[07:23:16] PROBLEM - puppet on ORES-web01.Experimental is CRITICAL: CRITICAL: Catalog fetch fail. Either compilation failed or puppetmaster has issues [07:52:41] RECOVERY - puppet on ORES-web01.Experimental is OK: OK: Puppet is currently enabled, last run 5 seconds ago with 0 failures [08:54:11] PROBLEM - puppet on ORES-web01.Experimental is CRITICAL: CRITICAL: Puppet has 8 failures. Last run 1 minute ago with 8 failures. Failed resources (up to 3 shown): Exec[set debconf flag seen for wireshark-common/install-setuid],Service[prometheus-node-exporter],Exec[apply_tc_config],Exec[enable_sites_local] [09:34:16] PROBLEM - puppet on ORES-web01.Experimental is CRITICAL: CRITICAL: Puppet has 17 failures. Last run 10 minutes ago with 17 failures. Failed resources (up to 3 shown): Package[ntp],Service[systemd-timesyncd],Service[exim4],Package[update-manager-core] [09:53:41] RECOVERY - puppet on ORES-web01.Experimental is OK: OK: Puppet is currently enabled, last run 38 seconds ago with 0 failures [13:22:31] o/ [13:23:34] halfak [13:23:41] * Vermont pings halfak [13:34:21] Vermont ^^ [13:34:45] halfak [13:34:52] o/ [13:35:34] What's the deal with cloud vps puppet? [13:56:57] halfak: Amir1: https://github.com/marcotcr/lime [13:58:13] nice [14:30:35] 10Scoring-platform-team (Current), 10Scap, 10Patch-For-Review: [blocked] Support git-lfs - https://phabricator.wikimedia.org/T180627#4080752 (10Halfak) [14:31:58] 10Scoring-platform-team (Current), 10ORES, 10articlequality-modeling, 10draftquality-modeling, and 2 others: Discuss surfacing ORES for AFC/NPP - https://phabricator.wikimedia.org/T190464#4080756 (10Halfak) a:03Halfak [14:32:49] 10Scoring-platform-team, 10drafttopic-modeling, 10revscoring, 10artificial-intelligence: Investigate runtime of tune with high number of estimators - https://phabricator.wikimedia.org/T190288#4080762 (10Halfak) [14:33:06] 10Scoring-platform-team (Current), 10editquality-modeling, 10artificial-intelligence: Train/test damaging & goodfaith models for Latvian Wikipedia - https://phabricator.wikimedia.org/T163006#4080767 (10Halfak) a:03Halfak [14:33:16] 10Scoring-platform-team (Current), 10editquality-modeling, 10artificial-intelligence: Train and test damaging/goodfaith model for arwiki - https://phabricator.wikimedia.org/T189710#4080769 (10Halfak) a:03Halfak [14:35:17] 10Scoring-platform-team (Current), 10Scap, 10Patch-For-Review: [Blocked] Support git-lfs - https://phabricator.wikimedia.org/T180627#4080777 (10Halfak) [14:36:04] 10Scoring-platform-team, 10Scap, 10Patch-For-Review: [Blocked] Support git-lfs - https://phabricator.wikimedia.org/T180627#3764014 (10Halfak) [14:37:05] 10Scoring-platform-team (Current), 10ChangeProp, 10ORES, 10Services (done): Change ORES rules to send all events to new "/precache" endpoint - https://phabricator.wikimedia.org/T158437#4080785 (10Halfak) [14:37:07] 10Scoring-platform-team (Current), 10ORES: Investigate 3x scoring load after ORES changeprop migrated to new precache endpoint - https://phabricator.wikimedia.org/T187928#4080784 (10Halfak) 05Open>03declined [14:38:08] 10Scoring-platform-team (Current), 10ChangeProp, 10ORES: Drop "non bot" condition from ORES changeprop rules - https://phabricator.wikimedia.org/T187927#4080787 (10awight) a:03awight [14:38:27] 10Scoring-platform-team, 10WMF-Communications, 10artificial-intelligence: Pitch: How ML saved Wikipedia - https://phabricator.wikimedia.org/T187210#4080788 (10Halfak) [14:39:49] 10Scoring-platform-team, 10JADE, 10Wikimedia-Extension-setup, 10Wikimedia-extension-review-queue: Extension review for JADE - https://phabricator.wikimedia.org/T188307#4080796 (10awight) [14:40:17] o/ Amir1 [14:40:23] MONITOR THE COMMUNITY [14:40:28] hello! [14:40:29] or MONITORING FOR THE COMMUNITY [14:40:53] Monitoring the community so we can ping them if they are stalled :D [14:41:06] I was thinking of "dashboard" too [14:42:08] +1 [14:42:15] I like dashboard. [14:42:26] ORES support dashboard. [14:42:40] That sounds nice [14:43:51] The other thing: Should it such a huge thing having its own VPS (and built by flask) or something very simple (probably written in php) [14:44:01] Dashboard sounds kinda real-time to me. What about "ORES support checklist" so that it's more obvious that the goal is to complete support? [14:44:29] another thing: should we connect to wikilabels database to get data or we use (and build) API in wikilables to support that [14:44:52] 10Scoring-platform-team, 10JADE, 10Wikimedia-Extension-setup, 10Wikimedia-extension-review-queue: Extension review for JADE - https://phabricator.wikimedia.org/T188307#4080833 (10awight) [14:46:57] 10Scoring-platform-team, 10JADE, 10Wikimedia-Extension-setup, 10Wikimedia-extension-review-queue: Design review for JADE - https://phabricator.wikimedia.org/T190690#4080835 (10awight) [14:47:21] Amir1, I think you can make a think flask app in ToolForge [14:47:26] 10Scoring-platform-team, 10JADE, 10Wikimedia-Extension-setup, 10Wikimedia-extension-review-queue: Extension review for JADE - https://phabricator.wikimedia.org/T188307#4080845 (10awight) [14:47:37] Using our public APIs should work [14:47:48] If not, we should extend our public APIs [14:47:55] awight, +1 for checklost [14:47:57] *i [14:48:04] checklost lol [14:48:07] ftw [14:48:17] hmm, okay [14:49:53] * halfak looks at awight's pull requests [14:50:22] halfak: https://github.com/mediawiki-utilities/python-mwreverts/pull/10 :D [14:50:43] 10Scoring-platform-team, 10revscoring, 10artificial-intelligence: Document revscoring pop-rate parameter - https://phabricator.wikimedia.org/T190691#4080849 (10awight) [14:52:18] 10Scoring-platform-team: Write JADE talk presentation for Wikimania - https://phabricator.wikimedia.org/T190692#4080860 (10awight) [14:53:01] Amir1, *merged* [14:54:57] Thank you! [14:56:42] 10Scoring-platform-team: Write JADE talk presentation for Wikimania - https://phabricator.wikimedia.org/T190692#4080874 (10awight) a:03awight [14:59:10] awight, looks like https://github.com/wiki-ai/editquality/pull/145 is failing [14:59:18] But maybe that was a temp travis issue? [15:00:18] 10Scoring-platform-team, 10JADE, 10ORES: Read and present JADE data in ORES - https://phabricator.wikimedia.org/T190472#4080888 (10awight) [15:03:02] 10Scoring-platform-team (Current), 10Research Ideas, 10artificial-intelligence: [Epic] Paid editing (COI) detection model - https://phabricator.wikimedia.org/T120170#4080892 (10awight) [15:04:40] halfak: Looks like a real failure, ty [15:05:19] Unfortunately, the test passes for me locally. I think it's an unsorted list comparison. [15:37:08] halfak: OK, the latest patch ^ should fix PR#145 [15:39:33] wiki-ai/editquality#256 (isolate_codegen - a88d650 : Adam Wight): The build was fixed. https://travis-ci.org/wiki-ai/editquality/builds/358466556 [15:49:10] 10Scoring-platform-team (Current), 10User-Ladsgroup: Build ORES support checklist - https://phabricator.wikimedia.org/T189954#4080977 (10Ladsgroup) [15:50:09] Amir1: Can I ask a favor, to check the puppet staleness and maybe force a run on ores-misc-01? [15:50:28] I added a new SSH pubkey on Thursday, but it's still not working... [15:50:32] sure [15:50:36] ty! [15:50:56] apergos: Hey, we don't make "All pages with complete page edit history " (combined) for wikidatawiki anymore? [15:51:14] I need it for https://phabricator.wikimedia.org/T189962 [15:51:16] not for some time [15:51:25] pick up the separate files, they have all the content [15:51:37] yeah [15:52:12] just do the same