[09:41:18] <jsahu>	 In mac how to get the suggestion
[09:43:13] <sjoerddebruin>	 Hm.
[11:11:33] <icinga-wm>	 PROBLEM - wikidata.org dispatch lag is higher than 300s on www.wikidata.org is CRITICAL: HTTP CRITICAL: HTTP/1.1 200 OK - pattern not found - 1953 bytes in 0.163 second response time
[11:16:33] <icinga-wm>	 RECOVERY - wikidata.org dispatch lag is higher than 300s on www.wikidata.org is OK: HTTP OK: HTTP/1.1 200 OK - 1933 bytes in 0.133 second response time
[11:27:25] <bharatahs>	 Hello,
[11:27:25] <bharatahs>	 I looking at deriving the count for all Wikidata entities with a coordinate location and to check the language coverage in top 10 spoken languages of the world. 
[11:28:09] <bharatahs>	 Should I be using query.wikidata.org for this or is there any alternate way to do this ?
[11:43:36] <icinga-wm>	 PROBLEM - wikidata.org dispatch lag is higher than 300s on www.wikidata.org is CRITICAL: HTTP CRITICAL: HTTP/1.1 200 OK - pattern not found - 1954 bytes in 0.114 second response time
[12:03:11] <Amir1>	 harej: please slow down the bot
[12:03:12] <Amir1>	 https://www.wikidata.org/w/index.php?title=Special:Contributions/Harej&offset=&limit=250&target=Harej
[12:03:21] <Amir1>	 250 in one minute is waaay too much
[12:03:37] <Amir1>	 60 per minute 
[12:07:05] <muelli>	 can I create a query that understands something like "ip address in ip network", e.g. "192.1.2.1 in 192.0.2.2/24?"
[12:07:05] <Amir1>	 Lydia_WMDE: https://www.wikidata.org/wiki/User_talk:Harej#Slow_down
[12:07:28] <harej>	 Amir1: i've stopped it for now; will throttle it to edit at a slower rate
[12:07:35] <harej>	 thank you for letting me know
[12:07:38] <Amir1>	 Thanks
[12:13:42] <icinga-wm>	 RECOVERY - wikidata.org dispatch lag is higher than 300s on www.wikidata.org is OK: HTTP OK: HTTP/1.1 200 OK - 1921 bytes in 0.112 second response time
[12:19:21] <harej>	 Amir1: question: at any point during my edits, was the lag above 5 seconds? Of course it's understandable if the rate of editing was still too fast but I'd be particularly concerned if my script wasn't even respecting maxlag even though it was supposed to
[12:23:54] <harej>	 if maxlag was otherwise being respected, do you think it would be helpful if I set it from its current five seconds (the WDI default) to something extra-strict like 0.5 seconds? My goal is to make the most intelligent use of spare capacity, rather than resorting to guesswork.
[12:39:46] <Incabell>	 hey everyone, quick question: once a language code is added and thus supported as a monolingual text value how can I get it recognized in Wikidata. My ticket has been resolved (https://phabricator.wikimedia.org/T167811) but still when trying to add -xpu- in the respective statement instead of -mis- like it is now, it can't find the language code in the drop down. is there anything else I should be doing to make this happen? Thanks for any 
[12:39:46] <Incabell>	 help!
[12:55:25] <sjoerddebruin>	 Hm, indeed.
[12:56:35] <sjoerddebruin>	 I think this is https://phabricator.wikimedia.org/T124758, Incabell 
[13:11:01] <sjoerddebruin>	 Lucas_WMDE: is autocompletion for properties on the query service "broken"? Most properties are missing.
[13:21:45] <Lucas_WMDE>	 sjoerddebruin: a patch to make it use elastic search was merged recently, but I don’t think that was deployed already
[13:22:23] <sjoerddebruin>	 Lucas_WMDE: P31 isn't even included...
[13:22:50] <Lucas_WMDE>	 weird, works for me
[13:22:57] <Lucas_WMDE>	 you’re searching for “instance of”?
[13:23:40] <sjoerddebruin>	 I'm in Dutch...
[13:24:23] <sjoerddebruin>	 But now it works, weirdly.
[13:24:30] <Lucas_WMDE>	 “is ee” found P31 (is een)
[13:24:33] <Lucas_WMDE>	 strange
[13:24:45] <sjoerddebruin>	 It doesn't work all the time...
[13:25:02] <Lucas_WMDE>	 is it possible that the search was overloaded because of the bot?
[13:25:07] <Lucas_WMDE>	 doesn’t feel very likely to me
[13:25:34] <sjoerddebruin>	 Oh wait, it's been a while for me. I was searching with a P in front.
[13:26:26] <sjoerddebruin>	 Wasting your time here.
[13:26:51] <sjoerddebruin>	 I need some caffeine. 
[13:27:48] <Lucas_WMDE>	 okay :D
[13:27:49] <Lucas_WMDE>	 phew
[13:38:47] <sjoerddebruin>	 Now let's work on these backlogs, pfff.
[13:48:09] <Incabell>	 sjoerddebruin, thanks!
