README.md

metadata

title: OpenTriage AI Engine
emoji: 🧠
colorFrom: purple
colorTo: pink
sdk: docker
app_port: 7860
pinned: false

🧠 OpenTriage AI Engine (Microservice)

This is the AI Brain of the OpenTriage platform. It is a specialized Python microservice acting as a "Sidecar" to the main TypeScript backend.

It hosts the heavy-lifting AI logic, including RAG (Retrieval-Augmented Generation), Issue Triage, and Mentor Matching, keeping the main application fast and lightweight.

🚀 Features

• 🔍 AI Triage: Automatically classifies GitHub issues by complexity and type.
• 📚 RAG Chatbot: "Chat with Repo" functionality using vector search.
• 🤝 Mentor Match: Connects contributors to mentors based on tech stack analysis.
• 🎉 Hype Generator: Generates celebratory messages for PR merges.

---

🔗 How to Connect

This service exposes a REST API via FastAPI. It is designed to be called by the Main Backend, not directly by users.

Base URL

Your API is live at: https://[YOUR_USERNAME]-opentriage-ai-engine.hf.space

(Check the "Embed this space" menu in the top right to get your exact direct URL)

API Endpoints

Method	Endpoint	Description
GET	/health	Server health check (returns 200 OK)
POST	/triage	Classifies an issue description
POST	/chat	General AI assistant response
POST	/rag/chat	Context-aware repository Q&A
POST	/mentor-match	Finds best matching mentors

---

🛠 Deployment Configuration

1. Environment Variables (Secrets)

You must set these in the Settings tab of this Space under "Variables and secrets":

Secret Name	Required?	Description
OPENROUTER_API_KEY	Yes	Required for all AI generation
MONGO_URL	Optional	If your RAG needs persistent vector storage

2. Docker Configuration

This Space uses a custom Dockerfile to ensure all Python scientific libraries (NumPy, Scikit-learn, LangChain) are installed in a compatible environment.

• Port: 7860 (Standard HF Space port)
• User: Runs as non-root user user (ID 1000) for security.

---

💻 Local Development

If you want to run this brain locally:

# 1. Create venv
python -m venv venv
source venv/bin/activate

# 2. Install
pip install -r requirements.txt

# 3. Run
uvicorn main:app --reload --port 8000