app.py

import numpy as np
import pandas as pd
import joblib
from flask import Flask, request, jsonify
from datetime import datetime

# Initialize Flask app
Superkart_api = Flask("SuperKart Sales Prediction")

# Load the trained SuperKart sales model from the correct path
model = joblib.load("backend_files/Superkart_Sales_prediction_model_v1_0.joblib")

# Feature Engineering
current_year = datetime.now().year
perishable = {'Meat','Dairy','Fruits and Vegetables','Frozen Foods','Seafood','Breads','Breakfast'}

def apply_feature_engineering(dataset):

    if 'Product_Id' in dataset.columns:
        dataset['Product_Category'] = dataset['Product_Id'].str[:2].map({
            'DR':'Drinks','NC':'Non-Consumable','FD':'Food & Veg'
        }).fillna('Other')
    if 'Store_Establishment_Year' in dataset.columns:
        dataset['Store_Age'] = current_year - dataset['Store_Establishment_Year']
    if 'Product_Type' in dataset.columns:
        dataset['Food_Type'] = dataset['Product_Type'].apply(
            lambda t: 'Perishable' if t in perishable else 'Non-Perishable'
        )
    return dataset

# Home route
@Superkart_api.get('/')
def home():
    return "Welcome to the SuperKart Sales Prediction!"

# Single prediction endpoint
@Superkart_api.post('/v1/sales')
def predict_sales():
    data = request.get_json()
    sample = {
        'Product_Id': data.get('Product_Id'),
        'Product_Weight': data.get('Product_Weight'),
        'Product_Sugar_Content': data.get('Product_Sugar_Content'),
        'Product_Allocated_Area': data.get('Product_Allocated_Area'),
        'Product_Type': data.get('Product_Type'),
        'Product_MRP': data.get('Product_MRP'),
        'Store_Id': data.get('Store_Id'),
        'Store_Establishment_Year': data.get('Store_Establishment_Year'),
        'Store_Size': data.get('Store_Size'),
        'Store_Location_City_Type': data.get('Store_Location_City_Type'),
        'Store_Type': data.get('Store_Type'),
    }
    input_df = pd.DataFrame([sample])
    input_df = apply_feature_engineering(input_df)
    predicted_sales = model.predict(input_df)[0]
    return jsonify({'Predicted Sales': round(float(predicted_sales), 2)})

# Batch prediction endpoint
@Superkart_api.post('/v1/salesbatch')
def predict_sales_batch():
    file = request.files['file']
    input_df = pd.read_csv(file)
    input_df = apply_feature_engineering(input_df)
    predicted_sales = model.predict(input_df).tolist()

    if 'Product_Id' in input_df.columns:
        product_ids = input_df['Product_Id'].tolist()
        output_dict = dict(zip(product_ids, [round(float(x), 2) for x in predicted_sales]))
    else:
        output_dict = [round(float(x), 2) for x in predicted_sales]

    return jsonify(output_dict)

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