Page MenuHomePhabricator

ErfgoedBot categorisation is broken for ir (fa)
Closed, ResolvedPublic

Event Timeline

Question: how easily reversible/recoverable is it when the categorisation goes wrong. I.e. must we manually revert this or is it enought to fix the base cause (since a similar thing happened to Kosovo).

@JeanFred what Lokal_Profil asked, plus, do you have any idea what may be happening here? :)

Is this possibility due to the "broken file links" cat being on one of the lists and the bot then adding the cat due to it being mapped to the commons equivalent on Wikidata?

This might be fixed by blacklisting any Wikimedia administration categories, P31:Q15647814.

This might be fixed by blacklisting any Wikimedia administration categories, P31:Q15647814.

The categorisation bot could possibly even populate such a list on startup by making the appropriate sparql query?

So this is still happening :-( https://commons.wikimedia.org/w/index.php?diff=207277931

Question: how easily reversible/recoverable is it when the categorisation goes wrong. I.e. must we manually revert this or is it enought to fix the base cause (since a similar thing happened to Kosovo).

No easy way I’m afraid :-( Here it’s not too bad as we can easily move things back from this fairly empty wrong category with Cat-a-lot or smth.

Well at least now we know the likely cause: https://commons.wikimedia.org/w/index.php?diff=207581449, i.e. one of the upper categories (categories in which the lists are) likely contains the "broken file links" category.

@LilyOfTheWest: Could you please confirm that https://fa.wikipedia.org/wiki/رده:لیست_یادمان‌ها_در_ایران is in a category like "Broken file links" (or is it possibly https://fa.wikipedia.org/wiki/رده:صفحه‌های_دارای_پیوند_خراب_به_پرونده) ?

@JeanFred: This is likely to happen whenever anyone edits the list and puts an erroneous filename. The maintenance category might then itself be mapped via Wikidata. I would say we either blacklist this particular one or go further and skip any categories which are marked as hidden on the wikipedia (should be able to get that info via pywikibot)