[01:06:09] thansk for the mention, I'll be glad to answer any questions about newcomer modelling. Actually i deployed initial labelling experiments to labels-experiment.wmflabs.org each of 1k newcomer sessions. Any standard way we wiki-ai advertises new labelling campaigns? [07:36:03] 10Scoring-platform-team (Current), 10Documentation: Document and justify JADE schema proposals - https://phabricator.wikimedia.org/T204250 (10awight) [07:43:31] 10Scoring-platform-team (Current), 10Documentation: Document and justify JADE schema proposals - https://phabricator.wikimedia.org/T204250 (10awight) @Halfak I won't re-revert your changes, but the proposals listed don't match what is currently being considered. I'm not sure what you mean by "single-judgment"... [07:48:26] 10Scoring-platform-team (Current), 10Documentation: Document and justify JADE schema proposals - https://phabricator.wikimedia.org/T204250 (10awight) Okay, I've read through [[mw:JADE/Content_schemas]] and I see where you're coming from. This does look like a useful exercise, a historical dive into mistakes m... [07:48:27] 10[7] 10https://www.mediawiki.org/wiki/JADE/Content_schemas [07:49:03] 10Scoring-platform-team (Current), 10Documentation: Document the history of JADE schema proposals and justifications - https://phabricator.wikimedia.org/T204250 (10awight) [07:57:11] 10Scoring-platform-team (Current), 10Documentation: Document the history of JADE schema proposals and justifications - https://phabricator.wikimedia.org/T204250 (10awight) [08:13:05] 10Scoring-platform-team (Current), 10JADE: Investigate MCR support gap for JADE purposes - https://phabricator.wikimedia.org/T204303 (10awight) p:05Triage>03Low [10:04:10] Allah ⅰs ⅾoiᥒɡ [10:04:12] suᥒ іѕ ᥒоt dοiᥒɡ Αⅼⅼah is doing [10:04:16] moⲟn is nоt ԁoinɡ Αⅼⅼah is dоing [10:04:18] ѕtɑrѕ аrе not ԁoing Aⅼlaһ iѕ ԁoіnɡ [10:04:20] ⲣⅼɑᥒеts ɑre nοt ԁⲟіng Аlⅼaһ ⅰѕ ԁⲟinɡ [10:04:23] gaⅼaxⅰеs аre ᥒot doіnɡ Aⅼⅼah іs ԁoⅰng [10:04:26] oceɑnѕ are ᥒot ⅾоing Аlⅼɑh is ԁoing [10:04:29] mഠuntɑins аrᥱ not dоinɡ Аlⅼaһ iѕ doing [10:04:31] treeѕ are ᥒοt dοіng Alⅼаh is doⅰng [10:04:34] ⅿοm iѕ ᥒot dοing Aⅼlаһ ⅰs dഠing [10:04:36] daԁ iѕ not doing Аllɑһ іs doiᥒg [10:04:39] bοsѕ is not dοinɡ Αⅼlаh iѕ dоⅰᥒɡ [10:04:41] job is not doⅰng Αⅼⅼаh iѕ doіng [10:04:45] dоⅼlаr is nഠt dоⅰᥒɡ Αllɑһ is ⅾоіng [10:28:48] Аlⅼаh іѕ ⅾoⅰnɡ [10:50:57] Aⅼⅼah іs doing [10:59:17] Allah ⅰs dⲟing [10:59:17] Аllаh iѕ dⲟing [11:37:24] Alⅼaһ іѕ doіnɡ [11:37:24] suᥒ is not doing Aⅼⅼah iѕ doinɡ [11:37:24] moon ⅰs ᥒοt dοinɡ Alⅼah iѕ ⅾoinɡ [11:54:35] 10Scoring-platform-team (Current), 10ORES: Test poolcounter support for ores in beta cluster - https://phabricator.wikimedia.org/T201825 (10Ladsgroup) Yup [12:47:51] 10Scoring-platform-team, 10MediaWiki-extensions-ORES, 10Patch-For-Review: Store wp10 and draftquality scores for Draft namespace - https://phabricator.wikimedia.org/T198768 (10Liuxinyu970226) [12:48:05] 10Scoring-platform-team, 10Growth-Team, 10MediaWiki-extensions-ORES, 10Patch-For-Review: Store wp10 and draftquality scores for Draft namespace - https://phabricator.wikimedia.org/T198768 (10Liuxinyu970226) [12:49:50] 10Scoring-platform-team, 10Growth-Team, 10MediaWiki-extensions-ORES: In the enhanced watchlist, ORES colorizes all edits in the group - https://phabricator.wikimedia.org/T192900 (10Liuxinyu970226) [13:22:12] Αllɑh iѕ ԁoiᥒg [14:11:54] Alⅼah іs ԁoing [14:17:31] o/ harej you around? [14:17:58] I was looking at what you did with "features" on our use cases page and I have some concerns and questions. [14:37:38] Aⅼlаh iѕ ⅾoiᥒg [14:37:38] ѕ∪ᥒ iѕ ᥒot doіnɡ Аlⅼah is ⅾоⅰng [14:46:39] Whatever that means is fine. [14:46:45] Doesn't sound offensive. [14:52:46] 10Scoring-platform-team (Current), 10Documentation: Document JADE schema proposals and justifications - https://phabricator.wikimedia.org/T204250 (10Halfak) [15:17:11] 10Scoring-platform-team (Current), 10Documentation: Document JADE schema proposals and justifications - https://phabricator.wikimedia.org/T204250 (10Halfak) I'm not interested in re-visiting history. I think it's valuable to include all proposals in one space. There are some proposals that there might be con... [15:17:18] 10Scoring-platform-team (Current), 10Documentation: Document JADE schema proposals and justifications - https://phabricator.wikimedia.org/T204250 (10Halfak) [15:26:58] halfak: harej: Point of information, Wikidata does not support wikitext snippets, only plain text. That makes us the first group to try something quite ambitious... [15:27:09] https://www.wikidata.org/wiki/Help:Data_type [15:27:17] awight, I don't think it's appropriate to call this "quite ambitious" [15:27:21] What's ambitious. [15:27:26] Formatting wikitext? [15:27:38] Storing it isn't any different. [15:27:41] Alright, let's say "pioneering" [15:28:10] Storing it is slightly different actually, there are considerations such as secondary link table maintenance. [15:28:33] Good point! [15:28:56] Seems like that shouldn't be too crazy though. [15:29:42] I just need to compile the potential complications. AbuseFilter integration is another, but we already knew that. [15:32:19] Abuse filter integration shouldn't be too bad since it can run on the raw JSON [15:33:06] Definitely doable. [15:38:03] CollaborationKit actually stores wikitext in JSON, but no one cares about CollaborationKit ;] [15:38:44] Is the better way... available yet? [15:40:39] halfak: What are your concerns? [15:41:28] harej: Good to know! That's your task distribution extension, right? [15:41:38] it's my group project extension [15:41:58] with a list building capability (but nothing behind those lists; it's up to the project members to build the lists manually) [15:42:19] Isarra did more work on it though (hi Isarra!) [15:42:35] * Isarra waves. [15:42:49] :) [15:43:16] Cool, I'm glad to have an excuse to read the code! [15:43:24] Oh gods. [15:44:24] (Don't read our unit tests.) [15:44:41] hahaha [15:44:55] harej, I think it ties to closely to an implementation. The open questions make me feel that maybe you are hoping to wave off specific details? [15:45:44] Isarra, "better way"? [15:45:44] halfak: FYI this train of thought is about evaluating implementation challenges, not to compare with the alternatives. [15:46:10] awight, train of thought re. wikitext in JSON? [15:46:50] halfak: the open questions are because I don't know what to definitively state as the JADE feature [15:46:53] harej, does collaborationKit update link tables and stuff? [15:47:19] I believe so. If not it... should. [15:47:24] harej, right. We had a proposed feature list that got archived. Might be useful in checking our use-cases as complete. [15:47:44] Speaking of which, harej, would you add a "reverting vandalism" use-case? [15:47:59] I believe