fastapi_ai_assistant

This repository contains a FastAPI application that demonstrates the OpenAI Assistants QuickStart guide, as shown in the official documentation. It implements the step-by-step instructions for creating an Assistant (including Code Interpreter), creating Threads for conversation, adding Messages, and streaming a Run using Server-Sent Events (SSE).

---

Overview

• OpenAI Assistants (Beta)
  This app uses the OpenAI Python Beta SDK to create, configure, and run a GPT-based Assistant.
• FastAPI
  We wrap the Assistants API logic in a FastAPI server to provide an HTTP interface, including SSE endpoints for real-time token streaming.
• Streaming Chat UI
  A simple front-end using Alpine.js & TailwindCSS to let users enter a prompt, see the streamed tokens in real time, and stop the generation if desired.

---

Key Steps from the QuickStart

This application follows the four major steps from OpenAI’s Assistants QuickStart:

1. Create an Assistant
   Configure a custom name, instructions, and attach any tools (e.g., Code Interpreter) to the GPT model.
2. Create a Thread
   Start a conversation “session” for the user.
3. Add Messages
   As the user talks to the Assistant, each message is stored in the Thread.
4. Run the Assistant
   Generate a response by calling the model (and any tools) on the entire conversation context. In this example, we use streaming to emit tokens to the client in real time.

---

Features

1. Real-Time Streaming
   Uses Server-Sent Events (SSE) via sse_starlette to stream tokens as they are generated by the Assistant.
2. Code Interpreter Tool
   Demonstrates how to enable the built-in code_interpreter tool for advanced math, scripting, and code execution.
3. FastAPI + Alpine.js
   Lightweight Python backend with a minimal, reactive JavaScript frontend.
4. Stop Generation
   A Stop button in the UI closes the SSE connection, immediately stopping further token streaming.

---

Installation

1. Clone the repository:

   git clone https://github.com/your-username/fastapi_ai_assistant.git
   cd fastapi_ai_assistant

2. Install dependencies (e.g., requirements.txt):

   pip install -r requirements.txt

3. Configure OpenAI (Beta) environment:

   • Set your OPENAI_API_KEY environment variable:

     export OPENAI_API_KEY="sk-..."

---

Usage

1. Run the FastAPI application:

   python main.py

   or

   uvicorn main:app --host 0.0.0.0 --port 8000

2. Navigate to:

   http://localhost:8000/
   

3. Chat with the AI Assistant:

   • Enter a message and press Send.
   • Watch as the response arrives token-by-token.
   • Click Stop to end the generation at any time.

---

Code Structure

fastapi_ai_assistant/
├── templates/
│   └── index.html        # Contains the chat UI (HTML, Alpine.js, TailwindCSS)
├── main.py               # FastAPI app with SSE and Assistants logic
├── README.md             # This file
└── requirements.txt      # Python dependencies (optional)

• main.py implements:

  • Assistant Creation: assistant = client.beta.assistants.create(...)
  • Thread Management: thread = client.beta.threads.create()
  • Message Creation: client.beta.threads.messages.create(...)
  • Run Streaming: Uses the runs.stream(...) context manager and a custom SSE event handler.
  • FastAPI Endpoints:
    • GET /: Serves the chat UI.
    • GET /chat-stream?prompt=...: Streams the AI response via SSE.

• index.html:

  • A Tailwind-styled chat interface with Alpine.js for reactivity.
  • Subscribes to the /chat-stream endpoint using EventSource.
  • Displays real-time tokens as they arrive and offers a Stop button.

---

Customization

• Model and Tools:
  Change the model parameter or add more tools in main.py to suit your needs.
  E.g., tools=[{"type": "code_interpreter"}, {"type": "another_tool"}].
• Instructions:
  Customize the Assistant’s global instructions (role: system) to define its personality or domain knowledge.
• UI/Styling:
  Modify index.html or integrate a different frontend approach.

---

Contributing

Feel free to open issues or submit pull requests if you have suggestions or improvements.

---

License

This project is provided under the MIT License.
See the LICENSE file for more details (if present).

---

Happy coding! Check out the OpenAI Assistants QuickStart docs for more information. If you have questions or run into any issues, please open an issue.