import gradio as gr
import numpy as np
import pandas as pd
import pickle
from sklearn.preprocessing import MinMaxScaler
from sklearn.linear_model import LinearRegression
from sklearn.model_selection import train_test_split

# Classification Model Imports
with open('decisionTree.pkl', 'rb') as file:
    loadedDecisionTree = pickle.load(file)
with open('linearReg.pkl', 'rb') as file:
    loadedLinearReg = pickle.load(file)
with open('mlpRegression.pkl', 'rb') as file:
    loadedmlpRegression = pickle.load(file)

# Regression Model Imports
with open('randForest.pkl', 'rb') as file:
    loadedForest = pickle.load(file)
with open('knn.pkl', 'rb') as file:
    loadedknn = pickle.load(file)
with open('mlpClassification.pkl', 'rb') as file:
    loadedmlpClassification = pickle.load(file)

races = ['White', 'Black', 'Chinese', 'Other', 'Korean', 'Hispanic', 'Asian', 'Intermediate', 'Malay', 'Indian']

modelChoice = gr.inputs.Dropdown(choices=["Decision Tree Regression", "Linear Regression", "MLP Neural Network Regression","Random Forest Classification", "K Nearest Neighbors Classification", "MLP Neural Network Classification"], label="Choose your Model")

# Define the input sliders for continuous features
age = gr.inputs.Slider(minimum=20, maximum=100, step=1, label="Age")
height = gr.inputs.Slider(minimum=140, maximum=200, step=1, label="Height (cm)")
weight = gr.inputs.Slider(minimum=30, maximum=150, step=1, label="Weight (kg)")
inr = gr.inputs.Slider(minimum=0, maximum=4, step=0.01, label="INR on Reported Therapeutic Dose of Warfarin")

# Define the input dropdowns for categorical features
gender = gr.inputs.Dropdown(choices=["Male", "Female"], label="Gender")
race = gr.inputs.Dropdown(choices=races, label="Race")
diabetes = gr.inputs.Dropdown(choices=["No", "Yes"], label="Diabetes")
simvastatin = gr.inputs.Dropdown(choices=["No", "Yes"], label="Simvastatin (Zocor)")
amiodarone = gr.inputs.Dropdown(choices=["No", "Yes"], label="Amiodarone (Cordarone)")
vkorc1 = gr.inputs.Dropdown(choices=["A/A", "A/G", "G/G"], label="VKORC1 genotype")

# Define the output textbox
therapeutic_dose = gr.outputs.Textbox(label="If Classification output = Yes or No (Above 50 mg/week) If Regression output = recomended dose (mg/week)")

# Define the function to make predictions
def predict_therapeutic_dose(modelChoice, gender, race, age, height, weight, diabetes, simvastatin, amiodarone, inr, vkorc1):
    # Converting values to values that the models understand
    gender = {"Male": 1, "Female": 0}[gender]
    race = races.index(race)
    diabetes = {"No": 0, "Yes": 1}[diabetes]
    simvastatin = {"No": 0, "Yes": 1}[simvastatin]
    amiodarone = {"No": 0, "Yes": 1}[amiodarone]
    vkorc1 = {"A/A": 1, "A/G": 2, "G/G": 3}[vkorc1]

    if modelChoice == "Decision Tree Regression":
        predicted_dose = loadedDecisionTree.predict(np.array([[gender, race, age, height, weight, diabetes, simvastatin, amiodarone, inr, vkorc1]]))
    elif modelChoice == "Linear Regression":
        predicted_dose = loadedLinearReg.predict(np.array([[gender, race, age, height, weight, diabetes, simvastatin, amiodarone, inr, vkorc1]]))
    elif modelChoice == "MLP Neural Network Regression":
        predicted_dose = loadedmlpRegression.predict(np.array([[gender, race, age, height, weight, diabetes, simvastatin, amiodarone, inr, vkorc1]]))
    elif modelChoice == "Random Forest Classification":
        predicted_dose = loadedForest.predict(np.array([[gender, race, age, height, weight, diabetes, simvastatin, amiodarone, inr, vkorc1]]))
    elif modelChoice == "K Nearest Neighbors Classification":
        predicted_dose = loadedknn.predict(np.array([[gender, race, age, height, weight, diabetes, simvastatin, amiodarone, inr, vkorc1]]))
    else:
        predicted_dose = loadedmlpClassification.predict(np.array([[gender, race, age, height, weight, diabetes, simvastatin, amiodarone, inr, vkorc1]]))        

    if predicted_dose[0] == 0:
      predicted_dose = "No"
    elif predicted_dose[0] == 1:
      predicted_dose = "Yes"

    return predicted_dose

# Create the Gradio interface
gr.Interface( fn=predict_therapeutic_dose, inputs=[modelChoice, gender, race, age, height, weight, diabetes, simvastatin, amiodarone, inr, vkorc1], outputs=therapeutic_dose ).launch(debug=True)