[08:12:56] <richelle>	 @meanwhile the first place where we were going through some sort of thing I want
[08:15:34] <richelle>	 maybe I should get out
[08:15:56] <richelle>	 !
[12:13:16] <Pyb>	 French Wikisource starts using Wikidata for the header template on Author pages :) E.g. https://fr.wikisource.org/wiki/Auteur:Henri_Mouhot
[12:15:28] <tseppelt>	 Pyb: this is great!+
[12:21:29] <Nemo_bis>	 well done Pyb 
[12:22:52] <hoo>	 Do we have the number of main snaks that are wikibase-items?
[12:23:06] <hoo>	 addshore: ^
[12:23:34] <Nemo_bis>	 Added to https://fr.wikisource.org/wiki/Cat%C3%A9gorie:Mod%C3%A8les_utilisant_les_donn%C3%A9es_wikidata
[12:23:51] <addshore>	 hoo: not that I know of
[12:24:16] <hoo>	 addshore: mh... do you have a rough estimate at hands?
[12:24:18] <addshore>	 oh hoo, a new one https://grafana.wikimedia.org/dashboard/db/wikidata-edits
[12:24:28] <addshore>	 hoo: I dont even have an estimate :/
[12:25:12] <addshore>	 hoo: I could work it out from finding all properties that have datatype item and then counting all items that have claims with that prop
[12:25:22] <hoo>	 hm
[12:25:44] <hoo>	 I'm just trying to figure memory requirements for something
[12:26:00] <hoo>	 I'll calculate with 100M then to play it safe
[12:26:02] <addshore>	 hmmm
[12:26:04] <hoo>	 it'll grow anyway
[12:26:17] <hoo>	 that'll probably give me a hear attack, though
[12:26:34] <addshore>	 :D
[12:27:01] <addshore>	 670 properties are of item type
[12:29:52] <addshore>	 hoo: how if your sparql foo?
[12:30:10] <hoo>	 I could try that
[12:30:20] <addshore>	 this lists all properties that are items SELECT ?s WHERE {?s wikibase:propertyType wikibase:WikibaseItem}
[12:30:30] <addshore>	 this lists all statements that use a given proeprty SELECT (count(distinct(?s)) AS ?scount) WHERE {?s wdt:P373 ?o}
[12:30:54] <addshore>	 just got to combine the two right ;) in SQL I would do a WHERE IN ( stuff ), just gotta see if thats possible in sparql
[12:31:04] <addshore>	 I have a feeling it may not work though ;)
[12:31:45] <hoo>	 I'm just going to guess
[12:31:49] <addshore>	 :D
[12:31:59] <addshore>	 tarrow: how is your sparql foo coming along?
[12:32:02] <hoo>	 it's probably not going to be nice, anyway
[12:32:28] <tarrow>	 What do you need?
[12:32:35] <addshore>	 read up ^^ ;p
[12:33:02] <addshore>	 and IN or something doesnt work I wonder if there is any nice way to run a looped query and union :P or if this should be in a script rather than a single sparql query...
[12:33:47] <tarrow>	 um...
[12:34:36] <tarrow>	 so you want to find all properties that have a datatype item and then find all of those items with that type of property?
[12:34:53] <addshore>	 yup ;)
[12:35:04] <hoo>	 addshore: If I don't grossly miscalculate my stuff, 100M such snaks should be "fine"
[12:35:40] <tarrow>	 let me have a look
[12:35:58] <addshore>	 right now a script seems nost realistic, but would be cool if it could be done in a single query
[12:36:51] <tarrow>	 I think it can be
[12:37:02] <tarrow>	 is it any data type?
[12:37:10] <tarrow>	 or just wikibaseItem?
[12:37:15] <addshore>	 SELECT ?s WHERE {?s wikibase:propertyType wikibase:WikibaseItem}
[12:37:18] <addshore>	 just Wikibase Item
[12:37:22] <addshore>	 for now ;)
[12:37:25] <tarrow>	 sure
[12:38:18] <tarrow>	 and any property?
