[01:01:14] YuviPanda or legoktm, have a look at https://github.com/mediawiki-utilities/python-mwapi/pull/27? [01:01:32] I'm going to start using this all the time! [01:06:23] halfak: it's always JSON right? [01:06:29] That's right [01:06:34] We only support JSOn [01:07:10] We'll actually overwrite the format if you try to set it [01:09:34] heh :D [08:42:32] Hello halfak and all! [13:19:58] Extension:WikiLabels day! i'll be bother you guys with questions, hopefully. [13:20:02] *ing [13:30:27] \o/ [13:30:31] Yay! [13:30:47] bmansurov, I'm heading into the university, but I'll be back online in about 1 hour [13:30:58] cool [14:23:29] legoktm: hey, do you have about five minutes? [14:53:56] o/ [14:57:26] halfak: do we have the less version of https://labels.wmflabs.org/gadget/WikiLabels.css ? [14:57:54] or was it just css? [14:57:57] bmansurov, less? [14:58:01] :) Just CSS [14:58:10] http://lesscss.org/ [14:58:13] ok [14:58:19] Is less becoming a standard? [14:58:36] hmm, hard to say, but it's certainly easier to maintain than css [14:58:51] i think mediawiki and its extensions generally use less [14:59:04] I'm happy to convert then [14:59:24] * halfak has opinions, but opinions should always yield to conventions. [14:59:37] halfak: o/ [14:59:44] o/ Amir1 [14:59:46] I've so many things to discuss [15:00:01] halfak: ok, i'll use less in the extension [15:01:04] Amir1, that PR. It looks wrong. [15:01:10] I'm not sure I see the problem it is solving [15:01:12] Today I started with working on the dump extractor, at first I couldn't run it at all, after several hours of investigation it turned out one blank line was missing from the docsctring [15:02:03] because I was couldn't copy it from raw (the server was blocked in Iran) I copied from github page directly and it deleted one line [15:02:33] so I made a PR in the docopt to warn properly in these cases [15:03:19] https://github.com/docopt/docopt/pull/306 then it seemed --start and --end doesn't work, so I checked and there was a bug in our code: https://github.com/wiki-ai/editquality/pull/5 [15:03:37] otherwise it gets every edit no matter of timestamp [15:03:53] so now the result is ready [15:04:05] I put it in the phab card [15:04:41] then I worked with ORES GUI, I added proper error handling [15:04:44] it's okay now [15:05:21] halfak: also, is https://labels.wmflabs.org/gadget/WikiLabels.js just one file in development or are you concatenating multiple files and outputting them as one? [15:05:28] I want to clean up the code. but I want to ask your opinion halfak, Which one is better: Pure JS (jquery included) or JS /PHP? [15:05:29] Amir1, that window is a deque. It should only fill with a max of `revert_radius` revisions. [15:05:57] strangely it gets everything, you can test it [15:06:33] I even thought maybe the deque isn't properly invoked (e.g. append and pop instead of append and popleft) [15:06:38] but it didn't work out [15:07:27] maybe the system throw out newer edits [15:08:02] bmansurov, concat [15:08:10] See wikilabels/wsgi/static/js [15:08:42] Amir1, I have a very uninformed opinion, but JS/PHP sounds awful. [15:09:02] okay :) [15:09:06] Amir1, I think that the problem might be elsewhere. (other than the window) [15:09:09] I'll have a look. [15:10:03] thanks halfak [15:10:32] halfak: sorry, is that a path of some url? [15:10:54] It's for the wikilabels repo [15:11:41] lastly I was working on this bug in ORES extension playing nice with "r", it turned out the ores tables are not being populated again... [15:12:14] spent about one hour one that and I couldn't figure out it why [15:12:35] halfak: thanks [15:12:37] but I tested my patch with a simple injection in the database [15:12:42] and it works [15:13:01] the "r" flag won't be shown if the edit is marked as patrolled [15:13:26] I'm about to push the change into gerrit [15:13:34] but I need someone to review it [15:14:06] (sorry, sort of in and out for the next 30 minutes [15:14:18] Working through messages slowly [15:14:29] it's okay halfak [15:14:43] I can check again window bug again for you [15:15:26] I'm so frustrated with lots of bugs [15:15:28] Amir1, I'm second guessing myself. I see where the issue is. [15:15:40] The whole window gets dumped even if there are not revisions within the bounds. [15:15:50] I like your solution better now. :) [15:15:56] I think I'll merge it. [15:16:03] \o/ [15:18:33] finding bugs in mediawiki is really hard, their logs in vagrant is huge [15:22:08] (03PS1) 10Ladsgroup: Don't show the r flag when edit is patrolled [extensions/ORES] - 10https://gerrit.wikimedia.org/r/264293 [15:30:55] I go for a walk [15:50:16] Hi to halfak and the channel. [15:51:28] o/ pipivoj [15:52:00] o/ back to you [15:53:06] I don't know is there some prerequisite background experience for people to join the team. [15:53:45] I