[08:25:28] <wikibugs>	 10Scoring-platform-team (Current), 10drafttopic-modeling: Implement English pronoun count features in topic models - https://phabricator.wikimedia.org/T242345 (10kevinbazira) I've added English pronoun count features to drafttopic. PR: https://github.com/wikimedia/drafttopic/pull/43
[14:25:49] <wikibugs>	 10ORES, 10Scoring-platform-team: [Discuss] Future ORES architecture - https://phabricator.wikimedia.org/T226193 (10Ottomata) https://flyte.org/
[14:36:19] <halfak>	 o/ kevinbazira 
[14:36:25] <halfak>	 I just saw your PR.  Having a look. 
[14:37:03] <kevinbazira>	 o/ halfak
[14:37:12] <halfak>	 Are you working on rebuilding the model now? 
[14:37:32] <kevinbazira>	 Which model should I rebuild?
[14:40:07] <halfak>	 either drafttopic or articletopic -- take your pic. 
[14:40:12] <halfak>	 Eventually we'll want to rebuild both. 
[14:40:23] <halfak>	 But we need one now to find out if the theory holds in practice. 
[14:42:55] <kevinbazira>	 Alright, should I rebuild the model before the review?
[14:46:28] <halfak>	 Yes.  We don't know if the change works yet.  We have a hypothesis: "Pronoun counts & proportion should help us identify biographies of women".  We need to test it empirically. 
[14:47:04] <halfak>	 Personally, I can't imagine this *wouldn't* work.  But hey, it's always good to know what we're getting before we commit to adding new features. 
[14:48:00] <kevinbazira>	 I see. Let me start the rebuild process.
[15:10:02] <kevinbazira>	 halfak, while rebuilding the drafttopic model, I'm getting a TypeError with load_kv(). Do you mind taking a look at it via a short video call?
[15:13:12] <halfak>	 Sure.  Call when ready
[15:13:15] <halfak>	 kevinbazira, ^ 
[15:13:31] <halfak>	 But you probably need to upgrade your version of revscoring. 
[15:13:37] <halfak>	 pip install revscoring==2.6.3
[15:14:11] <kevinbazira>	 Alright, let me first upgrade revscoring!
[15:58:52] <wikibugs>	 10ORES, 10Scoring-platform-team: [Discuss] Future ORES architecture - https://phabricator.wikimedia.org/T226193 (10Nuria) It is probably wise to start here with the build piece and do a design document on how to move the build of models to happen in the cluster. Some projects that do that very thing for scikit...
[20:22:29] <halfak>	 Arg. Totally got lost in mwtext land again.  I want to to be flexible for non-english Wikis and I want a set of tests that work before I declare victory.  
[20:22:44] <halfak>	 I think I can make it today.  Then I'll set up a demo for how it works.  
[20:24:01] <halfak>	 I need to change locations.  back in 30 
[21:55:47] <wikibugs>	 10MediaWiki-extensions-ORES, 10Scoring-platform-team, 10Discovery-Search, 10Growth-Team, 10NewcomerTasks 1.1: Expose ORES drafttopic data in ElasticSearch via a custom CirrusSearch keyword - https://phabricator.wikimedia.org/T240559 (10Tgr) >>! In T240559#5770430, @dcausse wrote: > I suggest a keyword sl...
[22:48:13] <halfak>	 Finally!  https://github.com/mediawiki-utilities/python-mwtext
[22:48:29] <halfak>	 Turns out this works great for korean, arabic, & vietnamese -- I'm pretty sure.