src/app.py

import os
os.environ["XDG_CONFIG_HOME"] = "/tmp/.streamlit"
os.makedirs("/tmp/.streamlit", exist_ok=True)
import streamlit as st
import joblib
import pandas as pd

os.environ["XDG_CONFIG_HOME"] = "/tmp/.streamlit"
# Load the trained model
model_path = os.path.join(os.path.dirname(__file__), "model.pkl")
model, expected_columns = joblib.load(model_path)

st.title("💳 Loan Default Prediction App")
st.markdown("Enter applicant information to predict the likelihood of default.")

# Collect input
input_data = {
    "RevolvingUtilizationOfUnsecuredLines": st.slider("Revolving Utilization", 0.0, 1.0, 0.3),
    "age": st.slider("Age", 18, 100, 45),
    "NumberOfTime30-59DaysPastDueNotWorse": st.number_input("30-59 Days Past Due", 0, 100, 0),
    "DebtRatio": st.slider("Debt Ratio", 0.0, 2.0, 0.8),
    "MonthlyIncome": st.number_input("Monthly Income", 0, 100000, 5000),
    "NumberOfOpenCreditLinesAndLoans": st.number_input("Open Credit Lines", 0, 50, 6),
    "NumberOfTimes90DaysLate": st.number_input("Times 90 Days Late", 0, 100, 0),
    "NumberRealEstateLoansOrLines": st.number_input("Real Estate Loans", 0, 50, 1),
    "NumberOfTime60-89DaysPastDueNotWorse": st.number_input("60-89 Days Past Due", 0, 100, 0),
    "NumberOfDependents": st.number_input("Number of Dependents", 0, 20, 2)
}

if st.button("Predict"):
    input_df = pd.DataFrame([input_data])
    prediction = model.predict(input_df)[0]
    result = "🚨 Will Default" if prediction == 1 else "✅ Will Not Default"
    st.subheader(f"Prediction: {result}")