[08:39:27] All lady's please stand up [08:40:33] Can we show pics [09:59:04] hoo, I think handling claims/other should be better handled in lower level on SqlUsageTracker rather then on the sccribuntu side [10:00:19] eranroz: hm… how would it know when to filter the X/O usage? [10:00:51] based only on disabledUsageAspects [10:01:32] So you would just disable X and O in these cases? [10:02:53] it should corase grain the usage e.g if C in disabledUsageAspects => replace C to O [10:08:14] So that would make disabledUsageAspects a map [ 'disableAspect' => 'replacement' ]? [10:12:16] eranroz: ^ [10:13:09] yes [10:14:10] eranroz: Sounds good to me… do you want to open a bug for that? Or just create a patch? [10:14:46] hoo: I'll just create a bug for it [10:15:01] Thanks :) [10:15:09] just want to avoid the configuration prune to errors with incompatible configurations between disabledUsageAspects and fineGrainedLuaTracking [10:16:31] Yeah, that makes sense to me :) [11:15:15] WDQS can be used for soo much more... I just built a search & replace system for OSM. The query produces a set of edits, and users review them one by one, and click "save" under their own account (oauth). http://tinyurl.com/yava4nq7 [11:16:17] the same approach can be used for Wikidata - curated reviewing and saving of the changes [11:17:23] cc: SMalyshev & Jonas_WMDE [11:21:45] whoa [11:23:00] yurik: I’ve used WDQS to output a QuickStatements command string a few times (nested GROUP_CONCATs with separator="\t" inside and separator="\n" outside) [11:23:36] Lucas_WMDE, this is a bit more involved - it edits OSM directly :) [11:23:51] yurik: yeah, I just thought I’d mention it [11:23:58] :) [11:24:28] hehe, i have used mediawiki markup to do some crazy text manipulations, but i haven't done much in sparql for that yet :) [11:26:01] Lucas_WMDE, https://wiki.openstreetmap.org/wiki/Wikidata%2BOSM_SPARQL_query_service#Using_Editor_Capabilities - docs [15:25:03] Lydia_WMDE: https://gerrit.wikimedia.org/r/379669 This is not even tracked, but C. Scott said in the Scrum of Scrums he feels blocked by us. What to do? How can we track this? [15:25:38] Thiemo_WMDE: we can take it into the next sprint [15:26:27] he is blocked indeed but it is not mega urgent to do before the next sprint [15:26:59] Lydia_WMDE: I added it to the "Parking lot". [15:27:06] perfect [17:25:43] Hi there [17:26:47] I requested a flood flag but that my first time, I hope it's the right way : https://www.wikidata.org/wiki/Wikidata:Bureaucrats%27_noticeboard [18:38:18] Hey folks. I'm looking to split bot edits from "human" edits in an analysis. It looks like a lot of bot accounts are unflagged and maybe even a lot of bots edit via IP. Is this against policy or a normal thing on Wikidata? [18:38:31] From https://www.wikidata.org/wiki/Wikidata:Bots#Approval_process, it looks like it's against policy but maybe there's wiggle room? [18:41:51] hi halfak [18:42:25] hard question, what is "human" ? a lot of human edits are semi-automated (with QuickStatement or Petscan) [18:42:55] VIGNERON, right! I'm splitting those out too. In this case, I'm looking for fully-automated bots. [18:43:01] and some bot account use the exact same tools (only the scale differs, and not always) [18:43:07] the frontier can be quite fuzzy... [18:44:00] maybe you can make a distinction with the frequency (more/less than X edits by minutes) [18:47:49] Thank VIGNERON. Bot detection is a bit of a bigger project. I'm just trying to get a sense now for how strictly the bot policy is enforced. [18:52:35] I'm worried there are a lot of bots running as IP editors [18:52:40] Based on what we're seeing in the data. [19:00:26] Oh ! interresting [19:02:12] but for the policy, I'm not sure (in fact, right now, I've request a flood flag for a semi-automated import - with my human account - and I'm note exactly sure of the procedure and customs) [19:58:33] *sigh* https://www.wikidata.org/w/index.php?title=Q831366&action=history