search

cake-contrib / Cake.AddinDiscoverer

Tool to aid with discovering information about Cake Addins
MIT License
5 stars 6 forks source link

Add blacklist for contributors #226

Closed pascalberger closed 11 months ago

pascalberger commented 11 months ago

We should implement a blacklist to filter out bots from contributors list.

Some bots I found:

Jericho commented 11 months ago

I just noticed that GitHub users (as returned from the GitHub API) have a property called "Type" which contains either the string User or Bot as you can see in this screenshot:

image

I could automatically filter out all bots based on the content of this "Type" property rather than manually maintain an exclusion list.

pascalberger commented 11 months ago

Would definitely make sense to filter based on available metadata. Not sure though if this catches all bots. I would expect for example cake-contrib-bot to be marked as user. For these cases we still can maintain a manual blacklist

Jericho commented 11 months ago

Publishing a new version of AddinDisco that automatically filters based on the "Type" property and also introduces an exclusion list:

https://github.com/cake-contrib/Cake.AddinDiscoverer/blob/develop/Source/Cake.AddinDiscoverer/exclusionlist.json#L74-L75

Jericho commented 11 months ago

AddinDisco was able to filter out 6 bots based on the GitHub user metadata in the latest PR.

Specifically:

but it did not detect the other 5 that you had listed:

@pascalberger Please accept the current PR and close the associated issue. In the mean time I will manually add these 5 to the exclusion list. When this is all done, AddinDisco will submit a new PR to remove the 5 excluded GitHub users..

pascalberger commented 11 months ago

@Jericho PR for website has been merged

Jericho commented 11 months ago

ok thanks. Triggering a new build of AddindDisco, we should get a PR to remove the excluded contributors in a few minutes

Jericho commented 11 months ago

Only 145 GitHub API requests left. Therefore skipping contributors files synchronization.

The latest build completed and unfortunately AddinDisco skipped the contributors file synchronization because we were about to trigger GitHub's abuse detection. I'll launch a new build later today.