Automatic Lab Notebook (ALN)

Automatically generate structured, AI-enhanced summaries of Jupyter Notebook changes on each commit. Designed to improve research reproducibility, documentation, and collaboration with minimal manual effort.

---

Features

• Automatic Summarization: Uses LLMs to compare notebook versions and generate concise summaries.
• Version Comparison: Detects and compares .ipynb changes between commits.
• New Notebook Detection: Gracefully handles newly created notebooks by summarizing content.
• Improvement Suggestions: Provides suggestions to enhance code readability, structure, or clarity.
• GitHub Actions Integration: Fully automated pipeline runs on notebook updates.
• Artifact Output: Summaries saved as markdown files and uploaded as artifacts.

---

Quick Start

1. Fork or Clone this repository.

2. Add your API Key (e.g., Gemini API):

   • Store the key as a GitHub secret:
     • Go to Settings → Secrets → Actions.
     • Add a new secret named: GEMINI_API_KEY.

3. Use the correct folder structure:

   • Your Jupyter notebooks should live in the root or subdirectories.
   • The ALN workflow will detect and summarize any notebook changed in main.

4. Push a Commit with a .ipynb file:

   • The GitHub Action will:
     • Compare notebook changes
     • Generate summaries
     • Upload markdown artifacts under aln_output/

---

Tech Stack

• Python 3.9
• google-generativeai (Gemini API)
• nbconvert
• pytz
• git, bash, and GitHub Actions

---

License

This project is licensed under the MIT License.

---

Acknowledgements

• Gemini by Google
• GitHub Actions

---

Contributing

Pull requests and issues are welcome. If you have ideas to improve accuracy, formatting, or add support for more file types, feel free to open a PR.