[13:48:23] <sjoerddebruin>	 My pleassure. :)
[13:53:37] <MichaelSchoenitz>	 Technical Advice IRC meeting starting at 3 pm UTC/5 pm CEST in channel #wikimedia-tech, hosts: @addshore & @Tobi_WMDE_SW - all questions welcome, more infos: https://www.mediawiki.org/wiki/Technical_Advice_IRC_Meeting
[14:24:26] <nikki>	 Incabell: by the way, I noticed that the text you requested xpu for is actually english, which doesn't seem right
[14:24:57] <nikki>	 if the english is a translation of the original, there should probably be a "literal translation" qualifier with the english text
[14:25:49] <nikki>	 and if the original text can't be entered, the only thing I can think of is to use unknown value... not sure if you can select a language then though
[14:26:16] <Incabell>	 Hey nikki: That was not my edit. I was still waiting to be able to use the language code, but you're right. It should have a different qualifier.
[14:27:01] <Incabell>	 I think I can't select anything anymore once I've entered unknown value. And also that wouldn't be a true statement because it is known, it's just that I can't enter the language.
[14:27:38] <nikki>	 I'm talking about https://www.wikidata.org/w/index.php?title=Q1146714&diff=500152501&oldid=493623495 which is clearly english
[14:28:44] <Incabell>	 I know. It's a translation of the original text. Preferably it should have the text in its original language as a value
[14:29:24] <Incabell>	 I don't think I properly get your point :/ sorry
[14:29:57] <nikki>	 what I'm saying is that that text is in english, so it's wrong to say it's in another language, it should be marked as english
[14:30:09] <Incabell>	 yes. I agree.
[14:30:20] <nikki>	 and since that doesn't appear to be the actual inscription, only a translation, it shouldn't be the value of the inscription property, it should only be a qualifier
[14:30:55] <Incabell>	 what I was asking for in the ticket is what I want to use ti enter the text in its original language
[14:32:23] <nikki>	 oh, you have the original text too?
[14:32:45] <nikki>	 I wasn't able to find it on any of the wikipedia pages I looked at
[14:33:45] <Incabell>	 the problem is that it's written in the phonecian alphabet. I don't know how this is usually handled on wikidata. I was thinking maybe it's possible to use an image as a value
[14:34:21] <nikki>	 nope :( properties have a particular data type
[14:34:44] <nikki>	 although you could request a new one if you wanted
[14:35:24] <Incabell>	 maybe that wouldn't be too bad for inscriptions although it wouldn't really help with the machine readability :/
[14:35:52] <Incabell>	 I'll check later how other items handle it when the text is written in a non-modern alphabet
[14:36:28] <nikki>	 if the script is in unicode, it should be possible to enter it, if you can find a unicode version
[14:38:10] <Incabell>	 looks like it is! https://en.wikipedia.org/wiki/Phoenician_alphabet#Unicode \o/
[14:38:18] <Incabell>	 cool
[14:39:55] <nikki>	 that's good :D
[14:43:01] <Incabell>	 although I can't seem to find anything written in the phoenician alphabet online:(
[14:43:30] <Incabell>	 Thanks for your help though! I will have to look into this in more detail later.
[14:44:56] <nikki>	 let me know if you managed to enter it, would be cool to see :)
[14:50:54] <MichaelSchoenitz>	 Technical Advice IRC meeting starting in 10 minutes in channel #wikimedia-tech, hosts: @addshore & @Tobi_WMDE_SW - all questions welcome, more infos: https://www.mediawiki.org/wiki/Technical_Advice_IRC_Meeting
[14:59:24] <DanielK_WMDE>	 Incabell: beware that (some version of) mysql's utf8 support fails for anything that's not in "Basic Multilingual Plane". I encountered this problem with cuniform and gothic. I'd imagine phoenician would have the same issue.
[14:59:53] <DanielK_WMDE>	 this isn't a problem on the wmf cluster, but may be for other wikis - particularly for development and testing environments
[15:00:30] <DanielK_WMDE>	 (some parts of) traditional chines also has this problem, btw.
[15:00:54] <DanielK_WMDE>	 https://www.sttmedia.de/unicode-basiclingualplane
[15:02:16] <Incabell>	 DanielK_WMDE, :( but thanks for the info.  
[15:02:22] <Incabell>	 nikki, will do :)
[16:06:52] <pintoch>	 does anybody use Wikidata-Toolkit to edit Wikidata?
[16:08:35] <pintoch>	 I'm running into this issue: https://github.com/Wikidata/Wikidata-Toolkit/issues/271
[16:08:40] <pintoch>	 Tpt[m]: any thoughts?
[16:09:11] <SMalyshev>	 DanielK_WMDE: any objections/comments for https://gerrit.wikimedia.org/r/#/c/368120/?
[16:13:30] <Tpt[m]>	 pintoch: I have started to investigate on this error. It is because the API returns the error "you should POST your request" even if it seems that Wikidata-Toolkit actually use a POST request
[16:13:41] <Tpt[m]>	 I should spend more time to figure out what it's not working
[16:14:26] <DanielK_WMDE>	 SMalyshev: why does the OWL file contain the path of the OWL file, but no OWL?
[16:14:54] <SMalyshev>	 DanielK_WMDE: actual file is in the docs folder, it's only a link to put it into public space
[16:15:08] <DanielK_WMDE>	 oh, so this is a symlink?
