[08:23:25] Hi, what is the best practice for P131 (located in the administrative territorial entity) for building entries? [08:24:23] Is it good to save different entries for the admin levels (county, city, quarter) ? [08:24:47] Or just the lowest possible level? [08:27:15] In example Q6232341 there was alreadyan entry Massachusetts. I added Boston. [08:28:10] Should I remove now the Massachusetts entry, because Boston is located in there? [08:30:09] I'm preparing a bulk upload and there are hundreds of similar cases... [08:33:45] If noone moans here, I will just add an entry on city level without checking if there are already entries on lower or higher level! [08:54:52] Hi, Arch2all :) [08:55:38] In theory, the correct thing is to use the lower level [08:56:19] However, sometimes these levels aren't very well modeled or standardized [08:58:01] For the case of Q6232341, Boston is perfectly defined and the value "Massachusetts" (Q771) can be removed [08:58:09] I would say that in practice we normally only have a single level and prefer the lowest [08:58:46] but we have a *lot* of data, so it's not perfect [09:00:47] if something isn't standardised, it's probably not really an administrative division [09:02:05] Ok, is there already a bot around, which checks P131 for multiple entries (with hierarchical relationship) and remove the higher ones? [09:03:14] Asking this, because it's make bulk upload much more complicated, if I have to do this n my own [09:04:32] I don't think there's a bot; if that were the case, there wouldn't be so many redundancies right now :) [09:05:33] And that's a good idea, but we also have to consider that people add those redundant statements because of limits in the Query Service, the quality extensions, etc., which are unable to iterate through so many levels [09:05:50] That's what I mean with the "In theory" [09:08:22] @abian That's why I asked for "best practice" ;) [09:08:52] :D [09:09:11] * abian likes the idea of removing everything redundant when nobody's looking [09:10:57] so, would it be accepted (and appreciated) by the community, if I create a bot, that set the P131 entries to lowest level? [09:11:42] I would appreciate it, and also with subclass of (P279) and instance of (P31) [09:13:28] Having said that, the quality extensions have to be fixed or they'll mark violations where they aren't [09:15:59] Which wrong violations are marked for P131 by the quality extensions? [09:17:04] I'm not sure, I know the ones with P279 and P31 since the P279 hierarchies are really deep [09:19:15] Another problem is the fact the administrative hierarchies are not always strict trees. For example the counties, townships, towns in US are often overlapping :( [09:19:49] Well, but that's not a problem [09:19:59] I made this as an example for Spain... https://www.wikidata.org/wiki/Help:P131/Q29 [09:22:14] I still think, that this is a problem for queries. Example: [09:24:01] You can't ask when two entities overlap in the geography if they aren't parent-child, that's true [09:25:04] A building entry has P131 with town XY, that covers county a and county B. But actually the building lies in county B. If You do a SPARQL query for county A You will get a false positive. [09:26:14] Yeah, that's what I mean; when there are more than one, I would choose the higher level [09:28:08] The point is that maybe towns don't belong to counties and they're parallel organizations [09:29:19] In the same way that on https://www.wikidata.org/wiki/Help:P131/Q29 the entities with the XOR are parallel to Q162620 entities [09:30:48] Yep, the American administrative system is a horrible chaos seen from a database architect's perspective ;) [09:30:57] xD [09:55:08] can someone explain how to define properly the license? https://www.wikidata.org/wiki/Q23946080 [09:55:27] https://www.wikidata.org/w/index.php?title=Q23946080&diff=718669596&oldid=718409411 gives an error [10:28:35] yannf: I have no idea how to say that something is public domain, I've used Q19652 (public domain) as the value in the past, which also gives a constraint violation but there's not a lot I can do until someone can tell me the right way [10:29:12] nikki, yes, but with which property? [10:29:26] "domain public" is not a license [10:29:47] that's the whole point of https://www.wikidata.org/wiki/Wikidata:Property_proposal/copyright_status [10:30:01] with license, because nobody has told me what I can use instead [10:31:00] you shouldn't add instance of license to templates though, a wikimedia template isn't a license, we have properties like "template's main topic" to link templates to what they represent [10:31:50] nikki, do you have an example? [10:32:16] https://www.wikidata.org/wiki/Q5271108 [10:33:00] ok, thanks [10:33:10] I will use this as a model [11:24:29] Lucas_WMDE: Are checks that time out still being considered violations? [11:30:09] abian: what kinds of checks are you thinking about? [11:30:39] query timeouts are no longer considered violations since recently: https://phabricator.wikimedia.org/T199788 [11:30:47] I don’t think we have any other timeouts… [11:30:53] Yeah, that's what I wondered [11:31:37] Perfect then; Arch2all, you can run your bot without these problems :) [11:41:00] PROBLEM - Response time of WDQS eqiad on einsteinium is CRITICAL: CRITICAL: 10.00% of data above the critical threshold [300000.0] https://grafana.wikimedia.org/dashboard/db/wikidata-query-service?orgId=1&panelId=13&fullscreen [11:41:38] @abian Which problems do You mean? [11:42:07] Having to write redundant statements so that these violations don't appear [11:45:20] PROBLEM - WDQS HTTP Port on wdqs1003 is CRITICAL: HTTP CRITICAL: HTTP/1.1 502 Bad Gateway - 380 bytes in 0.000 second response time [11:46:30] RECOVERY - WDQS HTTP Port on wdqs1003 is OK: HTTP OK: HTTP/1.1 200 OK - 434 bytes in 0.039 second response time [11:58:35] Hello! Anyone can mark this page for translation please? https://www.wikidata.org/wiki/Module:Databox/doc [11:59:06] Or approve me as a translator admin? https://www.wikidata.org/wiki/Wikidata:Requests_for_permissions/Translation_administrator/Benoit_Rochon [11:59:09] Thank you ! [12:06:55] !admin [12:06:55] Please visit https://www.wikidata.org/wiki/WD:AN [12:07:03] RECOVERY - Response time of WDQS eqiad on einsteinium is OK: OK: Less than 5.00% above the threshold [120000.0] https://grafana.wikimedia.org/dashboard/db/wikidata-query-service?orgId=1&panelId=13&fullscreen [12:10:02] you would need a bureaucrat to be made a translation admin, admins can't do that [12:10:18] good point! [12:10:29] :) [12:11:58] But if an admin can mark that page for translation, that would give a hand for now. I'm going to a presentation for Databox and people in the room don't speak English! [12:12:07] https://www.wikidata.org/wiki/Module:Databox/doc [12:12:36] Merci shön ! [12:15:08] I don't know if there's anyone around in here at the moment who can do that :( [12:15:16] maybe try posting on https://www.wikidata.org/wiki/Wikidata:Translators%27_noticeboard? [12:15:38] Ash_Crow: is translation admin I think [12:16:02] and also quiddity [12:17:50] I think admins can grant themselves the flag [12:18:22] "Administrators do not need to undergo another discussion to become translation administrator; they can self-grant the rights to their account if necessary. " [12:21:33] I am no longer admin since 2013! [12:23:53] Yeah, I mean there are more admins who can help you although they're not translation admins explicitly [12:27:39] I won't provide any example :) [12:45:57] hallo [12:48:13] hi windows [12:48:22] hallo andre [13:27:30] any wikidata admins and os'ers around [13:41:34] sjoerddebruin: ^ [13:41:42] Chrissymad: hi [13:42:08] sjoerddebruin: i'm not sure if what i just nominated actually needs os because it appears to potentially be a checking account number or gibberish but i don't know which lol [13:42:26] https://www.wikidata.org/wiki/Wikidata:Requests_for_deletions#Bulk_deletion_request [13:44:33] Chrissymad: done, I wonder where you would expect to find information about our oversighters? :) [13:45:02] sjoerddebruin: lol i was in a rush and didn't think to look at userrights [13:45:04] :P [13:47:50] Why are june and july missing on https://www.wikidata.org/wiki/Wikidata:Requests_for_deletions/Archive/2018... [13:48:23] * Chrissymad thinks jimbo broke it [13:50:18] because https://www.wikidata.org/w/index.php?title=Wikidata:Requests_for_deletions/Archive/2018/06&redirect=no is missing [13:50:41] now done [13:50:55] :D [14:23:59] Lucas_WMDE: I don't suppose you can help me bring https://phabricator.wikimedia.org/T173026 to the attention of someone who can fix it? it seems like it should be straightforward to fix if I found the actual problem (not that I can test it), so it'd be a shame if it goes unnoticed for another year :P [14:24:54] * nikki never knows who to ask, so I'm asking you [14:28:56] nikki: sounds like you hit the nail on the head with that comment [14:29:04] do you want to try submitting a patch and I’ll review it? :) [14:29:18] oh, wait, this is in DataValues, I don’t think I even have merge rights there [14:30:51] bleh, and