[01:13:45] (03CR) 10Hoo man: "recheck" [extensions/Capiunto] - 10https://gerrit.wikimedia.org/r/213556 (owner: 10Hoo man) [02:01:37] (03PS1) 10Hoo man: Generalize the action=query&meta=wikibase API, expose siteid [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/214517 [02:18:40] (03CR) 10Hoo man: [C: 032] Default mw.wikibase.* lua functions to the connected item of a page [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/210907 (https://phabricator.wikimedia.org/T99078) (owner: 10Aude) [02:22:02] (03Merged) 10jenkins-bot: Default mw.wikibase.* lua functions to the connected item of a page [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/210907 (https://phabricator.wikimedia.org/T99078) (owner: 10Aude) [02:23:29] (03PS2) 10Hoo man: Define resource loader module position explicitly [extensions/Capiunto] - 10https://gerrit.wikimedia.org/r/214490 (owner: 10Bene) [02:26:26] (03CR) 10jenkins-bot: [V: 04-1] Define resource loader module position explicitly [extensions/Capiunto] - 10https://gerrit.wikimedia.org/r/214490 (owner: 10Bene) [02:27:16] (03CR) 10Hoo man: [C: 04-1] "This changes the loading position from to to bottom, so it's not just making something explicit." [extensions/Capiunto] - 10https://gerrit.wikimedia.org/r/214490 (owner: 10Bene) [02:27:36] 10MediaWiki-extensions-WikibaseClient, 10Wikidata, 3Wikidata-Sprint-2015-05-05: default mw.wikibase.label() to the connected item of a page - https://phabricator.wikimedia.org/T99078#1319527 (10hoo) 5Open>3Resolved [02:33:47] (03CR) 10Hoo man: [C: 032] "Will remove the setting in a follow up" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/212787 (owner: 10Thiemo Mättig (WMDE)) [02:36:25] (03Abandoned) 10Hoo man: Cache supported entity data formats [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/206734 (owner: 10Hoo man) [02:37:49] (03Merged) 10jenkins-bot: usePropertyInfoTable is not a repo but a lib setting [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/212787 (owner: 10Thiemo Mättig (WMDE)) [02:44:54] (03PS1) 10Hoo man: Remove the usePropertyInfoTable setting [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/214518 [02:44:56] (03PS1) 10Hoo man: Remove the DummyPropertyInfoStore [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/214519 [02:48:13] (03Abandoned) 10Hoo man: Make label/description uniqueness checks case insensitive for properties [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/206386 (https://phabricator.wikimedia.org/T97129) (owner: 10Hoo man) [02:53:39] (03PS8) 10Hoo man: Rename SiteLinkLookup::getEntityIdForSiteLink to getItemIdForSiteLink [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/202715 [02:54:39] (03CR) 10Hoo man: "Manually rebased (per Jeroen)" [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/202715 (owner: 10Hoo man) [03:01:38] 10Wikidata, 10MediaWiki-Maintenance-scripts, 10MediaWiki-Sites: Investigate exportSites.php and importSites.php behaviour regarding site identifiers - https://phabricator.wikimedia.org/T100750#1319532 (10hoo) 3NEW [03:02:14] 10Wikidata, 10MediaWiki-Maintenance-scripts, 10MediaWiki-Sites: Investigate exportSites.php and importSites.php behaviour regarding site identifiers - https://phabricator.wikimedia.org/T100750#1319540 (10hoo) [03:05:38] 10Wikidata, 10Wikimedia-Site-requests: Change the language code of bhwiki from bh to bho in the site and site_identifiers table. - https://phabricator.wikimedia.org/T100751#1319541 (10hoo) 3NEW [03:06:03] 10Wikidata, 10MediaWiki-Maintenance-scripts, 10MediaWiki-Sites: Investigate exportSites.php and importSites.php behaviour regarding site identifiers - https://phabricator.wikimedia.org/T100750#1319532 (10hoo) [03:06:05] 10Wikidata, 10Wikimedia-Site-requests: Change the language code of bhwiki from bh to bho in the site and site_identifiers table. - https://phabricator.wikimedia.org/T100751#1319541 (10hoo) [04:06:40] 10Wikidata, 6Release-Engineering, 10Wikidata-Page-Banner: Setup Jenkins - https://phabricator.wikimedia.org/T100495#1319614 (10Jdlrobson) No jenkins bot on https://gerrit.wikimedia.org/r/#/c/213576/ Thus amongst other things I have to submit patch sets. [04:08:33] 10Wikidata, 6Release-Engineering, 10Wikidata-Page-Banner: Setup Jenkins - https://phabricator.wikimedia.org/T100495#1319615 (10Jdlrobson) (Note I also have a bug somewhere about setting this up automatically for new extensions) [04:44:44] 10Wikidata, 3Roadmap, 7notice: Deploy Wikibase usage tracking to remaining s7 wikis - https://phabricator.wikimedia.org/T100659#1319693 (10Bugreporter) [04:49:50] 10Wikidata, 10Wikidata-Page-Banner, 5Patch-For-Review: Convert pagebanner markup to use a template - https://phabricator.wikimedia.org/T98152#1319704 (10Jdlrobson) 5Open>3Resolved a:3Jdlrobson [06:20:40] 10Wikidata, 6Search-and-Discovery, 10Wikidata-Query-Service, 3Search-and-Discovery-Wikidata-Query-Service-Sprint: BlazeGraph uses old xsd:dateTime standard - https://phabricator.wikimedia.org/T94539#1319741 (10Smalyshev) [08:01:35] (03CR) 10Bene: "Ok, I will make this position => top to stay with the current behaviour and just get rid of nasty PHP warnings. We can still check if posi" [extensions/Capiunto] - 10https://gerrit.wikimedia.org/r/214490 (owner: 10Bene) [08:03:35] (03PS3) 10Bene: Define resource loader modules explicitely [extensions/Capiunto] - 10https://gerrit.wikimedia.org/r/214490 [08:06:12] (03CR) 10jenkins-bot: [V: 04-1] Define