[00:41:31] -ops [00:41:31] Se ha avisado a los operadores del canal registrados para atender usuarios. Por favor, espera hasta ser atendido/a. [08:24:49] Module:Lexeme: [08:24:50] -- Note: All functions here are hacks [08:35:20] :') [08:36:01] 100 million qualifiers soon https://grafana.wikimedia.org/dashboard/db/wikidata-datamodel-statements?refresh=30m&panelId=11&fullscreen&orgId=1&from=now-90d&to=now [09:13:31] The Query Service is starting to look slick btw ;) [09:47:56] Hallo [09:48:04] Hello. [09:48:28] Something is suspicious about the message https://translatewiki.net/w/i.php?title=MediaWiki:Apihelp-wbleditsenseelements-example-1/he&action=edit [09:48:38] The English text is: [09:48:39] Edits Sense with the ID $1 to have glosses $2 in languages $3 respectively [09:49:02] The message documentation (qqq) lists parameters in a way that seems wrong. [09:49:26] those parameters seem random [09:49:37] I think that it shows the default API message parameters, and not what is actually required in the message. [09:50:56] ... one of my favorite things to do: complaining about qqq in MediaWiki extensions on IRC ... [09:50:58] And then the person who added that API leaves :P [09:51:13] https://phabricator.wikimedia.org/T199206 [09:51:15] Hey... that's a clue :) [10:06:06] sjoerddebruin: I think I fixed it: https://translatewiki.net/wiki/MediaWiki:Apihelp-wbleditsenseelements-example-1/qqq . But please make a sanity check. [11:16:17] Hallo. [11:16:23] Another qqq question about Lexeme. [11:16:43] Wikibase-lexeme-summary-update-sense-glosses doesn't have useful qqq. [11:17:30] To fix it, just edit https://translatewiki.net/wiki/MediaWiki:Wikibase-lexeme-summary-update-sense-glosses/qqq [11:18:04] The {{Identical}} template is not useful there, and the parameter is not documented. [11:18:04] 10[2] 04https://www.wikidata.org/wiki/Template:Identical [11:20:07] leszek_wmde: ^ [11:22:11] aharoni: not nice [11:22:13] looking [11:22:23] thanks [11:24:21] aharoni: what would be the right thing to do, as in practice currently wikibase-lexeme-summary-update-sense-glosses and wikibase-lexeme-summary-update-sense-elements are supposed to emit the same message? [11:24:50] aharoni: should I just copy over the message to another? [11:24:53] looking... [11:25:29] aharoni: oh wait. just noticed that that "show identical messages" just given me a red link [11:25:37] so that is the part to fix [11:28:10] leszek_wmde: if the message is the same and the parameter has the same semantics, you can copy the qqq [11:28:32] and you don't need to do it in Gerrit. You can simply edit the corresponding translatewiki.net pages. [11:28:54] aharoni: got it. so no sole "Identical|foobar" as qqq content? [11:29:24] aharoni: out for a meeting now. will be fixing after it [11:29:30] I can do it :) [11:30:11] The Identical template can be there, but it's good to copy the whole qqq string. It's a bit unintuitive, but in localization, repetition is not necessarily a bad thing. [11:32:27] leszek_wmde: done, thanks for the hints [12:12:51] aharoni: oh hi :) you never responded again in https://phabricator.wikimedia.org/T144272, I'm wondering if those things would be acceptable in the language-data database you mentioned [12:14:36] nikki: oh, sorry, I missed that Phab comment. Let me think about it. [12:16:14] ok :) [13:23:46] any news on allowing trusted users longer query timeouts? Here's a query I can't get to run, even limiting to 100 results. [13:23:48] http://tinyurl.com/ycnpfrra [13:26:20] minus doesn't seem like the right choice there [13:27:01] that tells it to find *all* items without a hebrew description, independent of the surrounding query [13:27:31] nikki: oh! you're right! [13:29:36] filter not exists might work better, but I usually find optional and filter (!bound(...)) work best [13:30:17] it seems like even that is too difficult though :( [13:31:32] you might also be interested