GitHub

Session 5

September 17, 2024

Review & Homework

• If you turned an existing project into a git repo, how did it go?

• Ways to create a git repo:

  • run git init from the command line inside of project directory
  • create a new RStudio project with “use git” checked
  • run usethis::use_git() in the R console in a RStudio project

Today’s goals

1. Practice using git with RStudio
2. Explore commit history
3. Put a local repo on GitHub
4. Use git push and git pull to synchronize with GitHub

Working with git in RStudio (Live Coding)

The most common git commands can be used with the RStudio GUI.

Demo save, add, commit cycle with Git pane AND git menu in RStudio

Sharing with GitHub

• All contributors have a local copy of the repo.
• You might also have a copy in multiple places (on multiple computers, on university HPC).
• GitHub has a central repo that allows you to synchronize and merge work from multiple sources.
• You can also access the repo through github.com web UI

Figure 2 from (Bryan 2018)

Connect With GitHub (Live Coding)

• Set up your computer to “talk” to GitHub with steps in git_sitrep()
• Connect your local git repo to a remote GitHub repo with use_github()
• Explore the GitHub interface

Demo “starting from scratch” by running gitcreds::gitcreds_delete(). Remember to delete the PAT you create and show on camera.

Remind learners that repos can be private. use_github(private = TRUE)

Work With RStudio and GitHub (Live Coding)

• Push local changes to GitHub
• Pull GitHub changes to local

Demo save, add, commit, push and check GitHub for changes. Then, edit README on github and pull changes.

Ignoring files

• If you enabled git with Rstudio or usethis, a .gitignore file was added for you, explore the contents
• .gitignore also prevents files from being synced to GitHub
  • Use to keep large and non “diff-able” files local only to avoid GitHub file size limits
  • Use to prevent sensitive information from getting on GitHub (e.g. sensitive data, API tokens)
  • Cons: if ignored files are inputs, less reproducible and no backup on GitHub
• If a file is already tracked by git, you can’t (easily) retroactively remove it from git history and GitHub.

Remind learners that it is best practice to consider outputs as disposable, so might be good to .gitignore them if they are non-diff-able like .png files or .pdfs generated from Quarto. Can use .gitkeep to get git to track an empty folder.

Resources

• #git channel in UA Data Science Slack

• happygitwithr.com

• usethis::use_github()

Your Tasks

• Put your research compendium on GitHub (e.g. with use_github()) and share the link in our Slack channel (if it can be made public)
• Explore other research compendia shared in Slack

If there is extra time, encourage learners to use it to practice these new skills with a real project repo while you’re all still there to help troubleshoot.

References

Bryan, Jennifer. 2018. “Excuse Me, Do You Have a Moment to Talk about Version Control?” The American Statistician 72 (1): 2027. https://doi.org/10.1080/00031305.2017.1399928.