[16:15:14] <SMalyshev>	 yes
[16:15:24] <DanielK_WMDE>	 ah, gerrit doesn't make that obvious
[16:15:27] <SMalyshev>	 not that gerrit would tell you that :)
[16:15:32] <DanielK_WMDE>	 right
[16:15:41] <DanielK_WMDE>	 i cannot verify that that path is correct
[16:15:44] <DanielK_WMDE>	 otherwise, lgtm
[16:16:03] <SMalyshev>	 ok, cool, I'll then submit it for swat for deploy
[16:16:16] <DanielK_WMDE>	 make sure we actually follow symlinks ther e:)
[16:16:37] <SMalyshev>	 feel free to comment/+1 :)
[16:16:38] <SMalyshev>	 yeah sure
[16:17:13] <SMalyshev>	 DanielK_WMDE: that place already has tons of symlinks
[16:21:24] <SMalyshev>	 DanielK_WMDE: would also appreciate another look on https://gerrit.wikimedia.org/r/#/c/339575/ - I think the OrderedStreamingForkController issues are cleaned up and everything should work now
[16:25:14] <DanielK_WMDE>	 SMalyshev: i'll try to have another look, but i'm a bit swamped atm
[16:26:36] <SMalyshev>	 DanielK_WMDE: not super-urgent, but would like to get that in sometime soon, since it'd allow to run relforge tests easier. So when you have time :)
[16:30:31] <pintoch>	 Tpt[m]: I am also investigating but my Java is a bit rusty ^^
[16:31:32] <pintoch>	 I'm trying to dump the requests made by the toolkit…
[16:32:19] <pintoch>	 I have to admit I'm a bit surprised how low-level the underlying network API is
[16:33:41] <SoniWP_>	 @abartov I noticed a weird issue.
[16:34:34] <SoniWP_>	 query.wikidata.org generates shortURLs from tinyurl.com But that happens to be a blacklisted website for Wikimedia filters
[16:36:21] <SoniWP_>	 I think that 1) tinyurl must be unblocked from Wikidata projects
[16:36:22] <SoniWP_>	 and/or 2) Someone can host a URL shortener from a WMF related/affliated project that will not be blocked. Ideally a website that only creates shortURLs from some specific whitelisted domains (so cannot be used for spam generation)
[16:37:14] <harej>	 we have w.wiki but i'm not sure when that's going to be public
[16:38:48] <sjoerddebruin>	 We are aware of that SoniWP_ 
[16:38:54] <sjoerddebruin>	 The first thing will never happen.
[16:39:10] <sjoerddebruin>	 The second is something they are working on
[16:40:15] <SoniWP_>	 I see. So I'm going to assume Wikidata query will be one of the first places to start using the shortener?
[16:40:33] <sjoerddebruin>	 Yes, indeed.
[16:41:06] <SoniWP_>	 Got it. That's good to hear. (Hopefully it'll be soon though :) )
[16:41:56] <SMalyshev>	 hopefully yes
[16:48:49] <DanielK_WMDE>	 SMalyshev: gah, i missed the meeting! sorry! should i still join?
[16:49:11] <SMalyshev>	 DanielK_WMDE: sure, we've about the middle now...
[17:02:22] <Tpt[m]>	 pintoch:
[17:02:23] <Tpt[m]>	 > I have to admit I'm a bit surprised how low-level the underlying network API is
[17:02:24] <Tpt[m]>	 Yes, it is using the jdk API that is quite low level and painful
[17:02:41] <Tpt[m]>	 We should maybe consider to migrate to something better
[17:03:07] <pintoch>	 I'm dumping the requests with a proxy and they appear to be correct POST requests
[17:03:22] <Tpt[m]>	 yes, it looks stange
[17:03:30] <Tpt[m]>	 The jdk is going to have a new API but it will still be experimental in java 9
[17:04:10] <Tpt[m]>	 There is maybe something else wrong with the request and the MW API is returning a wrong error message?
[17:04:35] <pintoch>	 probably…
[17:17:09] <pintoch>	 Tpt[m]: it's amazing! just with a simple python server that does a faithful MITM between the Java program and the Wikidata API, the bug disappears :-D
[17:17:51] <pintoch>	 I guess the requests library makes nicer POST requests than that Java library… oO
[17:18:11] <Tpt[m]>	 probably
[17:18:20] <Tpt[m]>	 the JDK library is very bad
[17:29:58] <Tpt[m]>	 pintoch: I have found the buggy line: https://github.com/Wikidata/Wikidata-Toolkit/pull/272/commits/07a1aa3b34de2a79f20f633726dd7f47412a21db
[17:31:01] <pintoch>	 apparently settings "Content-Length" like that does not do anything
[17:31:22] <pintoch>	 does it fix the bug if you remove it?
[17:33:37] <pintoch>	 for me it doesn't…
[17:35:04] <pintoch>	 ah but you've made other changes
[17:35:08] <pintoch>	 let me try your branch
[17:37:42] <pintoch>	 Tpt[m]: it works!! \o/
[17:37:49] <pintoch>	 thanks a lot <3
[17:41:56] <Tpt[m]>	 Great!
[18:11:47] <SoniWP_>	 Hi
[18:11:57] <SoniWP_>	 Quick question regarding Wikidata queries
[18:12:10] <SoniWP_>	 is there a simple way to say "Property does not exist"
[18:34:40] <pintoch>	 SoniWP_: what do you mean?
