[05:33:29] Is the password reset function working? As I requested a new password within the last 24 hours but so far no email [08:39:25] WikidataFacts: did you saw we have a new top cocktail? :P [08:40:04] * harmonia doesn't drink alcohol but find these queries pretty funny [08:40:15] wow, 88 results now? nice [08:44:28] harmonia: me neither, actually, but it’s still fun to write a query that returns (mostly) proper English :D [08:44:57] WikidataFacts: yep :) [08:46:33] hello dachary [08:46:47] I'm almost done with the "official" cocktails. [08:47:51] * harmonia is learning the existence of official cocktails [08:47:57] harmonia: \o [08:48:19] https://www.wikidata.org/wiki/Q2536409 [08:48:54] they require a lot of new items (translations needed!) https://www.wikidata.org/w/index.php?limit=50&title=Special%3AContributions&contribs=user&target=Sjoerddebruin&namespace=&tagfilter=&newOnly=1&year=2016&month=-1 [08:48:56] O_O [08:49:06] so many things I don't know :p [08:59:19] not sure about https://www.wikidata.org/wiki/Q26882302 [09:16:57] sjoerddebruin: seems fine to me? or do you mean the details [09:17:15] if the statements look correct, then I shouldn't worry [09:17:23] well, I wouldn't say "material used: coffee" [09:17:46] this may just be an American thing but we consider "coffee" to be drip coffee, different from espresso [09:18:16] hmm yeah [09:19:06] and I am not sure sure how Wikidata classifies these different things: whether "coffee" refers to drip coffee, or if "drip coffee" and "espresso" are both subclasses of the general concept of "coffee" [09:20:17] other coffee drinks aren't helping either... [09:20:50] That's the annoying thing with Wikidata: people just do what comes naturally, and here you are, someone who actually cares about the data model [09:21:09] There are various unincorporated places that Wikidata casually calls "cities," and here I am like "ahhhhhhhh" [09:21:19] pdf document says "ground coffee", so gonna make a item for that [09:34:18] hare: https://www.wikidata.org/wiki/Q26882413 [09:34:32] makes sense [09:34:56] :) [09:41:41] does nikki approve too? :) [09:43:09] well, it's a thing, right? [09:43:34] yeah, and https://www.wikidata.org/wiki/Q1155703 is more about the process [09:44:01] I've never heard of that so I wouldn't know :) [09:44:53] :O [09:45:57] I don't drink coffee so I'm not very familiar with the terms for it [09:46:25] You can talk Japanese though? [09:47:45] I can read it better than I can speak it :P my grammar is pretty bad [09:55:11] huh... the "wikivoyage" header isn't translated into japanese even though all the other sitelink headers are [10:05:01] "on the rocks" is even on the frontpage :3 [10:09:15] nikki: i think that is a different product https://www.wikidata.org/w/index.php?title=Q10749869&type=revision&diff=376318107&oldid=376318105 [10:09:51] http://www.realreviews.in/food/food-products-beverages/entry/difference-between-vanilla-extract-and-vanilla-essence/ [10:10:07] https://de.wikipedia.org/wiki/Vanille_(Gew%C3%BCrz)#Vanilleextrakt seems to be wrong then :/ [10:13:11] I removed it anyway [10:13:55] the vanilla extract I have says "vanilla-extrakt" but it's not something I've come across here much. I had to buy it from some expensive shop [10:15:07] If it's expensive, it's mostly good. Vanilla essence is recreated. [10:15:17] it seems like german recipes just use vanilla sugar, which I've never seen in the uk [10:16:09] having a mixture of british and german recipes is sometimes rather annoying :P [10:16:18] yeah, most Dutch recipes use vanilla sugar too [10:19:49] and yeah, I know that it's quite expensive, I just meant that I had to go to a shop where everything is expensive to find it [10:20:32] still didn't find it in a real store here [10:20:37] should make it myself sometime [10:21:08] vanilla extract smells *wonderful* [10:21:11] and is full of alcohol! [10:21:44] it's rather strong, I think there are easier ways to consume alcohol if that's your aim :P [10:21:51] that must be the reason why we didn't have it at school :P [10:21:55] haha [10:22:13] we only had rum essence, great in brownies [10:22:36] we didn't do much cooking at school... [10:23:09] Well I was lucky to have like 4 years of cooking every week. [10:23:54] wow [10:24:36] and three years of school cafeteria, so yeah I've had fun [10:31:34] we had something like 2 hours a week of "design and technology", which included "food", "resistant materials" (aka woodwork... although we did once use perspex), "textiles", "electronics" and "graphic design". we would alternate between areas every few months and then after three years we had to pick one of them to do for the last two years [10:32:43] so basically half a year at most and even then most of the "food" classes didn't involve actual cooking [10:50:56] DanielK_WMDE: Around? [10:50:59] Let's hope I don't choose the wrong item. https://www.wikidata.org/wiki/Q26882990 [11:00:56] (WikidataFacts: look, we can make it even more complicated :P) [11:28:13] DanielK_WMDE: Jenkins is crazy today [11:28:18] failed to merge [11:30:42] *sigh [11:32:13] DanielK_WMDE: btw. I'm playing with the Special:EntityUsage [11:32:30] https://usercontent.irccloud-cdn.com/file/obp0HTFr/ [11:32:32] It's the result [11:32:52] one is for "S" aspect and one is for "O" aspect [11:33:09] I want to group them but we would lose aspects [11:33:26] why ouwld you lose aspects? [11:33:36] (I thought about something like this too http://dev.mysql.com/doc/refman/5.7/en/group-by-functions.html#function_group-concat but that only works in mysql) [11:33:48] you can group them programmatically, or with GROUP_CONCAT [11:34:36] programmatically I need to rewrite output result function of PageQuery [11:34:46] almost all of it [11:34:49] Amir1: SQLite has group_concat too: https://www.sqlite.org/lang_aggfunc.html#groupconcat [11:35:06] I checked for Postgres and it didn't have [11:35:06] Wikibase only supports MySQL and SQLite [11:35:19] if that's only the case I'm totally okay [11:36:58] looks like it can be emulated in postgres, too: https://explainextended.com/2009/03/04/aggregate-concatenation/ [11:38:14] yeah it has another method called agg_concat or something [11:38:29] but the name is not the same it might fail in postgres [11:38:51] Amir1: oh, even better! we have DatabaseBase::buildGroupConcatField [11:39:00] yesssss [11:39:05] :) [11:39:19] * Amir1 going back to building it [11:39:29] thanks for the +2 [11:39:37] np! [11:40:14] ha, group_concat syntax isn't the same for sqlite and mysql anyway. excellent that we have something for that in the db abstraction layer [11:56:04] official cocktails: done [11:58:20] sjoerddebruin: do we have pictures? [11:58:45] we lack pictures :( [11:58:54] out of curiosity, how much time is appropriate between a property proposal (where there seems to be solid consensus) and the actual creation request ? a week ? [11:59:11] Alphos: a week yes [11:59:21] ok great :) [11:59:35] there should be at least one, perhaps two new properties in the next week or two ^^ [12:00:59] :O [12:02:34] I am so against this propery. And still I proposed it xD https://www.wikidata.org/wiki/Wikidata:Property_proposal/Nationalencyklopedin_Online_ID I don't like us linking to "close source" competitors but... [12:02:43] oh you! [12:07:18] * harmonia is happy because the number of constraints violations for given names and surnames slowly but steadily reduced [12:07:56] harmonia: yes, it was pretty bad some months ago [12:08:17] i did so much work since mid-August :) [12:08:28] less than 300 false given names now! [12:09:21] Still want to make https://www.wikidata.org/wiki/User:Sjoerddebruin/Dutch_people_without_first_name empty. :P [12:09:38] family names are still a mess between disambiguation and family name [12:09:49] (no thanks to Yellowcard last year><) [12:10:02] but hey! slowly but steadily [12:16:38] "1 dash Angostura bitters, 3 centiliters gin, 0.75 centiliters Cointreau, 12 centiliters pineapple juice, 1 centiliter grenadine, 1.5 centiliters lime juice, 0.75 centiliters Bénédictine, and 1.5 centiliters cherry liqueur, served with maraschino cherry and pineapple in a highball glass" [12:24:04] DanielK_WMDE: sorry to bother you again, that method builds something like this: [12:24:34] > echo wfGetDB( DB_SLAVE )->buildGroupConcatField( '|', 'wbc_entity_usage', 'eu_aspect'); [12:24:34] (SELECT GROUP_CONCAT(eu_aspect SEPARATOR '|') FROM `wbc_entity_usage` ) [12:24:34] (the whole query) [12:25:06] but QueryPage asks only array of fields, etc. [12:25:06] getQueryInfo [12:25:23] and makes the string itself in reallyDoQuery [12:25:30] I have two options: [12:25:53] 1- Do a nasty hack to strip select from wfGetDB( DB_SLAVE )->buildGroupConcatField( '|', 'wbc_entity_usage', 'eu_aspect') [12:26:05] and add it into getQueryInfo [12:26:10] or 2- [12:26:25] rewrite reallyDoQuery in the subclass [12:26:29] 1 or 2? [12:26:56] Is my question clear? [12:40:07] I did not get any answer to https://lists.wikimedia.org/pipermail/wikidata/2016-September/009521.html and that makes me wonder if it's a dumb question and people are too polite to tell me ? ;-) [12:41:05] I don't know... [12:43:47] sjoerddebruin: :-) [12:45:42] * dachary found https://www.mediawiki.org/wiki/API:Delete [12:47:54] but it looks like it's for mediawiki pages, not wikibase [12:48:40] But we delete items with the normal interface too [12:49:09] how do you do that ? [12:49:30] being an admin a "Remove" entry shows in the page menu or something similar ? [13:01:34] Would it be possible to upgrade the user permissions to allow deletion of items / property on test.wikidata.org ? [13:10:35] Lydia_WMDE: the sprint you've added duplicatereferences starts on 20-09, does this mean we'll have to wait at least another week? [13:10:57] I've seen another one complaining yesterday and today someone in the project chat again. [13:11:15] sjoerddebruin: hmmm no. let me check. [13:13:16] sjoerddebruin: bah ok so the times in phabricator are all wrong it seems. i'll sort it out with the team when back in the office on friday [13:13:29] okay :( [13:13:39] * harmonia offers tea to Lydia_WMDE [13:14:37] thx harmonia :) [13:14:49] sjoerddebruin: *hug* i added it to the other one now as well [13:18:33] * Alphos offers matcha-vanilla flavored non-fat milk shake to Lydia_WMDE, and one-ups harmonia in the process :p [13:18:45] lol [13:18:55] :D [13:19:05] * Lydia_WMDE hands out a bunch of little kittens and wins [13:19:06] :P [13:19:11] Amir1: sorry, I was off for a bit. [13:19:13] Alphos: do we have a item for that including ingredients? :) [13:19:25] Lydia_WMDE : i'm not allergic to matcha, vanilla, or milk :/ [13:19:28] please don't try to strip stuff from the generated query, that's horrible [13:19:34] we could split the method in core, that would be nice [13:19:46] i like kittens, they just make my face bloated when i don't have antihistaminics :-( [13:19:53] ah damn [13:20:06] * Lydia_WMDE gets all of the kittens again and locks them away [13:20:21] yeah, I was thinking about it too [13:20:42] Amir1: I'd like if we could just use GROUP_CONCAT directly. not sure how do best handle the syntax difference between mysql and sqlite though [13:20:43] I do it and you review? [13:21:01] We can do it in core [13:21:08] yea, make a