[03:53:58] I would like to make interlanguage link between ja:野間昭典 and en:Akinori Noma (Q16733439). [07:06:43] :join #wikipedia-fr [07:06:56] hello [07:09:27] :P [07:10:12] :me doesn't know if anyone had seen https://twitter.com/Harmonia_Amanda/status/774999797019402240 [07:10:30] i made a mini-SPARQL tutorial yesterday [07:11:37] So https://www.wikidata.org/wiki/Q21010880 is still alive? :P [07:14:46] sjoerddebruin: Wikidata says so [07:14:53] Wikidata is never wrong [07:14:57] obviously [07:14:58] :p [07:15:46] So now we should do things like https://www.wikidata.org/w/index.php?title=Q3362446&type=revision&diff=375736597&oldid=354611345? [07:16:22] sjoerddebruin: exactly [07:18:15] sjoerddebruin: do you think the tutorial is somewhat clear? I wrote it in one go, so no real deep thinking... [07:18:31] It's good. [07:18:38] ok, great [07:18:51] I think I'll do more like this [07:18:59] with "complicated" example [07:19:07] like how to query qualifiers [07:19:18] things like that [07:25:51] and hello WikidataFacts [07:25:55] hi! [07:26:30] (I’ll be away pretty soon though) [07:26:57] I'm at work [07:26:58] :O [07:27:13] I think I've cleaned half of it already. [07:27:17] i wait for an email from my boss [07:27:26] sjoerddebruin: French actresses? [07:27:36] It looks a lot cleaner already. :P [07:28:02] I was sad when I finished the tutorial and i saw all that incomplete data [07:28:10] so thank you thank you thank you! [07:28:41] https://www.wikidata.org/wiki/Q731059 is record holder on nlwiki [07:30:20] record holder? [07:30:44] of being alive [07:30:56] ah [07:30:58] :D [07:31:10] I have a list of people who are probably dead with a sitelink to nlwiki. https://www.wikidata.org/wiki/User:Sjoerddebruin/Vermoedelijk_dode_mensen_op_nlwiki [07:31:57] :) [07:32:14] i don't want to start on birth/death dates [07:32:22] i'm already packed with the names [07:32:26] I've murdered a lot of people already. [07:32:32] too many things to do on WD [07:32:35] sjoerddebruin: XD [07:33:32] sjoerddebruin: Alphos yesterday said that when you do a SPARQL query, in French it's not "run", it's "exécuter" (execute) and it was somewhat funny when you queried dead people [07:33:45] :P [07:34:59] (why, but why, do bad, flimsy newspapers write so much crap about education?) [07:35:12] * harmonia is wary and it's only Monday morning [07:43:02] DuplicateReferences is still broken :( [07:43:32] sjoerddebruin: yes :'( [08:59:30] Entity suggestions for films are coming close to the book ones. :( [09:04:50] sjoerddebruin: *sigh* [09:05:05] I lost my good suggestions for theatre too, last week [09:05:25] and I was so happy when it started working [09:05:39] Yeah, reforms are needed and I'm working on getting priority for that. [09:06:05] But see the suggestions of https://www.wikidata.org/wiki/Q2511420 [09:06:56] sjoerddebruin: hmm, I have "cast member" and "screenwriter" that's not so bad [09:07:12] Is your screen so small, click for more suggestions. :P [09:08:00] I still want a way to force something to appear, like if I want to add the same property to a bunch of items, not having to search for it every time would be awesome [09:08:43] nikki: WikidataUsefuls [09:08:51] nikki: you can personalize it [09:09:19] I disable it frequently but when I know I'll work on a theme, I reactive it [09:10:04] harmonia: can you give a theatre item? [09:10:36] hm, isn't that the one where you need to give it the value too? if so, that's no good :( [09:10:44] an incomplete one, I suppose? [09:10:50] Yeah. [09:12:51] sjoerddebruin: https://www.wikidata.org/wiki/Q25345358 [09:14:49] yeah, also close to movies [09:15:14] i don't know what properties to use [09:15:51] I really do just want to influence the properties listed... I would write my own gadget but I have absolutely no idea how to change what the property suggester lists >_< [09:16:31] sjoerddebruin: I just listed the properties on the Wikiproject page but i haven't written a full tutorial on what to use and when yet [09:17:36] Link? "wikiproject plays" gives me nothing [09:22:09] sjoerddebruin: Theatre [09:24:13] yeah, useless suggestions [09:24:32] the suggester should take the value of P31 more serious [09:26:18] ^yep [09:28:34] well it is