[18:34:51] <pintoch>	 all statements are associated to valid properties
[18:35:09] <SMalyshev>	 SoniWP_: setting the value to "no value" probably would work if I get what you mean
[18:35:54] <SoniWP_>	 https://www.irccloud.com/pastebin/CH8AR6xS/
[18:36:02] <SoniWP_>	 This is essentially what I used
[18:36:05] <SoniWP_>	 It seems to work
[18:59:08] <SMalyshev>	 SoniWP_: oh, you mean query when there's no value... Yeah, MINUS or OPTIONAL/!bound would work. Also probably FILTER NOT EXISTS but it could be slower (depends on query)
[18:59:58] <SoniWP_>	 MINUS worked for me. And yea, any optimised query will be good, given my final result is expected to be in the range of 50k results :)
[19:04:28] <SoniWP_>	 Hi again.
[19:04:51] <SoniWP_>	 I would like some help with queries. Basically this is my query so far - http://tinyurl.com/yaumyl5m
[19:05:09] <SoniWP_>	 https://www.irccloud.com/pastebin/h9Lv9dF4/Query
[19:05:26] <SoniWP_>	 And I would like it to show numbers of each settlement type as well
[19:05:42] <SoniWP_>	 Rather than just what are the different settlement types available.
[19:05:49] <SoniWP_>	 What will be my way of doing it?
[19:06:25] <SoniWP_>	 SELECT ?settlement COUNT(DISTINCT (?settlement)) or something along those lines?
[19:07:36] <WikidataFacts>	 SoniWP_: http://tinyurl.com/y979yt3l, or more efficiently, http://tinyurl.com/y7khw7sx
[19:08:29] <SoniWP_>	 @WikidataFacts That is two queries with just numbers
[19:08:45] <SoniWP_>	 Also is this a bot :P
[19:08:47] <WikidataFacts>	 ah FUCK
[19:08:51] <WikidataFacts>	 I fucking hate NoScript
[19:09:01] <WikidataFacts>	 I keep forgetting that it “protects” me from XSS attacks with that shit
[19:09:02] <WikidataFacts>	 sorry
[19:09:07] <WikidataFacts>	 correct queries coming up
[19:09:07] <SoniWP_>	 NP
[19:09:29] <SoniWP_>	 (Going afk for a few minutes, can I ping you when I get back? Ty)
[19:09:33] <WikidataFacts>	 sure
[19:38:02] <SoniWP_>	 Hi 
[19:38:07] <SoniWP_>	 @WikidataFacts 
[19:38:14] <WikidataFacts>	 SoniWP_: http://tinyurl.com/ybfylwcj, or more efficiently, http://tinyurl.com/y7hc6kju – and a slightly different suggestion, http://tinyurl.com/ybbbwb2y, which doesn’t count classes that aren’t subclasses of “human settlement”
[19:38:15] <WikidataFacts>	 (e. g. some items are apparently instance of some subclass of human settlement and also instance of lake?)
[19:43:58] <nikki>	 WikidataFacts: do you have some examples? that sounds wrong
[19:44:38] <WikidataFacts>	 nikki: I just saw that the results of the first two queries included those classe
[19:44:40] <WikidataFacts>	 *classes
[19:44:49] <SoniWP_>	 http://tinyurl.com/yd4xuxww
[19:44:51] <SoniWP_>	 @nikki
[19:45:12] <WikidataFacts>	 SoniWP_: you beat me to it :)
[19:45:14] <SoniWP_>	 This looks like a nice way to clean up all the weirdly labelled locations
[19:45:25] <WikidataFacts>	 (you can remove the /wdt:P279* in L7 btw)
[19:45:55] <SoniWP_>	 That was deliberate. I didnt want to skip weird subclassing if I changed Lake to something else
[19:46:21] <WikidataFacts>	 ok
[19:47:46] <SoniWP_>	 I plan to have these two queries as a way to figure out weird location parameters. First use the first query to find all labels that human settlements also use (city/town/lakes etc). And then query for those labels to find all the items that are categorised that way
[19:48:08] <nikki>	 ah, not too many, thankfully
[19:48:37] <nikki>	 I was worried there would be a load of badly linked svwiki/cebwiki articles :P
[19:48:53] <SoniWP_>	 There's 14 items that are in the list which are classified as "village development committee of Nepal"
[19:49:03] <SoniWP_>	 I was only searching for locations in India
[19:50:27] <WikidataFacts>	 here’s a more generic query for “instance of human settlement and also something else” btw: http://tinyurl.com/ybkp7a9t (though I can’t find any way to add labels to it without timeout)
[19:55:16] <muelli>	 I'm looking at https://de.wikipedia.org/wiki/Liste_der_Hochschulen_in_Deutschland and there is that table with the "state" the university is located.  When trying to recreate that table from wikidata, should that state be inferred from the university's location?  Or should each university be extended with a property?
[19:57:21] <nikki>	 inferred from the location (if you're using sparql, it shouldn't be difficult to follow the p131 statements, if not, I'm not sure what the best way to do it would be)
[19:57:33] <SoniWP_>	 @WikidataFacts From the previous query, I found something more interesting. 