resource loader modules explicitely [extensions/Capiunto] - 10https://gerrit.wikimedia.org/r/214490 (owner: 10Bene) [08:47:09] 10Wikidata, 10Continuous-Integration-Config, 10Wikidata-Page-Banner: Setup CI configuration for WikidataPageBanner extension - https://phabricator.wikimedia.org/T100495#1320006 (10hashar) p:5Triage>3Normal [08:52:13] (03PS1) 10Dimitri.schmidt: tooltips now clickable and hoverable via javascript [extensions/WikidataQualityConstraints] (v1) - 10https://gerrit.wikimedia.org/r/214580 [09:16:11] 10MediaWiki-extensions-WikibaseRepository, 10Wikidata: wbgetclaims api module should provide resolving of redirects - https://phabricator.wikimedia.org/T97928#1320045 (10Lokal_Profil) >>! In T97928#1318803, @JanZerebecki wrote: > This task is about resolving entity references in the result not about the workin... [09:32:03] 10Wikidata, 3Roadmap, 7notice: Deploy Wikibase usage tracking to remaining s7 wikis - https://phabricator.wikimedia.org/T100659#1320054 (10aude) [09:32:34] (03PS1) 10Hashar: Npm configuration for banana/jshint [extensions/WikidataPageBanner] - 10https://gerrit.wikimedia.org/r/214582 (https://phabricator.wikimedia.org/T100495) [09:34:43] 10Wikidata, 10Continuous-Integration-Config, 10Wikidata-Page-Banner, 5Patch-For-Review: Setup CI configuration for WikidataPageBanner extension - https://phabricator.wikimedia.org/T100495#1320060 (10hashar) a:3hashar [09:34:59] any clue what this WikidataPageBanner is ? [09:35:27] I sent a basic patch that adds an npm entry point to run jshint and lint the i18n files https://gerrit.wikimedia.org/r/214582 [09:52:25] hashar: i think it is to support the header images on wikivoyage (on mobile + desktop) [09:52:32] https://phabricator.wikimedia.org/T93106 [09:52:33] [13WikibaseDataModelJavaScript] 15tobijat pushed 1 new commit to 06master: 02http://git.io/vk41l [09:52:33] 13WikibaseDataModelJavaScript/06master 14f27c4e0 15Tobi Gritschacher: Merge pull request #32 from wmde/1.1.0... [09:52:44] thanks Tobi_WMDE_SW :) [09:54:26] (03PS1) 10Dominic.sauer: Add changes introduced by displaying icons [extensions/WikidataQualityConstraints] - 10https://gerrit.wikimedia.org/r/214585 [09:54:59] [13WikibaseDataModelJavaScript] 15tobijat tagged 061.0.2 at 06master: 02http://git.io/vk4Ml [09:55:09] aude: https://github.com/wmde/WikibaseDataModelJavaScript/releases/tag/1.0.2 [10:00:14] (03PS1) 10WikidataBuilder: New Wikidata Build - 2015-05-29T10:00:01+0000 [extensions/Wikidata] - 10https://gerrit.wikimedia.org/r/214587 [10:01:25] (03CR) 10JanZerebecki: [C: 032] New Wikidata Build - 2015-05-29T10:00:01+0000 [extensions/Wikidata] - 10https://gerrit.wikimedia.org/r/214587 (owner: 10WikidataBuilder) [10:02:51] aude: it is not like we already have a page banner extension :-D [10:03:08] aude: anyway I got some basic CI config :} [10:05:51] (03Merged) 10jenkins-bot: New Wikidata Build - 2015-05-29T10:00:01+0000 [extensions/Wikidata] - 10https://gerrit.wikimedia.org/r/214587 (owner: 10WikidataBuilder) [10:56:33] 10Wikidata, 3Roadmap, 7notice: Deploy arbitrary access to Wikiquote and Wikivoyage - https://phabricator.wikimedia.org/T100774#1320284 (10aude) 3NEW a:3aude [11:00:35] 10Wikidata, 3Roadmap, 7notice: Deploy arbitrary access to s7 wikis (arwiki, cawiki, eswiki, huwiki, kowiki, rowiki, ukwiki, viwiki) - https://phabricator.wikimedia.org/T100776#1320309 (10aude) 3NEW a:3aude [11:46:22] (03PS3) 10Dominic.sauer: Add changes for showing icon and messages in UI on items page. [extensions/WikidataQualityExternalValidation] - 10https://gerrit.wikimedia.org/r/214329 [11:46:31] (03PS23) 10Dominic.sauer: WIP add js script for showing icons on claims [extensions/WikidataQuality] - 10https://gerrit.wikimedia.org/r/207764 [11:47:24] (03CR) 10jenkins-bot: [V: 04-1] WIP add js script for showing icons on claims [extensions/WikidataQuality] - 10https://gerrit.wikimedia.org/r/207764 (owner: 10Dominic.sauer) [11:51:02] (03CR) 10Dominic.sauer: [C: 032 V: 032] Fix for older php version. [extensions/WikidataQualityExternalValidation] - 10https://gerrit.wikimedia.org/r/213824 (owner: 10Soeren.oldag) [11:52:54] (03PS4) 10Dominic.sauer: Add changes for showing icon and messages in UI on items page. [extensions/WikidataQualityExternalValidation] - 10https://gerrit.wikimedia.org/r/214329 [12:05:49] (03CR) 10Hoo man: [C: 032 V: 032] Define resource loader modules explicitely [extensions/Capiunto] - 10https://gerrit.wikimedia.org/r/214490 (owner: 10Bene) [12:06:57] Lydia_WMDE: https://wikitech.wikimedia.org/wiki/Deployments [12:08:52] 10Wikidata, 3Roadmap, 7notice: Deploy arbitrary access to s7 wikis (arwiki, cawiki, eswiki, huwiki, kowiki, rowiki, ukwiki, viwiki) - https://phabricator.wikimedia.org/T100776#1320446 (10aude) [12:08:54] 10Wikidata, 3Roadmap, 7notice: Deploy arbitrary access to Wikiquote and Wikivoyage - https://phabricator.wikimedia.org/T100774#1320447 (10aude) [12:08:57] 10MediaWiki-extensions-WikibaseClient, 10Wikidata, 3Roadmap, 7Tracking, 7notice: Track Wikidata entity usage on client pages - https://phabricator.wikimedia.org/T49288#1320445 (10aude) [12:15:49] 10Wikidata: don't take case into consideration for item labels - https://phabricator.wikimedia.org/T97129#1320460 (10matej_suchanek) [12:27:02] Lydia_WMDE: https://wikitech.wikimedia.org/wiki/Deployments#Long_term_callouts :) [12:30:28] 10Wikidata, 3Roadmap, 7notice: Deploy usage tracking to s3 wikis (part 