# Import necessary libraries
import numpy as np
import joblib  # For loading the serialized model
import pandas as pd  # For data manipulation
from flask import Flask, request, jsonify  # For creating the Flask API

# Initialize Flask app with a name reflective of the project
extraalearn_api = Flask("ExtLearn")

# Load the trained lead conversion model (ensure the file name matches your saved model)
model = joblib.load("extlearn_model.joblib")

# Define a route for the home page
@extraalearn_api.get('/')
def home():
    return "Welcome to the ExtraaLearn Lead Conversion Prediction System"

# Define an endpoint to predict status (converted/not converted) for a lead
@extraalearn_api.post('/v1/predict')
def predict_conversion():
    # Get JSON data from the request body
    data = request.get_json()

    # Extract relevant features based on the ExtraaLearn dataset
    # These must match the exact feature names used during model training
    sample = {
        'age': data['age'],
        'current_occupation': data['current_occupation'],
        'first_interaction': data['first_interaction'],
        'profile_completed': data['profile_completed'],
        'website_visits': data['website_visits'],
        'time_spent_on_website': data['time_spent_on_website'],
        'page_views_per_visit': data['page_views_per_visit'],
        'last_activity': data['last_activity'],
        'print_media_type1': data['print_media_type1'],
        'print_media_type2': data['print_media_type2'],
        'digital_media': data['digital_media'],
        'educational_channels': data['educational_channels'],
        'referral': data['referral']
    }

    # Convert the extracted data into a DataFrame for the model pipeline
    input_data = pd.DataFrame([sample])
    
    # Calculate the engineered feature 'age_time_interaction'
    input_data['age_time_interaction'] = input_data['age'] * input_data['time_spent_on_website']

    # Make a prediction (1 for converted, 0 for not converted)
    prediction = int(model.predict(input_data)[0])

    # Optional: Get the probability of conversion
    probability = model.predict_proba(input_data)[0][1]

    # Return the prediction and probability as a JSON response
    return jsonify({
        'Status_Prediction': prediction,
        'Conversion_Probability': round(float(probability), 4),
        'Message': 'High Potential Lead' if prediction == 1 else 'Low Potential Lead'
    })

# Run the Flask app
if __name__ == '__main__':
    extraalearn_api.run(debug=True)