[19:57:43] <SoniWP_>	 http://tinyurl.com/y7p2deb7 Locations in India AND a neighbouring country
[19:59:27] <WikidataFacts>	 muelli: I agree with nikki, but I’m not sure what the start property is. I looked at https://www.wikidata.org/wiki/Q1388301 as a random example and it doesn’t have a “location” or “located in administrative territorial entity”, just two “headquarters location”s
[20:00:07] <muelli>	 WikidataFacts: hm. it has a "coordinate location" thugh
[20:00:08] <nikki>	 one of the village development committee ones I looked at appeared to be vandalism or someone who didn't know what they were doing
[20:00:30] <WikidataFacts>	 SoniWP_: nice :) but I think this one’s even better: settlements in India and a NON-neighbouring country :D http://tinyurl.com/y7qrajvn
[20:01:01] <SoniWP_>	 Ah filter. Thats what I wasnt able to figure out.
[20:01:11] <SoniWP_>	 Thats why I went for neighbouring country rather than minus
[20:01:22] <SoniWP_>	 not equals, I mean
[20:01:36] <WikidataFacts>	 muelli: yeah, but we can’t currently use that to find the containing country, at least not in the query service
[20:02:16] <muelli>	 ah. hm. so what was nikki referring to then?
[20:03:11] <WikidataFacts>	 muelli: usually you can follow the chain of P131 “located in the administrative territorial entity”
[20:03:27] <WikidataFacts>	 e. g. https://www.wikidata.org/wiki/Q309988 is located in Karlsruhe, which is located in Karlsruhe Government Region, which is located in Baden-Württemberg
[20:03:58] <WikidataFacts>	 you follow that until you find something that’s an instance of state of Germany
[20:04:02] <WikidataFacts>	 that’s possible to do in a query
[20:05:24] <muelli>	 ah. okay, yeah, I get that.  Then the universities would need to get added that  located in the administrative territorial entity (P131) property.
[20:05:50] <muelli>	 with my question "should that state be inferred from the university's location?" I meant the coordinates
[20:05:59] <WikidataFacts>	 oh, okay
[20:06:23] <SoniWP_>	 What does this line do
[20:06:28] <SoniWP_>	 https://www.irccloud.com/pastebin/labMNFXl/
[20:06:35] <muelli>	 because it's somewhat redundant, maybe.
[20:07:18] <WikidataFacts>	 SoniWP_: it adds Label variables (e. g. ?objectLabel, ?country2Label)
[20:07:25] <muelli>	 like you can probably infer that P131 from the coordinate location, right? (Assuming that we don't have vertically separated areas of administration -.-)
[20:07:38] <WikidataFacts>	 usually it’s written in one line to save some space, e. g. SERVICE wikibase:label { bd:serviceParam wikibase:language "en,fr". }
[20:07:49] <WikidataFacts>	 muelli: or temporally separated?
[20:07:50] <nikki>	 we don't have boundaries for the areas, so not very easily
[20:07:57] <WikidataFacts>	 yeah
[20:08:01] <SoniWP_>	 So if I have a Wikidata item with no french or english labels, would those be also returned by the query?
[20:08:10] <WikidataFacts>	 in theory we can, with geoshapes on commons, but we can’t use those in the query service yet
[20:08:18] <WikidataFacts>	 and even then, that’s inefficient
[20:08:27] <WikidataFacts>	 SoniWP_: yes, it would just show the Q-number instead
[20:08:29] <WikidataFacts>	 but you can also add more languages
[20:08:40] <nikki>	 we would also need to have the data, and I really doubt we would have the shapes for every administrative unit in every country... might work for germany though
[20:09:07] <SoniWP_>	 Cool. I was slightly worried because I didnt understand what that did
[20:11:40] <muelli>	 hm. assuming that "sovereign states" are an "administrative territorial entity" then "country" is redundant to P131, no?
[20:12:14] <nikki>	 yes, but it's quite handy
[20:12:23] <muelli>	 ;-
[20:12:25] <muelli>	 ;-)
[20:12:30] <WikidataFacts>	 I don’t think it’s redundant for historical entities
[20:12:30] <nikki>	 I guess it's a compromise between duplicating all of the levels in p131 and not duplicating any of them :)
[20:12:38] <WikidataFacts>	 but I’m not sure how we handle “country” for those anyways
[20:12:43] <muelli>	 WikidataFacts: then you have the end time qualifier
[20:12:44] <nikki>	 badly, probably
[20:13:07] <WikidataFacts>	 muelli: but following a chain and looking at qualifiers at every point in the chain isn’t possible in a query
[20:13:14] <muelli>	 oh
[20:13:30] <WikidataFacts>	 nikki: yeah… Magnus used to have a list years ago of people whose lifetime didn’t intersect with the time of their country
[20:13:31] <muelli>	 well. that's a technical limitation that can be overcome, I assume.
[20:13:45] <WikidataFacts>	 IIRC Germany was by far the longest entry, due to both coverage and history
[20:14:12] <SoniWP_>	 @WikidataFacts I found an interesting query
[20:14:14] <SoniWP_>	 http://tinyurl.com/y8yecckc
[20:14:16] <WikidataFacts>	 “but surely Goethe must have been from Germany!”…
[20:14:23] <SoniWP_>	 List of settlements in two countries
[20:14:37] <SoniWP_>	 (The query has duplicates, but still 2.5K entries)
[20:14:59] <muelli>	 Can I somehow conveniently browse subclasses of "university" Q3918 ?