in https://phabricator.wikimedia.org/T167361 [13:31:56] not being able to select labels/descriptions for a particular language is a problem I keep having :/ [13:32:56] nikki: yeah. :-/ [13:33:00] FILTER NOT EXISTS would work here (though in practice it's still to slow) [13:33:14] I can run queries manually that are slightly over a minute [13:33:23] but I think this one might be to heavy [13:35:50] I'm running it (unrestricted) with LIMIT 1 now… let's see how this goes [13:35:53] Wonder if a separate small replicated/clone server could be run where fancypants users can query all they want without it disturbing others [13:35:54] but probably not to well [13:36:20] I mean, I'd be happy to wait for someone else's 2 min query to finish, most of the time, if I want something extreme myself [13:36:48] item [13:36:49] [13:36:53] After 64s [13:37:31] adding minus { ?item schema:description "מאמר מדעי"@he } might help [13:38:08] If that description is widely used it might [13:38:19] it is, that's why I suggested it :) [13:38:25] LIMIT 2 takes 94s already [13:38:58] hah, that works very well [13:39:02] nice, nikki :) [13:39:10] :D [13:39:21] (I used it with FILTER NOT EXISTS, though) [13:40:50] if my experience is anything to go by, minus is probably slightly faster for that, but what sparql likes or doesn't like is mostly a mystery to me [13:42:09] like I still find it puzzling how combining two very fast selections (in the order of milliseconds) suddenly becomes unfathomably difficult for it [13:43:37] Well, suddenly having to collect large temporal results in memory and mangling them together later can be very heavy [13:44:42] * temporary [13:45:02] I don't remember any examples off the top of my head, but I think I've seen that happen even when the queries I'm trying to combine don't have that many results [13:46:38] I can usually get around it by using with {} as %something for one or both of them though [13:46:43] In that case blazegraph might just be stupid [13:47:00] that's my personal opinion :D [13:47:02] yeah, if that works easily, the optimizer is probably to blame [14:00:44] excluding the specific common Hebrew label does speed things up. But what if I do want to look for items without any Hebrew label? What's an effective syntax? [14:00:59] Technical Advice IRC meeting starting in 60 minutes in channel #wikimedia-tech, hosts: @leszek_wmde & @jakob_WMDE - all questions welcome, more infos: https://www.mediawiki.org/wiki/Technical_Advice_IRC_Meeting [14:14:07] abartov: You can do both… exclude the common label and all hebrew ones [14:14:12] that will speed things up [14:14:26] I tested it, but sadly I don't have the query saved :S [14:14:56] hoo: Thanks, but I'm not sure about the syntax for 'no value' with a language tag. []@he? [14:15:46] no wait, that just means I don't care what the value is. [14:18:16] http://tinyurl.com/y7hljhga That works [14:18:21] there might be faster ways to do it [14:50:48] Technical Advice IRC meeting starting in 10 minutes in channel #wikimedia-tech, hosts: @leszek_wmde & @jakob_WMDE - all questions welcome, more infos: https://www.mediawiki.org/wiki/Technical_Advice_IRC_Meeting [17:18:18] I made a topic on https://www.mediawiki.org/wiki/Talk:Phabricator/Help for my email problem, perhaps they can help me :s [17:27:46] Some properties in https://www.wikidata.org/wiki/Wikidata:Property_proposal/Natural_science are "ready" is anyone around with the create right [18:36:54] Search acting weird? [22:00:48] Technical Advice IRC meeting starting in 60 minutes in channel #wikimedia-tech, hosts: @CFisch_WMDE & @chiborg - all questions welcome, more infos: https://www.mediawiki.org/wiki/Technical_Advice_IRC_Meeting [22:51:06] Technical Advice IRC meeting starting in 10 minutes in channel #wikimedia-tech, hosts: @CFisch_WMDE & @chiborg - all questions welcome, more infos: https://www.mediawiki.org/wiki/Technical_Advice_IRC_Meeting