_explicit_count = 0 wayne_exp_popularity_count = 0 wayne_ok_count = 0 wayne_ok_popularity_count = 0 wayne_explicit_len = 0 wayne_ok_len = 0 for track in top_wayne_tracks: print(track['name'], track['explicit'], track['popularity'], track["duration_ms"]) if True: wayne_explicit_count = wayne_explicit_count + 1 wayne_exp_popularity_count = wayne_exp_popularity_count + int(track['popularity']) wayne_avg_pop = wayne_exp_popularity_count / wayne_explicit_count wayne_explicit_len = wayne_explicit_len + int(track["duration_ms"]) if not track['explicit']: wayne_ok_count = wayne_ok_count + 1 wayne_ok_popularity_count = wayne_ok_popularity_count + track['popularity'] wayne_ok_avg_pop = wayne_ok_popularity_count / wayne_ok_count wayne_ok_len = wayne_ok_len + track["duration_ms"] if wayne_explicit_count > 0: print("The average popularity for Lil Wayne's explicit songs is", wayne_avg_pop) #1 minute is 60000 milliseconds, who knew? wayne_explicit_mins = int(wayne_explicit_len) / 60000 print("Lil Wayne has", wayne_explicit_mins, "minutes of explicit songs") if wayne_ok_count > 0: print("The average popularity for Lil Wayne's non-explicit songs is", wayne_ok_avg_pop) wayne_ok_mins = int(wayne_ok_len) / 60000 print("Lil Wayne has", wayne_ok_mins, "minutes of explicit songs") Explanation: Will the world explode if a musician swears? Get an average popularity for their explicit songs vs. their non-explicit songs. How many minutes of explicit songs do they have? Non-explicit? End of explanation mama_exp_count = 0 mama_exp_pop_count = 0 mama_ok_count = 0 mama_ok_pop_count = 0 mama_exp_len = 0 mama_ok_len = 0 for track in top_mama_tracks: print(track['name'], track['explicit'], track['popularity'], track["duration_ms"]) if True: mama_exp_count = mama_exp_count + 1 mama_exp_pop_count = mama_exp_pop_count + int(track['popularity']) mama_avg_pop = int(mama_exp_pop_count) / int(mama_exp_count) mama_exp_len = mama_exp_len + int(track["duration_ms"]) if not track['explicit']: mama_ok_count = mama_ok_count + 1 mama_ok_pop_count = mama_ok_pop_count + int(track['popularity']) mama_ok_avg_pop = int(mama_ok_pop_count) / int(mama_ok_count) mama_ok_len = mama_ok_len + int(track["duration_ms"]) if mama_exp_count > 0: #1 minute is 60000 milliseconds, who knew? print("The average popularity for Lil Mama's xplicit songs is", mama_avg_pop) mama_exp_mins = int(mama_exp_len) / 60000 print("Lil Mama has", mama_exp_mins, "minutes of explicit songs") if mama_ok_count > 0: print("The average popularity for Lil Mama's non-explicit songs is", mama_ok_avg_pop) mama_ok_mins = int(mama_ok_len) / 60000 print("Lil Mama has", mama_ok_mins, "minutes of non-explicit songs") Explanation: QUESTION: Why does this return both true and not true statements for non-explicit statements? for track in top_mama_tracks: print(track['name'], track['explicit']) if True: print(track['name'], "is explicit and has a popularity of", track['popularity']) if not track['explicit']: print(track['name'], "is not explicit and has a popularity of", track['popularity']) End of explanation #We need to bypass the limit. And find out response = requests.get('https://api.spotify.com/v1/search?query=biggie&type=artist') biggie_data = response.json() biggie_artists = biggie_data['artists'] biggie_names = biggie_artists['items'] biggie_count= 0 for name in biggie_names: print(name['name']) biggie_count = biggie_count + 1 print("There are a total number of", biggie_count, "biggie artists") response = requests.get('https://api.spotify.com/v1/search?query=lil&type=artist') lil_data = response.json() lil_x_artists = lil_data['artists'] lil_x_names = lil_x_artists['items'] lil_x_count= 0 for name in lil_x_names: print(name['name']) lil_x_count = biggie_count + 1 print("There are a total number of", lil_x_count, "lil artists") Explanation: 7) Since we're talking about Lils, what about Biggies? How many total "Biggie" artists are there? How many total "Lil"s? If you made 1 request every 5 seconds, how long would it take to download information on all the Lils vs the Biggies? End of explanation response = requests.get('https://api.spotify.com/v1/search?query=biggie&type=artist&limit=50') b_data = response.json() b_artists = b_data['artists'] b_names = b_artists['items'] b_pop_count = 0 b_number = 0 for names in b_names: print(names['name'], names['popularity']) b_number = b_number + 1 b_pop_count = b_pop_count + int(names['popularity']) avg_b_pop = b_pop_count / int(b_number) print("The Biggies' average popularity is", avg_b_pop) lil_pop_count = 0 lil_number = 0 for names in lil_names: print(names['name'], names['popularity']) lil_number = lil_number + 1 lil_pop_count = lil_pop_count + int(names['popularity']) avg_lil_pop = lil_pop_count / int(lil_number) print("The Lils average popularity is", avg_lil_pop) print("The Lils are far more popular") Explanation: 8) Out of the top 50 "Lil"s and the top 50 "Biggie"s, who is more popular on average? End of explanation <END_TASK>
15,965
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Convolutional Neural Networks Step1: Run the next cell to load the "SIGNS" dataset you are going to use. Step2: As a reminder, the SIGNS dataset is a collection of 6 signs representing numbers from 0 to 5. <img src="images/SIGNS.png" style="width Step3: In Course 2, you had built a fully-connected network for this dataset. But since this is an image dataset, it is more natural to apply a ConvNet to it. To get started, let's examine the shapes of your data. Step5: 1.1 - Create placeholders TensorFlow requires that you create placeholders for the input data that will be fed into the model when running the session. Exercise Step7: Expected Output <table> <tr> <td> X = Tensor("Placeholder Step9: Expected Output Step11: Expected Output Step13: Expected Output Step14: Run the following cell to train your model for 100 epochs. Check if your cost after epoch 0 and 5 matches our output. If not, stop the cell and go back to your code! Step15: Expected output
<ASSISTANT_TASK:> Python Code: import math import numpy as np import h5py import matplotlib.pyplot as plt import scipy from PIL import Image from scipy import ndimage import tensorflow as tf from tensorflow.python.framework import ops from cnn_utils import * %matplotlib inline np.random.seed(1) Explanation: Convolutional Neural Networks: Application Welcome to Course 4's second assignment! In this notebook, you will: Implement helper functions that you will use when implementing a TensorFlow model Implement a fully functioning ConvNet using TensorFlow After this assignment you will be able to: Build and train a ConvNet in TensorFlow for a classification problem We assume here that you are already familiar with TensorFlow. If you are not, please refer the TensorFlow Tutorial of the third week of Course 2 ("Improving deep neural networks"). 1.0 - TensorFlow model In the previous assignment, you built helper functions using numpy to understand the mechanics behind convolutional neural networks. Most practical applications of deep learning today are built using programming frameworks, which have many built-in functions you can simply call. As usual, we will start by loading in the packages. End of explanation # Loading the data (signs) X_train_orig, Y_train_orig, X_test_orig, Y_test_orig, classes = load_dataset() Explanation: Run the next cell to load the "SIGNS" dataset you are going to use. End of explanation # Example of a picture index = 6 plt.imshow(X_train_orig[index]) print ("y = " + str(np.squeeze(Y_train_orig[:, index]))) Explanation: As a reminder, the SIGNS dataset is a collection of 6 signs representing numbers from 0 to 5. <img src="images/SIGNS.png" style="width:800px;height:300px;"> The next cell will show you an example of a labelled image in the dataset. Feel free to change the value of index below and re-run to see different examples. End of explanation X_train = X_train_orig/255. X_test = X_test_orig/255. Y_train = convert_to_one_hot(Y_train_orig, 6).T Y_test = convert_to_one_hot(Y_test_orig, 6).T print ("number of training examples = " + str(X_train.shape[0])) print ("number of test examples = " + str(X_test.shape[0])) print ("X_train shape: " + str(X_train.shape)) print ("Y_train shape: " + str(Y_train.shape)) print ("X_test shape: " + str(X_test.shape)) print ("Y_test shape: " + str(Y_test.shape)) conv_layers = {} Explanation: In Course 2, you had built a fully-connected network for this dataset. But since this is an image dataset, it is more natural to apply a ConvNet to it. To get started, let's examine the shapes of your data. End of explanation # GRADED FUNCTION: create_placeholders def create_placeholders(n_H0, n_W0, n_C0, n_y): Creates the placeholders for the tensorflow session. Arguments: n_H0 -- scalar, height of an input image n_W0 -- scalar, width of an input image n_C0 -- scalar, number of channels of the input n_y -- scalar, number of classes Returns: X -- placeholder for the data input, of shape [None, n_H0, n_W0, n_C0] and dtype "float" Y -- placeholder for the input labels, of shape [None, n_y] and dtype "float" ### START CODE HERE ### (≈2 lines) X = tf.placeholder(shape=[None, n_H0, n_W0, n_C0], dtype=np.float32) Y = tf.placeholder(shape=[None, n_y], dtype=np.float32) ### END CODE HERE ### return X, Y X, Y = create_placeholders(64, 64, 3, 6) print ("X = " + str(X)) print ("Y = " + str(Y)) Explanation: 1.1 - Create placeholders TensorFlow requires that you create placeholders for the input data that will be fed into the model when running the session. Exercise: Implement the function below to create placeholders for the input image X and the output Y. You should not define the number of training examples for the moment. To do so, you could use "None" as the batch size, it will give you the flexibility to choose it later. Hence X should be of dimension [None, n_H0, n_W0, n_C0] and Y should be of dimension [None, n_y]. Hint. End of explanation # GRADED FUNCTION: initialize_parameters def initialize_parameters(): Initializes weight parameters to build a neural network with tensorflow. The shapes are: W1 : [4, 4, 3, 8] W2 : [2, 2, 8, 16] Returns: parameters -- a dictionary of tensors containing W1, W2 tf.set_random_seed(1) # so that your "random" numbers match ours ### START CODE HERE ### (approx. 2 lines of code) W1 = tf.get_variable('W1', [4, 4, 3, 8], initializer=tf.contrib.layers.xavier_initializer(seed = 0)) W2 = tf.get_variable('W2', [2, 2, 8, 16], initializer=tf.contrib.layers.xavier_initializer(seed = 0)) ### END CODE HERE ### parameters = {"W1": W1, "W2": W2} return parameters tf.reset_default_graph() with tf.Session() as sess_test: parameters = initialize_parameters() init = tf.global_variables_initializer() sess_test.run(init) print("W1 = " + str(parameters["W1"].eval()[1,1,1])) print("W2 = " + str(parameters["W2"].eval()[1,1,1])) Explanation: Expected Output <table> <tr> <td> X = Tensor("Placeholder:0", shape=(?, 64, 64, 3), dtype=float32) </td> </tr> <tr> <td> Y = Tensor("Placeholder_1:0", shape=(?, 6), dtype=float32) </td> </tr> </table> 1.2 - Initialize parameters You will initialize weights/filters $W1$ and $W2$ using tf.contrib.layers.xavier_initializer(seed = 0). You don't need to worry about bias variables as you will soon see that TensorFlow functions take care of the bias. Note also that you will only initialize the weights/filters for the conv2d functions. TensorFlow initializes the layers for the fully connected part automatically. We will talk more about that later in this assignment. Exercise: Implement initialize_parameters(). The dimensions for each group of filters are provided below. Reminder - to initialize a parameter $W$ of shape [1,2,3,4] in Tensorflow, use: python W = tf.get_variable("W", [1,2,3,4], initializer = ...) More Info. End of explanation # GRADED FUNCTION: forward_propagation def forward_propagation(X, parameters): Implements the forward propagation for the model: CONV2D -> RELU -> MAXPOOL -> CONV2D -> RELU -> MAXPOOL -> FLATTEN -> FULLYCONNECTED Arguments: X -- input dataset placeholder, of shape (input size, number of examples) parameters -- python dictionary containing your parameters "W1", "W2" the shapes are given in initialize_parameters Returns: Z3 -- the output of the last LINEAR unit # Retrieve the parameters from the dictionary "parameters" W1 = parameters['W1'] W2 = parameters['W2'] ### START CODE HERE ### # CONV2D: stride of 1, padding 'SAME' Z1 = tf.nn.conv2d(X, W1, strides=[1, 1, 1, 1], padding='SAME') # RELU A1 = tf.nn.relu(Z1) # MAXPOOL: window 8x8, sride 8, padding 'SAME' P1 = tf.nn.max_pool(A1, ksize=[1, 8, 8, 1], strides=[1, 8, 8, 1], padding='SAME') # CONV2D: filters W2, stride 1, padding 'SAME' Z2 = tf.nn.conv2d(P1, W2, strides=[1, 1, 1, 1], padding='SAME') # RELU A2 = tf.nn.relu(Z2) # MAXPOOL: window 4x4, stride 4, padding 'SAME' P2 = tf.nn.max_pool(A2, ksize=[1, 4, 4, 1], strides=[1, 4, 4, 1], padding='SAME') # FLATTEN P2 = tf.contrib.layers.flatten(P2) # FULLY-CONNECTED without non-linear activation function (not not call softmax). # 6 neurons in output layer. Hint: one of the arguments should be "activation_fn=None" Z3 = tf.contrib.layers.fully_connected(P2, 6, activation_fn=None) ### END CODE HERE ### return Z3 tf.reset_default_graph() with tf.Session() as sess: np.random.seed(1) X, Y = create_placeholders(64, 64, 3, 6) parameters = initialize_parameters() Z3 = forward_propagation(X, parameters) init = tf.global_variables_initializer() sess.run(init) a = sess.run(Z3, {X: np.random.randn(2,64,64,3), Y: np.random.randn(2,6)}) print("Z3 = " + str(a)) Explanation: Expected Output: <table> <tr> <td> W1 = </td> <td> [ 0.00131723 0.14176141 -0.04434952 0.09197326 0.14984085 -0.03514394 <br> -0.06847463 0.05245192] </td> </tr> <tr> <td> W2 = </td> <td> [-0.08566415 0.17750949 0.11974221 0.16773748 -0.0830943 -0.08058 <br> -0.00577033 -0.14643836 0.24162132 -0.05857408 -0.19055021 0.1345228 <br> -0.22779644 -0.1601823 -0.16117483 -0.10286498] </td> </tr> </table> 1.2 - Forward propagation In TensorFlow, there are built-in functions that carry out the convolution steps for you. tf.nn.conv2d(X,W1, strides = [1,s,s,1], padding = 'SAME'): given an input $X$ and a group of filters $W1$, this function convolves $W1$'s filters on X. The third input ([1,f,f,1]) represents the strides for each dimension of the input (m, n_H_prev, n_W_prev, n_C_prev). You can read the full documentation here tf.nn.max_pool(A, ksize = [1,f,f,1], strides = [1,s,s,1], padding = 'SAME'): given an input A, this function uses a window of size (f, f) and strides of size (s, s) to carry out max pooling over each window. You can read the full documentation here tf.nn.relu(Z1): computes the elementwise ReLU of Z1 (which can be any shape). You can read the full documentation here. tf.contrib.layers.flatten(P): given an input P, this function flattens each example into a 1D vector it while maintaining the batch-size. It returns a flattened tensor with shape [batch_size, k]. You can read the full documentation here. tf.contrib.layers.fully_connected(F, num_outputs): given a the flattened input F, it returns the output computed using a fully connected layer. You can read the full documentation here. In the last function above (tf.contrib.layers.fully_connected), the fully connected layer automatically initializes weights in the graph and keeps on training them as you train the model. Hence, you did not need to initialize those weights when initializing the parameters. Exercise: Implement the forward_propagation function below to build the following model: CONV2D -&gt; RELU -&gt; MAXPOOL -&gt; CONV2D -&gt; RELU -&gt; MAXPOOL -&gt; FLATTEN -&gt; FULLYCONNECTED. You should use the functions above. In detail, we will use the following parameters for all the steps: - Conv2D: stride 1, padding is "SAME" - ReLU - Max pool: Use an 8 by 8 filter size and an 8 by 8 stride, padding is "SAME" - Conv2D: stride 1, padding is "SAME" - ReLU - Max pool: Use a 4 by 4 filter size and a 4 by 4 stride, padding is "SAME" - Flatten the previous output. - FULLYCONNECTED (FC) layer: Apply a fully connected layer without an non-linear activation function. Do not call the softmax here. This will result in 6 neurons in the output layer, which then get passed later to a softmax. In TensorFlow, the softmax and cost function are lumped together into a single function, which you'll call in a different function when computing the cost. End of explanation # GRADED FUNCTION: compute_cost def compute_cost(Z3, Y): Computes the cost Arguments: Z3 -- output of forward propagation (output of the last LINEAR unit), of shape (6, number of examples) Y -- "true" labels vector placeholder, same shape as Z3 Returns: cost - Tensor of the cost function ### START CODE HERE ### (1 line of code) cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=Z3, labels=Y)) ### END CODE HERE ### return cost tf.reset_default_graph() with tf.Session() as sess: np.random.seed(1) X, Y = create_placeholders(64, 64, 3, 6) parameters = initialize_parameters() Z3 = forward_propagation(X, parameters) cost = compute_cost(Z3, Y) init = tf.global_variables_initializer() sess.run(init) a = sess.run(cost, {X: np.random.randn(4,64,64,3), Y: np.random.randn(4,6)}) print("cost = " + str(a)) Explanation: Expected Output: <table> <td> Z3 = </td> <td> [[-0.44670227 -1.57208765 -1.53049231 -2.31013036 -1.29104376 0.46852064] <br> [-0.17601591 -1.57972014 -1.4737016 -2.61672091 -1.00810647 0.5747785 ]] </td> </table> 1.3 - Compute cost Implement the compute cost function below. You might find these two functions helpful: tf.nn.softmax_cross_entropy_with_logits(logits = Z3, labels = Y): computes the softmax entropy loss. This function both computes the softmax activation function as well as the resulting loss. You can check the full documentation here. tf.reduce_mean: computes the mean of elements across dimensions of a tensor. Use this to sum the losses over all the examples to get the overall cost. You can check the full documentation here. Exercise: Compute