[07:54:11] what's wrong with the badges? I can't see them on Wikipedia anymore [07:57:06] ok found https://phabricator.wikimedia.org/T172592 [08:05:46] https://fi.wikipedia.org/wiki/Special:RecentChanges why there is no other languages anymore in the sidebar for the recent changes special page? https://fi.wikipedia.org/wiki/MediaWiki:Recentchangestext was not modified recently [08:28:54] Stryn: maybe removed with the recent changes (they have been working on that beta feature for months now)? [08:29:09] https://phabricator.wikimedia.org/T172461 [08:29:35] thanks [13:00:17] :/ https://www.wikidata.org/w/index.php?title=Q504867&type=revision&diff=532926150&oldid=532925963 [13:00:58] oh that is correct hmmm [13:01:23] where is my coffee [15:41:55] Dispatch going down nicely, in half since the last 12 hours. [15:42:52] The amount of changes is much lower than before, still 11 hours. Sitelinks? [16:04:27] https://grafana.wikimedia.org/dashboard/db/wikidata-dispatch?refresh=1m&orgId=1&from=now-3h&to=now :) [16:05:33] I wonder why this is happening – the change to increase dispatch batch size wasn’t merged yet as far as I’m aware [16:05:50] Is it some kind of cron job? [16:05:51] perhaps it’s just a different kind of changes now in the backlog [16:06:10] Ah. [16:06:50] Wonder how low we can get.... [16:24:16] Oh, there we go. [16:24:50] Now it can increase again. [16:27:46] https://www.wikidata.org/wiki/Special:Contributions/Mindsfer somebody who understands Russian should check those edits [16:38:14] looks like spam to me [19:34:32] Lucas_WMDE: regarding type constraints --- how many levels of P279 are tested by the checkConstraints gadget? [19:35:19] MisterSynergy: I don’t remember what the cutoff is, but after that, it falls back to SPARQL [19:35:28] but the fallback is broken in the version currently deployed on Wikidata [19:35:39] there’s already a fix, and it was even deployed for a short time last week [19:35:55] but then Wikidata had to be rolled back due to unrelated issues, so type constraints are broken again for the time being :/ [19:36:11] k thnx [19:36:13] :( [19:36:34] I oft see type constraint violations highlighted, although all claims are just fine [19:36:47] (I checked the config repo, the cutoff is 10 entities – total, not just depths) [19:36:56] typically this involves indirected subclassing (i.e. more than one level) [19:37:45] the task is https://phabricator.wikimedia.org/T169326 btw [19:38:09] does it go 10 entities on the first subclass it finds, or does it go one entity per subclass and continues until the target entity is found (or the counter is at 10)? [19:38:56] it does a depth-first search and counts all the entities it checks [19:39:03] could also be related to items that have more than one subclass; maybe the limit of 10 is consumed by a wrong subclass relation, which tests up to the "entity" item [19:39:12] yes, exactly [19:39:18] not good :-) [19:39:24] (I also have a longer comment with explanation in the task) [19:39:40] but if it is able to fallback to SPARQL, this should be okay [19:39:58] (assuming this comes back soon) [19:40:17] setting the limit to 0 would actually fix the type checks… at a cost of a lot more SPARQL queries [19:40:41] perhaps I’ll check if that’s okay – at least it wouldn’t require a config deployment, just a config change [19:41:47] a *code deployment, I mean [20:35:02] the free knowledge base with 31,006,315 data items that anyone can edit [20:39:13] WikidataFacts: I have two complex constraints now for checks on the value of the coordinates (for these two, that the latitude is below a certain value) and I suspect I'll be able to add at least one more soon... I was wondering if we could make a proper constraint for that, what do you think? [20:39:29] sjoerddebruin: did we miss the 30 millionth item between all the round Q numbers, or did I just not hear about it? [20:39:49] nikki: processed half the geonames thingies now [20:39:49] (for NL) [20:39:56] good :D [20:39:56] nikki: complex constraints are a SPARQL query that you run once and get all the violations, right? [20:40:03] WikidataFacts: i think you've missed it [20:40:12] k [20:40:16] sparql yes, but I think a bot updates the page daily [20:40:37] yeah, but that’s completely different from how the other constraints work [20:41:07] you can’t check a complex constraint on a single item [20:41:21] and you probably don’t want to run the query on every constraint check [20:41:22] I know, what I mean is that the check I'm doing is quite straightforward, it's a bit like a range constraint [20:42:25] so I'm wondering if making a new constraint "coordinate latitude is in range: x to y" (and the same for longitude) would be a good idea or not [20:42:46] I think once we have caching of constraint check results, complex constraints should be much more natural to implement [20:43:12] have a job that periodically runs the query and writes the results to the same cache as the other constraint check results, something like that [20:43:33] nikki: hm, what property would that be? [20:44:18] https://www.wikidata.org/wiki/Property_talk:P804 and https://www.wikidata.org/wiki/Property_talk:P3230 [20:45:05] so my latitude range would be -90 to -52.9 for the first one and -90 to -60 for the second [20:45:19] okay, so you’re not checking the lat/lon of the value, but of the value of a different statement (P625)… [20:45:25] yeah [20:46:03] (sorry, I'm quite worn out and probably being even more unclear than normal >_<) [20:55:33] I guess that would even be possible as part of the normal range constraint [20:55:39] oh? [20:55:49] we’d just need new properties for minimum/maximum latitude/longitude as qualifiers [20:55:56] and implement it, of course [20:56:10] https://www.wikidata.org/wiki/User:Pasleim/projectmerge/cebwiki-nlwiki :( [20:57:36] I could imagine it as qualifiers on a required statement constraint, not sure how it would work with the existing range constraint though [20:57:50] sjoerddebruin: fun :( [20:58:20] oh, right, the different property part is a bit weird [20:58:26] that’s not like a normal range constraint [20:58:30] yeah [20:59:34] on a related note, I have a potential constraint: items with a gnis antarctica id should have a scar id... but only if the latitude is below 60 S [20:59:47] I can't decide whether to do a complex constraint with no exceptions or a normal one with 953 exceptions :P [21:00:04] noooo please no constraint with 953 exceptions :D [21:00:22] (oh, the complex one would have a couple of exceptions, but nowhere near 953)