patch, let's see if it's easy. [13:21:13] i can't promise it will go in though [13:21:16] awwww [13:21:21] also, i'll be offline tomorrow [13:21:59] Amir1: depending on the underlying DB, it may not be possible to emulate a GROUP_CONCAT without the entire SELECT... [13:22:16] that's the nasty part [13:22:33] yeah [13:23:01] the strange thing is that the buildGroupConcatField don't get options part [13:23:08] like order by, etc. [13:24:15] this is straightforward to fix [14:58:12] hoo: second try, hey. [14:58:18] hey sjoerddebruin :) [14:59:06] ^^ [14:59:13] After some more "research", I think https://www.wikidata.org/wiki/Property:P18 is to blame for most wrong suggestions. [14:59:34] sjoerddebruin: Could be… also it doesn't make sense to have that in there [14:59:42] the same could be said about https://www.wikidata.org/wiki/Property:P373, but I wouldn't know the impact then [15:00:04] thinking about it, what good guess could you do by knowing that there's an image on that Item? :P [15:00:18] yeah, we can probably kick both out [15:00:29] Lydia_WMDE: Opinions ^? [15:00:51] i'm wondering if you still get a suggestion to add a commons category when a image exists and vice-versa [15:02:28] Well, you would still get these properties suggested [15:02:45] but you wouldn't get suggestions based on these properties [15:03:04] that's would be *great* ! [15:03:14] ah, what about https://www.wikidata.org/wiki/Property:P17 then? 5,5 million uses on various subjects [15:03:55] probably as well [15:04:20] I think a lot should still be converted to https://www.wikidata.org/wiki/Property:P495, like companies [15:07:58] hoo: maybe we should start with those three I've named and see what the effects are? Waiting for Lydia_WMDE's approval of course. [15:08:47] We can easily play with this… should we remove to much, we can easily re-import the data derived from the dump [15:08:52] does the property suggester take values used for the property into account or only the properties themselves? [15:09:40] nikki: Only the properties themselves, except for subclass of and instance of (these are "classifying properties) [15:10:13] how much is the influence of those? [15:10:27] as I saw more relevant properties when I've added https://www.wikidata.org/wiki/Q2536409, despite the low use (most of them were added this week) [15:10:37] I think equal with the fact that an other property is present [15:10:39] ah... I was wondering because I've noticed that identifiers for geographical objects in big countries tend to get suggested a lot (presumably because big countries have more objects that can have identifiers) [15:10:57] whereas they're pretty much (by definition) limited to a specific country [15:11:26] I thought about also making for example "occupation" classifying for example, but not sure about the impact [15:11:30] Oh, like https://www.wikidata.org/wiki/Property:P442 popping up every time [15:11:35] (nikki) [15:11:39] yes, exactly! [15:12:18] would it make sense to make P17 as classifying property then? [15:13:23] sjoerddebruin: Doubt it [15:13:41] P17 is used on a whole set of subjects [15:13:44] with the same values [15:13:54] Hm, okay. That's something for the new version then [15:14:15] I've created https://www.wikidata.org/wiki/User:Sjoerddebruin/Entity_suggester a few days ago [15:15:41] Because I personally think the project can really benefit from a smarter suggester. We want people to add references for example, but you get really bad suggestions before saving and mostly no references when you edit a existing statement. [15:16:08] +1 [15:16:16] I agree [15:16:18] * harmonia endorses this :p [15:16:19] yeah, the suggester saves a lot of effort searching for stuff, when it works [15:17:09] speaking of which, having to repeatedly