Gather all the info your LLM needs to understand your repo in one command.
Create a virtual environment and then run:
pip install -r ./requirements.txt
Create a personal access token on Github and add it to your environment variables.
$env:GITHUB_TOKEN="your-token"
export GITHUB_TOKEN="your-token"
Run python main.py
and paste your target repo's GitHub URL when prompted:
python main.py https://github.com/username/repository
Your results will be saved as outputs/{REPO_NAME}_analysis.txt
To analyze a local directory, provide the path to the directory:
python main.py /path/to/your/directory
You can now analyze specific subdirectories within a local directory. Use the following syntax:
python main.py /path/to/base/directory [subdir1] [subdir2] [subdir3]
For example:
python main.py ../mycoolapp/cloud checkout functions admin
This will create separate analysis files for each subdirectory in the outputs
folder:
outputs/checkout_analysis.txt
outputs/functions_analysis.txt
outputs/admin_analysis.txt
Update the value of ADD_INSTRUCTIONS
in main.py
if you wish to add LLM instructions to the file.
The analysis output includes:
- README content
- Repository/directory structure
- File contents (with optional summarization for large files)
When analyzing GitHub repositories, ensure that your GITHUB_TOKEN
environment variable is set correctly.