[20:15:23] <muelli>	 I mean, I can probably come up with a simple SPARQL Query, but I was wondering whether there was already something available.
[20:15:37] <WikidataFacts>	 muelli: not sure, but the query would be http://tinyurl.com/yd72p5fc
[20:15:39] <nikki>	 there's https://tools.wmflabs.org/wikidata-todo/tree.html?q=3918&rp=279
[20:15:53] <SoniWP_>	 @muelli I did almost exactly something similar with SPARQL. http://tinyurl.com/yb4tprgv
[20:16:00] <WikidataFacts>	 or https://angryloki.github.io/wikidata-graph-builder/?property=P279&item=Q3918&mode=reverse
[20:16:16] <SoniWP_>	 Ooh, I like that. Ty :)
[20:16:29] <muelli>	 ah nikki. that looks interesting. It's not limited to classes it seems, though.
[20:16:40] <muelli>	 ah WikidataFacts. that looks amazing.
[20:17:03] <WikidataFacts>	 actually, for this shallow tree, I think the tree nikki linked to looks much better :)
[20:17:10] <SoniWP_>	 @WikidataFacts Can you look at my query from above? The settlements from two trees.
[20:17:11] <WikidataFacts>	 I can barely read anything in the angryloki tree
[20:17:26] <muelli>	 except it doesn't label Q3803846 for me. Probably because it has a Frenhc label only. But I'd be more happy with the French label rather than the Q number
[20:18:12] <SoniWP_>	 *two countries. Sorry I'm getting confused :)
[20:18:35] <WikidataFacts>	 SoniWP_: which query? I think I got the wrong link
[20:19:19] <SoniWP_>	 @WikidataFacts http://tinyurl.com/yacedqhj
[20:19:39] <WikidataFacts>	 ah, okay
[20:20:36] <muelli>	 hm. this one has wrongly set the subclass of property, no?  https://www.wikidata.org/wiki/Q1811146  Because it seems to refer to an actual university rather than a class of universities.
[20:21:22] <WikidataFacts>	 SoniWP_: a lot of the results seem to be due to history
[20:21:44] <WikidataFacts>	 muelli: at a glance – agreed
[20:21:50] <SoniWP_>	 Like previously in a country, and currently in another?
[20:22:03] <WikidataFacts>	 yeah
[20:22:08] <WikidataFacts>	 look at the list for Constantinople
[20:23:33] <SoniWP_>	 @WikidataFacts we could filter out anything that has a start time or an end time then?
[20:23:45] <muelli>	 can I quickly determine the number of instances of such a class? e.g. https://www.wikidata.org/wiki/Q1811146  Like is there is a quick link or so from the wikidata page s.t. I don't need to assemble a SPARQL Query now?
[20:24:26] <muelli>	 SoniWP_: No need to prefix the names with that @ sign.  Let's save that byte for something else.
[20:24:35] <nikki>	 hm. those soviet union ones look like another statement needs marking as preferred
[20:24:49] <WikidataFacts>	 nikki: oh, good point
[20:25:15] <SoniWP_>	 @muelli Autocomplete from IRCCloud. Preferred to use tab to notify properly :)
[20:26:16] <nikki>	 quite a few disputes by the look of it too
[20:26:58] <SoniWP_>	 Any way I can filter out all settlements with a start time?
[20:27:16] <SoniWP_>	 So we only have disputed teritorry and mislabeled ones?
[20:28:56] <SoniWP_>	 Also +1 to @muelli's question. Will be a helpful thing to check.
[20:29:53] <WikidataFacts>	 muelli: not that I know of… but I’m the guy who would immediately reach for SPARQL anyways :)
[20:30:24] <muelli>	 yeah, I guess it's coming much more naturally after a while.
[20:30:56] <SoniWP_>	 Okay cool. 
[20:30:56] <WikidataFacts>	 though occasionally I do use “what links here” if I suspect it’s a short list
[20:31:07] <WikidataFacts>	 and I know there’s a gadget, EasyQuery or something like that
[20:31:28] <SoniWP_>	 And can you help me figure out how to better my previous "Part of multiple countries" query
[20:32:34] <Ortano>	 Hi, I want to undo a merge of two wikidata items. Undoing the merge in the target item was successful. However, when I want to restore the per-merge version of the source-item, i receive an error: Undo failed Invalid content data. Can anyone tell me, what I am doing wrong?
[20:33:53] <Ortano>	 The target item of the merge was https://www.wikidata.org/wiki/Q27721 and the source item is https://www.wikidata.org/w/index.php?title=Q38906046&redirect=no .
[20:35:01] <Ortano>	 So this restoration operation fails: https://www.wikidata.org/w/index.php?title=Q38906046&action=edit&restore=553616257
[20:36:06] <WikidataFacts>	 SoniWP_: here’s a suggestion: remove all countries with end time or “dissolved or abolished” http://tinyurl.com/ya2nweux
[20:36:13] <WikidataFacts>	 (slightly easier than going for qualifiers on the settlements)
[20:37:31] <SoniWP_>	 I like that. I doubt it will work for all places but thats a start (Places that were part of one contry but given to another, say)
[20:38:19] <SoniWP_>	 Also thanks for the less than symbol. I couldnt figure out the STR part so that confused me
[20:39:01] <WikidataFacts>	 yeah, directly comparing the variables doesn’t work
[20:39:39] <SoniWP_>	 So that solves problem 1
[20:39:50] <SoniWP_>	 Now I am thinking of how to filter out disputed teritorry
[20:40:16] <muelli>	 another thing I'm interested in is seeing what the most used properties of the universities are, s.t. I know what queries make most sense. E.g. if only one university has their, say, budget, set,  I'm less inclined to ask to know about that for the instances of universities. But if all universities have a property, I'm much more interested in comparing it.