search for new properties (because they're too new and unused to get suggested) is another problem I have :P [15:17:14] Indeed… I would really love for it to be more reliable [15:17:30] nikki: Yeah, I only update it every 4 weeks right now [15:17:37] but I could probably do it more often [15:17:57] I've scripted most of the (rather ugly) manual process by now, so it's less effort for me [15:18:02] depends on the activities onwiki, if someone does a big bot run... [15:18:07] but still some manual work needed [15:20:00] I can already feel the impact of hare's work for example :) [15:21:46] But also things like https://www.wikidata.org/wiki/Property:P1435 show up in every geographical item, despite they only make sense in a smaller group of them [15:39:01] Is there a magic approval yet, hoo? :P [15:48:40] sjoerddebruin: hoo: let's try with those two i'd say! [15:48:51] thanks, will do [15:48:59] image and commons category? [15:49:04] yeah [15:49:11] ok, i'm curious [15:49:20] * Lydia_WMDE too [15:50:34] done [15:50:37] please verify :) [15:53:38] nothing bad so far [15:54:13] the human properties on https://www.wikidata.org/wiki/Q2108187 are gone [15:57:30] Should order some pizza first before I starve [16:35:25] umm...https://www.wikidata.org/wiki/Q4627834 [16:35:34] sjoerddebruin: Want to deal wih that? :p [16:40:51] Lydia_WMDE: https://gerrit.wikimedia.org/r/#/c/310580/ [16:41:10] I've got the special page ready, only some tests and we are good for review [16:41:18] \o/ [16:41:26] (Also the list=wbeulistentityusage got merged today) [16:41:48] excellent [16:49:44] Josve05a: weird stufff... [16:50:06] refer to https://www.wikidata.org/wiki/Q21694049 .... [16:50:24] some weird sitelink move/split? [16:50:42] that maniac did something wrong, repairing now [17:26:40] DanielK_WMDE: https://gerrit.wikimedia.org/r/#/c/309972/6/repo/tests/phpunit/includes/Notifications/JobQueueChangeNotificationSenderTest.php [17:26:47] Are you ok with the workaround in setUp? [17:26:52] If so, could you +2? [19:23:44] So, now I have time for a cocktail. :P [19:41:24] sjoerddebruin: Hou jij https://www.wikidata.org/wiki/Wikidata:Database_reports/without_claims_by_site/nlwiki ook nog in de gaten? [19:41:40] multichill: ik was druk met cocktails https://twitter.com/WikidataFacts/status/775748843807903744 [19:41:45] sjoerddebruin it's been so hot this afternoon, the only cocktail i can come up with has nothing but ice cubes and water in it [19:42:34] Alphos: wonder if that is a legit cocktail [19:42:46] hmm, cocktail... [19:42:57] at this point, i don't care, it does the job :p [19:43:00] http://www.urbandictionary.com/define.php?term=Water%20on%20the%20rocks [19:43:01] I already had a beer and a glass of wine ;-) [19:43:10] :D [19:43:35] i've had two glasses of wine at dinner too [20:16:46] Should we have a property for "serving suggestion"? [20:38:54] hashar CI has a segmentation fault :( https://gerrit.wikimedia.org/r/#/c/297563/ [20:42:29] Hi - when I retrieve Omar Khayyam's date of birth through the SPARQL endpoint, I get May 30 1048. But the wikidata page says May 24 1048. I can provide details but wanted to make sure it is appropriate to ask about this here. [20:44:00] Guest94140: there could be multiple statements, without given ranking it selects the first added [20:44:26] hm, only one statement :/ [20:44:50] https://query.wikidata.org/#%23%20query%20for%20www.dageshify.com%2FautoMath%20%0A%23%20%22do%22%20stands%20for%20%22date%20of%22%2C%20as%20in%20%22date%20of%20birth%22%20or%20%22date%20of%20death%22.%0ASELECT%0A%20%3Fdob_year%20%3Fdob_month%20%3Fdob_day%20%3Fdob_calendarModel%20%3Fdob_precision%0A%0AWHERE%20%7B%0A%20%20wd%3AQ35900%20p%3AP569%20%3Fdob_statement%20.