the cost below using the function above. End of explanation # GRADED FUNCTION: model def model(X_train, Y_train, X_test, Y_test, learning_rate = 0.009, num_epochs = 100, minibatch_size = 64, print_cost = True): Implements a three-layer ConvNet in Tensorflow: CONV2D -> RELU -> MAXPOOL -> CONV2D -> RELU -> MAXPOOL -> FLATTEN -> FULLYCONNECTED Arguments: X_train -- training set, of shape (None, 64, 64, 3) Y_train -- test set, of shape (None, n_y = 6) X_test -- training set, of shape (None, 64, 64, 3) Y_test -- test set, of shape (None, n_y = 6) learning_rate -- learning rate of the optimization num_epochs -- number of epochs of the optimization loop minibatch_size -- size of a minibatch print_cost -- True to print the cost every 100 epochs Returns: train_accuracy -- real number, accuracy on the train set (X_train) test_accuracy -- real number, testing accuracy on the test set (X_test) parameters -- parameters learnt by the model. They can then be used to predict. ops.reset_default_graph() # to be able to rerun the model without overwriting tf variables tf.set_random_seed(1) # to keep results consistent (tensorflow seed) seed = 3 # to keep results consistent (numpy seed) (m, n_H0, n_W0, n_C0) = X_train.shape n_y = Y_train.shape[1] costs = [] # To keep track of the cost # Create Placeholders of the correct shape ### START CODE HERE ### (1 line) X, Y = create_placeholders(n_H0, n_W0, n_C0, n_y) ### END CODE HERE ### # Initialize parameters ### START CODE HERE ### (1 line) parameters = initialize_parameters() ### END CODE HERE ### # Forward propagation: Build the forward propagation in the tensorflow graph ### START CODE HERE ### (1 line) Z3 = forward_propagation(X, parameters) ### END CODE HERE ### # Cost function: Add cost function to tensorflow graph ### START CODE HERE ### (1 line) cost = compute_cost(Z3, Y) ### END CODE HERE ### # Backpropagation: Define the tensorflow optimizer. Use an AdamOptimizer that minimizes the cost. ### START CODE HERE ### (1 line) optimizer = tf.train.AdamOptimizer(learning_rate).minimize(cost) ### END CODE HERE ### # Initialize all the variables globally init = tf.global_variables_initializer() # Start the session to compute the tensorflow graph with tf.Session() as sess: # Run the initialization sess.run(init) # Do the training loop for epoch in range(num_epochs): minibatch_cost = 0. num_minibatches = int(m / minibatch_size) # number of minibatches of size minibatch_size in the train set seed = seed + 1 minibatches = random_mini_batches(X_train, Y_train, minibatch_size, seed) for minibatch in minibatches: # Select a minibatch (minibatch_X, minibatch_Y) = minibatch # IMPORTANT: The line that runs the graph on a minibatch. # Run the session to execute the optimizer and the cost, the feedict should contain a minibatch for (X,Y). ### START CODE HERE ### (1 line) _ , temp_cost = sess.run([optimizer, cost], feed_dict={X: minibatch_X, Y: minibatch_Y}) ### END CODE HERE ### minibatch_cost += temp_cost / num_minibatches # Print the cost every epoch if print_cost == True and epoch % 5 == 0: print ("Cost after epoch %i: %f" % (epoch, minibatch_cost)) if print_cost == True and epoch % 1 == 0: costs.append(minibatch_cost) # plot the cost plt.plot(np.squeeze(costs)) plt.ylabel('cost') plt.xlabel('iterations (per tens)') plt.title("Learning rate =" + str(learning_rate)) plt.show() # Calculate the correct predictions predict_op = tf.argmax(Z3, 1) correct_prediction = tf.equal(predict_op, tf.argmax(Y, 1)) # Calculate accuracy on the test set accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float")) print(accuracy) train_accuracy = accuracy.eval({X: X_train, Y: Y_train}) test_accuracy = accuracy.eval({X: X_test, Y: Y_test}) print("Train Accuracy:", train_accuracy) print("Test Accuracy:", test_accuracy) return train_accuracy, test_accuracy, parameters Explanation: Expected Output: <table> <td> cost = </td> <td> 2.91034 </td> </table> 1.4 Model Finally you will merge the helper functions you implemented above to build a model. You will train it on the SIGNS dataset. You have implemented random_mini_batches() in the Optimization programming assignment of course 2. Remember that this function returns a list of mini-batches. Exercise: Complete the function below. The model below should: create placeholders initialize parameters forward propagate compute the cost create an optimizer Finally you will create a session and run a for loop for num_epochs, get the mini-batches, and then for each mini-batch you will optimize the function. Hint for initializing the variables End of explanation _, _, parameters = model(X_train, Y_train, X_test, Y_test) Explanation: Run the following cell to train your model for 100 epochs. Check if your cost after epoch 0 and 5 matches our output. If not, stop the cell and go back to your code! End of explanation fname = "images/thumbs_up.jpg" image = np.array(ndimage.imread(fname, flatten=False)) my_image = scipy.misc.imresize(image, size=(64,64)) plt.imshow(my_image) Explanation: Expected output: although it may not match perfectly, your expected output should be close to ours and your cost value should decrease. <table> <tr> <td> **Cost after epoch 0 =** </td> <td> 1.917929 </td> </tr> <tr> <td> **Cost after epoch 5 =** </td> <td> 1.506757 </td> </tr> <tr> <td> **Train Accuracy =** </td> <td> 0.940741 </td> </tr> <tr> <td> **Test Accuracy =** </td> <td> 0.783333 </td> </tr> </table> Congratulations! You have finised the assignment and built a model that recognizes SIGN language with almost 80% accuracy on the test set. If you wish, feel free to play around with this dataset further. You can actually improve its accuracy by spending more time tuning the hyperparameters, or using regularization (as this model clearly has a high variance). Once again, here's a thumbs up for your work! End of explanation <END_TASK>
15,966
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Softmax exercise Adapt from the Stanford CS231n assignment1, find the original version on the course website. In this exercise we will Step5: Load CIFAR-10 data Load the data and split into training / validation / testing datasets. See notebook 00 for more details. Step7: Softmax Classifier Step8: Sanity Check In a randomly initiated weight, the loss should be close to $-\log(\frac{1}{10}) = -\log(0.1) \approx 2.303$ Step10: Vectorized loss function Step12: Stochastic Gradient Descent We now have vectorized and efficient expressions for the loss, the gradient and our gradient matches the numerical gradient. We are therefore ready to do SGD to minimize the loss.
<ASSISTANT_TASK:> Python Code: import random import time import numpy as np import matplotlib.pyplot as plt from cs231n.data_utils import load_CIFAR10 from cs231n.gradient_check import grad_check_sparse # plotting setting %matplotlib inline plt.rcParams['figure.figsize'] = (10.0, 8.0) # set default size of plots plt.rcParams['image.interpolation'] = 'nearest' plt.rcParams['image.cmap'] = 'gray' Explanation: Softmax exercise Adapt from the Stanford CS231n assignment1, find the original version on the course website. In this exercise we will: implement a fully-vectorized loss function for the SVM implement the fully-vectorized expression for its analytic gradient check your implementation using numerical gradient use a validation set to tune the learning rate and regularization strength optimize the loss function with SGD visualize the final learned weights End of explanation def get_CIFAR10_data(num_training=49000, num_validation=1000, num_test=1000): Load the CIFAR-10 dataset from disk and perform preprocessing to prepare it for the linear classifier. These are the same steps as we used for the SVM, but condensed to a single function. # Load the raw CIFAR-10 data cifar10_dir = 'cs231n/datasets/cifar-10-batches-py' X_train, y_train, X_test, y_test = load_CIFAR10(cifar10_dir) # subsample the data mask = range(num_training, num_training + num_validation) X_val = X_train[mask] y_val = y_train[mask] mask = range(num_training) X_train = X_train[mask] y_train = y_train[mask] mask = range(num_test) X_test = X_test[mask] y_test = y_test[mask] # Preprocessing: reshape the image data into rows X_train = np.reshape(X_train, (X_train.shape[0], -1)) X_val = np.reshape(X_val, (X_val.shape[0], -1)) X_test = np.reshape(X_test, (X_test.shape[0], -1)) # Normalize the data: subtract the mean image mean_image = np.mean(X_train, axis = 0) X_train -= mean_image X_val -= mean_image X_test -= mean_image # add bias dimension and transform into columns X_train = np.hstack([X_train, np.ones((X_train.shape[0], 1))]).T X_val = np.hstack([X_val, np.ones((X_val.shape[0], 1))]).T X_test = np.hstack([X_test, np.ones((X_test.shape[0], 1))]).T return X_train, y_train, X_val, y_val, X_test, y_test # Invoke the above function to get our data. X_train, y_train, X_val, y_val, X_test, y_test = get_CIFAR10_data() print('Train data shape:', X_train.shape) print('Train labels shape:', y_train.shape) print('Validation data shape:', X_val.shape) print('Validation labels shape:', y_val.shape) print('Test data shape:', X_test.shape) print('Test labels shape:', y_test.shape) class LinearClassifier: def __init__(self): self.W = None def train( self, X, y, learning_rate=1e-3, reg=1e-5, num_iters=100, batch_size=200, verbose=False, seed=None ): Train this linear classifier using stochastic gradient descent. Inputs: - X: D x N array of training data. Each training point is a D-dimensional column. - y: 1-dimensional array of length N with labels 0...K-1 for K classes. - learning_rate: (float) learning rate for optimization. - reg: (float) regularization strength. - num_iters: (integer) number of steps to take when optimizing. - batch_size: (integer) number of training examples to use at each step. - verbose: (boolean) If true, print progress during optimization. Outputs: A list containing the value of the loss function at each training iteration. dim, num_train = X.shape # assume y takes values 0...K-1 where K is number of classes num_classes = np.max(y) + 1 if self.W is None: # lazily initialize W self.W = np.random.randn(num_classes, dim) * 0.001 batch_rs = np.random.RandomState(seed) # Run stochastic gradient descent to optimize W loss_history = [] for it in range(num_iters): batch_ix = batch_rs.choice( np.arange(num_train), size=batch_size, replace=True ) X_batch = X[:, batch_ix] y_batch = y[batch_ix] # evaluate loss and gradient, internally use self.W loss, grad = self.loss(X_batch, y_batch, reg) loss_history.append(loss) # perform parameter update self.W -= grad * learning_rate if verbose and it % 100 == 0: print('iteration %d / %d: loss %f' % (it, num_iters, loss)) return loss_history def predict(self, X): Use the trained weights of this linear classifier to predict labels for data points. Inputs: - X: D x N array of training data. Each column is a D-dimensional point. Returns: - y_pred: Predicted labels for the data in X. y_pred is a 1-dimensional array of length N, and each element is an integer giving the predicted class. score_pred = self.W.dot(X) # shape: C x N y_pred = score_pred.argmax(axis=0) return y_pred def loss(self, X_batch, y_batch, reg): Compute the loss function and its derivative. Subclasses will override this. Inputs: - X_batch: D x N array of data; each column is a data point. - y_batch: 1-dimensional array of length N with labels 0...K-1, for K classes. - reg: (float) regularization strength. Returns: A tuple containing: - loss as a single float - gradient with respect to self.W; an array of the same shape as W raise NotImplementedError Explanation: Load CIFAR-10 data Load the data and split into training / validation / testing datasets. See notebook 00 for more details. End of explanation def softmax_loss_naive(W, X, y, reg): Softmax loss function, naive implementation (with loops) Inputs: - W: C x D array of weights - X: D x N array of data. Data are D-dimensional columns - y: 1-dimensional array of length N with labels 0...K-1, for K classes - reg: (float) regularization strength Returns: a tuple of: - loss as single float - gradient with respect to weights W, an array of same size as W # Initialize the loss and gradient to zero. loss = 0.0 dW = np.zeros_like(W) num_train = X.shape[1] f = W.dot(X) # shape: C x N p = np.zeros(num_train, dtype=np.float) for i in range(num_train): f_i = f[:, i].copy() # shape C x 1 f_i -= np.max(f_i) # improve numerical stability f_i = np.exp(f_i) x_i = X[:, i] all_class_p_i = f_i / np.sum(f_i) p[i] = all_class_p_i[y[i]] # Update gradient # all_class_p_i no used later, don't copy dw_x_weight_i = all_class_p_i dw_x_weight_i[y[i]] -= 1 dW -= dw_x_weight_i[:, np.newaxis] * x_i[np.newaxis, :] loss += np.mean(-np.log(p)) # Add regularization loss += 0.5 * reg * np.sum(W * W) # Gradient # ref: http://ufldl.stanford.edu/wiki/index.php/Softmax_Regression dW /= -num_train dW += reg * W return loss, dW Explanation: Softmax Classifier End of explanation %%time # Generate a random softmax weight matrix and use it to compute the loss. W = np.random.randn(10, 3073) * 0.0001 loss, grad = softmax_loss_naive(W, X_train, y_train, 0.0) # As a rough sanity check, our loss should be something close to -log(0.1). print('loss: %f' % loss) print('sanity check: %f' % (-np.log(0.1))) %%time # Complete the implementation of softmax_loss_naive and implement a (naive) # version of the gradient that uses nested loops. loss, grad = softmax_loss_naive(W, X_train, y_train, 0.0) # As we did for the SVM, use numeric gradient checking as a debugging tool. # The numeric gradient should be close to the analytic gradient. f = lambda w: softmax_loss_naive(w, X_train, y_train, 0.0)[0] grad_numerical = grad_check_sparse(f, W, grad, 5) # increase 5 here to check for more times Explanation: Sanity Check In a randomly initiated weight, the loss should be close to $-\log(\frac{1}{10}) = -\log(0.1) \approx 2.303$ End of explanation def softmax_loss_vectorized(W, X, y, reg): Softmax loss function, vectorized version. Inputs and outputs are the same as softmax_loss_naive. # Initialize the loss and gradient to zero. loss = 0.0 dW = np.zeros_like(W) num_train = X.shape[1] _train_ix = np.arange(num_train) # for sample coord 0...N-1 f = W.dot(X) # shape: C x N f -= np.max(f, axis=0) # improve numerical stability f = np.exp(f) p = f / np.sum(f, axis=0) # shape: C x N # loss function loss += np.mean(-np.log(p[y, _train_ix])) loss += 0.5 * reg * np.sum(W * W) # gradient # ref: http://ufldl.stanford.edu/wiki/index.php/Softmax_Regression dW_x_weight = p # no use p later, don't copy dW_x_weight[y, _train_ix] -= 1 # CxD -= CxN dot NxD dW -= dW_x_weight.dot(X.T) dW /= -num_train dW += reg * W return loss, dW # Now that we have a naive implementation of the softmax loss function and its gradient, # implement a vectorized version in softmax_loss_vectorized. # The two versions should compute the same results, but the vectorized version should be # much faster. tic = time.time() loss_naive, grad_naive = softmax_loss_naive(W, X_train, y_train, 0.00001) toc = time.time() print('naive loss: %e computed in %fs' % (loss_naive, toc - tic)) tic = time.time() loss_vectorized, grad_vectorized = softmax_loss_vectorized(W, X_train, y_train, 0.00001) toc = time.time() print('vectorized loss: %e computed in %fs' % (loss_vectorized, toc - tic)) # The losses should match but your vectorized implementation should be much faster. # The loss is a single number, so it is easy to compare the values computed # by the two implementations. The gradient on the other hand is a matrix, so # we use the Frobenius norm to compare them. grad_difference = np.linalg.norm(grad_naive - grad_vectorized, ord='fro') print('Loss difference: %f' % np.abs(loss_naive - loss_vectorized)) print('Gradient difference: %f' % grad_difference) Explanation: Vectorized loss function End of explanation class Softmax(LinearClassifier): A subclass that uses the Softmax + Cross-entropy loss function def loss(self, X_batch, y_batch, reg): return softmax_loss_vectorized(self.W, X_batch, y_batch, reg) # Now implement SGD in LinearSVM.train() function and run it with the code below softmax = Softmax() tic = time.time() loss_hist = softmax.train( X_train, y_train, learning_rate = 5e-8, reg=1e3, # better params: # learning_rate=5e-7, reg=5e4, num_iters=1500, seed=9527, verbose=True ) toc = time.time() print('That took %fs' % (toc - tic)) # A useful debugging strategy is to plot the loss as a function of # iteration number: plt.plot(loss_hist) plt.xlabel('Iteration number') plt.ylabel('Loss value') plt.show() y_train_pred = softmax.predict(X_train) print('training accuracy: {:.3%}'.format(np.mean(y_train == y_train_pred))) y_val_pred = softmax.predict(X_val) print('validation accuracy: {:.3%}'.format(np.mean(y_val == y_val_pred))) # evaluate on test set # Evaluate the best svm on test set y_test_pred = softmax.predict(X_test) test_accuracy = np.mean(y_test == y_test_pred) print('softmax on raw pixels final test set accuracy: %f' % (test_accuracy, )) # Visualize the learned weights for each class w = softmax.W[:,:-1] # strip out the bias w = w.reshape(10, 32, 32, 3) w_min, w_max = np.min(w), np.max(w) classes = ['plane', 'car', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck'] for i in range(10): plt.subplot(2, 5, i + 1) # Rescale the weights to be between 0 and 255 wimg = 255.0 * (w[i].squeeze() - w_min) / (w_max - w_min) plt.imshow(wimg.astype('uint8')) plt.axis('off') plt.title(classes[i]) plt.show() Explanation: Stochastic Gradient Descent We now have vectorized and efficient expressions for the loss, the gradient and our gradient matches the numerical gradient. We are therefore ready to do SGD to minimize the loss. End of explanation <END_TASK>
15,967
