[02:02:32] <wikibugs>	 10Scoring-platform-team (Current), 10Security-Team, 10Wikilabels, 10Security: Wikilabels does not show the exact wikitext on diff: it parses <ref>s as HTML - https://phabricator.wikimedia.org/T250313 (10sbassett) p:05Triage→03Medium
[09:38:53] <wikibugs>	 10Jade, 10Scoring-platform-team (Current), 10MW-1.35-notes (1.35.0-wmf.30; 2020-04-28): Can endorse two different labels for the same facet via jadeproposeandendorse.  That should be impossible. - https://phabricator.wikimedia.org/T247974 (10kevinbazira)
[13:48:23] <halfak>	 o/
[15:06:19] <wikibugs>	 10Scoring-platform-team, 10Wikilabels: Internal links on comment/summary point to Wikilabels instead of the target wiki - https://phabricator.wikimedia.org/T250704 (10He7d3r)
[15:20:25] <wikibugs>	 10Scoring-platform-team, 10Wikilabels: Internal links on comment/summary point to Wikilabels instead of the target wiki - https://phabricator.wikimedia.org/T250704 (10He7d3r)
[15:20:38] <halfak>	 chtnnh, looks like we need to update https://quarry.wmflabs.org/query/43261 based on feedback. 
[15:20:48] <halfak>	 See Helder's last comment in https://phabricator.wikimedia.org/T246667
[15:21:52] <chtnnh>	 let me have a look halfak
[15:34:28] <halfak>	 Looks like we're blocked on review for the ORES deployment. 
[15:34:32] <halfak>	 https://phabricator.wikimedia.org/T250536
[15:36:58] <chtnnh>	 why are we blocked halfak
[15:37:11] <halfak>	 Waiting on review from accraze or kevinbazira
[15:37:32] <chtnnh>	 hasnt accraze already reviewed
[15:37:37] <chtnnh>	 i saw it on gerrit
[15:38:30] <chtnnh>	 https://gerrit.wikimedia.org/r/c/mediawiki/services/ores/deploy/+/589713
[15:39:28] <halfak>	 Aha!  He just forgot to submit. 
[15:39:36] <halfak>	 OK moving this to beta. 
[15:39:44] <chtnnh>	 doesnt seem like hes online tho
[15:39:55] <chtnnh>	 yay :D
[15:40:01] <halfak>	 It's OK.  So long as he +2s I can submit. 
[15:40:09] <chtnnh>	 great!
[15:52:01] <halfak>	 ORES seems to be struggling in beta.  Looks like we only have 8GB of RAM there and we're trying to make it work with a reduced worker pool
[15:52:22] <chtnnh>	 ouch 
[15:52:44] <chtnnh>	 but we just added 2 models right?
[15:52:59] <halfak>	 Right. 
[15:53:22] <chtnnh>	 how many models in total
[15:53:36] <halfak>	 Hmm.  I just tried restarting our celery workers and it failed.  But not because of memory as far as I can tell. 
[15:53:57] <chtnnh>	 is this somehow related to the memory issue you were working on?
[15:54:19] <halfak>	 I don't know. 
[15:57:08] <halfak>	 Hmm.  Celery seems to start up but it doesn't work any workers. 
[16:00:31] <halfak>	 Yeah.  This is super weird. 
[16:00:36] <halfak>	 uwsgi is also not starting workers. 
[16:00:48] <chtnnh>	 im sorry i am not able to understand all this :( havent dealt with this part of the codebase before
[16:02:33] <chtnnh>	 but is it possible that a mistake in our code may have caused this?
[16:03:27] <wikibugs>	 10ORES, 10Scoring-platform-team (Current): ORES beta won't start workers - https://phabricator.wikimedia.org/T250712 (10Halfak)
[16:04:14] <wikibugs>	 10ORES, 10Scoring-platform-team (Current): ORES beta won't start workers - https://phabricator.wikimedia.org/T250712 (10Halfak) I tried running `sudo service ores-celery-workers restart` and `sudo service uwsgi-ores restart`.  Everything proceeded as expected except that workers didn't start up.
[16:09:52] <wikibugs>	 10ORES, 10Scoring-platform-team (Current): ORES beta won't start workers - https://phabricator.wikimedia.org/T250712 (10Halfak) Aha!  I found this in the syslog:  ` Apr 20 16:08:56 deployment-ores01 celery-ores-worker[12127]: Traceback (most recent call last): Apr 20 16:08:56 deployment-ores01 celery-ores-work...
