[21:54:55] <aetilley>	 halfak: So for what it's worth I seriously Overestimated the time it would take me to code up this algorithm.
[21:55:26] <aetilley>	 I started on the "hard thresholding" variant after breakfast this morning and I'm pretty much done.
[21:55:29] <halfak>	 yay!
[21:55:34] <halfak>	 Woot!
[21:55:41] <aetilley>	 I need to test and debug it, but I should have something for you by tomorrow.
[21:55:51] <halfak>	 Great news.
[21:55:58] <aetilley>	 numpy is awesome.
[21:56:06] <halfak>	 Except for the compile time ;)
[21:56:15] <aetilley>	 ah
[21:59:32] <ToAruShiroiNeko_>	 indeed
[21:59:42] <ToAruShiroiNeko_>	 numpy is amazing but it has big upkeep :p
[21:59:47] <ToAruShiroiNeko_>	 hello halfak
[22:00:05] <ToAruShiroiNeko_>	 [2015-10-14, 20:24:34] * +halfak has a look.
[22:01:00] <halfak>	 ToAruShiroiNeko_, yeah.  Got some relevant revids loaded into my sample data, but I didn't get any tests in place yet.
[22:01:12] <halfak>	 I've gotta demonstrate the problem and then I'll work on the solution.
[22:40:01] <White_Cat>	 oh ok
[22:40:21] <White_Cat>	 I was trying to ask if the tr campaign is realy done
[22:58:38] <halfak>	 Arg.  ToAruShiroiNeko_ ping me if you want me to reply!
[22:59:29] <halfak>	 I haven't had a chance to run the query for trwiki yet.  Been focused on revscoring productionization stuff and other work.