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Opinion Game - One Topic in the network These codes were developed as a part of the our opinion game model. Import Python built-in functions we need to run and plot the game Step1: Set up inline matplotlib This tells Python to plot the figure right in here, in the notebook. Step2: Import Game Modules From a Given Path User have to edit the path and put the correct one on his/her machine. Step3: Setting Up Game Parameters Step4: seed PRNG Step5: Set up the state of the system State of the system includes Step6: User Defined States Can go in the following cell Step7: Run The Game Step8: Reshape the history Since here we have just one topic, the matrix can be converted to a 2D array, rather than 3. This makes the plotting job easier Step9: Convert to Panda DataFrame and then plot. Step10: Plot the numpy array
<ASSISTANT_TASK:> Python Code: import numpy as np from numpy.random import randn import pandas as pd from pandas import Series, DataFrame import matplotlib.pyplot as plt import matplotlib.animation as animation import matplotlib.image as mpimg from matplotlib import rcParams import seaborn as sb Explanation: Opinion Game - One Topic in the network These codes were developed as a part of the our opinion game model. Import Python built-in functions we need to run and plot the game End of explanation %matplotlib inline #rcParams['figure.figsize'] = 5, 4 sb.set_style('whitegrid') Explanation: Set up inline matplotlib This tells Python to plot the figure right in here, in the notebook. End of explanation import sys # search path for modules sys.path.append('/Users/hn/Documents/GitHub/PyOpinionGame/') import opiniongame.config as og_cfg import opiniongame.IO as og_io import opiniongame.coupling as og_coupling import opiniongame.state as og_state import opiniongame.adjacency as og_adj import opiniongame.selection as og_select import opiniongame.potentials as og_pot import opiniongame.core as og_core import opiniongame.stopping as og_stop import opiniongame.opinions as og_opinions Explanation: Import Game Modules From a Given Path User have to edit the path and put the correct one on his/her machine. End of explanation config = og_cfg.staticParameters() path = '/Users/hn/Documents/GitHub/PyOpinionGame/' # path to the 'staticParameters.cfg' staticParameters = path + 'staticParameters.cfg' config.readFromFile(staticParameters) config.threshold = 0.0001 config.Kthreshold = 0.00001 config.startingseed = 20 config.learning_rate = 0.1 tau = 0.62 #tip of the tent potential function config.printOut() Explanation: Setting Up Game Parameters End of explanation print("SEEDING PRNG: "+str(config.startingseed)) np.random.seed(config.startingseed) Explanation: seed PRNG: must do this before any random numbers are ever sampled during default generation End of explanation # These are the default matrices for the state of the system: # If you want to change them, you can generate a new one in the following cell default_weights = og_coupling.weights_no_coupling(config.popSize, config.ntopics) default_initialOpinions = og_opinions.initialize_opinions(config.popSize, config.ntopics) default_adj = og_adj.make_adj(config.popSize, 'full') Explanation: Set up the state of the system State of the system includes: Weight Matrix (Matrix of the coupling wieghts between topic) Initial Opinions of agents Adjacency matrix of the network End of explanation # user_weights = # user_initialOpinions = # user_adj = state = og_state.WorldState(adj=default_adj, couplingWeights=default_weights, initialOpinions=default_initialOpinions, initialHistorySize=100, historyGrowthScale=2) state.validate() ufuncs = og_cfg.UserFunctions(og_select.FastPairSelection, og_stop.iterationStop, og_pot.createTent(tau)) Explanation: User Defined States Can go in the following cell: End of explanation # # run the game # state = og_core.run_until_convergence(config, state, ufuncs) # Extract and save the history of the game rdict = {} rdict['history'] = state.history og_io.saveMatrix('output.mat', rdict) Explanation: Run The Game End of explanation time, population_size, no_of_topics = rdict['history'].shape evolution = rdict['history'].reshape(time, population_size) Explanation: Reshape the history Since here we have just one topic, the matrix can be converted to a 2D array, rather than 3. This makes the plotting job easier: End of explanation df = pd.DataFrame(data=evolution, index=None, columns=None) fig = plt.figure() ax = df.plot(legend=False) ax.set_xlabel("Time") ax.set_ylabel("Opinions") ax.set_title('Evolution of Opinions') plt.show() Explanation: Convert to Panda DataFrame and then plot. End of explanation fig1 = plt.figure() plt.plot(evolution) plt.xlabel('Time') plt.ylabel('Opinionds') plt.title('Evolution of Opinions') plt.show() Explanation: Plot the numpy array End of explanation <END_TASK>
15,968
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: LogGabor user guide Table of content What is the LogGabor package? Installing Importing the library Properties of log-Gabor filters Testing filter generation Testing on a sample image Building a pyramid An example of fitting images with log-Gabor filters Importing the library Step1: To install the dependencies related to running this notebook, see Installing notebook dependencies. Back to top Step2: Perspectives Step3: Back to top performing a fit Step4: With periodic boundaries, check that the filter "re-enters" the image from the other border Step5: Back to top TODO
<ASSISTANT_TASK:> Python Code: %load_ext autoreload %autoreload 2 from LogGabor import LogGabor parameterfile = 'https://raw.githubusercontent.com/bicv/LogGabor/master/default_param.py' lg = LogGabor(parameterfile) lg.set_size((32, 32)) Explanation: LogGabor user guide Table of content What is the LogGabor package? Installing Importing the library Properties of log-Gabor filters Testing filter generation Testing on a sample image Building a pyramid An example of fitting images with log-Gabor filters Importing the library End of explanation import os import numpy as np np.set_printoptions(formatter={'float': '{: 0.3f}'.format}) %matplotlib inline import matplotlib.pyplot as plt fig_width = 12 figsize=(fig_width, .618*fig_width) Explanation: To install the dependencies related to running this notebook, see Installing notebook dependencies. Back to top End of explanation def twoD_Gaussian(xy, x_pos, y_pos, theta, sf_0): FT_lg = lg.loggabor(x_pos, y_pos, sf_0=np.absolute(sf_0), B_sf=lg.pe.B_sf, theta=theta, B_theta=lg.pe.B_theta) return lg.invert(FT_lg).ravel() # Create x and y indices x = np.arange(lg.pe.N_X) y = np.arange(lg.pe.N_Y) x, y = xy = np.meshgrid(x, y) #create data x_pos, y_pos, theta, sf_0 = 14.6, 8.5, 12 * np.pi / 180., .1 data = twoD_Gaussian(xy, x_pos, y_pos, theta=theta, sf_0=sf_0) # plot twoD_Gaussian data generated above #plt.figure() #plt.imshow(data.reshape(lg.pe.N_X, lg.pe.N_Y)) #plt.colorbar() # add some noise to the data and try to fit the data generated beforehand data /= np.abs(data).max() data_noisy = data + .25*np.random.normal(size=data.shape) # getting best match C = lg.linear_pyramid(data_noisy.reshape(lg.pe.N_X, lg.pe.N_Y)) idx = lg.argmax(C) initial_guess = [idx[0], idx[1], lg.theta[idx[2]], lg.sf_0[idx[3]]] print ('initial_guess :', initial_guess, ', idx :', idx) import scipy.optimize as opt popt, pcov = opt.curve_fit(twoD_Gaussian, xy, data_noisy, p0=initial_guess) data_fitted = twoD_Gaussian(xy, *popt) extent = (0, lg.pe.N_X, 0, lg.pe.N_Y) print ('popt :', popt, ', true : ', x_pos, y_pos, theta, sf_0) fig, axs = plt.subplots(1, 3, figsize=(15, 5)) _ = axs[0].contourf(data.reshape(lg.pe.N_X, lg.pe.N_Y), 8, extent=extent, cmap=plt.cm.viridis, origin='upper') _ = axs[1].imshow(data_noisy.reshape(lg.pe.N_X, lg.pe.N_Y), cmap=plt.cm.viridis, extent=extent) _ = axs[2].contourf(data_fitted.reshape(lg.pe.N_X, lg.pe.N_Y), 8, extent=extent, cmap=plt.cm.viridis, origin='upper') for ax in axs: ax.axis('equal') Explanation: Perspectives: Better fits of the filters Basically, it is possible to infer the best possible log-Gabor function, even if it's parameters do not fall on the grid Defining a reference log-gabor (look in the corners!) End of explanation from LogGabor import LogGaborFit lg = LogGaborFit(parameterfile) lg.set_size((32, 32)) x_pos, y_pos, theta, sf_0 = 14.6, 8.5, 12 * np.pi / 180., .1 data = lg.invert(lg.loggabor(x_pos, y_pos, sf_0=np.absolute(sf_0), B_sf=lg.pe.B_sf, theta=theta, B_theta=lg.pe.B_theta)) data /= np.abs(data).max() data_noisy = data + .25*np.random.normal(size=data.shape) data_fitted, params = lg.LogGaborFit(data_noisy.reshape(lg.pe.N_X, lg.pe.N_Y)) data_fitted.shape params.pretty_print() extent = (0, lg.pe.N_X, 0, lg.pe.N_Y) print ('params :', params, ', true : ', x_pos, y_pos, theta, sf_0) fig, axs = plt.subplots(1, 3, figsize=(15, 5)) _ = axs[0].contourf(data.reshape(lg.pe.N_X, lg.pe.N_Y), 8, extent=extent, cmap=plt.cm.viridis, origin='upper') _ = axs[1].imshow(data_noisy.reshape(lg.pe.N_X, lg.pe.N_Y), cmap=plt.cm.viridis, extent=extent) _ = axs[2].contourf(data_fitted.reshape(lg.pe.N_X, lg.pe.N_Y), 8, extent=extent, cmap=plt.cm.viridis, origin='upper') for ax in axs: ax.axis('equal') Explanation: Back to top performing a fit End of explanation data_fitted, params = lg.LogGaborFit(data_noisy.reshape(lg.pe.N_X, lg.pe.N_Y), do_border=False) extent = (0, lg.pe.N_X, 0, lg.pe.N_Y) print ('params :', params, ', true : ', x_pos, y_pos, theta, sf_0) fig, axs = plt.subplots(1, 3, figsize=(15, 5)) _ = axs[0].contourf(data.reshape(lg.pe.N_X, lg.pe.N_Y), 8, extent=extent, cmap=plt.cm.viridis, origin='upper') _ = axs[1].imshow(data_noisy.reshape(lg.pe.N_X, lg.pe.N_Y), cmap=plt.cm.viridis, extent=extent) _ = axs[2].contourf(data_fitted.reshape(lg.pe.N_X, lg.pe.N_Y), 8, extent=extent, cmap=plt.cm.viridis, origin='upper') for ax in axs: ax.axis('equal') Explanation: With periodic boundaries, check that the filter "re-enters" the image from the other border: End of explanation %load_ext watermark %watermark -i -h -m -v -p numpy,matplotlib,scipy,imageio,SLIP,LogGabor -r -g -b Explanation: Back to top TODO: validation of fits Back to top more book keeping End of explanation <END_TASK>
15,969
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Porting Bike-Sharing project-1 to RNN Step1: Load and prepare the data A critical step in working with neural networks is preparing the data correctly. Variables on different scales make it difficult for the network to efficiently learn the correct weights. Below, we've written the code to load and prepare the data. You'll learn more about this soon! Step2: Checking out the data This dataset has the number of riders for each hour of each day from January 1 2011 to December 31 2012. The number of riders is split between casual and registered, summed up in the cnt column. You can see the first few rows of the data above. Below is a plot showing the number of bike riders over the first 10 days or so in the data set. (Some days don't have exactly 24 entries in the data set, so it's not exactly 10 days.) You can see the hourly rentals here. This data is pretty complicated! The weekends have lower over all ridership and there are spikes when people are biking to and from work during the week. Looking at the data above, we also have information about temperature, humidity, and windspeed, all of these likely affecting the number of riders. You'll be trying to capture all this with your model. Step3: Dummy variables Here we have some categorical variables like season, weather, month. To include these in our model, we'll need to make binary dummy variables. This is simple to do with Pandas thanks to get_dummies(). Step4: Scaling target variables To make training the network easier, we'll standardize each of the continuous variables. That is, we'll shift and scale the variables such that they have zero mean and a standard deviation of 1. The scaling factors are saved so we can go backwards when we use the network for predictions. Step5: Splitting the data into training, testing, and validation sets We'll save the data for the last approximately 21 days to use as a test set after we've trained the network. We'll use this set to make predictions and compare them with the actual number of riders. Step6: We'll split the data into two sets, one for training and one for validating as the network is being trained. Since this is time series data, we'll train on historical data, then try to predict on future data (the validation set). Step7: Lets build the get_batches the input of each time step will be one row of data ( one hour prediction array of 53 features) Step8: Lets test the batches above Step9: Build the network Lets build a RNN with tensorflow Inputs Step10: Lstm Step11: Output Step12: Validation Accuracy Step13: Trainning Step14: Check out your predictions Here, use the test data to view how well your network is modeling the data. If something is completely wrong here, make sure each step in your network is implemented correctly. Step15: Create a graph to compare the data and predictions
<ASSISTANT_TASK:> Python Code: %matplotlib inline %config InlineBackend.figure_format = 'retina' import sys import numpy as np import pandas as pd import matplotlib.pyplot as plt Explanation: Porting Bike-Sharing project-1 to RNN End of explanation data_path = 'Bike-Sharing-Dataset/hour.csv' rides = pd.read_csv(data_path) rides.head() Explanation: Load and prepare the data A critical step in working with neural networks is preparing the data correctly. Variables on different scales make it difficult for the network to efficiently learn the correct weights. Below, we've written the code to load and prepare the data. You'll learn more about this soon! End of explanation rides[:24*10].plot(x='dteday', y='cnt') Explanation: Checking out the data This dataset has the number of riders for each hour of each day from January 1 2011 to December 31 2012. The number of riders is split between casual and registered, summed up in the cnt column. You can see the first few rows of the data above. Below is a plot showing the number of bike riders over the first 10 days or so in the data set. (Some days don't have exactly 24 entries in the data set, so it's not exactly 10 days.) You can see the hourly rentals here. This data is pretty complicated! The weekends have lower over all ridership and there are spikes when people are biking to and from work during the week. Looking at the data above, we also have information about temperature, humidity, and windspeed, all of these likely affecting the number of riders. You'll be trying to capture all this with your model. End of explanation dummy_fields = ['season', 'weathersit', 'mnth', 'hr', 'weekday'] for each in dummy_fields: dummies = pd.get_dummies(rides[each], prefix=each, drop_first=False) rides = pd.concat([rides, dummies], axis=1) fields_to_drop = ['instant', 'dteday', 'season', 'weathersit', 'weekday', 'atemp', 'mnth', 'workingday', 'hr'] data = rides.drop(fields_to_drop, axis=1) data.head() Explanation: Dummy variables Here we have some categorical variables like season, weather, month. To include these in our model, we'll need to make binary dummy variables. This is simple to do with Pandas thanks to get_dummies(). End of explanation quant_features = ['casual', 'registered', 'cnt', 'temp', 'hum', 'windspeed'] # Store scalings in a dictionary so we can convert back later scaled_features = {} for each in quant_features: mean, std = data[each].mean(), data[each].std() scaled_features[each] = [mean, std] data.loc[:, each] = (data[each] - mean)/std Explanation: Scaling target variables To make training the network easier, we'll standardize each of the continuous variables. That is, we'll shift and scale the variables such that they have zero mean and a standard deviation of 1. The scaling factors are saved so we can go backwards when we use the network for predictions. End of explanation # Save data for approximately the last 21 days test_data = data[-21*24:] # Now remove the test data from the data set data = data[:-21*24] # Separate the data into features and targets target_fields = ['cnt', 'casual', 'registered'] features, targets = data.drop(target_fields, axis=1), data[target_fields] test_features, test_targets = test_data.drop(target_fields, axis=1), test_data[target_fields] Explanation: Splitting the data into training, testing, and validation sets We'll save the data for the last approximately 21 days to use as a test set after we've trained the network. We'll use this set to make predictions and compare them with the actual number of riders. End of explanation # Hold out the last 60 days or so of the remaining data as a validation set train_features, train_targets = features[:-60*24], targets[:-60*24] val_features, val_targets = features[-60*24:], targets[-60*24:] train_targets.head() Explanation: We'll split the data into two sets, one for training and one for validating as the network is being trained. Since this is time series data, we'll train on historical data, then try to predict on future data (the validation set). End of explanation # each element of x is an array with 53 features and each element of y is an array with 3 targets # each x is one hour features def get_batches(x, y, n_seqs, n_steps): '''Create a generator that returns batches of size n_seqs x n_steps from arr. Arguments --------- array x and array y: Array you want to make batches from n_seqs: Batch size, the number of sequences per batch n_steps: Number of sequence steps per batch ''' # Get the number of hours per batch and number of batches we can make hours_per_batch = n_seqs * n_steps n_batches = len(x)//hours_per_batch # convert from Pandas to np remove the index column x = x.reset_index().values[:,1:] y = y.reset_index().values[:,1:] # make only full batches x, y = x[:n_batches*hours_per_batch], y[:n_batches*hours_per_batch] # TODO: this needs to be optmized # x_temp will be ( n rows x n_steps wide) where each element is an array of 53 features # this first look splits the x with n rows and n_steps wide x_temp = [] y_temp = [] for st in range(0, n_batches*hours_per_batch, n_steps ): x_temp.append( x[st:st+n_steps] ) y_temp.append( y[st:st+n_steps] ) x = np.asarray(x_temp ) y = np.asarray(y_temp ) # this splits x in n_seqs rows so