[16:10:15] <wikibugs>	 10ORES, 10Scoring-platform-team (Current): ORES Beta startup errors not being routed to our app logging.  - https://phabricator.wikimedia.org/T250712 (10Halfak)
[16:10:47] <halfak>	 Looks like I need to update the revscoring wheel. 
[16:12:24] <chtnnh>	 sounds like an important step
[16:35:46] <chtnnh>	 did it work after you rebuilt the wheel?
[16:41:28] <chtnnh>	 halfak ^
[16:44:22] <wikibugs>	 10ORES, 10Scoring-platform-team (Current): ORES Beta startup errors not being routed to our app logging. - https://phabricator.wikimedia.org/T250712 (10Halfak) a:03Halfak
[16:50:20] <wikibugs>	 10Scoring-platform-team, 10Wikilabels: Internal links on comment/summary point to Wikilabels instead of the target wiki - https://phabricator.wikimedia.org/T250704 (10Halfak) p:05Triage→03Low
[16:52:16] <wikibugs>	 10Scoring-platform-team, 10revscoring, 10Documentation, 10artificial-intelligence: Scoring example no longer works - https://phabricator.wikimedia.org/T250635 (10Halfak) p:05Triage→03Medium
[16:54:44] <wikibugs>	 10Scoring-platform-team, 10revscoring, 10artificial-intelligence: Template matching regex features only matches prefix - https://phabricator.wikimedia.org/T250522 (10Halfak) p:05Triage→03Lowest
[16:57:03] <wikibugs>	 10Scoring-platform-team, 10Wikilabels: Labeling gadget has non-functional expansion buttons - https://phabricator.wikimedia.org/T250176 (10Halfak) I think getting rid of the button would probably be the best solution here.
[16:59:45] <wikibugs>	 10Scoring-platform-team, 10Wikilabels: Labeling gadget has non-functional expansion buttons - https://phabricator.wikimedia.org/T250176 (10Halfak) p:05Triage→03Low
[17:00:12] <wikibugs>	 10Scoring-platform-team (Current), 10Discovery-Search, 10Elasticsearch, 10revscoring, 10artificial-intelligence: Improve the performance and quality of tokenization in revscoring - https://phabricator.wikimedia.org/T248480 (10Halfak)
[17:04:11] <wikibugs>	 10Jade, 10Scoring-platform-team (Current): Address UI issues identified in user-testing - https://phabricator.wikimedia.org/T247897 (10Halfak)
[17:08:01] <wikibugs>	 10Jade, 10Scoring-platform-team: Apply i18n to rendered edit comments in Jade - https://phabricator.wikimedia.org/T250721 (10Halfak)
[17:08:47] <wikibugs>	 10Jade, 10Scoring-platform-team: Apply i18n to rendered edit comments in Jade - https://phabricator.wikimedia.org/T250721 (10Halfak) p:05Triage→03High
[17:11:09] <wikibugs>	 10Jade, 10Scoring-platform-team: [Spike] What facilities are available to us when rendering edit comments? - https://phabricator.wikimedia.org/T250723 (10Halfak)
[17:11:31] <wikibugs>	 10Jade, 10Scoring-platform-team: [Spike] What facilities are available to us when rendering edit comments? - https://phabricator.wikimedia.org/T250723 (10Halfak) p:05Triage→03High
[17:14:24] <wikibugs>	 10Jade, 10Scoring-platform-team, 10MediaWiki-Page-Diffs: [Spike] How can we render custom diffs for Jade pages? - https://phabricator.wikimedia.org/T250725 (10Halfak)
[17:14:44] <wikibugs>	 10Jade, 10Scoring-platform-team, 10MediaWiki-Page-Diffs: [Spike] How can we render custom diffs for Jade pages? - https://phabricator.wikimedia.org/T250725 (10Halfak) p:05Triage→03Medium
[17:30:42] <wikibugs>	 (03PS1) 10Halfak: Updates for revscoring==2.6.9 [research/ores/wheels] - 10https://gerrit.wikimedia.org/r/591114 (https://phabricator.wikimedia.org/T250536)
[17:32:48] <halfak>	 Hey folks.  I'd love a quick review of https://gerrit.wikimedia.org/r/#/c/research/ores/wheels/+/591114 if you have a chance. 
