AI-powered tool for analyzing GitHub repositories, extracting build instructions from README files, and automating build execution. Supports multiple LLM backends including DSPy, OpenAI, and Ollama.
- Extracts build instructions from repository README files
- Executes automated dependency installation and build steps
- Compares repositories to identify missing files
- Continuous keyword-based scanning and auditing of public repositories
- Local LLM support via Ollama for offline analysis
- Python 3.10+
- GitHub API token
- OpenAI API key (optional, used as fallback)
- Ollama (optional, for local LLM inference)
git clone https://github.com/Jimmyu2foru18/github-analyzer.git
cd github-analyzer
pip install -r requirements.txtCopy config.yaml and fill in the required values:
GITHUB_TOKEN: "your_github_token"
OPENAI_API_KEY: "your_openai_api_key"
MODEL_NAME: "gpt-4.0-mini"
BASE_DIRECTORY: "github-repos"
LOG_DIRECTORY: "logs"
OLLAMA_BASE_URL: "http://localhost:11434/api"
OLLAMA_MODEL: "llama3"Environment variables override config file values:
GITHUB_TOKENOPENAI_API_KEYOLLAMA_BASE_URLOLLAMA_MODEL
python main.pyEnter a GitHub repository URL when prompted. Optionally enter a second repository URL to compare file structures.
python scanner_main.pyEnter a keyword to search for and audit relevant repositories.
| Module | Responsibility |
|---|---|
github_service.py |
GitHub API interactions: download, search, README extraction |
auto_builder.py |
Build automation and AI-driven analysis orchestration |
dspy_analyzer.py |
DSPy-based README analysis |
ollama_service.py |
Local LLM inference via Ollama |
scanner_service.py |
Automated scanning and auditing workflow |
config.py |
Configuration loading and validation |
logger.py |
Logging configuration |
- Read repository README
- Attempt DSPy analysis
- Fallback to Ollama if DSPy fails
- Fallback to OpenAI if Ollama fails
- Parse structured build instructions
- Execute dependency installation and build steps
MIT