[03:11:05] If I want to make an external web page (no back-end, being served via github for example), that needs the user to login into Wikidata so that they can make edits using the tool, then is OAuth the only choice I have? [09:25:32] Need assistance with creating a page for a musical artist [09:47:51] One minute. They stayed *one* minute [09:47:53] * reosarevok sighs [10:15:47] Lucas_WMDE: Check out the nice big warning at https://www.wikidata.org/wiki/Property_talk:P350 [10:16:11] *sigh* [10:16:17] but that’s a different thing again, isn’t it? [10:16:40] Where are you hiding? I'm in the main hallway [10:16:40] the fact that Module:Constraints produces this warning doesn’t necessarily mean that KrBot still doesn’t support mandatory constraints with exceptions [10:16:50] Wikicite room [10:18:09] multichil: https://www.wikidata.org/wiki/Template_talk:Constraint#Incompatibility_between_exceptions_and_mandatory [10:18:22] and https://www.wikidata.org/wiki/Template_talk:Constraint#Mandatory_constraints_with_exceptions [10:24:40] Lucas_WMDE: Thanks, I'll probably go do something else now ;-) [10:28:39] SMalyshev: so say I have a Wikidata query that is computationally simple but the result is large because it involves several million Wikidata items. Right now what happens is that I try and the response gets cut off. Do I have any options to instead ~stream~ the data? [10:29:17] harej: not really... you can use limit/offset probably [10:30:38] basically increment the offset until the result set is zero? [10:30:47] harej: yes [10:30:58] or less than limit, essentially [10:31:46] What if for my purposes I need all the results? [10:32:04] well, you'll get all results, once you're done [10:32:35] so... say i do 100,000 at a time. limit 100,000 offset 0, then offset 100,001, then 200,001, something like that? [10:32:41] yes [10:33:00] that would be slower but should work if the query is simple I think [10:33:10] query is just "give me all items with this property" [10:33:28] hmm... for this you may be ok with LDF [10:33:43] That was the thing I was told about yesterday and then promptly forgot [10:34:04] just put property in the pattern and you'd get the rest [10:34:24] any examples I can use as my starting point? [10:34:27] another alternative would be dump + wikidata toolkit or plain old grep/awk :) [10:34:49] well right now I am downloading the dump but downloading 23 GB takes a while even on good UAB wi-fi which only goes offline sometimes! [10:35:57] harej: https://query.wikidata.org/bigdata/ldf?subject=&predicate=http%3A%2F%2Fwww.wikidata.org%2Fprop%2Fdirect%2FP527&object= [10:36:27] that eventually gives you all items with P527 [10:36:47] (non-deprecated ones, if you care, use p: one instead) [10:37:06] can I get this in JSON? [10:37:23] hmm I think so [10:39:24] https://www.mediawiki.org/wiki/Wikidata_Query_Service/User_Manual#Linked_Data_Fragments_endpoint [10:55:46] I need to talk to somebody that knows how Wikidata UI (and specifically entity selector) works inside. Who should I be talking to? [10:56:54] 16 million items with PubMed IDs by the way. https://query.wikidata.org/bigdata/ldf?subject=&predicate=http%3A%2F%2Fwww.wikidata.org%2Fprop%2Fdirect%2FP698&object= [10:58:29] harej: so I assume it worked? [10:59:18] SMalyshev: the query does; now I'm trying to get it in JSON via content negotiation (alas) and seeing how the data is structured [10:59:51] ok, the data should be JSON-LD [11:00:03] not sure I remember more details though :) [11:00:58] SMalyshev: not sure who you need to talk to, but perhaps Thiemo? [11:01:12] Lucas_WMDE: ok, thanks [11:01:41] I need to do some things with entity selector, so whoever knows it best (on the UI side) is who I need [11:10:58] SMalyshev: when there is no nextpage