autodg 1.0.0

A modular Python API analyzer for Flask, FastAPI, Django, and DRF with AI-powered documentation generation

API Analyzer

A modular, extensible Python API analyzer that scans codebases and generates comprehensive documentation with AI-powered explanations.

Features

• Multi-Framework Support: Flask, FastAPI, Django, Django REST Framework
• AI-Powered Documentation: Supports Ollama, OpenAI, Claude, and Gemini
• Call Graph Analysis: Visualizes function dependencies
• Database Operation Detection: Identifies DB queries and operations
• MermaidJS Diagrams: Auto-generated architecture diagrams
• File-wise Documentation: AI overviews for every Python file
• Progress Tracking: Real-time progress with ETA

Installation

pip install autodg

Optional Dependencies (AI Features)

Note: AI-powered documentation generation is an experimental feature. For the best local experience, we recommend running Ollama with the llama3.1 model.

To enable support for specific cloud LLM providers, install the package with the corresponding extras:

# OpenAI Support
pip install autodg[openai]

# Claude Support
pip install autodg[claude]

# Gemini Support
pip install autodg[gemini]

# Support for all providers
pip install autodg[all]

Quick Start

autodg --paths /path/to/your/project --output ./docs

Configuration

Create a config.yaml file:

llm:
  provider: ollama  # Options: ollama, openai, claude, gemini
  
  ollama:
    host: http://localhost:11434
    model: llama3.1

Usage

# Basic usage
autodg --paths /path/to/project --output ./output

# With LLM explanations
autodg --paths /path/to/project --output ./output --ollama True

Supported Frameworks

• Flask: Routes, Blueprints
• FastAPI: APIRouter, nested routers, class-based routes
• Django: path(), re_path(), FBV, CBV
• DRF: Routers, ViewSets

Output

The analyzer generates:

• request_docs/: Documentation for each API endpoint (optional ai explanations)
• files/: AI-powered overviews of each Python file (optional ai explanations)

License

MIT License - see LICENSE file for details.

Author

Dhinagaran S ([e-mail])