import gradio as gr
import pickle
import numpy as np

# Load model and scaler
with open("log_reg.pkl", "rb") as f:
    model = pickle.load(f)

with open("scaler.pkl", "rb") as f:
    scaler = pickle.load(f)

def predict_creditworthiness(age, income, debt, payment_history, employment_status):
    try:
        debt_to_income = debt / (income + 1)
        payment_map = {"Bad": 0, "Average": 1, "Good": 2}
        payment_num = payment_map.get(payment_history, 1)
        employed_num = 1 if employment_status == "Employed" else 0

        input_data = np.array([[age, income, debt, debt_to_income, payment_num, employed_num]])
        
        # --- Debug prints ---
        print("Input data", input_data)
        print("Input shape", input_data.shape)
        print("Scaler expects", scaler.mean_.shape)   # Should be (6,)

        input_scaled = scaler.transform(input_data)
        proba = model.predict_proba(input_scaled)[0, 1]
        pred_class = model.predict(input_scaled)[0]
        credit_status = "Good Credit" if pred_class == 1 else "Bad Credit"
        return f"Prediction: {credit_status} (Probability of good credit: {proba:.2f})"
    except Exception as e:
        # Print error for debug
        print("Prediction error:", str(e))
        return f"Error: {str(e)}"

iface = gr.Interface(
    fn=predict_creditworthiness,
    inputs=[
        gr.Number(label="Age", value=30, precision=0),
        gr.Number(label="Income", value=50000, precision=2),
        gr.Number(label="Debt", value=5000, precision=2),
        gr.Dropdown(label="Payment History", choices=["Bad", "Average", "Good"], value="Average"),
        gr.Radio(label="Employment Status", choices=["Employed", "Unemployed"], value="Employed"),
    ],
    outputs="text",
    title="Credit Scoring Model (Logistic Regression)",
    description="Enter your financial details to predict creditworthiness."
)

if __name__ == "__main__":
    iface.launch()