[12:39:14] <addshore>	 well, where the property is a WikibaseItem / has the datatype wikibaseitem :P
[12:40:10] <addshore>	 im guessing a filter? or something
[12:58:43] <jzerebecki>	 DanielK_WMDE: https://phabricator.wikimedia.org/T119707
[13:01:45] <jzerebecki>	 benestar|cloud: https://phabricator.wikimedia.org/T119707
[13:02:42] <addshore>	 jzerebecki: cant view it
[13:04:47] <addshore>	 ty
[13:54:53] <frimelle>	 I think there could be something broken here: https://www.wikidata.org/wiki/User:Joshbaumgartner/property_available_summary/2400-2499
[14:29:13] <Lydia_WMDE>	 Pyb: \o/
[15:37:11] <Krinkle>	 DanielK_WMDE: #wikimedia-releng
[17:37:19] <tobias47n9e>	 Friday is the last day: (╯°□°）╯︵ ┻━┻ https://www.wikidata.org/wiki/Wikidata:Wikimania_2016
[17:45:46] <hoo>	 SMalyshev: Is it just me or is WDQS dead?
[17:45:59] <SMalyshev>	 shouldn't be... let me see
[17:46:07] <hoo>	 I might have killed it myself, though (by doing some awful queries)
[17:48:48] <Pyb>	 don't kill it, there is a workshop about wikidata & sparql in 1 hour ;)
[17:48:56] <hoo>	 :P
[17:49:07] <hoo>	 ah, it just responded
[17:49:19] <SMalyshev>	 yeah definitely appears to be slow but let me see why
[17:49:28] <hoo>	 It's back to normal now
[17:49:32] <SMalyshev>	 Running query count=29
[17:49:42] <SMalyshev>	 somebody is having fun there...
[17:49:48] <hoo>	 Might have been me
[17:49:59] <hoo>	 well, in case my queries kept running after I got timeouts from nginx
[17:50:40] <SMalyshev>	 the query should time out after 30 secs... but that doesn't prevent one to run 100 of them in parallel...
[17:50:51] <SMalyshev>	 so far we don't have real concurrency controls there
[17:51:07] <hoo>	 ok
[17:52:55] <SMalyshev>	 Running query count=32
[17:53:05] <SMalyshev>	 hoo: are you sending new ones?
[17:53:28] <hoo>	 Yeah, but not more than one at a time
[17:53:32] <hoo>	 always wait for them to time out
[17:53:47] <hoo>	 or error out, fwiw
[17:55:01] <SMalyshev>	 ok, let me restart it maybe... it looks like there are so many of them that the engine can't time them out in time... will need to investigate that
[17:55:10] <hoo>	 Ok
[17:55:18] <hoo>	 I'll be more careful after
[18:01:57] <SMalyshev>	 hoo: I'm still nseeing a ton of parallel queries
[18:02:05] <SMalyshev>	 PREFIX wikibase: <http://wikiba.se/ontology#> PREFIX wd: <http://www.wikidata.org/entity/> PREFIX wdt: <http://www.wikidata.org/prop/direct/> PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#> PREFIX p: <http://www.wikidata.org/prop/> PREFIX v: <http://www.wikidata.org/prop/statement/> prefix schema: <http://schema.org/> Select ?jeu ?jeuLabel ?jeuarticleen ?jeuarticlefr {
[18:02:05] <SMalyshev>	 ?jeuarticleen schema:about ?jeu . ?jeuarticleen schema:inLanguage "en" . ?jeuarticleen wikibase:badge wd:Q17437796 . ?jeu wdt:P279*/wdt:P31 wd:Q7889 . filter not exists { ?jeuarticlefr schema:about ?jeu . ?jeuarticlefr schema:inLanguage "fr" . ?jeuarticlefr wikibase:badge wd:Q17437796 . } SERVICE wikibase:label { bd:serviceParam wikibase:language "fr" . } }
[18:02:11] <SMalyshev>	 is this yours?