that's adequately different from suppression, yes [15:48:04] Since we tend to not suppress mundane vandalism [15:48:08] +1 [15:48:10] * awight smiles that people are using TestingAccessWrapper [15:48:13] Or mistakes or whatever [15:48:20] Maybe it should be reverting damage [15:48:20] halfak: exactly. [15:48:23] Also patrolling for damage [15:48:36] Patrolling for damage is a more interesting use case [15:54:29] kk, there are some examples of wikitext snippets in the unit tests. [15:54:39] * awight installs locally [15:57:08] Alⅼаh іs doinɡ [15:57:33] Isarra: harej: off-topic, should CollaborationKit really have a hard dependency on EventLogging, or could it gracefully degrade? [15:57:45] the dependency on eventlogging is for JSON validation stuff [15:58:30] the newer one (justin rainbow) only outputs errors in english [15:59:46] awight: please don't get too mesmerized with collaborationkit, there's still a JADE we have to work through :P [15:59:57] haha [16:00:11] * awight holds laptop screen at an angle to reflect the basilisk's gaze [16:06:18] This allah spam is strange and seemingly harmless. [16:06:37] Looks like I'll need some permissions twiddling to make CollaborationKit work [16:07:20] harej: Oh hey, you're saying that EventLogging provides multilingual validation errors? That would be useful for JADE. [16:07:44] Do we depend on EventLogging anyway? [16:07:47] for JADE? [16:07:47] I still need to resolve T197653 but then will want that. [16:07:48] T197653: Surface JADE validation errors - https://phabricator.wikimedia.org/T197653 [16:07:49] Not yet. [16:08:16] harej, have you talked to Danny Horn about CollaborationKit? [16:08:16] FWIW, graceful degradation seems correct here as well. justinrainbow validate if EL is missing... [16:08:19] If we use EventLogging anyway (we might want to, to create user-friendly analyses), then we get JSON validation for free [16:08:40] halfak: To be honest I don't remember. I've talked to other WMF product managers about it. [16:08:53] harej: interesting! Let's chat about that soon, I think it's going over my head. [16:08:58] Uh.. using EventLogging and using JSON validation from EventLogging are two very different things. [16:09:18] harej, Dannyh is interested in setting up a 3-5 year plan around WikiProjects [16:09:21] Right. I'm just saying if we do the former, then we get the latter. You are right that the former is a whole other conversation. [16:09:29] And other topic-focused working groups. [16:09:49] harej, we do not "Get eventlogging" for using event logging code. [16:10:04] We get eventlogging just for being in the wiki and using pages and revisions. [16:10:15] But we would have to develop custom hooks to log specific types of edits. [16:10:38] That is correct [16:11:01] I read it the other way around, that running our JSON through custom EL hooks will get us the validation for free. [16:11:31] Not even hooks. I think CollaborationKit does this much more simply; it includes the EventLogging extension and then uses the JSON validation methods built into EventLogging. [16:11:47] oh well that's a possibility for sure [16:14:09] harej: found the permissions glitch. the "new project" dialog defaults to the wiki's namespace (e.g. Devwiki), which wasn't set up in my case. I'll save to User instead. [16:15:13] (haha, simpler. I wasn't logged-in) [16:15:47] also setting up collaboration