[01:18:03] (03CR) 10Catrope: [C: 04-1] Make filters thresholds more configurable (034 comments) [extensions/ORES] - 10https://gerrit.wikimedia.org/r/348496 (https://phabricator.wikimedia.org/T162760) (owner: 10Sbisson) [10:51:35] (03PS3) 10Sbisson: Make filters thresholds more configurable [extensions/ORES] - 10https://gerrit.wikimedia.org/r/348496 (https://phabricator.wikimedia.org/T162760) [10:51:46] (03CR) 10Sbisson: Make filters thresholds more configurable (034 comments) [extensions/ORES] - 10https://gerrit.wikimedia.org/r/348496 (https://phabricator.wikimedia.org/T162760) (owner: 10Sbisson) [13:26:57] o/ [14:37:23] I'm training the same model over and over again to test out the refactor I've been working on. This feels so silly :) [14:37:47] Also, I'm so tired of waiting [14:38:30] I'm learning some interesting stuff from the thresholds set. It turns out that most of the interesting thresholds are very small proportions. [14:39:33] I wonder if I should resort to just using linspace. [14:39:35] hmm [14:51:42] It looks like using more observations makes a big difference. [14:51:43] :) [15:02:48] 06Revision-Scoring-As-A-Service, 10Collaboration-Community-Engagement, 06Community-Liaisons, 10Edit-Review-Improvements-RC-Page, 06Collaboration-Team-Triage (Collab-Team-Q4-Apr-Jun-2017): Communicate new beta prefs and changes to ORES users specifically - https://phabricator.wikimedia.org/T163153#3193884 (... [15:08:14] 10Revision-Scoring-As-A-Service-Backlog, 07Easy, 07artificial-intelligence: Text complexity scoring - https://phabricator.wikimedia.org/T155843#3193950 (10Basvb) The textstat package looks like a good idea for English. For other languages this might be a bit more difficult to use. Maybe using overall word fr... [15:23:28] 10Revision-Scoring-As-A-Service-Backlog, 10MediaWiki-Watchlist, 10ORES, 05codfw-rollout: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3194027 (10Anomie) It looks like this may have something to do with ORES: when I enable the beta feature I see the reported problem... [15:24:49] 10Revision-Scoring-As-A-Service-Backlog, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, 10ORES, 05codfw-rollout: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3194034 (10Anomie) [15:26:20] 10Revision-Scoring-As-A-Service-Backlog, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, 10ORES, and 2 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3194042 (10Ladsgroup) a:03Ladsgroup [15:27:02] 10Revision-Scoring-As-A-Service-Backlog, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, 10ORES, and 2 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3194045 (10Ladsgroup) I'm afk at the moment. Will look into it as soon as I get a pc [16:35:39] 10Revision-Scoring-As-A-Service-Backlog, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, 10ORES, and 2 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3194352 (10Ladsgroup) It happened before but on eqiad and got resolved {T144233} So my first hypothesis i... [16:39:11] 10Revision-Scoring-As-A-Service-Backlog, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, 10ORES, and 2 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3194362 (10Anomie) Seems likely to be that somehow duplicate entries got added to the database: ``` mysql... [16:49:43] Amir1, when you have a minute, I want to talk about what happened with those dupes. [16:49:57] I'm guessing it was an old thing with replication to CODFW [16:50:07] halfak: hey, sure [16:50:15] I'm all ears [16:50:56] Oh. I just wanted to know what happened and how it happened. [16:51:06] I've been looking for the old task for the original bug. [16:52:21] I made a link in the phab card [16:52:25] you can find it easily [16:52:37] Great [16:53:36] 06Revision-Scoring-As-A-Service, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, 10ORES, and 2 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3194435 (10Halfak) [17:01:51] 06Revision-Scoring-As-A-Service, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, 10ORES, and 2 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3194452 (10Ladsgroup) I need to example the situation in four examples. Note db1089 is in eqiad, db2048 is in cod... [17:10:44] 06Revision-Scoring-As-A-Service, 10MediaWiki-JobQueue, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, and 3 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3194458 (10Krinkle) [17:12:57] 06Revision-Scoring-As-A-Service, 10MediaWiki-JobQueue, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, and 3 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3194462 (10Ladsgroup) Yes, Jobrunner request scores twice. Compare [[https://logstash.wikimedia.org... [17:18:14] halfak: Did you see my comments in the phab card? [17:19:33] Yikes. Just dragged it from Done to Active [17:19:54] 06Revision-Scoring-As-A-Service, 10MediaWiki-JobQueue, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, and 3 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3194508 (10Anomie) I don't think it's edits made on codfw, since e.g. [[https://en.wikipedia.org/w/... [17:21:54] 06Revision-Scoring-As-A-Service, 10MediaWiki-JobQueue, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, and 3 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3194515 (10Ladsgroup) To increase the strangeness of the issue. This doesn't happen in Persian Wiki... [17:24:04] 06Revision-Scoring-As-A-Service, 10MediaWiki-JobQueue, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, and 3 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3194532 (10Ladsgroup) (edit conflict) >>! In T163337#3194508, @Anomie wrote: > So it looks to me li... [17:26:36] 06Revision-Scoring-As-A-Service, 10MediaWiki-JobQueue, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, and 3 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3193929 (10Joe) @Ladsgroup is this happening for new jobs enqueued now? [17:27:34] 