to them that you would to the one file: get rid of the header and footer and use the page content however [15:52:20] awight: it says puppet has been ran one minute ago [15:52:46] apergos: yeah, I'm feeding them inside a bash loop [15:52:52] ok [15:52:56] hargh [15:53:01] awight: it ran without any errors [15:53:10] bear in mind if your data source can take multiple files at once you could feed them in parallel [15:53:31] if it takes multipel writes at once I should day [15:53:32] say [15:53:45] let me check [15:54:02] hmm, it accepts that [15:54:03] nice [15:54:08] win! [15:54:23] each of them has header and footer just like the combined file did [16:28:10] awight: https://gerrit.wikimedia.org/r/#/c/421196/ :D [16:33:27] loonch [16:38:42] 10Scoring-platform-team, 10Cleanup, 10Gerrit, 10Repository-Admins, 10Release-Engineering-Team (Watching / External): Investigate and clean up deprecated ORES repos - https://phabricator.wikimedia.org/T179610#4081256 (10greg) [16:44:11] halfak|Lunch: for when you're back, I just realized ores scorer (ui thingy) is wikimedia-ui (old thing) ewwww. Do we have a ticket to use WikimediaUI? [16:44:49] -._o_.- [16:44:57] Amir1, I think you implemented that :P [16:45:28] Stupid young Amir did that, I'm someone else now :D [16:47:30] Amir1: CR! thanks for the reminder [16:49:55] (03CR) 10Awight: [C: 032] "Great, thanks!" [extensions/ORES] - 10https://gerrit.wikimedia.org/r/421196 (https://phabricator.wikimedia.org/T189948) (owner: 10Ladsgroup) [16:50:34] Thank you! [16:57:42] halfak|Lunch: I just saw that ACL response period is now open [16:58:27] plz forward [16:59:14] (03Merged) 10jenkins-bot: Clean up old config style [extensions/ORES] - 10https://gerrit.wikimedia.org/r/421196 (https://phabricator.wikimedia.org/T189948) (owner: 10Ladsgroup) [17:03:13] (03CR) 10jenkins-bot: Clean up old config style [extensions/ORES] - 10https://gerrit.wikimedia.org/r/421196 (https://phabricator.wikimedia.org/T189948) (owner: 10Ladsgroup) [17:06:42] I go grab something to eat and come back soon [17:38:06] awight, are you still working on huwiki? [17:40:24] awight, I'm wondering if I should work on arwiki and lvwiki now or if we should finish our discussion re. simplifications to codegen. [17:40:26] What do you think? [18:20:19] halfak: sorry, I'm blocked on huwiki by a mysteriously malfunctioning new ssh key for wmflabs. [18:20:57] Discussion sounds good. [18:21:06] awight: you got blocked? [18:23:13] 10Scoring-platform-team (Current), 10ChangeProp, 10ORES: Drop "non bot" condition from ORES changeprop rules - https://phabricator.wikimedia.org/T187927#4081768 (10awight) [18:23:26] awight, wondering if we should implement something that requires us to change the yaml docs right now or not [18:23:28] ^ for each wiki [18:23:48] Vermont: it's not a very interesting blockage, just an access thing... [18:24:16] halfak: Was this the change you were suggesting last week? [18:24:22] awight, yeah [18:24:44] I liked the idea very much, except for the additional boilerplate required for the common use cases. [18:24:45] * Vermont is disappointed [18:24:51] I think I have a compromise (halfak) [18:25:12] that we create partial templates which describe the most common flows, and these are described using your explicit syntax. [18:25:13] cool [18:25:24] Ooh. I like it. [18:25:26] :D [18:25:46] k I can try to make that happen. [18:26:17] Amir1 should be in on codegen discussions, btw... [18:26:41] I'm about to make some progress on some stuff [18:26:45] you might like it [18:26:51] :) [18:26:54] which stuff? [18:27:06] halfak: want to merge editquality#145? [18:27:09] ores support checklist [18:27:15] oh, neat [18:28:34] https://usercontent.irccloud-cdn.com/file/66Fnk8Rf/image.png [18:29:12] awight: halfak ^ [18:29:28] I also made a flask app it's about to finish [18:29:46] well, the hard part is left, storage of the data [18:30:09] * halfak goes into next meeting. [18:30:17] awight, merged, BTW [18:31:12] wiki-ai/wikilabels#330 (progress_days - becb44f : Amir Sarabadani): The build passed. https://travis-ci.org/wiki-ai/wikilabels/builds/358542443 [18:33:22] halfak: Thanks! [18:35:49] awight: Thanks! [18:35:56] We should deploy it :D [18:36:26] I'm working on the app, hopefully it'll be done soon [18:46:53] What app is that? [18:47:59] srrodlun_, around? [19:06:03] o/ Vermont & srrodlund. [19:06:10] I wanted you both to meet each other. [19:06:29] Vermont has been working with us for a while and was looking to do some ORES-y work today. I thought docs would be a good opportunity. [19:06:49] srrodlund, is our resident tech writer. She's been working with us to assess ORES docs and prioritize new work. [19:07:21] I was thinking of asking Vermont to try to answer a basic question about ORES using the available docs and document his progress. [19:07:58] We can then use his note to prioritize new work on ORES docs and maybe Vermont can help us iterate -- to make sure they are understandable to someone outside of the team :) [19:10:32] @vermont Nice to meet you! [19:10:33] 04Error: Command “vermont” not recognized. Please review and correct what you’ve written. [19:10:46] @Vernont [19:10:46] 04Error: Command “vernont” not recognized. Please review and correct what you’ve written. [19:12:19] @shuddap AsimovBot [19:12:20] 04Error: Command “shuddap” not recognized. Please review and correct what you’ve written. [19:12:21] halfak: Early adopter issues, https://github.com/halfak/deep_merge/pull/1 [19:12:35] I see you're busy :p [19:12:47] lol [19:12:56] Sorry about that one [19:12:59] lol @ that random print [19:13:38] 0.0.2 uploaded to pypi [19:13:46] ty [19:13:59] > Makefile => surprise! [19:24:47] awight, looks like docs meeting isn't happening [19:26:40] that works for me [19:26:49] my day is only going to get asymptotically crazier. [19:26:58] oh damn. [19:27:14] * awight heads to infinity [19:27:21] awight, do you think we can get this new template stuff squared away by EOD so I can work on the next round of wikis? [19:27:36] We're currently squared, I'd say. [19:27:49] I'm working on huwiki, not the new syntax [19:27:55] (got the ssh key working) [19:28:11] OK so I should implement with the old syntax? [19:28:33] Please do [19:29:41] * awight scratches head that huwiki is no longer included in the Makefile [19:33:24] * awight eyes deep_merge with suspicion [19:34:21] deep_merge.merge modifies its first parameter? [19:34:22] please do! I kind of hacked it together quickly :\ [19:34:30] hehe ty for the service [19:34:39] * awight puts on Laika-sized spacesuit [19:34:39] yes. Stuff on the right takes priority [19:34:59] https://en.wikipedia.org/wiki/West_Siberian_Laika ? [19:35:43] https://en.wikipedia.org/wiki/Laika [19:36:59] I would prefer that merge() either modifies its first param, *or* returns a result without modifying the params. [19:37:07] The latter is nicer :) [19:37:31] Oh for cute [19:37:52] awight, oh! It does modify the first param [19:37:53] Sorry [19:37:55] * awight consults midwestern glossary [19:37:56] for the lack of clarity [19:38:11] growing pains, not a problem [19:38:15] If you want it to create a brand new dict, you can always pass that brand new dict as the first param [19:38:19] I'll workaround for now [19:38:26] nice hack [19:38:32] three = merge({}, one, two) [19:38:44] javascripty :) [19:39:02] * halfak interprets that positively -- through great effort [19:39:08] lolol [19:40:07] I like the Ruby solution to this question [19:40:21] Most