[10:13:27] Hello. Any new news on Wikibase federation? Inability to use Wikidata's data on non-WMF wikis really is obviously a disincentive to add to Wikidata [10:30:16] Doesn't seem to be stopping people too much ;) [11:26:58] jubo2: "Really Soon Now" (tm) [11:27:28] depending on what precisely you want the federation to do. [11:33:42] DanielK_WMDE: I'd like to use Wikidata's data on a non-WMF Wikibase Client together with a non-WMF Wikibase Repository (used to store information that falls out of the scope of Wikidata) [11:34:05] So to fetch information from multiple repositories [11:34:44] DanielK_WMDE: Got any link on the federation work/progress/aims ? [11:34:51] jubo2: so you'd want to have local Q-items along with wikidat Q-items, and perhaps also local Properties alongside wikidata Properties [11:35:06] Yes [11:35:28] This capability is built into wikibase already "in principle", with two things to be figured out: caching/puriging, and UI. [11:35:37] Which are, unfortunately, not trivial things [11:35:53] yes I understand they are non-trivial issues [11:37:05] jubo2: tehre is a stub ticket here https://phabricator.wikimedia.org/T196997 which links to a dicument that provdes an overview of features sets and use cases. [11:37:24] I'm not aware of any concrete planning. I know that the Wikidata team is looking ito doing this "soon". [11:38:31] jubo2: for context - I designed the internal federation mechanism, but I'm no longer part of the wikidata team [11:39:34] :( [14:04:08] Technical Advice IRC meeting starting in 60 minutes in channel #wikimedia-tech, hosts: @chiborg & @amir1 - all questions welcome, more infos: https://www.mediawiki.org/wiki/Technical_Advice_IRC_Meeting [14:09:34] gehel: I am trying to do the sparql install, and I get https://pastebin.com/raw/DUpjnneY [14:09:47] Is it possible there is a problem in the dump ? [16:07:45] !admin I need to protect a page [16:07:45] Attention requested HakanIST sjoerddebruin revi [16:07:55] link please [16:08:34] Ineed: please provide a link and a reason [16:08:41] https://www.wikidata.org/wiki/Q145 [16:08:48] And https://www.wikidata.org/wiki/Q146 [16:09:25] And https://www.wikidata.org/wiki/Q30 [16:09:47] for what reason? [16:10:14] Usa is a country along with Uzk [16:10:16] I don't see an obvious vandalism or such [16:10:16] UK [16:10:30] And Cat is an ani al [16:10:33] Animal [16:10:47] I can't see any immediate vandalism, and that is required to protect a page [16:11:08] Please have a look at [[WD:PPP]] [16:11:08] 10[1] 1010https://www.wikidata.org/wiki/Wikidata:Page_protection_policy - Redirección desde 10https://www.wikidata.org/wiki/WD:PPP?redirect=no [16:11:28] > Admins will protect pages for short amounts of time to prevent vandalism or spam which repeatedly occurs on them. [16:12:08] Wait [16:12:33] So create protection protects a page that hasn't even been made yet [16:12:41] It can [16:12:46] join [16:12:57] You've already joined the #wikidata, jdd [16:13:23] Does create protection [16:13:36] protect create [16:13:46] Protect pages that haven't been made [16:13:59] It protects the page from being created by unauthorized people [16:14:10] (i.e. without accounts) [16:14:15] Therefore it hasn't been made [16:15:07] Ineed: why do you always connect on this channel using different name? [16:15:08] So anyway [16:15:17] I don't know [16:15:25] I want those three pages [16:15:30] nope [16:15:35] no protection will be made to those pages [16:15:39] Swmi [16:15:43] there is no reason to protect it. [16:15:44] Ok [16:15:54] But why does dog [16:15:58] it seems your block ended [16:17:53] My block? [16:17:55] Oh yes [16:17:59] 3 days ago [16:20:28] McDonald's was protected a few weeks ago [16:20:32] Now it is not [16:21:12] Please protect it again [16:21:13] Ok [16:21:22] https://www.wikidata.org/w/index.php?title=Q38076&direction=prev&oldid=687499864 so this doesn't happen [16:23:22] Admin [16:23:25] Please [16:23:30] Protect McDonald s [16:24:22] No ongoing vandalism, no protection. [16:24:34] Ok [16:27:53] I merged a paged [16:27:55] Page [16:28:41] Stryn [16:28:44] Please [16:28:54] Copyright law © [16:29:07] Why did you say thay [16:29:08] That [16:29:11] hi [16:29:21] Stryn [16:29:31] Ineed: you are going to be blocked on Commons if you continue uploading files that are not your [16:29:33] is there a software that helps with mapping things to wikidata id? [16:30:14] I uploaded 3 others that aren't mine [16:30:24] A minpin thing [16:30:30] And music [16:30:47] mark them for deletion then [16:31:24] I have a database of actors and films. I have already mapped all the films, but not the actors [16:32:14] Ok [16:32:21] Show me what is [16:32:33] What it is frickel [16:34:11] frickel what does it look like [16:34:24] it looks like a wiki [16:34:30] !ops [16:34:59] so its not really a database i guess 😅 [16:35:11] Frickel [16:35:38] !admin block me on wikimedia commons [16:35:38] Attention requested HakanIST sjoerddebruin revi [16:35:47] !ops remove Ineed [16:35:54] I [16:36:04] AlexZ: ^ [16:36:09] Have been breaking the copyright restrictions [16:36:16] You are wasting everyone's time here [16:36:20] please leave. [16:36:34] Block my commons account [16:36:39] Please [16:36:50] Wikidata and Wikimedia Commons is a different wiki, with different admins. [16:36:56] Now leave, please. [16:37:13] Ok [16:38:42] thanks [16:38:52] Np [16:38:57] Actually I am Commons admin [16:38:59] but who cares [17:15:25] frickel: OpenRefine is a software that helps map things to Wikidata ids [17:16:21] there are quite a lot of tutorials about that, see for instance https://www.wikidata.org/wiki/Wikidata:Tools/OpenRefine/Editing/Tutorials/Video [18:13:13] pintoch: thanks! [18:49:35] pintoch: You around? [18:49:46] yes [18:49:54] https://www.wikidata.org/w/index.php?title=Property%3AP143&type=revision&diff=706174496&oldid=705730012 [18:50:00] Looked a bit at the imported from [18:50:54] nice! now I hope the server will not blow up in fire when KrBot updates the violation report :-D [18:51:09] The violation report was already broken [18:51:39] ah, that solves the issue I guess ^^ [18:52:02] I think at least because it has always been empty [18:52:05] https://www.wikidata.org/w/index.php?title=Wikidata:Database_reports/Constraint_violations/P143&oldid=662002573 [18:52:08] And no statistics [18:52:17] I think it only looks at the main usage, not references usage [18:52:37] So the main source of problems seems to be lack of documentation [18:53:00] As far as I can see the scope has always been only stuff imported from Wikimedia projects, but people just started using it for other things too [18:54:15] So it's probably a matter of properly documenting the scope and checking if we don't have huge disagrements over that [18:55:49] pintoch: We could even setup an abusefilter rule that imported from can only be added if the same user did the previous edit [18:55:58] That will prevent a lot of people from adding it later [18:58:13] Nazral: an error is always possible! SMalyshev probably knows more about that than I do [19:00:42] it would be nice if there was a way to publish a quickedit batch for others to review before starting it [19:03:11] frickel: Ask Magnus [19:04:12] i could just paste it into a github gist [19:10:33] multichill: hmm why? [19:18:30] pintoch: Why what? [19:19:12] The abusefilter? Currently seems one of the few ways to tell a user something is wrong before doing the edit [19:19:38] Interface probably messes it up? [19:19:52] but why would we refuse "imported from" if the last edit was not made by the same user? [19:20:16] Because the user didn't import that data [19:20:53] You add some statement, second edit is to add the reference with imported from. It's always done by some automated tool. [19:21:07] Should never be added later than that [19:21:44] yeah but say you are cleaning up some item and moving things around, you might encounter a situation where you want to do that for some reason [19:22:07] So in the 0.0000001% edge case you would get a warning [19:22:08] an abusefilter sounds like a pretty rigid constraint no? [19:22:21] You can always save it anyway [19:22:27] You just get a warning the first time you try [19:22:29] oh right, yes just a warning [19:22:34] yes why not [19:22:52] But i'm not sure how good/bad the interface currently is [19:23:10] but then if you just create a claim with a reference in one edit, this is not going to get caught, right? [19:23:11] I think the user just gets some nasty API warning [19:23:24] yes it's not extremely user friendly [19:23:44] I'd rather put my money on better documentation [19:29:10] does anyone here have expperience with open refine? [19:29:39] pintoch: ^ [19:30:23] pintoch: I conntected 600 names with corresponding wikidata entities. I'd like to check if all these people have the statement occupation -> voice actor [19:30:24] multichill: agreed - and maybe we could propose a label change? (something with "Wikimedia" in it?) [19:30:54] imported from Wikimedia project? [19:31:08] That would make it a lot clearer and in line with the other property [19:31:09] frickel: you can do something like this https://www.wikidata.org/wiki/Wikidata:Tools/OpenRefine#/media/File:GIF_screencast_of_data_extension_on_OpenRefine_with_Wikidata.gif [19:31:21] multichill: that would be ideal I think [19:32:33] frickel: once you have fetched the "occupation" statement from your column with people, you can create a text facet on this new column [19:33:41] but i can only do this when I'm done right? [19:34:22] you can do this at any point - but it will only work for the cells that have been matched so far [19:34:50] if you want to redo it later on it is always possible to fetch the column again [19:39:34] also, when creating a facet on the new column, you might want to switch to "records mode" which will ensure that your filter will apply consistently to people with multiple occupations [19:39:52] cool [19:40:19] another thing. can I create a new wikidata entity for every voice actor in my database? [19:41:10] 99% of them will later be referenced to wikidata items that have a wikipedia article [19:42:38] if these actors have a wikipedia article, they *should* already have items [19:43:12] no, thats not what I mean [19:43:38] I mean they will be connected with films that have a wikidata entity [19:43:50] that have wikipedia articles [19:44:10] so i