[09:12:10] would it be a good practice for a bot to ping the user who authored a changed if such a change needs to be fixed ? With a proper explanation that is, for instance: "You did XYZ and URL documents that it should be otherwise, could you please fix that ?" [09:12:19] s/changed/change/ [09:13:53] I have the feeling that bots are not expected to do that. They are expected to produce violation reports instead. Not sure :-) [09:38:38] dachary, editors are not necessarily the one who watch constraint reports, so I think it totally would be a good practice - as long as community approves the models [09:39:30] it coud also be very pedagogical as the use is likely to do the same mistake again [09:40:18] it could be built from a "bad smells" database [10:30:35] Is there a template genius online who knows how to fix the formatting of Bacdive ID in this template? https://en.wikipedia.org/wiki/Template_talk:Taxonbar#Bacdive [10:32:02] dachary: I wanted to ask you if we should ask the people of appdata.xml to include the Wikidata ID in the specification (Example here: https://github.com/chergert/gnome-builder/blob/master/data/org.gnome.Builder.appdata.xml.in) [10:37:24] dachary : isn't that what a watchlist is for ? [10:45:16] .+prenom Janka [13:23:33] dachary: TomT0m : Take for example https://www.wikidata.org/wiki/User_talk:Thayts#Deprecated_rank . That is easy to detect with a bot and to leave a note, but isn't it much nicer if a human leaves a note? [13:50:38] multichill, I'd say that depends if someone is willing to do this :) Plus it's always possible that the bots puts the name of someone if the contributor has any question [14:30:00] Alphos: you're right, very good point :-) [14:31:43] multichill: it is nicer indeed. I react very differently when a bot tells me something or when a human being does the same. [14:34:10] tobias47n9e-c: I don't know about appdata.xml [14:34:13] * dachary looking it up [14:38:07] tobias47n9e-c: after reading https://people.freedesktop.org/~hughsient/appdata/ I see what you mean. It would make sense: maintainers of the file could lookup wikidata to see if more information is available there. Or update wikidata with a script using appdata.xml whenever they change something. [14:39:40] https://www.wikidata.org/wiki/Wikidata_talk:WikiProject_Informatics/FLOSS#propose_a_wikidata_id_to_appdata.xml as a reminder [14:43:24] dachary: It would also have advantages for e.g. The Gnome software center. Through the Wikidata ID a program could fetch labels, categories and descriptions in more languages. [14:43:41] * dachary nods [14:44:04] dachary: I am in the #freedesktop channel on freenode, but I haven't gotten an answer today. [14:47:34] tobias47n9e-c: ack [16:05:08] https://www.wikidata.org/wiki/Special:Contributions/FLOSSbot [16:05:09] hm [16:05:17] Is that the right property [16:28:26] We're a few thousand away from #20,000,000! [21:25:50] any wikidata property creators want to process a unanimous proposal? :) https://www.wikidata.org/wiki/Wikidata:Property_proposal/OpenCitations_bibliographic_resource_ID [21:43:59] <1,000 away from the 20 millionth item https://www.wikidata.org/wiki/Special%3AStatistics [22:15:56] is there a way to look for items which have more than multiple statements (say P1324) and none of them is preferred ? [22:15:56] P1324 Masterwork From Distant Lands - https://phabricator.wikimedia.org/P1324 [22:25:15] dachary count, group by ? [22:26:16] Alphos: really ? [22:26:29] I realize my question is very poorly phrasesd, trying again [22:26:38] is there a way to look for items which have multiple statements (say P1324) and none of them is preferred ? [22:26:39] P1324 Masterwork From Distant Lands - https://phabricator.wikimedia.org/P1324 [22:28:37] dachary: http://tinyurl.com/zbu97am ? [22:28:44] dachary http://tinyurl.com/gvrfp7z [22:28:53] * dachary looking [22:29:05] * Alphos hi5s WikidataFacts [22:29:16] nice thought with the group_concat :) [22:29:44] I’m actually not sure I like it, it takes up a lot of space and doesn’t get any form of prettifying (e. g. turning it into links) [22:30:22] Alphos: I suppose your much simpler query works too, due to the way wdt: works depending on ranks :) [22:30:52] anybody approximately awake and with solid shell-fu around ? trying to see if i made some kind of mistakes, my shell times are long gone https://gist.github.com/alaefin/f0b1bd17d91a2563bd40fe60981ed35e sql is solid, not sure about the .sh [22:31:00] WikidataFacts : my thought exactly [22:31:07] Alphos: WikidataFacts clever, thanks :-) [22:31:11] dachary yw [22:32:23] Alphos: will your $( /usr/bin/mysql ... output values with tab/space ? [22:32:48] dachary newline afaik [22:32:49] I think not because dbname