[01:54:50] evening wikipedians! [13:38:07] o/ [14:19:05] o/ [14:21:55] Hey tarrow! Lots of discussions re. citation metadata (an your work on pubmed metadata) at the dev. summit. [14:22:03] Did you get CC'ed on some new stuff last week? [14:22:07] Awesome! [14:23:38] I saw a couple of phabricator things on new utils. Not taken a deep look yet though [14:23:55] Is there a link to the discussions / any conclusions people came to? [14:32:50] tarrow, I don't think there is a centralized summary yet. [14:33:04] We'll need to bother DarTar and harej about that when they get online. [14:33:31] ah cool, no worries. Just interested if I should be doing different things. :) [14:48:01] * halfak records all of the discussion he had in the last week. [14:48:08] I think that might be all I get done today. [15:39:20] * halfak waits eagerly for our new SE to join us today [15:57:34] \o [17:00:14] hi halfak [17:00:23] * YuviPanda is in sunnyvale [17:00:40] o/ [17:02:04] YuviPanda, what are you doing in sunnyvale? [17:03:26] sitting in a park listening to birds! [17:04:15] Cool! [17:04:16] :) [17:06:06] halfak: :D I was too late to catch the train with leila, so am chilling here waiting for the next one [17:06:43] halfak: I left some more comments / filed more issues on WikiChatter. recent changes on that library make it fail for some teahouse pages [17:07:17] Gotcha. Will be trying to catch up on documenting a bunch of dev summit/allstaff stuff today before I think about new work. [17:07:32] halfak: cool. we should try to get the wikichatter person on here [17:09:00] Yeah +1. I'm onboarding him today (for another project) so I'll include logging on here. :) [17:09:08] YuviPanda: Working on resolving those issues right now [17:09:26] Oh, hi also [17:09:42] haha hello kjschiroo :D [17:09:45] Woot! [17:10:02] * halfak ducks into meeting [17:10:25] https://tools.wmflabs.org/paws/public/YuviPanda/TeaHouse.ipynb is the notebook I'm using to test this [17:10:41] kjschiroo: welcome to the channel! [17:11:06] Thanks! [17:11:49] How does halfak do the "* halfak ducks into meeting" [17:18:55] kjschiroo: type /me somethingsomething [17:19:01] Also, hello :) [17:19:21] * kjschiroo trying it [17:19:27] neat [17:19:30] Hi [17:40:04] halfak: ping [17:40:59] * leila waves and says hi to those around. [17:41:17] morning from Chad's desk [17:41:36] morning Ironholds. :-) won't bug you unless you announce you want to be bugged. :D [17:41:44] leila, I am fine with being bugged! [17:41:53] ooki. :D heading there in 5 min Ironholds. [17:42:01] bahaha [18:01:35] YuviPanda: I'm getting really odd behavior on issue #16 you submitted to wikichatter. [18:01:42] o/ Guerillero [18:01:45] was making lunch. [18:01:49] Now heading to meeting :( [18:02:54] ok [18:03:03] ping me when you are libre [18:10:55] kjschiroo: sure, go on. what odd behavior? [18:12:11] YuviPanda: mwparserfromhell appears to care if you pass it the file to the text vs the text itself [18:12:20] aah [18:12:56] kjschiroo: is there a reason you're re-parsing wikitext a lot instead of just passing wikicode objects around? [18:13:00] YuviPanda: I posted an example on the issue. Would you be willing to take a quick look at it and see if you are getting the same results? It just seems really strange [18:13:08] sure looking [18:14:37] YuviPanda: I am concerned with the dependencies that it may introduce. Most of the time it doesn't make sense to think about it in wikicode objects so I don't want all the code to have to think of them as such [18:14:59] I really only need it to get sections and headings [18:20:10] kjschiroo: you're already heavily dependent on mwparser, and there are a lot of times when you convert wikicode to string, pass it to something, and then it reparses that into a wikicode object [18:22:27] kjschiroo: I can repro it too [18:22:36] kjschiroo: let's take it to ##earwig channel and ask earwig! [18:24:24] kjschiroo: there's also a nice-ish way to make sure you aren't forcing people to use wikicode objects without reparsing. let me make a small PR [18:25:07] How do I get to that channel? I'm new to ircs [18:25:39] kjschiroo: type [18:25:41] kjschiroo: /join ##earwig [18:29:13] YuviPanda: thanks! also PR? [18:29:57] Pull Request [18:31:06] poor kjschiroo is already being drowned in acronyms and earwigs [18:31:15] ha [18:44:18] kjschiroo: https://github.com/kjschiroo/WikiChatter/pull/18 [18:44:31] YuviPanda: looking at it right now [18:46:48] kjschiroo: that doesn't actualy fix the issue tho, just makes it more efficient :) [18:55:59] o/ schana [19:04:10] hey halfak :). You off in minnesnowta again? [19:04:44] Yup. And it's coooold. I heard that you east coasters got dumped on last weekend too. [19:04:53] yup. 8F before I left :/ [19:05:03] that's approximately 50F less than I'm comfortable with [19:05:06] Was -15F when I woke up this morning. [19:05:15] Dog still wanted to go outside. [19:05:17] :) [19:05:26] bragging ;p. [19:05:32] fine, I _won't_ show you my statistics tattoo [19:05:38] I live on mars. [19:05:46] Whaaa. Statistics tatoo? [19:05:48] Want [19:05:50] See [19:05:51] are you Matt Damon? [19:05:52] PLZ <3 [19:06:04] halfak, it's the PDF of the log-normal distribution :D [19:06:22] bearloga got the posterior probability statement for Bayesian theory, our friend Ted got the Poisson PDF, iirc [19:06:23] :D! I'm assuming it's not logged [19:06:34] not logged? [19:06:42] Otherwise it'll just look normal! [19:06:49] * bearloga groans [19:06:49] ohhh. GROAN :D [19:06:49] I suppose you could tell by the axis scaling [19:06:56] I mean it's the full function [19:07:01] wooo! First groan of the day! [19:07:20] halfak, it could be worse. Somebody pointed out it was actually a really bad distribution for me to like [19:07:24] unlike me, it's never negative [19:07:30] halfak: is groans per minute a kpi for you? [19:07:56] Now that's a good proposal. It's progress towards (dad|professor)-jokes [19:08:20] Ironholds, I lol'd & groaned [19:08:38] * Ironholds bows [19:09:08] o/ Ellery [19:09:44] Re. discussion modeling, I'd like to talk to you about some work that we're doing with (Probabilistic context-free grammars) PCFGs [19:10:12] (1) Build markov model of sentences that exhibit some property [19:10:43] (2) Run new sentences through state flow to obtain likelihood of observed sentence exhibiting said property [19:11:03] (3) Use as model itself or just source of signal for larger model [19:24:16] YuviPanda: Merged your PR. [23:43:42] we found a bug in mwparserfromhell! [23:43:45] this is exciting [23:44:52] \o/ [23:45:10] :D [23:48:28] YuviPanda: Finding that it's from hell, isn't a bug [23:51:54] Reedy, Ha! I actually filed a bug to rename it to "mwparser" because it is awesome. [23:53:28] halfak: what happened to it? :D [23:53:53] https://github.com/earwig/mwparserfromhell/issues/123 [23:53:56] "wont fix" [23:54:13] I even found the guy who owner an old package called "mwparser" who was willing to give up the name.