the return is a batch of n_seqs rows with n_steps wide # where each element is an array of 53 features (one hour from our data) for sq in range(0,(n_batches*hours_per_batch)//n_steps, n_seqs ): yield x[sq:sq+n_seqs,:,:], y[sq:sq+n_seqs,:,:] Explanation: Lets build the get_batches the input of each time step will be one row of data ( one hour prediction array of 53 features) End of explanation print(train_features.tail()) batches = get_batches(train_features, train_targets, 20, 96) x, y = next(batches) print(x.shape) # x, y = next(batches) # print(x.shape) Explanation: Lets test the batches above End of explanation import tensorflow as tf num_features = 56 num_targets = 3 batch_size = 10 # one step for each hour that we want the sequence to remember num_steps = 50 lstm_size = 256 num_layers = 2 learning_rate = 0.0005 keep_prob_val = 0.75 # Declare placeholders we'll feed into the graph inputs = tf.placeholder(tf.float32, [batch_size, None, num_features], name='inputs') targets = tf.placeholder(tf.float32, [batch_size, None, num_targets], name='targets') # Keep probability placeholder for drop out layers keep_prob = tf.placeholder(tf.float32, name='keep_prob') learningRate = tf.placeholder(tf.float32, name='learningRate') Explanation: Build the network Lets build a RNN with tensorflow Inputs End of explanation # # Use a basic LSTM cell # lstm = tf.contrib.rnn.BasicLSTMCell(lstm_size) # # Add dropout to the cell # drop = tf.contrib.rnn.DropoutWrapper(lstm, output_keep_prob=keep_prob) # # Stack up multiple LSTM layers, for deep learning # #cell = tf.contrib.rnn.MultiRNNCell([drop] * num_layers) # initial_state = cell.zero_state(batch_size, tf.float32) #Replaced the code above because TF with GPU was complaining def lstm_cell(): cell = tf.contrib.rnn.BasicLSTMCell(lstm_size) return tf.contrib.rnn.DropoutWrapper(cell, output_keep_prob=keep_prob) cell = tf.contrib.rnn.MultiRNNCell([lstm_cell() for _ in range(num_layers)], state_is_tuple = True) initial_state = cell.zero_state(batch_size, tf.float32) Explanation: Lstm End of explanation outputs, final_state = tf.nn.dynamic_rnn(cell, inputs, dtype=tf.float32) # this is one thing that I still dont fully understood, I had to set the activation_fn=None so the # fully connected layer dont use any activation funcition, this seems to work predictions = tf.contrib.layers.fully_connected(outputs, 3, activation_fn=None) cost = tf.losses.mean_squared_error(targets, predictions) optimizer = tf.train.AdamOptimizer(learning_rate).minimize(cost) Explanation: Output End of explanation correct_pred = tf.equal(tf.cast(tf.round(predictions), tf.int32), tf.cast(tf.round(targets), tf.int32)) accuracy = tf.reduce_mean(tf.cast(correct_pred, tf.float32)) Explanation: Validation Accuracy End of explanation epochs = 100 saver = tf.train.Saver() #validation accuracy to plot val_accuracy=[] training_loss=[] with tf.Session() as sess: sess.run(tf.global_variables_initializer()) iteration = 1 for e in range(epochs): state = sess.run(initial_state) for ii, (x, y) in enumerate(get_batches(train_features, train_targets, batch_size, num_steps), 1): feed = {inputs: x, targets: y, keep_prob: keep_prob_val, initial_state: state} loss, state, _ = sess.run([cost, final_state, optimizer], feed_dict=feed) if iteration%5==0: print("Epoch: {}/{}".format(e, epochs), "Iteration: {}".format(iteration), "Train loss: {:.3f}".format(loss)) training_loss.append(loss) if iteration%25==0: val_acc = [] val_state = sess.run(cell.zero_state(batch_size, tf.float32)) for x, y in get_batches(val_features, val_targets, batch_size, num_steps): feed = {inputs: x, targets: y, keep_prob: 1, initial_state: val_state} batch_acc, val_state = sess.run([accuracy, final_state], feed_dict=feed) val_acc.append(batch_acc) val_accuracy.append( np.mean(val_acc) ) print("Val acc: {:.3f}".format(np.mean(val_acc))) iteration +=1 saver.save(sess, "checkpoints/bike-sharing.ckpt") plt.plot(val_accuracy, label='Accuracy') plt.legend() _ = plt.ylim() plt.plot(training_loss, label='Loss') plt.legend() _ = plt.ylim() Explanation: Trainning End of explanation test_acc = [] #with tf.Session(graph=graph) as sess: with tf.Session() as sess: saver.restore(sess, tf.train.latest_checkpoint('checkpoints')) test_state = sess.run(cell.zero_state(batch_size, tf.float32)) for ii, (x, y) in enumerate(get_batches(test_features, test_targets, batch_size, num_steps), 1): feed = {inputs: x, targets: y, keep_prob: 1, initial_state: test_state} batch_acc, test_state = sess.run([accuracy, final_state], feed_dict=feed) test_acc.append(batch_acc) print("Test accuracy: {:.3f}".format(np.mean(test_acc))) Explanation: Check out your predictions Here, use the test data to view how well your network is modeling the data. If something is completely wrong here, make sure each step in your network is implemented correctly. End of explanation with tf.Session() as sess: saver.restore(sess, tf.train.latest_checkpoint('checkpoints')) test_state = sess.run(cell.zero_state(batch_size, tf.float32)) fig, ax = plt.subplots(figsize=(8,4)) mean, std = scaled_features['cnt'] batch = get_batches(test_features, test_targets, batch_size, num_steps) x,y = next(batch) feed = {inputs: x, targets: y, keep_prob: 1, initial_state: test_state} pred = sess.run([predictions], feed_dict=feed) pred = pred[0].reshape(500,-1) pred[:,0] *= std pred[:,0] += mean lf = pred[:,0] # predictions = network.run(test_features).T*std + mean ax.plot(lf, label='Prediction') ax.plot((test_targets['cnt']*std + mean).values, label='Data') ax.set_xlim(right=len(lf)) ax.legend() dates = pd.to_datetime(rides.ix[test_data.index]['dteday']) dates = dates.apply(lambda d: d.strftime('%b %d')) ax.set_xticks(np.arange(len(dates))[12::24]) _ = ax.set_xticklabels(dates[12::24], rotation=45) Explanation: Create a graph to compare the data and predictions End of explanation <END_TASK>
15,970
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: ЦАО Step1: Собираем ссылки на все квартиры первых тридцати страниц выдачи Step2: Cтандартный блок, в котором мы получаем по ссылке текст страницы в удобном формате Step3: Prices Функция для извлечения цены квартиры из странички, взято из лекции Step4: Функция для сбора всех цен с сайта в выбранном диапазоне ссылок ( l < r ) Step5: Теперь используем эти функции и составим колонку для всех цен Step6: Dist Извлечем координаты так же, как это было продемонстировано на лекции Step7: Определим функцию получения координат из текста страницы Step8: И функцию для получения всех координат в выбранном диапазоне ссылок (такая схема из двух функций будет повторяться везде) Step9: Теперь нам предстоит использовать полученные координаты для вычисления расстояния до центра города. Воспользуемся для этого формулой для вычисления расстояния между двумя точками, расположенными на сфере. Запишем отдельную функцию Step10: Rooms Перейдем к числу комнат, тоже сделано по аналогии с лекционным ноутбуком Step11: Страницы о многокомнатных квартирах ( > 5 ) не содержат информации о количестве комнат в явном виде, такие квартиры фигурируют как "многокомнатные". Так что соответствующие объекты у меня получат в столбец Rooms значение 'mult'. Это более достоверная информация, чем, например, 6. Step12: Metrdist и Walking Теперь начнем разбираться с расстоянием до метро, заодно и со способом до него добраться (пешком или на машине) Step13: Тут у нас появляется отличие от того, что было раньше Step14: Walking Аналогично поступим с колонкой walking Step15: Теперь перейдем к работе с данными, которые мы будем извлекать из таблицы. Это будущие колонки Step16: Информация по Brick и New содержится между следующими двумя обязательными полями Step17: Ситуация следующая Step18: New Теперь перейдем к New - тут либо один вариант из двух (1 - новостройка, 0 - вторичка), либо nan. Step19: Теперь переходим дальше Floor, Nfloors Step20: Информация по Floor и Nfloors тоже содержится между обязательными полями "Этаж" и "Тип продажи". Step21: Floor указан обязательно, и он лежит во второй ячейке, Nfloors может не быть, но если есть - то он в третьей. Floor Step22: Nfloors Step23: Перейдем дальше Totsp, Livesp Step24: Всех полей, что находятся справа от "Общая площадь", может не быть. Так что ограничим нашу область поиска только слева. Step25: Totsp и Livesp будут обязательно. Тут мы столкнемся с проблемой Step26: Totsp Step27: Livesp Step28: Переходим к опциональным полям Kitsp, Bal, Tel Step29: Нужный нам подблок информации мы снова можем ограничить только слева названием поля "Общая площадь" Step30: Kitsp Если площадь кухни не указана, то она фигурирует в таблице в виде прочерка. При этом она может быть float с запятой в качестве разделителя, как и предыдущие площади. Учтем это Step31: Bal Балкона может не быть, балкон может быть, а еще балконов может быть больше одного. Информации о балконе может не быть. Так что будем готовы к значениям > 1. Прочерк будет говорить о том, что информации нет, слово "нет" - о том, что есть информация о том, что нет балкона. Step32: Tel Телефон может быть, может не быть, и о нем может не быть информации. О том, что он есть, нам сообщает слово "да", о том, что его нет - слово "нет". Step33: Последние штрихи перед сборкой DataFrame N Step34: В этом ноутбуке квартиры только из ЦАО, укажу это District Step35: Собираем все воедино DataFrame
<ASSISTANT_TASK:> Python Code: district = 'http://www.cian.ru/cat.php?deal_type=sale&district%5B0%5D=13&district%5B1%5D=14&district%5B2%5D=15&district%5B3%5D=16&district%5B4%5D=17&district%5B5%5D=18&district%5B6%5D=19&district%5B7%5D=20&district%5B8%5D=21&district%5B9%5D=22&engine_version=2&offer_type=flat&p={}&room1=1&room2=1&room3=1&room4=1&room5=1&room6=1' Explanation: ЦАО End of explanation links = [] for page in range(1, 30): page_url = district.format(page) search_page = requests.get(page_url) search_page = search_page.content search_page = BeautifulSoup(search_page, 'lxml') flat_urls = search_page.findAll('div', attrs = {'ng-class':"{'serp-item_removed': offer.remove.state, 'serp-item_popup-opened': isPopupOpen}"}) flat_urls = re.split('http://www.cian.ru/sale/flat/|/" ng-class="', str(flat_urls)) for link in flat_urls: if link.isdigit(): links.append(link) Explanation: Собираем ссылки на все квартиры первых тридцати страниц выдачи End of explanation flat_url = 'http://www.cian.ru/sale/flat/' + str(links[0]) + '/' #flat_url = 'http://www.cian.ru/sale/flat/150531912/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') Explanation: Cтандартный блок, в котором мы получаем по ссылке текст страницы в удобном формате End of explanation def getPrice(flat_page): price = flat_page.find('div', attrs={'class':'object_descr_price'}) price = re.split('<div>|руб|\W', str(price)) price = "".join([i for i in price if i.isdigit()][-4:]) return int(price) Explanation: Prices Функция для извлечения цены квартиры из странички, взято из лекции End of explanation def getAllPrices(l, r): prices = [] for i in range(l, r): flat_url = 'http://www.cian.ru/sale/flat/' + str(links[i]) + '/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') prices.append(getPrice(flat_page)) return prices Explanation: Функция для сбора всех цен с сайта в выбранном диапазоне ссылок ( l < r ) End of explanation prices = getAllPrices(0, len(links)) Explanation: Теперь используем эти функции и составим колонку для всех цен End of explanation flat_url = 'http://www.cian.ru/sale/flat/' + str(links[0]) + '/' #flat_url = 'http://www.cian.ru/sale/flat/150531912/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') coords = flat_page.find('div', attrs={'class':'map_info_button_extend'}).contents[1] coords = re.split('&amp|center=|%2C', str(coords)) coords coords_list = [] for item in coords: if item[0].isdigit(): coords_list.append(item) lat = float(coords_list[0]) lon = float(coords_list[1]) lat lon Explanation: Dist Извлечем координаты так же, как это было продемонстировано на лекции End of explanation def getCoords_at(flat_page): coords = flat_page.find('div', attrs={'class':'map_info_button_extend'}).contents[1] coords = re.split('&amp|center=|%2C', str(coords)) coords_list = [] for item in coords: if item[0].isdigit(): coords_list.append(item) lat = float(coords_list[0]) lon = float(coords_list[1]) return lat, lon Explanation: Определим функцию получения координат из текста страницы End of explanation def getAllCoordinates(l, r): coordinates = [] for i in range(l, r): flat_url = 'http://www.cian.ru/sale/flat/' + str(links[i]) + '/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') coordinates.append(getCoords(flat_page)) return coordinates coordinates = getAllCoordinates(0, len(links)) Explanation: И функцию для получения всех координат в выбранном диапазоне ссылок (такая схема из двух функций будет повторяться везде) End of explanation from math import radians, cos, sin, asin, sqrt AVG_EARTH_RADIUS = 6371 def haversine(point1, point2): # извлекаем долготу и широту lat1, lng1 = point1 lat2, lng2 = point2 # переводим все эти значения в радианы lat1, lng1, lat2, lng2 = map(radians, (lat1, lng1, lat2, lng2)) # вычисляем расстояние по формуле lat = lat2 - lat1 lng = lng2 - lng1 d = sin(lat * 0.5) ** 2 + cos(lat1) * cos(lat2) * sin(lng * 0.5) ** 2 h = 2 * AVG_EARTH_RADIUS * asin(sqrt(d)) return h MSC_POINT_ZERO = (55.755831, 37.617673) distance = [] for i in range(0, len(coordinates)): distance.append(haversine(MSC_POINT_ZERO, coordinates[i])) Explanation: Теперь нам предстоит использовать полученные координаты для вычисления расстояния до центра города. Воспользуемся для этого формулой для вычисления расстояния между двумя точками, расположенными на сфере. Запишем отдельную функцию End of explanation flat_url = 'http://www.cian.ru/sale/flat/' + str(links[2]) + '/' #flat_url = 'http://www.cian.ru/sale/flat/150844464/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') rooms_n = flat_page.find('div', attrs={'class':'object_descr_title'}) rooms_n = html_stripper(rooms_n) rooms_n re.split('-|\n', rooms_n) Explanation: Rooms Перейдем к числу комнат, тоже сделано по аналогии с лекционным ноутбуком End of explanation def getRoom(flat_page): rooms_n = flat_page.find('div', attrs={'class':'object_descr_title'}) rooms_n = html_stripper(rooms_n) room_number = '' flag = 0 for i in re.split('-|\n', rooms_n): if 'много' in i: flag = 1 break elif 'комн' in i: break else: room_number += i if (flag): room_number = 'mult' room_number = "".join(room_number.split()) return room_number def getAllRooms(l, r): rooms = [] for i in range(l, r): flat_url = 'http://www.cian.ru/sale/flat/' + str(links[i]) + '/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') rooms.append(getRoom(flat_page)) return rooms rooms = getAllRooms(0, len(links)) Explanation: Страницы о многокомнатных квартирах ( > 5 ) не содержат информации о количестве комнат в явном виде, такие квартиры фигурируют как "многокомнатные". Так что соответствующие объекты у меня получат в столбец Rooms значение 'mult'. Это более достоверная информация, чем, например, 6. End of explanation #flat_url = 'http://www.cian.ru/sale/flat/' + str(links[2]) + '/' flat_url = 'http://www.cian.ru/sale/flat/150387502/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') metro = flat_page.find('div', attrs={'class':'object_descr_metro'}) metro = re.split('metro_name|мин', str(metro)) metro re.split('metro_name|мин', str(metro)) Explanation: Metrdist и Walking Теперь начнем разбираться с расстоянием до метро, заодно и со способом до него добраться (пешком или на машине) End of explanation def getMetroDistance(flat_page): metro = flat_page.find('div', attrs={'class':'object_descr_metro'}) metro = re.split('metro_name|мин', str(metro)) if (len(metro) > 2): # если оба поля не были заполнены, то предыдущий сплит даст размерность 2 metro_dist = 0 power = 0 # кусок metro[1] после сплита будет содержать в конце кучу хлама, потом количество минут (если есть) flag = 0 for i in range(0, len(metro[1])): if metro[1][-i-1].isdigit(): flag = 1 metro_dist += int(metro[1][-i-1]) * 10 ** power power += 1 elif (flag == 1): break else: metro_dist = np.nan return metro_dist def getAllMetroDistances(l, r): metro_distance = [] for i in range(l, r): flat_url = 'http://www.cian.ru/sale/flat/' + str(links[i]) + '/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') metro_distance.append(getMetroDistance(flat_page)) return metro_distance metro_distances = getAllMetroDistances(0, len(links)) Explanation: Тут у нас появляется отличие от того, что было раньше: некоторые поля на Циане являются обязательными для заполнения, а некоторые нет, и вот эти поля обязательными не являются. Класть в данные по квартире, по которой нет этой информации, дефолтный 0 не будет правильным, потому что люди могли не указать как 3 минуты пешком (ну мало ли, лень было вписывать что-то в ячейки, квартира и так продастся), а могли не указать 30 минут на машине. Если уравнивать такие вещи (и тем более все их забивать нулями), то позже это плохо скажется на модели. Так что будем класть в такие ячейки nan'ы. Metrdist Функция для извлечения расстояния до метро в минутах End of explanation def getMetroWalking(flat_page): metro = flat_page.find('div', attrs={'class':'object_descr_metro'}) metro = re.split('metro_name|мин', str(metro)) if (len(metro) > 2): # если оба поля не были заполнены, то предыдущий сплит даст размерность 2 if 'пешк' in metro[2]: walking = 1 elif 'машин' in metro[2]: walking = 0 else: walking = np.nan # да, проверка на то, отсутствовали ли оба поля была. мне лично не попадались ситуации, где бы не # было заполнено только значение поля "пешком/на машине", но вдруг они есть? на такой случай проверка else: walking = np.nan return walking def getAllMetroWalking(l, r): metro_walking = [] for i in range(l, r): flat_url = 'http://www.cian.ru/sale/flat/' + str(links[i]) + '/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') metro_walking.append(getMetroWalking(flat_page)) return metro_walking walking = getAllMetroWalking(0, len(links)) Explanation: Walking Аналогично поступим с колонкой walking End of explanation #flat_url = 'http://www.cian.ru/sale/flat/' + str(links[2]) + '/' flat_url = 'http://www.cian.ru/sale/flat/150387502/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') table = flat_page.find('table', attrs = {'class':'object_descr_props'}) table = html_stripper(table) table Explanation: Теперь перейдем к работе с данными, которые мы будем извлекать из таблицы. Это будущие колонки: Bricks - кирпичный ли дом, монолитный (варианта жб на Циане нет)/иное New - новостройка/вторичка Totsp, Livsp, Kitsp - общая, жилая и площадь кухни соответственно Tel - наличие телефона Bal - наличие балкона Floor - этаж Nfloors - всего этажей в доме. Среди вышеперечисленного обязательны к заполнению только поля Totsp, Livsp и Floor. В остальных местах могут быть nan'ы. Bricks End of explanation building_block = re.split('Этаж|Тип продажи', table)[1] building_block Explanation: Информация по Brick и New содержится между следующими двумя обязательными полями: "Этаж" и "Тип продажи" (то есть на наличие этих слов в этом месте можно положиться) End of explanation def getBrick(flat_page): table = flat_page.find('table', attrs = {'class':'object_descr_props'}) table = html_stripper(table) brick = np.nan building_block = re.split('Этаж|Тип продажи', table)[1] if 'Тип дом' in building_block: if (('кирпич' in building_block) | ('монолит' in building_block)): brick = 1 elif (('панельн' in building_block) | ('деревян' in building_block) | ('сталин' in building_block) | ('блочн' in building_block)): brick = 0 return brick def getAllBricks(l, r): bricks = [] for i in range(l, r): flat_url = 'http://www.cian.ru/sale/flat/' + str(links[i]) + '/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') bricks.append(getBrick(flat_page)) return bricks bricks = getAllBricks(0, len(links)) Explanation: Ситуация следующая: поле "Тип дома" может быть, его может не быть (тогда brick = nan). Если оно есть, то оно может содержать информацию по полю Brick, а может и не содержать. Поле Brick может задавать два значения: 1- если это кирпичный или монолитный (или кирпично-монолитный), и 0 - если это что-то другое (и мы заранее можем посмотреть на сайте на то, какие еще бывают варианты: панельный, блочный, деревянный и сталинский). Запишем эти размышления в функцию End of explanation def getNew(flat_page): table = flat_page.find('table', attrs = {'class':'object_descr_props'}) table = html_stripper(table) new = np.nan building_block = re.split('Этаж|Тип продажи', table)[1] if 'Тип дом' in building_block: if 'новостр' in building_block: new = 1 elif 'втор' in building_block: new = 0 return new def getAllNew(l, r): new = [] for i in range(l, r): flat_url = 'http://www.cian.ru/sale/flat/' + str(links[i]) + '/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') new.append(getNew(flat_page)) return new new = getAllNew(0, len(links)) Explanation: New Теперь перейдем к New - тут либо один вариант из двух (1 - новостройка, 0 - вторичка), либо nan. End of explanation #flat_url = 'http://www.cian.ru/sale/flat/' + str(links[2]) + '/' flat_url = 'http://www.cian.ru/sale/flat/150387502/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') table = flat_page.find('table', attrs = {'class':'object_descr_props'}) table = html_stripper(table) table Explanation: Теперь переходим дальше Floor, Nfloors End of explanation building_block = re.split('Этаж|Тип продажи', table)[1] building_block floor_block = re.split('\xa0/\xa0|\n|\xa0', building_block) floor_block Explanation: Информация по Floor и Nfloors тоже содержится между обязательными полями "Этаж" и "Тип продажи". End of explanation def getFloor(flat_page): table = flat_page.find('table', attrs = {'class':'object_descr_props'}) table = html_stripper(table) floor_is = 0 building_block = re.split('Этаж|Тип продажи', table)[1] floor_block = re.split('\xa0/\xa0|\n|\xa0', building_block) for i in range(1, len(floor_block[2]) + 1): if(floor_block[2][-i].isdigit()): floor_is += int(floor_block[2][-i]) * 10**(i - 1) return floor_is def getAllFloors(l, r): floors = [] for i in range(l, r): flat_url = 'http://www.cian.ru/sale/flat/' + str(links[i]) + '/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') floors.append(getFloor(flat_page)) return floors floors = getAllFloors(0, len(links)) Explanation: Floor указан обязательно, и он лежит во второй ячейке, Nfloors может не быть, но если есть - то он в третьей. Floor End of explanation def getNFloor(flat_page): table = flat_page.find('table', attrs = {'class':'object_descr_props'}) table = html_stripper(table) floors_count = np.nan building_block = re.split('Этаж|Тип продажи', table)[1] floor_block = re.split('\xa0/\xa0|\n|\xa0', building_block) if floor_block[3].isdigit(): floors_count = int(floor_block[3]) return floors_count def getAllNFloors(l, r): nfloors = [] for i in range(l, r): flat_url = 'http://www.cian.ru/sale/flat/' + str(links[i]) + '/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') nfloors.append(getNFloor(flat_page)) return nfloors nfloors = getAllNFloors(0, 20) Explanation: Nfloors End of explanation #flat_url = 'http://www.cian.ru/sale/flat/' + str(links[2]) + '/' flat_url = 'http://www.cian.ru/sale/flat/150387502/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') table = flat_page.find('table', attrs = {'class':'object_descr_props'}) table = html_stripper(table) table Explanation: Перейдем дальше Totsp, Livesp End of explanation space_block = re.split('Общая площадь', table)[1] space_block Explanation: Всех полей, что находятся справа от "Общая площадь", может не быть. Так что ограничим нашу область поиска только слева. End of explanation def myStrToFloat(string): delimiter = 0 value = 0 for i in range(0, len(string)): if string[i] == ',': delimiter = i for i in range(0, delimiter): value += int(string[delimiter - i - 1]) * 10 ** i for i in range(1, len(string) - delimiter): value += (int(string[delimiter + i]) * (10 ** (i - 2))) return value Explanation: Totsp и Livesp будут обязательно. Тут мы столкнемся с проблемой: Циан записывает float через запятую, Python - через точку, поэтому дефолтная float(str) не сработает. Набросаем свой конвертор. End of explanation def getTotsp(flat_page): table = flat_page.find('table', attrs = {'class':'object_descr_props'}) table = html_stripper(table) space_block = re.split('Общая площадь', table)[1] total = re.split('Площадь комнат', space_block)[0] total_space = re.split('\n|\xa0', total)[2] if total_space.isdigit(): total_space = int(total_space) else: total_space = myStrToFloat(total_space) return total_space def getAllTotsp(l, r): totsp = [] for i in range(l, r): flat_url = 'http://www.cian.ru/sale/flat/' + str(links[i]) + '/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') totsp.append(getTotsp(flat_page)) return totsp totsp = getAllTotsp(0, len(links)) Explanation: Totsp End of explanation def getLivesp(flat_page): table = flat_page.find('table', attrs = {'class':'object_descr_props'}) table = html_stripper(table) space_block = re.split('Общая площадь', table)[1] living = re.split('Жилая площадь', space_block)[1] living_space = re.split('\n|\xa0', living)[2] if living_space.isdigit(): living_space = int(living_space) else: living_space = myStrToFloat(living_space) return living_space def getAllLivesp(l, r): livesp = [] for i in range(l, r): flat_url = 'http://www.cian.ru/sale/flat/' + str(links[i]) + '/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') livesp.append(getLivesp(flat_page)) return livesp livesp = getAllLivesp(0, len(links)) Explanation: Livesp End of explanation #flat_url = 'http://www.cian.ru/sale/flat/' + str(links[2]) + '/' flat_url = 'http://www.cian.ru/sale/flat/150387502/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') table = flat_page.find('table', attrs = {'class':'object_descr_props'}) table = html_stripper(table) table Explanation: Переходим к опциональным полям Kitsp, Bal, Tel End of explanation space_block = re.split('Общая площадь', table)[1] space_block optional_block = re.split('Жилая площадь', space_block)[1] optional_block Explanation: Нужный нам подблок информации мы снова можем ограничить только слева названием поля "Общая площадь" End of explanation def getKitsp(flat_page): table = flat_page.find('table', attrs = {'class':'object_descr_props'}) table = html_stripper(table) space_block = re.split('Общая площадь', table)[1] optional_block = re.split('Жилая площадь', space_block)[1] kitchen_space = np.nan if 'Площадь кухни' in optional_block: kitchen_block = re.split('Площадь кухни', optional_block)[1] if re.split('\n|\xa0', kitchen_block)[2] != '–': if re.split('\n|\xa0', kitchen_block)[2].isdigit(): kitchen_space = int(re.split('\n|\xa0', kitchen_block)[2]) else: kitchen_space = myStrToFloat(re.split('\n|\xa0', kitchen_block)[2]) return kitchen_space def getAllKitsp(l, r): kitsp = [] for i in range(l, r): flat_url = 'http://www.cian.ru/sale/flat/' + str(links[i]) + '/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') kitsp.append(getKitsp(flat_page)) return kitsp kitsp = getAllKitsp(0, len(links)) Explanation: Kitsp Если площадь кухни не указана, то она фигурирует в таблице в виде прочерка. При этом она может быть float с запятой в качестве разделителя, как и предыдущие площади. Учтем это End of explanation def getBal(flat_page): table = flat_page.find('table', attrs = {'class':'object_descr_props'}) table = html_stripper(table) space_block = re.split('Общая площадь', table)[1] optional_block = re.split('Жилая площадь', space_block)[1] balcony = np.nan if 'Балкон' in optional_block: balcony_block = re.split('Балкон', optional_block)[1] if re.split('\n', balcony_block)[1] != 'нет': if re.split('\n', balcony_block)[1] != '–': balcony = int(re.split('\n', balcony_block)[1][0]) else: balcony = 0 return balcony def getAllBal(l, r): bal = [] for i in range(l, r): flat_url = 'http://www.cian.ru/sale/flat/' + str(links[i]) + '/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') bal.append(getBal(flat_page)) return bal bal = getAllBal(0, len(links)) Explanation: Bal Балкона может не быть, балкон может быть, а еще балконов может быть больше одного. Информации о балконе может не быть. Так что будем готовы к значениям > 1. Прочерк будет говорить о том, что информации нет, слово "нет" - о том, что есть информация о том, что нет балкона. End of explanation def getTel(flat_page): table = flat_page.find('table', attrs = {'class':'object_descr_props'}) table = html_stripper(table) space_block = re.split('Общая площадь', table)[1] optional_block = re.split('Жилая площадь', space_block)[1] telephone = np.nan if 'Телефон' in optional_block: telephone_block = re.split('Телефон', optional_block)[1] if re.split('\n', telephone_block)[1] == 'да': telephone = 1 elif re.split('\n', telephone_block)[1] == 'нет': telephone = 0 return telephone def getAllTel(l, r): tel = [] for i in range(l, r): flat_url = 'http://www.cian.ru/sale/flat/' + str(links[i]) + '/' flat_page = requests.get(flat_url) flat_page = flat_page.content flat_page = BeautifulSoup(flat_page, 'lxml') tel.append(getTel(flat_page)) return tel tel = getAllTel(0, len(links)) Explanation: Tel Телефон может быть, может не быть, и о нем может не быть информации. О том, что он есть, нам сообщает слово "да", о том, что его нет - слово "нет". End of explanation N = [] for i in range(0, len(links)): N.append(i) Explanation: Последние штрихи перед сборкой DataFrame N End of explanation district = [] for i in range(0, len(links)): district.append('CAD') Explanation: В этом ноутбуке квартиры только из ЦАО, укажу это District End of explanation data = dict([('New', new), ('Bal', bal), ('Tel', tel), ('Walk', walk), ('Metrdist', metrdist), ('Nfloors', nfloors), ('Floor', floor), ('Totsp', totsp), ('Livesp', livesp), ('Kitsp', kitsp), ('N', N), ('Price', prices), ('Rooms', rooms), ('Distance', distance), ('Brick', bricks), ('District', district)]) df = pd.DataFrame(data) df.T df.to_csv('cian.csv', index=False) Explanation: Собираем все воедино DataFrame End of explanation <END_TASK>
15,971
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Table of Contents <p><div class="lev1"><a href="#Data-preparation"><span class="toc-item-num">1 - </span>Data preparation</a></div><div class="lev1"><a href="#Data-pipeline"><span class="toc-item-num">2 - </span>Data pipeline</a></div><div class="lev2"><a href="#Selecting-appliances"><span class="toc-item-num">2.1 - </span>Selecting appliances</a></div><div class="lev2"><a href="#Selecting-time-sequences"><span class="toc-item-num">2.2 - </span>Selecting time sequences</a></div><div class="lev2"><a href="#Selecting-houses"><span class="toc-item-num">2.3 - </span>Selecting houses</a></div><div class="lev2"><a href="#Dealing-with-unbalanced-data Step1: Data preparation We have power data for 6 houses from main meters that measure the aggregate power consumption of the house and from each appliance separately. For example here we show the aggregated data (sum of all appliances in the house) for house 2. Step2: One contribution to this aggregated data is the cycles of the fridge plotted below. Step3: Of course there are many other appliances that contributes to the aggregated data observed above. The list of appliances listed in this house is shown here Step4: The sum of all these appliances is highly correlated to the aggregated data. Step5: We can visualize this looking at the main meters for one day and the sum of all the appliances for one day. Step6: The correlation is high because the 9 appliances listed above consume a significant proportion of the energy and represent a range of different power 'signatures'. Here is one signature from the fridge and one signature from the microwave. Step7: Obviously the microwave signature is a pic of short duration while the fridge cycle is a bit longer. In this study we try to retrieve from the aggregated data the times an appliance switches on, the times it swithes off and the average power comsuption in each cycle. We call this the rectangle method. Essentially, we want to retrieve from the aggregated data, this type of representation Step8: Data pipeline Selecting appliances We train each neural network per appliance. This is different from the CO and FHMM methods. Selecting time sequences We down sample the main meters and the submeters to 6 samples per seconds in order to have the aggregated and the submeter sequences properly aligned. We throw away any activation shorter than some threshold duration to avoid spurious spikes. Selecting houses We choose to train the algorithm on house 1,2,3 and 6 and test the data on house 6. Dealing with unbalanced data Step9: Standardisation of the input data (aggregated data) A typical step in the data pipeline of neural network is the standardization of data. For each sequences of 512 samples (= 85 seconds) we substract the mean to center the sequence. Furthermore every input sequence is divided by the standard deviation of a random sample in the training set. In this case we cannot divide each sequence by its own standard deviation because it would delete information about the scale of the signal. This is done via Step11: Output data (start time, end time and average power) The output of the neural network is 3 neurons Step14: Implementation strategy for real time data augmentation While the neural network runs an NVIDIA GeForce GT 750M (GPU) we maintain the CPU busy doing the data augmentation in real time (load aggregated data, create the synthetic data, preprocess the mini-batch to be fed to the neural network). For this we create a python generator that creates a queue of 50 mini-batch and feed them successively to the GPU for training. The pipeline class can be found in neuralnilm.data.datapiline at https Step15: Network architecture We use a convolutional neural network (ConvNet) to take advantage of the translation invariance. We want the ConvNet to recognize target appliance activation anywhere in the sequence. For this project we have tried multiple architecture that are reported later on. These architecture all have a first convolutional layer of filter size 3 and stride 1. We have played with both the filter size and the number of output filters on the first layer. We have found that 16 filters is a reasonable number -- increasing the number of filters in the first layer did not improve significantly the scores. The best neural network we found consist of Input layer Step16: We found by experimenting that the best optimizer is Adam (http Step17: where dw is the gradient of the lost function with respect to the considered weight and w the considered weight. The hyperparameter for this optimizer are beta1, beta2 and eps. We set beta_1=0.9, beta_2=0.999, epsilon=1e-08 and the learning rate = 1e-3. Experimenting with ConvNet architectures For the implementation of the ConvNet we use Keras (http Step18: One convolutional layer Here we present the results from the ConvNet detailed in the previous cell. We first load the ConvNet architecture, the weights and the history of the training and validation losses (the network has been trained previously using GPU). Step19: This is a very good F1 score, much better than with any other method tested before. Here we show the evolution of the training loss and the validation loss. Step20: The training and validation loss during training decreases simulateanously. There is no overfitting. We stop the simulation after 20 epochs but the model was still improving. 20 epochs on GPU took about 1.5 hours. Two convolutional layers Since we are not overfitting we did not add any Pooling layer or Dropout to the previous network. However we experimented an other network with an additional convolutional layer and train the model over 30 epoch. The model is the following Step21: There are 2 convolutional layers with 16 filters and 2 dense layers and the output layer. In this case the results are fairly similar to the network presented above. Step22: Results for the microwave Here we run the ConvNet with one layer on the microwave. We train using house 1 and 2 and test on house 5.
