[07:39:41] My chat client says that I am banned from #wikidata - is just lurking a crime? [13:36:37] leszek_wmde: https://gerrit.wikimedia.org/r/398443 [17:59:55] Woah, click *Usage history for last 50 days* https://www.wikidata.org/wiki/Property_talk:P497 [18:03:31] Oh, added with 828 edits per minute... Okay! http://wikidata.wikiscan.org/?menu=dates&date=20171215&list=users&filter=all&sort=weight [18:03:48] Wonder when he started... http://wikidata.wikiscan.org/gimg.php?type=edits&date=20171214&size=big [18:04:50] wow, that is fast [18:04:53] #totallynotabot [18:05:04] And now we end up with items like this: https://www.wikidata.org/wiki/Q45728947 [18:06:04] Worst Dispatch was half an hour though... [18:06:36] I don't think it's healthy to do 1 million edits without approval tbh [18:07:20] "I may edit some entities" https://www.wikidata.org/wiki/Wikidata:Bureaucrats%27_noticeboard#Flood_flag_request [18:10:47] sjoerddebruin: what the fuck? do the edits make sense, at least? [18:10:53] Is the Wikidata search API mutlilingual? I can search in any locale but the results are always returned in English [18:11:13] DanielK_WMDE: well, as you can see they only contain P31:Q5 and some identifier. [18:11:29] And he also added P31:Q5 to items that already had it [18:11:29] Well he did edit some entities :D [18:11:43] tobias47n9e: try the uselang parameter [18:12:32] Oh, these are all newly created? That's much dodgier, yeah :/ [18:12:40] 303,338 new items, yes. [18:13:32] So for all we know they might be dupes, since we haven't seen their code and have no idea how they're avoiding dupe creating, right? [18:13:57] I've removed the mandatory parameter of the unique value constraint as it was like 500 violations. [18:16:05] DanielK_WMDE: That works. Is there a reason why it is not documented here: https://www.wikidata.org/w/api.php?action=help&modules=wbsearchentities ? [18:17:15] sjoerddebruin: is there a way to at least force the person to add the rest of the info available on that site, to make the items a tiny bit easier to disambiguate and identify? [18:17:28] It seems part of the process. [18:17:34] See https://www.wikidata.org/wiki/Wikidata:Data_Import_Hub#Discussion_16 [18:17:43] "Done: Import most male in the database." [18:17:49] Oh, the gendergap community would love that! [18:17:50] https://www.wikidata.org/wiki/Q45728947 has at least a gender that has not been marked [18:18:25] Oh, "to do more claims" I see [18:18:36] Can he be blocked until we see a bot request= [18:18:38] *? [18:19:05] Well, I also left a message on the Bureaucrats' noticeboard [18:19:16] He should do this on a separate account. [18:19:35] The bot policy doesn't seem to apply to people doing it on their personal accounts (including semi-automatic edits) [18:20:37] That is something I was actually curious about - QS and PetScan are basically botting, but don't require community [18:20:39] approval [18:20:46] Indeed. [18:21:16] Which is nice in that they really help bring those items with no claim numbers down, but eh [18:21:25] It's something I proposed during some problems on Wikidata session at WikidataCon, we need a policy and approval process for large tasks. [18:21:40] It wouldn't hurt for 100 items imo, but this is.... [18:22:16] I think the flag should be removed from this user, flood flag is not used for discussion https://www.wikidata.org/w/index.php?limit=200&title=Special%3AContributions&contribs=user&target=Fantasticfears&namespace=0&nsInvert=1&tagfilter=&start=&end= [18:22:18] Yeah. I've done what was basically a bot run recently with QS tbh removing "Category:" from the English label of P31:street items and stuff like that [18:22:30] I'm pretty sure it's all good, but it was also several thousand items [18:23:17] Stryn: maybe ping one of the two bureaucrats that are active? [18:23:18] And while that was fine, it's pretty much as easy to change all English labels to "LOL"... [18:23:29] (the other one quietly did some deletions while almost losing his or her role) [18:23:37] ah I don't know why but I remembered you're a 'crat :P [18:23:51] Oversighter. :P [18:23:56] yeah [18:23:57] I mean, I can make myself crat... [18:45:04] i'd like to write a gadget to migrate {{Property proposal}} to claims on the property, so I'm looking for a wikicode parser in javascript that I could call from a script in common.js. Any ideas? [18:45:04] 10[3] 10https://www.wikidata.org/wiki/Template:Property_proposal [18:46:55] https://commons.wikimedia.org/wiki/MediaWiki:Gadget-libWikiDOM.js [18:54:56] sjoerddebruin: thanks! :) [19:01:28] tobias47n9e: because uselang is a generic parameter that overrides the user's UI language. it works in all api modules, and not just there. it actually works for all requests! [19:01:47] so, it is documented, just not specifically for this