%0A%20%20OPTIONAL%20%7B%20%0A%20%20%20%20%3Fdob_statement%20%3Fps [20:44:57] could it be conversion from Julian to Georgian calendar? [20:45:01] that's the sparql query [20:45:20] perhaps - the calendar model is proleptic Julian [20:45:58] I am not sure what Proleptic means or how to convert Proleptic Julian to non-Proleptic Julian [20:46:31] But there is no Georgian involved [20:47:06] maybe SMalyshev can give more details [20:49:32] thanks, I pinged them, marked as away but I will keep an eye out [21:08:57] Jonas_WMDE: yeah segfault :((( [21:09:11] Jonas_WMDE: Hoo mentioned it earlier today. I am not sure what to do with it. [21:43:42] sjoerddebruin: sorry, was offline for a moment, what's the question? [21:43:57] see the messages of Guest94140, SMalyshev [21:44:21] hi - [21:44:59] when I retrieve Omar Khayyam's date of birth through the SPARQL endpoint, I get May 30 1048. But the wikidata page says May 24 1048. [21:45:17] Guest94140: so the difference is probably because in wdqs and generally in RDF/SPARQL dates are gregorian [21:45:22] here is the sparql query: [21:45:25] https://query.wikidata.org/#%23%20query%20for%20www.dageshify.com%2FautoMath%20%0A%23%20%22do%22%20stands%20for%20%22date%20of%22%2C%20as%20in%20%22date%20of%20birth%22%20or%20%22date%20of%20death%22.%0ASELECT%0A%20%3Fdob_year%20%3Fdob_month%20%3Fdob_day%20%3Fdob_calendarModel%20%3Fdob_precision%0A%0AWHERE%20%7B%0A%20%20wd%3AQ35900%20p%3AP569%20%3Fdob_statement%20.%0A%20%20OPTIONAL%20%7B%20%0A%20%20%20%20%3Fdob_statement%20%3Fps [21:45:41] the calendar model when retrieved through sparql is proleptic julian [21:45:43] or rather proleptic Gregorian (which means extending Gregorian calendar to dates where it didn't exist as if it did) [21:46:00] oh - let me check julian or gregorian [21:46:03] Guest94140: not sure what you mean by that [21:46:42] the calendar model returned by the sparql query is proleptic julian [21:47:19] Q1985786 [21:48:47] Guest94140: ah, ok, so this is a bit confusing part. The value you see as xsd:dateTime is always gregorian [21:48:54] that's what the standard requires [21:49:10] however, the preferred calendar for it - i.e. for display, etc. - is Julian [21:49:26] so if you'd look on it in Wikidata, you'd see Julian date [21:50:09] okay - wikidata shows it in the preferred calendar? [21:50:14] yes [21:50:38] ok. but sparql returns it in non-proleptic gregorian. [21:51:32] no, it's proleptic gregorian [21:51:48] it can't be non-proleptic because we wouldn't be able to represent BC dates then [21:53:41] ok - got it - sparql returns proleptic gregorian, wikidata ui returns what is specified by calendar model .. now one more complication .. wikipedia shows a different date altogether. [21:53:53] is that also because of calendar model issues? [21:54:11] or did they get the date from a different source? [21:54:12] wikipedia has different data, it's probably not syncronized with wikidata [21:54:37] unless specifically synced with templates, wikipedia and wikidata are different databases [21:54:44] so one of them may be wrong [21:55:23] ok, great. and there's no way to tell when querying wikidata which value is used by wikipedia, right? [21:55:54] right [21:56:04] some wikipedia articles do use values from wikidata by template [21:56:10] but many others don't [21:56:34] excellent. This is really interesting. Thanks a lot! [21:57:27] no problem, you're welcome [22:08:26] so for thos reading their IRC backlog in the morning. The Jenkins Zend 5.5 jobs are segfaulting [22:08:41] but I got a couple core dump files with a huge stacktrace :D Bug is https://phabricator.wikimedia.org/T142158 [23:15:14] huh, gc bugs... that sucks [23:17:11] question: when entity on wikidata is edited, are page props updated immediately or later? [23:17:44] specifically sitelinks/statements counts?