[20:41:08] <SoniWP_>	 @WikidataFacts Does a disputed teritorry property already exist? I kind of want to try that as a filter, but cannot locate the property
[20:43:04] <muelli>	 What does "STR(?country2) < STR(?country)" do? STR sounds like "string representation" to me.
[20:43:36] <WikidataFacts>	 yes, it converts the entity URI to a string
[20:43:41] <WikidataFacts>	 because URIs aren’t comparable
[20:43:58] <WikidataFacts>	 with ?country != ?country2, you get each result twice, with ?country and ?country2 swapped
[20:44:06] <WikidataFacts>	 using < removes one of those duplicates
[20:44:59] <WikidataFacts>	 Ortano: if it was a very old revision I would suspect that it was stored in the database using an older content model that’s no longer supported
[20:45:07] <WikidataFacts>	 but since the revision is only a week old, that seems unlikely
[20:47:22] <WikidataFacts>	 Ortano: aha! rollback and then restore worked
[20:47:31] <WikidataFacts>	 no idea why
[20:48:26] <nikki>	 SoniWP_: we have "statement disputed by" which we use as a qualifer
[20:48:47] <SoniWP_>	 @muelli I think you need something like this for how to find the most used properties of universities
[20:48:48] <SoniWP_>	 http://tinyurl.com/y88ksf2u
[20:49:02] <SoniWP_>	 (Not the cleanest query, but it works, I think)
[20:49:45] <muelli>	 cool SoniWP_. Thanks!
[20:49:50] <Ortano>	 WikidataFacts: great! thanks.
[20:50:29] <SoniWP_>	 @muelli See if you can find a way to add the label names to each of the properties. I kindof want to know how to do that. So if anyone can help with that, it sounds great
[20:50:42] <muelli>	 can it somehow be improved to show the "name" of the property? It's quite... involved to find out what these properties are right now
[20:50:45] <muelli>	 ah, yeah
[20:56:04] <muelli>	 hm. I'm confused why the propertyLabel shows the URL rather than the actual label of the property.
[21:06:16] <SoniWP_>	 @muelli I think I figured out why, just because of a completely different reason
[21:07:10] <SoniWP_>	 @muelli See an updated list of "Properties of universities" at http://tinyurl.com/yd7ozjff
[21:07:33] <SoniWP_>	 I was trying to remove duplicates and was trying some sort of string matching at ?property
[21:08:23] <SoniWP_>	 When I realised that for property, wdt:Pxyz is not the string. The full URL is (www.wikidata.org/prop/direct/Pxyz)
[21:08:55] <SoniWP_>	 And thats basically why this is acting so weird. Because wdt:Pxyz is treated differently than wd:Pabc
[21:09:04] <SoniWP_>	 Thats my theory on whats going on
[21:09:38] <muelli>	 hrm. filtering duplicates with string matching sounds wrong to me.
[21:09:45] <muelli>	 hm SoniWP_. but it still doesn't show "located in the administrative territorial entity" for P131.
[21:09:48] <SoniWP_>	 Can anyone else look over http://tinyurl.com/yd7ozjff and http://tinyurl.com/y88ksf2u to see if thats correct?
[21:10:41] <SoniWP_>	 @muelli It does. I was just trying a quick and dirty way to get only items of the type wdt:Pxyz
[21:11:08] <SoniWP_>	 Hopefully someone else here can figure out how to get the name of the ?property from the ?property
[21:11:24] <SoniWP_>	 @WikidataFacts maybe, if they're not too busy?
[21:12:52] <WikidataFacts>	 SoniWP_: http://tinyurl.com/y7cf2kda
[21:13:42] <SoniWP_>	 Lovely. There you go @muelli 
[21:13:53] <muelli>	 cool.
[21:14:26] <SoniWP_>	 Also @WikidataFacts do you know of any way I can filter out all the disputed settlements from the query i had?
[21:14:50] <SoniWP_>	 I was hoping to find a property called disputed teritorry or something
[21:14:57] <WikidataFacts>	 no idea how that’s modeled
[21:15:05] <WikidataFacts>	 “statement disputed by” qualifier? (I think that property exists)
[21:15:38] <muelli>	 multiple P1336s
[21:15:55] <WikidataFacts>	 oh
[21:19:01] <muelli>	 What's the difference between  ?object wdt:P31/wdt:P279* wd:Q3918.   and     ?object wdt:P31/wdt:P279 wd:Q3918.   i.e. the missing "*".  I assume the asterisk version to be more inclusive and to be the thing one generally wants.
[21:19:12] <WikidataFacts>	 yeah
[21:19:21] <WikidataFacts>	 wdt:P31/wdt:P279 is a path with one P31 and then one P279
[21:19:32] <WikidataFacts>	 i. e. instance of a subclass of wd:Q3918
[21:19:44] <WikidataFacts>	 but not directly instance of Q3918, or instance of subclass of subclass of wd:Q3918
[21:20:03] <SoniWP_>	 Nice.