better now than it was two years ago [09:28:57] when we couldn't even find "female"/"male" when we searched a value for "gender" [09:28:59] but still [09:29:03] it could be better [09:29:38] well that is the suggester for values, those come from a other database [10:08:46] https://tools.wmflabs.org/wikishootme/ is pretty grey. [10:11:13] ah, there it is... [10:21:28] that page needs a key :/ [10:24:24] key? [10:26:11] showing what all the dots mean [10:26:27] I guess legend would be another word [10:26:52] Top right is a button [10:28:10] it doesn't seem to include the light green circles [10:28:48] also it won't stay open >_< [10:30:54] what kind of device are you? :O [10:31:34] I'm a person :P but I'm using a normal computer [10:32:03] ::P [10:32:24] nikki: do you have a reference for your P31:Q5? :p [10:32:43] we don't accept controbersial statements without references [10:32:44] :p [10:33:09] hah :P [10:50:46] https://www.wikidata.org/wiki/Q21279585#P527 :3 [11:16:59] sjoerddebruin: what exactly is going on there [11:17:29] "there are four football pitches with grass and two with astroturf"? [11:17:34] yes [11:18:43] I could make separate items for them, but that makes it harder to add those combined courts. [11:45:12] Glorian_WMDE: http://www.wikihow.com/Peel-a-Banana [11:46:09] 🤔 [14:10:08] why do people who haven't read a pitch for a property proposal answer negatively while showing they didn't understand a single part of the pitch ? [14:11:11] i'm proposing two properties related to territorial intersection, and several people say we could just use inclusion properties :/ have they never seen a venn diagram ? [14:12:51] Alphos: maybe not [14:13:05] not everyone on WD know ontologies [14:13:40] never fear, i just replied pretty much the same thing as the pitch and my blogpost intro [15:36:20] Is there a tool which can isplay a map with all items with gps coordinates, to see which countries has most/least item coverage? [15:48:58] Josve05a: Query Service maybe? [15:50:01] hmm, yeah. But was more thinking of a visual thing...I'll keep looking or resort to making up something myself... [15:50:21] Josve05a: http://addshore.com/2015/04/wikidata-map-visualizations/ and https://commons.wikimedia.org/wiki/Category:Wikidata_visualizations [15:50:36] ooh [15:50:37] there are three million coordinates on wikidata: http://tinyurl.com/j3yfmkr – I don’t think the average browser could handle that [15:50:50] https://commons.wikimedia.org/wiki/File:Wikidata_Map_April_2016_Normal.png :) [15:51:34] If you go somewhat larger, you can see the best country. ;) https://upload.wikimedia.org/wikipedia/commons/d/d4/Wikidata_Map_April_2016_Huge.png [15:52:28] Josve05a: well, WDQS does provide some sort of visualisation... [15:53:13] Next month a new map, addshore? [16:34:34] sjoerddebruin: ooooh [16:34:52] I want one on my wall. [16:36:21] sjoerddebruin: https://phabricator.wikimedia.org/T145414 [17:36:02] SMalyshev: stupid question... we have both a production and a master branch on the wdqs deploy repo. Which one is expected to be deployed? [17:38:27] SMalyshev: forget it, production is an unpruned branch... [17:52:58] PROBLEM - WDQS HTTP on wdqs1001 is CRITICAL: HTTP CRITICAL: HTTP/1.1 503 Service Temporarily Unavailable - 387 bytes in 0.014 second response time [17:53:37] PROBLEM - WDQS SPARQL on wdqs1001 is CRITICAL: HTTP CRITICAL: HTTP/1.1 503 Service Temporarily Unavailable - 387 bytes in 0.003 second response time [17:55:54] ^ that's me... [18:01:08] gehel: production is the branch [18:01:21] sorry missed the message [18:02:10] gehel: looks like template is not deployed on wdq1 :( [18:02:29] SMalyshev: fyi https://phabricator.wikimedia.org/T145426 and https://phabricator.wikimedia.org/T145424 [18:02:39] SMalyshev: yep, template has been created in /srv/deployment/wdqs/wdqs-cache/revs/49f0f8bfdf7cb545394c36bf0011667cdb9715e7/.git/config-files/srv/deployment/wdqs/wdqs/RWStore.properties but not symplinked where it should [18:03:14] SMalyshev: rolling back to HEAD^1 for the moment, I'll check with tyler later (he's busy atm) [18:03:19] gehel: hmm weird. Can it be that it can not overwrite existing files? [18:04:23] SMalyshev: that's my hypothesis... [18:05:36] gehel: