[01:57:25] Are there any AutoWikiBot-like tools that let you make a lot of repetitive Wikidata edits quickly? [02:00:46] harej: there are [02:00:48] !! [02:01:23] harej: I personally prefer autolist [02:02:37] harej: https://www.wikidata.org/wiki/Wikidata:Tools/External_tools [02:36:43] So, I see there is "cause of" and "has cause" [02:37:22] if Q1 is the cause of Q2, do I reciprocate and label Q2 as having the cause Q1? [14:25:14] Lydia_WMDE: by the way, I've now added a section to the bot requests page for wikispecies sitelinks, so if anyone else asks about it, you can tell them to look there :) [14:25:26] nikki: cool :) [14:25:32] thx! [14:28:21] hehe, no problem. I like seeing things getting linked [14:48:43] aude: if you have any specific questions youd like answered from the analytics side that I might be able to slip into the presentation let me know! [14:56:44] addshore: please add a kitten :) [14:57:01] otherwise, don't have specific questions [15:01:26] okay [15:09:22] Is there a reason, the warning "Don't try to merge items until phab:T115892 has been solved. All merges fail, it only removes items from the base item, but doesn't add anything to the target item." appears for me on the item Q7251 but not on Q7259? [15:10:30] oh gah [15:10:33] caching? [15:11:02] * aude sees no such message [15:11:07] Might be, but than Q7252 gives me "Due to issues, the merge feature is temporarily disabled until further notice. Please see phab:T115892 for more information. Moving sitelinks isn't an alternative, just remember the two items and merge them after the issues have been resolved. " [15:11:15] Which is just a different message [15:11:21] are you logged in? [15:11:25] No [15:11:29] oh [15:11:49] i still don't see the messages logged out [15:11:53] but suspect caching :( [15:12:36] It is [15:12:37] frimelle: when i look at Q7251 in firefox, i see it :( [15:12:53] Yes, in firefox. After purge it's gone. [15:13:28] Cache. :( [15:13:30] anons can't even merge [15:13:36] afaik [15:14:00] They can, I think. [15:14:06] they could use the api though, right? [15:14:11] Everyone: https://www.wikidata.org/wiki/Special:ListGroupRights [15:14:34] But aude, yes seems to be the cache. Cool, thanks :) [15:16:45] sjoerddebruin: i see [15:49:23] Does anyone have an overview what kind of games are made so far using Wikidata beside the ones by Magnus? [15:51:47] Not only games, but here is a overview of tools. http://tools.wmflabs.org/hay/directory/ [15:52:18] there are not much beside magnus ones [15:59:32] Okay, cool, thanks a lot! [16:11:10] addshore: asked for clarification on your mm-api ticket :) [16:13:34] JohnFLewis: replied [16:14:21] addshore: hm, okay :) I'll reply now along with what Krenair's brought up [16:32:44] addshore: fyi I've subscribed to the read-only task so I'll keep an eye on it and so [16:33:09] should be fairly easy to do, if noone jumps on it ill take a look in a week or so [16:56:13] hoo: will you be around for a bit? (and not busy?) [16:56:36] * aude totally forgot that we are deploying to wikipedias and can probably run home + eat in time to be back [16:57:52] aude: I'm busy, but can react in case of an emergency [16:58:02] but should listen to other stewards, if possible [16:58:55] ok [16:59:14] everything probably will be ok and i'll probably/hope back in time [16:59:28] :) [16:59:31] * aude no longer lives 10 min from the office :( [16:59:41] How long do you need to get home? [17:02:47] aude: hope its not too long! [17:12:47] hey. i'm feeling slightly silly here. i'm tryin to play with our sparql endpoint, but my sparql fook is somewhat weak... [17:13:01] i'm trying to list all cities in saxony: http://pastebin.com/Uw1mP40h [17:13:30] but this keeps timing out. did I get the qury wrong, or are we just very very slow right now? [17:15:00] Ough... that should really work [17:16:16] originally, i wanted the ten larges cities. but when that times out, i went for a simpler query. didn't help, though [17:19:00] DanielK_WMDE: It works if you only go down up to one located in terittory level [17:19:19] ?districts wdt:P131 wd:Q1202 . ?city wdt:P131 ?districts . [17:19:22] Like that [17:19:42] hoo: when i tried that, i got one result (Dreseden). Which is clearly wrong. [17:20:00] I get 28 [17:20:05] now i'm getting more too... [17:20:21] oh, that was with the population thing included. maybe i missed something there [17:20:59] hoo: err. i get 9 if i just remove the "*" from my original query. whut? [17:21:20] yeah... so people sometime put in the state [17:21:23] oh, right, your query does two steps. [17:21:27] and sometimes apparently the Landkreis [17:21:32] but then you miss the major cities, which are not located in districts [17:21:38] which would mean you miss the largest cities... [17:22:07] :P [17:22:16] this is a very basic use case :/ [17:22:53] DanielK_WMDE: * queries are not very fast, unfortunately... [17:23:25] SMalyshev: i can understand that when the pathes get long, or the