because i've iterated through all the results, is nextpage blank or just not there? [11:45:11] harej: hmm I do not really remember... probably the easiest thing is to take something like https://query.wikidata.org/bigdata/ldf?object=%22topic%2FSeptember%22 and look how it looks like in JSON [11:46:15] SMalyshev: https://www.wikidata.org/wiki/Wikidata:SPARQL_query_service/Federation_report [11:46:56] harej: looks like no next page [11:47:08] multichil: coolio! thanks [11:48:08] multichil: britishart thing works for me though [11:49:03] hmm wait maybe not... looks like it does not return sparql result format. weird [11:50:35] looks like the correct URL is http://collection.britishart.yale.edu/openrdf-sesame/repositories/ycba [11:51:33] multichil: maybe you could publish the script that made the table? I'd run it periodically [12:23:22] also it looks like some sparql endpoints do not understand nested queries :( [12:56:21] SMalyshev: Scheduled to run once a week [13:51:52] Plz [13:52:44] Im gone in fife min [13:52:53] ? [13:53:12] Yee [13:53:18] ? [13:53:29] I kno it u [13:54:58] Just u. Yes or N. Im tired. [13:55:20] good night then :) [13:55:25] Lydia_WMDE: you mentioned that it is possible to apply tags when editing items with the API, right? [13:55:34] Omg [13:56:00] I c signs 4real [13:56:18] I cannot figure out how to do that in the same request: there is no "tags" parameter in "wbeditentity" according to https://www.wikidata.org/w/api.php?action=help&modules=wbeditentity [13:57:37] My ears r pinging baddd. [13:58:06] don't drink and chat [13:58:30] U kno i dnt drink. Bye [13:58:49] Not u. Srry to bother [14:29:35] I am Artix/mechquester. ask me anything. I'll be here for 30 minutes [14:29:51] multichil: sounds good! [14:30:43] someone12: who are you? :-) [14:30:58] relevant to those who know. [14:33:36] enjoy your 30 minutes of silence then :) [16:47:51] SMalyshev: the LDF output doesn't seem to be in any particular order. But is it at least in a deterministic order so that I can rely on the pagination without having to sort it first? [17:21:37] Also, can you do more than 100 results per page? 13 million triples with only 100 results per page = 130,000 GET requests [18:17:22] Lucas_WMDE: Still at the hackathon? [18:17:30] yeah [18:17:37] main hallway [18:17:43] just gone for a few minutes, brb [18:19:29] Lucas_WMDE: Can you have a look at http://tinyurl.com/y9znv2c4 ? If you remove the comments, it's not returning anything. [18:19:34] Missing something here..... [18:20:25] the optimizer might be messing with the query [18:21:15] oh, no, wait [18:21:33] STR(something) as ?itemLabel can’t be right, labels always have language tags [18:21:43] Something with joining the two strings [18:24:45] so the main problem, I think, is that you’re trying to find ?item inside the lov.okfn.org service [18:24:54] which doesn’t know anything about Wikidata [18:25:05] but independently of that, the optimizer is also causing trouble… [18:25:16] in conclusion: http://tinyurl.com/y88k5ozg [18:25:26] multichil: that one should work better [18:26:57] or, slightly cleaned up: http://tinyurl.com/ych2vtsc [20:43:58] pintoch: i thought so but can't find any useful docs. Maybe Amir1 can help? [20:44:15] Lydia_WMDE: sup? I'm in the hotel lobby [20:44:41] Amir1: pintoch was looking for documentation about how to apply edit tags to an edit done via the api [20:45:53] Lydia_WMDE pintoch: in the code, it's possible. we don't have API for that (yet) we need to make it once we got the tables less horrible [20:46:37] Amir1: how does quickstatements do it? [20:46:52] abuse filter I think [20:46:57] can't say for sure [20:48:54] ok [20:49:01] thx [20:49:16] let me double check [21:01:14] Lydia_WMDE: it's probably from OAuth [21:01:24] ok cool