1, aawiki - lvwiki) + frwiki + jawiki - https://phabricator.wikimedia.org/T100783#1320471 (10aude) 3NEW a:3aude [12:31:34] 10Wikidata, 3Roadmap, 7notice: Deploy usage tracking to s3 wikis (part 2, maiwiki - zuwiki) + dewiki + ruwiki - https://phabricator.wikimedia.org/T100784#1320480 (10aude) 3NEW a:3aude [12:32:26] 10Wikidata, 3Roadmap, 7notice: Deploy usage tracking to enwiki + s2 wikis - https://phabricator.wikimedia.org/T100785#1320488 (10aude) 3NEW a:3aude [12:33:18] 10Wikidata, 3Roadmap, 7notice: Deploy arbitrary access to s3 wikis (part 1, aawiki - lvwiki) + frwiki + jawiki - https://phabricator.wikimedia.org/T100786#1320495 (10aude) 3NEW a:3aude [12:34:09] 10Wikidata, 3Roadmap, 7notice: Deploy arbitrary access to s3 wikis (part 2, maiwiki - zuwiki) + dewiki + ruwiki - https://phabricator.wikimedia.org/T100787#1320502 (10aude) 3NEW a:3aude [12:35:01] 10Wikidata, 3Roadmap, 7notice: Deploy arbitrary access to enwiki + s2 wikis - https://phabricator.wikimedia.org/T100788#1320509 (10aude) 3NEW a:3aude [12:43:34] 10Wikidata, 3Roadmap, 7notice: Deploy usage tracking and arbitrary access to zhwiki - https://phabricator.wikimedia.org/T98286#1320540 (10aude) tentatively we have deployment of usage tracking to zhwiki on July 28 and arbitrary access on August 18. if handling of language variants is ready sooner, we could... [13:22:27] (03CR) 10Amire80: [C: 031] Generalize the action=query&meta=wikibase API, expose siteid [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/214517 (owner: 10Hoo man) [13:22:52] (03CR) 10Amire80: "Thanks. This will be useful for ContentTranslation, among other things." [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/214517 (owner: 10Hoo man) [13:34:09] 10Wikidata: Suggester styling issues on test.wikidata - https://phabricator.wikimedia.org/T100794#1320602 (10aude) 3NEW [13:36:23] 10Wikidata: Suggester styling issues on test.wikidata - https://phabricator.wikimedia.org/T100794#1320617 (10aude) also the rank selector is not styled correctly. I can reproduce that locally but not on beta.wikidata. [13:41:56] 10Wikidata: Suggester styling issues on test.wikidata - https://phabricator.wikimedia.org/T100794#1320629 (10aude) now can reproduce the styling issue in the rank selector on beta [14:08:44] (03PS1) 10Dominic.sauer: Add overwriteException flag to save violations in order to do not overwrite status of violations with status exception. [extensions/WikidataQuality] - 10https://gerrit.wikimedia.org/r/214611 [14:09:45] (03CR) 10jenkins-bot: [V: 04-1] Add overwriteException flag to save violations in order to do not overwrite status of violations with status exception. [extensions/WikidataQuality] - 10https://gerrit.wikimedia.org/r/214611 (owner: 10Dominic.sauer) [14:22:48] JeroenDeDauw: oh my autoloader problem went away when i upgraded to composer alpha 10. it was a result of the older composer version. [14:26:48] (03PS1) 10Aude: Remove position=top from rank selector style module [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/214615 (https://phabricator.wikimedia.org/T100798) [14:51:49] (03PS1) 10Aude: Separate module definition for addModuleStyles [extensions/Wikibase] (wmf/1.26wmf6) - 10https://gerrit.wikimedia.org/r/214619 (https://phabricator.wikimedia.org/T97410) [14:55:59] (03PS1) 10Soeren.oldag: MwTimestamp is now used instead of DateTime for import date in DumpMetaInformation. [extensions/WikidataQualityExternalValidation] - 10https://gerrit.wikimedia.org/r/214622 [16:04:01] (03PS1) 10Soeren.oldag: Made comparer more robust. [extensions/WikidataQualityExternalValidation] (v1) - 10https://gerrit.wikimedia.org/r/214635 [16:06:24] (03CR) 10Soeren.oldag: [C: 032 V: 032] Made comparer more robust. [extensions/WikidataQualityExternalValidation] (v1) - 10https://gerrit.wikimedia.org/r/214635 (owner: 10Soeren.oldag) [16:28:39] 10Wikidata, 3Roadmap, 7notice: Enable arbitrary access on Wikisource and itwiki projects - https://phabricator.wikimedia.org/T98756#1321199 (10greg) [17:20:48] 10Wikidata, 10Wikidata-Quality-External-Validation, 5Patch-For-Review, 3Wikidata-Sprint-2015-05-05: Setup CI for WikidataQualityExternalValidation - https://phabricator.wikimedia.org/T97528#1321270 (10JanZerebecki) a:3JanZerebecki [17:20:57] 10Wikidata, 10Wikidata-Quality-Constraints, 5Patch-For-Review, 3Wikidata-Sprint-2015-05-05: Setup CI for WikidataQualityConstraints - https://phabricator.wikimedia.org/T97527#1321272 (10JanZerebecki) a:3JanZerebecki [17:49:25] is it a bad idea to merge https://www.wikidata.org/wiki/Q623334 and https://www.wikidata.org/wiki/Q17560541 ? [17:53:24] I'm not sure what the latter is supposed to represent really, but they seem like they would be separate things, since multiple file formats can use the same extension [17:54:51] neither are about a single file format; both are about executable file formats that are used with the .exe extension, as far as I can tell [17:55:14] or, well, one might be about the file extension as used for executables [17:57:41] it's kind of annoying how we have to have both Q778586 (executable) and Q17560478 (executable file format) [17:57:52] for example [17:58:59] I think it's a good question though, the more I think about it, the more confused I get :P [17:59:59] there is p1195 for the extension and two of the three things using it already have that, so I do wonder if q17560541 really serves a useful purpose [18:00:54] wait, didn't