Movie Recommendation System with Chat Interface

A web-based movie recommendation system that uses OpenAI's GPT, Nemo Guardrails, and SingleStore to provide personalized movie recommendations and engage in general conversation about movies.

Features

• Interactive chat interface for movie recommendations
• Natural language processing for understanding user queries
• Integration with SingleStore for efficient movie data storage and retrieval
• Real-time response with loading indicators
• Support for both movie-specific and general queries
• Caching system for improved performance

Prerequisites

• Python 3.8 or higher
• SingleStore database instance
• OpenAI API key
• Git (for cloning the repository)

Installation

1. Clone the repository:

git clone <repository-url>
cd singlestore-swarm-agent

2. Create and activate a virtual environment:

# On macOS/Linux
python -m venv venv
source venv/bin/activate

# On Windows
python -m venv venv
.\venv\Scripts\activate

3. Install the required packages:

pip install -r requirements.txt

4. Create a .env file in the project root with the following variables:

OPENAI_API_KEY=your_openai_api_key
SINGLESTORE_HOST=your_singlestore_host
SINGLESTORE_USER=your_singlestore_user
SINGLESTORE_PASSWORD=your_singlestore_password
SINGLESTORE_DATABASE=your_singlestore_database

Running the Application

1. Make sure your virtual environment is activated:

source venv/bin/activate  # On macOS/Linux
.\venv\Scripts\activate   # On Windows

2. Start the FastAPI server:

uvicorn app:app --reload

3. Open your web browser and navigate to:

http://localhost:8000

Project Structure

• app.py: FastAPI application and route handlers
• singlestore_swarm.py: Core logic for movie recommendations and database interactions
• templates/: HTML templates for the web interface
• nemo-configs/: Configuration files for Nemo Guardrails
• requirements.txt: Python package dependencies

Code Flow

1. User sends a message through the web interface
2. The message is processed by FastAPI and sent to the chat endpoint
3. Nemo Guardrails analyzes the query to determine if it's movie-related
4. If movie-related:
   • Query is processed by SingleStore for movie recommendations
   • Results are formatted and returned to the user
5. If general query:
   • Query is handled by OpenAI's GPT
   • Response is cached for future use
6. Response is sent back to the web interface

Troubleshooting

• If you encounter database connection issues, verify your SingleStore credentials in the .env file
• For OpenAI API errors, check if your API key is valid and has sufficient credits
• If the web interface is not loading, ensure all static files are properly served

Contributing

Feel free to submit issues and enhancement requests!

License

This project is licensed under the MIT License - see the LICENSE file for details.

Note: This project includes dependencies that may be licensed under different terms. Please refer to each dependency's documentation for their specific license terms.