module [19:02:01] the module just returns output in the user's ui language [19:03:16] DanielK_WMDE: Oh i see. Thanks :) [19:21:43] hi guys, i've seen there's a copy of wikidata in tool labs (or cloud) [19:22:11] is that a database I can send queries to connecting with python? [19:23:14] i haven't used sparql yet, perhaps it works different than mysql [19:23:35] and i can only use the real wikidata database [19:24:33] marmick: what do you mean by "real"? [19:24:44] marmick: if you want to query content, you will have to use SPARQL. [19:25:01] SQL will only get you meta-data (like who edited which page when) [19:25:21] i mean the same from here: https://query.wikidata.org/ [19:25:50] when i use tool labs, i connect to database copies of the different wikipedia language editions [19:26:08] sometimes using the sql in the command line, sometimes creating python scripts [19:26:20] i'd like to know if i could do the same with wikidata [19:33:30] one option is to use a sparql client (https://pypi.python.org/pypi/sparql-client) or sparql endpoint (https://rdflib.github.io/sparqlwrapper/) [19:34:11] but i see there is this other client library: https://pypi.python.org/pypi/Wikidata/0.6.1 [19:36:16] DanielK_WMDE: I understand I will need to use SPARQL as a query language [19:37:22] but perhaps this python client is an option? [19:37:36] i mean python library* [19:38:28] marmick: sure you can use any of these libraries [19:38:47] i just need to read, not to update wikidata. that's why i said that accessing a copy of wikidata might be a good option (the same way i do with the wikipedias) [19:39:02] marmick: I suppose so. But it probably won't help with the difficult part: understanding the data model, so you can write the queries you need. [19:39:12] I recommend playing with the examples on the WDQS homepage [19:39:44] yup, i know, i need to learn the data model and the query lang. but now i'm concerned about the efficiency and how to integrate it with a python script that already queries to the wikipedia language editions [19:39:53] marmick: the query service is a copy of a copy of a copy anyway, no updates are possible via SPARQL. [19:39:56] efficiency, performance... [19:40:28] oh, if you want to integrate with a python bot, you should definitly use a sparql lib for python [19:40:40] not a bot [19:40:56] i'm analyzing wikipedia articles and it might be helpful to check parallel data from wikidata [19:41:11] but it seems the same case [19:47:21] thanks DanielK_WMDE! :) [19:58:14] SMalyshev, ping, how do you run blazegraph locally during the development? Docs mention tools/runBlazegraph, but that doesn't exist [19:58:38] i am able to use mvn install but that takes forever, and requires all tests to pass [20:37:59] Maybe a very silly question, but: what is the actual use of having items for Wikipedia disambiguation pages? [20:38:17] It seems like you would rarely want a translation for those, and otherwise they seem useless :/ [20:41:59] I would use Cognate nowadays, but sometimes they are like [X] (disambugation) [20:53:06] * reosarevok isn't sure what Cognate is exactly [20:53:21] Oh, https://www.mediawiki.org/wiki/Extension:Cognate ? [20:55:53] Dunno, I just get kinda confused about us having those items - they kinda don't fit in. Wikipedia categories and list items are sliiightly better, because at least translations make sense [20:57:40] yurik: hmm. that should be there... let me check [20:58:09] yurik: it's war/runBlazegraph.sh [20:59:03] yurik: where docs says it's in tools? Must be out of date [20:59:09] SMalyshev, yes, i saw it, but it doesn't run [20:59:19] yurik: and you can use mvn install -DskipTests=yes [20:59:23] that should be much faster [20:59:39] yurik: by "doesn't run" you mean.... ? [21:00:39] i mean that i have to do mvn install first, before it works :) [21:01:47] SMalyshev, and no, skip tests still runs them [21:02:57] SMalyshev, also, the bigger issue - war/runBlazegraph.sh -d pauses on start, but i cannot attach to it from IDEA. Any idea why? :) [21:03:17] I would like to be able to debug blazegraph when it starts [21:03:26] yurik: no IDEA ;) [21:03:29] or at least attach it during execution [21:03:38] yurik: it works for me with eclipse [21:03:45] wrong port or something maybe? [21:04:06] yurik: no, skip tests doesn't run tests. maybe wrong syntax or something? it definitely does not [21:07:28] SMalyshev, https://gist.github.com/nyurik/bd7ca1fbac508722d7c6441a76a30304 [21:07:43] this is running with my modifications [21:08:20] it seems like lots of unit tests, no? [21:10:09] the run takes a few minutes at least [21:12:33] hmm maybe skipTests=true? [21:12:40] I don't remember which one is it. [21:13:05] http://www.mkyong.com/maven/how-to-skip-maven-unit-test/ this says -DskipTests should work [21:13:30] or maven.test.skip