[00:37:23] <wikibugs>	 10Scoring-platform-team, 10Scap, 10Patch-For-Review, 10Release-Engineering-Team (Kanban): Support git-lfs - https://phabricator.wikimedia.org/T180627#4128798 (10awight)
[00:51:20] <wikibugs>	 (03PS1) 10Awight: Update assets submodule with word2vec bin [services/ores/deploy] - 10https://gerrit.wikimedia.org/r/425939
[01:03:33] <wikibugs>	 (03PS4) 10Awight: Point submodules at gerrit [services/ores/deploy] - 10https://gerrit.wikimedia.org/r/425717 (https://phabricator.wikimedia.org/T180627)
[01:03:35] <wikibugs>	 (03PS6) 10Awight: Add the assets submodule, git-lfs enabled [services/ores/deploy] - 10https://gerrit.wikimedia.org/r/419613 (https://phabricator.wikimedia.org/T180627)
[01:03:37] <wikibugs>	 (03PS2) 10Awight: Update assets submodule with word2vec bin [services/ores/deploy] - 10https://gerrit.wikimedia.org/r/425939
[01:14:24] <wikibugs>	 (03PS7) 10Awight: [DNM] Add the assets submodule, git-lfs enabled [services/ores/deploy] - 10https://gerrit.wikimedia.org/r/419613 (https://phabricator.wikimedia.org/T180627)
[13:42:10] <halfak>	 o/
[13:51:22] <Vermont>	 hi
[14:16:34] <paladox>	 i think we may get puppet errors today
[14:16:38] <paladox>	 due to the openstack upgrade
[14:18:38] <halfak>	 Oh?  Did it cause some alerts paladox?
[14:18:49] <paladox>	 halfak not yet, but just in case :)
[14:19:00] <paladox>	 it will take all day according to the email labs sent
[14:20:27] <wikibugs>	 10Scoring-platform-team, 10editquality-modeling, 10artificial-intelligence: [research] Why is the japanese 'reverted' model so bad? - https://phabricator.wikimedia.org/T133405#4129739 (10Halfak) We just got pinged to re-consider this here: https://www.mediawiki.org/wiki/Topic:Ub6ir6tww9z81960  @Ladsgroup, fr...
[16:29:57] <halfak>	 WOah VSP
[16:30:15] <halfak>	 Or Vermont 
[16:30:35] <halfak>	 Oh.   Probably not you?
[16:35:38] <codezee>	 halfak: Using the previous abstract and current points I've put together a final version of the new abstract. Since abstracts need to go on 16th, it'd great if you can spare 5min to glance through that for any issues...
[16:35:55] <codezee>	 its on overleaf
[17:02:10] <halfak>	 Thanks.  I'll have a look in an hour
[18:04:02] <halfak>	 Got out of meeting.  Doing some family tech support and them I'm looking at the abstract
[18:33:00] <codezee>	 np
[18:56:17] <Vermont>	 halfak: VSP is me.
[18:56:24] <halfak>	 o/ Vermont 
[18:56:32] <halfak>	 Saw you getting disconnected a lot 
[18:56:35] <halfak>	 Connection issues?
[18:59:07] <Vermont>	 Yep.
[18:59:30] <Vermont>	 Now I have to deal with getting it unbanned from all 56 channels it was in.
[19:01:00] <halfak>	 codezee, {{done}} 
[19:01:00] <AsimovBot>	 You rule, halfak!
[19:01:01] <halfak>	 please review
[19:01:07] <halfak>	 lol
[19:01:12] <codezee>	 halfak: thanks! looking
[19:05:59] <codezee>	 I still have a long way to go in learning to write better :)
[19:06:54] <halfak>	 Na.  I incorporated most of your points.  Do you like my version better?
[19:10:38] <codezee>	 yes, its concise, one thing that I always get confused about is usage of past and present, like I think it should be "we developed an automatic topic modeling strategy"
[19:11:04] <halfak>	 Ahh yes.  It should be consistent across the board. 
[19:11:24] <halfak>	 So rather than "develop" it should be the present tense of the paper "describe"
[19:15:14] <halfak>	 {{adjusted}}
[19:15:15] <AsimovBot>	 10[1] 04https://meta.wikimedia.org/wiki/Template:adjusted
[19:15:18] <halfak>	 :) 
[19:15:21] <halfak>	 codezee, ^ 
[19:18:21] <codezee>	 does not exist :/
[19:19:18] <codezee>	 since "in this paper..." is occuring twice I'll go ahead and replace the 2nd one with further
[19:21:48] <halfak>	 Cool
[19:31:36] <codezee>	 halfak: are you aware of past works trying to use categories for topics? if you know you can throw the keywords i'll google
[19:32:57] <halfak>	 codezee, https://scholar.google.com/scholar?hl=en&q=wikipedia+category+structure&btnG=&as_sdt=1%2C24&as_sdtp=
[19:33:38] <halfak>	 I've got to change locations.  Back in ~25 mins
[19:33:51] <codezee>	 thanks...i'll dive into those
[20:12:58] <halfak>	 https://www.mediawiki.org/wiki/Topic:Ub58fvti46x4e5lf
[20:19:08] <codezee>	 i see cscw has a system design category whose description exactly fits the topic of this paper...
[20:21:59] <codezee>	 i'll set aside sometime to read this, interesting insight into ores tech specs
[20:22:19] <halfak>	 :) It felt good to write that out. 
[20:22:30] <halfak>	 I still feel like I'm behind on ORES system paper essays. 
[20:22:36] <halfak>	 But it's getting close.  
[20:22:45] <halfak>	 If I can button this up next week, it'll be a small miracle!
[20:24:40] <codezee>	 although i can't add much to the tech specs but feel free to ping for looking over it for any minor errors...
[20:28:30] <ebernhardson>	 perhaps interesting, slides for my talk at haystack: https://commons.wikimedia.org/wiki/File:From_Clicks_to_Models_The_Wikimedia_LTR_Pipeline.pdf
[20:36:28] <codezee>	 ebernhardson: a question, what kind of data was being shuffled b/w JVM and python? during label generation
[20:38:25] <halfak>	 :) 
[20:52:35] <ebernhardson>	 codezee: in that case it's ~40% of search sessions over the last 90 days
[20:52:47] <ebernhardson>	 codezee: in general though any time in pyspark you actually load the data into python (as opposed to using the sql dataframe) it's going to suffer a fairly significant overhead in my experience
[20:53:30] <ebernhardson>	 not sure these came through, hotel wifi is horrible ...
[20:53:35] <ebernhardson>	 codezee: in that case it's ~40% of search sessions over the last 90 days
[20:53:40] <ebernhardson>	 codezee: in general though any time in pyspark you actually load the data into python (as opposed to using the sql dataframe) it's going to suffer a fairly significant overhead in my experience
[21:45:33] <halfak>	 Heading out.  Will be doing a bit of writing review over the weekend.  codezee, I think I'll want to take a full pass of the paper on Sunday if you can have it ready. 
[21:45:46] <halfak>	 Otherwise, see y'all on Monday!