Fast Standard Code Comment

A FastAPI-based web service that automatically generates standardized documentation comments for code blocks using a local LLM (Ollama). It detects the programming language, references the appropriate comment convention (JSDoc, TSDoc, PEP 257, etc.), and returns the original code enriched with professional inline documentation.

Tech Stack

• Python 3.14
• FastAPI + Uvicorn — REST API server
• LangChain + langchain-ollama — LLM agent orchestration
• Ollama — local LLM runtime (default model: qwen2.5:7b)
• codelang-detect — automatic programming language detection
• docling — web crawling for convention reference URLs
• Pydantic — request/response validation

How to Build and Run

1. Install Ollama and pull the model:

   ollama pull qwen2.5:7b

2. Install dependencies (using uv or pip):

   uv sync
   # or
   pip install -e .

3. Start the server:

   python main.py

   The server listens on http://127.0.0.1:8085.

4. Send a code block for commenting via POST /invoke:

   curl -X POST http://127.0.0.1:8085/invoke \
     -H "Content-Type: application/json" \
     -d '{"code_block": "def add(a, b):\n    return a + b"}'

   The response contains the original code with generated documentation comments.