<ASSISTANT_TASK:> Python Code: # for plotting purposes %matplotlib inline from matplotlib.pylab import plt from matplotlib import rcParams dark_colors = ["#A51C30", "#808080", (0.8509803921568627, 0.37254901960784315, 0.00784313725490196), (0.4588235294117647, 0.4392156862745098, 0.7019607843137254), (0.9058823529411765, 0.1607843137254902, 0.5411764705882353), (0.4, 0.6509803921568628, 0.11764705882352941), (0.9019607843137255, 0.6705882352941176, 0.00784313725490196), (0.6509803921568628, 0.4627450980392157, 0.11372549019607843), (0.4, 0.4, 0.4)] # dark_colors = ['#A51C30', '#808080'] rcParams['figure.figsize'] = (13, 6) rcParams['figure.dpi'] = 150 rcParams['axes.color_cycle'] = dark_colors rcParams['lines.linewidth'] = 2 rcParams['axes.facecolor'] = "white" rcParams['axes.titlesize'] = 20 rcParams['axes.labelsize'] = 17.5 rcParams['xtick.labelsize'] = 15 rcParams['ytick.labelsize'] = 15 rcParams['legend.fontsize'] = 17.5 rcParams['patch.edgecolor'] = 'none' rcParams['grid.color']="gray" rcParams['grid.linestyle']="-" rcParams['grid.linewidth'] = 0.3 rcParams['grid.alpha']=1 rcParams['text.color'] = "444444" rcParams['axes.labelcolor'] = "444444" rcParams['ytick.color'] = "444444" rcParams['xtick.color'] = "444444" Explanation: Table of Contents <p><div class="lev1"><a href="#Data-preparation"><span class="toc-item-num">1 - </span>Data preparation</a></div><div class="lev1"><a href="#Data-pipeline"><span class="toc-item-num">2 - </span>Data pipeline</a></div><div class="lev2"><a href="#Selecting-appliances"><span class="toc-item-num">2.1 - </span>Selecting appliances</a></div><div class="lev2"><a href="#Selecting-time-sequences"><span class="toc-item-num">2.2 - </span>Selecting time sequences</a></div><div class="lev2"><a href="#Selecting-houses"><span class="toc-item-num">2.3 - </span>Selecting houses</a></div><div class="lev2"><a href="#Dealing-with-unbalanced-data:-selecting-aggregated-data-windows"><span class="toc-item-num">2.4 - </span>Dealing with unbalanced data: selecting aggregated data windows</a></div><div class="lev2"><a href="#Synthetic-aggregated-data"><span class="toc-item-num">2.5 - </span>Synthetic aggregated data</a></div><div class="lev1"><a href="#Standardisation-of-the-input-data-(aggregated-data)"><span class="toc-item-num">3 - </span>Standardisation of the input data (aggregated data)</a></div><div class="lev1"><a href="#Output-data-(start-time,-end-time-and-average-power)"><span class="toc-item-num">4 - </span>Output data (start time, end time and average power)</a></div><div class="lev1"><a href="#Scores---performance-evaluation-of-the-neural-network"><span class="toc-item-num">5 - </span>Scores - performance evaluation of the neural network</a></div><div class="lev1"><a href="#Implementation-strategy-for-real-time-data-augmentation"><span class="toc-item-num">6 - </span>Implementation strategy for real time data augmentation</a></div><div class="lev1"><a href="#Network-architecture"><span class="toc-item-num">7 - </span>Network architecture</a></div><div class="lev1"><a href="#Loss-function-and-optimizer"><span class="toc-item-num">8 - </span>Loss function and optimizer</a></div><div class="lev2"><a href="#Loss-function"><span class="toc-item-num">8.1 - </span>Loss function</a></div><div class="lev2"><a href="#Optimizer"><span class="toc-item-num">8.2 - </span>Optimizer</a></div><div class="lev1"><a href="#Experimenting-with-ConvNet-architectures"><span class="toc-item-num">9 - </span>Experimenting with ConvNet architectures</a></div><div class="lev2"><a href="#One-convolutional-layer"><span class="toc-item-num">9.1 - </span>One convolutional layer</a></div><div class="lev1"><a href="#Two-convolutional-layers"><span class="toc-item-num">10 - </span>Two convolutional layers</a></div><div class="lev2"><a href="#Results-for-the-microwave"><span class="toc-item-num">10.1 - </span>Results for the microwave</a></div> This notebook is the report on the convolutional neural network implementation applied to energy disaggregation. All the codes are in the repository: https://github.com/tperol/neuralnilm. We did not transfer all the codes to this final repository in order to keep this clean. The data can also be found in https://github.com/tperol/neuralnilm. Some of the codes (mostly preprocessing of the data, although it was done for an other dataset) has been forked from Jack Kelly github Repository Neuralnilm (https://github.com/JackKelly/neuralnilm). We have implemented here all the neural network architectures tested as well as the preprocessing and the score metrics. The data preprocessing has been slightly edited from Jack Kelly's code to be fed to the Neural Network through a python generator that does the data augmentation on CPU while the GPU is training the network (see details later). End of explanation from __future__ import print_function import nilmtk from nilmtk.utils import print_dict from nilmtk import DataSet # import dataset dataset = DataSet('./redd_data/redd.h5') dataset.set_window("2011-04-01", "2011-05-01") # look at house 2 BUILDING = 2 elec = dataset.buildings[BUILDING].elec # plot mains elec.mains().plot(kind = 'sum'); Explanation: Data preparation We have power data for 6 houses from main meters that measure the aggregate power consumption of the house and from each appliance separately. For example here we show the aggregated data (sum of all appliances in the house) for house 2. End of explanation elec['fridge'].plot(); Explanation: One contribution to this aggregated data is the cycles of the fridge plotted below. End of explanation elec.submeters().appliances Explanation: Of course there are many other appliances that contributes to the aggregated data observed above. The list of appliances listed in this house is shown here: End of explanation corr = elec.correlation_of_sum_of_submeters_with_mains() print print('------------------------------------------------------------------------------') print('the correlation between the sum of the submeters and the main meter is:', corr) print('------------------------------------------------------------------------------') Explanation: The sum of all these appliances is highly correlated to the aggregated data. End of explanation dataset = DataSet('./redd_data/redd.h5') dataset.set_window("2011-04-19", "2011-04-20") dataset.buildings[2].elec.mains().plot(kind= 'sum'); plt.figure(figsize = (16,8)) dataset.buildings[2].elec.submeters().plot(kind = 'sum'); Explanation: We can visualize this looking at the main meters for one day and the sum of all the appliances for one day. End of explanation elec['fridge'].get_activations()[0].plot(); elec['microwave'].get_activations()[0].plot(); Explanation: The correlation is high because the 9 appliances listed above consume a significant proportion of the energy and represent a range of different power 'signatures'. Here is one signature from the fridge and one signature from the microwave. End of explanation dataset.buildings[2].elec.plot_when_on() Explanation: Obviously the microwave signature is a pic of short duration while the fridge cycle is a bit longer. In this study we try to retrieve from the aggregated data the times an appliance switches on, the times it swithes off and the average power comsuption in each cycle. We call this the rectangle method. Essentially, we want to retrieve from the aggregated data, this type of representation: End of explanation # here are the code that perform the preprocessing of the mini batches import nilmtk from nilmtk.utils import print_dict from nilmtk import DataSet from neuralnilm.data.loadactivations import load_nilmtk_activations from neuralnilm.data.syntheticaggregatesource import SyntheticAggregateSource from neuralnilm.data.realaggregatesource import RealAggregateSource from neuralnilm.data.stridesource import StrideSource from neuralnilm.data.datapipeline import DataPipeline from neuralnilm.data.processing import DivideBy, IndependentlyCenter # ------------ # create dictionary with train, unseen_house, unseen_appliance # ------------ def select_windows(train_buildings, unseen_buildings): windows = {fold: {} for fold in DATA_FOLD_NAMES} def copy_window(fold, i): windows[fold][i] = WINDOWS[fold][i] for i in train_buildings: copy_window('train', i) copy_window('unseen_activations_of_seen_appliances', i) for i in unseen_buildings: copy_window('unseen_appliances', i) return windows def filter_activations(windows, activations): new_activations = { fold: {appliance: {} for appliance in APPLIANCES} for fold in DATA_FOLD_NAMES} for fold, appliances in activations.iteritems(): for appliance, buildings in appliances.iteritems(): required_building_ids = windows[fold].keys() required_building_names = [ 'UK-DALE_building_{}'.format(i) for i in required_building_ids] for building_name in required_building_names: try: new_activations[fold][appliance][building_name] = ( activations[fold][appliance][building_name]) except KeyError: pass return activations NILMTK_FILENAME = './redd_data/redd.h5' SAMPLE_PERIOD = 6 STRIDE = None APPLIANCES = ['fridge'] WINDOWS = { 'train': { 1: ("2011-04-19", "2011-05-21"), 2: ("2011-04-19", "2013-05-01"), 3: ("2011-04-19", "2013-05-26"), 6: ("2011-05-22", "2011-06-14"), }, 'unseen_activations_of_seen_appliances': { 1: ("2011-04-19", None), 2: ("2011-04-19", None), 3: ("2011-04-19", None), 6: ("2011-05-22", None), }, 'unseen_appliances': { 5: ("2011-04-19", None) } } # get the dictionary of activations for each appliance activations = load_nilmtk_activations( appliances=APPLIANCES, filename=NILMTK_FILENAME, sample_period=SAMPLE_PERIOD, windows=WINDOWS ) # ------------ # get pipeline for the fridge example # ------------ num_seq_per_batch = 16 target_appliance = 'fridge' seq_length = 512 train_buildings = [1, 2, 3, 6] unseen_buildings = [5] DATA_FOLD_NAMES = ( 'train', 'unseen_appliances', 'unseen_activations_of_seen_appliances') filtered_windows = select_windows(train_buildings, unseen_buildings) filtered_activations = filter_activations(filtered_windows, activations) synthetic_agg_source = SyntheticAggregateSource( activations=filtered_activations, target_appliance=target_appliance, seq_length=seq_length, sample_period=SAMPLE_PERIOD ) real_agg_source = RealAggregateSource( activations=filtered_activations, target_appliance=target_appliance, seq_length=seq_length, filename=NILMTK_FILENAME, windows=filtered_windows, sample_period=SAMPLE_PERIOD ) # ------------ # needed to rescale the input aggregated data # rescaling is done using the a first batch of num_seq_per_batch sequences sample = real_agg_source.get_batch(num_seq_per_batch=1024).next() sample = sample.before_processing input_std = sample.input.flatten().std() target_std = sample.target.flatten().std() # ------------ pipeline = DataPipeline( [synthetic_agg_source, real_agg_source], num_seq_per_batch=num_seq_per_batch, input_processing=[DivideBy(input_std), IndependentlyCenter()], target_processing=[DivideBy(target_std)] ) # ------------ # create the validation set # ------------ num_test_seq = 101 X_valid = np.empty((num_test_seq*num_seq_per_batch, seq_length)) Y_valid = np.empty((num_test_seq*num_seq_per_batch, 3)) for i in range(num_test_seq): (x_valid,y_valid) = pipeline.train_generator(fold = 'unseen_appliances', source_id = 1).next() X_valid[i*num_seq_per_batch: (i+1)*num_seq_per_batch,:] = x_valid[:,:,0] Y_valid[i*num_seq_per_batch: (i+1)*num_seq_per_batch,:] = y_valid X_valid = np.reshape(X_valid, [X_valid.shape[0],X_valid.shape[1],1]) Explanation: Data pipeline Selecting appliances We train each neural network per appliance. This is different from the CO and FHMM methods. Selecting time sequences We down sample the main meters and the submeters to 6 samples per seconds in order to have the aggregated and the submeter sequences properly aligned. We throw away any activation shorter than some threshold duration to avoid spurious spikes. Selecting houses We choose to train the algorithm on house 1,2,3 and 6 and test the data on house 6. Dealing with unbalanced data: selecting aggregated data windows We first exctract using NILMTK libraries the appliance (here the fridge) activations in the time series. We concatenate the times series from house 1,2,3, and 6 for the training set and will test on house 5. We feed to our neural network algorithm (detailed later) balanced mini-batches of data sequences of aggregated data in which the fridge is activated and sequences in which it is not activated. This is a way to deal with unbalanced data -- there are more sequences where the fridge is not activated than sequences with the fridge activated. Most of the data pipeline used is borrowed from https://github.com/JackKelly/neuralnilm. Synthetic aggregated data We use the method from Jack Kelly to create synthetic data. To create a single sequence of synthetic data, we start with two vectors of zeros: one vector will become the input to the net; the other will become the target. The length of each vector defines the ‘window width’ of data that the network sees. We go through five appliance classes and decide whether or not to add an activation of that class to the training sequence. There is a 50% chance that the target appliance will appear in the sequence and a 25% chance for each other ‘distractor’ appliance. For each selected appliance class, we randomly select an appliance activation and then randomly pick where to add that activation on the input vector. Distractor appliances can appear anywhere in the sequence (even if this means that only part of the activation will be included in the sequence). The target appliance activation must be completely contained within the sequence (unless it is too large to fit). We ran neural networks with and without synthetic aggregated data. We found that synthetic data acts as a regulizer, it improves the scores on useen house. All the codes that perform the operation described in this section 2 are in https://github.com/tperol/neuralnilm. End of explanation # needed to rescale the input aggregated data # rescaling is done using the a first batch of num_seq_per_batch sequences sample = real_agg_source.get_batch(num_seq_per_batch=1024).next() sample = sample.before_processing input_std = sample.input.flatten().std() target_std = sample.target.flatten().std() # ------------ Explanation: Standardisation of the input data (aggregated data) A typical step in the data pipeline of neural network is the standardization of data. For each sequences of 512 samples (= 85 seconds) we substract the mean to center the sequence. Furthermore every input sequence is divided by the standard deviation of a random sample in the training set. In this case we cannot divide each sequence by its own standard deviation because it would delete information about the scale of the signal. This is done via: End of explanation def scores(Y_pred, Y_test, activation_threshold = 0.1 ,plot_results= True, print_results = False): a function that computes the classification scores with various metrics return: dictionary with the various scores # post process the data np.putmask(Y_pred[:,0], Y_pred[:,0] <=0, 0) np.putmask(Y_pred[:,1], Y_pred[:,1] >=1, 1) np.putmask(Y_pred[:,0],Y_pred[:,1] < Y_pred[:,0],0) np.putmask(Y_pred[:,1],Y_pred[:,1] < Y_pred[:,0],0) np.putmask(Y_pred[:,1],Y_pred[:,2] < activation_threshold,0) np.putmask(Y_pred[:,0],Y_pred[:,2] < activation_threshold,0) # find negative in prediction pred_negatives = (Y_pred[:,0] ==0) &(Y_pred[:,1] ==0) pred_positives = ~pred_negatives obs_negatives = (Y_test[:,0] ==0) &(Y_test[:,1] ==0) obs_positives = ~obs_negatives TP = obs_positives[pred_positives].sum() FN = obs_positives[pred_negatives].sum() TN = obs_negatives[pred_negatives].sum() FP = obs_negatives[pred_positives].sum() recall = TP / float(TP + FN) precision = TP / float(TP+ FP) f1 = 2* precision*recall / (precision + recall) accuracy = (TP + TN)/ float(obs_negatives.sum() +obs_positives.sum() ) if print_results: print('number of Predicted negatives:',pred_negatives.sum() ) print('number of Predicted positives:',pred_positives.sum() ) print('number of Observed negatives:', obs_negatives.sum() ) print('number of Observed positives:', obs_positives.sum() ) print('f1:', f1) print('precision :' ,precision) print('recall : ', recall) print('accuracy:', accuracy) results = { 'accuracy': accuracy, 'f1_score': f1, 'precision': precision, 'recall_score': recall} if plot_results: pd_results = pd.DataFrame.from_dict(results, orient = 'index') pd_results = pd_results.transpose() sns.barplot(data = pd_results) return results Explanation: Output data (start time, end time and average power) The output of the neural network is 3 neurons: start time, end time and average power. We rescale the time to the interval [0,1]. Therefore if the fridge starts in the middle of the input sequences the output of the first neuron is 0.5. If its stops after the end of the input window the ouput of the second neuron is set to 1. The third neuron is the average power during the activation period. Of course this is set to 0 when it is not activated during the input sequence. We also post process the data by setting any start time lower than 0 to 0 and end time higher than 1 to 1. We create a average power threshold set to 0.1 that indicates if the appliance was active or not (under the threshold the appliance is considered off, above it is considered on). Here we show as an example the input data and the ouput calculated by a trained