the latest version uses scalar type hints, I’m not sure if we’re allowed to use that in production :( [14:35:17] if by submitting a patch you mean sending a pull request on github, I can probably do that [14:35:25] yeah, that would be it [14:35:36] and then I’ll find out who I need to bug to get it merged ^^ [14:36:24] or I could create the pull request as well if you prefer [14:36:35] that would be easier :) [14:36:39] ok :) [14:36:42] if you can test it, that would be even better [14:53:46] nikki: https://github.com/DataValues/Geo/pull/137 [14:54:05] jakob_WMDE: can you take a look at that pull request ^ ? you’re one of the five people in the GitHub org, I have no idea who else is allowed to merge it ^^ [14:54:26] thanks :D [14:54:44] that is some weird org [15:28:18] Lucas_WMDE: I'll have a look in a bit! [15:29:09] thx [15:57:40] Lucas_WMDE: merged, but also unsure how to proceed with the < php7 backport [15:58:59] ok [15:59:05] I guess JeroenDeDauw would know? [16:49:35] PROBLEM - puppet last run on wdqs1009 is CRITICAL: CRITICAL: Catalog fetch fail. Either compilation failed or puppetmaster has issues [17:04:59] I really miss the gender game [17:05:46] The Wikidata gender game is no longer a thing? [17:06:07] It is not loading anymore for 2/3 years now :| [17:06:19] Now I have some 300± backlog again [17:07:12] Maybe we need some less specific game: give a sparql query and the easy buttons you want. :) [17:09:57] RECOVERY - puppet last run on wdqs1009 is OK: OK: Puppet is currently enabled, last run 1 minute ago with 0 failures [17:18:27] it seems to load for me [17:21:27] maybe it is just slow if you only want one language [17:21:40] it doesn't use sparql afaik [17:26:37] sparql would probably also be slow, given the number of items :/ [17:32:21] It's doable for specific wiki's. http://tinyurl.com/ycnh5rwv [18:38:36] I get an error everytime I try to subscribe to the mailing list here: https://lists.wikimedia.org/mailman/listinfo/wikidata-tech [18:39:42] and the error is.......................? [18:41:18] Unrelated: do you inexplicably see headings in Polish? https://www.wikidata.org/wiki/Property:P3499 [18:41:44] it's Forbidden [18:42:17] Forbidden [18:42:17] You don't have permission to access /mailman/subscribe/wikidata-tech on this server. [18:44:46] worked for me [18:45:00] davidwbarratt who just left: Might be https://gerrit.wikimedia.org/r/#/c/operations/puppet/+/432168/ [18:47:31] davidwbarratt: Might be https://gerrit.wikimedia.org/r/#/c/operations/puppet/+/432168/ [18:47:38] Try again later? [18:48:22] andre__ same error [18:49:18] davidwbarratt, ??? [18:49:31] "same error" referring to what? [18:50:11] andre__ an HTTP 403 error "Forbidden" [18:50:25] davidwbarratt: See what I wrote: "Try again later" [18:50:36] "Later" is not "right now" :P [18:50:36] oh you said "later" haha [18:50:40] I missed that. :) [18:50:42] heh [18:57:58] abian: yes, it's a bug that's already been fixed, but lots of pages are still cached [18:58:18] purging should fix it [18:58:37] Cool, nikki, thanks! [18:58:51] I was already filling a task... :) [19:02:38] https://www.google.com/search?q=site%3Awikidata.org+%22Typ+danych%22&oq=site%3Awikidata.org+%22Typ+danych%22 [19:02:45] Google doesn't find many [19:03:35] But neither P3499 :( [19:05:08] https://www.google.com/search?q=site%3Awikidata.org+%22Wikis%C5%82ownik%22+%22Deklaracje%22 [19:49:23] Good evening. One user ran an unapproved bot making a huge mess. Do we have tools these days to rollback all the edits? [19:49:30] Fun stuff like https://www.wikidata.org/w/index.php?title=Q908135&type=revision&diff=718943744&oldid=711998743 [19:49:51] Oh :S [19:50:15] Arch2all was here this morning [19:50:27] And we talked about this [19:50:34] Good evening, multichill :) [19:50:39] And you didn't point out the bot policy? [19:51:06] Sadly not, I wrongly thought they were aware of it [19:52:34] The user seems to not understand located in the administrative territorial entity (P131) so basically replaced it with something lower all over the place [19:52:47] https://www.wikidata.org/w/index.php?title=Q1156823&type=revision&diff=718963627&oldid=705751468 for US [19:53:08] Lot's of changes in Germany, but not sure what the lowest level is there. Municipality? [19:53:30] sjoerddebruin: ^ [19:54:06] You can check the log if you want to read the conversation... https://bit.ly/WikidataIRCLog [19:54:19] https://www.wikidata.org/w/index.php?title=Q1513292&type=revision&diff=718963883&oldid=705751698 <- is this valid for example? [19:54:48] He started talking at 08:23 (UTC) [19:54:50] ortsteil is administrative territorial entity according to it's wikidata item [19:55:01] Yeah, I read it already [19:55:05] * multichill blames nikki [19:55:21] er... what did I do? [19:55:33] I would say that in practice we normally only have a single level and prefer the lowest [19:55:58] Well, that is technically correct [19:56:00] And than you end up with junk like https://www.wikidata.org/w/index.php?title=Q17342544&curid=18939167&diff=718953175&oldid=705742909 :P [19:56:15] Yup, but nobody explained what the lowest level is..... [19:56:41] I never said "replace administrative divisions with non-administrative places" :P [19:57:00] that website has Kootwijk as a quarter... [19:57:20] should instead have respected constraints [19:57:38] https://www.wikidata.org/w/index.php?diff=718965707 [19:57:52] https://www.wikidata.org/w/index.php?title=Property%3AP131&type=revision&diff=699129481&oldid=698877378 that explains the lack of constraint violations [19:58:15] BTW, hey, sjoerddebruin, so long without reading you by here :) [19:58:24] I was in South Africa for two weeks ;) [19:58:36] multichill: ugh [19:58:44] Hahaha, I see :) [19:59:25] https://www.wikidata.org/wiki/Q2983893 seems like a mess [19:59:37] https://www.wikidata.org/wiki/Q2333944 and then https://www.wikidata.org/wiki/Q15830667 [20:06:10] for the places I'm familiar with, big cities are divided into stadtbezirke and then stadtteile, smaller places are divided into areas that don't seem to have a consistent name... I've seen them called stadtteile, ortsteile, ortschaften and probably other names [20:07:56] But are these real administrative divisions? Do they have their own goverment? [20:08:14] Can you vote on people? Do they raise taxes? [20:11:14] nikki: ? [20:11:57] I don't know [20:14:56] if I had to be able to answer those questions to use p131, I would have to stop using it [20:15:56] I mostly assume that if it exists officially with a known border, it's probably fine, and I don't seem to get constraint violations at least :P [20:16:59] +1 (and with a known name) [20:17:36] I was also assuming that, yes, otherwise I wouldn't know what to search for :) [20:19:24] Hehe, we get that kind of stuff in high school. ;-) [20:19:58] So this user seems to have introduced about a thousand constraint violations [20:24:22] it seems like osm also considers them administrative, with admin_level=9 for the stadtbezirke and admin_level=10 for the others [20:25:38] and learning that stuff in school wouldn't have helped me much since I switched countries :D [20:28:40] Look at what I've just found... https://bots.wmflabs.org/logs/%23wikidata/20161011.txt :D [20:31:28] I killed a first batch of errors for the Netherlands. I appreciate if someone else could follow up with this user how to clean this up. I don't expect to be online before the weekend [20:33:51] Maybe it's good that he understands the possible issues here and revert their own edits [20:34:30] Via undone, not revert, because he's not a rollbacker [20:34:34] https://www.wikidata.org/wiki/User_talk:Arch2all#Radio_Kootwijk_located_in_the_administrative_territorial_entity [20:34:38] *undo [20:34:53] Wouldn't be the first time I assigned someone rollback after a bot error [20:36:18] You're being very hard with him :( [20:36:35] He has good faith, I'm sure [20:37:06] I'll be the grumpy bad cop, you can be the good cop abian [20:40:27] Lol [20:59:51] multichill: did you see my answers to your question in my talkpage? ) [21:00:51] Oh. Just saw the not being online thing. Well, no hurry :D [21:01:46] Just don't be grumpy bad with me D [21:01:52] * :D [21:04:02] I notice that sometimes I seem to miss pings. I think I got them on mobile and later don't see it on laptop [21:04:55] Don't worry about the report, it will eventually be empty. About the dates: I usually get the start and end date from the source data and if it's both a year and the same year, I'll add it to Wikidata [21:05:21] So the 1880-1885 stuff gets skipped, the 1881-1881 does gets added [21:05:42] reosarevok: And now I am going offline ;-) [21:11:28] ... https://www.wikidata.org/w/index.php?title=Q2416723&action=history [21:32:57] Does wikidata have a non-free media policy like Wikipedia? (i.e. things that are "fair use" can be uploaded) or does all media have to come from commons [21:33:09] only commons [21:33:20] sjoerddebruin ah, ok, thanks [21:33:39] there are only a few wikipedia with fair use though [23:58:09] well, the fair use thing is mainly US-based and does not apply to most other counties :x [23:58:12] +r