from statistics import mode
from cProfile import label
from joblib import load

import matplotlib.pyplot as plt
import gradio as gr
import numpy as np

def getdata(Age,Sex,CP,Trtbps,Chol,Fbs,Restecg,Thalachh,Oldpeak,Slp,Caa,Thall,Exng):

    if Sex == "Male":
        Sex = 1
    else: 
        Sex = 0

    if CP == "Typical Angina":
        CP = 0
    elif CP == "Atypical Angina":
        CP = 1
    elif CP == "Non-anginal Pain":
        CP = 2
    else: 
        CP = 3

    if Fbs == "True":
        Fbs = 1
    else: 
        Fbs = 0

    if Restecg == "Normal":
        Restecg = 0
    elif Restecg == "ST-T wave normality":
        Restecg = 1
    else: 
        Restecg = 2

    if Exng == "Yes":
        Exng = 1
    else: 
        Exng = 0

    a = [Age,Sex,CP,Trtbps,Chol,Fbs,Restecg,Thalachh,Exng,Oldpeak,Slp,Caa,Thall]
    arr = np.array([a])

    return arr

def getfig(X_test):
    X_pca = load('X_pca.data')
    y = load('y.data')

    pca = load('pca.dim')
    u_pca  = pca.transform(X_test)

    fig = plt.figure(figsize=(5,4))

    plt.scatter(X_pca[:, 0], X_pca[:, 1], c = y, cmap = plt.cm.Spectral, s =  10, label='More Chance AH')
    plt.scatter(u_pca[:, 0], u_pca[:, 1], c = 'g', cmap = plt.cm.Spectral, s =  40)
    plt.annotate('You', (u_pca[:, 0]+10, u_pca[:, 1]+5), bbox = dict(boxstyle="round",ec='g',fc='lightgreen'))
    plt.title(f"PCA, Exp. Variance: {np.round(np.sum(pca.explained_variance_ratio_), 4)}")
    plt.legend()
    plt.xlabel("PC 1")
    plt.ylabel("PC 2")

    return fig

def greet(Age,Sex,CP,Trtbps,Chol,Fbs,Restecg,Thalachh,Oldpeak,Slp,Caa,Thall,Exng):
    
    X_test = getdata(Age,Sex,CP,Trtbps,Chol,Fbs,Restecg,Thalachh,Oldpeak,Slp,Caa,Thall,Exng)
    
    scaler = load('stdscaler.model') 
    x_std = scaler.transform(X_test)

    log_reg = load('log_reg.model') 
    y_lr  = log_reg.predict(x_std)
    
    kmeans = load('kmeans.model') 
    y_km  = kmeans.predict(x_std)

    tree = load('tree.model') 
    y_tree  = tree.predict(x_std)

    nb = load('nb.model') 
    y_bayes  = nb.predict(X_test)

    forest = load('forest.model') 
    y_forest  = forest.predict(X_test)

    r = [y_lr[0], y_km[0], y_tree[0], y_bayes[0], y_forest[0]]

    f = mode(r)

    if f == 0:
        x = "You have less chance of heart attack"
    else:
        x = "You have more chance of heart attack"

    fig = load('dime.fig')

    fig2 = getfig(X_test)

    return x, fig, fig2


interface = gr.Interface(
    title = "HeartAttack prediction - UMG <br> Project Coeur ❤",
    description = "<h3>The idea is to classify between 0 = less chance of heart attack and 1 = more chance of heart attack, according to the data provided by the user.</h3>"+
    "<b>Models:</b> Logistic Regression, K-means, Decision Trees, Naive Bayes and Random Forest"+
    "<br><b>Metrics:</b> Accuracy: 0.82, Precision: 0.775, Recall: 0.939, F1 Score: 0.849 <br>  <br><b>Please provide the requested data:</b>",
    article='Step-by-step on GitHub <a href="https://github.com/Adrian8aS/Machine-Learning-App-Gradio/blob/21246d9ba87859e9068369b89d48b4c6ee13dfe5/Proyecto%20integrador.ipynb"> notebook </a> <br> ~ Project Coeur',
    allow_flagging = "never",
    fn = greet, 
    inputs = [
        gr.Number(label="Age of the patient"),
        gr.Radio(["Male", "Female"], label="Sex of the patient"),
        gr.Radio(["Typical Angina", "Atypical Angina", "Non-anginal Pain", "Asymptomatic"], label="Chest pain type"),
        gr.Number(label="Resting blood pressure (in mm Hg)"),
        gr.Number(label="Cholestoral in mg/dl fetched via BMI sensor"),
        gr.Radio(["True", "False"], label="Fasting blood sugar > 120 mg/dl"),
        gr.Radio(["Normal", "ST-T wave normality", "Left ventricular hypertrophy"], label="Resting electrocardiographic results"),
        gr.Number(label="Maximum heart rate achieved"),
        gr.Number(label="Previous peak"),
        gr.Radio([0, 1, 2], label="Slope"),
        gr.Radio([0, 1, 2, 3, 4], label="Number of major vessels"),
        gr.Radio([0, 1, 2, 3], label="Thalium Stress Test result"),
        gr.Radio(["Yes", "No"], label="Exercise induced angina")
    ],
    outputs = [gr.Text(label="Prediction"), 'plot', 'plot'],
    examples = [[41,"Female","Typical Angina",130,204,"False","Normal",172,1.4,2,0,2,"No"],
                [45,"Male","Non-anginal Pain",110,264,"False","ST-T wave normality",132,0.2,1,0,3,"No"]]
)

interface.launch(share = False)