app.py

import streamlit as st
import pandas as pd
import numpy as np
import joblib

# page config
st.set_page_config(
    page_title="Loan_Prediction App",
    page_icon="🗃️",
    layout="wide"
)

@st.cache_resource
def load_model():
    return joblib.load("Loan.joblib")

st.title("🗃️ Loan Prediction App")

# ---- Input UI ----
col1, col2 = st.columns(2)

with col1:
    LoanId = st.text_input("Loan ID")
    Gender = st.selectbox("Gender", ["Male", "Female"])
    Martial_Status = st.selectbox("Marital Status", ["Yes", "No"])
    Employed = st.selectbox("Employed", ["Yes", "No"])
    ApplicantIncome = st.number_input("Applicant Income", min_value=0)

with col2:
    CoapplicantIncome = st.number_input("Coapplicant Income", min_value=0)
    Loan_Amount = st.number_input("Loan Amount", min_value=0)
    Loan_Amount_Term = st.number_input("Loan Amount Term", min_value=0)
    Credit_History = st.selectbox("Credit History", [1.0, 0.0])

# ---- Encoding for model ----
def encode_inputs():
    gender_map = {"Male": 1, "Female": 0}
    yesno_map = {"Yes": 1, "No": 0}

    return [
        hash(LoanId) % 10000,  # convert alphanumeric LoanID -> numeric
        gender_map[Gender],
        yesno_map[Martial_Status],
        yesno_map[Employed],
        ApplicantIncome,
        CoapplicantIncome,
        Loan_Amount,
        Loan_Amount_Term,
        Credit_History
    ]

# ---- Predict ----
if st.button("Predict Loan Status", type="primary"):
    with st.spinner("Predicting..."):

        model = load_model()
        features = np.array(encode_inputs()).reshape(1, -1)

        prediction = model.predict(features)

        if prediction[0] == 1:
            st.success("Loan Approved ✅")
        else:
            st.error("Loan Rejected ❌")