[06:29:36] <icinga-wm>	 PROBLEM - https://grafana.wikimedia.org/dashboard/db/ores-extension grafana alert on einsteinium is CRITICAL: CRITICAL: https://grafana.wikimedia.org/dashboard/db/ores-extension is alerting: Failure rate alert.
[06:35:36] <icinga-wm>	 RECOVERY - https://grafana.wikimedia.org/dashboard/db/ores-extension grafana alert on einsteinium is OK: OK: https://grafana.wikimedia.org/dashboard/db/ores-extension is not alerting.
[15:02:07] <wikibugs>	 10Scoring-platform-team (Current), 10MediaWiki-extensions-ORES, 10Wikimedia-Site-requests, 10Patch-For-Review, and 2 others: Deploy ORES review tool in Turkish Wikipedia - https://phabricator.wikimedia.org/T139992#3949064 (10Liuxinyu970226)
[15:05:09] <wikibugs>	 10Scoring-platform-team, 10editquality-modeling, 10artificial-intelligence: Deploy edit quality models for nowiki - https://phabricator.wikimedia.org/T130291#3949075 (10Liuxinyu970226)
[15:06:04] <wikibugs>	 10Scoring-platform-team, 10editquality-modeling, 10artificial-intelligence: Deploy edit quality models for viwiki - https://phabricator.wikimedia.org/T130300#3949076 (10Liuxinyu970226)
[15:06:35] <wikibugs>	 10Scoring-platform-team (Current), 10editquality-modeling, 10Russian-Sites, 10User-Ladsgroup, 10artificial-intelligence: Deploy edit quality models for ruwiki - https://phabricator.wikimedia.org/T130293#3949080 (10Liuxinyu970226)
[15:06:39] <wikibugs>	 10Scoring-platform-team (Current), 10Wikilabels, 10editquality-modeling, 10Russian-Sites, 10artificial-intelligence: Complete ruwiki edit quality campaign - https://phabricator.wikimedia.org/T130270#3949082 (10Liuxinyu970226)
[15:07:43] <wikibugs>	 10Scoring-platform-team, 10ORES, 10articlequality-modeling, 10Turkish-Sites, 10artificial-intelligence: Generate ranked list of likely SM articles for Turkish Wikipedia - https://phabricator.wikimedia.org/T168075#3949089 (10Liuxinyu970226)
[15:09:15] <wikibugs>	 10Scoring-platform-team, 10Commons, 10editquality-modeling, 10artificial-intelligence: Train 'reverted' model for commonswiki - https://phabricator.wikimedia.org/T116938#3949096 (10Liuxinyu970226)
[15:19:04] <halfak>	 o/
[15:22:14] <awight|distract>	 hola!
[15:22:44] <awight|distract>	 Sorry about my multiple handicaps today, but I think I can make up some time on Friday.
[15:35:20] <halfak>	 awight|distract, understood
[15:35:54] <halfak>	 I'm working on next year's budget today. 
[15:36:10] <halfak>	 Also I've got a lecture to give in a couple of hours
[15:37:10] <awight|distract>	 Fun!
[18:06:16] <DebianTUX>	 halfak: hey bro! whats up? i'v just cloned wikilabels and the Campaigns.create method is missing info_url in the new_campaign utility. may I write a pull request to fix this?
[18:06:40] <DebianTUX>	 s/info_url/info_url parameter
[18:06:56] <halfak>	 Yes please :) 
[18:26:45] <halfak>	 I'm AFK for a while to give a lecture. 
[18:26:58] <halfak>	 I should be =back online in a couple of hours
[18:26:59] <halfak>	 o/
[20:07:13] <DebianTUX>	 halfak: finished that pull request :P (the simplest pull request of the project, ever)
[20:52:24] <ebernhardson>	 a python/ai question: Any good way to evaluate what "perfect accuracy" means for a dataset? My problem is that the inputs are user sessions. If two users were shown the same documents in the same order for the same query they will have the same input vector, but the target can vary based on what they actually clicked. Because of this one input vector has multiple expected outcomes (of which it 
[20:52:30] <ebernhardson>	 can only predict one). How could i measure this?
