Spaces:
Sleeping
Sleeping
| from matplotlib import pyplot as plt | |
| from utils import * | |
| import warnings | |
| from sklearn.kernel_ridge import KernelRidge | |
| from sklearn.linear_model import LinearRegression | |
| from sklearn.neighbors import KDTree | |
| from sklearn.neighbors import NearestNeighbors | |
| #### Define outputs #### | |
| write_file = True | |
| plot_profile = True | |
| #### Define outputs #### | |
| # load data | |
| pre = "/plp_user/agar_sh/PBML/pytorch/TPH/MLP/profiles/" | |
| with open(pre + 'x_p.pkl', 'rb') as file: | |
| x_p = pickle.load(file) | |
| with open(pre + 'y_p.pkl', 'rb') as file: | |
| y_p = pickle.load(file) | |
| # calculate baselines | |
| linear_reg = LinearRegression().fit(x_p["train"], y_p["train"]) | |
| krr = KernelRidge(alpha=0.01, kernel="rbf") | |
| krr.fit(x_p["train"], y_p["train"]) | |
| tree = KDTree(x_p["train"], leaf_size=2) | |
| nbrs = NearestNeighbors(n_neighbors=1, algorithm='ball_tree').fit(x_p["train"]) | |
| # read the simulation parameters and ensure correct foramtting | |
| f_nn = "my_simulation_parameters.txt" | |
| with open(f_nn) as fw: | |
| lines = fw.readlines() | |
| for line in lines: | |
| l = line.rstrip() | |
| if "r_list" in l: | |
| if not l[-1] == ",": | |
| raise Exception("Ensure there is a comma after last parameter value in " + f_nn) | |
| r_list = [float(p) for p in l.split("=")[1].split(",")[:-1]] | |
| elif "t_list" in line: | |
| if not l[-1] == ",": | |
| raise Exception("Ensure there is a comma after last parameter value in " + f_nn) | |
| t_list = [float(p) for p in l.split("=")[1].split(",")[:-1]] | |
| elif "v_list" in line: | |
| if not l[-1] == ",": | |
| raise Exception("Ensure there is a comma after last parameter value in " + f_nn) | |
| v_list = [float(p) for p in l.split("=")[1].split(",")[:-1]] | |
| if not len(r_list) == len(v_list) and len(r_list) == len(t_list): | |
| raise Exception("Ensure equal number of values for all parameters in " + f_nn) | |
| # Check parameter ranges and print appropriate warnings | |
| for i in range(len(r_list)): | |
| if r_list[i]<0 or r_list[i]>9.5: | |
| warnings.warn('RaQ/Ra is outside the range of the training dataset') | |
| if t_list[i]<1e+6 or t_list[i]>5e+9: | |
| warnings.warn('FKT is outside the range of the training dataset') | |
| if v_list[i]<1 or v_list[i]>95: | |
| warnings.warn('FKV is outside the range of the training dataset') | |
| ### calculates y points ### | |
| num_points = 128 | |
| y_prof = np.concatenate(([1], np.linspace(0+1/(num_points*2),1-1/(num_points*2),num_points-2)[::-1], [0])) | |
| ### calculates y points ### | |
| ### calculates temperature profile ### | |
| x_in = get_input(r_list, t_list, v_list) | |
| y_pred = {} | |
| y_pred["linear"] = linear_reg.predict(x_in) | |
| y_pred["krr"] = krr.predict(x_in) | |
| distances, indices = nbrs.kneighbors(x_in) | |
| y_pred["neighbor"] = y_p["train"][indices,:].reshape(x_in.shape[0],128) | |
| dist, ind = tree.query(x_in, k=3) | |
| y_pred["interp"] = np.zeros_like(y_pred["krr"]) | |
| for i in range(y_pred["interp"].shape[0]): | |
| weights = 1.0/dist[i,:] | |
| weights /= weights.sum(axis=0) | |
| y_pred["interp"][i,:] = y_p["train"][ind[i,0],:]*weights[0] \ | |
| + y_p["train"][ind[i,1],:]*weights[1] \ | |
| + y_p["train"][ind[i,2],:]*weights[2] | |
| ### calculates temperature profile ### | |
| ### writes out temperature profile ### | |
| if write_file: | |
| for algorithm in ["linear", "krr", "interp", "neighbor"]: | |
| for i in range(len(r_list)): | |
| fname = "outputs/" + algorithm + "_profile_raq_ra" + str(r_list[i]) + "_fkt" + str(t_list[i]) + "_fkv" + str(v_list[i]) | |
| f = open(fname + ".txt", "wb") | |
| for j in range(len(y_prof)): | |
| f.writelines([str(y_prof[j]).encode('ascii'), | |
| " ".encode('ascii'), | |
| str(y_pred[algorithm][i,j]).encode('ascii'), | |
| "\n".encode('ascii')]) | |
| f.close() | |
| ### writes out temperature profile ### | |
| ### plots temperature profile ### | |
| for algorithm in ["linear", "krr", "interp", "neighbor"]: | |
| for i in range(len(r_list)): | |
| fname = "outputs/" + algorithm + "_profile_raq_ra" + str(r_list[i]) + "_fkt" + str(t_list[i]) + "_fkv" + str(v_list[i]) | |
| plt.figure() | |
| plt.plot(y_pred[algorithm][i,:], y_prof, 'k-', linewidth=3.0, label="pointwise neural network") | |
| plt.ylim([1,0]) | |
| plt.xlabel("Temperature") | |
| plt.ylabel("Depth") | |
| plt.legend() | |
| plt.grid() | |
| plt.savefig(fname + ".png") | |
| ### plots temperature profile ### |