Customer Support Assistant

An AI-powered customer support system designed to streamline ticket management and enhance user interactions through intelligent automation.

🚀 Features

• User Management: Secure registration, authentication, and role-based access control.
• Ticketing System: Create, view, and manage support tickets efficiently.
• Messaging: Facilitate communication between users and support agents within tickets.
• AI Integration: Leverage Groq's LLaMA 3 model to generate context-aware responses.
• Token Management: Issue and revoke access tokens with expiration handling.
• Pagination: Efficient data retrieval with paginated endpoints for scalability.
• Feedback Loop: Reply to AI-generated messages with human responses and feed them back to the model.

🧰 Technologies Used

• Backend: Python, FastAPI
• Database: PostgreSQL, SQLAlchemy
• ORM: SQLAlchemy with scoped sessions
• AI Model: Groq's LLaMA 3
• Authentication: JWT Tokens
• Environment Management: Poetry

📦 Installation

1. Clone the repository:

   git clone https://github.com/RusabKhan/customer_support_assistant.git
   cd customer_support_assistant

2. Install Poetry (if not already installed):

   curl -sSL https://install.python-poetry.org | python3 -

3. Install dependencies:

   poetry install

4. Activate the virtual environment:

   poetry shell

5. Configure environment variables:

   Create a .env file in the root directory:

   DB_DIALECT=postgresql
   DB_USER=your_db_username
   DB_PASSWORD=your_db_password
   DB_HOST=localhost
   DB_PORT=5432
   DB_NAME=your_db_name
   GROQ_API_KEY=your_groq_api_key
   SECRET_KEY=your_secret_key

6. Run the application:

   uvicorn main:app --reload

✅ Note: Alembic migrations are triggered automatically when you start the FastAPI app, so you don’t need to run them manually.

🗃️ Database Management

Database sessions are managed using a custom DB utility class found in src/utils/db.py. This class uses the with statement to ensure proper session management:

• Commits transactions on success.
• Rolls back transactions on exception.
• Automatically flushes and disposes the session on exit.

This ensures atomicity and clean handling of transactions without leaking sessions or failing silently.

🧪 API Endpoints

• User Endpoints:

  • POST /users/ – Register a new user.
  • POST /login/ – Authenticate and receive a JWT token.

• Ticket Endpoints:

  • POST /tickets/ – Create a new support ticket.
  • GET /tickets/ – Retrieve a list of tickets with pagination.
  • GET /tickets/{ticket_id}/ – Retrieve a specific ticket.

• Message Endpoints:

  • POST /tickets/{ticket_id}/messages/ – Add a message to a ticket.
  • GET /tickets/{ticket_id}/messages/ – Retrieve messages for a ticket.

• AI Integration:

  • GET /tickets/{ticket_id}/ai-response/ – Generate an AI response for a ticket.
  • POST /tickets/{ticket_id}/ai-feedback/ – Submit feedback or a follow-up to the AI-generated response.

🐳 Docker Support

To run the project in Docker:

docker-compose up --build

Then go to http://localhost:8000

🤝 Contributing

Contributions are welcome! Please fork the repository and submit a pull request for any enhancements or bug fixes.

🪪 License

This project is licensed under the MIT License.