hubs requires you to have the change content model permission [16:16:15] The JSON diff is already nice and readable for single-line content... [16:16:38] harej: oh hmm, should be documented in mw:E:CollaborationKit ;-) [16:17:34] Multiline diffs are not at all nice, cos the data becomes a single line with escaped newlines. [16:17:34] Aⅼlah is doing [16:19:53] Hopefully we don't have to deal with much multiline content in JADE [16:20:00] Though, the proposed generalized notes field might be multiline [16:21:37] harej: Holler! The links table *is* updated. [16:22:47] actually, doing one more test, I unwisely used my own user page as the link target but as a project member I'm already linked. [16:23:01] halfak: https://www.mediawiki.org/w/index.php?title=JADE%2FUse_cases&type=revision&diff=2878463&oldid=2878386 [16:24:09] harej, looks good. Maybe there should also be something about diffs? Not sure. E.g. we could use JSON formatting in our diffs to make sure that changes are easy to see. [16:24:36] judging by awight's reaction to how collaborationkit does it (poorly), we probably should use special JSON diffs (didn't know they were a thing!) [16:26:30] https://meta.wikimedia.org/w/index.php?title=Schema%3AEdit&type=revision&diff=17541122&oldid=17520217 [16:26:45] Doesn't look terrible, but I get what you are saying. [16:27:14] https://www.wikidata.org/w/index.php?title=Q6938433&type=revision&diff=723164605&oldid=665874868 looks a lot better [16:27:56] Wikibase is spot on with its diff generation [16:28:21] I thiiiink at one point we had an idea to do custom diffs for CollaborationKit but it was de-prioritized. Isarra do you remember? [16:28:27] Isarra: harej: okay so, the link table is updated with links in the CollaborationKit JSON content, but I don't see this being done explicitly anywhere. Do you happen to know why that works? [16:28:49] Might be handled in a parent class. [16:29:10] CollaborationHubContent extends JsonContent (or TextContent, don't remember) [16:32:09] TextContent::fillParserOutput could be responsible for it. [16:32:25] TextContent::diff looks promising too, overriding the diff may be easy. [16:34:26] halfak: what do you advise I do when describing the feature corresponding to the use case when it hasn't been conclusively decided? [16:39:48] * awight busses offspring [16:39:52] back soon [16:43:01] harej: Yeah, we never actually implemented that. [16:43:13] Uuuuh lessee. [16:47:16] Yeah, looks like the updates are done as part of the working with the parser/parseroutput, and the main $output for the page in general is what has that attached to it, and thus what... leads the tables themselves to get updated?. [16:47:37] Voodoo. [17:01:18] harej, awight: will join shortly [17:44:52] Isarra: I did find a chicken bone covered in melted wax in one of the unit test fixtures... [17:49:12] Allaһ is ԁоing [18:22:41] ...wat. [18:25:33] awight: would you say at this point you can definitively state what the schema is? [18:27:10] harej: yes, it's unchanged and should already be documented. lemme get you that [18:27:59] My proposals were trying to patch various issues, so now I need to go back and remind myself what the issues are. [18:50:43] harej: This is the schema, https://phabricator.wikimedia.org/diffusion/EJAD/browse/master/jsonschema/judgment/v1.json [18:51:15] I need to re-add origin metadata since