set gets very large [17:23:31] neither should be the case here... [17:23:58] maybe there's a loop or something in the data? [17:25:39] quite possible :/ [17:26:34] How many of them should there be? [17:26:40] SMalyshev: uh, blazegraph doesn't check against circles when evaluating pathes?... [17:26:51] thaqt would be pretty horr [17:26:57] ...horrible [17:27:22] DanielK_WMDE: I have no idea, didn't check that. Maybe it has limit but timeout kicks in sooner [17:27:38] this one: http://tinyurl.com/pmu2aa6 - produces 35 cities [17:27:42] this is a bit embarressing... i was trying to show off to my daughter, who needed to name the four largest cities in saxony for her homework >_< [17:27:46] is it too little or about ok? [17:27:55] 35 sounds about right [17:28:15] SMalyshev: https://de.wikipedia.org/wiki/Kategorie:Gemeinde_in_Sachsen probably these [17:28:22] but data might be incomplete on WD [17:28:27] so the 35 I got might match [17:28:39] 35 is what I get from autolist [17:29:05] hoo: many of those arent cities [17:29:12] SMalyshev: if there is really no check for loops, we need to add it, either in blazegtaph or via pre-processing somehow. [17:29:16] they are "municipality" or "town" or something [17:29:35] DanielK_WMDE: yes, probably. Mind filing a task in phabricator? :) [17:29:47] sure... [17:30:12] it looks like some loop or so because if I unroll it manually it works just fine and I don't believe there's hierarchy deeper than 3-4 levels [17:30:26] SMalyshev: btw, the * in the instanceof part of the query didn't seem to have a big impact on performance [17:30:31] that seems to work just fine [17:30:36] so either there are tons of cities on level 8 or it's looping [17:31:18] *sigh* [17:31:35] ok, i'll verify that my original query works with manually unwinding the path, and then file a ticket. [17:32:34] I wonder how to identify these loops (for content maintenance reasons) [17:32:59] would manually unroll the queries and then look for stuff that comes up on several levels? :/ [17:34:54] hoo: write a script that traverses transitive properties and repots loops. all you need to do is keep track of your current path. [17:35:37] i suppose if you try to parallelizse that, i comes down to pretty much what you suggested... [17:38:01] DanielK_WMDE: Yeah... but probably hard to do in pure SPARQL [17:38:28] I was thinking about doing it purely in Sparql, but that might not work within our limitations [17:38:33] or in the real world at all :P [17:39:31] hoo, SMalyshev: something is still not right: http://pastebin.com/8HApR3ZV is missing some obvious results (Leipzig, Chemnitz, ...). Maybe I need to add more levels? Or is there something wrong with the population filter? [17:40:43] DanielK_WMDE: Leipzig doesn't have a population in WD [17:41:16] the primary sources tool suggest a few values, but these are quite old [17:41:21] and entirely unsourced [17:45:13] hoo: yay :D [17:45:18] SMalyshev: https://phabricator.wikimedia.org/T116298 [17:50:14] Building a loopfinder sounds like a nice pet project [17:50:18] * hoo should do that at some point [17:50:35] but obviously it would be nice if someone were to beat me to it ;) [17:54:27] hoo: i'll file a ticket [17:54:51] hoo: also, it's apparently impossible to edit references at the meoment. the "save" button stays disabled. [17:54:59] is this a known problem? or is it just me? [17:56:14] looking [17:56:38] Looks good to me [17:57:42] https://www.wikidata.org/w/index.php?title=Q4115189&action=history [17:59:34] hoo: can you add a Title snak to the reference for the population number of https://www.wikidata.org/wiki/Q2079? [17:59:47] The title would be the one for this document: http://www.statistik.sachsen.de/download/010_GB-Bev/Bev_Z_Gemeinde_akt.pdf [18:00:12] hoo: if it works for your, please comment on https://phabricator.wikimedia.org/T116300 [18:01:14] So "Bevölkerung des Freistaates Sachsen jeweils am Mona [18:01:14] tsende ausgewählter Berichtsmonate nach Gemeinden"? [18:02:08] https://www.wikidata.org/w/index.php?title=Q2079&type=revision&diff=261069756&oldid=261067595 [18:02:11] DanielK_WMDE: ^ Like that? [18:03:02] hoo: yes... that's odd. it just worked for you? which browser? [18:03:09] DanielK_WMDE: Latest firefox [18:03:12] 41 [18:03:13] DanielK_WMDE: thanks [18:03:22] sorry, had to go to a meeting [18:03:45] hoo: FF41 here, too. wtf? [18:03:49] mh [18:03:52] any JS errors? [18:04:44] DanielK_WMDE: I've got Leipzig on your query [18:04:59] * aude back [18:06:13] hoo: didn't check the console tbh. my daughter is telling me she's tired and wants me to read her a bed time story :) will you comment on the task? i'll investigate later [18:06:42] Already did [18:06:51] SMalyshev: yea, i added a population statement now. it seems that our data has a lot of gaping holes... [18:06:56] hoo: thanks [18:07:14] DanielK_WMDE: I wouldn't be surprised, demographic/economic data are far from complete [18:07:29] yea, looks like it. so that's not a software problem [18:30:19] Chemnitz btw does not have a link to Saxony as it seems [18:30:31] not to the current Saxony at the least [18:35:15] ... :/ [18:35:58] https://www.wikidata.org/wiki/Q8172 and https://www.wikidata.org/wiki/Q1022 located in each other [18:36:35] and we actually have 515 such pairs [18:37:49] https://www.wikidata.org/wiki/Q18715197 is actually located within itself [18:37:52] SMalyshev: That's clearly wrong... fixed [18:37:54] Yay [18:37:57] ... :P [18:38:09] it's bigger on the inside [18:41:29] wow we have 109 entries which are located within themselves [18:42:07] Technically that's not wrong :P [18:42:09] e.g. in Saxony: https://www.wikidata.org/wiki/Q2773 [18:42:39] * aude found items with stuff like P31:museum and P31:art museum [18:42:46] Well, but I guess we want *proper* supersets there [18:42:51] The item is located on the territory of the following administrative unit (upper level administrative subdivision). [18:42:58] when art museum is a subclass of museum [18:43:01] the docs say "upper" so it's wrong [18:43:20] aude: having two paths is not a problem. Having cycle is [18:43:54] ah [18:43:58] * aude had some trouble with https://gist.github.com/filbertkm/c07e9cdfbf9daabde5a8 [18:44:08] for finding all museums in a place (including subclass) [18:44:27] I think we need to have property definition as antireflexive and have bots check it [19:24:44] SMalyshev: can you make a list? I could probably fix some of them [19:25:24] nikki: this one for starters: http://tinyurl.com/qdx844z [19:25:30] these are self-referencing [19:26:29] nikki: this is for two levels: http://tinyurl.com/pxvukbk [19:27:14] aude: wdt:P131* probably is to blame, due to abundance of loops in P131 [19:27:55] SMalyshev: could be but when i removed P31:museum from the items in question [19:27:58] then the query worked [19:28:15] otherwise got a concurrent... something exception [19:29:15] also removed (?Removed claim: located in the administrative territorial entity (P131): New York (Q1384)) (current) [19:29:58] aude: that query works for me right now as found in https://gist.github.com/filbertkm/c07e9cdfbf9daabde5a8 [19:30:11] Total results: 55, duration: 594 ms. [19:30:45] i fixed the items :) [19:30:55] ahh ok :) [19:31:13] P131 definitely needs some cleanup and probably a bot to watch over it [19:31:49] yeah [19:35:23] I guess P150 will have the same issue, since that's a very similar property in the other direction [19:35:28] * nikki is fixing some anyway :) [19:40:49] https://www.wikidata.org/wiki/Q542740 is weird... [19:41:24] it seems like a bunch of gemeinden were merged together and kept the name of one of them [19:58:47] Right. https://www.wikidata.org/w/index.php?title=Q15687022&action=history [21:07:33] Change on 12www.mediawiki.org a page Extension:Wikibase Client was modified, changed by 176.92.148.38 link https://www.mediawiki.org/w/index.php?diff=1921291 edit summary: [-14676] Replaced content with "Hatsune Miku[[:File:]]" [21:07:34] 10[1] 04https://www.wikidata.org/wiki/:File: [21:08:00] Change on 12www.mediawiki.org a page Extension:Wikibase Client was modified, changed by Sjoerddebruin link https://www.mediawiki.org/w/index.php?diff=1921292 edit summary: [+14676] Undo revision 1921291 by [[Special:Contributions/176.92.148.38|176.92.148.38]] ([[User talk:176.92.148.38|talk]]) [21:08:03] 10[2] 04https://www.wikidata.org/wiki/Special:Contributions/176.92.148.3813 => [21:08:06] 10[3] 04https://www.wikidata.org/wiki/User_talk:176.92.148.38 [22:23:01] question: I have chemical exposure items that correspond to a chemical. For example, for rhodium, there is the item "rhodium exposure". The exposure item will list things like symptoms of exposure, since it was decided this is superior to having a separate property for "symptom of exposure". In any case, to link these things together I initially considered [22:23:01] using the "cause of" property. Example would be rhodium --> cause of --> rhodium exposure. [22:23:25] But in retrospect I feel like this is an awkward construction; why would a chemical cause its own exposure? [22:23:31] Is there a better way to link the two items together? [22:23:46] mh [22:23:55] sounds like you want an "effects" property [22:24:08] or maybe even "exposure effects" one [22:24:13] probably the first, though [22:24:26] is that an existing property? [22:25:20] ...doesn't look like it [22:28:17] In the meantime, if the exposure item is understood to be a medical condition, would it be okay to say that [chemical] is the cause of [chemical exposure]? [22:28:36] * harej looks to see how they do this for biological causes of sickness [22:29:41] HIV (Q15787) is labeled as the cause of (P1542) acquired immunodeficiency syndrome (Q12199), so I think I will proceed with that [22:31:17] likewise with rhinovirus and common cold [22:45:21] harej: mh... I'd still prefer to have a separate property [22:45:52] having HIV cause a symptom is different from a chemical having effects on exposure