app.py

import pickle
import pandas as pd
import shap
from shap.plots._force_matplotlib import draw_additive_plot
import gradio as gr
import numpy as np
import matplotlib.pyplot as plt

# load the model from disk
loaded_model = pickle.load(open("db_xgb.pkl", 'rb'))

# Setup SHAP
explainer = shap.Explainer(loaded_model) # PLEASE DO NOT CHANGE THIS.

# Create the main function for server
def main_func(HighBP, HighChol, CholCheck, BMI, Smoker, Stroke, HeartDiseaseorAttack, PhysActivity, Fruits, Veggies, HvyAlcoholConsump, AnyHealthcare, NoDocbcCost, GenHlth, MentHlth, PhysHlth, DiffWalk, Sex, Age, Education, Income):
    new_row = pd.DataFrame.from_dict({'HighBP': HighBP, 'HighChol': HighChol, 'CholCheck': CholCheck, 'BMI': BMI, 'Smoker': Smoker, 'Stroke': Stroke, 'HeartDiseaseorAttack': HeartDiseaseorAttack, 'PhysActivity':PhysActivity, 'Fruits':Fruits, 'Veggies':Veggies, 'HvyAlcoholConsump': HvyAlcoholConsump, 'AnyHealthcare': AnyHealthcare, 'NoDocbcCost': NoDocbcCost, 'GenHlth': GenHlth, 'MentHlth': MentHlth, 'PhysHlth': PhysHlth, 'DiffWalk': DiffWalk, 'Sex': Sex, 'Age': Age, 'Education': Education, 'Income': Income}, 
                                     orient = 'index').transpose()
    
    prob = loaded_model.predict_proba(new_row)
    
    shap_values = explainer(new_row)
    # plot = shap.force_plot(shap_values[0], matplotlib=True, figsize=(30,30), show=False)
    # plot = shap.plots.waterfall(shap_values[0], max_display=6, show=False)
    plot = shap.plots.bar(shap_values[0], max_display=6, order=shap.Explanation.abs, show_data='auto', show=False)

    plt.tight_layout()
    local_plot = plt.gcf()
    plt.close()
    
    return {"Low Chance": float(prob[0][0]), "High Chance": 1-float(prob[0][0])}, local_plot

# Create the UI
title = "**Diabetes Predictor & Interpreter** 🪐"
description1 = """This app takes info from subjects and predicts their diabetes likelihood. Do not use for medical diagnosis."""

description2 = """
To use the app, click on one of the examples, or adjust the values of the factors, and click on Analyze. 🤞
""" 

with gr.Blocks(title=title) as demo:
    gr.Markdown(f"## {title}")
    gr.Markdown(description1)
    gr.Markdown("""---""")
    gr.Markdown(description2)
    gr.Markdown("""---""")
    
    input_block_left() == gr.Column(
        HighBP = gr.Radio(label="Do you have high Blood Pressure?", choices=["No", "Yes"], default="Yes", description="0 = no high BP, 1 = high BP"),
        HighChol = gr.Radio(label="Do you have high Cholesterol?",  choices=["No", "Yes"], default="Yes", description="0 = no high cholesterol, 1 = high cholesterol"),
        CholCheck = gr.Radio(label="Did you have a Cholestorol check in past 5 years?", choices=["No", "Yes"], default="Yes", description="No = not checked in 5 years, Yes = checked in 5 years"),
        BMI = gr.Number(label="BMI", minimum=0, maximum=98, default=1),
        Smoker = gr.Radio(label="Are you a smoker?", choices=["No", "Yes"], default="Yes", description="No = never smoked, Yes = smoked at least 100 cigarettes"),
        Stroke = gr.Radio(label="Have you had a stroke?", choices=["No", "Yes"], default="Yes", description="No = never had a stroke, Yes = had a stroke"),
        HeartDiseaseorAttack = gr.Radio(label="Do you have coronary heart disease or myocardial infarction?", choices=["No", "Yes"], default="Yes", description="No = no CHD/MI, Yes = CHD/MI"),
        PhysActivity = gr.Radio(label="Physical Activity in the past 30 days?", choices=["No", "Yes"], default="Yes", description="No = no activity, Yes = active"),
        Fruits = gr.Radio(label="Do you consume fruit 1 or more times per day?", choices=["No", "Yes"], default="Yes", description="No = less than daily, Yes = daily"),
        Veggies = gr.Radio(label="Do you consume Vegetables 1 or more times per day?", choices=["No", "Yes"], default="Yes", description="No = less than daily, Yes = daily"),
        HvyAlcoholConsump = gr.Radio(label="Do you drink often? (adult men having more than 14 drinks/week and adult women having more than 7 drinks/week)", choices=["No", "Yes"], default="Yes", description="No = not heavy drinker, Yes = heavy drinker"),
        AnyHealthcare = gr.Radio(label="Do you have any kind of health care coverage? (e.g., health insurance, prepaid plans such as HMO)", choices=["No", "Yes"], default="Yes", description="No = no coverage, Yes = coverage"),
        NoDocbcCost = gr.Radio(label="Was there a time in the past 12 months when you needed to see a doctor but could not because of cost?", choices=["No", "Yes"], default="Yes", description="No = no barrier, Yes = cost barrier")
    )
                    
