[00:03:43] (03CR) 10Catrope: [C: 032] Allow filter UI to be turned off [extensions/ORES] - 10https://gerrit.wikimedia.org/r/359364 (https://phabricator.wikimedia.org/T167908) (owner: 10Catrope) [00:05:04] (03CR) 10Catrope: [C: 032] "One minor thing we may or may not care about: the entry for the "r" in the RC legend doesn't go away when you set $wgOresUiEnabled = false" [extensions/ORES] - 10https://gerrit.wikimedia.org/r/359364 (https://phabricator.wikimedia.org/T167908) (owner: 10Catrope) [00:06:49] (03Merged) 10jenkins-bot: Allow filter UI to be turned off [extensions/ORES] - 10https://gerrit.wikimedia.org/r/359364 (https://phabricator.wikimedia.org/T167908) (owner: 10Catrope) [00:11:30] And with that I'm AFK [00:11:32] o/ [00:11:34] have a good one! [00:17:26] 10Scoring-platform-team, 10editquality-modeling, 10User-Ladsgroup, 10artificial-intelligence: Flagged revs approve model to fiwiki - https://phabricator.wikimedia.org/T166235#3480201 (10awight) Announced as https://meta.wikimedia.org/wiki/Research_talk:Automated_classification_of_edit_quality/Work_log/2017... [00:36:01] (03CR) 10Awight: "@Catrope: It doesn't seem like we can rely on having BetaFeatures loaded at extension.json callback time? If not, then fixing might cause" [extensions/ORES] - 10https://gerrit.wikimedia.org/r/359364 (https://phabricator.wikimedia.org/T167908) (owner: 10Catrope) [00:37:49] halfak: When you’re ready for some light reading, https://meta.wikimedia.org/wiki/Research_talk:Automated_classification_of_edit_quality/Work_log/2017-07-26 [07:49:24] 10Scoring-platform-team-Backlog, 10ORES, 10Operations: Reimage ores* hosts with Debian Stretch - https://phabricator.wikimedia.org/T171851#3480582 (10MoritzMuehlenhoff) p:05Triage>03Normal [13:08:55] 10Scoring-platform-team-Backlog, 10MediaWiki-Recent-changes, 10MediaWiki-extensions-ORES, 10Wikidata: ORES and Enhanced Recent Changes: Backgound color stucks after patrolling - https://phabricator.wikimedia.org/T171953#3481172 (10Raymond) [15:10:53] 10Scoring-platform-team, 10ORES, 10Operations, 10Patch-For-Review, 10User-Joe: Stress/capacity test new ores* cluster - https://phabricator.wikimedia.org/T169246#3481458 (10Joe) [15:25:38] 10Scoring-platform-team, 10ORES, 10Operations, 10Patch-For-Review, 10User-Joe: Stress/capacity test new ores* cluster - https://phabricator.wikimedia.org/T169246#3481494 (10Halfak) I've completed a few tests. TL;DR: we need to up our celery worker count before we'll get an accurate reflection of the cap... [18:00:52] (03PS1) 10Sbisson: Messages for low and high accuracy likelybad filters [extensions/ORES] - 10https://gerrit.wikimedia.org/r/368463 (https://phabricator.wikimedia.org/T164997) [18:02:22] (03CR) 10jerkins-bot: [V: 04-1] Messages for low and high accuracy likelybad filters [extensions/ORES] - 10https://gerrit.wikimedia.org/r/368463 (https://phabricator.wikimedia.org/T164997) (owner: 10Sbisson) [18:06:29] (03PS2) 10Sbisson: Messages for low and high accuracy likelybad filters [extensions/ORES] - 10https://gerrit.wikimedia.org/r/368463 (https://phabricator.wikimedia.org/T164997) [18:08:30] (03Abandoned) 10Catrope: Add job to purge old scores (older than $wgRCMaxAge) from ores_classification [extensions/ORES] - 10https://gerrit.wikimedia.org/r/341946 (https://phabricator.wikimedia.org/T159753) (owner: 10Catrope) [18:22:53] (03CR) 10Catrope: [C: 04-1] "It would be good to explain in the qqq that on any given wiki, either the low or the high message will be used depending on whether the ma" [extensions/ORES] - 10https://gerrit.wikimedia.org/r/368463 (https://phabricator.wikimedia.org/T164997) (owner: 10Sbisson) [18:23:24] 10Scoring-platform-team, 10ORES, 10Operations, 10Patch-For-Review, 10User-Joe: Stress/capacity test new ores* cluster - https://phabricator.wikimedia.org/T169246#3482118 (10Halfak) I scheduled some time with @joe for running another test with more workers on Monday at 1300 UTC. [18:29:15] (03PS3) 10Sbisson: Messages for low and high accuracy likelybad filters [extensions/ORES] - 10https://gerrit.wikimedia.org/r/368463 (https://phabricator.wikimedia.org/T164997) [18:36:36] o/ fajne [18:36:39] How's it going? [18:37:15] the yes-no thing? not very well [18:37:27] Oh? Maybe I can help. [18:39:24] may be there is a better fit person for js and html tasks out there? this little thing can take me till the end of my internship :) [18:40:40] ha. It's a simple change that will likely be a bit more complicated to test. Maybe I can help or would you really rather look at some modeling work instead? [18:40:52] We could work together to get a new language implemented. [18:41:01] I think we have some language assets that are waiting to be used. [18:42:08] Oh! We could also experiment with training the draftquality model on a balanced training set. [18:42:34] https://phabricator.wikimedia.org/T168909 [18:42:40] let me try it till the end of the day (with your help), and when i give up, i'll get back to pythonland [18:42:46] OK cool :) [18:44:27] i changed the words in en.json file.. i need to change the css file.. [18:51:44] Are the new terms rendering too wide in the buttons? [18:53:28] 10Scoring-platform-team, 10editquality-modeling, 10User-Ladsgroup, 10artificial-intelligence: Flagged revs approve model to fiwiki - https://phabricator.wikimedia.org/T166235#3482194 (10Halfak) Nice work on this. I'm starting to think that we're going to want some tuning to get this working better. @Zach... [19:02:37] halfak: they sure will be [19:02:51] gotcha [19:03:29] and i cannot get where is the translation hiding [19:12:42] ? [19:12:51] fajne, which translation are you looking for? [19:13:10] of the yes-no [19:13:35] https://github.com/wiki-ai/wikilabels-wmflabs-deploy/blob/master/forms/i18n/damaging_and_goodfaith/en.json [19:13:48] I think I link to that in the task :) [19:15:28] aha, i found this json. let me look at the task again [19:21:54] and how do i test it ? [19:22:44] If you pull down the wikilabels-wmflabs-deploy repo and modify that file, then you can run the wikilabels server from that directory. [19:29:22] 10Scoring-platform-team, 10editquality-modeling, 10User-Ladsgroup, 10artificial-intelligence: Flagged revs approve model to fiwiki - https://phabricator.wikimedia.org/T166235#3482308 (10awight) If we do this again, noting one minor thing I messed up: We should have thrown out multi-revision approvals, but... [19:36:40] 10Scoring-platform-team, 10editquality-modeling, 10User-Ladsgroup, 10artificial-intelligence: Flagged revs approve model to fiwiki - https://phabricator.wikimedia.org/T166235#3482326 (10Zache) >>! In T166235#3482194, @Halfak wrote: > Nice work on this. I'm starting to think that we're going to want some t... [20:32:55] o/ [20:40:45] halfak: pushing an update now that adds a more descriptive user agent to the dashboard ores requests. It'll now have server url plus environment, so you'll be able to sort outreachdashboard.wmflabs.org production from dashboard.wikiedu.org prodction from dev and test environments. [20:57:43] ragesoss, sounds good. Thank you :) [20:58:10] np :) [22:29:25] 10Scoring-platform-team-Backlog, 10ORES: [Spike] [bike shed] Name the Meta ORES service - https://phabricator.wikimedia.org/T153146#3482968 (10Halfak) Thinking about measuring the purity of a chemical. There are lots of types of [spectrometry](https://en.wikipedia.org/wiki/Spectrometry), e.g. [mass spectromet... [22:38:08] (03PS4) 10Catrope: Messages for low and high accuracy likelybad filters [extensions/ORES] - 10https://gerrit.wikimedia.org/r/368463 (https://phabricator.wikimedia.org/T164997) (owner: 10Sbisson) [22:38:13] (03CR) 10Catrope: [C: 032] Messages for low and high accuracy likelybad filters [extensions/ORES] - 10https://gerrit.wikimedia.org/r/368463 (https://phabricator.wikimedia.org/T164997) (owner: 10Sbisson) [22:45:03] (03Merged) 10jenkins-bot: Messages for low and high accuracy likelybad filters [extensions/ORES] - 10https://gerrit.wikimedia.org/r/368463 (https://phabricator.wikimedia.org/T164997) (owner: 10Sbisson) [22:58:39] halfak: till what hour you are around? [22:58:46] Just about to step away :/ [22:59:30] hm, ok, see you on monday then. can i shoot you an email on this weekend? [23:00:45] Sure! I'll be around tomorrow morning until 10AM PDT