network. We compare this with the real appliance activation. <img src='./figures/output_example.png'> As we can see here the network does a very good job at detecting the activation of the fridge. The red line is the aggregated data. In the flat region it would be impossible to detect the activation of the fridge with human eye. We would tend to put an activation in the step region. However the network does a very accurate prediction of the activation of the fridge ! Scores - performance evaluation of the neural network Because of the dimension of the ouput we choose classification score metrics. When the starting time and the ending time are both 0 we call this a negative. We also call negative if the power average is lower than threshold. Otherwise this is positive (the appliance is activated). We call TP true positive, TN true negative, FP false positive and FN false negative. The various metrics/scores used in this study are $$ recall = \frac{TP}{TP + FN} $$ $$ precision = \frac{TP}{TP + FP} $$ $$ F1 = 2 * \frac{precision* recall}{precision + recall} $$ $$ accuracy = \frac{TP + TN}{P + N} $$ where P is the number of positives and N the number of negatives. The code to compute can be found in neuralnilm/scores and reproduced partially here. End of explanation def train_generator(self, fold='train', enable_all_appliances=False, source_id=None, reset_iterator=False, validation=False ): class method that yield generator for training the neural network parameters: ---------- fold: 'train' for training or 'unseen_appliance' for testing/validation source_id: if None we use both real and synthetic data with 50:50 ratio if 1 , only use the real data if 0, only use the synthetic data output: --------- generator (X_train,Y_train): a tuple with X_train being the mini-batch for training and Y_train being the mini-batch of labels while 1: batch_iter = self.get_batch(fold, enable_all_appliances, source_id, reset_iterator,validation) X_train = batch_iter.input input_dim = X_train.shape[1] Y_train = self._get_output_neurons(batch_iter) yield (np.reshape(X_train, [self.num_seq_per_batch, input_dim, 1]), Y_train.astype(np.float32)) def _get_output_neurons(self, new_batch): class method that generates the indices needed for the generator Once the network has seen all the sample, the algorithm feed again the mini-batch by reinitialization of the generator (there is no StopIteration error raised) Suffling of the mini-batch is the default strategy implemented batch_size = new_batch.target.shape[0] neural_net_output = np.empty((batch_size, 3)) for b in range(batch_size): seq = new_batch.target[b] # case 1 and 2: if the signal start at 0 if seq[0] > 0: start = 0 stop_array = np.where(seq > 0)[0] # case 2: signal stops after 1 # set stop to the last element if len(stop_array) == 0: stop = seq[-1] # case 1: signal stops before 1 else: stop = stop_array[-1] # calculate avg power avg_power = np.mean(seq[start:stop + 1]) # case 3: signal starts after 0 and before 1 else: start_array = np.where(seq > 0)[0] if len(start_array) == 0: # case 5: there is no signal in the window start = 0 stop = 0 avg_power = 0 else: start = start_array[0] # find stop stop_array = np.where(seq > 0)[0] # case 4: signal stops after 1 # set to the last element if len(stop_array) == 0: stop = seq[-1] else: stop = stop_array[-1] avg_power = np.mean(seq[start:stop + 1]) start = start / float(new_batch.target.shape[1] - 1) stop = stop / float(new_batch.target.shape[1] - 1) if stop < start: raise ValueError("start must be before stop in sequence {}".format(b)) neural_net_output[b, :] = np.array([start, stop, avg_power]) return neural_net_output Explanation: Implementation strategy for real time data augmentation While the neural network runs an NVIDIA GeForce GT 750M (GPU) we maintain the CPU busy doing the data augmentation in real time (load aggregated data, create the synthetic data, preprocess the mini-batch to be fed to the neural network). For this we create a python generator that creates a queue of 50 mini-batch and feed them successively to the GPU for training. The pipeline class can be found in neuralnilm.data.datapiline at https://github.com/tperol/neuralnilm and is partially reproduced here. We do the same to generate the validation and test set. End of explanation w_ahead = w + mu * v # evaluate dw_ahead v = mu * v - learning_rate * dw_ahead x += v Explanation: Network architecture We use a convolutional neural network (ConvNet) to take advantage of the translation invariance. We want the ConvNet to recognize target appliance activation anywhere in the sequence. For this project we have tried multiple architecture that are reported later on. These architecture all have a first convolutional layer of filter size 3 and stride 1. We have played with both the filter size and the number of output filters on the first layer. We have found that 16 filters is a reasonable number -- increasing the number of filters in the first layer did not improve significantly the scores. The best neural network we found consist of Input layer: one channel and lenght of 512 samples 1D convolutional layer (filter size = 3, stride = 1 , number of filters = 16, activation function = relu, border mode = valid, weight initialization = normal distribution) Fully connected layer (N = 1024, activation function = relu, weight initialization = normal distribution) Fully connected layer (N = 512, activation function = relu, weight initialization = normal distribution) Fully connected layer (N= 3, activation function = relu) The ouput has 3 neurons activated by a relu activation function since the output cannot be negative. We have tried other networks that are reported later in this notebook. However this is the layout of the best one we found. Loss function and optimizer Loss function Since the output neurons spans the real axis there is no other choice than using a L2 norm for the loss function. This is (predicted start time - true start time)$^2$ + (predicted end time - true end time)$^2$ + (predicted average power - true average power)$^2$. The total loss function is the sum of the loss function for all the sample in a mini-batch. Optimizer We have tried various optimizer to find the best one. We used a classical Stochastic Gradient Descent to update the weights where we feed one mini-batch chosen randomly to the neural network and then update each weight $$w_j = w_j - \eta \frac{\partial L}{\partial w_j} $$ where $L$ is the loss function evaluated for the given mini-batch. The gradient of the loss function is calculated using the backpropagation algorithm (not detailed here for simplicity). At each epoch we decrease the learning rate $eta$ to allow the algorithm to converge towards a local minimum. We tried a variaton of SGD by using the momentum method . This method has some physical interpretation where $\mu$ is the friction coefficient. In this case the weights are update using $$w_j = w_j + \mu v - \eta \frac{\partial L}{\partial w_j} $$ where $v$ is the velocity. An other tested implementation is the Nesterov momentum in which, at a given position in the landscape of weight we look one step ahead with the momentum and then evaluate the gradient there to calculate the new value of the weight. A pseudo code for this would be End of explanation m = beta1*m + (1-beta1)*dw v = beta2*v + (1-beta2)*(dw**2) w += - learning_rate * m / (np.sqrt(v) + eps) Explanation: We found by experimenting that the best optimizer is Adam (http://arxiv.org/pdf/1412.6980v8.pdf). A pseudo code for Adam is End of explanation # import Keras related libraries from keras.layers import Input, Dense, Flatten, MaxPooling1D, AveragePooling1D, Convolution1D from keras.models import Model import keras.callbacks from keras.callbacks import ModelCheckpoint import time from keras.models import model_from_json import pickle # ------------ exp_number = 13 output_architecture = './tmpdata/convnet_architecture_exp' + str(exp_number) + '.json' best_weights_during_run = './tmpdata/weights_exp' + str(exp_number) + '.h5' final_weights = './tmpdata/weights_exp' + str(exp_number) + '_final.h5' loss_history = './tmpdata/history_exp' + str(exp_number) + '.pickle' # ------------ # ------------ # a class used to record the training and validation loss # at the end of each epoch class LossHistory(keras.callbacks.Callback): def on_train_begin(self, logs={}): self.train_losses = [] self.valid_losses = [] def on_epoch_end(self, epoch, logs = {}): self.train_losses.append(logs.get('loss')) self.valid_losses.append(logs.get('val_loss')) # ------------ # input sequence input_seq = Input(shape = (seq_length, 1)) # first convolutional layer conv1_layer = Convolution1D(nb_filter = 16, filter_length = 3, border_mode='valid', init = 'normal', activation = 'relu') conv1 = conv1_layer(input_seq) # flatten the weights flat = Flatten()(conv1) # first dense layer dense1 = Dense(1024, activation = 'relu')(flat) # second dense layer dense2 = Dense(512, activation = 'relu', init= 'normal')(dense1) # output layer predictions = Dense(3, activation = 'linear')(dense2) # create the model model = Model(input=input_seq, output=predictions) # compile the model -- define the loss and the optimizer model.compile(loss='mean_squared_error', optimizer='Adam') # record the loss history history = LossHistory() # save the weigths when the vlaidation lost decreases only checkpointer = ModelCheckpoint(filepath=best_weights_during_run, save_best_only=True, verbose =1 ) # fit the network using the generator of mini-batches. model.fit_generator(pipeline.train_generator(fold = 'train'), \ samples_per_epoch = 30000, \ nb_epoch = 20, verbose = 1, callbacks=[history, checkpointer], validation_data = (x_valid,y_valid), max_q_size = 50) losses_dic = {'train_loss': history.train_losses, 'valid_loss':history.valid_losses} # save history losses_dic = {'train_loss': history.train_losses, 'valid_loss':history.valid_losses} with open(loss_history, 'wb') as handle: pickle.dump(losses_dic, handle) print('\n saving the architecture of the model \n') json_string = model.to_json() open(output_architecture, 'w').write(json_string) print('\n saving the final weights ... \n') model.save_weights(final_weights, overwrite = True) print('done saving the weights') print('\n saving the training and validation losses') print('This was the model trained') print(model.summary()) Explanation: where dw is the gradient of the lost function with respect to the considered weight and w the considered weight. The hyperparameter for this optimizer are beta1, beta2 and eps. We set beta_1=0.9, beta_2=0.999, epsilon=1e-08 and the learning rate = 1e-3. Experimenting with ConvNet architectures For the implementation of the ConvNet we use Keras (http://keras.io). This is a library implemented on top of Theano and Tensorflow (in this case we use Theano to take advantage of the GPU, GPU training is not yet available on Mac OS using TensorFlow). Here we show the code ran using GPU and Keras to train our best ConvNet. End of explanation from keras.models import model_from_json # load experiments number exp_number = 13 # load the model architecture output_architecture = './tmpdata/convnet_architecture_exp' + str(exp_number) + '.json' # load the weights for the lowest validation loss during training best_weights_during_run = './tmpdata/weights_exp' + str(exp_number) + '.h5' # load the final weights at the end of the 20 epochs final_weights = './tmpdata/weights_exp' + str(exp_number) + '_final.h5' model = model_from_json(open(output_architecture).read()) # load intermediate or final weights model.load_weights(best_weights_during_run) model.compile(loss='mean_squared_error', optimizer='sgd') # print the summary of the architecture model.summary() # load the loss summary (training and validation losses) import pickle losses = pickle.load( open('./tmpdata/history_exp' + str(exp_number) + '.pickle' , 'rb')) # load the test set test_set = pickle.load( open('./tmpdata/TestSet.pickle', 'rb')) X_test = test_set['X_test'] Y_test = test_set["Y_test"] # Here we predict the output from the neural network and show the scores import neuralnilm.scores Y_pred = model.predict(X_test) scores(Y_pred, Y_test) Explanation: One convolutional layer Here we present the results from the ConvNet detailed in the previous cell. We first load the ConvNet architecture, the weights and the history of the training and validation losses (the network has been trained previously using GPU). End of explanation from neuralnilm.scores import plot_loss plot_loss(losses) Explanation: This is a very good F1 score, much better than with any other method tested before. Here we show the evolution of the training loss and the validation loss. End of explanation exp_number = 14 output_architecture = './tmpdata/convnet_architecture_exp' + str(exp_number) + '.json' best_weights_during_run = './tmpdata/weights_exp' + str(exp_number) + '.h5' final_weights = './tmpdata/weights_exp' + str(exp_number) + '_final.h5' model = model_from_json(open(output_architecture).read()) # load intermediate or final weights model.load_weights(best_weights_during_run) model.compile(loss='mean_squared_error', optimizer='sgd') model.summary() # load the loss summary import pickle losses = pickle.load( open('./tmpdata/history_exp' + str(exp_number) + '.pickle' , 'rb')) # load the test set test_set = pickle.load( open('./tmpdata/TestSet.pickle', 'rb')) X_test = test_set['X_test'] Y_test = test_set["Y_test"] Explanation: The training and validation loss during training decreases simulateanously. There is no overfitting. We stop the simulation after 20 epochs but the model was still improving. 20 epochs on GPU took about 1.5 hours. Two convolutional layers Since we are not overfitting we did not add any Pooling layer or Dropout to the previous network. However we experimented an other network with an additional convolutional layer and train the model over 30 epoch. The model is the following: End of explanation import neuralnilm.scores Y_pred = model.predict(X_test) scores(Y_pred, Y_test) from neuralnilm.scores import plot_loss plot_loss(losses) Explanation: There are 2 convolutional layers with 16 filters and 2 dense layers and the output layer. In this case the results are fairly similar to the network presented above. End of explanation from keras.models import model_from_json exp_number = 13 output_architecture = './tmpdata/convnet_architecture_exp' + str(exp_number) + '.json' best_weights_during_run = './tmpdata/weights_exp' + str(exp_number) + '.h5' final_weights = './tmpdata/weights_exp' + str(exp_number) + '_final.h5' model = model_from_json(open(output_architecture).read()) # load intermediate or final weights model.load_weights(best_weights_during_run) model.compile(loss='mean_squared_error', optimizer='sgd') model.summary() # load the loss summary import pickle losses = pickle.load( open('./tmpdata/history_exp' + str(exp_number) + '.pickle' , 'rb')) # load the test set test_set = pickle.load( open('./tmpdata/TestSet_microwave.pickle', 'rb')) X_test_microwave = test_set['X_test'] Y_test_microwave = test_set["Y_test"] from neuralnilm.scores import scores Y_pred_microwave = model.predict(X_test_microwave) scores(Y_pred_microwave, Y_test_microwave) Explanation: Results for the microwave Here we run the ConvNet with one layer on the microwave. We train using house 1 and 2 and test on house 5. End of explanation <END_TASK>
15,972
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Could not find kernel error? If you ever see a "could not find a kernel for this notebook" error message, it will offer you a pull down menu for you to pick a fitting kernel. Remember, kernels are the notebook's way to find the correct interpretor for the code you write into notebook cells. And these days this can be R, Julia, Python and several other things (Find the available kernels list here). Review on your own time Step1: Q. What will be printed? Q. What went wrong? You can force variables to be certain types Step2: Q. What will this produce? Step3: Review continued Step4: It's using the exact same library twice, you just told Python 2 different ways to get to it. And there's even a way to prove it Step5: As you can see it's the same memory address (but this number is not necessarily the same on your computer), meaning the Python interpreter uses the exact same object twice, you just gave it 2 different names. Another syntax is available to import modules Step6: or specific functions within a module Step7: Today Step8: Without a while loop Step9: Recall the Gravitational Force Equation $$F(r) = G \frac{m_1 m_2}{r^2}$$ Step10: Q. What will this loop do ("trace" it)? The increment could have been done in shorthand Step11: Q. What about this one? Can you predict any problems it may cause? Infinite loops Step12: Q. How does this work? INTERLUDE Step13: Q. What is the value of this? Step14: Q. What is the value of this? See how readable Python is? Boolean expressions can be combined with "and", "or" and "not" to form compound conditional expressions. Step15: Q. How about this? Step16: Back to while loops Example - User Input Step17: Q. What happens if you enter a letter instead of a number? Lists (Section 2.2) Step18: Q. What will this print? Step19: Lesson learned Step20: Q. What will the next line produce? Step21: List operations Step22: Q. What will this produce? Step23: Q. What will this give us? Creating lists with while loops Step24: Q. What is massRatio?
