🦦 fastapi-gpt-websocket-http-stream-chat 🦦

A Proof of Concept (PoC) that leverages websockets and HTTP streaming to relay OpenAI's GPT chat completions in real-time using the FastAPI framework. While FastAPI is used here, any API framework supporting websockets and HTTP streaming can be adapted for this purpose.

Local Development Setup 🛠

If you prefer to set up your local environment without using Docker, you can follow these steps:

1. Create a Virtual Environment:

   python3 -m venv .venv

   Activate the Virtual Environment:

   • For macOS and Linux:

     source .venv/bin/activate

   • For Windows:

     .venv\Scripts\activate

2. Install the Dependencies:

   Ensure you're inside the project directory and your virtual environment is activated.

   pip install -r requirements.txt

3. Run the Application:

   uvicorn main:app --reload

   This will start the FastAPI application with hot-reloading enabled. You can access the application at http://localhost:8000.

Docker Deployment 🐳

Production Setup 🚀

1. Build the Docker Image:

   From the project root, execute:

   docker build -t app .

2. Run the Container:

   docker run -d -p 8000:8000 app

   The application will be available at: http://localhost:8000.

Development Setup 💻

FastAPI development is made smoother with hot-reloading. I have added a dedicated Dockerfile.dev for this.

1. Build the Development Image:

   From the project root, execute:

   docker build -f Dockerfile.dev -t app-dev .

2. Run the Container with Volume Mount:

   This mounts the current project directory to the container, allowing you to observe changes in real-time:

   docker run -p 8000:8000 -v $(pwd):/app app-dev

   Navigate to http://localhost:8000. Any source code modifications will instantly reflect in the application.