Hi! > repository. They can use javascript and connect to the github api > however, so it might be possible to build a useful browser interface > that filtered based on the '-' separated name components. You could at
Exactly! I've (ab)used github API for a bit, and found this (please correct me if the figures look weird, I trusted API passed through several quick-n-dirty scripts so mistakes possible): - we have 1780 repos - 900 of them are extensions, so putting extensions in separate group cuts it by half - 102 repos do not have "-" in their name, meaning those would be hardest to classify - 1016 have "mediawiki" prefix - 243 are "operations" - 76 are "analytics" - 48 are "labs" - 20 are "pywikibot" - 18 are "thumbor" - 18 are "integration" - 15 are "apps" - 10 are "phabricator" The rest of the groups are smaller or don't look like a good unifying principle immediately. Thus, even with very naive approach we could categorize at least 1464 or over 80% of the repos, leaving just about 300 non-obvious ones, of which probably about 100-200 will end up in "other". Not ideal, but at least better than 1780 :) The list of names is at https://gist.github.com/smalyshev/dfa72c79d9f750058262b902f47c0130 if anybody wants to play with it and save time on extracting it :) -- Stas Malyshev [email protected] _______________________________________________ Wikitech-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/wikitech-l