maybe let's try renaming that file in deploy repo and see if it works then [18:06:16] DanielK_WMDE: thanks will look [18:07:27] DanielK_WMDE: for T145426 we need either some tool that can read RDF (WDTK unfortunately can't) or T144103 [18:07:28] T144103: Create .nt (NTriples) dumps for wikidata data - https://phabricator.wikimedia.org/T144103 [18:07:28] T145426: Provide a way to add new unit normalizations to the query service without a full reload - https://phabricator.wikimedia.org/T145426 [18:08:22] RECOVERY - WDQS HTTP on wdqs1001 is OK: HTTP OK: HTTP/1.1 200 OK - 9706 bytes in 0.015 second response time [18:08:48] Did I miss some new English language rules? https://lists.wikimedia.org/pipermail/wikitech-ambassadors/2016-September/001471.html [18:08:54] SMalyshev: why can't we just run a sparql query to find values that have a "new" unit, but no normalized value associated? [18:09:01] RECOVERY - WDQS SPARQL on wdqs1001 is OK: HTTP OK: HTTP/1.1 200 OK - 9706 bytes in 0.010 second response time [18:09:27] SMalyshev: we can do that with LIMIT 100, add normalized values for the ones we found, and rinse and repeat until nothing is found any more. [18:09:32] DanielK_WMDE: not sure what you mean by "new unit"? [18:09:52] SMalyshev: we added a conversion for unit foo. We can query for all values that use unit foo, right? [18:10:08] sjoerddebruin: lol, they sent the Czech version, my translation :-P [18:10:34] DanielK_WMDE: that would be a lot of values. Probably in hundreds of thousands. Updating that many entities at once would not be fast. [18:10:56] DanielK_WMDE: note that we don't have any code to "add normalized values", we just have code to "update whole entity" [18:11:02] SMalyshev: yea, so do small chunks. 100 at a time. the sparql query can exclude values that already have a normalized value [18:11:14] yea, i know - hence the two options in the ticket [18:11:22] SMalyshev: I need to take a short break. Rollback completed, wdqs up and running with latest code, just missing the scap changes [18:11:32] we can either update the entire entity, or write specialiezd code to just insert the triples we need [18:13:30] DanielK_WMDE: right. so finding those triples is the trick. If we have code to find those triples, we don't even need sparql - we can just run it on dump and bulk-load the result [18:14:45] SMalyshev: runnign that on an rdf dump would be tricky, because there is no locality to the data. you would have to track a lot of state. [18:14:53] SMalyshev: could be done with a JSON dump though [18:15:04] or based on sparql (why not?) [18:15:46] DanielK_WMDE: I don't think we'd need too much state, only to remember value IDs maybe. But I didn't write the code so I'm not sure. Also, doing it on .ttl dump would suck, .nt would be way better [18:16:03] DanielK_WMDE: that'd require JSON->RDF thing which we don't have [18:17:09] it would be a partial JSON->RDF converter, really [18:17:22] with what we have I *think* the easiest way is a) .nt dump b) extractor script that goes over it and extracts statements relevant for the task. I'll need to look more into it to be sure though [18:17:23] a full JSON->RDF converter would be pretty easy to do. btw [18:17:34] DanielK_WMDE: I'm afraid partial is not easier than full one :) [18:17:41] no, but more useful [18:17:47] it would be nearly the same code [18:18:10] a full converter can probably be done in 100 lines. 200 if you want nice parameters etc. plus 500 lines of tests ;) [18:18:36] (note that JSON dumps can be read line by line. we guarantee that each line is one entity.) [18:18:50] so, read line, decode into entity, generate rdf, repeat [18:18:55] hmm... I doubt it'd be as easy as 100 lines, we have a lot of infrastructure that needs to be set up if we use existing rdfbuilder code [18:19:18] DanielK_WMDE: that assumes JSON has full entity? [18:19:26] yes. [18:19:30] it always does [18:19:35] we don't have partial JSON [18:20:34] but we don't need full entity... not sure that existing flavor flags are selective enough to generate that [18:20:54] probably can be done but not sure how easy. I feel I need to experiment some on it [18:21:41] we just need to generate normalized values and statements linking to them [18:21:54] I kind of think doing it based on SPARQL would be easier to use. slightly harer to code. [18:22:11] it has the advantage that we can just run the script until the triple store is consistent [18:22:14] let me see [18:22:29] with a JSON dump, you are always out of sync, and miss some statements (or include deleted or changed ones) [18:22:47] so we have about 300 convertable units right now, and about half of those are actually used [18:22:58] SMalyshev: no need to rush, i added these tasks so we don't forget, not because they need to be doen NOW [18:23:54] I want to just estimate how many items use, say, metre [18:24:08] we have 12k values using metre [18:25:48] SMalyshev: that's not horrible [18:25:52] but getting to entities from there may not be easy... [18:25:58] but these are the ones that don't need conversion :P [18:26:20] we don't really need to get to entitites. just to statements. [18:26:22] DanielK_WMDE: well, metre is a good top bound [18:26:34] 372596 statements [18:26:43] you think so? try foot, km, and mile [18:26:45] that seems to be a lot... [18:27:14] 12k values are used in 372k statements? huh... really? [18:27:45] mile is 537, foot is 6176, km is 38141 [18:27:58] thanks [18:28:05] so 38k for km. [18:28:13] values, not statements, right? [18:29:15] statements [18:29:22] oh, ok [18:29:56] yea, so... fetching the statements and generating the triples for the normalized value shouldn't be terribly hard, right? [18:30:01] it looks weird for meters, but I looked at results and seems legit... dunno [18:30:36] square kilometre also seems to be very common btw [18:30:42] http://tinyurl.com/zjxhxlx [18:31:20] WikidataFacts: good point [18:31:41] but I’m getting way lower counts with this: http://tinyurl.com/jm7gazf [18:32:08] The most popular unit is the number 1? [18:32:16] with labels: http://tinyurl.com/htmzxd8 [18:32:20] * Harmonia_Amanda wonders what WikidataFacts thought of her little tutorial [18:32:27] sjoerddebruin: I guess that must be “no unit”? [18:32:35] probably [18:32:43] sjoerddebruin: the number 1 is "The Unit" (basically: no unit. a count) [18:33:01] The bottom of the list is interesting. [18:33:01] oh right, the word “unit” also has that meaning :) [18:33:03] "English Wikipedia" [18:33:15] "Same-sex marriage in Connecticut" [18:33:40] how can I fix those easily? [18:33:43] Days (a song) [18:35:34] WikidataFacts: that counts values, not statements, though. Values can be shared [18:35:56] oh, they can? I didn’t know that [18:36:21] yes, same value always produces same node, so "1 meter" is always the same value [18:36:30] same for references btw [18:36:37] SMalyshev: *nearly* always :P [18:36:56] * DanielK_WMDE still likes the dedupe solution [18:37:23] DanielK_WMDE: dedupe doesn't produce different values, it can just produce the same one twice [18:38:09] sjoerddebruin: http://tinyurl.com/zwn54j6 – statements with rare units [18:38:11] SMalyshev: ah, well, yea, it will have the same id, so it's "the same" in the triple store. not the same bytes in the dump. semantics :P [18:38:12] DanielK_WMDE: value ID is still value-hash, so same value => same hash. And since there's no duplicate triples, same value => same node [18:38:24] DanielK_WMDE: exactly :) [18:38:49] WikidataFacts: <3 [18:39:01] for square km, 60k statements [18:39:08] so metre still wins :) [18:39:11] “items with rare units” times out, but fortunately the statement link redirects to the item [18:40:23] DanielK_WMDE: anyway, so if we say items-per-unit is bound by ~10k then we'd still have about a million entities [18:40:42] that's of course very generous top bound, I'd expect real one be tens of thousands at most [18:41:16] SMalyshev: issue with scap is more complex than I thought. See #scap3 for details [18:43:23] SMalyshev: well, that makes all the difference, right? re-importing a million items would suck. 30k would be ok i guess. [18:44:35] DanielK_WMDE: I think if we could extract just the new stuff it'd be better. Bulk-importing a ton of data is mmuch faster than updating items one by one [18:44:52] so I'll look into how to do that [18:45:07] if we just import new data even a million is not that big of a deal [18:45:24] SMalyshev: generating just the