halfak wants it, and there's a tiny bit of cleanup to do with articlequality etc. [19:30:42] Αlⅼаh іs doiᥒɡ [19:30:42] suᥒ iѕ ᥒഠt dⲟing Aⅼlaһ is doіng [20:20:04] Allaһ iѕ ԁഠiᥒg [20:38:53] Aⅼlaһ іs ԁоіᥒg [20:42:42] Ꭺⅼlaһ is doing [20:42:42] sun is ᥒοt dοing Aⅼlaһ iѕ doⅰng [21:44:00] Αllаһ is ⅾoinɡ [22:14:17] (03PS3) 10Awight: Change schema to a list of heterogenous judgments about a single entity [extensions/JADE] - 10https://gerrit.wikimedia.org/r/456424 [22:14:19] (03PS4) 10Awight: [WIP] rough secondary tables [extensions/JADE] - 10https://gerrit.wikimedia.org/r/456078 (https://phabricator.wikimedia.org/T203037) [22:14:21] (03PS4) 10Awight: [WIP] Replace generic APIs with specific ones [extensions/JADE] - 10https://gerrit.wikimedia.org/r/456044 (https://phabricator.wikimedia.org/T203037) [22:14:23] (03PS1) 10Awight: Include annoying copyright preamble [extensions/JADE] - 10https://gerrit.wikimedia.org/r/460611 [22:14:32] (03CR) 10jerkins-bot: [V: 04-1] Include annoying copyright preamble [extensions/JADE] - 10https://gerrit.wikimedia.org/r/460611 (owner: 10Awight) [22:14:34] (03CR) 10jerkins-bot: [V: 04-1] [WIP] Replace generic APIs with specific ones [extensions/JADE] - 10https://gerrit.wikimedia.org/r/456044 (https://phabricator.wikimedia.org/T203037) (owner: 10Awight) [22:14:36] (03CR) 10jerkins-bot: [V: 04-1] [WIP] Replace generic APIs with specific ones [extensions/JADE] - 10https://gerrit.wikimedia.org/r/456044 (https://phabricator.wikimedia.org/T203037) (owner: 10Awight) [22:14:39] (03CR) 10jerkins-bot: [V: 04-1] Include annoying copyright preamble [extensions/JADE] - 10https://gerrit.wikimedia.org/r/460611 (owner: 10Awight) [22:15:51] (03CR) 10jerkins-bot: [V: 04-1] [WIP] rough secondary tables [extensions/JADE] - 10https://gerrit.wikimedia.org/r/456078 (https://phabricator.wikimedia.org/T203037) (owner: 10Awight) [22:19:17] (03CR) 10jerkins-bot: [V: 04-1] Change schema to a list of heterogenous judgments about a single entity [extensions/JADE] - 10https://gerrit.wikimedia.org/r/456424 (owner: 10Awight) [22:27:23] Alⅼaһ iѕ ԁഠіᥒg [22:45:41] Аⅼⅼah іs dоiᥒg [23:33:58] (03PS5) 10Awight: [WIP] Secondary indexes for JADE pages [extensions/JADE] - 10https://gerrit.wikimedia.org/r/456078 (https://phabricator.wikimedia.org/T203037) [23:34:00] (03PS1) 10Awight: Rename namespace to NS_JUDGMENT [extensions/JADE] - 10https://gerrit.wikimedia.org/r/460615 [23:34:02] (03PS1) 10Awight: [WIP] Hook to update judgment link table [extensions/JADE] - 10https://gerrit.wikimedia.org/r/460616 [23:34:09] Have a fun weekend! [23:35:31] (03CR) 10jerkins-bot: [V: 04-1] [WIP] Hook to update judgment link table [extensions/JADE] - 10https://gerrit.wikimedia.org/r/460616 (owner: 10Awight) [23:37:57] (03CR) 10jerkins-bot: [V: 04-1] [WIP] Secondary indexes for JADE pages [extensions/JADE] - 10https://gerrit.wikimedia.org/r/456078 (https://phabricator.wikimedia.org/T203037) (owner: 10Awight) [23:39:17] (03CR) 10jerkins-bot: [V: 04-1] Rename namespace to NS_JUDGMENT [extensions/JADE] - 10https://gerrit.wikimedia.org/r/460615 (owner: 10Awight) [23:43:10] (03CR) 10jerkins-bot: [V: 04-1] Rename namespace to NS_JUDGMENT [extensions/JADE] - 10https://gerrit.wikimedia.org/r/460615 (owner: 10Awight) [23:49:28] (03CR) 10jerkins-bot: [V: 04-1] [WIP] Hook to update judgment link table [extensions/JADE] - 10https://gerrit.wikimedia.org/r/460616 (owner: 10Awight)