AI Chatbot

A Flask-based AI chatbot application that uses OpenAI's GPT-4 model to provide intelligent responses while maintaining a knowledge base of previous interactions. The application includes features for user feedback and response streaming.

Features

• Interactive chat interface
• Integration with OpenAI's GPT-4 model
• PostgreSQL database for storing chat history and knowledge base
• Real-time response streaming
• User feedback system
• Docker support for easy deployment
• Custom data enrichment for API requests

Prerequisites

• Python 3.x
• PostgreSQL database
• OpenAI API key
• Docker (optional, for containerized deployment)

Environment Variables

The following environment variables need to be set:

DB_HOST=localhost
DB_NAME=chatdb
DB_USER=chatuser
DB_PASS=yourpassword
DB_PORT=5432
OPENAI_API_KEY=your_openai_api_key

Installation

1. Clone the repository:

git clone <repository-url>
cd aichatbot

2. Install dependencies:

pip install -r requirements.txt

3. Set up the PostgreSQL database and create the required tables:

# The tables will be automatically created when the application starts

Running the Application

Local Development

python app.py

Using Docker

docker build -t aichatbot .
docker run -p 5000:5000 aichatbot

Using Gunicorn (Production)

gunicorn -c gunicorn.conf.py app:app

API Endpoints

• GET /: Main chat interface
• POST /chat: Send a message and receive a response
• GET /chat: Stream response for a given question
• POST /feedback: Submit feedback for a chat response

Project Structure

.
├── app.py              # Main application file
├── Dockerfile          # Docker configuration
├── gunicorn.conf.py    # Gunicorn configuration
├── requirements.txt    # Python dependencies
├── static/            # Static files
└── templates/         # HTML templates

Database Models

Chat

• Stores chat history with questions, answers, and user feedback
• Fields: id, question, answer, feedback

KnowledgeBase

• Maintains a knowledge base of previous Q&A pairs
• Fields: id, question, answer

Contributing

1. Fork the repository
2. Create your feature branch
3. Commit your changes
4. Push to the branch
5. Create a new Pull Request

License

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