[01:49:47] YuviPanda, ping [01:53:49] Ironholds: pong [01:54:05] Going to sleep soon [01:54:06] Sup [01:59:09] YuviPanda, how do I access python3 on stat3? Alternately, do you wanna test something for me? :P [02:06:46] wait, worked it out [04:47:02] Ironholds: poke? [13:53:14] Ironholds: went to sleep [13:53:21] Ironholds: you can by just typing 'python3'? [14:09:19] Ugh... no virtualenv on stat1003 :( [14:09:29] How did this happen? [14:10:23] hey YuviPanda, how had would it be to get virtualenv installed quickly on stat1003? [14:10:31] or rather, is quickly possible? [14:10:47] halfak: good question. technically, it's a one line puppet patch [14:10:53] halfak: the question is if ops will allow it [14:11:06] We've been using virtualenv for years [14:11:11] The upgrade squashed it. [14:11:14] this is a regression. [14:11:29] it wasn't in the puppet manifest, so someone must've installed it by hand [14:11:36] I can make a patch, yeah [14:11:53] OK. I'll try to figure out how to bootstrap in the meantime. [14:27:28] halfak: cool :) I suspect ottomata would have to merge it, and I guess he won't be around till Monday, so I'll delay making the patch until then [14:27:54] OK. no worries. Thanks. [14:27:59] Shall I make an RT? [14:28:08] halfak: sure [14:43:50] halfak: btw, I've restarted doing my devlogs again http://yuvi.in/blog/category/devlog/ [14:44:37] Nice. Yours is fancier than mine. I do mine on a per-project basis. [14:44:50] https://github.com/halfak/Difference-Engine/blob/master/WORKLOG.rst [14:48:54] My worklog is my rubber ducky that collects my thoughts when simply thinking about things isn't working out for me. [14:50:08] :D [14:50:11] right [14:53:29] halfak: another idea I had was to make/ maintain a labsdb database that's wikidata in a more usable format [15:05:05] halfak: btw, quarry has reached 99 users :) [15:07:01] Nice! [15:07:16] :) Did it ever go out via @WikiResearch? [15:09:06] halfak: it did, yeah. [15:11:06] OK. I didn't think it did -- and that we could pull in even more users. [16:27:17] YuviPanda, I think that we need to get together with milimetric briefly to chat about indexes and labsdb early next week. [16:27:39] I don't want to propose that another index land just on analytics-store until we re-assess the situation [16:27:49] And I've got some that are going to be important. [16:33:47] * Ironholds groans [16:33:56] I just looked at my own code from 6 months ago. Why do I do this. [16:34:01] Why do I do the things I know will hurt. [16:34:26] (good news: my image composition library is now faster, can accept more varied input, has half the code length of the previous version, and contains an entire 10 lines of overlap). [16:40:07] halfak: cool and yeah totally [17:19:02] halfak, can I suggest a new section for the NDA doc? [17:20:23] which is "What do we give people NDAs over?" - the systems access versus specific datasets thing. [17:53:46] halfak: btw, if you remove quotes from .my.cnf, it works with pymysql [17:53:48] I'll file a bug [17:54:24] halfak: or it might be intended behavior, I'm unsure if you usually put quotes around these things. mysqldb works fine with it... [17:57:03] halfak: indeed, it might be intended behavior [18:22:27] YuviPanda, the quoted style works with the mysql client. [18:22:44] Strangely though, I thought the same thing and tried that earlier without success. [18:23:02] It seems that it's ConfigParser [18:23:08] It doesn't understand the quotes. [18:23:28] that's why I thought to try it when I first started using ConfigParser to parse the file manually. [18:23:39] * halfak tries again [18:24:21] halfak: pymysql also uses RawConfigParser, I wonder if that might cause problems [18:24:38] Never used that one [18:25:01] It would make sense though for the error [18:25:14] '""' != '' [18:28:40] yeah [18:33:56] Bah! Works now. [18:34:04] rm quotes == the win [18:34:19] I must have messed something up when I tried last time. [18:34:33] Cool. Good to know. [18:55:13] halfak: btw, I'm playing with ipython notebooks. Not over the web, but might end up with a simple setup for running kernels on tools and also publishing notbooks to tools [18:55:23] *notebooks [18:56:09] I need to pick up notebooks again. [18:56:18] Ggploit for python is starting to look solid. [18:56:35] I need to publish into the wiki though. [18:56:43] At least eventually [18:57:04] So long as I could export svgs, I could still work in ipython [18:58:31] halfak: bokeh was quite nice as well [18:58:53] I played around a bit with bokeh. At the time, it was missing some critical functionality. [21:50:27] halfak: ah, I see. I haven't played around with any of them [21:51:03] I'm going to have a hard time moving away from data.tables [21:51:09] They are pretty freakin' awesome. [22:12:46] YuviPanda, https://meta.wikimedia.org/wiki/Research:Revision_scoring_as_a_service [22:13:05] Similar to your idea for a revert detection service. [22:13:16] I think that the revert detection service would be much easier. [22:13:43] We could wrap it up into the same service. [22:13:53] * halfak needs to go to bed. [22:13:56] o/ [22:15:46] halfak: \o/ good night [22:36:44] * Ironholds yawns