[00:04:53] 10Scoring-platform-team, 10ORES, 10articlequality-modeling, 10artificial-intelligence: articlequality makefile has hardcoded path - https://phabricator.wikimedia.org/T192324#4134905 (10awight) [00:06:28] 10Scoring-platform-team, 10ORES, 10articlequality-modeling, 10artificial-intelligence: articlequality makefile has hardcoded path - https://phabricator.wikimedia.org/T192324#4134916 (10awight) There are other assumptions that prevent building anywhere but stat1005.eqiad.wmnet [00:21:05] gtg for a few hours. [03:19:07] (03PS1) 10Awight: Rename articlequality; remove repos from requirements [services/ores/deploy] - 10https://gerrit.wikimedia.org/r/427043 [13:38:29] o/ [13:50:21] 10Scoring-platform-team (Current), 10editquality-modeling, 10artificial-intelligence: Nulls appear in labeled data (merge_labels issue) - https://phabricator.wikimedia.org/T192362#4136034 (10Halfak) [13:50:52] 10Scoring-platform-team (Current), 10editquality-modeling, 10artificial-intelligence: Nulls appear in labeled data (merge_labels issue) - https://phabricator.wikimedia.org/T192362#4136047 (10Halfak) @awight, can you have a look at this? [15:42:22] https://www.mediawiki.org/wiki/Topic:Ubg7ie7lui72gbpd -- ORES open transparent process/system [15:51:45] akosiaris, it looks like both Adam and I will be AFK for much of next week. [15:51:59] Monday, Tuesday, Wednesday, Thursday at least. [15:52:07] So you might be on the hook if ORES explodes. [15:52:24] Amir1 should be around for Monday and Wednesday. [16:33:15] Meanwhile, the ORES updates are not at all ready to deploy... [16:51:38] How do I get to phabricator from stat100*…. I tried https?_proxy with no luck. [17:13:49] halfak: ^ have any clues? [17:20:57] 10Scoring-platform-team: Drafttopic: add article text fetching utility - https://phabricator.wikimedia.org/T183355#4136839 (10Sumit) 05Open>03declined [17:21:17] 10Scoring-platform-team: Drafttopic: Add utility to extract dependents - https://phabricator.wikimedia.org/T183392#4136841 (10Sumit) 05Open>03declined [17:29:39] OK I see how you manage this—you use the repos directly and not the phabricator submodules. [17:35:29] 10Scoring-platform-team (Current), 10editquality-modeling, 10User-Ladsgroup, 10User-Tgr, 10artificial-intelligence: Train/test damaging and goodfaith model for Hungarian Wikipedia - https://phabricator.wikimedia.org/T185903#4136904 (10Bencemac) @awight Thanks, I am looking forward to! :) [17:41:03] Sorry awight_ [17:41:11] forgot to say I was AFK for lunch [17:42:13] no worries! [18:17:32] Officially back now [18:17:52] Been going through emails :| [18:26:28] This is the first time I’ve built the other models, it seems… bit of a learning curve. [18:27:32] 10Scoring-platform-team (Current), 10articlequality-modeling, 10User-Ladsgroup, 10artificial-intelligence: Train and test wp10 model for fawiki - https://phabricator.wikimedia.org/T190050#4137249 (10Halfak) This is somehow missing a ton of observations for the lower classes. Those should come out of one o... [18:41:49] Argh, I thought I had “deletedtext” on enwiki, but I don’t see how that would have been possible now. [18:41:57] Without it, I can’t build draftquality [18:42:53] * awight_ copies file from ~halfak [18:43:06] more fun [18:43:07] ERROR 1045 (28000): Access denied for user 'research'@'10.64.53.30' (using password: YES) [18:43:38] I guess the password changed recently? [18:46:34] awight_, huh? [18:46:42] Are you trying to log into the mysql slaves? [18:47:02] If so, you should be using the common file. I forget where it is. Maybe /etc/mysql [18:51:05] This is coming from a makefile rule, [18:51:05] mysql -h analytics-store.eqiad.wmnet -u research enwiki | bzip2 -c > datasets/enwiki.draft_quality.201606.tsv.bz2 [18:51:06] ERROR 1045 (28000): Access denied for user 'research'@'10.64.53.30' (using password: YES) [18:52:27] cool, thanks for the pointer to /etc, the pw was in there. [21:47:50] 10Scoring-platform-team (Current), 10editquality-modeling, 10artificial-intelligence: Nulls appear in labeled data (merge_labels issue) - https://phabricator.wikimedia.org/T192362#4137768 (10awight) Confirmed broken. This line in human_labeled, datasets/enwiki.human_labeled_revisions.20k_2015.json:{"aut... [22:00:26] https://meta.wikimedia.org/wiki/Research:ORES:_Facilitating_re-mediation_of_Wikipedia%27s_socio-technical_problems [22:00:53] OK I have most of the references formatted somewhat reasonably. I think I can clearly see the two biggest holes in the paper: (1) related work and (2) conclusions. [22:01:33] I'd also like to add a nice and clean example of an ORES prediction request. [22:01:53] But that will wait until tomorrow. For now, I must go pedal hard. [22:01:54] o/ [22:07:47] neat! [22:48:08] 10Scoring-platform-team, 10ORES: Update deprecated cv_train usage - https://phabricator.wikimedia.org/T192412#4137856 (10awight) [23:45:55] 10Scoring-platform-team (Current), 10editquality-modeling, 10artificial-intelligence: Nulls appear in labeled data (merge_labels issue) - https://phabricator.wikimedia.org/T192362#4138026 (10awight) It's an edge case that can only happen when no autolabeled file is given, and we're only passing human labeled...