    input_block_right == gr.Column(
        GenHlth = gr.Slider(label="In general, rank your health on a scale: 1(excellent)-5(poor)", minimum=1, maximum=5, default=1, step=1, description="1 = excellent, 5 = poor"),
        MentHlth = gr.Number(label="Poor Mental Health in the Past 30 Days", minimum=0, maximum=30, default=1, description="Days not good out of last 30"),
        PhysHlth = gr.Number(label="Poor Physical Health in the Past 30 Days", minimum=0, maximum=30, default=1, description="Days not good out of last 30"),
        DiffWalk = gr.Radio(label="Do you have serious difficulty walking or climbing stairs?", choices=["No", "Yes"], default="Yes", description="No = no difficulty, Yes = difficulty"),
        Sex = gr.Radio(label="Sex", choices=["Female", "Male"], default="Male", description="Female or Male"),
        Age = gr.Number(label="Age", minimum=1, maximum=100, default=1),
        Education = gr.Dropdown(label="Education Level", choices=["Never attended school", "Grades 1-8", "Grades 9-11", "Grade 12 or GED", "College 1-3 years", "College 4+ years"], default="Never attended school", description="Education level"),
        Income = gr.Dropdown(label="Income Level", choices=["< $10,000", "$10,000 - $24,999", "$25,000 - $49,999", "$50,000 - $74,999", "$75,000 or more"], default="< $10,000", description="Income level")
    )
    
    input_block == gr.Row([input_block_left, input_block_right])
        
    output_block == gr.Column(
        gr.Label(label="Predicted Label"),
        gr.Plot(label="SHAP Plot")
    )

    submit_btn = gr.Button("Analyze")

    submit_btn.click(
        main_func,
        [HighBP, HighChol, CholCheck, BMI, Smoker, Stroke, HeartDiseaseorAttack, PhysActivity, Fruits, Veggies, HvyAlcoholConsump, AnyHealthcare, NoDocbcCost, GenHlth, MentHlth, PhysHlth, DiffWalk, Sex, Age, Education, Income],
        [label,local_plot], api_name="Diabetes_Predictor"
    )
    
    gr.Markdown("### Click on any of the examples below to see how it works:")
    gr.Examples([[0,0,1,0,22,0,0,0,1,1,1,0,0,1,3,25,23,1,1,21,5,3], [1,1,1,1,30,1,1,1,0,0,0,1,1,0,2,20,23,0,0,21,3,2]], [HighBP, HighChol, CholCheck, BMI, Smoker, Stroke, HeartDiseaseorAttack, PhysActivity, Fruits, Veggies, HvyAlcoholConsump, AnyHealthcare, NoDocbcCost, GenHlth, MentHlth, PhysHlth, DiffWalk, Sex, Age, Education, Income], [label,local_plot], main_func, cache_examples=True)

demo.launch()