[18:02:15] <hoo>	 Nope
[18:02:35] <SMalyshev>	 ok, then somebody else is DOSing it
[18:03:54] <SMalyshev>	 Accepted query count=157
[18:03:56] <SMalyshev>	 wow
[18:04:19] <SMalyshev>	 Running query count=15
[18:05:45] <SMalyshev>	 hmm... somebody really needs some data NOW
[18:09:19] <addshore>	 :P
[18:17:11] <addshore>	 SMalyshev: how do you get those numbers? ;)
[18:17:48] <SMalyshev>	 addshore: it's internal GUi for blazegraph
[18:18:06] <addshore>	 *wonders why that isnt one of the stats it spits out internally too* :P
[18:18:19] <SMalyshev>	 addshore: yes, it's one of the counters
[18:18:26] <addshore>	 is it O_o
[18:19:15] <hoo>	 SMalyshev: Do you think my queries were related to the troubles (I was only running a small amount of them, as said earlier)
[18:19:30] <hoo>	 Can non-parallel queries actually be a problem
[18:19:32] <hoo>	 ?
[18:19:40] <SMalyshev>	 hoo: no it looks like somebody else is doing it
[18:19:49] <SMalyshev>	 hoo: because it keeps coming
[18:19:51] <hoo>	 Ok
[18:20:29] <hoo>	 Do we have any instance in labs still running that can be used for testing?
[18:24:09] <SMalyshev>	 hoo: there's http://wdqs-test.wmflabs.org/ but it is out of date and being used for debugging some other issue right now... so it may bounce up and down 
[19:13:06] <Sylvain_WMFr>	 hi
[19:13:06] <Sylvain_WMFr>	 is the sparql endpoint down ?
[19:13:08] <Sylvain_WMFr>	 we are at a Wikidata workshop about sparql, if we cannot use it it's not good
[19:13:08] <hoo>	 Looks like it, yes
[20:04:27] <tobias47n9e>	 Friday is the last day: (╯°□°）╯︵ ┻━┻ https://www.wikidata.org/wiki/Wikidata:Wikimania_2016
[20:21:46] <hoo>	 Wow, wikidata has a lot of circles :P
[20:21:47] <hoo>	 https://tools.wmflabs.org/hoo/Wikidata-circle-finder.php
[20:30:05] <hoo>	 moved to https://tools.wmflabs.org/hoo/Wikidata-cycle-finder.php
[20:32:15] <nikki>	 hoo: would be nice if the results were links :)
[20:33:05] <hoo>	 nikki: Yeah, indeed
[20:35:12] <multichill>	 hey hoo and nikki, do you know if dates are supposed to be localised? I'm getting "1660s" even when I set my language to Dutch
[20:35:56] <hoo>	 multichill: Where?
[20:36:07] <multichill>	 Wikidata on an item
[20:36:24] <hoo>	 they are, but it's obscure how it works
[20:36:29] <hoo>	 so could be, that that is expected
[20:36:31] <hoo>	 doubt it, though
[20:36:32] <nikki>	 I remember being annoyed by the weird way the dates were displayed for japanese
[20:37:15] <hoo>	 nikki: Done :)
[20:37:19] <hoo>	 Now off to have some food
[20:37:36] <nikki>	 so in my experience they're partly localised, but not properly
[20:37:52] <nikki>	 thanks hoo :)
[20:43:50] <nikki>	 heh, yeah, for japanese I get things like "dd mm月 yyyy" instead of "yyyy年mm月dd日", it seems like it just translates the month name and puts it where the english word goes
[20:46:42] <nikki>	 and even for english it still displays things like "20th century" as "20. century", which looks like denglisch
[20:47:10] <multichill>	 Hmm
[20:47:19] <multichill>	 I feel an i18n coming up
[20:49:14] <nikki>	 ah, I think it's https://phabricator.wikimedia.org/T63958
[20:50:32] <multichill>	 Looks like it.
[20:51:01] <multichill>	 Bug should be blocked by the task where MwDateFormatParser is implemented
[20:51:33] <multichill>	 Can't find that one
[21:08:10] <speckmade>	 I'd like to hear thoughts/opinions/about consensus on what to do in cases like:
[21:08:10] <speckmade>	 If we have an object for a wikipedia article for something like "is a website and the associated software it runs on",
[21:08:10] <speckmade>	 do we create an object for the site and one for the system that runs it?
[21:09:07] <speckmade>	 wspecially: why should i not do that?
