[02:42:37] <icinga-wm>	 PROBLEM - ORES web node labs ores-web-05 on ores.wmflabs.org is CRITICAL: CRITICAL - Socket timeout after 10 seconds
[02:44:37] <icinga-wm>	 RECOVERY - ORES web node labs ores-web-05 on ores.wmflabs.org is OK: HTTP OK: HTTP/1.1 200 OK - 370 bytes in 6.031 second response time
[04:38:36] <grrrit-wm>	 (03CR) 10Ladsgroup: [C: 032] Some more CI tests [extensions/ORES] - 10https://gerrit.wikimedia.org/r/301786 (https://phabricator.wikimedia.org/T140455) (owner: 10Ladsgroup)
[04:40:40] <grrrit-wm>	 (03Merged) 10jenkins-bot: Some more CI tests [extensions/ORES] - 10https://gerrit.wikimedia.org/r/301786 (https://phabricator.wikimedia.org/T140455) (owner: 10Ladsgroup)
[07:00:39] <grrrit-wm>	 (03PS1) 10Ladsgroup: Small fixes to tests [extensions/ORES] - 10https://gerrit.wikimedia.org/r/301916 
[07:29:25] <grrrit-wm>	 (03PS11) 10Ladsgroup: Integrate with Special:Contributions [extensions/ORES] - 10https://gerrit.wikimedia.org/r/264608 (https://phabricator.wikimedia.org/T122537) (owner: 10Awight)
[07:30:33] <grrrit-wm>	 (03CR) 10Ladsgroup: "PS11 is rebase only" [extensions/ORES] - 10https://gerrit.wikimedia.org/r/264608 (https://phabricator.wikimedia.org/T122537) (owner: 10Awight)
[07:54:53] <grrrit-wm>	 (03PS12) 10Ladsgroup: Integrate with Special:Contributions [extensions/ORES] - 10https://gerrit.wikimedia.org/r/264608 (https://phabricator.wikimedia.org/T122537) (owner: 10Awight)
[08:04:09] <grrrit-wm>	 (03PS13) 10Ladsgroup: Integrate with Special:Contributions [extensions/ORES] - 10https://gerrit.wikimedia.org/r/264608 (https://phabricator.wikimedia.org/T122537) (owner: 10Awight)
[08:06:07] <grrrit-wm>	 (03CR) 10Ladsgroup: [C: 031] "You can test it in http://mw-revscoring.wmflabs.org/wiki/Special:Contributions/Someone" [extensions/ORES] - 10https://gerrit.wikimedia.org/r/264608 (https://phabricator.wikimedia.org/T122537) (owner: 10Awight)
[15:45:16] <halfak>	 o/
[16:00:12] <halfak>	 I've got some emails to work on before I start looking at the spam/vandalism new page creation problem. 
[16:00:26] <halfak>	 I'm overdue for some writing about that. 
[16:02:36] <grrrit-wm>	 (03CR) 10Thiemo Mättig (WMDE): [C: 031] Small fixes to tests [extensions/ORES] - 10https://gerrit.wikimedia.org/r/301916 (owner: 10Ladsgroup)
[16:04:26] <halfak>	 So, my UI has decided that only the number 3 will be shown in menues
[16:04:38] <halfak>	 All other letters/numbers are transparent 
[16:04:43] <halfak>	 Probably going to reboot soon. 
[16:52:43] <halfak>	 o/ sabya 
[16:53:05] <sabya>	 o/ halfak 
[16:53:22] <halfak>	 How's life and hacking?
[16:54:15] <sabya>	 pretty much help up at work coz of a release.
[16:54:45] <halfak>	 Gotcha.  Sort of crunch time then?
[16:55:15] <sabya>	 yeah!
[16:57:13] <halfak>	 sabya, godspeed with your efforts.  
[16:57:18] <halfak>	 Oh one thing I'd like to tell you about
[16:57:32] <halfak>	 So the engineering we need to do to eventually incorporate your work with hash vectors into revscoring...
[16:57:48] <halfak>	 We're going to need that for a new model for detecting harassment too. 
[16:58:15] <halfak>	 So I'm hoping to pick up that work soon.  Maybe the timing will work out where we can meet in the middle with our work :) 
[16:59:06] <halfak>	 This new model is a harassment ("personal attack" specifically) classifier that uses a basic bag-of-words approach. 
[16:59:10] <sabya>	 sounds interesting :) can you explain a bit more? 
[16:59:30] <halfak>	 https://phabricator.wikimedia.org/T132580
[16:59:39] <halfak>	 This is the task for doing the implementing in revscoring. 
[17:00:15] <halfak>	 So, essentially, this would allow us to capture sparse vectors in a `revscoring.Feature` and dependency injection pattern. 
[17:04:30] <sabya>	 do you plan to use HV there as well or simple occurance of words, (bi|tri|n)grams?
[17:05:53] <halfak>	 sabya, I want us to be able to support both, but for right now, I want to just replicate the work on the harassment model that is just simple unigrams. 
[17:08:10] <sabya>	 ok
[17:13:09] <sabya>	 i'd like to know how are you planning it. could you update the phab task when you get a chance?
[17:13:18] <sabya>	 halfak:^
[17:14:03] <halfak>	 The TLDR right now is that I like option #2 of a "FeatureSet".  
[17:14:48] <halfak>	 So a FeatureSet will actually represent an ordered list of values (our vector) that will be expanded before being sent to the estimator.
[17:15:07] <halfak>	 But from the point of view of the dependency injection system, it'll look a lot like `revscoring.Feature`
[17:16:59] <sabya>	 k
[17:17:36] <halfak>	 Anyway, I think that the next step you are doing re. feature selection will fit nicely into this.  
[17:18:01] <halfak>	 We're going to need a way to "train" a FeatureSet though.  That'll be interesting.  :) 
[17:30:17] <sabya>	 got it :)
[19:39:38] <grrrit-wm>	 (03CR) 10Aude: [C: 032] Small fixes to tests [extensions/ORES] - 10https://gerrit.wikimedia.org/r/301916 (owner: 10Ladsgroup)
[19:40:20] <grrrit-wm>	 (03Merged) 10jenkins-bot: Small fixes to tests [extensions/ORES] - 10https://gerrit.wikimedia.org/r/301916 (owner: 10Ladsgroup)