probably don't have space in them [22:33:11] each dbname on its line, and they're only [a-z_]+ [22:34:10] echo -e '\t' [22:34:13] if it works, we will have weekly redirects conflicts reports from now on, and visualization when i finally get to doing that bit :D [22:34:24] otherwise you'll have \t instead of a tab [22:34:33] dachary not in .sh [22:34:56] then it depends on what /bin/sh really is [22:34:56] -e '\t' in sh outputs the literal "-e" [22:35:04] i tried ;) [22:35:22] it did feel weird, but ¯\_(ツ)_/¯ [22:35:30] :-) [22:35:42] you don't need the trailing ; but that's cosmetic and won't hurt [22:36:05] /bin/echo -e 'a\tb' outputs an actual tab over here, but just `echo` in bash and dash doesn’t [22:36:06] it's dumb, but i NEVER remember for sure, or if it's needed inside loops [22:36:25] so inside loops, i always put trailing ; :D [22:36:37] better safe than sorry [22:37:00] :-) [22:37:01] for some reason, outside loops, i don't mind one-liners with ; between commands ^^° [22:37:21] otherwise looks good, is this script giving you trouble ? [22:37:41] not so far, i just want to make sure it works as intended before starting it tonight ^^' [22:38:56] that very ugly feeling, down your guts, that something is going to go VERY wrong at some point, because murphy's law [22:40:05] wish me luck, i'm doing it >_> [22:40:22] shellcheck has a few complaints (mostly quoting), but none that seem important (e. g. you know that $(date +%G-%V) won’t contain spaces) [22:40:31] good luck :D [22:40:34] ERROR 2005 (HY000): Unknown MySQL server host '.labsdb' (0) [22:40:41] murphy's law didn't even wait :D [22:41:01] $i empty in line 6? [22:41:16] worse than that : stupidity [22:41:25] didn't think of updating the live script XD [22:41:30] haha :D [22:43:13] welp, so far, it seems like it's doing its thing :) https://tools.wmflabs.org/wikidata-redirects-conflicts-reports/reports/2016-35/ [22:43:44] (pay no attention to -2016-35.tsv, forgot to remove that) [22:44:26] come to think of it, maybe i could replace all these ; with && [22:45:33] I don't get why http://tinyurl.com/jm8b8ta shows OpenVZ still. I set one of the 1324 as preferred. [22:45:43] * dachary scratching his head [22:45:43] (not to mention add /usr/bin/time before the sql call so i can add that to the statistics too) [22:46:25] dachary: I don’t see it over here… did you get cached results perhaps? [22:46:36] dachary : there are two big issues in computer science : the naming of stuff, cache invalidation, and off-by-one errors [22:46:48] dachary : your issue is [1] in the list above ;) [22:46:48] 10[1] 04https://www.wikidata.org/wiki/Template:dubious [22:47:12] cached... yes [22:47:24] sorry for the noise :-) [22:47:24] AsimovBot i know it's dubious, it's a joke on "two" and "off-by-one errors" ! *slaps the bot* [22:47:35] dachary bah, don't worry ^^ [22:48:41] ok, so enwiki takes a little over 4 minutes to complete \o/ [22:49:09] and dewiki about 30 seconds :D [22:50:51] When we get to Wikidata item #20,000,000, will we know which one it is? [22:51:57] https://www.wikidata.org/w/index.php?title=Wikidata%3AWikiProject_Informatics%2FFLOSS&type=revision&diff=373011238&oldid=372839581 query archived, thanks again ! [22:58:03] 20,000,000 items! \o/ [22:58:39] Now, which one is the 20 millionth? [22:59:06] not sure… it went past fairly fast, I have four candidates from the recent changes [23:01:03] how would you choose the statement which has the preferred rank in a template ? I'm thinking https://en.wikipedia.org/wiki/Template:Infobox_software [23:02:18] like what's done for #property:P856 but for #property:P1324 and only using the preferred statement (by default it uses all of them, I think) [23:02:18] P856 Second run of convertNamespaceFromWikitext.php on Catalan Beta - https://phabricator.wikimedia.org/P856 [23:02:18] P1324 Masterwork From Distant Lands - https://phabricator.wikimedia.org/P1324 [23:02:25] or maybe not... [23:03:49] oh... https://en.wikipedia.org/w/index.php?title=Template%3AInfobox_software&type=revision&diff=737565982&oldid=737501452 did it already. That's fast. [23:04:02] process complete in 12 minutes and 53 seconds :) https://tools.wmflabs.org/wikidata-redirects-conflicts-reports/reports/2016-35/ [23:07:36] * dachary reading https://en.wikipedia.org/wiki/Template_talk:Infobox_software#Repository_URL [23:07:48] Alphos: that's fast [23:08:53] dachary especially if you compare it to my last attempt (which gave the same results but with more comprehensive info), which took about 70 minutes for arwiki alone (biggest wiki attempted back then) [23:25:40] https://github.com/alaefin/wikidata-redirects-conflicts-reports/commit/69e968543cf3b571c15f0afd4cc8ec6b0a4c3712 and it's committed !