functions have "merge" and "merge!" variants [19:40:33] the -"!" variant does crazy shit like modifying params [19:41:21] cool cos usages are always clean... if somewhat magical [19:43:07] The overrides for each merge are slick, btw [19:45:04] halfak: It looks like the param modification actually jumps across dict references [19:45:29] e.g. if key *not in d1: d1[key] = d2[key] [19:46:08] that will make d1[key] a reference, and then merging a third dict in causes d2 to be modified. [19:47:09] awight, oh good catch. [19:47:45] Wait. How would d2 get modified in that case? [19:48:00] d1[key] becomes a reference, but recursion stops there. [19:50:02] merge({}, d2, d3) can cause "d1" to be, for example, d1 = {"key1": &d2["key1"]} [19:50:21] right. [19:50:34] then, d1 is merged with d3 and if d3 contains key1, then d3["key1"] is merged onto d2["key1"] [19:50:50] aha! [19:51:07] I also have empirical evidence ;-) [19:51:20] ./utility autolabel --host=https://ar.wikipedia.org \ [19:51:20] [19:51:20] --trusted-groups=sysop,oversight,editor,bot,rollbacker,checkuser,abusefilter,bureaucrat,sysop,oversight,bot,rollbacker,checkuser,abusefilter,bureaucrat,bot,sysop,bureaucrat,autopatrolled,bot,bureaucrat,checkuser,reviewer,rollbacker,sysop,bot,sysop,autopatrolled,bot,bureaucrat,checkuser,reviewer,rollbacker,sysop,sysop,oversight,bot,rollbacker,checkuser,abusefilter,bureaucrat,bot,bureaucrat,sysop,sysop, [19:51:27] oversight,bot,rollbacker,checkuser,abusefilter,bureaucrat,sysop,oversight,bot,rollbacker,checkuser,abusefilter,bureaucrat,sysop,oversight,bot,rollbacker,checkuser,abusefilter,bureaucrat,sysop,oversight,bot,rollbacker,checkuser,abusefilter,bureaucrat,autopatrolled,sysop,oversight,bot,rollbacker,checkuser,abusefilter,bureaucrat,autopatrolled,sysop,oversight,bot,rollbacker,checkuser,abusefilter,bureaucra [19:51:33] t,flow-bot,sysop,oversight,bot,rollbacker,checkuser,abusefilter,bureaucrat,flow-bot,sysop,oversight,bot,rollbacker,checkuser,abusefilter,bureaucrat,sysop,oversight,bot,rollbacker,checkuser,abusefilter,bureaucrat,autopatrolled,bot,bureaucrat,checkuser,reviewer,rollbacker,sysop,sysop,oversight,trusted,bot,rollbacker,checkuser,abusefilter,bureaucrat,editor,templateeditor,interface-editor,autoreview,bot,b [19:51:39] Shall we document that "no inputs are safe" or should we preemtively make inputs safe? [19:51:39] ureaucrat,checkuser,editor,flow-bot,oversight,reviewer,rollbacker,sysop,autopatrolled,bot,bureaucrat,checkuser,reviewer,rollbacker,sysop,sysop,oversight,bot,rollbacker,checkuser,abusefilter,bureaucrat,abusefilter,bot,bureaucrat,checkuser,eliminator,interface-editor,oversight,rollbacker,sysop,abusefilter,bot,bureaucrat,checkuser,eliminator,interface-editor,oversight,rollbacker,sysop,sysop,bureaucrat,bo [19:51:45] t,oversight,checkuser,patroller,autopatrolled,abusefilter,arbcom,bureaucrat,checkuser,rollbacker,sysop,bot,sysop,oversight,bot,rollbacker,checkuser,abusefilter,bureaucrat,abusefilter,arbcom,bureaucrat,checkuser,rollbacker,sysop,bot,templateeditor,patroller,autopatrolled,abusefilter,arbcom,bureaucrat,checkuser,rollbacker,sysop,bot,sysop,oversight,trusted,bot,rollbacker,checkuser,abusefilter,bureaucrat, [19:51:49] holy moley [19:51:51] autopatrolled,bot,bureaucrat,patroller,rollbacker,sysop,sysop,oversight,trusted,bot,rollbacker,checkuser,abusefilter,bureaucrat,autopatrolled,bot,bureaucrat,patroller,rollbacker,sysop,sysop,oversight,trusted,bot,rollbacker,checkuser,abusefilter,bureaucrat,abusefilter,arbcom,bureaucrat,checkuser,rollbacker,sysop,bot,bot,bureaucrat,sysop,rollbackers,checkuser,bureaucrat,sysop,eliminator,bot,checkuser,bu [19:51:57] reaucrat,sysop,eliminator,bot \ [19:51:59] for example [19:52:01] * halfak kicks awight for spam ; [19:52:02] IMO, not