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("glioma_xgb.pkl", 'rb'))

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

# Create the main function for server
def main_func(Gender, Age_at_diagnosis, IDH1, TP53, ATRX, PTEN, EGFR, CIC, MUC16, PIK3CA, NF1, PIK3R1, FUBP1, RB1, NOTCH1, BCOR, CSMD3, SMARCA4, GRIN2A, IDH2, FAT4, PDGFRA):
    new_row = pd.DataFrame.from_dict({'Gender':Gender,
              'Age_at_diagnosis':Age_at_diagnosis,'IDH1':IDH1,'TP53':TP53,
              'ATRX':ATRX, 'PTEN':PTEN,'EGFR':EGFR,'CIC':CIC,
              'MUC16':MUC16,'PIK3CA':PIK3CA,'NF1':NF1,'PIK3R1':PIK3R1, 'FUBP1': FUBP1, 'RB1': RB1, 'NOTCH1': NOTCH1,
              'BCOR': BCOR, 'CSMD3': CSMD3, 'SMARCA4': SMARCA4, 'GRIN2A': GRIN2A, 'IDH2': IDH2, 'FAT4': FAT4, 'PDGFRA': PDGFRA}, 
               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 {"Chance of Having GBM Tumor": 1-float(prob[0][0]), "Chance of Having LGG Tumor": float(prob[0][0])}, local_plot

# Create the UI
title = "**Glioma Predictor & Interpreter** 🪐"
description1 = """This app takes info from subjects and predicts the severity of their brain tumor (LGG or GBM). 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("""---""")

    with gr.Row():
        Gender = gr.Radio(["Female", "Male"], label="Gender", type = "index")
        Age_at_diagnosis = gr.Number(label="Age at Diagnosis")
    with gr.Row():
        IDH1 = gr.Radio(["No", "Yes"], label="IDH1 Mutation", type="index") 
        TP53 = gr.Radio(["No", "Yes"], label="TP53 Mutation", type="index")
        ATRX = gr.Radio(["No", "Yes"], label="ATRX Mutation", type="index")
    with gr.Row():
        PTEN = gr.Radio(["No", "Yes"], label="PTEN Mutation", type="index")
        EGFR = gr.Radio(["No", "Yes"], label="EGFR Mutation", type="index")
        CIC = gr.Radio(["No", "Yes"], label="CIC Mutation", type="index")
    with gr.Row():
        MUC16 = gr.Radio(["No", "Yes"], label="MUC16 Mutation", type="index")
        PIK3CA = gr.Radio(["No", "Yes"], label="PIK3CA Mutation", type="index")
        NF1 = gr.Radio(["No", "Yes"], label="NF1 Mutation", type="index")
    with gr.Row():
        PIK3R1 = gr.Radio(["No", "Yes"], label="PIK3R1 Mutation", type="index")
        FUBP1 = gr.Radio(["No", "Yes"], label="FUBP1 Mutation", type="index")
        RB1 = gr.Radio(["No", "Yes"], label="RB1 Mutation", type="index")
    with gr.Row():
        NOTCH1 = gr.Radio(["No", "Yes"], label="NOTCH1 Mutation", type="index")
        BCOR = gr.Radio(["No", "Yes"], label="BCOR Mutation", type="index")
        CSMD3 = gr.Radio(["No", "Yes"], label="CSMD3 Mutation", type="index")
    with gr.Row():
        SMARCA4 = gr.Radio(["No", "Yes"], label="SMAECA4 Mutation", type="index")
        GRIN2A = gr.Radio(["No", "Yes"], label="GRIN2A Mutation", type="index")
        IDH2 = gr.Radio(["No", "Yes"], label="IDH2 Mutation", type="index")
        FAT4 = gr.Radio(["No", "Yes"], label="FAT4 Mutation", type="index")
        PDGFRA = gr.Radio(["No", "Yes"], label="PDGFRA Mutation", type="index")






    
    submit_btn = gr.Button("Analyze")

    with gr.Column(visible=True) as output_col:
        label = gr.Label(label = "Predicted Label")
        local_plot = gr.Plot(label = 'Grade:')

    submit_btn.click(
        main_func,
        [Gender, Age_at_diagnosis, IDH1, TP53, ATRX, PTEN, EGFR, CIC, MUC16, PIK3CA, NF1, PIK3R1, FUBP1, RB1, NOTCH1, BCOR, CSMD3, SMARCA4, GRIN2A, IDH2, FAT4, PDGFRA],
        [label,local_plot], api_name="Glioma_Predictor"
    )
    
    gr.Markdown("### Click on any of the examples below to see how it works:")
    gr.Examples([["Male",24,"Yes","No","Yes","Yes","Yes","No","Yes","Yes","Yes","Yes","Yes","No","No","No","No","Yes","No","Yes","No","Yes"], ["Male",70,"No","No","No","No","No","No","No","No","No","Yes","No","Yes","No","No","No","No","No","No","No", "No"]], [Gender, Age_at_diagnosis, IDH1, TP53, ATRX, PTEN, EGFR, CIC, MUC16, PIK3CA, NF1, PIK3R1, FUBP1, RB1, NOTCH1, BCOR, CSMD3, SMARCA4, GRIN2A, IDH2, FAT4, PDGFRA], [label,local_plot], main_func, cache_examples=True)

demo.launch()