[01:10:20] 10Jade, 10Scoring-platform-team (Current), 10Patch-For-Review: Handle empty Jade pages - https://phabricator.wikimedia.org/T246033 (10ACraze) Just added a patchset that contains a shortcut for us to create an empty entity page on the fly by creating a page with an empty object: `{}` Here's a gif showing the... [01:14:49] 10Jade, 10Scoring-platform-team (Current), 10Patch-For-Review: Handle empty Jade pages - https://phabricator.wikimedia.org/T246033 (10ACraze) The above patchset doesn't get us 100% to our goal, but it's a step in the right direction. Currently looking at using parser hooks to auto-create pages within the `J... [12:39:23] 10MediaWiki-extensions-ORES, 10Scoring-platform-team, 10Discovery-Search, 10NewcomerTasks 1.1, and 3 others: Expose ORES drafttopic data in ElasticSearch via a custom CirrusSearch keyword - https://phabricator.wikimedia.org/T240559 (10Johan) I've added an item to https://meta.wikimedia.org/wiki/Tech/News/2... [12:40:15] 10MediaWiki-extensions-ORES, 10Scoring-platform-team, 10Discovery-Search, 10NewcomerTasks 1.1, and 3 others: Expose ORES drafttopic data in ElasticSearch via a custom CirrusSearch keyword - https://phabricator.wikimedia.org/T240559 (10kostajh) Looks good to me, thanks @Johan [14:20:55] Async Update 2020.03.05 [14:20:55] Reviewed Andy's patchset 577009 (Handle initial revision diff) 577013 (Handle empty EntityContent object) Gave Andy more usability feedback on the Jade diff view based on the latest changes he has made. Moved moveEndorsementDialog errorMessage to the top of dialog buttons. Made moveEndorsementDialog errorMessage span entire section width. No blockers for now. [14:20:56] Have a great day everyone 😊 [15:00:18] 10Jade, 10Scoring-platform-team (Current), 10MW-1.35-notes (1.35.0-wmf.23; 2020-03-10), 10Patch-For-Review: Handle empty Jade pages - https://phabricator.wikimedia.org/T246033 (10Jdforrester-WMF) [15:00:37] 10Jade, 10Scoring-platform-team (Current), 10MW-1.35-notes (1.35.0-wmf.23; 2020-03-10): Implement Diff view for Jade Entity UI - https://phabricator.wikimedia.org/T245316 (10Jdforrester-WMF) [15:00:45] 10Jade, 10Scoring-platform-team (Current), 10Design, 10MW-1.35-notes (1.35.0-wmf.23; 2020-03-10), 10Patch-For-Review: Implement CSS styles for Jade Entity UI - https://phabricator.wikimedia.org/T242648 (10Jdforrester-WMF) [15:18:28] wikimedia/articlequality#285 (enwiki_image_model - 69435af : Aaron Halfaker): The build passed. https://travis-ci.org/wikimedia/articlequality/builds/658733151 [15:25:32] 10Scoring-platform-team (Current), 10articlequality-modeling, 10artificial-intelligence: Improve ORES articlequality feature extraction for images - https://phabricator.wikimedia.org/T180822 (10Halfak) I rebuilt the models. See the result in this PR: https://github.com/wikimedia/articlequality/pull/104 It... [16:20:09] 10Scoring-platform-team, 10articlequality-modeling, 10artificial-intelligence: Add features for English Language idioms to articlequality models - https://phabricator.wikimedia.org/T247000 (10Halfak) [16:20:22] 10Scoring-platform-team, 10articlequality-modeling, 10artificial-intelligence: Add features for English Language idioms to articlequality models - https://phabricator.wikimedia.org/T247000 (10Halfak) [16:20:24] 10Scoring-platform-team (Current), 10articlequality-modeling, 10editquality-modeling, 10revscoring, and 2 others: Add English Language idioms to revscoring - https://phabricator.wikimedia.org/T205545 (10Halfak) [16:25:06] 10Scoring-platform-team, 10articlequality-modeling, 10artificial-intelligence: Add features for English Language idioms to articlequality models - https://phabricator.wikimedia.org/T247000 (10Halfak) See https://github.com/wikimedia/articlequality/blob/master/articlequality/feature_lists/tests/test_enwiki.py... [17:06:39] 10Scoring-platform-team, 10Discovery-Search, 10Growth Design, 10Growth-Team (Current Sprint), 10MW-1.35-notes (1.35.0-wmf.22; 2020-03-03): Newcomer tasks: UX changes for ORES topics - https://phabricator.wikimedia.org/T244421 (10Etonkovidova) Checked in betalabs - all look identical to the screenshots @R... [17:15:49] Y: Merged changes to articlequality by Haksoat and rebuild the AQ model. I also researched compare and tried to dig into i18n looking for date string formatting and getting specific i18n error messages. I helped chtnnh set up his environment and start working on ptwiki's articlequality model. [17:15:49] T: Continued working with chtnnh on text complexity features. I will continue my work on Jade docs. I have a meeting with product folks re. topic modeling products so I'll have something to report on that. [17:16:01] My async ^ [17:43:50] o/ [17:43:55] quick async update on my end [17:44:09] Y: Fixed the diff widget and got fixes pushed out to beta -- [17:44:21] it now uses revid to build the diff & headers [17:44:32] added handling for initial revisions to show the '(No Difference)' message similar to Special:Diff [17:45:06] Also got a shortcut in place to create Jade pages on the fly, it's not 100% what we want yet, but it's a step in the right direction. [17:45:27] Here's a gif of the empty page creation flow: [17:45:39] https://phab.wmfusercontent.org/file/data/7t3ts6puugy3hsmw37hw/PHID-FILE-mwq7lhpunnqspycird23/empty-page-hack.gif [17:46:23] T: Kevin pointed out some potential issues with the header links that I'm going to look at. [17:47:40] also plan to work more on the empty page workflow using parser hooks [18:52:19] 10Scoring-platform-team (Current), 10articlequality-modeling, 10artificial-intelligence: Improve ORES articlequality feature extraction for images - https://phabricator.wikimedia.org/T180822 (10HAKSOAT) I think its valuable too. I hope the performance didn't drop either? [18:57:44] @halfa [18:57:44] 04Error: Command “halfa” not recognized. Please review and correct what you’ve written. [18:58:04] Hello halfak I've seen the comments [18:58:07] Will work on it [18:58:11] Thanks [19:08:00] Awesome. I hope they are helpful! [19:15:31] 10Scoring-platform-team (Current), 10articlequality-modeling, 10artificial-intelligence: Add features for English Language idioms to articlequality models - https://phabricator.wikimedia.org/T247000 (10Halfak) [19:16:36] 10Scoring-platform-team (Current), 10articlequality-modeling, 10artificial-intelligence: Improve ORES articlequality feature extraction for images - https://phabricator.wikimedia.org/T180822 (10Halfak) No drop. I think it's good to merge. [19:17:05] 10Scoring-platform-team (Current), 10articlequality-modeling, 10artificial-intelligence: Improve ORES articlequality feature extraction for images - https://phabricator.wikimedia.org/T180822 (10Halfak) @Ragesoss, any thoughts as the original filer? [19:34:01] 10Scoring-platform-team (Current), 10articlequality-modeling, 10artificial-intelligence: Improve ORES articlequality feature extraction for images - https://phabricator.wikimedia.org/T180822 (10Ragesoss) Lovely! [19:37:10] 10Scoring-platform-team, 10Discovery-Search, 10NewcomerTasks 1.1, 10Growth-Team (Current Sprint): Once the ORES articletopic - ElasticSearch pipeline is set up, update data about all articles - https://phabricator.wikimedia.org/T243357 (10EBernhardson) >>! In T243357#5933137, @EBernhardson wrote: > I've al... [19:45:32] 10Scoring-platform-team, 10Wikilabels, 10articlequality-modeling, 10artificial-intelligence: Build article quality model for ptwikipedia - https://phabricator.wikimedia.org/T246663 (10Chtnnh) a:03Chtnnh @Halfak @GoEThe I will be working on this task together with Halfak. Hope to finish it as soon as poss... [20:06:08] halfak I just opened a PR for the english idioms work [20:06:22] Travis is quite slow so I'll just wait [20:06:32] But here's a link: https://github.com/wikimedia/articlequality/pull/105 [20:06:54] Nice! [20:17:12] Tests passed [20:42:56] haksoat, I'm rebuilding. Fingers crossed for a significant improvement. [20:49:29] 10Scoring-platform-team (Current), 10articlequality-modeling, 10artificial-intelligence: Add features for English Language idioms to articlequality models - https://phabricator.wikimedia.org/T247000 (10Halfak) https://github.com/wikimedia/articlequality/pull/105 [20:51:57] halfak great! [20:53:27] Oooh. For some reason, that idiom scanning is really slow. [20:53:35] I'm looking into it. [20:55:04] Ooops. Okay. [20:55:27] Uh oh. It looks like we have some problematic idiom matches showing up. E.g. 'of a', 'if only', 'Iron Curtain', 'According to' [20:55:33] These just look like common phrases. [20:58:25] Hmm. I'll look into this more later. [20:58:45] I wonder if our regex-based strategy is just too limited to do this scan. [21:02:35] Perhaps. [21:03:11] How do I get to reproduce it to get those problematic idioms, so I could take a look as well [22:02:38] haksoat, I'll get you a gist [22:03:56] https://gist.github.com/halfak/b9ce3f174a066e4851d04a2de7d2437d [22:04:09] In this example, I'm using the article on Alan Turing. [22:04:21] It looks like some of the idioms that get matches showed up in the edit comments. [22:04:27] *html comments. [22:04:36] [22:08:00] Taking a look... [22:13:13] The issue is with the idiom list itself [22:13:21] halfak [22:13:34] Right. I think we might need to filter some out. [22:13:38] I can find some of the results are standalone idioms in the list [22:13:47] Yes [22:14:26] Perhaps take away lines with just two words [22:14:58] But there are some valuable idioms that fall in that category there I think [22:17:12] Actually a lot... Haha [22:17:34] Right. Hmm. It seems many of these are just... phrases and not really idioms [22:25:51] True [22:26:12] Trying to find a pattern though. So we could probably filter them out. [22:27:27] I wonder if we could filter them out by how common they are in a few target pages. [22:27:40] E.g. we could take some featured articles and see which ones are common and just remove those. [22:32:08] That may work [22:32:51] How about the script that pulls them though, we'll need to keep in mind just in case we pull with it again. [22:33:28] Or we could overlook it for now, so we get this to work fine ASAP [22:34:54] Just finished building the model. No improvement. I could see why though. In the FA quality Alan Turing article, there's 76 "idioms"! [22:36:34] Some idioms get picked up in references. E.g. 'Anon (2017). "Turing, Alan Mathison". Who's Who. ukwhoswho.com (online Oxford University Press ed.). A & C Black, an imprint of Bloomsbury Publishing plc. ' contains "Who's Who". [22:38:03] Oooops [22:41:29] I think that latter example might be OK. [22:41:41] It's uncommon and I think the model should still be able to learn from it. [22:52:27] Hmmmm. Alright. I don't have a lot of experience with NLP and building models though, so I'll like to see how we resolve this. [22:59:35] I've got to run now but I can take another look in my morning tomorrow to advise. [22:59:43] Have a good one! [22:59:44] o/