import gradio as gr
import joblib

# Load the trained model
model = joblib.load("loan_classifier.joblib")

# Load Standared Scaler
scalar = joblib.load("std_scaler.bin")


def predict_loan_status(
    int_rate,
    installment,
    log_annual_inc,
    dti,
    fico,
    revol_bal,
    revol_util,
    inq_last_6mths,
    delinq_2yrs,
    pub_rec,
    installment_to_income_ratio,
    credit_history,
):
    input_dict = {
        "int.rate": int_rate,
        "installment": installment,
        "log.annual.inc": log_annual_inc,
        "dti": dti,
        "fico": fico,
        "revol.bal": revol_bal,
        "revol.util": revol_util,
        "inq.last.6mths": inq_last_6mths,
        "delinq.2yrs": delinq_2yrs,
        "pub.rec": pub_rec,
        "installment_to_income_ratio": installment_to_income_ratio,
        "credit_history": credit_history,
    }
    # Convert the dictionary to a 2D array
    input_array = [list(input_dict.values())]
    scaled_array = scalar.transform(input_array)
    prediction = model.predict(scaled_array)[0]

    if prediction == 0:
        return "Loan fully paid"
    else:
        return "Loan not fully paid"


inputs = [
    gr.Slider(0.06, 0.23, step=0.01, label="Interest Rate"),
    gr.Slider(100, 950, step=10, label="Installment"),
    gr.Slider(7, 15, step=0.1, label="Log Annual Income"),
    gr.Slider(0, 40, step=1, label="DTI Ratio"),
    gr.Slider(600, 850, step=1, label="FICO Score"),
    gr.Slider(0, 120000, step=1000, label="Revolving Balance"),
    gr.Slider(0, 120, step=1, label="Revolving Utilization"),
    gr.Slider(0, 10, step=1, label="Inquiries in Last 6 Months"),
    gr.Slider(0, 20, step=1, label="Delinquencies in Last 2 Years"),
    gr.Slider(0, 10, step=1, label="Public Records"),
    gr.Slider(0, 5, step=0.1, label="Installment to Income Ratio"),
    gr.Slider(0, 1, step=0.01, label="Credit History"),
]
outputs = [gr.Label(num_top_classes=2)]

title = "Loan Approval Classifier"
description = (
    "Enter the details of the loan applicant to check if the loan is approved or not."
)
gr.Interface(
    fn=predict_loan_status,
    inputs=inputs,
    outputs=outputs,
    title=title,
    description=description,
).launch()