modifying parameters is the best solution [19:52:02] :P [19:52:42] make "overwrite" make a copy [19:53:03] Seems like an easier solution [19:53:03] * awight chuckles that a bot might have died due to my spam [19:53:18] What does that mean? [19:54:12] overwrite should do more than just return v2 [19:54:37] should return copy.deepcopy(v2) [19:54:46] I'm not sure there's a lot of use for an overwrite... [19:54:56] Also, that list seems like it's an append issue. [19:55:04] Seems like you want merge_list to use overwrite. [19:55:24] Even within deep_merge, you assume that the return value is useful and the first param was disposable [19:55:33] When one dict has a key and another one doesn't "overwrite()" us used [19:55:54] [spam] https://commons.wikimedia.org/wiki/Category:Laika#/media/File:Laika_ac_Laika_(6836480284).jpg [19:56:12] https://github.com/halfak/deep_merge/blob/master/deep_merge/merge.py#L9 [19:56:48] https://github.com/halfak/deep_merge/blob/master/deep_merge/merge.py#L76 is what I mean by not dogfooding [19:57:29] How does overwrite() help? I think you're suggesting we replace line 42 with d1[key] = overwrite(d1[key], d2[key])? [19:57:43] ah and overwrite returns a deep copy [19:57:51] yeah that's better! [19:58:15] right--cos even if we "return d" and don't modify the first param, the result still would have contained these broken references. [19:59:29] btw, you don't have to push updates to pypi for me [19:59:35] I have pip installed -e .'d [20:00:07] awight, other people. Hugely popular package ;) [20:00:10] lolol [20:00:26] hehe I was happy that you kept up the joke, though! [20:03:21] * halfak gets the last couple pieces of the edit type puzzle together [20:04:36] 10Scoring-platform-team (Current), 10edittypes-modeling, 10artificial-intelligence: Train edit types model on labeled data for English Wikipedia - https://phabricator.wikimedia.org/T121715#4082153 (10Halfak) Diyi's old work: https://github.com/diyiy/Wiki_Semantic_Intention The repo I started a while ago: ht... [20:24:56] halfak: awight_ around? [20:25:03] o/ [20:25:31] this is output of the flask app I wrote today, it's completely automatic [20:25:34] https://usercontent.irccloud-cdn.com/file/RlHos5Kq/image.png [20:26:16] halfak: It's not very nice design though [20:26:21] Amir1: Bravo! [20:26:58] It's supremely usable [20:27:15] Thanks. I make a repo in github, publish it and make it in toolforge and call it a day [20:27:27] Looks good :) [20:27:30] feel free to add/change things mercilessly [20:27:47] Once you have a stale-detector, that would be worth noting in the chart [20:28:04] Yeah, that makes the chart even bigger [20:28:09] What are "not applicable"vs blank? [20:28:20] Just got back. [20:28:30] I see I was pinged 6 times [20:28:32] not applicable means since we have damaging, reverted doesn't make sense [20:28:42] srrodlund: pinging [20:28:43] blank means we can have it but we don't [20:28:56] Amir1: OIC, that's blocked out. works for me now [20:29:12] Vermont hey hey! [20:29:24] yeah, let me make it live and you edit it mercielssly [20:29:58] So, I should be doing...? [20:30:22] I believe halfak has information for you [20:31:10] Vermont, I was thinking of a good question to start with [20:31:23] I like "What is ORES and how do I use it?" [20:31:32] I'm hoping to have to write up a journal entry. [20:31:46] Well I can answer that pretty simply [20:31:51] That probably starts with "So I type ORES into Wikipedia's search box and find nothing useful" [20:32:17] To do this [20:32:23] What sort of response do you want and for what purpose [20:32:43] let me redesign the table to follow DRY and get less space [20:34:01] Vermont, record your steps and the answers you come across. [20:34:05] Or don't