I just paste two items related to .exe files? [18:03:05] err I mean there are three things using q17560541, which seems to just mean an executable (one of the two you linked... or both?) with p1195 set to "exe" [18:03:23] see, I'm confused :P [18:05:57] oh, you mean like the ones about MZ, New Executable, and Portable Executable? [18:06:55] yeah [18:07:10] 10Wikidata, 3Wikidata-Sprint-2015-05-05: when the autoloader of Wikibase is loaded the Mediawiki installer fails - https://phabricator.wikimedia.org/T100813#1321380 (10JanZerebecki) 3NEW [18:08:41] 10Wikidata, 3Wikidata-Sprint-2015-05-05: when the autoloader of Wikibase is loaded the Mediawiki installer fails - https://phabricator.wikimedia.org/T100813#1321391 (10JanZerebecki) [18:09:51] those are some (most?) of the file formats used for files with names ending in .exe, for DOS, 16-bit Windows, and 32-bit Windows executables respectively (and probably also for various OS/2 things) [18:10:29] 10Wikidata, 3Wikidata-Sprint-2015-05-05: when the autoloader of Wikibase is loaded the Mediawiki installer fails - https://phabricator.wikimedia.org/T100813#1321380 (10JanZerebecki) Quick hack I used to get around this: sed --in-place -e 's/||/\&\&/; s/!//' extensions/Wikibase/Wikibase.composer.php [18:22:24] Lydia_WMDE: hey :) I wanna try to rebuild an existing Infobox on dewiki using wikidata/lua. Do you have an idea which one I should use? [18:48:34] (03CR) 10Gilles: [C: 04-1] "Position is now required for modules added via addModuleStyles." [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/214615 (https://phabricator.wikimedia.org/T100798) (owner: 10Aude) [18:53:26] (03PS1) 10Gilles: Make rank selector CSS selector as specific as it needs to be [extensions/Wikibase] - 10https://gerrit.wikimedia.org/r/214671 (https://phabricator.wikimedia.org/T100798) [19:03:34] (03PS1) 10Sumit: WikidataPageBanner allow named parameters [extensions/WikidataPageBanner] - 10https://gerrit.wikimedia.org/r/214672 [19:09:12] * SamB_laptop just suggested merging an article on eswiki, and fixed up one of the relevant templates, without really being able to read or write spanish ... hopefully the english edit summaries aren't too useless [19:10:24] (because of an interwiki conflict) [19:10:56] (though one of the items didn't actually possess interwiki links) [19:14:51] (03CR) 10Gilles: [C: 031] Separate module definition for addModuleStyles [extensions/Wikibase] (wmf/1.26wmf6) - 10https://gerrit.wikimedia.org/r/214619 (https://phabricator.wikimedia.org/T97410) (owner: 10Aude) [19:26:46] could someone block https://www.wikidata.org/wiki/Special:Contributions/187.246.224.207? [19:36:46] nikki: blocked [19:36:53] thanks! [19:43:05] Is there a command line tool to explore wikidata via its API? [19:49:55] RailRoadRunner: what should it be able to do? [19:50:37] benestar: eg, "tool sun" would return "Q525 star at the centre of the Solar System\nQ328268 Japanese commercial TV station in Kobe, Hyogo Prefecture\n" (where \n is newline) for example. [19:51:11] benestar: For starters, do a wbsearchentities. Then, if given an item (Q*), return its triples. That sort of thing. [19:51:20] so just doing a search on wikidata dn return the descriptions? [19:51:45] do you have an exact usecase? [19:52:18] benestar: Yes, starting with that, and then allowing specified objects to be queried. My use case: I prefer to browse data on the command line, that's all. [19:52:45] benestar: Ultimately stuff like "*,P123,foo" meaning "show me all Q objects whose P123 value is foo" [19:53:22] there is/will be a SPARQL command line tool to query Wikidata [19:53:33] benestar: Perhaps going as far as TREE and CLAIM.... ok, that's waht I want. Is there one now? [19:53:33] that should be able to handle those triples [19:54:17] benestar: One big issue right now is finding the Q identity of your object. Does "label" have a P property number (I suppose it must, but it's not P0 or P1). [19:54:42] benestar: So is there one or will there be one one day? [19:54:52] you mean in SPARQL/RDF triples? [19:54:59] labels are rdfs:label [19:55:20] benestar: Oh, you don't have to give a property name when querying the API? You can use rdfs:label? [19:55:43] if you use the sparql endpoint, yes [19:55:53] the mediawiki wikibase api has much less functionality though [19:56:17] https://www.wikidata.org/w/api.php <- is this the right or wrong endpoint to use for SPARQL? [19:56:49] https://github.com/mariushoch/asparagus [19:57:11] RailRoadRunner --^ that is the ci tool created on the hackathon by hoo|away [19:57:20] http://wdqs-beta.wmflabs.org/ is the url to the sparql endpoint which is still beta [19:57:45] the url you linked only provides access to the mediawiki api which supports editing but no complex queries [19:58:23] benestar: Yes I saw that earlier, but quasi-rejected it since it's indirect... but wikidata links to it so I suppose it's OK. [19:58:50] what do you mean saying "indirect"? [19:59:06] it will once become the official query endpoint [19:59:22] Oh, but wmflabs.org is an independent organization, right? Not part of the Wikimedia Foundation? [19:59:46] wmflabs.org is the server of Wikimedia Foundation for community and testing things [19:59:50] see wikitech.wikimedia.org [19:59:52] benestar: Nope, I'm wrong, ignore that. I thought it was a separate company just offering help [20:02:02] Hmmm, SPARQL seems