06Revision-Scoring-As-A-Service, 10MediaWiki-JobQueue, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, and 3 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3194542 (10Halfak) p:05Triage>03Unbreak! [17:29:16] 06Revision-Scoring-As-A-Service, 10MediaWiki-JobQueue, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, and 3 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3193929 (10Krinkle) >>! In T163337#3194462, @Ladsgroup wrote: > Yes, Jobrunner request scores twice... [17:30:35] 06Revision-Scoring-As-A-Service, 06Collaboration-Team-Triage, 10Edit-Review-Improvements: ORES highlights are completely disabled even when ERI is disabled - https://phabricator.wikimedia.org/T163025#3194562 (10jmatazzoni) Hi Amir. The idea is that as we roll out the New Filters interface to more pages (Watc... [17:36:56] 06Revision-Scoring-As-A-Service, 10MediaWiki-JobQueue, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, and 3 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3194661 (10Anomie) >>! In T163337#3194539, @Joe wrote: > @Ladsgroup is this happening for new jobs... [17:40:03] 06Revision-Scoring-As-A-Service, 10MediaWiki-JobQueue, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, and 4 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3194671 (10Paladox) [17:42:36] 06Revision-Scoring-As-A-Service, 10MediaWiki-JobQueue, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, and 4 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3194686 (10Krinkle) >>! In T163337#3194508, @Anomie wrote: > The earliest recent enwiki revision I... [17:45:39] 06Revision-Scoring-As-A-Service, 10MediaWiki-JobQueue, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, and 4 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3194696 (10Ladsgroup) A note on other wikis. All were zero except: nlwiki (486), plwiki (156) and w... [17:52:05] 06Revision-Scoring-As-A-Service, 10MediaWiki-JobQueue, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, and 4 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3194717 (10Joe) On possible explanation is this is due to jobs that already were running in eqiad b... [18:27:03] (03CR) 10Catrope: [C: 032] Make filters thresholds more configurable [extensions/ORES] - 10https://gerrit.wikimedia.org/r/348496 (https://phabricator.wikimedia.org/T162760) (owner: 10Sbisson) [18:28:43] (03Merged) 10jenkins-bot: Make filters thresholds more configurable [extensions/ORES] - 10https://gerrit.wikimedia.org/r/348496 (https://phabricator.wikimedia.org/T162760) (owner: 10Sbisson) [18:39:04] 06Revision-Scoring-As-A-Service, 10Edit-Review-Improvements-RC-Page, 10ORES, 06Collaboration-Team-Triage (Collab-Team-Q4-Apr-Jun-2017), and 2 others: Tweak ORES-Related Preferences for Watchlist and RC Page ahead of next release - https://phabricator.wikimedia.org/T162831#3195041 (10jmatazzoni) [18:57:30] 06Revision-Scoring-As-A-Service, 10Edit-Review-Improvements-RC-Page, 10ORES, 06Collaboration-Team-Triage (Collab-Team-Q4-Apr-Jun-2017), 13Patch-For-Review: Conform ORES sensitivity levels to the new ERI standards - https://phabricator.wikimedia.org/T160575#3195117 (10jmatazzoni) [18:57:33] 06Revision-Scoring-As-A-Service, 10Edit-Review-Improvements-RC-Page, 10ORES, 06Collaboration-Team-Triage (Collab-Team-Q4-Apr-Jun-2017), and 2 others: Tweak ORES-Related Preferences for Watchlist and RC Page ahead of next release - https://phabricator.wikimedia.org/T162831#3195115 (10jmatazzoni) 05Open>0... [18:58:15] https://movielens.org/ [19:00:22] 06Revision-Scoring-As-A-Service, 10Edit-Review-Improvements-RC-Page, 10ORES, 06Collaboration-Team-Triage (Collab-Team-Q4-Apr-Jun-2017), and 2 others: Manage ORES preferences on Watchlist (and Contributions) - https://phabricator.wikimedia.org/T160475#3195142 (10jmatazzoni) [19:03:47] halfak: I'm currently deleting old rows in ores_classification [19:04:09] Great. Do you think this'll be the last pass or are those old jobs still executing? [19:04:46] I'm guessing there will be 90M rows freed [19:05:10] halfak: We should wait and see for sure, but new ones are done for now [19:05:20] we shouldn't see any duplicates [19:05:31] the root cause was completely different thing [19:06:07] Amir1, gotcha. Thanks for monitoring and taking care of this. [19:35:04] 06Revision-Scoring-As-A-Service, 10Research Ideas, 07artificial-intelligence: [Epic] Article importance prediction model - https://phabricator.wikimedia.org/T155541#3195271 (10Halfak) Met with @Milimetric, @nettrom, and @JAllemandou today. Here's our notes: https://etherpad.wikimedia.org/p/measuring_importa... [19:36:45] 06Revision-Scoring-As-A-Service, 10MediaWiki-JobQueue, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, and 4 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3195285 (10Ladsgroup) >That seems like a lot. Presumably they weren't all running parallelly on eqi... [19:56:52] I'm heading to bed o/ [20:01:51] o/ Amir1 [20:02:14] sorry I missed your signoff. I hope I don't pull you back D: [20:08:06] 06Revision-Scoring-As-A-Service, 10MediaWiki-JobQueue, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, and 4 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3193929 (10Catrope) The 90M rows of old data are known: {T159753}. I was going to clean them up, bu... [20:29:55] 06Revision-Scoring-As-A-Service, 10MediaWiki-JobQueue, 10MediaWiki-Watchlist, 10MediaWiki-extensions-ORES, and 4 others: Watchlist entries duplicated several times - https://phabricator.wikimedia.org/T163337#3195521 (10Catrope) >>! In T163337#3195406, @Catrope wrote: > The 90M rows of old data are known: {... [22:00:37] 10Revision-Scoring-As-A-Service-Backlog, 10OOjs-UI, 10ORES: On labels.wmflabs.org, make the blue buttons more visible when they have been selected - https://phabricator.wikimedia.org/T163222#3190410 (10Volker_E) I don't know the exact implementation, but before we're going to invent another unique button gro...