app.py

import gradio as gr
import joblib

# Define a function for the Gradio interface
def predict_risk(Age, hOCP, hMisCrg, hHRT, CA125, HE4, FBS, USG):
    # Load the trained model
    model = joblib.load('RDF_OvCa_Final.joblib')
    
    # Create a DataFrame with the input data
    input_data = pd.DataFrame({
        'Age': [Age],
        'hOCP': [hOCP],
        'hMisCrg': [hMisCrg],
        'hHRT': [hHRT],
        'CA125': [CA125],
        'HE4': [HE4],
        'FBS': [FBS],
        'USG': [USG]
    })
    
    # Predict the probability of malignancy (class 1)
    probability = model.predict_proba(input_data)[:, 1][0]
    
    # Scale the risk score to -1 to +1
    risk_score = 2 * probability - 1
    
    # Determine the predicted status
    status = "malignant" if probability >= cutoff else "benign"
    
    result = f"The OvaCa Risk Score is {risk_score:.2f}. Based on it, the probability is more of {status} ({1 if status == 'malignant' else 0})."
    
    return result

# Define the Gradio interface
iface = gr.Interface(
    fn=predict_risk,
    inputs=[
        gr.inputs.Number(label="Age of patient in years"),
        gr.inputs.Radio([0, 1], label="History of OCP intake (0: No 1: Yes)"),
        gr.inputs.Radio([0, 1], label="History of Miscarriage (0: No 1: Yes)"),
        gr.inputs.Radio([0, 1], label="History of HRT (0: No 1: Yes)"),
        gr.inputs.Number(label="Serum CA125 level"),
        gr.inputs.Number(label="Serum HE4 level"),
        gr.inputs.Number(label="Serum Fasting Blood Sugar Level"),
        gr.inputs.Radio([0, 1], label="USG Finding (0: Absent or Single Finding 1: More Than 1 Findings)")
    ],
    outputs=gr.outputs.Textbox(label="Result (Risk Score on a Scale of -1 to +1, where >0 ~ Malignant)")
)

# Launch the Gradio interface
iface.launch()