[12:05:23] there seems to be a little bug with dates, is there not? [12:06:20] I enter something like "3. century", and it ends up with "1. century", and I have to edit it to correct the value set up at the creation of the property [12:06:41] uh? that's new [12:06:49] never had this problem before :s [12:06:55] and hello Rama [12:07:57] Oh also, German speakers are the only people I've ever seen using a period after a number to create a cardinal. This should not be hard-coded, it's almost as weird as the mental date formats at the Americans insist on using even though we told them not to. [12:08:40] or their insane units for length, temperature, volume-and-mass-somehow-conflated-for-no-reason, for that matters... [12:08:45] hello Harmonia_Boulot! [12:14:50] Harmonia_Boulot: oh do you think it would be feasible to rename an item? I mistyped a reference number the other day... [12:15:09] "rename" an item? [12:15:21] you can change the label all you want? [12:15:43] not sure I understand what you want [12:23:47] oh gosh, I had not realised that the label could be edited that easily [12:23:50] sweet [12:24:25] ^^ [12:24:46] Harmonia_Boulot: do the Label and Description also have P-numbers? [12:24:54] Rama: and you can change labels/descriptions/aliases in all languages [12:24:55] no [12:25:05] why would they? [12:25:23] they are associated with a Qid [12:25:29] so I could call them from {{#Property}}... [12:25:45] you can call them [12:25:53] but not with {{#Property}} [12:26:20] you can call all labels, descriptions, aliases associated with a Qid by language [12:26:22] ah, do you know how? [12:26:39] or do I have to code a Lua method for that? [12:27:08] urgh, I never work with such templates so no, i don't know like that [12:27:27] there seems to be bugs in #Property, the documentation claims that is imports the property in the relevant language, but in fact it just concatenates everything [12:33:18] Rama: entities.getLabel(entity) en lua ? [12:33:50] yup, works [12:33:53] sweet [12:34:07] now the template is getting a bit more mutli-lingual [12:34:10] :) [12:34:46] only a bit though, there are things I really should be importing in English because I send them through Commons templates [12:35:26] this is still a bit rough but it is more and more functionnal [12:35:26] Rama: Which language version are you working on? It might already be in the local Wikidata module [12:35:59] Ainali: just switching back and forth from English to French on Commons, for testing purposes [12:36:13] only English on Wikidata [12:37:11] Rama: you have activated the LabelLister gadget, of course? [12:37:28] Rama: to edit more easily labels, descriptions and aliases in all languages? [12:38:00] hmm, no idea [12:38:13] it's fairly easy as it is, anyway [12:38:26] Rama: On commons, this should work: {{#invoke:Wikidata|getLabel|entity=Q42}} [12:39:10] Rama: activate it [12:39:28] Ainali: ah, thank you. I had written my own in between ^^ [12:42:06] ah but that's not what I need, it is the description I am after [12:42:33] it doesn't seem to be available though the Wikidata module, weird [12:42:48] well mine has it now, so no sweat [12:53:53] ah, just fixed my {{author:}} anguishes thanks to Ainali's tip [12:53:54] 10[7] 04https://www.wikidata.org/wiki/Template:author: [12:54:12] \o/ [14:50:50] WikidataFacts: I have a challenge for you: making this somehow work again by dividing it into smaller queries: [14:50:50] https://query.wikidata.org/#SELECT%20%3Fitem%20%3Fpmcid%20%3Fcites%20WHERE%20%7B%0A%20%20%20%20%3Fitem%20wdt%3AP932%20%3Fpmcid%20.%0A%20%20%20%20%3Fitem%20p%3AP2860%20%3Fscites%20.%0A%20%20%20%20%3Fscites%20ps%3AP2860%20%3Fcites%20.%0A%20%20%20%20%3Fscites%20prov%3AwasDerivedFrom%20%5B%20pr%3AP248%20wd%3AQ229883%20%5D%20.%0A%7D [14:52:12] One idea of mine is having one query for ?item values that are < Q28000000 and another that is >= Q28000000 [14:55:47] harej: apparently there is *exactly one* reference that matches the last part – or am I doing something wrong? http://tinyurl.com/jdnckn9 [14:56:05] if true, that same node is presumably used on all of those statements [14:56:09] so that might work [14:59:46] well, there's literally over 2 million P2860 statements that are attributed to PubMed Central. Does the hash returned in that query mean "it is done exactly one way with no variations"? [15:00:46] (This might not be true in the long term as we do more sophisticated metadata in citations but for now it's true) [15:01:17] harej: yes, afaik if the references are identical they get the same node [15:01:40] but right now I have two versions of the query running that haven’t timed out yet [15:02:59] ah, no, there were results, which I can see in the dev tools network tab, but apparently the website isn’t yet done parsing them [15:03:24] the error I've been getting has not been a timeout but an unexpected end to the JSON [15:03:30] hm, okay [15:03:56] which I guess makes sense -- it's not a computationally complex query, but a simple one with a *lot* of results [15:05:40] oh yes, apparently [15:06:14] 1168753 results [15:07:39] oh, but the query still doesn’t *complete*, it just sends the partial result after 30 seconds elapse [15:07:53] yup, got 1735772 this time [15:10:24] try doing a count? [15:10:34] you mean in the query? [15:11:53] okay, now I see a timeout, so now I can check if the query timeouts or not, good idea [15:14:58] I am moving the mediawiki/extensions/Wikibase jenkins jobs to disposable instances [15:15:04] that is a noop :} [15:15:59] harej: counting the statements doesn’t timeout yet: http://tinyurl.com/heaxt5c [15:16:33] (result is 2204561 btw) [15:17:02] That sounds right [15:17:23] but then just finding items with those statements (even discarding the PCMID part) times out [15:37:29] oh for goodness sake [15:37:54] the British English "no result" message lacks " You may create a new item" [15:40:57] harej: I can’t make it not timeout, sorry [15:41:16] Thank you for trying, at the very least :] [15:41:55] it’s weird, there are only some 300k items with PMCID (and counting them was almost instantaneous), I thought that filtering for that first would help… but apparently not [16:39:23] halfak: o/ [16:39:49] hey glorian_wd [16:39:57] so yeah, thinking about the quality of references. [16:40:08] It seems that we have "external" as a reference quality criteria [16:40:13] Wikipedia has "Reliable Sources" [16:40:27] https://en.wikipedia.org/wiki/Wikipedia:Identifying_reliable_sources [16:42:33] yeah, I am just thinking if we just explicitly define the reader's experience as "external reference" instead of "high quality reference" [16:42:48] but we dont need to change the criteria [16:43:20] Readers don't know what we mean by "external reference" [16:44:02] I think the point of the "readers experience" column is try to think about how an outside would describe an item. [16:46:45] what would you say if we write it as "... external references (references other than Wikimedia projects)"? so we explain it in the sentence. [16:47:03] maybe not an elegant solution? [16:47:17] I don't think readers know what "Wikimedia projects" are either. [16:48:50] Right [16:49:27] References to a non-wiki website? [16:50:04] Even so, *why* do we want that? [16:50:12] I don't think a reader experiences anything so specific [16:50:21] okay [16:50:22] I think they experience "good references" [16:50:34] I think the idea is that Wikidata wants to make sure the information it has is valid outside of the wikimedia bubble [16:50:51] The goal of "external" is not to just *be external* but to have reliable sources -- of which Wikipedia is not. [16:51:13] +1 harej [16:51:36] We could say "reliable references" or maybe "good references" [16:54:08] halfak: valid references? [16:55:34] I'm confused about what the actual problem is with "quality" [16:55:53] valid (adj): having a sound basis in logic or fact; reasonable or cogent. [16:59:03] halfak, based on that definition, what do you think about "valid references"? [17:00:09] I don't think a reader would know the "sound basis" for references. [17:09:36] halfak: so, I am just thinking if we say, "high quality" references, it may open rooms for misinterpretation. Although you might argue that, when we say "high quality", we refer it as external references, which presumably, reliable [19:35:16] https://www.wikidata.org/w/index.php?title=Q160556&type=revision&diff=449889769&oldid=440485797 *coughs* [19:37:24] hoo: reference URL: http://www.der-postillon.com/2017/01/ber-fertiggestellt.html ? [19:38:18] that “service entry: 2018” (P729) also looks rather suspicious [19:38:21] Damn, I missed that news :P [20:31:26] Hey folks. glorian_wd and I are trying to figure out how wbs_propertypairs is generated and what we should expect to find in it. [20:32:09] Ask hoo. [20:46:31] hoo is on first [20:46:31] https://en.wikipedia.org/wiki/Who's_on_First%3F [20:46:36] * halfak is hilarious [20:50:20] Damn timezomes [20:52:31] halfak: lol [21:55:04] I'm trying to fetch claims from wikidata using RepoApi.js, but it seems that the done/then of the defered object doesn't get called [21:55:08] e.g new wikibase.api.RepoApi( wikibase.client.getMwApiForRepo() ).getClaims(mw.config.get('wgWikibaseItemId')).done(function(d){console.log(d);}) [21:55:47] nvm