app.py

import gradio as gr
from joblib import load
import pandas as pd
import numpy as np
from sklearn.preprocessing import StandardScaler

# Load the imputed features and target variable from the CSV file
df_imputed_features = pd.read_csv('imputed_data.csv')

# Separate features and target variable
X = df_imputed_features.drop(columns=['Therapeutic Dose of Warfarin'])
Y = df_imputed_features['Therapeutic Dose of Warfarin']

# Load the scaler
scaler = StandardScaler()
X_scaled = scaler.fit_transform(X)

# Define the mappings
age_dict = {"10 - 19": 0, "20 - 29": 1, "30 - 39": 2, "40 - 49": 3, "50 - 59": 4, "60 - 69": 5, "70 - 79": 6, "80 - 89": 7, "90+": 8}
race_dict = {"Caucasian": 0, "Chinese": 1, "Indian": 2, "Japanese": 3, "Korean": 4, "Han Chinese": 5, "Hispanic": 6, "African-American": 7, "Asian": 8, "Black": 9, "Malay": 10, "White": 11, "Other": 12, "Other Mixed Race": 13}
diabetes_dict = {'0.0': 0, '1.0': 1}
simvastatin_dict = {'0.0': 0, '1.0': 1}
amiodarone_dict = {'0.0': 0, '1.0': 1}
cyp2C9_genotypes_dict = {'*1/*1': 0, '*1/*2': 1, '*1/*3': 2, '*2/*2': 3, '*2/*3': 4, '*3/*3': 5}
VKORC1_genotype_dict = {'AA': 0, 'AG': 1, 'GG': 2}

age = gr.Dropdown(choices=list(age_dict.keys()), label="Age")
gender = gr.Radio(["male", "female"], label="Gender")
race = gr.Dropdown(choices=list(race_dict.keys()), label="Race")
weight = gr.Number(label="Weight")
height = gr.Number(label="Height")
diabetes = gr.Radio(["0.0", "1.0"], label="Diabetes")
simvastatin = gr.Radio(["0.0", "1.0"], label="Simvastatin")
amiodarone = gr.Radio(["0.0", "1.0"], label="Amiodarone")
INR_reported = gr.Number(label="INR on Reported Therapeutic Dose of Warfarin")
cyp2C9_genotypes = gr.Dropdown(choices=list(cyp2C9_genotypes_dict.keys()), label="Cyp2C9 genotypes")
VKORC1_genotype = gr.Dropdown(choices=list(VKORC1_genotype_dict.keys()), label="VKORC1 genotype")
model = gr.Dropdown(choices=["Linear Regression", "Ridge Regression", "Decision Tree", "KNN","ANN", "Random Forest"], label="Model")


# Modify the gender encoding and ensure all categorical variables are properly encoded
def multi_inputs(age, gender, race, weight, height, diabetes, simvastatin, amiodarone, INR_reported, cyp2C9_genotypes, VKORC1_genotype, model, dose):
    # Load the appropriate model based on the selected model
    if model == 'Linear Regression':
        model = load('Linear_regressor_model.pkl')
    elif model == 'Ridge Regression':
        model = load('best_ridge_regression_model.pkl')
    elif model == 'Decision Tree':
        model = load('best_decision_tree_model.pkl')
    elif model == 'KNN':
        model = load('best_knn_regressor_model.pkl')
    elif model == 'ANN':
        model = load('Best_ann_regressor.pkl')
    else:
        model = load("best_random_forest_model.pkl")

    # Map categorical variables to their encoded values
    age_encoded = age_dict.get(age)
    race_encoded = race_dict.get(race)
    gender_encoded = 0 if gender == "male" else 1  # Encoding gender
    diabetes_encoded = diabetes_dict.get(diabetes)
    simvastatin_encoded = simvastatin_dict.get(simvastatin)
    amiodarone_encoded = amiodarone_dict.get(amiodarone)
    cyp2C9_genotypes_encoded = cyp2C9_genotypes_dict.get(cyp2C9_genotypes)
    VKORC1_genotype_encoded = VKORC1_genotype_dict.get(VKORC1_genotype)

    # Transform input data
    inputs = [age_encoded, gender_encoded, race_encoded, weight, height, diabetes_encoded, simvastatin_encoded, amiodarone_encoded, INR_reported, cyp2C9_genotypes_encoded, VKORC1_genotype_encoded]

    # Predict therapeutic dose of warfarin
    input_data = np.array([inputs])
    output = model.predict(input_data)

    return output


# Create the Gradio interface
inputs = [age, gender, race, weight, height, diabetes, simvastatin, amiodarone, INR_reported, cyp2C9_genotypes, VKORC1_genotype, model]
outputs = gr.Textbox(label="Predicted Therapeutic Dose of Warfarin")
gr.Interface(fn=multi_inputs, inputs=inputs, outputs=outputs).launch(share=True, debug=True)