[06:59:40] (03PS1) 10Ashuro07: Merge branch 'master' of ssh://gerrit.wikimedia.org:29418/mediawiki/extensions/ORES into review/ashuro07/590259 [extensions/ORES] - 10https://gerrit.wikimedia.org/r/592521 [06:59:50] (03CR) 10jerkins-bot: [V: 04-1] Merge branch 'master' of ssh://gerrit.wikimedia.org:29418/mediawiki/extensions/ORES into review/ashuro07/590259 [extensions/ORES] - 10https://gerrit.wikimedia.org/r/592521 (owner: 10Ashuro07) [07:00:54] (03CR) 10Ashuro07: [C: 03+1] Merge branch 'master' of ssh://gerrit.wikimedia.org:29418/mediawiki/extensions/ORES into review/ashuro07/590259 [extensions/ORES] - 10https://gerrit.wikimedia.org/r/592521 (owner: 10Ashuro07) [07:01:44] (03Abandoned) 10Ashuro07: Merge branch 'master' of ssh://gerrit.wikimedia.org:29418/mediawiki/extensions/ORES into review/ashuro07/590259 [extensions/ORES] - 10https://gerrit.wikimedia.org/r/592521 (owner: 10Ashuro07) [07:08:18] (03PS4) 10Ashuro07: Upgrade tests to WebdriverIO v-5 [extensions/ORES] - 10https://gerrit.wikimedia.org/r/590259 (https://phabricator.wikimedia.org/T248223) [07:08:28] (03CR) 10jerkins-bot: [V: 04-1] Upgrade tests to WebdriverIO v-5 [extensions/ORES] - 10https://gerrit.wikimedia.org/r/590259 (https://phabricator.wikimedia.org/T248223) (owner: 10Ashuro07) [07:29:02] (03PS5) 10Ashuro07: Upgrade tests to WebdriverIO v-5 [extensions/ORES] - 10https://gerrit.wikimedia.org/r/590259 (https://phabricator.wikimedia.org/T248223) [07:31:12] (03CR) 10jerkins-bot: [V: 04-1] Upgrade tests to WebdriverIO v-5 [extensions/ORES] - 10https://gerrit.wikimedia.org/r/590259 (https://phabricator.wikimedia.org/T248223) (owner: 10Ashuro07) [07:37:33] (03PS6) 10Ashuro07: Upgrade tests to WebdriverIO v-5 [extensions/ORES] - 10https://gerrit.wikimedia.org/r/590259 (https://phabricator.wikimedia.org/T248223) [14:17:27] Hi halfak o/ [14:17:49] Hey kevinbazira! [14:18:08] Do you have a minute to jump on a short video call? [14:18:24] Yes. [14:18:26] Call when ready [14:18:35] Cool. Calling now ... [14:57:33] 10Scoring-platform-team (Current), 10drafttopic-modeling: Why does loading the drafttopic models take so much memory? - https://phabricator.wikimedia.org/T250435 (10akosiaris) If I have understood correctly what this does (haven't read the docs), the difference would be in how the memory is accounted for in th... [15:09:18] 10Jade, 10Scoring-platform-team: [Spike] What facilities are available to us when rendering edit comments? - https://phabricator.wikimedia.org/T250723 (10kevinbazira) a:03kevinbazira Based on [[ https://github.com/wikimedia/mediawiki-extensions-Wikibase/blob/3f2892b9b63d71ee657be8949f162e83979ba8fc/repo/Repo... [15:10:04] 10Jade, 10Scoring-platform-team (Current): [Spike] What facilities are available to us when rendering edit comments? - https://phabricator.wikimedia.org/T250723 (10kevinbazira) [16:43:07] 10ORES, 10Scoring-platform-team: ORES Beta startup errors not being routed to our app logging. - https://phabricator.wikimedia.org/T250712 (10Halfak) a:05Halfak→03None [17:06:05] 10Scoring-platform-team, 10artificial-intelligence: Add `words_to_watch` to articlequality and draftquality models in ptwiki - https://phabricator.wikimedia.org/T251171 (10Chtnnh) [17:11:35] 10Scoring-platform-team, 10artificial-intelligence: Add `words_to_watch` to articlequality and draftquality models in ptwiki - https://phabricator.wikimedia.org/T251171 (10Chtnnh) After adding `words_to_watch` to draftquality we did not achieve any significant fitness improvement. This is evident in the tuning... [17:42:53] posting our async update notes -- [17:43:09] kevinbazira - [17:43:12] Last Week: [17:43:13] I was mostly OOO for the Giant WMF holiday: April 22-26th [17:43:15] Jade [17:43:17] Focused on addressing UI issues identified in user-testing from https://phabricator.wikimedia.org/T247897 [17:43:19] MW Core [17:43:21] My first contributions to MediaWiki core got merged. Thanks to halfak, Volker_E and edsanders for the guidance! [17:43:23] 1. https://phabricator.wikimedia.org/T249804 [17:43:25] 2. https://phabricator.wikimedia.org/T250788 [17:43:27] T: [17:43:29] Jade [17:43:31] Looked into what facilities are available to us when rendering edit comments: https://phabricator.wikimedia.org/T250723 [17:43:33] - Based on this and this, Wikidata is using the FormatAutocomments MediaWiki hook to render localized strings for edit comments. [17:43:35] halfak - [17:43:37] Last week: New versions of revscoring (2.7.2 out!). Got new ptwiki models deployed to production. I discussed the models with ptwikipedians and helped them set up some user scripts. I also worked with chtnnh and haksoat a bit during the holiday days for the WMF. Specifically, we continued work on ptwiki's draftquality datasets and tokenization work. I also made some progress on the ORES paper [17:43:39] revision -- specifically, I have been working on discussions of decoupling product processes and applying participatory design to our modeling process. [17:43:41] T: Interview with an SWE candidate! This is a bit out of order so I've reached out to Erika to clarify. Otherwise, I have a heavy meeting day with sync ups with the Legal and Research teams. If I find some time I'll