Skip to content

Latest commit

 

History

History
109 lines (71 loc) · 3.29 KB

CONTRIBUTING.md

File metadata and controls

109 lines (71 loc) · 3.29 KB

Contributing to NexusAlign

Thank you ❤️ for your interest in contributing to NexusAlign!

Contributions of all kinds are welcome 👋 — including bug reports, new features, and suggestions.

Please read the following guidelines before contributing.

📚 Table of Contents

🐛 Reporting bugs

Before opening an issue:

  1. Search existing issues to avoid duplicates.
  2. Make sure you are using the latest version.
  3. Provide a minimal, reproducible example if reporting a bug.

A good issue should include:

  • A clear description of the problem.
  • Expected vs. actual behavior.
  • Environment details (Python version, torch version, OS, etc.).
  • A minimal code snippet or configuration (if applicable).

💡 Well-written issues save time for both maintainers and contributors.

🔀 Submitting Pull Requests

To increase the chance of your PR being reviewed and merged quickly:

  1. Open an issue first for non-trivial changes.
  2. Keep PRs focused — one logical change per PR.
  3. Write clear, descriptive commit messages.
  4. Ensure all tests pass.
  5. Run formatting and linting tools before submitting.

🛠️ Development Setup

pip install -e ".[dev]"

🧹 Code Style

We use ruff for formatting. Before committing, please run:

make format-check  # get format issue feedback
mkke format-fixup  # directly fix format issues

✅ Testing

If your PR introduces new functionality or changes behavior:

  1. Add appropriate tests.
  2. Ensure existing tests pass.
  3. Avoid breaking backward compatibility unless it has been discussed and agreed upon.

📖 Documentation

Good documentation is as valuable as good code.

Documentation contributions are highly appreciated, including:

  1. Fixing unclear explanations.
  2. Improving docstrings.
  3. Adding usage examples.
  4. Clarifying research assumptions or design choices.

🧪 Research Contributions

If you are contributing a new algorithm, model, or training method, please include:

  1. A short conceptual explanation.
  2. Links to relevant papers (if applicable).
  3. Design motivation and key trade-offs.
  4. Reproducible configuration (e.g., config files, commands, or scripts).

💡 We prefer implementations that are modular, well-documented, and easy to ablate.

🤝 Community Standards

Please follow our Code of Conduct in all interactions.

🙂 We aim to maintain a respectful, inclusive, and professional environment.

❓ Questions

If you're unsure whether a contribution aligns with the project direction, or you need guidance on where to start:

  • Feel free to open a discussion.
  • You can also start with smaller issues labeled as good first tasks (if available).

🚀 We deeply appreciate your interest in improving NexusAlign and look forward to collaborating with you.