complicated. [20:02:15] wmflabs.org runs on WMF hardware [20:02:33] Krenair: Thanks, my bad, I realize that now [20:02:51] it's open for any wikimedia developer to run relevant stuff on, and isn't heavily controlled like WMF production [20:03:52] This is NOT the right place to ask basic SPARQL questions, right? [20:06:46] RailRoadRunner: there are some tutorials [20:07:05] * benestar looks [20:07:12] benestar: I know SQL but don't understand how to convert my SQL queries to SPARQL or even if that makes sense. [20:07:22] rather not [20:07:34] SQL is for relational databases but SPARQL is for graph databases [20:07:36] and triple stores [20:08:06] benestar: Right, but (I guess I'm going to ask anyway!) consider: SELECT * FROM wikidata WHERE y='rdfs:label' and z RLIKE 'sun'; [20:08:15] RailRoadRunner: this are frimelle's slides on SPARQL https://commons.wikimedia.org/wiki/File:Wikidata%27s_SPARQL_introduction_presentation.pdf [20:08:20] Literally querying triples based on their x y and z values [20:08:37] yes, you just write x y z then ;) [20:08:49] so eg ?x rdfs:label "sun"@en [20:08:59] SELECT ?x WHERE { [20:09:05] ?x rdfs:label "sun"@en [20:09:06] } [20:09:11] would be the whole query [20:10:04] benestar: asparagus tells me "Query execution failed" [20:10:49] i didn't test that tool so not sure if it works ^^ [20:10:58] just try the web interface which gives you more details [20:11:25] Instead of @en it's the Filter thingy, I guess [20:12:18] frimelle: 'SELECT ?x WHERE { ?x rdfs:label "sun"};' gives me a prompt, and when I hit RETURN, I get "Query execution failed". [20:12:28] benestar: Trying now [20:12:46] FILTER (LANG(?x) = "en") . [20:13:12] frimelle: nope, when having a string you have to add @en [20:13:24] when querying the string you need to filter, that's right [20:13:27] Ah, right! Sorry! [20:13:46] RailRoadRunner: so you have to add "sun"@en [20:13:52] to specify the language [20:14:04] benestar: Oh, right, because lang= is a required field even in the wikidata API [20:14:19] sure, how should the tool know you speak english and not french? [20:14:40] SELECT ?x WHERE { ?x rdfs:label "Sun"@en } [20:14:42] frimelle: btw is there a way to disable case sensitivity in SPARQL? [20:14:54] and don't forget the prefixes... [20:14:58] Ja, that's what I just thought [20:15:06] I clicked "Execute" and its just hanging [20:15:14] It turns out adding a semicolon at the end breaks it [20:15:15] I have to look it up honestly, benestar [20:15:22] np [20:15:47] in production one will perhaps want to deal with Q-ids anyways [20:15:58] label/alias resolving is just not reliable enough [20:16:53] "Total results: 0, duration: 131 ms. Click on * to explore." [20:17:17] Oh never mind, needed "Sun" capitalized. [20:19:21] aude: when will you merge my patches? :) [20:19:34] "SELECT ?z WHERE {Q525 P398 ?z}" <- doesn't appear to work? [20:22:15] note that labels are not unique... [20:22:44] RailRoadRunner: missing some prefixes? [20:23:03] SMalyshev: How do you mean? Should I use the "add prefixes" button? [20:23:42] SMalyshev: You're saying Q525 does not uniquely identify the Sun? [20:23:51] RailRoadRunner: http://tinyurl.com/pgc9eoq is probably what you want [20:24:27] (of course, you can cut extra prefixes: PREFIX wd: [20:24:27] PREFIX wdt: [20:24:27] SELECT ?z WHERE {wd:Q525 wdt:P398 ?z} [20:24:58] SMalyshev: Oh, I get it! Q525 is the Sun only in wikidata.org's schema, but the query tool lets you use other object id things as well [20:26:18] RailRoadRunner: in RDF, everything is a URI (except for literals). So wd:Q525 is the same as http://www.wikidata.org/entity/Q525 [20:26:37] SMalyshev: Oh, so "Q525" is meaningless since it's not a URL. [20:26:46] RailRoadRunner: right. [20:27:07] SMalyshev: Can I use property names instead of values? like wdt:"child astronomical body"? [20:27:50] RailRoadRunner: not really. You can do a label lookup (i.e. ask "which property has this label" and then use that property's predicate) but it's much slower [20:28:19] RailRoadRunner: for the engine, everythins is the URL. so if we want to do by label, we need to build external GUI [20:28:26] (which we plan to but not yet) [20:28:33] SMalyshev: OK, so what's the SPARQL for "what property has label 'planet'?" [20:29:14] RailRoadRunner: I don't think there's such property... but let me do the sparql [20:29:21] SELECT ?x WHERE {?x rdfs:label "planet"@en} [20:29:33] But that gave me back an object, not a property? [20:30:27] RailRoadRunner: right. Are you trying to find items that are planets? [20:30:41] then you need P31, aka instance of [20:30:49] I'm trying to find the Sun's planets (pretending I don't know about the property "child astronomical body") [20:31:06] In other words, the P number for the property "planet" [20:31:15] (or hasPlanet or whatever) [20:32:05] RailRoadRunner: well, if you don't know the name of the property or it's P-id, I'm not sure I can write sparql for that [20:32:24] you need to know at least one of them (or look up using elastic search, etc.) [20:32:39] OK, I know the name of the property is "planet", I now want the SPARQL to find its P id, is that possible? [20:32:47] RailRoadRunner: yes [20:32:56] let me show [20:33:20] SELECT ?x WHERE {?x rdfs:alias "planet"@en} (is my guess) [20:33:41] So, RailRoadRunner, http://bit.ly/1eE6G2w would be the label look-up for the property "child astronomical body" [20:34:15] which returns the