# Import necessary libraries
import numpy as np
import joblib  # For loading the serialized machine learning model
import pandas as pd  # For data manipulation and analysis
from flask import Flask, request, jsonify  # Flask framework for creating the web API

# Initialize the Flask application with a descriptive name
total_sales_predictor_api = Flask("SuperKart Product Total Sales Prediction")

# Load the pre-trained regression model from a Joblib file
model = joblib.load("product_store_sales_total_prediction_model_v1_0.joblib")

# -----------------------------------------------
# Define a GET route for the root endpoint
# -----------------------------------------------
@total_sales_predictor_api.get('/')
def home():
    """
    Handles GET requests to the root URL ('/').
    Returns a simple welcome message to confirm the API is running.
    """
    return "Welcome to the SuperKart Product Total Sales Prediction API!"

# -----------------------------------------------
# Define a POST route for single record prediction
# -----------------------------------------------
@total_sales_predictor_api.post('/v1/totalsales')
def predict_total_sales():
    """
    Handles POST requests to '/v1/totalsales' endpoint.
    Accepts JSON input with product and store features,
    and returns the predicted 'Product_Store_Sales_Total' as JSON.
    """
    # Get JSON data from request body
    input_data = request.get_json()

    # Extract necessary input features from JSON
    sample = {
        'Product_Weight': input_data['Product_Weight'],                         # Product weight
        'Product_Allocated_Area': input_data['Product_Allocated_Area'],         # Shelf/display area allocated
        'Product_MRP': input_data['Product_MRP'],                               # Maximum Retail Price
        'Store_Establishment_Year': input_data['Store_Establishment_Year'],     # Store's opening year
        'Product_Sugar_Content': input_data['Product_Sugar_Content'],           # Sugar category (Low/Medium/High)
        'Product_Type': input_data['Product_Type'],                             # Category/type of product
        'Store_Id': input_data['Store_Id'],                                     # Store identifier
        'Store_Size': input_data['Store_Size'],                                 # Store size category
        'Store_Location_City_Type': input_data['Store_Location_City_Type'],     # Urban/rural/metro
        'Store_Type': input_data['Store_Type']                                  # Type of store
    }

    # Convert the single input sample into a DataFrame (expected by model)
    input_df = pd.DataFrame([sample])

    # Make prediction using the trained model
    predicted_sales = model.predict(input_df)[0]  # Predict returns a NumPy float

    # Round off the result to 2 decimal places and convert to Python float
    predicted_sales = round(float(predicted_sales), 2)

    # Return the prediction result as a JSON response
    return jsonify({'Predicted Product Store Sales Total': predicted_sales})

# ---------------------------------------------------
# Define a POST route for batch prediction via CSV
# ---------------------------------------------------
@total_sales_predictor_api.post('/v1/totalsalesbatch')
def predict_total_sales_batch():
    """
    Handles POST requests to '/v1/totalsalesbatch' endpoint.
    Accepts a CSV file of multiple records and returns a dictionary
    mapping 'Product_Id' to predicted total sales value.
    """
    # Retrieve the uploaded file from the request
    file = request.files['file']

    # Read the uploaded CSV into a DataFrame
    input_df = pd.read_csv(file)

    # Generate predictions using the trained model
    predicted_prices = model.predict(input_df).tolist()

    # Extract Product IDs to identify predictions
    product_ids = input_df['Product_Id'].tolist()

    # Pair each prediction with its corresponding Product ID
    output_dict = dict(zip(product_ids, [round(float(p), 2) for p in predicted_prices]))

    # Return the results as a dictionary in JSON response
    return output_dict

# ---------------------------------------------------
# Run the Flask app in debug mode if called directly
# ---------------------------------------------------
if __name__ == '__main__':
    total_sales_predictor_api.run(debug=True)