know I haven't been exactly verbose about mine, but if necessary I can be. [15:53:48] :) [15:55:14] When talking about the team, I don't mean the formal enlisting on some page at wiki, just offer my pair of hands (and the brain) to chip in with the help. [15:58:55] * halfak is in and out. should be back soon [16:04:04] (03PS1) 10Ladsgroup: Fix parsing results from ORES API [extensions/ORES] - 10https://gerrit.wikimedia.org/r/264295 [16:04:25] halfak: this was the bug that made me crazy [16:07:13] pipivoj, no requirements for contributing [16:07:19] But please be verbose [16:07:21] :) [16:08:51] Ok. At university I had one semester (one course) of "Pattern recognition" and my graduate thesis was SOM Toolbox (for Matlab) - 11 years ago. [16:10:11] But recently I've been completing the Andrew Ng's "Machine learning" at Coursera. 9 1/2 weeks out of 11 finished. [16:12:31] As for Python I am (mostly) able to read and understand the code - I've read the Dive into Python by Mark Pilgrim. [16:14:03] As for other I am versed especially in C, as well as C# and Java. JavaScript a bit. [16:17:49] pipivoj, what about interests? It seems like you might be interested in modeling stuff (as opposed to backend engineering or UX stuff) [16:18:10] Also, what's the SOM toolbox do? [16:18:54] Self Organizing Maps - Kohonen maps are similar to it if I recall. [16:19:52] As far as my work at thesis has been: it�s good for visualizing multidim data. [16:21:28] And yes, you got me there. I would be quite interested in modelling but not sure about what do you have in mind, since I'm not in sync with ML parlance. [16:21:58] What kind of modelling are you referring to? [16:22:43] Mostly, we're working with classifiers. There's a lot of work we do around engineering new ways to get features (predictors) with high signal and evaluation -- how we know the model works. [16:23:44] Yes, that's what I was thinking you were talking about. [16:26:07] Anyway, have to go now. Have quite a lot IRL obligations ATM, so will come later tonight (I'm UTC+1). Is 1:30 UTC ok? [16:27:51] * pipivoj will try be here at 1:30utc [16:50:12] * halfak becomes substantially more available for the next 40 minutes. [16:55:01] halfak: so you're online now? [16:55:11] Yes [16:55:22] Did I miss anything that you wanted me to look at? [16:55:28] I saw that you have data for urdu [17:00:11] Yeah I have it [17:00:16] it's in the phab card [17:00:18] :) [17:05:49] YuviPanda: hey, do you know how someone can get +2 right in a certain extension? [17:06:32] Images of the cerebration: https://commons.wikimedia.org/wiki/Category:Wikipedia_15_in_Iran [17:07:08] 15 candles even! [17:10:51] we were actually looking for "1" and "5" candles but we couldn't find it in the proper time [17:11:00] so we took 15 candles [17:14:47] Soon, wikipedia will become old enough that we *can't* have enough candles on the cake and still expect to eat it. [17:20:05] * halfak starts working on etherpad for meeting [17:20:08] https://etherpad.wikimedia.org/p/revscoring [17:20:39] * Amir1 hopes for that day [17:31:11] Hey [17:32:12] halfak yo [17:33:13] Hey [18:24:21] Yay [18:25:04] Sorry for the background noise if any. :-) [20:36:51] halfak: please tell me when the meeting with Lila is over [20:36:55] :) [20:37:06] Will do. In it now. [20:37:24] good luck [21:41:30] (03PS2) 10Ladsgroup: Don't show the r flag when edit is patrolled [extensions/ORES] - 10https://gerrit.wikimedia.org/r/264293 (https://phabricator.wikimedia.org/T121589) [22:22:43] (03CR) 10Legoktm: [C: 032] Fix parsing results from ORES API [extensions/ORES] - 10https://gerrit.wikimedia.org/r/264295 (owner: 10Ladsgroup) [22:22:50] halfak: Status: Not totally fubar [22:22:51] https://github.com/aetilley/pcfg/tree/OOP [22:23:29] \o/ [22:24:09] (03Merged) 10jenkins-bot: Fix parsing results from ORES API [extensions/ORES] - 10https://gerrit.wikimedia.org/r/264295 (owner: 10Ladsgroup) [22:24:51] (03CR) 10Legoktm: "Quick style comment, code looks fine but I haven't tested" (031 comment) [extensions/ORES] - 10https://gerrit.wikimedia.org/r/263184 (https://phabricator.wikimedia.org/T122535) (owner: 10Ladsgroup) [22:25:43] Amir1: you should be able to +2 now, might have to log in and out first though [22:26:07] thanks legoktm :) [22:27:33] Yes, I can +2 now [22:27:43] * Amir1 is dancing [22:28:17] \o/ [22:40:50] (03PS4) 10Ladsgroup: Add onOldChangesListRecentChangesLine hook [extensions/ORES] - 10https://gerrit.wikimedia.org/r/263184 (https://phabricator.wikimedia.org/T122535) [22:42:11] legoktm: is it okay now? https://gerrit.wikimedia.org/r/#/c/263184/3..4/includes/Hooks.php,unified [22:46:05] I've got to run, but I'll be back in a few hours to catch up with pipivoj. [22:46:11] o/ [22:57:39] I really need to sleep [22:57:43] o/ [23:50:38] Hello [23:51:19] o/ halfak and others