app.py

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

# Load trained model (Pipeline with preprocessing inside, or just model)
model = joblib.load("superkart_sales_predictor.joblib")

app = Flask(__name__)

# Final feature columns used during training
final_columns = [
    'Product_Id',
    'Store_Id',
    'Product_NumId',
    'Product_Code',
    'Store_Establishment_Year',
    'Product_Weight',
    'Product_MRP',
    'Product_Allocated_Area',
    'Product_Type',
    'Store_Size',
    'Store_Location_City_Type',
    'Store_Type',
    'Product_Sugar_Content',
    'Store_Age',
    'Is_No_Sugar'
]

# Mappings for categorical encoding (adjust based on your training pipeline)
store_size_map = {"Small": 0, "Medium": 1, "High": 2}
store_location_map = {"Tier 1": 0, "Tier 2": 1, "Tier 3": 2}
store_type_map = {
    "Grocery Store": 0,
    "Supermarket Type1": 1,
    "Supermarket Type2": 2,
    "Supermarket Type3": 3
}
product_type_map = {
    "Dairy": 0,
    "Soft Drinks": 1,
    "Meat": 2,
    "Fruits and Vegetables": 3,
    "Household": 4,
    "Baking Goods": 5,
    "Snack Foods": 6,
    "Frozen Foods": 7,
    "Breakfast": 8,
    "Health and Hygiene": 9,
    "Hard Drinks": 10,
    "Canned": 11,
    "Breads": 12,
    "Starchy Foods": 13,
    "Others": 14
}
sugar_content_map = {"No Sugar": 0, "Low": 1, "Regular": 2}

@app.route("/")
def home():
    return jsonify({"message": "SuperKart Sales Predictor API is running!"})

@app.route("/predict", methods=["POST"])
def predict():
    try:
        # Get JSON request
        data = request.json
        df = pd.DataFrame([data])

        # Add engineered columns
        df["Product_Id"] = 1
        df["Store_Id"] = 1
        df["Product_NumId"] = df.index + 1
        df["Product_Code"] = df.index + 100

        # Store age
        current_year = datetime.datetime.now().year
        df["Store_Age"] = current_year - df["Store_Establishment_Year"]

        # Encode categorical variables
        df["Store_Size"] = df["Store_Size"].map(store_size_map)
        df["Store_Location_City_Type"] = df["Store_Location_City_Type"].map(store_location_map)
        df["Store_Type"] = df["Store_Type"].map(store_type_map)
        df["Product_Type"] = df["Product_Type"].map(product_type_map)
        df["Product_Sugar_Content"] = df["Product_Sugar_Content"].map(sugar_content_map)

        # Fallback for unseen values → set as -1
        df.fillna(-1, inplace=True)

        # Binary flag for No Sugar
        df["Is_No_Sugar"] = df["Product_Sugar_Content"].apply(lambda x: 1 if x == 0 else 0)

        # Reorder columns
        df = df[final_columns]

        # Prediction
        prediction = model.predict(df)[0]

        return jsonify({"prediction": float(prediction)})

    except Exception as e:
        return jsonify({"error": str(e)})


if __name__ == "__main__":
    app.run(host="0.0.0.0", port=7860)