[17:32:57] <halfak>	 It's the wheels we need to move forward with a deployment. 
[17:35:13] <halfak>	 OK I'm running to lunch.  
[17:57:54] <wikibugs>	 (03CR) 10Accraze: [C: 03+2] Updates for revscoring==2.6.9 [research/ores/wheels] - 10https://gerrit.wikimedia.org/r/591114 (https://phabricator.wikimedia.org/T250536) (owner: 10Halfak)
[18:07:09] <accraze>	 posting our async update notes --
[18:07:31] <accraze>	 kevinbazira -
[18:07:33] <accraze>	 Last Week:
[18:07:35] <accraze>	     Focused on addressing UI issues identified in user-testing from https://phabricator.wikimedia.org/T247897
[18:07:37] <accraze>	         - Pushed a couple of patchsets
[18:07:39] <accraze>	 T:
[18:07:41] <accraze>	 Jade
[18:07:43] <accraze>	     Looking into rendering usernames in Jade edit comments: https://phabricator.wikimedia.org/T248135
[18:07:45] <accraze>	 MW Core
[18:07:47] <accraze>	     As advised by VolkerE on mw.notify, I have emphasized Base color (Base0) as foreground when setting the new background color and put the comments above each of the respective lines. 
[18:07:50] <accraze>	 halfak -
[18:07:52] <accraze>	 Last week: Did a lot of work for the tuning session.  Otherwise debugged memory usage issues in ORES and put together a capex proposal.  Otherwise, I onboarded Habeeb and tried to carry chtnnh's work to production. 
[18:07:54] <accraze>	 Today:  Well, I tried a beta deployment of ORES and everything is very weird.  See https://phabricator.wikimedia.org/T250712  I'll be digging into this today.  If I get past that, I'll get to work on the ORES paper rewrite that I'd been putting off for weeks now. 
[18:07:56] <accraze>	 haksoat -
[18:08:06] <accraze>	 Last week:
[18:08:07] <accraze>	   I spent a lot of time looking through the tokenizer modules on revscoring and the deltas package. I also did research on the possibility of using our generated regex match on elasticsearch.
[18:08:09] <accraze>	 Today:
[18:08:11] <accraze>	   I was able to come up with a short script that converts the 32-bit hex unicode values from Python into Java's supported 16-bit hex values. Currently, the regex pipe | doesn't seem to work yet, so I'm trying to understand why and fix it on the elasticsearch tokenizer.
[18:08:13] <accraze>	 and me -
[18:08:22] <accraze>	 Last week: Mostly continued work on secondary integrations for Jade and lots of code review.
[18:08:23] <accraze>	 Today: Continuing the ad hoc approach for jade 2ndary schemas and re-enabling hooks, also some more code review as well
[18:43:26] <chtnnh>	 hey halfak re: helders comments on the draftquality query
[18:43:50] <halfak>	 yeah. 
[18:44:05] <chtnnh>	 i checked out helders query, i think the 718 results include ER6 and ER20 in them
[18:44:13] <chtnnh>	 i ran another query to verify
[18:44:36] <halfak>	 Oh!  So we're already good with this query? 
[18:44:57] <chtnnh>	 https://quarry.wmflabs.org/query/44152
[18:45:08] <chtnnh>	 the thing is, we might be missing some of the results
[18:47:34] <halfak>	 Right.  I think that we need an "OR" condition to be able to catch this stuff too. 
[18:47:52] <chtnnh>	 yes exactly
[18:48:12] <chtnnh>	 also, the query you linked
[18:48:24] <chtnnh>	 https://quarry.wmflabs.org/query/43261
[18:48:34] <chtnnh>	 is not the one we linked in the Makefile 
[18:48:40] <halfak>	 Aha.  
[18:48:59] <chtnnh>	 let me show you the one we linked
[18:49:14] <wikibugs>	 (03PS1) 10Halfak: Bumps wheels for revscoring-2.6.9 [services/ores/deploy] - 10https://gerrit.wikimedia.org/r/591129
[18:50:04] <wikibugs>	 (03CR) 10Halfak: [V: 03+2 C: 03+2] "(Self-merging) a simple bump to wheels." [services/ores/deploy] - 10https://gerrit.wikimedia.org/r/591129 (owner: 10Halfak)
[18:52:13] <chtnnh>	 https://quarry.wmflabs.org/run/444292/output/0/json-lines?