[20:54:21] <ebernhardson>	 as an added difficulty, the input matrix is sparse and would be multiple terabytes to convert to dense :(
[20:55:19] <ebernhardson>	 each session is represented by a 123,915 dimensional vector
[20:56:56] <travis-ci>	 wiki-ai/editquality#105 (make_automated - 8bb9012 : Amir Sarabadani): The build passed. https://travis-ci.org/wiki-ai/editquality/builds/338199252
[21:25:15] <halfak>	 o/
[21:25:27] <halfak>	 I'm back from the lecture. 
[21:25:28] <halfak>	 Went well.  
[21:25:39] <halfak>	 I might need to lay down for a bit though. 
[21:25:57] <halfak>	 Amir1, thanks for the fiwiki merge :) 
[21:36:10] <Amir1>	 halfak: no worries, I wish I could be more helpful 
[21:36:22] <halfak>	 Yup.  Tried doing some reviewing and am feeling bad.  Gonna try a nap
[21:36:41] <halfak>	 Amir1, need something on your plate now?
[21:36:55] <Amir1>	 why not
[21:37:28] <halfak>	 heh.  So.  I'm looking for an analysis of an idea from your point of view. 
[21:38:15] <halfak>	 Let's say we do JADE as a simple MediaWiki extension with a ContentHandler and a JS UI like wikibase.   What would that take?
[21:38:20] <halfak>	 What are the hard parts?
[21:38:32] <halfak>	 And what would be difficult to maintain
[21:40:44] <halfak>	 E.g. wait is painful about having a custom UI on top of Wikibase?
[21:44:42] <halfak>	 Amir1, ^ 
[21:44:59] <halfak>	 *wait --> what
[21:45:13] * halfak types while laying down
[21:45:51] <Amir1>	 halfak: it is a little bit of effort 
[21:46:37] <Amir1>	 as the current UI is horrible and there is some plans to completely rip it off and make a new one from scratch (using vue or react)
[21:46:47] <Amir1>	 we can piggy back in their efforts
[21:47:14] <Amir1>	 but It should not be that hard
[21:47:15] <halfak>	 Interesting. 
[21:47:19] <Amir1>	 that's for sure
[21:47:31] <Amir1>	 as we are not planning to be as complex as Wikibase
[21:47:33] <halfak>	 What about using content handler and implementing schema validation?
[21:48:11] <Amir1>	 that part should not be hard
[21:48:13] <halfak>	 Oh!  One other Q.  Can I use the API to make structured edits in Wikibase?  E.g. change a specific property without editing the whole content?
[21:48:36] <Amir1>	 halfak: yup, action=wb*
[21:48:40] <Amir1>	 look at them
[21:48:57] <Amir1>	 https://www.wikidata.org/w/api.php
[21:49:01] <halfak>	 Interesting.  Can you lay out what would need to come together to have something like that for JADE?
[21:49:17] <halfak>	 Like, in an etherpad/wiki page, with commentary
[21:49:46] <halfak>	 I want to compare approaching this as MW-only to approaching it as a Server/MW interaction. 
[21:50:34] <Amir1>	 sure
[21:51:46] <Amir1>	 I will write this down tomorrow (ORES day)
[21:52:35] <halfak>	 Awesome.  Thanks :) 
[21:56:34] <Amir1>	 Thank you! I hope I can start working on JADE soon
[21:57:06] <Amir1>	 halfak: oh btw. Please review my phab card regarding edge cases and let me know we should we do with them 
[21:57:56] <halfak>	 Will do...  but probably not this week.  Feeling really icky today and otherwise laser focused on annual planning nonsense.