property P398 [20:34:30] Oh, so I was doing it right, but there just happens to be no property called "planet"? [20:34:46] PREFIX wd: [20:34:46] PREFIX wdt: [20:34:46] PREFIX wikibase: [20:34:46] PREFIX rdfs: [20:34:47] SELECT ?p ?pred WHERE { [20:34:48] ?p a wikibase:Property; rdfs:label "child astronomical body"@en . [20:34:49] ?p wikibase:directClaim ?pred . [20:34:50] } [20:35:23] SELECT ?x WHERE {?x rdfs:alias "child astronomical body"@en} seems simpler [20:35:32] so what you get is wd:398 - which is the property itself, and wdt:P398 - which is what you use to find claims with this [20:35:53] *alias -> label [20:36:23] RailRoadRunner: that finds you property. but property and predicate is not the same thing - property has a number of predicates, depending on the context (see https://www.mediawiki.org/wiki/Wikibase/Indexing/RDF_Dump_Format#Predicates [20:36:28] So the only thing that actually has a label "planet" is an object, not a property. [20:36:48] RailRoadRunner: right [20:36:59] "planet" is not a property, it's a thing [20:37:14] 10Wikidata, 3Roadmap, 7notice: Enable arbitrary access on Wikisource and itwiki projects - https://phabricator.wikimedia.org/T98756#1321952 (10Ricordisamoa) @greg: the new title clashes with {T98303} [20:37:17] SMalyshev: Well, ok, how about "hasPlanet"? [20:37:21] you can't reall say "Sun planet Pluto" [20:37:49] SMalyshev: Can I wildcard search for properties matching "*planet*"? [20:38:14] SELECT ?x WHERE {?x rdfs:label "*planet*"@en} [does not work] [20:38:14] RailRoadRunner: yes, but not with wdqs. It's the fulltext search thing. [20:38:51] SMalyshev: so no equivalent to RLIKE in SPARQL? Or at least not in that implementation? [20:38:55] so you can use existing wikidata search API... but you'll get a lot of stuff with planet in its name [20:39:12] RailRoadRunner: blazegraph supports fulltext search, but we didn't implement it [20:39:30] I'm not sure if we will, since we already have elastic so that would be duplicating it [20:39:39] (and elastic is much better at it) [20:39:39] OK, wildcard search is weaker than fulltext search, but also not available? [20:39:54] What is "elastic" in this context? [20:40:03] RailRoadRunner: ElasticSearch [20:40:20] the same thing you get when you type stuff in search box on wikidata [20:40:29] (and API behind it of course) [20:40:43] SMalyshev: Oh, but again, not a SPARQL query? In other words, you have to go outside SPARQL to find P and Q ids for wikidata? [20:41:10] RailRoadRunner: if you odn't know the exact names, yes, you have to, at least for now [20:42:00] we are thinking of making GUI to make it easier, but for now to use SPARQL engine you need to know either exact name or even better, Q/P-id [20:45:03] Fair enough [20:45:58] OK, I'll now try to find the diameters of the Sun's child astronomical bodies... [20:46:18] we're just starting with this thing :) a lot to do yet [20:48:06] diameters sounds like something we don't have yet, since they need units [20:48:39] nikki: Yes, I just noticed our planet doesn't have a diameter. [20:49:45] Wait, so wikidata has no information where the value would have units? [20:50:37] RailRoadRunner: mostly, yes. only some things (like population, etc.) are, but most quantities now aren't really there yet [20:50:52] this is supposed to change pretty soon according to wikidata team plans [20:51:22] SMalyshev: Well, population and eccentricity are unitless, but ... what sites do have triples where there is unit-full data? [20:52:04] I thought wikidata.org was the be-all and end-all of all semantic data in the Q2? [20:53:23] RailRoadRunner: units will be available next month or at least within this year [20:53:46] we know that they are really powerful and people are waiting for them basically since the beginning of Wikidata [20:54:27] benestar: OK, but I guess I'm confused. I thought wikidata had existed for years? Are there other collections of semantic data on the web? You can't be telling me that there is no semantic triple "Earth,circumference,40000km" (or the equivalent) anywhere on the web? [20:55:07] RailRoadRunner: there is dbpedia which exists longer than wikidata and tries to extract information out of Wikipedia's infoboxes [20:55:22] i guess freebase also had such information [20:55:37] benestar: Ah, OK. Is wikidata.org's ultimate goal to contain all the triples that dbpedia and freebase have? [20:56:31] And how does schema.org fit in to all this, if at all? [20:56:46] RailRoadRunner: not all triples [20:57:04] RailRoadRunner: schema.org defines semantics - i.e. collection of classes/predicates that people agree on using with certain sense [20:57:09] some information in freebase might not be imported to wikidata due to license or notability reasons [20:57:24] i.e. schema:lastModified means time of the last modification of something [20:57:32] benestar: OK, but most of it, yes? [20:57:38] yes [20:57:50] but afaik it doesn't have any data, just description of what certain URIs are supposed to mean [20:58:02] benestar: It seems like having no data with units is a major hole, so I'm having trouble accepting it. [20:58:04] so different data stores could use same URIs with same semantics [20:58:35] SMalyshev: When you say classes/predicates you mean items/properties? Or is there more to it? [21:00:28] unrelated question: does anyone happen to know how people do the line saying that the