<ASSISTANT_TASK:> Python Code: tempF = 212.0 tempC = (5 / 9) * (tempF - 32.0) tempC Explanation: Could not find kernel error? If you ever see a "could not find a kernel for this notebook" error message, it will offer you a pull down menu for you to pick a fitting kernel. Remember, kernels are the notebook's way to find the correct interpretor for the code you write into notebook cells. And these days this can be R, Julia, Python and several other things (Find the available kernels list here). Review on your own time: A few "last" things about types $T_C = \frac{5}{9} \left(T_F - 32 \right)$ End of explanation x = 45 type(x) # Gives (returns) the type of variable x = float(x) print(type(x)) x Explanation: Q. What will be printed? Q. What went wrong? You can force variables to be certain types End of explanation x = 26.9 int(x) Explanation: Q. What will this produce? End of explanation from math import * import math print(sqrt(2)) math.sqrt(2) Explanation: Review continued: One "last" note on modules End of explanation id(sqrt) id(math.sqrt) Explanation: It's using the exact same library twice, you just told Python 2 different ways to get to it. And there's even a way to prove it: With the id() function: End of explanation # Import math module and give it a new name import math as m # Note the use of "as", a reserved word m.sqrt(2) Explanation: As you can see it's the same memory address (but this number is not necessarily the same on your computer), meaning the Python interpreter uses the exact same object twice, you just gave it 2 different names. Another syntax is available to import modules: End of explanation # Import sqrt from math and give it a new name from math import sqrt as sq from math import pi as PIE sq(2) Explanation: or specific functions within a module: End of explanation x = 0 # Initialize the variable x to 0 while(x != 3): # While (as long as) x is not equal to 3 print("The value of x is", x) # Print this to the screen x += 1 # Increment x by 1 (add 1 to x) # REPEAT!!! print(x) # What is the value of x? Explanation: Today: Loops & Lists The point of loops is to compactly code repetitive tasks. For example, computing the gravitational force for multiple planetary masses. Loops are an essential programming tool (this is why we program!). Python supports two types of loops: while loops for loops While Loops (Section 2.1.2 in the book) Basic While Loop <Talk about how Python knows what's in the loop> End of explanation x = 0 # Initialize the variable x to 0 print("The value of x is", x) # Print this to the screen x += 1 # Increment x by 1 (add 1 to x) print("The value of x is", x) # Print this to the screen x += 1 # Increment x by 1 (add 1 to x) print("The value of x is", x) # Print this to the screen x += 1 # Increment x by 1 (add 1 to x) Explanation: Without a while loop End of explanation print('# Table of Gravitational Forces for Multiple Planet Masses\n') # Initialize variables - use meters and kilograms for units G = 6.67e-11 # Gravitational constant mass_earth = 5.97e24 # Earth mass mass_person = 70 # Person mass radius_earth = 6.37e6 # Earth radius # Begin calculation mass1 = mass_earth # Print a header print('# mass1/mass_earth Force') # The loop ends when conditional mass1 <= (10.0 * massEarth) is no longer true while(mass1 <= (10.0 * mass_earth)): # Note the colon! force = G * mass1 * mass_person / radius_earth**2 # All lines in the loop must be indented by # the same amount (iPython does it automatically) # print(str(mass1 / mass_earth) + " " + str(force)) print("{mass_ratio}\t{force:7.2f}".format(mass_ratio=mass1 / mass_earth, force=force)) mass1 = mass1 + mass_earth # Increment by Earth's mass # No indent! This line is executed after the loop is done print('# Done') Explanation: Recall the Gravitational Force Equation $$F(r) = G \frac{m_1 m_2}{r^2}$$ End of explanation # Note that I have to reset mass1 here!! mass1 = mass_earth print('# mass1/mass_earth Force') while(mass1 <= (10.0 * mass_earth)): force = G * mass1 * mass_person / radius_earth**2 print("{:18.1f} {:7.2f}".format(mass1 / mass_earth, force)) # mass1 = mass1 + mass_earth mass1 += mass_earth # Shorthand version of the line above. '# Done' Explanation: Q. What will this loop do ("trace" it)? The increment could have been done in shorthand End of explanation # How to prevent an infinite loop maxCount = 10 # A number that is more than your loop should ever do count = 0 # The current number your loop is on # Adding "and < maxCount" to the end of your conditional prevents infinite loops while(True and count < maxCount): print("Loop count: " + str(count)) count += 1 # Increment your current loop count Explanation: Q. What about this one? Can you predict any problems it may cause? Infinite loops End of explanation 5 <= 10 Explanation: Q. How does this work? INTERLUDE: Boolean (logic) expressions (Section 2.1.3) Boolean expressions are conditional statements. There are only two possible values: True or False I've capitalized True and False because these are reserved words in Python. Q. What is the value of this? End of explanation 5 >= 10 not 5 >= 10 Explanation: Q. What is the value of this? End of explanation 5 <= 10 and 5 >= 10 Explanation: Q. What is the value of this? See how readable Python is? Boolean expressions can be combined with "and", "or" and "not" to form compound conditional expressions. End of explanation 5 <= 10 or 5 >= 10 Explanation: Q. How about this? End of explanation import random minNumber = 1 maxNumber = 10 # Get a random number between 1 and 10 randomNumber = random.randint(minNumber, maxNumber) userGuess = -1 while(userGuess != randomNumber): userPrompt = "Guess a number between " + str(minNumber) + " and " + str(maxNumber) + ": " userGuess = input(userPrompt) # Prompt the user userGuess = int(userGuess) print("You have guessed the correct number! " + str(userGuess)) Explanation: Back to while loops Example - User Input End of explanation massRatio = [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0] massRatio massRatio[3] Explanation: Q. What happens if you enter a letter instead of a number? Lists (Section 2.2) End of explanation type(massRatio[3]) Explanation: Q. What will this print? End of explanation massRatio.append(11.0) massRatio # This inserts 4.5 into index 4 of the list: massRatio.insert(4, 4.5) massRatio del massRatio[4] Explanation: Lesson learned: Python is zero-index based Modifying lists End of explanation massRatio Explanation: Q. What will the next line produce? End of explanation # We can find out its length with len(object) len(massRatio) # Python uses [] to access elements and () to perform a function on an object. massRatio = massRatio + [12.0, 13.0, 14.0] massRatio massRatio.extend([15.0, 16.0, 17.0]) print("Extend", massRatio) massRatio.append([18.0, 19.0, 20.0]) print("Append", massRatio) print(massRatio[17][1]) Explanation: List operations End of explanation massRatio.index(12.0) # And, this fails massRatio.index(20.0) # We can check if there is an element in a list. The result of the check # is boolean: True or False. 14.0 in massRatio 99.0 in massRatio massRatio # Negative indices start counting from the right (the end) of a list: massRatio[-4] Explanation: Q. What will this produce? End of explanation # Initializations first massRatio = [] # Creates an empty list massRatioValue = 1.0 # For the conditional massRatioMax = 5.0 # Also for the conditional userInput = "BIG NOPE" # And the while loop while(userInput != "N" and massRatioValue <= massRatioMax): # Remember the colon! # Remember to indent! massRatio.append(massRatioValue) massRatioValue += 1.0 userInput = input("Add another mass ratio value? ") userInput = userInput.upper() print("Finished creating the list massRatio!") Explanation: Q. What will this give us? Creating lists with while loops End of explanation massRatio Explanation: Q. What is massRatio? End of explanation <END_TASK>
15,973
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Transform EEG data using current source density (CSD) This script shows an example of how to use CSD [1] [2] [3]_. CSD takes the spatial Laplacian of the sensor signal (derivative in both x and y). It does what a planar gradiometer does in MEG. Computing these spatial derivatives reduces point spread. CSD transformed data have a sharper or more distinct topography, reducing the negative impact of volume conduction. Step1: Load sample subject data Step2: Plot the raw data and CSD-transformed raw data Step3: Also look at the power spectral densities Step4: CSD can also be computed on Evoked (averaged) data. Here we epoch and average the data so we can demonstrate that. Step5: First let's look at how CSD affects scalp topography Step6: CSD has parameters stiffness and lambda2 affecting smoothing and spline flexibility, respectively. Let's see how they affect the solution
<ASSISTANT_TASK:> Python Code: # Authors: Alex Rockhill <aprockhill206@gmail.com> # # License: BSD (3-clause) import numpy as np import matplotlib.pyplot as plt import mne from mne.datasets import sample print(__doc__) data_path = sample.data_path() Explanation: Transform EEG data using current source density (CSD) This script shows an example of how to use CSD [1] [2] [3]_. CSD takes the spatial Laplacian of the sensor signal (derivative in both x and y). It does what a planar gradiometer does in MEG. Computing these spatial derivatives reduces point spread. CSD transformed data have a sharper or more distinct topography, reducing the negative impact of volume conduction. End of explanation raw = mne.io.read_raw_fif(data_path + '/MEG/sample/sample_audvis_raw.fif', preload=True) events = mne.find_events(raw) raw = raw.pick_types(meg=False, eeg=True, eog=True, ecg=True, stim=False, exclude=raw.info['bads']) raw.set_eeg_reference(projection=True).apply_proj() Explanation: Load sample subject data End of explanation raw_csd = mne.preprocessing.compute_current_source_density(raw) raw.plot() raw_csd.plot() Explanation: Plot the raw data and CSD-transformed raw data: End of explanation raw.plot_psd() raw_csd.plot_psd() Explanation: Also look at the power spectral densities: End of explanation event_id = {'auditory/left': 1, 'auditory/right': 2, 'visual/left': 3, 'visual/right': 4, 'smiley': 5, 'button': 32} epochs = mne.Epochs(raw, events, event_id=event_id, tmin=-0.2, tmax=.5, preload=True) evoked = epochs['auditory'].average() Explanation: CSD can also be computed on Evoked (averaged) data. Here we epoch and average the data so we can demonstrate that. End of explanation times = np.array([-0.1, 0., 0.05, 0.1, 0.15]) evoked_csd = mne.preprocessing.compute_current_source_density(evoked) evoked.plot_joint(title='Average Reference', show=False) evoked_csd.plot_joint(title='Current Source Density') Explanation: First let's look at how CSD affects scalp topography: End of explanation fig, ax = plt.subplots(4, 4) fig.subplots_adjust(hspace=0.5) fig.set_size_inches(10, 10) for i, lambda2 in enumerate([0, 1e-7, 1e-5, 1e-3]): for j, m in enumerate([5, 4, 3, 2]): this_evoked_csd = mne.preprocessing.compute_current_source_density( evoked, stiffness=m, lambda2=lambda2) this_evoked_csd.plot_topomap( 0.1, axes=ax[i, j], outlines='skirt', contours=4, time_unit='s', colorbar=False, show=False) ax[i, j].set_title('stiffness=%i\nλ²=%s' % (m, lambda2)) Explanation: CSD has parameters stiffness and lambda2 affecting smoothing and spline flexibility, respectively. Let's see how they affect the solution: End of explanation <END_TASK>
15,974
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Unsupervised learning Step1: First, we start with some exploratory clustering, visualizing the clustering dendrogram using SciPy's linkage and dendrogram functions Step2: Next, let's use the AgglomerativeClustering estimator from scikit-learn and divide the dataset into 3 clusters. Can you guess which 3 clusters from the dendrogram it will reproduce? Step3: Density-based Clustering - DBSCAN Another useful approach to clustering is Density-based Spatial Clustering of Applications with Noise (DBSCAN). In essence, we can think of DBSCAN as an algorithm that divides the dataset into subgroup based on dense regions of point. In DBSCAN, we distinguish between 3 different "points" Step4: Exercise <div class="alert alert-success"> <b>EXERCISE</b>
<ASSISTANT_TASK:> Python Code: from sklearn.datasets import load_iris iris = load_iris() X = iris.data[:, [2, 3]] y = iris.target n_samples, n_features = X.shape plt.scatter(X[:, 0], X[:, 1], c=y); Explanation: Unsupervised learning: Hierarchical and density-based clustering algorithms In a previous notebook, "08 Unsupervised Learning - Clustering.ipynb", we introduced one of the essential and widely used clustering algorithms, K-means. One of the advantages of K-means is that it is extremely easy to implement, and it is also computationally very efficient compared to other clustering algorithms. However, we've seen that one of the weaknesses of K-Means is that it only works well if the data can be grouped into a globular or spherical shape. Also, we have to assign the number of clusters, k, a priori -- this can be a problem if we have no prior knowledge about how many clusters we expect to find. In this notebook, we will take a look at 2 alternative approaches to clustering, hierarchical clustering and density-based clustering. Hierarchical Clustering One nice feature of hierachical clustering is that we can visualize the results as a dendrogram, a hierachical tree. Using the visualization, we can then decide how "deep" we want to cluster the dataset by setting a "depth" threshold. Or in other words, we don't need to make a decision about the number of clusters upfront. Agglomerative and divisive hierarchical clustering Furthermore, we can distinguish between 2 main approaches to hierarchical clustering: Divisive clustering and agglomerative clustering. In agglomerative clustering, we start with a single sample from our dataset and iteratively merge it with other samples to form clusters -- we can see it as a bottom-up approach for building the clustering dendrogram. In divisive clustering, however, we start with the whole dataset as one cluster, and we iteratively split it into smaller subclusters -- a top-down approach. In this notebook, we will use agglomerative clustering. Single and complete linkage Now, the next question is how we measure the similarity between samples. One approach is the familiar Euclidean distance metric that we already used via the K-Means algorithm. As a refresher, the distance between 2 m-dimensional vectors $\mathbf{p}$ and $\mathbf{q}$ can be computed as: \begin{align} \mathrm{d}(\mathbf{q},\mathbf{p}) & = \sqrt{(q_1-p_1)^2 + (q_2-p_2)^2 + \cdots + (q_m-p_m)^2} \[8pt] & = \sqrt{\sum_{j=1}^m (q_j-p_j)^2}.\end{align} However, that's the distance between 2 samples. Now, how do we compute the similarity between subclusters of samples in order to decide which clusters to merge when constructing the dendrogram? I.e., our goal is to iteratively merge the most similar pairs of clusters until only one big cluster remains. There are many different approaches to this, for example single and complete linkage. In single linkage, we take the pair of the most similar samples (based on the Euclidean distance, for example) in each cluster, and merge the two clusters which have the most similar 2 members into one new, bigger cluster. In complete linkage, we compare the pairs of the two most dissimilar members of each cluster with each other, and we merge the 2 clusters where the distance between its 2 most dissimilar members is smallest. To see the agglomerative, hierarchical clustering approach in action, let us load the familiar Iris dataset -- pretending we don't know the true class labels and want to find out how many different follow species it consists of: End of explanation from scipy.cluster.hierarchy import linkage from scipy.cluster.hierarchy import dendrogram clusters = linkage(X, metric='euclidean', method='complete') dendr = dendrogram(clusters) plt.ylabel('Euclidean Distance'); Explanation: First, we start with some exploratory clustering, visualizing the clustering dendrogram using SciPy's linkage and dendrogram functions: End of explanation from sklearn.cluster import AgglomerativeClustering ac = AgglomerativeClustering(n_clusters=3, affinity='euclidean', linkage='complete') prediction = ac.fit_predict(X) print('Cluster labels: %s\n' % prediction) plt.scatter(X[:, 0], X[:, 1], c=prediction); Explanation: Next, let's use the AgglomerativeClustering estimator from scikit-learn and divide the dataset into 3 clusters. Can you guess which 3 clusters from the dendrogram it will reproduce? End of explanation from sklearn.datasets import make_moons X, y = make_moons(n_samples=400, noise=0.1, random_state=1) plt.scatter(X[:,0], X[:,1]) plt.show() from sklearn.cluster import DBSCAN db = DBSCAN(eps=0.2, min_samples=10, metric='euclidean') prediction = db.fit_predict(X) print("Predicted labels:\n", prediction) plt.scatter(X[:, 0], X[:, 1], c=prediction); Explanation: Density-based Clustering - DBSCAN Another useful approach to clustering is Density-based Spatial Clustering of Applications with Noise (DBSCAN). In essence, we can think of DBSCAN as an algorithm that divides the dataset into subgroup based on dense regions of point. In DBSCAN, we distinguish between 3 different "points": Core points: A core point is a point that has at least a minimum number of other points (MinPts) in its radius epsilon. Border points: A border point is a point that is not a core point, since it doesn't have enough MinPts in its neighborhood, but lies within the radius epsilon of a core point. Noise points: All other points that are neither core points nor border points. A nice feature about DBSCAN is that we don't have to specify a number of clusters upfront. However, it requires the setting of additional hyperparameters such as the value for MinPts and the radius epsilon. End of explanation from sklearn.datasets import make_circles X, y = make_circles(n_samples=1500, factor=.4, noise=.05) plt.scatter(X[:, 0], X[:, 1], c=y); # %load solutions/20_clustering_comparison.py Explanation: Exercise <div class="alert alert-success"> <b>EXERCISE</b>: <ul> <li> Using the following toy dataset, two concentric circles, experiment with the three different clustering algorithms that we used so far: `KMeans`, `AgglomerativeClustering`, and `DBSCAN`. Which clustering algorithms reproduces or discovers the hidden structure (pretending we don't know `y`) best? Can you explain why this particular algorithm is a good choice while the other 2 "fail"? </li> </ul> </div> End of explanation <END_TASK>
15,975
<SYSTEM_TASK:> Given the following text description, write Python code to implement the functionality described below step by step <END_TASK> <USER_TASK:> Description: Estimating At-Work Subtour Frequency This notebook illustrates how to re-estimate a single model component for ActivitySim. This process includes running ActivitySim in estimation mode to read household travel survey files and write out the estimation data bundles used in this notebook. To review how to do so, please visit the other notebooks in this directory. Load libraries Step1: We'll work in our test directory, where ActivitySim has saved the estimation data bundles. Step2: Load data and prep model for estimation Step3: Review data loaded from the EDB The next step is to read the EDB, including the coefficients, model settings, utilities specification, and chooser and alternative data. Coefficients Step4: Utility specification Step5: Chooser data Step6: Estimate With the model setup for estimation, the next step is to estimate the model coefficients. Make sure to use a sufficiently large enough household sample and set of zones to avoid an over-specified model, which does not have a numerically stable likelihood maximizing solution. Larch has a built-in estimation methods including BHHH, and also offers access to more advanced general purpose non-linear optimizers in the scipy package, including SLSQP, which allows for bounds and constraints on parameters. BHHH is the default and typically runs faster, but does not follow constraints on parameters. Step7: Estimated coefficients Step8: Output Estimation Results Step9: Write the model estimation report, including coefficient t-statistic and log likelihood Step10: Next Steps The final step is to either manually or automatically copy the *_coefficients_revised.csv file to the configs folder, rename it to *_coefficients.csv, and run ActivitySim in simulation mode.
<ASSISTANT_TASK:> Python Code: import os import larch # !conda install larch -c conda-forge # for estimation import pandas as pd Explanation: Estimating At-Work Subtour Frequency This notebook illustrates how to re-estimate a single model component for ActivitySim. This process includes running ActivitySim in estimation mode to read household travel survey files and write out the estimation data bundles used in this notebook. To review how to do so, please visit the other notebooks in this directory. Load libraries End of explanation os.chdir('test') Explanation: We'll work in our test directory, where ActivitySim has saved the estimation data bundles. End of explanation modelname = "atwork_subtour_frequency" from activitysim.estimation.larch import component_model model, data = component_model(modelname, return_data=True) Explanation: Load data and prep model for estimation End of explanation data.coefficients Explanation: Review data loaded from the EDB The next step is to read the EDB, including the coefficients, model settings, utilities specification, and chooser and alternative data. Coefficients End of explanation data.spec Explanation: Utility specification End of explanation data.chooser_data Explanation: Chooser data End of explanation model.estimate(method='SLSQP') Explanation: Estimate With the model setup for estimation, the next step is to estimate the model coefficients. Make sure to use a sufficiently large enough household sample and set of zones to avoid an over-specified model, which does not have a numerically stable likelihood maximizing solution. Larch has a built-in estimation methods including BHHH, and also offers access to more advanced general purpose non-linear optimizers in the scipy package, including SLSQP, which allows for bounds and constraints on parameters. BHHH is the default and typically runs faster, but does not follow constraints on parameters. End of explanation model.parameter_summary() Explanation: Estimated coefficients End of explanation from activitysim.estimation.larch import update_coefficients result_dir = data.edb_directory/"estimated" update_coefficients( model, data, result_dir, output_file=f"{modelname}_coefficients_revised.csv", ); Explanation: Output Estimation Results End of explanation model.to_xlsx( result_dir/f"{modelname}_model_estimation.xlsx", data_statistics=False, ) Explanation: Write the model estimation report, including coefficient t-statistic and log likelihood End of explanation pd.read_csv(result_dir/f"{modelname}_coefficients_revised.csv") Explanation: Next Steps The final step is to either manually or automatically copy the *_coefficients_revised.csv file to the configs folder, rename it to *_coefficients.csv, and run ActivitySim in simulation mode. End of explanation <END_TASK>