Restaurant Recommender Web Application

A Flask-based web application that recommends restaurants based on user preferences using machine learning. The application uses a KMeans clustering algorithm to provide personalized restaurant recommendations based on city, cuisine, country, cost, and rating preferences.

Features

• User-friendly web interface
• ML-based restaurant recommendations using KMeans clustering
• Results displayed as an interactive table
• Docker support for easy deployment
• Personalized recommendations based on multiple criteria

Project Structure

/app
  ├── app.py                # Flask application
  ├── model_utiles.py       # ML model and recommendation logic
  ├── templates/
  │    ├── index.html       # Input form
  │    └── results.html     # Recommendation results
  ├── requirements.txt      # Python dependencies
├── Dockerfile             # Docker image setup
├── docker-compose.yml     # Docker Compose config
└── README.md             # Project documentation

Workflow

User Input (City, Cuisine, Country, Cost, Rating)
   ↓
Feature Encoding (Scaler + OneHot Encoding)
   ↓
KMeans Clustering → Predict User Cluster
   ↓
Filter Restaurants in Same Cluster
   ↓
Compute Distance to User Vector
   ↓
Rank by Nearest Distance
   ↓
Return Top-N Recommendations

Requirements

• Python 3.8+
• Flask
• pandas
• scikit-learn
• Docker (optional, for containerized deployment)
• Docker Compose (optional)

Installation & Setup

Option 1: Run Locally

1. Clone the repository:

   git clone <repository-url>
   cd restaurant-recommender

2. Install dependencies:

   pip install -r requirements.txt

3. Run the Flask application:

   python app.py

4. Access the application:

   • Open your browser and navigate to the address shown in the Flask console output
   • Default: http://localhost:5000 (or the port specified in your Flask app)

Option 2: Run with Docker

1. Build the Docker image:

   docker build -t restaurant-recommender:latest .

2. Run the Docker container:

   docker run -p 5000:5000 restaurant-recommender:latest

3. Access the application:

   • Open your browser at http://localhost:5000

Option 3: Run with Docker Compose

1. Start the application:

   docker-compose up -d

2. Access the application:

   • Open your browser at the port specified in docker-compose.yml
   • Check the compose file for the exact port mapping

3. Stop the application:

   docker-compose down

Usage

1. Enter your preferences:

   • Navigate to the home page
   • Fill in the form with your preferences:
     • City
     • Cuisine type
     • Country
     • Cost range
     • Minimum rating

2. Get recommendations:

   • Click the submit button
   • View personalized restaurant recommendations based on your inputs

3. Refine search:

   • Navigate back to adjust your preferences
   • Submit again for new recommendations

Configuration

• Port Configuration: Modify the port in app.py or docker-compose.yml as needed
• Model Parameters: Adjust clustering parameters in model_utiles.py
• Debug Mode: Currently enabled in Flask; disable for production deployment

Important Notes

Before deployment:

• Ensure model_utiles.py correctly loads your ML model and restaurant dataset
• Disable Flask debug mode for production environments
• Set appropriate environment variables for production
• Configure proper logging and error handling

Docker considerations:

• Modify Dockerfile if additional services or dependencies are required
• Update docker-compose.yml for multi-service deployments
• Ensure data persistence by configuring volumes if needed

Development

To contribute or modify the application:

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Test thoroughly (local and Docker environments)
5. Submit a pull request

Troubleshooting

• Port already in use: Change the port in Flask app or docker-compose configuration
• Model loading errors: Verify the model file path and format in model_utiles.py
• Missing dependencies: Run pip install -r requirements.txt again
• Docker build fails: Check Dockerfile syntax and ensure all required files exist