new stuff would be a lot better. but why not based on sparql? [18:45:34] i don't see how we could reliably get to a consistent state based on a dump [18:46:04] DanielK_WMDE: what sparql would do? you mean getting ids of items to check? [18:46:45] SMalyshev: no, not items! just get the triples that define the values themselves, and the triples that connect the values to statements. [18:46:48] DanielK_WMDE: it doesn't matter, newly edited entities are ok once we turn it on in Special:EntityData. [18:47:37] you'd still generate triples for values that no longer exist... but oh, well. [18:47:38] DanielK_WMDE: err not sure I understand. we don't have the triples that define values themselves, that's what we are just about to add [18:47:54] that was some creative use again https://www.wikidata.org/w/index.php?title=Q3145241&type=revision&diff=375876876&oldid=302922502 [18:48:01] SMalyshev: we have triples tha tdefine the unconverted value [18:48:15] from that, we can calculate the converted value, no? [18:48:34] sjoerddebruin: yeah people feel the need to add units to a lot of things that shouldn't really have them [18:49:06] DanielK_WMDE: ahh, that's what you mean! hmm... [18:49:08] WikidataFacts: why is Q185648 included in that query? [18:49:50] sjoerddebruin: because of this value deduplication, probably. here’s a better one: http://tinyurl.com/jytaqly [18:50:42] DanielK_WMDE: for units that may work, but will be duplicating the job we already did when dumping. but possible I guess [18:50:43] "passengers per hour per direction" is a real unit, woah [18:50:56] “Jupiter radius”, too (with reference!) [18:51:40] sjoerddebruin: yeah we need to somehow explain the concept of implied unit instead of specifying those as real units [18:52:25] SMalyshev: what will be duplicated? [18:52:46] DanielK_WMDE: the conversion work, we already have those converted values in the dump [18:53:09] if we already made a new dump, yes [18:53:12] but so what [18:58:25] ah, WikidataFacts already found the mojito item [18:58:34] that seems like a legit use of leaf as a unit though :P [18:58:41] yeah, I love it :D [19:06:07] ^^ [19:06:29] what to do with https://www.wikidata.org/wiki/Q16021 [19:07:02] Harmonia_Amanda: how would you like to approve a property creation proposal that has no objections? :] [19:07:37] hare: hmm, I'm watching James Bond right now, so in two hours? [19:07:46] Sure [19:08:40] Harmonia_Amanda: I like the tutorial… the only comment I have is that I wouldn’t have included the PREFIX [19:09:26] WikidataFacts: that's because Bob DuCharme was "angry" at me for skipping PREFIX in my previous blog article [19:09:34] oh, okay [19:09:43] so i thought I would explain it one [19:09:45] once* [19:09:54] prefixes are only needed if you don't use the WDQS UI right? [19:10:23] if you use an endpoint who doesn't declare them already [19:12:12] It might be useful to explain that PREFIX is a SPARQL thing that you normally need, but the WDQS UI does it for you, or something [19:12:38] hare: that's what I did [19:13:05] sjoerddebruin: more precisely, prefixes is only needed if you don't use WDQS :) [19:13:41] sjoerddebruin: if you query WDQS, by whatever means there are, standard prefixes come built in. You will still need to define non-standard ones though [19:13:43] hare: https://blog.ash.bzh/sunday-query-the-200-oldest-alive-french-actresses/ [19:14:12] gehel: so, do we have any resolution for scap issue? [19:15:50] SMalyshev: the issue is generating the file in the project directory [19:16:29] So we could generate it outside and have the startup script pass an absolute path to the config file [19:16:29] Okay. https://www.wikidata.org/w/index.php?title=Q15720641&type=revision&diff=257720886&oldid=224915100 [19:17:07] That would be more levels of indirection than I would like, but it should work [19:17:43] Scap3 will support config file on the project dir in the future, so we can cleanup at that point. [19:19:59] 19.000.000 albums of Kylie Minogue is not close I think. https://www.wikidata.org/w/index.php?title=Q13983&type=revision&diff=375884335&oldid=370487757 [19:27:41] +/-1 year...I hae that +/-1 thing on all "number" properties... [19:27:44] hate* [19:33:54] Josve05a: it's