[02:57:07] does WD have edit filters like WP does, for catching ultra-common types of vandalism [02:58:07] because if so i'd like to talk to someone who knows how to implement one [03:02:55] Dragonfly6-7 we have some filters, is there something specific you want to know how to implement? [03:03:42] vandalism to entries on names and surnames [03:03:47] changing how they display [03:03:57] i can show you a bunch of examples [03:05:11] okay, I don't have time to look into a new filter right now, but if you want to send the examples (here or PM) I can take a look when I have time unless someone beats me to it. [03:05:12] It would probably be added_lines like `{property for name/surname}\n{value}` and then you check the value [03:05:17] just off the top of my head [03:05:52] how many examples would you like [03:06:02] I'll put them in PM [03:06:09] idk, 5-10 should be enough [03:14:07] done [03:14:28] i find these when they show up in Commons' "wanted categories" page [17:23:30] What's the best way, via the API, to determine that this edit consists of adding a reference: https://www.wikidata.org/w/index.php?title=Q111372853&diff=next&oldid=1603415769 [17:24:44] Some edits make it unambiguous in the edit summary, via the `wbsetreference-add`. But edits like this just indicate 'claim changed' (`wbsetclaim-update`) in the edit summary. [17:28:32] I’m not sure there’s a great solution for that… I’d probably use action=compare with uselang=qqx and look for “(wikibase-diffview-reference)” in the output [17:28:34] https://www.wikidata.org/w/api.php?action=compare&format=json&uselang=qqx&fromrev=1603415769&torelative=next&prop=diff [17:28:41] or maybe there’s some library to compare entities client-side [17:29:42] Hmm... I want to do this at scale, in terms of calculating statistics for a large number of Wikidata edits to determine how many references were added. [17:30:20] ie, the wikidata stats on the Dashboard, like here: https://outreachdashboard.wmflabs.org/courses/WikiProject_Kerala/One_million_Malayalam_labels_(30_October_2020_-_October_29,_2021) [17:31:37] All these are currently based on edit summaries, but the 'refrences added' number is only from `wbsetreference-add` edits, which I'm now realizing will dramatically undercount them. [17:33:59] diff parsing sounds like a nightmare, but I could explore it if that's the only option. [17:39:05] the alternative would be to download the entity JSON of both revisions (Special:EntityData with ?revision=) and compare the references of each statement yourself, I think [17:50:50] I guess that would probably the right approach to capture any sort of multi-part edits accurately. [17:52:05] thanks Lucas_WMDE! [20:33:12] [Q] I'm reading this tutorial on using MWAPI: https://w.wiki/4RhQ One of the exmaples is: https://w.wiki/4zmZ It supposedly lists articles in Wikipedia with the word "cheese", but I'm wondering if (1) it lists those pages whose title contain the word "cheese" or (2) those pages whose title or body contain the word "cheese". Could anyone confirm if it's (1) or (2) and, if possible, provide more [20:33:15] information? [20:55:12] Looks like full text search, so the entire body: https://www.mediawiki.org/wiki/API:Search [20:56:14] Plus if you raise the LIMIT a bit you can see non-cheese entitled articles