[18:52:18] <chtnnh>	 halfak ^
[18:52:44] <halfak>	 That's the query output. 
[18:52:49] <halfak>	 It's not the query itself. 
[18:53:50] <halfak>	 Hmm.  I can't find the query we actually used. 
[18:55:06] <chtnnh>	 i have the query
[18:55:24] <chtnnh>	 let me link it up https://quarry.wmflabs.org/query/43197
[18:55:29] <chtnnh>	 its a fork of your query
[18:58:33] <halfak>	 I updated https://quarry.wmflabs.org/query/43197
[18:59:23] <halfak>	 What do you think?
[18:59:41] <halfak>	 Looks like we'll double our "spam" article count. 
[18:59:55] <chtnnh>	 we should ideally
[19:00:03] <chtnnh>	 the IF structure is killing me tho
[19:00:17] <chtnnh>	 no way we can OR this is in SQL?
[19:00:18] <chtnnh>	 xD
[19:01:38] <halfak>	 Our click dependency changed it's dang capitalization. 
[19:01:54] <halfak>	 chtnnh, not a nice way to do that :\
[19:02:51] <chtnnh>	 yeah i guess so
[19:02:54] <wikibugs>	 (03PS1) 10Halfak: Removes old Click dependency [research/ores/wheels] - 10https://gerrit.wikimedia.org/r/591136
[19:03:16] <halfak>	 accraze, around?  Looks like I missed something in that wheels change. https://gerrit.wikimedia.org/r/#/c/research/ores/wheels/+/591136
[19:03:30] <halfak>	 The capitalization changed so I didn't notice the old version there. 
[19:03:41] <accraze>	 ah ok, taking a look now halfak
[19:03:53] <halfak>	 Thanks!
[19:06:54] <wikibugs>	 (03CR) 10Accraze: [C: 03+2] Removes old Click dependency [research/ores/wheels] - 10https://gerrit.wikimedia.org/r/591136 (owner: 10Halfak)
[19:09:58] <haksoat>	 Some good news halfak
[19:10:28] <haksoat>	 I've found a way to tokenize on regex
[19:10:35] <haksoat>	 Tested with our regex
[19:10:52] <haksoat>	 Works, though, there are some things that still needs to be worked on
[19:11:14] <halfak>	 \o/
[19:11:30] <halfak>	 That's great news.  I wonder what kind of performance you are seeing. 
[19:14:47] <wikibugs>	 (03PS1) 10Halfak: Bumps wheels (remove double Click dep) [services/ores/deploy] - 10https://gerrit.wikimedia.org/r/591142
[19:15:33] <wikibugs>	 (03CR) 10Halfak: [V: 03+2 C: 03+2] "(self-merging) Bumping wheels to HEAD" [services/ores/deploy] - 10https://gerrit.wikimedia.org/r/591142 (owner: 10Halfak)
[19:16:09] <haksoat>	 I'm not running on the actual elasticsearch in the test environment, because it needs me to create an index, and I don't want to mess things up there
[19:16:18] <haksoat>	 Just using my local installation
[19:16:32] <haksoat>	 https://gist.github.com/HAKSOAT/16ca8ae47830cd51605f7f7c4c272ac1 the results halfak
[19:16:56] <haksoat>	 I intend polishing the regex more and also see how I can extract the named groups as well
[19:17:14] <haksoat>	 I currently don't know how to get that
[19:17:17] <halfak>	 That looks awesome.  Getting types eventually will be important. 
[19:17:26] <halfak>	 I see that there is a type. 
[19:17:37] <halfak>	 I wonder if ES can associate a token regex with a type
[19:17:41] <haksoat>	 Elasticsearch has its own type system
[19:17:56] <haksoat>	 But since I included named groups in the regex
[19:18:04] <haksoat>	 There should be a way to fetch it
[19:19:10] <haksoat>	 For performance, I'm yet to profile it since it's still a bit rough getting it to work as we want, but I think it's quite fast
[19:20:52] <rotuna>	 Hello!
[19:21:13] <halfak>	 How did this happen!?  I missed zipp too! 
[19:21:22] * halfak looks for duplicates again. 
[19:21:29] <halfak>	 Hello rotuna!