previous unsigned comment was added by [username here]? is it a template or something? [21:00:33] My world view is that 1) all knowledge can be modeled as semantic triples, and 2) much of it already has been. How true is this? [21:00:34] RailRoadRunner: well, so RDF is very basic - it's basically collection of triples Subject-Predicate-Object. But then we can assign special meaning to some URIs - i.e. rdfs:label means label, or rdfs:type means subject represents certain kind of object. Then we can define classes - like something is a wikidata entry, or something is a description of a painting, etc. [21:01:08] SMalyshev: Yes, and even rdfs:child or rdfs:parent to create a hierarchy. [21:01:13] RailRoadRunner: 1) yes, probably 2) some of it is, but it's right now very distributed. [21:01:52] And rdfs:implements or something too. [21:02:12] 1) correction - modelling some knowledge using triples can be very hard, but most factual knowledge is reasonably easy [21:02:31] SMalyshev: The semantic web has been around for ages, hasn't it? How can this not already be done? [21:02:46] RailRoadRunner: well, there's OWL where there's a very rich semantics of it but it's another big can of worms... [21:03:05] SMalyshev: OK, but people have used OWL to create semantic triples, yes? [21:03:16] RailRoadRunner: it's hard :) turns out humans know a lot of things and this knowledge is very disjoint [21:03:46] Let me ask it this way: is there any non-trivial site that has the triple "Earth, radius,4000 miles" (or whatever)? [21:03:47] SMalyshev is the update service for wdqs still down? [21:04:02] or get recent changes added into blazegraph again? [21:04:13] RailRoadRunner: OWL is rather for making sense of semantic triples. I.e. how we create links between entities, how we assign meanings, how we define constraints (so that we don't say "circumference of Earth is equal to duck" [21:04:24] oh, I found it, template:unsigned... I was searching for "previous", not "preceding"... [21:04:41] SMalyshev: Oh, but doesn't OWL also have classes and properties itself? [21:04:43] benestar: it's up though now is about 6-8 hrs behind... slowly catching up [21:05:01] SMalyshev: when will you give it its own server/hardware? [21:05:05] RailRoadRunner: yes, it has classes and it allows you to define your own classes [21:05:46] SMalyshev: Are you telling me the number of "famous" SPARQL endpoints on the web is fairly small? And that none of them will tell me the diameter of Earth? [21:05:58] RailRoadRunner: you have to distinguish between systems which define a grammar or some generic properties and those containing actual data [21:06:03] benestar: not a question for me :) If I get a real metal server, I can put it up there, but so far I don't have one [21:06:43] RailRoadRunner: probably one of them has, but I don't know which one out of the top of my head. Maybe freebase or dbpedia does [21:06:57] benestar: OK, I'm trying to get a handle on how much semantic data actually exists, and how I can query most of it? [21:07:10] pretty sure if you come back in half a year, wikidata one would too but right now due to units thing not yet [21:07:20] RailRoadRunner: dbpedia has a sparql endpoint which however is a bit broken afaik [21:07:31] freebase has its own query syntax and is going to be shut down anyways [21:07:37] SMalyshev: I feel like I've traveled backwards through a time warp. [21:07:58] wikidata will have quantities with units within half a year and will be the most reliable and useful database then hopefully :) [21:08:18] benestar: When did wikidata start? [21:09:03] benestar: actually about 3.5 hrs already... so maybe it''s catch up soon, it's on 2015-05-29T17:53:45Z now so not so far away [21:09:13] RailRoadRunner: the website was launched 30 October 2012 [21:09:17] is there anything non-developers can do to help quantities with units be done sooner? [21:09:25] the development startet about half a year before [21:09:44] the problem is even xlarge is kind of wimpy for the load as it seems [21:09:48] the basic idea exists since 2004 I guess [21:09:49] benestar: OK, I guess I'm just frustrated that what I consider to be basic information doesn't exist in semantic form somewhere obvious. [21:09:58] RailRoadRunner: :( [21:10:05] stay tuned, it will become available soon [21:10:10] nikki: not sure... [21:10:16] RailRoadRunner: believe me you're not the only one. that's why we're working on it :) [21:10:21] benestar: Semantic triples have been around for ages, though. [21:10:59] RailRoadRunner: that does not mean that open and free information is stored in this format ;) [21:11:27] "show me people who have co-starred with someone who has appeared in a movie whose title contains a word that is the name of a planet orbiting the sun and whose diameter is greater than 10000km" [21:11:42] (er, the planet's diameter, not the person's) [21:11:58] You're telling me there is no good way to get that information right now? [21:12:58] RailRoadRunner: well, not an easy way. You'll have to combine the databases (or have access to IBM Watson ;) [21:13:25] SMalyshev: but you can't even tell me which databases to combine? It's not like there are 5 total databases that will give me this info? [21:14:07] RailRoadRunner: sorry, out of the top of my head I can't... If I spent a day or two researching it I probably could but not on the spot [21:14:43] SMalyshev: I was