[21:59:39] <Amir1>	 okay, thanks :)
[21:59:45] <Amir1>	 I'm leaving for the day
[21:59:47] <Amir1>	 see you
[22:37:22] <halfak>	 o/
[22:59:04] <wikibugs>	 10Scoring-platform-team, 10MediaWiki-extensions-ORES, 10MediaWiki-extensions-PageAssessments: Implement ORES wp10 predictions in PageAssessments tool - https://phabricator.wikimedia.org/T156820#3950874 (10TBolliger)
[22:59:16] <awight>	 oh hey!
[22:59:23] <awight>	 That was quite the day
[22:59:40] <awight>	 For the last few hours, I’ve somehow had *5* kids to be responsible for
[22:59:57] <awight>	 Michelle finally showed up, soaked in blood and birth fluids
[23:00:02] <awight>	 That’s one helluva job.
[23:04:02] <apergos>	 oh my
[23:12:42] <paladox>	 awight lol
[23:14:18] * awight tosses a cup of placenta smoothie into the channel for effect
[23:16:24] <paladox>	 lol
[23:16:39] <apergos>	 it's not smoothie season
[23:16:42] <apergos>	 not til summer
[23:17:04] <paladox>	 it's smoothy's all day @ mcdonalds :)
[23:17:13] <paladox>	 apergos ^^
[23:17:36] <apergos>	 eewwww
[23:17:50] <paladox>	 at least it is here
[23:18:02] <Platonides>	 not knowing who Michelle was, it looked daunting:
[23:18:07] <Platonides>	 "Michelle finally showed up, soaked in blood"
[23:18:15] <paladox>	 apergos https://www.mcdonalds.com/gb/en-gb/product/mango-pineapple-iced-fruit-smoothie-regular.html
[23:18:23] <paladox>	 lol
[23:18:49] <Platonides>	 but seems that in spite of that, it actually was a happy event :)
[23:19:40] <apergos>	 i avoid the large chains like that... there's a book, 'fast food nation'
[23:19:48] <apergos>	 check it out if you have a strong stomach :-P
[23:20:02] <paladox>	 apergos lol really?
[23:20:13] <paladox>	 i have like 4 or 5 where i live
[23:20:26] <paladox>	 though i do live in quite a big town.
[23:23:48] <apergos>	 apparently there are 17 here in the whole country
[23:23:51] <apergos>	 which is more than I expected
[23:24:24] <paladox>	 apergos only 17 mcdonalds?
[23:24:31] <apergos>	 uh huh
[23:24:33] <paladox>	 which country is this?
[23:24:36] <apergos>	 greece
[23:24:40] <paladox>	 oh
[23:24:43] <apergos>	 should there be a bunch in every country?
[23:24:59] <paladox>	 apergos in the uk it has 100's or even thounsonds
[23:25:00] <apergos>	 if so, why? if not, why not? discuss ;-P
[23:25:03] <paladox>	 there's one every where
[23:25:30] <paladox>	 apergos heh
[23:26:05] <Platonides>	 you should start by referencing the number of mcdonalds per country :P
[23:26:43] <paladox>	 apergos in the town i currently live in, we have around 3-4 pizza hut's, a couple of dominos, 4-5 mcdonalds possibly more. 3 or 4 subways.
[23:26:51] <paladox>	 3-4 kfc's
[23:27:08] <paladox>	 1 toby carvery
[23:28:18] <paladox>	 apergos one of the mcondalds it really near me, though i have never been to it
[23:28:23] <paladox>	 as it's on the motorway
[23:29:26] <apergos>	 there must be a way to tie this back to machine learning at wikimedia
[23:29:33] <apergos>	 otherwise we are waaaaaay off topic
[23:29:39] <apergos>	 ( :-P )
[23:30:27] <paladox>	 apergos there's an offtopic channel
[23:30:31] <paladox>	 #wikimedia-offtopic
[23:30:41] <apergos>	 too many channels
[23:31:01] <paladox>	 heh