going away soon! <3 [19:33:59] yay [19:35:09] hmmmm, how can we improve this... https://www.wikidata.org/wiki/Q1635189 [19:36:34] sjoerddebruin: I keep going through that list and finding entries you already fixed :D [19:37:25] ugh, flickrwashing https://www.flickr.com/photos/95128916@N00/14440518720 [19:37:29] WikidataFacts: annoying huh [19:37:40] WikidataFacts: list? [19:37:52] Josve05a: http://tinyurl.com/zymf6gt [19:38:04] ah :p [19:38:16] a thousand High and Mighty Color songs… https://www.wikidata.org/w/index.php?title=Q24184996&type=revision&diff=375888216&oldid=371863889 [19:38:48] So, who wants to go to the kitchen and take a pinch of salt? [19:39:36] similarly: dash https://www.wikidata.org/wiki/Q26869687 (used in Bloody Mary: https://www.wikidata.org/wiki/Q207711) [19:40:08] Hm, I already seen that one before. [19:40:54] it’s kinda weird how much of that list is just obscure units being expressed in other obscure units [19:41:25] there is a lack of statements in food [19:43:24] Like, how can I indicate that a cocktail is served on the rocks? [19:43:26] Bonsoir. Je cherche qui pourrait créer les propriétés dont j'ai demandé la création (et qui ont été acceptées). La liste est ici : https://www.wikidata.org/wiki/Wikidata_talk:WikiProject_Equines [19:45:19] I hope I guessed this unit correctly – but I’m pretty sure Cameroon was wrong: https://www.wikidata.org/w/index.php?title=Q24276051&type=revision&diff=375889906&oldid=371721071 [19:46:15] WikidataFacts: that statement was correct, the painting was 143 times the length of that country [19:46:25] oh, silly me, I should’ve thought of that [19:48:10] gehel: but if it's outside the directory then puppet can make it [19:49:42] SMalyshev: yes, but the goal was to keep the content of that file close to the code, to have a single deployment unit [19:50:07] Which we loose if we move the content under Puppet control [19:50:29] yeah true [19:55:29] gehel: so, you'll make a patch? [19:59:45] SMalyshev: yep, I'll take care of that [19:59:54] gehel: cool, thx [20:04:19] no problem! [20:19:00] I'm gonna merge 3 items in to one :O [20:20:13] done. https://www.wikidata.org/w/index.php?title=Q1438321&action=history [20:20:22] jura's page is great! https://www.wikidata.org/wiki/User:Jura1/test32 [20:32:10] I'm interested in creating Wikidata items on industrial sectors, based on their NAICS code: https://www.naics.com/naics-drilldown-table/? [20:32:15] Any recommendation on how to proceed with this? [20:34:03] For context, there are a bunch of papers on Wikidata that, in the database I am working with, are associated with NAICS codes (and SIC codes, which is the predecessor system), and I think it would be useful to annotate the Wikidata items with this information [20:35:09] The way I would imagine it would work is, I would request properties for NAICS and SIC codes, and then associate them with existing items, and create new ones as necessary [20:35:18] Each item would be "instance of industrial sector" or something like that [20:36:36] The issue is that a lot of the categories are these amalgamations, like "investment banking and securities dealing" [20:37:19] While Wikidata would consider investment banking and securities trade to be, technically, two different concepts, NAICS likes to combine things together. So I would create an item representing the consolidated concept for NAICS purposes and then state that it consists of parts X and Y. Does that sound right? [21:08:29] i think i've hit a bit of a snag with blazegraph again https://query.wikidata.org/#ASK%20%7B%0A%20%20wd%3AQ618123%20wdt%3AP279%2B%20wd%3AQ5%20.