dbschenker / contribution-checker Goto Github PK

Provide a JSON schema that

• explains the output
• verifies the syntax
• allows others to create data structures from that

It would be great to have unit tests, e.g. with pytest.

The setup could be to create a mock Git repo, create some commits from different authors, and check the output against the JSON we'd expect.

To get a better overview about a set of repositories we need a way to specify multiple repositories. I would suggest an option -f, --file, which takes the name of a configuration file which contains all the information needed to do the analysis. This would be the expression for filtering by author email and a list of repository URLs.

To make this efficient, we might want to cache repositories at a defined location, so that the tool only transfers the minimal amount of data and doesn't check out all repositories from scratch with every invocation (see #4).

The report should then show a summary instead of the results for an individual repository.

For users, it may be helpful to have a package available on PyPI. With this action, it can be easily automated.

Remote repositories should be cached somehow, so that they don't need to be checked out again with each invocation of the checker. We could check them out to a well-defined space (e.g. ~/.local/state/contribution-checker/repos) and then just run a git pull if the checkout already exists.

https://github.com/platformdirs/platformdirs could be used to get the corresponding cache directory for all systems.

An important information can be how active we are in a project in relation to overall activity. It would be good to show that. One way would be to include commits by others in the plot of commits (via --plot).