450102 SE Open Source Software Development for Astronomy (2026W)

This practical seminar (inspired by the Code/Astro workshops, semaphorep.github.io/codeastro) enables students to develop and release an open-source software package for astronomical research. Working individually, participants take a small project from conception through to a public release, with the emphasis throughout on best software development practices rather than on producing any particular scientific result. The course is built around the principles of openness, inclusivity and reproducibility in research software.

Sessions combine short instructor demonstrations with hands-on development time. Core topics (and associated tools) include:

- Motivation: why and how research software becomes open source
- Version control and collaboration (git, GitHub)
- Environment and dependency management (uv)
- Testing (pytest)
- Code quality: linting, formatting, and type hints (ruff, Python typing)
- Documentation (sphinx)
- Continuous integration (GitHub Actions)
- Packaging: project structure and namespacing
- Publishing and citing a release
- Community building: contribution guidelines, codes of conduct, and welcoming new contributors

AI coding agents are demonstrated and their role in open-source workflows discussed, but their use is not required.

By the end of the course, students will be able to: use git and GitHub to version control and collaborate on a software project; manage Python environments and dependencies with modern tooling; write tests for a codebase; apply linting, formatting, and type hints to maintain code quality; write clear documentation; set up continuous integration; structure and package a piece of research software; and version, license, cite, and release it publicly. Students will also be able to critically evaluate the role of AI coding assistants in an open-source workflow, and will have practiced community-building skills such as writing contribution guidelines and welcoming new contributors.

Basic Python experience and familiarity with the command line are recommended. Roughly half the course uses Python-specific tooling; participants who prefer to work in another language may attend, but should expect to independently find and apply equivalent tools for those sessions throughout the semester.