[21:11:17] <Dereckson>	 each item should have one specific nature
[21:11:25] <Dereckson>	 I would create two items
[21:11:25] <speckmade>	 if wikidata is really primarily about storing interwiki links, then i should rather not - right?
[21:11:33] <speckmade>	 rather not split
[21:11:35] <Dereckson>	 one P31 software / the other P31 website
[21:11:44] <Dereckson>	 Who told you such a lie?
[21:11:49] <speckmade>	 :-)
[21:12:09] <Dereckson>	 "Wikidata is a free linked database that can be read and edited by both humans and machines."
[21:12:22] <Dereckson>	 We're not responsible for the homonimies of Wikipedia projects.
[21:12:59] <speckmade>	 i think i remember something from a page like "project goals" or the like.
[21:13:19] <Dereckson>	 Yes, one of the main goal is to store interwiki links.
[21:13:38] <Dereckson>	 but not at the cost of a lack of precision of the items
[21:13:44] <speckmade>	 it was talking about "phases" of the project and its use or something.
[21:14:03] <speckmade>	 so - where do i put the wikilinks then?
[21:14:13] <Dereckson>	 The relevant question in your case is "should we anchor the interwiki to the website item or the software item?"
[21:14:16] <speckmade>	 it refuses to put them with both items
[21:14:28] <speckmade>	 indeed. :-)
[21:14:31] <Dereckson>	 okay we concur about the real question :)
[21:14:51] <speckmade>	 dices?..
[21:15:14] <Dereckson>	 For stuff like Reddit, website seems currently to be prefered.
[21:15:30] <speckmade>	 well, in the one case at hand right now the article is mainly about one aspect - but i've seen other cases.
[21:16:01] <Dereckson>	 You could for each Wikipedia article evaluate if th earticle is mainly about website or software
[21:16:08] <Dereckson>	 and link accordingly
[21:16:15] <speckmade>	 this can also change.
[21:16:26] <Dereckson>	 Yes, you can split the interwiki links to disambig.
[21:16:39] <Dereckson>	 What's the website here?
[21:17:04] <speckmade>	 i'm currently thinking about WikiWikiWeb
[21:17:18] <nikki>	 given that wikidata items are the basis of the interwiki links, we can't just ignore the fact that lots of wikipedia articles don't split things up like we want to :/
[21:18:02] <Dereckson>	 en.wikipedia infobox chosen is the website one
[21:18:19] <Dereckson>	 fr. doesn't speak about the software
[21:18:49] <Dereckson>	 but as a template for the wiki software ...
[21:18:51] <Dereckson>	 how coherent
[21:19:09] <Dereckson>	 pt. speaks about the software
[21:19:23] <Dereckson>	 ru. website
[21:19:27] <speckmade>	 i mean, if i'm on an article on a website and a software and i want to jump to the equivalent in another language - and in that other language they have seperate articles... :-/
[21:19:42] <Dereckson>	 split the links according the major topic of each article
[21:19:50] <nikki>	 and I bet people on ptwiki want interwiki links to frwiki too
[21:20:13] <speckmade>	 there's cases when you have a company named after their main product where it's much more ambiguous than with Wiki.
[21:20:50] <speckmade>	 tricky.
[21:20:51] <Dereckson>	 If you really really really really really want to keep these interwikis together, link them all to the website
[21:21:00] <Dereckson>	 If you want purity of data first, split them
[21:21:06] <Dereckson>	 In both case, create two items
[21:21:37] <speckmade>	 i'm grateful for the insight, people! :-)
[21:22:14] <Dereckson>	 I sympathize, this kind of question is though
[21:32:23] <speckmade>	 don't we also have this thing that notability of an object is usually proven through interwiki links? too bad if they have to be with the other object on the main topic of that article...
[21:37:04] <nikki>	 there only needs to be one sitelink, and sitelinks aren't actually required, there are other ways for something to be notable
[21:38:35] <nikki>	 so it's not really a problem for us if we link one page to one item and another page to another item, but it might be a problem for the wikipedias (or other projects) if they want those pages to be interwiki links
[22:06:10] <tobias47n9e>	 Friday is the last day: (╯°□°）╯︵ ┻━┻ https://www.wikidata.org/wiki/Wikidata:Wikimania_2016