Credits page and API call limitation

[BenoitAnastay]

Since there's a lot of new contributors to the HA project the current code struggle to fetch all data because of the API rate limitation.

We should find a workaround, maybe refactoring this code with more cached data to lower the number of API calls or make it run as a background service since in takes hours to update a credit page.

NB : the last update is my commit back in 2022

[domdomegg]

I would propose trying to avoid the as much as possible GitHub API given this seems like a big bottleneck. Instead, we could clone all the repos (filtering for just the metadata, not the actual content so it's quick), then parse this. We can then take the unique committer emails, and find their most recent name - and just add this to the credits.

We'd lose the GitHub profile links, but maybe this is okay?

Alternatively we could then use the GitHub API just to search up emails - which I think might require fewer API requests total so we'd have more headroom on the API limit (but I think would still require multiple hours, just fewer hours 😅). It would also be a lot easier to cache.

[kylehakala]

What about a hybrid of a couple things where the existing list of credited users could be cached into a GitHub Actions cache file (which could be referenced by future GH action workflows) and then only commits made “since the last cache file was created” are processed and then compared against the list of cached users?