[21:20:05] <WikidataFacts>	 * marks a path component of any length (including zero: for one or more there’s +)
[21:20:13] <muelli>	 SoniWP_: ^^ That's for you then. Because you seem to like the non-asterisk version.
[21:20:19] <WikidataFacts>	 so wdt:P31/wdt:P279* means instance of, or instance of subclass, or of subclass of subclass, …
[21:20:44] <SoniWP_>	 I like the asterisk, just forget it sommetimes :)
[21:20:58] <muelli>	 WikidataFacts: Is it possible to add, say, up to five items to the properties list to be able to see who is actually using the properties?  I mean I could probably SPARQL together a query to search for the properties, but if it's possible in one query, then I'm more happy, I guess.
[21:22:40] <WikidataFacts>	 you can add one example object: http://tinyurl.com/yau9xqzz
[21:22:45] <WikidataFacts>	 but I’m not sure how to easily get more samples
[21:23:47] <muelli>	 I've seen a GROUP_CONCAT or something somewhere.
[21:24:03] <WikidataFacts>	 hm, yeah, but that’ll get out of hand for the top properties
[21:24:26] <muelli>	 hm, yeah, needed to be limited to up to five items
[21:25:10] <WikidataFacts>	 http://tinyurl.com/y8zurx9j
[21:25:22] <WikidataFacts>	 unfortunately GROUP_CONCAT of entity URIs doesn’t display very nicely in the query UI
[21:26:43] <muelli>	 ah. right. but that also shows only those properties with at most five users.
[21:27:06] <SoniWP_>	 I dont think this query does what you think it does @WikidataFacts 
[21:27:15] <muelli>	 thanks WikidataFacts, btw, for keeping up with my ignorance reg. reading the SPARQL spec.  I mean all the syntax and keywords is buried in there and I would just need to go and read it...
[21:27:30] <WikidataFacts>	 muelli: np, it takes a lot of time to learn all that :D
[21:27:33] <WikidataFacts>	 SoniWP_: why not?
[21:28:03] <SoniWP_>	 What @muelli wants is "I want the five properties most used together for universities, sorted in Descending order"
[21:28:27] <SoniWP_>	 I think
[21:28:59] <SoniWP_>	 Is that correct?
[21:29:04] <WikidataFacts>	 SoniWP_: I think the query does do what I think it does – see the comment at the top
[21:29:15] <WikidataFacts>	 that may not be what muelli wanted, but that’s a different question :P
[21:29:24] <SoniWP_>	 I see :P
[21:29:29] <WikidataFacts>	 adding a limit to the GROUP_CONCAT isn’t possible unfortunately
[21:29:32] <muelli>	 ;-)
[21:29:42] <WikidataFacts>	 (well, hm, I suppose it is… ’ang on a minute)
[21:29:44] <SoniWP_>	 What did @muelli want then?
[21:30:03] <muelli>	 Maybe there is a better SAMPLE() that has a limit
[21:30:43] <WikidataFacts>	 regexes to the rescue :P http://tinyurl.com/y82qyb2t
[21:30:59] <WikidataFacts>	 all the properties, with up to five sample entities
[21:31:13] <WikidataFacts>	 (not necessarily random ones, mind)
[21:31:56] <SoniWP_>	 I'm still trying to figure out what @muelli wants btw :P
[21:32:02] <muelli>	 woahr. that looks.. like a hack.
[21:32:56] <muelli>	 SoniWP_: The list of properties sorted by the number of times they are used. Like we had in that other query. But in addition a few example items to click on to see which items are actually using the property.
[21:33:43] <SoniWP_>	 Aaah
[21:33:55] <SoniWP_>	 That makes a lot more sense than what I thought
[21:35:39] <muelli>	 I'm a bit surprised that GROUP_CONCAT works on the item itself, having assumed it to work on string types only. I somehow expected to provide a serialisation to a string. I'm happy, of course, that it seems rather to be easy.
[21:36:01] <WikidataFacts>	 yeah, not sure if that’s in line with the spec or if Blazegraph is just being generous to us and adds an implicit STR()
[21:39:41] <muelli>	 can I somehow run queries across wikidata and dbpedia?  Or more conveniently import from dbpedia?
[21:40:33] <WikidataFacts>	 you can use a federated query, like https://www.wikidata.org/wiki/Wikidata:SPARQL_query_service/queries/examples#Universities_ranked_by_PageRank_on_English_Wikipedia_.28federated_query.29
[21:41:20] <muelli>	 that's probably what I want.
[21:47:55] <muelli>	 is there a list of these services? So that I might be able to locate dbpedia.
[21:50:49] <WikidataFacts>	 in principle it can be any URI, but we have a whitelist so WDQS isn’t abused to DoS-attack random servers out there
[21:50:49] <WikidataFacts>	 https://www.mediawiki.org/wiki/Wikidata_query_service/User_Manual#Federation
[22:03:03] <muelli>	 ah, cool.
[22:59:31] <YairRand>	 https://rawgit.com/YairRand/universe_tree/master/main.html < tool for mapping "part of" relations
[23:02:52] <YairRand>	 the tree is currently a bit of a mess, due to some mistakes on various items