speaking more metaphorically. I meant: do there exist a small list of databases that I can cross-query to answer my question? The answer appears to be 'no' though :( [21:14:50] * nikki thinks that if there were something that could really do all this stuff already, there'd be no need for wikidata [21:15:19] When I first saw semantic triples, I realized this was the future. I just can't believe it hasn't been done already. [21:15:26] RailRoadRunner: I don't think there is right now. We're moving there but no, not yet [21:16:43] Is wikidata in "competition" with schema.orG? [21:17:30] RailRoadRunner: nope [21:17:38] they are working together at some point actually [21:17:44] RailRoadRunner: no, schema.org defines various meanings to URIs basically. We use some of them in our data model, as well as other ones - from w3c.org and others [21:19:26] so schema.org does not have any data - it just establishes how to represent certain kinds of data so that different databases could use the same language [21:19:34] So schema.org will ultimately be a collection of classes (like Person) and properties (like birthDate), but the properties will be a subset of wikidata's properties, and the classes will be a small subset of wikidata's objects, namely those objects that are classes (or a subset thereof)? [21:20:48] Oh, I get it! Currently, wikidata lets you assign ANY propetty to ANY object. schema.org places restrictions on this. [21:21:19] RailRoadRunner: right now we don't use schema:birthDate directly, but if you look at https://www.wikidata.org/wiki/Property:P569 you see it under "equivalent property". [21:21:23] RailRoadRunner: see http://wdqs-beta.wmflabs.org/#PREFIX%20wd%3A%20%3Chttp%3A%2F%2Fwww.wikidata.org%2Fentity%2F%3E%0APREFIX%20wdt%3A%20%3Chttp%3A%2F%2Fwww.wikidata.org%2Fprop%2Fdirect%2F%3E%0APREFIX%20wikibase%3A%20%3Chttp%3A%2F%2Fwikiba.se%2Fontology%23%3E%0APREFIX%20p%3A%20%3Chttp%3A%2F%2Fwww.wikidata.org%2Fprop%2F%3E%0APREFIX%20v%3A%20%3Chttp%3A%2F%2Fwww.wikidata.org%2Fprop%2Fstatement%2F%3E%0APREFIX%20q%3A%20%3Chttp%3A%2F%2Fwww.wikidat [21:21:24] a.org%2Fprop%2Fqualifier%2F%3E%0APREFIX%20rdfs%3A%20%3Chttp%3A%2F%2Fwww.w3.org%2F2000%2F01%2Frdf-schema%23%3E%0A%0ASELECT%20%3Fx%20%3Furl%20WHERE%20{%0A%20%20%3Fx%20wdt%3AP1628%20%3Furl%20FILTER%20REGEX%28%20str%28%20%3Furl%20%29%2C%20%22schema.org%22%20%29%0A%20%20} [21:21:43] http://tinyurl.com/qad3vsf [21:21:56] this will give you all properties which have an equivalent on schema.org on wikidata [21:21:58] RailRoadRunner: schema.org just establishes the common vocabulary. [21:22:25] benestar: btw, wd:P50 is broken :) [21:22:43] SMalyshev: I fixed it on wikidata thus asking how big the lag is ;) [21:22:55] it uses https instead of http [21:22:58] benestar: ah, awesome :) [21:23:33] it'll catch up soon. if people didn't insist on editing it all the time it'd go faster of course ;) [21:24:12] benestar: Show those objects/properties whose "equivalent property" value matches schema.org? [21:24:30] yes, properties which have an equivalent on schema.org [21:25:30] benestar: right now it's 3 hrs behind. I expect by the end of the day it'll catch up unless there's a ton of edit or we hit some bug :) [21:25:31] OK, I'm now going to try to find the moons of the planets of the Sun. [21:25:41] that should be doable ) [21:25:51] SMalyshev: \o/ [21:27:49] SELECT ?z2 WHERE {SELECT ?z WHERE {wd:Q525 wdt:P398 ?z} wdt:P398 ?z} [21:27:53] That does not work [21:28:28] Even when I change the last ?z to ?z2 [21:29:34] SELECT ?z2 WHERE { [21:29:34] wd:Q525 wdt:P398 ?z. [21:29:34] ?z wdt:P398 ?z2 [21:29:34] } [21:29:42] RailRoadRunner: just connect both triples with a dot [21:30:00] or even easier - SELECT ?z2 WHERE { [21:30:00] wd:Q525 wdt:P398/wdt:P398 ?z2 [21:30:00] } [21:30:17] sparql allows yo to chain stuff [21:30:22] * SMalyshev out for lunch [21:30:28] SMalyshev: so "wdt:P398/wdt:P398 means P398 of P398 of?) [21:30:49] RailRoadRunner: yes, it's traversing P398 link twice [21:31:23] And the other version does a subselect of sorts. [21:34:00] OK, now what if I want the labels for the objects I'm getting? [21:34:15] SELECT ?z.label ?z2.label WHERE {wd:Q525 wdt:P398 ?z. ?z wdt:P398 ?z2} [21:34:15] <- does not work [21:35:33] Or do I need to look for the label property of my results? [21:38:43] What is the P value for the name or label property? It doesn't appear to show up when I search for something on wikidata. [21:39:11] it's not a property [21:39:24] nikki: OK, how do I get it then? [21:40:02] in sparql? no idea, I haven't used that yet [21:40:34] RailRoadRunner: if you wan't labels, you need to use label property. like this: [21:40:37] SELECT ?z ?z2 WHERE {wd:Q525 wdt:P398 ?z. ?z wdt:P398 ?z2} gives me two columns of object ids which is great, but I'd like to get a little bit more about these. [21:41:24] It seems to me like the metric base units would make an ideal "default" unit, but that's just me. [21:41:36] http://tinyurl.com/p2fnmof [21:42:20] SMalyshev: I could've done that as a subquery too, right? rdfs:label is just a regular property? [21:42:43] RailRoadRunner: also note if you query potentially has a lot of results makes sense to use limits, like this: http://tinyurl.com/nm75zul [21:43:09] SMalyshev: I just learned that :) Froze my browser for a second [21:43:15] RailRoadRunner: well, subquery is usually worse, I'm not sure query optimizer will like it much [21:43:46]