[20:02:37] <aetilley>	 halfak: I'm not sure if you want to be bothered with such things, but the last sentence of the docstring for class Dependent is clipped.
[20:02:59] <halfak>	 aetilley, a wonderful opportunity for a pull request. 
[20:03:05] <aetilley>	 haha
[20:03:06] <halfak>	 Do you know how the sentence ought to end?
[20:03:08] <YuviPanda>	 halfak: http://worrydream.com/ClimateChange/ is... really intense.
[20:03:43] <aetilley>	 "Dependents"?
[20:04:38] <halfak>	 aetilley, can you link?
[20:04:45] <aetilley>	 or rather the canonical path to class Dependents, which I admit I'm frogetting the form for.
[20:05:03] <aetilley>	 link what?
[20:05:25] <aetilley>	 halfak: ^
[20:05:42] <halfak>	 where the doc cuts off
[20:06:52] <aetilley>	 I'm not sure what you're asking for.  May I just don't know how to do what you're asking.
[20:07:00] <halfak>	 Where are the docs bad?
[20:07:17] <aetilley>	 Line 24 of dependent.py
[20:07:18] <halfak>	 Got it https://github.com/wiki-ai/revscoring/blob/master/revscoring/dependencies/dependent.py#L24
[20:07:59] <aetilley>	 In the future I'm happy to make a pull request for something like this.
[20:08:14] <aetilley>	 I wasn't sure if you had a running list of minor errata that you liked to deal with in batch.
[20:08:54] <halfak>	 aetilley, na.  But I did just fix the issue you pointed out :) 
[20:09:40] <aetilley>	 Amir1: I've started looking at Kian but it's slow going.  Where would you recommend I start?
[20:15:31] * aetilley is realizing he needs to spike neural networks.
[23:22:15] <Vito>	 (crossposting) I have a simple question: did you use abusefilters to get (in some way) badwords and heuristics for ORES?
[23:22:48] <YuviPanda>	 ToAruShiroiNeko: halfak ^
[23:23:09] <halfak>	 o/  in meeting.  will respond soon
[23:23:37] <Vito>	 no pb, it's anything but urgent
[23:29:48] <halfak>	 OK.  So.  We don't use abuse filter rules straight up, but I would like to extend our badwords detection with abuse filter rules. 
[23:30:16] <halfak>	 FWIW, we used to use a stemming/sets based detection strategy, but now we use regexes which are easily compatible with abuse filter. 
[23:30:21] <halfak>	 Vito, ^ 
[23:30:40] <halfak>	 Vito, do you do much work with abuse filter?
[23:30:55] <Vito>	 halfak: great, abusefilter should be a pit of informations
[23:31:01] <halfak>	 +1
[23:31:10] <halfak>	 Here's the format that we capture badwords in: https://github.com/wiki-ai/revscoring/blob/master/revscoring/languages/english.py#L28
[23:31:16] <Vito>	 yep, I manage most of filters at it.wiki
[23:31:27] <halfak>	 We also do informals: https://github.com/wiki-ai/revscoring/blob/master/revscoring/languages/english.py#L110
[23:31:43] <halfak>	 Oh great!  I wonder if you would be able to help us merge in what you've got. 
[23:31:54] <halfak>	 https://github.com/wiki-ai/revscoring/blob/master/revscoring/languages/italian.py
[23:32:21] <halfak>	 We made these regexes by processing the history if itwiki and looking for words added by edits that tend to get reverted. 
[23:34:45] <halfak>	 brb.  need to let the dog out. 
[23:34:46] <Vito>	 https://it.wikipedia.org/wiki/Speciale:FiltroAntiAbusi/45 <-- is full of informals
[23:34:55] <Vito>	 see ay!
[23:34:57] <Vito>	 *ya
[23:53:33] <halfak>	 Booo.  I don't have the perms to view the rules. 
[23:53:37] <halfak>	 I'll just query the DB :) 
[23:59:59] <YuviPanda>	 heh