import gradio as gr
import pickle
import numpy as np

# Load the preprocessor and model
with open("processor.pkl", "rb") as f:
    processor = pickle.load(f)

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

# The original FastAPI endpoint URL (not active now)
# API_URL = "http://4.186.39.227/predict"

# Function to preprocess input data and predict customer segment
def predict_customer_segment(data):
    # Preprocess the data
    processed_data = processor.transform(np.array([[
        data["CUST_ID"], data["BALANCE"], data["BALANCE_FREQUENCY"], data["PURCHASES"],
        data["ONEOFF_PURCHASES"], data["INSTALLMENTS_PURCHASES"], data["CASH_ADVANCE"],
        data["PURCHASES_FREQUENCY"], data["ONEOFF_PURCHASES_FREQUENCY"], data["PURCHASES_INSTALLMENTS_FREQUENCY"],
        data["CASH_ADVANCE_FREQUENCY"], data["CASH_ADVANCE_TRX"], data["PURCHASES_TRX"],
        data["CREDIT_LIMIT"], data["PAYMENTS"], data["MINIMUM_PAYMENTS"], data["PRC_FULL_PAYMENT"],
        data["TENURE"]
    ]]))

    # Predict the cluster label using the model
    cluster_label = model.predict(processed_data)[0]

    # Map the cluster label to a human-readable description
    cluster_descriptions = {
        0: "Low Value Customer",
        1: "Medium Value Customer",
        2: "High Value Customer",
        3: "Premium Customer"
    }

    return cluster_descriptions.get(cluster_label, "Unknown Cluster")

# Function to handle predictions from form input
def predict_from_form(CUST_ID, BALANCE, BALANCE_FREQUENCY, PURCHASES, ONEOFF_PURCHASES, INSTALLMENTS_PURCHASES, CASH_ADVANCE, PURCHASES_FREQUENCY, ONEOFF_PURCHASES_FREQUENCY, PURCHASES_INSTALLMENTS_FREQUENCY, CASH_ADVANCE_FREQUENCY, CASH_ADVANCE_TRX, PURCHASES_TRX, CREDIT_LIMIT, PAYMENTS, MINIMUM_PAYMENTS, PRC_FULL_PAYMENT, TENURE):
    data = {
        "CUST_ID": CUST_ID,
        "BALANCE": BALANCE,
        "BALANCE_FREQUENCY": BALANCE_FREQUENCY,
        "PURCHASES": PURCHASES,
        "ONEOFF_PURCHASES": ONEOFF_PURCHASES,
        "INSTALLMENTS_PURCHASES": INSTALLMENTS_PURCHASES,
        "CASH_ADVANCE": CASH_ADVANCE,
        "PURCHASES_FREQUENCY": PURCHASES_FREQUENCY,
        "ONEOFF_PURCHASES_FREQUENCY": ONEOFF_PURCHASES_FREQUENCY,
        "PURCHASES_INSTALLMENTS_FREQUENCY": PURCHASES_INSTALLMENTS_FREQUENCY,
        "CASH_ADVANCE_FREQUENCY": CASH_ADVANCE_FREQUENCY,
        "CASH_ADVANCE_TRX": CASH_ADVANCE_TRX,
        "PURCHASES_TRX": PURCHASES_TRX,
        "CREDIT_LIMIT": CREDIT_LIMIT,
        "PAYMENTS": PAYMENTS,
        "MINIMUM_PAYMENTS": MINIMUM_PAYMENTS,
        "PRC_FULL_PAYMENT": PRC_FULL_PAYMENT,
        "TENURE": TENURE
    }
    return predict_customer_segment(data)

# Define Gradio interface components
with gr.Blocks() as demo:
    gr.Markdown("# Customer Segmentation Model")
    gr.Markdown("### Predict the customer segment based on the provided details")

    with gr.Row():
        CUST_ID = gr.Textbox(label="Customer ID", lines=1, max_lines=1)
        BALANCE = gr.Number(label="Balance")
        BALANCE_FREQUENCY = gr.Number(label="Balance Frequency")
        PURCHASES = gr.Number(label="Purchases")
        ONEOFF_PURCHASES = gr.Number(label="One-off Purchases")
        INSTALLMENTS_PURCHASES = gr.Number(label="Installments Purchases")
        CASH_ADVANCE = gr.Number(label="Cash Advance")
        PURCHASES_FREQUENCY = gr.Number(label="Purchases Frequency")
        ONEOFF_PURCHASES_FREQUENCY = gr.Number(label="One-off Purchases Frequency")
        PURCHASES_INSTALLMENTS_FREQUENCY = gr.Number(label="Purchases Installments Frequency")
        CASH_ADVANCE_FREQUENCY = gr.Number(label="Cash Advance Frequency")
        CASH_ADVANCE_TRX = gr.Number(label="Cash Advance Transactions")
        PURCHASES_TRX = gr.Number(label="Purchases Transactions")
        CREDIT_LIMIT = gr.Number(label="Credit Limit")
        PAYMENTS = gr.Number(label="Payments")
        MINIMUM_PAYMENTS = gr.Number(label="Minimum Payments")
        PRC_FULL_PAYMENT = gr.Number(label="Percent Full Payment")
        TENURE = gr.Number(label="Tenure")

    predict_button = gr.Button("Predict", variant="primary")
    prediction_output = gr.Textbox(label="Predicted Customer Segment", lines=1, max_lines=1)

    predict_button.click(predict_from_form, inputs=[
        CUST_ID, BALANCE, BALANCE_FREQUENCY, PURCHASES, ONEOFF_PURCHASES, INSTALLMENTS_PURCHASES,
        CASH_ADVANCE, PURCHASES_FREQUENCY, ONEOFF_PURCHASES_FREQUENCY, PURCHASES_INSTALLMENTS_FREQUENCY,
        CASH_ADVANCE_FREQUENCY, CASH_ADVANCE_TRX, PURCHASES_TRX, CREDIT_LIMIT, PAYMENTS,
        MINIMUM_PAYMENTS, PRC_FULL_PAYMENT, TENURE
    ], outputs=prediction_output)

demo.launch()