%0A%20%20%7D [21:08:45] how are geographical entities subclasses of "human being" ? :/ [21:09:20] (subclasses of subclasses of … but still) [21:12:03] Alphos: hmm shouldn't be happening. I don't think it's a real result. it looks like a bug [21:12:15] please submit an issue to phabricator [21:12:18] phab+T then ? :) [21:12:21] right on it :) [21:18:14] So many dupes! :O https://www.wikidata.org/wiki/Special:Contributions/Josve05a [21:19:12] Good job! [21:23:44] I can't gt https://www.wikidata.org/wiki/User:Josve05a/dupes to work :/ [21:23:46] get* [21:25:15] sparql must be one line afaik [21:25:45] no... [21:26:06] (at least for listeria) [21:26:34] I copied the code from https://www.wikidata.org/wiki/User:Jura1/test32 but changed the yearfrom one years limit to 300...so it would search for all people between 1700 and now...seems like it is hitting the "timeout" :/ [21:26:53] so it should work, just not with this "big" query... :/ [21:27:12] Let's see if {{Wikidata list end}} is required. ;) [21:27:13] 10[3] 10https://www.wikidata.org/wiki/Template:Wikidata_list_end [21:27:43] :( [21:28:30] WQS: http://tinyurl.com/jr4vj6y : Query timeout limit reached [21:28:47] Probably too broad yeah [21:29:10] meh [21:29:31] ok...will do another way...will take a year to do this however... [21:31:21] damn...can't do {{#expr:{{SUBPAGENAME}}+1}} within templates... grunt [21:31:21] 10[4] 04https://www.wikidata.org/wiki/Template:SUBPAGENAME [21:32:22] SMalyshev : done :) [21:32:32] thanks [21:32:37] no, thank you :) [21:36:47] sjoerddebruin: creating yearly/decade subpages https://www.wikidata.org/wiki/User%3AJosve05a%2Fdupes%2F2000s [21:37:01] yeah, that's the only way afaik :( [21:37:31] WikidataFacts : https://www.wikidata.org/wiki/Q15720641 :p [21:37:44] without even using wdqs, but i cheated, i'm french ;p [21:39:23] Alphos: yeah, I have no clue if that’s a well-known thing in France [21:39:34] for all I know, it’s as famous as “all of Gaul? No!” [21:39:37] trust me, it is ^^ [21:39:41] good to know :) [21:39:56] it's about as famous as "all of gaul" indeed ^^ [21:40:11] ok, maybe not that much, but at most one order of magnitude less [21:41:13] WikidataFacts second one is positive integers, innit ? [21:42:00] or negative ones, but not sure if that set has an item ^^ [21:42:11] Alphos: that would probably also be correct, but the actual item is https://www.wikidata.org/wiki/Q1193504 [21:42:40] but https://www.wikidata.org/wiki/Q21199 has “cardinality of this set: aleph null” [21:42:49] hmmm...sjoerddebruin do you know if I can do multiple such list on the same wikidata page? [21:42:53] lists* [21:43:13] you could transclude them on one I think? [21:43:16] * Alphos jumps to oeis ! [21:43:38] "At this moment, only one list per page is possible." [21:44:03] * Josve05a grunts [21:44:27] WikidataFacts now i'm wondering if they were described by ramanujan, given their name >_> [21:47:33] interesting http://lu.is/blog/2016/09/12/copyleft-and-data-database-law-as-poor-platform/ [21:49:39] WikidataFacts even more efficient, but won't allow a fallback simply : http://tinyurl.com/hrxabhp [21:50:35] 1.6 seconds seems about equivalent to the label service version… [21:52:58] i had 1.4 instead of 1.8 ;) [21:53:11] and fallback http://tinyurl.com/jn5bjzc 1.7 [21:53:42] but rdfs:label will be usually measurably faster than SERVICE{} [21:53:52] oooh, COALESCE looks very useful [21:54:11] I’ll have to remember to use that instead of IF(BOUND(?x),?x,IF…) chains [21:54:12] thanks :) [21:54:26] ^_^ [21:55:13] and I think you only notice the label service’s slowness with >1000 labels [21:55:27] 1.4 vs 1.8 is in range of normal variation, I think [21:55:30] in exchange (but another time, i'm really exhausted, just finishing reading my twitter feed) you'll have to explain me how and when to use [] ^^ [22:03:52] Alphos: you can try to read this later :) http://tinyurl.com/gne7m6j (don’t pay attention to the actual wdt: and wd:, I just threw random entities together) [22:04:43] but i thought i saw nested [] at some point :x [22:04:53] and it might have been in one of yours too [22:05:08] yes, you can nest it as deeply as you want [22:05:09] hang on [22:05:55] by the way, one result : https://www.wikidata.org/wiki/Q4934675 <3 [22:05:57] oh yeah, this one’s a bit large, huh https://twitter.com/WikidataFacts/status/773248512167469056 [22:06:08] Alphos: LIMIT 1, because I don’t care about the results :D [22:06:33] oh right, in the last one, each of them is a distinct dummy ? [22:06:47] but the issue is you can't reuse said dummies, right ? [22:06:54] (dummy ~ tmp ) [22:07:18] yes, exactly, they’re all unnamed [22:08:07] and by the way, http://tinyurl.com/zcb5yee i fixed seconds * 60 = minutes to seconds / 60 = minutes ;p [22:08:31] http://tinyurl.com/jgyg4kv [22:08:54] …whoops [22:09:09]