Upload folder using huggingface_hub

Dockerfile

@@ -0,0 +1,24 @@
+# ================================
+# Dockerfile for SuperKart Price Prediction API
+# ================================
+
+# Use a lightweight official Python image
+FROM python:3.9-slim
+
+# Set the working directory inside the container
+WORKDIR /app
+
+# Copy all backend and model files into the container
+COPY . .
+
+# Install dependencies from requirements.txt
+RUN pip install --no-cache-dir --upgrade -r requirements.txt
+
+# Expose port 7860 (Hugging Face Spaces default for Docker)
+EXPOSE 7860
+
+# Run the Flask app using Gunicorn with 4 workers
+# -w 4 : 4 worker processes for handling multiple requests
+# -b 0.0.0.0:7860 : bind to all network interfaces on port 7860
+# app:superkart_api : points to "superkart_api" instance in app.py
+CMD ["gunicorn", "-w", "4", "-b", "0.0.0.0:7860", "app:superkart_api"]

app.py

@@ -0,0 +1,88 @@
+# ==============================
+# SuperKart Price Prediction API (Flask Backend)
+# ==============================
+
+# Import necessary libraries
+from flask import Flask, request, jsonify  # For creating the Flask API
+
+# Initialize the Flask application
+superkart_api = Flask("SuperKart Price Predictor")
+
+# Load the trained SuperKart model pipeline (includes preprocessing + tuned XGBoost model)
+model = joblib.load("deployment_files/xgb_tuned_superkart_model_v1_0.joblib")
+
+# ---------------------------------
+# Root Endpoint (GET request)
+# ---------------------------------
+@superkart_api.get('/')
+def home():
+    """
+    This function handles GET requests to the root URL ('/').
+    It simply returns a welcome message confirming the API is running.
+    """
+    return "Welcome to the SuperKart Price Prediction API!"
+
+# ---------------------------------
+# Single Prediction Endpoint (POST request)
+# ---------------------------------
+@superkart_api.post('/v1/predict')
+def predict_price():
+    """
+    Handles POST requests to '/v1/predict'.
+    Expects JSON input with product features and returns
+    the predicted selling price as JSON.
+    """
+    # Get JSON data from request body
+    product_data = request.get_json()
+
+    # Convert JSON to a DataFrame (single row)
+    input_data = pd.DataFrame([product_data])
+
+    # Make prediction using the trained pipeline
+    predicted_price = model.predict(input_data)[0]
+
+    # Ensure prediction is a standard Python float and round
+    predicted_price = round(float(predicted_price), 2)
+
+    # Return prediction as JSON response
+    return jsonify({'Predicted Price': predicted_price})
+
+# ---------------------------------
+# Batch Prediction Endpoint (POST request)
+# ---------------------------------
+@superkart_api.post('/v1/predictbatch')
+def predict_price_batch():
+    """
+    Handles POST requests to '/v1/predictbatch'.
+    Expects a CSV file upload with multiple product entries and
+    returns predicted prices for all rows in JSON format.
+    """
+    # Get uploaded CSV file
+    file = request.files['file']
+
+    # Read CSV into DataFrame
+    input_data = pd.read_csv(file)
+
+    # Make predictions
+    predicted_prices = model.predict(input_data).tolist()
+
+    # Round predictions to 2 decimal places
+    predicted_prices = [round(float(p), 2) for p in predicted_prices]
+
+    # If 'id' column exists in data, map predictions to IDs
+    if 'id' in input_data.columns:
+        product_ids = input_data['id'].tolist()
+        output_dict = dict(zip(product_ids, predicted_prices))
+    else:
+        # Otherwise return as a simple list
+        output_dict = {"Predicted Prices": predicted_prices}
+
+    # Return JSON response
+    return jsonify(output_dict)
+
+# ---------------------------------
+# Run the Flask application
+# ---------------------------------
+if __name__ == '__main__':
+    # Run API in debug mode (for development)
+    superkart_api.run(debug=True)

requirements.txt

@@ -0,0 +1,14 @@
+# This file specifies all the dependencies required to run the SuperKart Price Prediction API
+# It ensures that the environment is consistent across development, testing, and deployment.
+
+pandas==2.2.2          # For handling tabular data and preprocessing
+numpy==2.0.2           # For numerical computations
+scikit-learn==1.6.1    # For preprocessing pipeline utilities
+xgboost==2.1.4         # The core ML model (XGBoost Regressor)
+joblib==1.4.2          # For loading/saving serialized models
+Werkzeug==2.2.2        # Flask dependency (request/response utilities)
+flask==2.2.2           # Flask framework for serving the API
+gunicorn==20.1.0       # WSGI server for production deployment
+requests==2.28.1       # For making HTTP requests (testing endpoints, etc.)
+uvicorn[standard]      # Optional: ASGI server (if switching to FastAPI later)
+streamlit==1.43.2      # For interactive frontend (if we build a Streamlit dashboard)