Support multiple "open source" hosts

[andrew]

Hosts like https://opencollective.com/numfocus don't currently get indexed as they aren't part of the opensource host but we'd like to see them if possible.

Could do with a list of known "open source" hosts from @poohlaga to test with

[gusaus]

Does this issue (with Open Collective's Discovery platform) seem related? opencollective/opencollective#6373

If yes, could the solution described in this comment help resolve? opencollective/opencollective#6373 (comment)