[10:10:50] 10Revision-Scoring-As-A-Service-Backlog, 10Bad-Words-Detection-System, 10revscoring: Add language support for Tagalog - https://phabricator.wikimedia.org/T149475#2753796 (10Pokefan95) [10:12:30] 10Revision-Scoring-As-A-Service-Backlog, 10Bad-Words-Detection-System, 10revscoring: Add language support for Tagalog - https://phabricator.wikimedia.org/T149475#2753808 (10Pokefan95) Tagalog Wikipedia (and other sister Tagalog wikis) are always hit by vandalism, and it would be a great idea if the Tagalog c... [10:44:56] 10Revision-Scoring-As-A-Service-Backlog, 10Bad-Words-Detection-System, 10revscoring: Add language support for Tagalog - https://phabricator.wikimedia.org/T149475#2753812 (10Pokefan95) I notified the Tagalog Wikipedia community about this, which can be found at https://tl.wikipedia.org/wiki/Usapang_Wikipedia:... [11:25:23] 10Revision-Scoring-As-A-Service-Backlog, 10Bad-Words-Detection-System, 10revscoring: Add language support for Tagalog - https://phabricator.wikimedia.org/T149475#2753848 (10Pokefan95) Notified the Tagalog Wikibooks and Tagalog Wikitionary communites as well. [16:18:19] o/ [16:18:42] I'm going to try to extract some sentences from featured articles today :) [16:32:27] halfak: hey [16:32:31] Wo! [16:32:33] *Yo [16:32:47] I will join you in ores hack session in five minutes. [16:32:54] Great! [16:33:04] Finishing a patch for WMDE [16:37:16] I actually ended up sending a few emails, so I actually haven't started gathering sentences yet :/ [16:38:36] I think my plan is to start by gathering sentences from featured articles and then comparing PCFG scores against sentences from withheld featured articles. [16:59:11] halfak: I want to write a semi-announcement for the community about mw.config.get('oresData') and tell them that now they can write up gadgets on it. I want to ask you if that's okay or not [16:59:26] Yeah! That's a great idea. [16:59:41] Do you think that maybe you could hack together a minimal gadget to go with the announcement? [16:59:52] I already did [16:59:58] Great! Yeah. [17:00:02] https://phabricator.wikimedia.org/T144922#2736504 [17:00:07] So that people could test it out and see how you access/use oresData [17:00:11] the rainbow gadget :D [17:00:33] So yeah, that. It should go on the wiki so that people can install it the usual way. [17:00:51] Maybe even set up a repo for it. [17:00:56] oh, a real gadget [17:01:03] yeah... [17:01:09] Trail blazing, you know. [17:01:13] pave the way for more work. [17:01:20] I can do it [17:01:57] wiki-ai/mw-gadget-ORESRainbow [17:02:05] Or something like that [17:02:10] ORESPride? [17:02:18] :))) [17:02:22] :D [17:02:53] I will write it and probably make a repo in github [17:03:22] Cool! I'm looking forward to installing it and reviewing PRs. [17:46:26] I just pushed deltas-0.4.0 with improved sentence parsing [20:00:01] WIP PR here: https://github.com/wiki-ai/revscoring/pull/291 [20:00:10] Seems to work OK [20:00:25] Sentences get split in weird places, but some stuff is working pretty well [20:00:27] wiki-ai/revscoring#832 (sentence_datasources - 7642854 : halfak): The build passed. https://travis-ci.org/wiki-ai/revscoring/builds/171668527 [20:00:33] Thanks tr [20:00:35] avis [20:03:46] OK. WIP PR is up. I'll be working to extract sentences tomorrow. [20:04:00] If everything goes OK, I'll take away the (WIP) and call it good. [20:04:04] Have a good one, folks [20:05:33] halfak: you too [20:30:57] I still havent gotten used to PR being pull request rather than public relations :/ [20:47:18] (03CR) 10Ladsgroup: [C: 032] "I added as much as reviewer I could. No one reviewed in the past five days, it's just tests. I +2 it and if anyone disagree, I revert." [extensions/ORES] - 10https://gerrit.wikimedia.org/r/317172 (https://phabricator.wikimedia.org/T146560) (owner: 10Ladsgroup) [20:55:53] (03Merged) 10jenkins-bot: Add CacheTest.php (was Extensive CI tests, part III) [extensions/ORES] - 10https://gerrit.wikimedia.org/r/317172 (https://phabricator.wikimedia.org/T146560) (owner: 10Ladsgroup)