review Andy's notes on the Jade secondary integrations and try to push forward a proposal. [17:43:45] haksoat - [17:43:47] Last week: [17:43:49] I was able to get Elasticsearch to tokenize text based on passed in regular expressions. However, speed wasn't impressive. Also did some research on the kinds of engines that power regular expressions. [17:43:51] Today: [17:43:53] Reading extensively on the NFA (Nondeterministic Finite Automaton) engine which is what powers our tokenizer and Elasticsearch's tokenizer, in Python and Java respectively. [17:43:58] and me - [17:44:00] Last week: Got the majority of the ad-hoc Jade secondary schema/integration work done, it's not a great solution, but it should work for the MVP. There's still some work to be done related to scripts that seed the jade_facet table on install and also need to fix some jenkins errors for both of the WIP patchsets. [17:44:02] T: Taking a little break from Jade while I think about the 2ndary schema stuff a bit more. Will be looking at automating docs for ORES and writing a blubber file so we can eventually move to k8s [19:31:15] wikimedia/editquality#721 (master - b4ee594 : Andy Craze): The build was fixed. https://travis-ci.org/wikimedia/editquality/builds/680259899 [20:01:00] 10Scoring-platform-team (Current), 10Wikilabels, 10editquality-modeling, 10artificial-intelligence: Create follow-up edit quality campaign for ptwikipedia - https://phabricator.wikimedia.org/T246668 (10ACraze) Reviewed & merged that PR @Halfak [20:53:46] Thanks accraze! Above and beyond! [21:22:13] 10Scoring-platform-team, 10artificial-intelligence: Add `words_to_watch` to articlequality and draftquality models in ptwiki - https://phabricator.wikimedia.org/T251171 (10Halfak) @He7d3r, I'm surprised this didn't work. I would expect that many vandalism or spam articles would have phrases from words_to_watc... [21:42:09] 10Scoring-platform-team, 10artificial-intelligence: Add `words_to_watch` to articlequality and draftquality models in ptwiki - https://phabricator.wikimedia.org/T251171 (10He7d3r) That is odd. Does this tuning report reflect only the changes in the ptwiki features, or does it also include other articles to the... [21:46:35] 10Scoring-platform-team, 10artificial-intelligence: Add `words_to_watch` to articlequality and draftquality models in ptwiki - https://phabricator.wikimedia.org/T251171 (10He7d3r) @GoEThe Correct me if I'm mistaken, but I believe a reasonable amount of new articles having vandalism or spam would contain expres... [21:48:27] 10Scoring-platform-team, 10artificial-intelligence: Add `words_to_watch` to articlequality and draftquality models in ptwiki - https://phabricator.wikimedia.org/T251171 (10Halfak) It does include the new articles matched beyond the ER# tags. Could it be possible that we're not matching the features effectivel... [21:56:06] 10Scoring-platform-team, 10artificial-intelligence: Add `words_to_watch` to articlequality and draftquality models in ptwiki - https://phabricator.wikimedia.org/T251171 (10He7d3r) It could be. For example, @Darwinius noticed that images loaded from Wikidata are not counted: https://www.mediawiki.org/wiki/ORES/... [21:56:46] halfak, ^ Is it even possible to detect if the article has images from Wikidata? [21:56:53] Hmm. I don't think so. [21:57:28] Maybe create a feature for the number of templates wich might add images from wikidata? [21:57:51] as part of the language specific features [21:58:07] Sounds reasonable to me. [22:01:10] halfak, changing the subject a little: is it possible to get the "feature importance" from the GradientBoosting models? [22:01:50] Hmm. It is. We don't output that, but we can look at it. I have the model loaded. I'll see if I can check it for the word2watch. [22:02:59] wikimedia/ores#1432 (no_cache_default - 591eabd : halfak): The build was fixed. https://travis-ci.org/wikimedia/ores/builds/680303799 [22:03:09] damn right [22:06:37] In part it could (?) be useful for answering questions such as https://pt.wikipedia.org/wiki/User_talk:Chtnnh?diff=58099803 [22:09:02] on how the model gets to its predictions (e.g. which features are more relevant) [22:09:34] 10Scoring-platform-team, 10artificial-intelligence: Add `words_to_watch` to articlequality and draftquality models in ptwiki - https://phabricator.wikimedia.org/T251171 (10GoEThe) I would imagine that we would, unless people are too loose with the deletion reason and are using that as a catch-all reason to del... [22:17:52] 10Scoring-platform-team, 10artificial-intelligence: Add `words_to_watch` to articlequality and draftquality models in ptwiki - https://phabricator.wikimedia.org/T251171 (10Halfak) Here's a random sample of 100 articles from the dataset. Columns are: label, rev_id, words_to_watch detected ` OK 54564977 ['cham... [22:26:46] 10Scoring-platform-team, 10artificial-intelligence: Add `words_to_watch` to articlequality and draftquality models in ptwiki - https://phabricator.wikimedia.org/T251171 (10Halfak) Here's the "feature importances" as reported for @chtnnh's model: ` 0.0 feature.ptwiki.revision.category_links 0.0 feature.(ptwik... [22:26:57] OK with that I'm out of here. Have a good one, folks.