[19:21:37] <chtnnh>	 hello rotuna
[19:21:40] <halfak>	 ls
[19:26:01] <rotuna>	 I found this job opening [https://boards.greenhouse.io/wikimedia/jobs/2117176] and applied, I thought it would be good to talk to the current team to get know more about the team and the work you guys do.
[19:26:02] <wikibugs>	 (03PS1) 10Halfak: Removes double zipp [research/ores/wheels] - 10https://gerrit.wikimedia.org/r/591148
[19:26:45] <halfak>	 rotuna, yup!  That's us.  Here's our public face: https://www.mediawiki.org/wiki/Wikimedia_Scoring_Platform_team
[19:26:51] <halfak>	 Oh!  I need to add haksoat 
[19:27:10] <halfak>	 haksoat, do you have a photo of you up on commons? 
[19:27:29] <rotuna>	 Yup! I found this chat through that link. :D
[19:28:50] <rotuna>	 I can't really find the code though, I keep finding that seems to be the trello boards. 
[19:29:03] <halfak>	 Ha!  Those still exist!? 
[19:29:17] <halfak>	 See the "Projects" section :) 
[19:30:16] <haksoat>	 Not yet
[19:30:20] <halfak>	 rotuna, can you see this? https://github.com/orgs/wikimedia/teams/scoring-platform/repositories
[19:30:22] <haksoat>	 I could add a picture
[19:30:48] <rotuna>	 No, GitHub is trying to go Jedi on me 
[19:30:55] <halfak>	 Jedi? 
[19:31:19] <halfak>	 This is the board we work from: https://phabricator.wikimedia.org/tag/scoring-platform-team-current/
[19:31:25] <halfak>	 Here's our backlog: https://phabricator.wikimedia.org/tag/scoring-platform-team
[19:32:18] <rotuna>	 The 404 image is GitHub trying to tell me that isn't what I'm looking for. I've always seen that as Github trying to go Jedi on me [ Bad joke, sorry ] 
[19:32:27] <halfak>	 Oh!  Ha. 
[19:33:44] <halfak>	 There's really not a good way to find a list of our repos.  That's annoying. 
[19:34:20] <halfak>	 https://github.com/wikimedia/ores & https://github.com/wikimedia/revscoring represent the frameworks we are building on top of. 
[19:34:30] <halfak>	 We have repos for each type of model we support. 
[19:34:55] <halfak>	 https://github.com/wikimedia/editquality & https://github.com/wikimedia/articlequality & https://github.com/wikimedia/draftquality & https://github.com/wikimedia/drafttopic are examples. 
[19:34:57] <rotuna>	 Found it 
[19:35:28] <halfak>	 accraze, sorry to bother you again.  I don't know how I missed this one: https://gerrit.wikimedia.org/r/#/c/research/ores/wheels/+/591148
[19:35:33] <rotuna>	 Oh I just found ores
[19:35:37] <halfak>	 But I did a scan and I think that is the last one. 
[19:36:37] <wikibugs>	 (03CR) 10Accraze: [C: 03+2] Removes double zipp [research/ores/wheels] - 10https://gerrit.wikimedia.org/r/591148 (owner: 10Halfak)
[19:37:24] <halfak>	 accraze, I want to talk to you about a good way to run CI on our wheels.  These kind of issues should be easy to catch.  I'll make an item for the backlog :) 
[19:37:27] <halfak>	 More automation!
[19:37:46] <accraze>	 cool yeah was just going to say we should automate all of the wheels stuff
[19:38:59] <wikibugs>	 10Scoring-platform-team, 10Continuous-Integration-Infrastructure: CI should check to see if our wheels are good - https://phabricator.wikimedia.org/T250746 (10Halfak)
[19:39:13] <haksoat>	 halfak: For some reason, I can't log into commons.wikimedia.org for an upload
[19:39:25] <haksoat>	 But I can log into my regular wiki account
[19:39:53] <halfak>	 If you are already logged into your wiki account, you should be automatically logged into commons. 
[19:40:03] <halfak>	 Oh wait.  wikitech and commons are separate. 
[19:40:10] <halfak>	 wikipedia and commons use the same account. 
[19:40:12] <chtnnh>	 yeah
[19:40:19] <chtnnh>	 wikitech has a separate registration
[19:40:40] <haksoat>	 Haha
[19:40:44] <haksoat>	 Exactly
[19:40:54] <haksoat>	 Changed passwords now... I'm in
[19:41:49] <halfak>	 rotuna, ORES is pretty old tech at this point.  We're gearing up to revamp it on top of a nice open source framework. 