come across. [20:34:09] okay [20:52:14] halfak: awight What do you think of this? [20:52:17] https://usercontent.irccloud-cdn.com/file/BlmPClRh/image.png [20:52:39] s/done// [20:53:17] s#not applicable#n/a# [20:53:43] kk [20:54:53] halfak: [20:54:56] https://usercontent.irccloud-cdn.com/file/bIOIQ7Ak/image.png [20:55:14] It needs some CSS clean up, things like that [20:58:13] Solid [20:58:28] Maybe s/Labeling campaign/labeling/ [21:00:28] Amir1: that's even better! Pure-number cell data is more readable, and it's clear how to get to the goal. [21:00:43] it's also consistent with the ORES/Support table, fwiw [21:01:32] collapsing damaging and goodfaith is rad [21:01:56] Maybe replace "100%" with a check mark? [21:02:35] left-to-right (in English) layout from basic to advanced models is really nice [21:13:53] awight: halfak https://tools.wmflabs.org/ores-support-checklist/ [21:14:37] Things to do: 1- link to the source code 2- make the table sortable 3- add cronjob to update the data every three hours [21:14:55] 4- automatic update of the tool (using cron) [21:15:06] 5- clean up the source code [21:15:07] https://github.com/wiki-ai/ores-support-checklist [21:15:13] halfak: Spooky time. Why is d2 modified during append_list? https://github.com/adamwight/deep_merge/commit/dc70bb7a864f8acd4846bc32e7972810425fcef4#diff-baafecea32b5f63c1a6ce6cd177bd305R77 [21:15:36] I'm calling it a day, already worked for 11 hours today, it's 23:15 :/ [21:15:38] Amir1: rockin'! [21:15:51] that's too many hours [21:16:29] 10Scoring-platform-team (Current), 10User-Ladsgroup: Build ORES support checklist - https://phabricator.wikimedia.org/T189954#4082429 (10Ladsgroup) https://tools.wmflabs.org/ores-support-checklist/ Things to do: 1- link to the source code 2- make the table sortable 3- add cronjob to update the data every thre... [21:16:35] The official repo and tool are rocking it; excessive overtime is not rocking [21:17:56] I had some undertime for ORES before, it's okay [21:18:00] see you tomorrow [21:18:01] o/ [21:23:08] see ya [21:23:15] o/ have a good one :) [21:32:58] halfak: https://github.com/halfak/deep_merge/pull/2 [21:33:10] The tests should be wired into CI [21:33:15] want to move this to wiki-ai? [21:33:52] Confirmed that it solves my generate_make problems. [21:34:24] Still, I might write a "merge_set" function to use instead of append_list. [21:35:08] err--on second thought, maybe overwrite is the right choice for the wiki_defaults lists [21:36:21] yes, definitely. trusted_groups is the only key [22:05:03] 10Scoring-platform-team, 10Research, 10Research-outreach: Organize a technical workshop for ORES - https://phabricator.wikimedia.org/T141310#4082696 (10leila) Halfak says this is unlikely to happen. I'll decline it. Feel free to bring it back to life, you two! :) [22:05:14] 10Scoring-platform-team, 10Research, 10Research-outreach: Organize a technical workshop for ORES - https://phabricator.wikimedia.org/T141310#4082697 (10leila) 05Open>03declined [22:06:16] wiki-ai/editquality#259 (awight-huwiki - 53d62e5 : Adam Wight): The build failed. https://travis-ci.org/wiki-ai/editquality/builds/358631161 [22:07:58] 10Scoring-platform-team: Report revscoring utilities errors in a summary at the end of the run, rather than inline warnings - https://phabricator.wikimedia.org/T190747#4082717 (10awight) [22:14:13] * halfak builds tuning reports for arwiki and lvwiki [22:23:15] Amir1, has taken over stat1005 :P [22:23:46] I need to run. Time for evening errands. [22:23:51] Have a good one folks! [22:23:58] godspeed awight :) [22:24:51] o/ [22:29:31] 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#4082811 (10awight) I've made merge_labels smarter, it should do the correct inter... [22:30:36] \o/