Collaborative software development in Python

Dark Energy school: session on Software Engineering

Course preparation

0. Install Python version 3.5 or newer. Install the dependencies for this course using pip install -r requirements.txt. Run python check_setup.py to verify your installation.
1. Install an editor, such as Atom or VS Code
2. Ensure you can execute commands on the terminal / shell. E.g., commands such as python, git, and pytest should run.
3. Read How To Write Good Scientific Software.

Schedule

15:15–17:00

End with 15 mins. Q&A, if time allows.

Outcomes

General

• Discuss best practices for design & collaborative coding.
• How do you best organize a remote collaboration? Meetings, PRs, sprints, chat?
• Discuss classes vs functions; when is OOP needed in practice? Keep it simple—the truth about clever code. Premature optimization.
• How to package your code (vs just keeping it in a notebook)

Students will be able to do the following:

Git/GitHub

1. Make and navigate branches on Git
2. Make a git commit to a project
3. Submit your change as a pull request on GitHub
4. Review someone else's pull request

Modular APIs, testing

1. Write a function in Python
2. Write a docstring for that function
3. Write a test for that function + corner cases
4. Execute those tests locally
5. Have Travis-CI run these tests online
6. Develop your own Python package, installed using pip install -e ..

Reading from last year

Pre-lesson readings & post-lesson resources:

• Practical Lessons in Peer Code Review - Salsita Software blog http://blog.salsitasoft.com/practical-lessons-in-peer-code-review/
• Better Learning through Code Reviews - Capgemini Engineering blog https://capgemini.github.io/learning/better-learning-code-reviews/
• Why code reviews matter (and actually save time!) - Atlassian Agile Coach https://www.atlassian.com/agile/code-reviews
• 7 ways to up-level your code review skills - Asana blog https://blog.asana.com/2016/12/7-ways-to-uplevel-your-code-review-skills/

Resources

• SciPy lectures