[19:42:06] <halfak>	 Right now we're looking at KubeFlow, but there are other frameworks we want to look at before we settle. 
[19:44:19] <wikibugs>	 (03PS1) 10Halfak: Bumps wheels to HEAD (rm double zipp) [services/ores/deploy] - 10https://gerrit.wikimedia.org/r/591151
[19:45:31] <wikibugs>	 (03CR) 10Halfak: [V: 03+2 C: 03+2] "(Self-merging) Bumps wheels to HEAD" [services/ores/deploy] - 10https://gerrit.wikimedia.org/r/591151 (owner: 10Halfak)
[19:52:08] <halfak>	 ITS ALIVE
[19:52:08] <halfak>	 https://ores-beta.wmflabs.org/v3/scores/ptwiki/7239
[19:52:47] <haksoat>	 halfak: https://commons.wikimedia.org/wiki/File:Habeeb_Shopeju_2020.jpg
[19:53:17] <halfak>	 Nice photo!  
[19:53:34] <haksoat>	 :) Thanks
[19:54:00] <chtnnh>	 great photo haksoat <3
[19:55:06] <halfak>	 haksoat, https://www.mediawiki.org/wiki/Wikimedia_Scoring_Platform_team#Team
[19:56:11] <haksoat>	 Thanks chtnnh
[19:57:04] <halfak>	 Looks like we are good on beta.  I'm going to try to send this to prod in 3 minutes. 
[19:59:11] <chtnnh>	 yay :D
[20:04:32] <halfak>	 Interesting that, while I'm doing this deploy, top on ores1001 reports 50GB free while grafan reports 22gb free. 
[20:05:53] <chtnnh>	 that sounds good
[20:05:55] <halfak>	 Looks like a lag of ~3 minutes. 
[20:06:08] <chtnnh>	 ouch
[20:20:00] <halfak>	 Make your user/common.js page look like this: https://pt.wikipedia.org/wiki/Usu%C3%A1rio(a):EpochFail/common.js
[20:20:06] <halfak>	 chtnnh, ^ 
[20:20:49] <halfak>	 It works!
[20:21:00] <chtnnh>	 user/common.js ?
[20:21:10] <halfak>	 User:Chtnnh/common.js
[20:21:15] <halfak>	 Or whatever you have as a username. 
[20:22:20] <chtnnh>	 i should add importScript("User:EpochFail/ArticleQuality.js") to my page?
[20:23:29] <chtnnh>	 yes i see ORES predicted quality
[20:23:31] <chtnnh>	 xD
[20:32:10] <halfak>	 \o/  
[20:33:28] <rotuna>	 I'm trying to look into this: https://phabricator.wikimedia.org/T242013 Could anyone tell me how to access stat1007.eqiad.wmnet?
[20:37:31] <chtnnh>	 halfak ^
[20:38:18] <halfak>	 rotuna, stat1007 is an internal server that only staff have access to.  Do you need a file discussed in the ticket? 
[20:38:52] <rotuna>	 It would be nice if you could send it to me, yes. 
[20:41:15] <halfak>	 Which file? 
[20:41:40] <halfak>	 Oh ha.  You want isaacj's code, probably. 
[20:42:02] <halfak>	 isaacj, any chance we could get you to host your work on https://phabricator.wikimedia.org/T242013 publicly? 
[20:43:23] <isaacj>	 halfak: yep, that's easy. I'll throw it into a repo and ping
[20:44:30] <halfak>	 Thanks!
[20:44:37] <isaacj>	 rotuna: just in case you sign off, i'll put it here: https://github.com/geohci/miscellaneous-wikimedia
[20:50:20] <rotuna>	 Thanks!
[20:50:37] <rotuna>	 Yeah, I just wanted the code 
[21:13:00] <isaacj>	 rotuna: https://github.com/geohci/miscellaneous-wikimedia/blob/master/topic-classification/topic_classification.md
[21:32:57] <wikibugs>	 10Scoring-platform-team (Research), 10revscoring, 10artificial-intelligence: Implement native NN model in revscoring - https://phabricator.wikimedia.org/T242013 (10Isaac) FYI -- I gathered the code for this and put it in this repository: [[https://github.com/geohci/miscellaneous-wikimedia/blob/master/topic-c...