guess they are relevant [19:44:21] I am not certain that this qualifies as a structural need… it would be safer to ask on wiki [19:45:14] I suspect that many voice actors would not be notable [19:45:48] what constitutes of sometjing is notable? [19:46:13] the criteria are here: https://www.wikidata.org/wiki/Wikidata:Notability [19:48:32] _can be described using serious and publicly available references_ most of the voice actors have an entry in _Deutsche Synchronkartei_ (P4727) [19:48:57] The notability criteria are not particularly clear, although I guess that's true of every wikiproject :) [19:49:34] i could extract that information and add it [19:54:18] there is https://www.wikidata.org/wiki/Wikidata:Dataset_Imports to discuss these things [20:02:23] pintoch: https://www.wikidata.org/wiki/Property_talk:P143#Making_the_usage_of_this_property_clearer care to do the languages you understand? :-) [20:02:28] I also updated the usage instructions [20:04:40] multichill: thx, done [20:06:28] great and thanks reosarevok ;-) [20:06:39] np :) [20:06:57] Maybe add usage instructions in French and German too? [20:07:35] I would add them in Estonian but my Estonian is a bit too meh [20:07:40] (same for my Asturian) [20:07:43] So Spanish will have to do :) [20:10:17] It's used 14,917,344 times so no way to get an overview of biggest mistakes, right [20:10:18] ? [20:12:22] multichill: the property creation with the original name ;) [20:12:36] (if you mean biggest mistakes in use, maybe not :D ) [20:12:45] Oh god I'm turning into SothoTalKer. I'm sorry :D [20:13:48] what does that mean, now :p [20:19:36] SMalyshev: Do you know if https://query.wikidata.org/bigdata/ldf?subject=&predicate=http%3A%2F%2Fwww.wikidata.org%2Fprop%2Freference%2FP143&object= can be used in combination with SPARQL? [20:21:00] can I close openrefine at any time and contunue later? [20:21:29] continue [20:21:53] frickel: yes you can [20:22:15] thank you and good night ^^ [20:22:25] thanks, you too! :) [20:28:56] multichill: in combination no, probably not... but it's essentially the same as SPARQL triple. That's what it is underneath [20:30:57] Just wondering if it would work better in large sets [20:32:43] we could probably consider to run a bot to migrate "imported from" to "stated in" for non-Wikimedia projects? [20:34:25] I would be very careful with that [20:34:40] Some specific subsets it will work, some will just be junk [20:36:40] pintoch: http://tinyurl.com/ybkrjj3g playing around a bit [20:37:26] pintoch: Examples like https://www.wikidata.org/wiki/Q735944 could be done automatically [20:38:37] Pff, but no link to where the data came from.... [20:50:40] multichill: I still think it would help if we didn't have so many references using imported from in the first place, and if they weren't treated like proper references by the ui [20:52:06] Yeah. We haven't really paying enough attention [20:52:19] Take for example http://tinyurl.com/ybpmtv7m 47.000 viaf id's imported from viaf? [20:52:52] 'cause it's not surprising to me that people see "imported from some wikipedia" everywhere and then think they should add those references too [20:53:08] Of course. We both probably did the same at some point [20:54:18] but speaking of which, I'm currently working on a bot to add statements and references to a load of those cebwiki items, which also replaces imported from cebwiki if it finds a match [20:57:21] currently a bit stuck on matching coordinates... I've come to hate decimal representations of coordinates :P [20:58:56] I think we have some (LUA) code for that [20:59:10] Matching coordinates is generally a PITA [21:00:26] I'm using the gns data, which geonames imported but changed the number of decimal places, which cebwiki imported from geonames, which we imported from cebwiki... along the way it's picked up various changes so doesn't quite match the original >_< [21:01:26] I wonder if I need to do another request for permissions for it [21:02:03] nikki: https://www.wikidata.org/w/index.php?title=Q6018&type=revision&diff=318049851&oldid=306761593 <- DOH! [21:02:37] heh :) [21:02:47] at least they're fairly easy to find [21:03:34] nikki: oh that's very brave to tackle this problem - thanks, it's a real issue for me [21:04:03] the cebwiki items? [21:04:09] yeah [21:05:12] I'm sure there'll still be a lot of merging to do :( [21:06:28] Looks like running over viaf and updating references on identifiers seems to be a good start [21:06:55] my main aim right now is to have all the basic info like country and coordinates, so that I can stop adding them manually [21:09:33] also to reduce the insane number of constraint violations for the geonames property :P [21:16:14] seems like there are a lot for data.bnf.fr too [21:19:14] 320k [21:20:30] Fun stuff. [21:21:41] oh fun [21:21:44] https://www.wikidata.org/wiki/Q1913787#P40 [21:21:48] and stuff like that [21:23:01] Not sure if I shared http://tinyurl.com/ybjz5c5d [21:23:07] Full of these kind of weird things [21:24:24] hm. imported from mix'n'match [21:24:59] https://www.wikidata.org/w/index.php?title=Q825036&action=historysubmit&type=revision&diff=706257225&oldid=704750690 could be done by a bot [23:57:28] Hello