[08:07:41] PROBLEM - puppet on ORES-web02.Experimental is CRITICAL: connect to address 10.68.23.111 port 5666: Connection refusedconnect to host ores-web-02.ores.eqiad.wmflabs port 5666: Connection refused [08:08:43] PROBLEM - check disk on ORES-web02.Experimental is CRITICAL: connect to address 10.68.23.111 port 5666: Connection refusedconnect to host ores-web-02.ores.eqiad.wmflabs port 5666: Connection refused [08:09:32] PROBLEM - check load on ORES-web02.Experimental is CRITICAL: connect to address 10.68.23.111 port 5666: Connection refusedconnect to host ores-web-02.ores.eqiad.wmflabs port 5666: Connection refused [08:09:32] PROBLEM - check users on ORES-web02.Experimental is CRITICAL: connect to address 10.68.23.111 port 5666: Connection refusedconnect to host ores-web-02.ores.eqiad.wmflabs port 5666: Connection refused [08:12:19] 10Scoring-platform-team, 10articlequality-modeling, 10artificial-intelligence: Add wp10 model in PetScan - https://phabricator.wikimedia.org/T197871#4314385 (10Magnus) I run PetScan. The reason this fails is because neither ores_classification nor oresc_model tables are available on the euwiki database repl... [09:07:44] PROBLEM - puppet on ORES-web02.Experimental is CRITICAL: connect to address 10.68.23.111 port 5666: Connection refusedconnect to host ores-web-02.ores.eqiad.wmflabs port 5666: Connection refused [09:08:44] PROBLEM - check disk on ORES-web02.Experimental is CRITICAL: connect to address 10.68.23.111 port 5666: Connection refusedconnect to host ores-web-02.ores.eqiad.wmflabs port 5666: Connection refused [09:09:34] PROBLEM - check load on ORES-web02.Experimental is CRITICAL: connect to address 10.68.23.111 port 5666: Connection refusedconnect to host ores-web-02.ores.eqiad.wmflabs port 5666: Connection refused [09:09:34] PROBLEM - check users on ORES-web02.Experimental is CRITICAL: connect to address 10.68.23.111 port 5666: Connection refusedconnect to host ores-web-02.ores.eqiad.wmflabs port 5666: Connection refused [09:16:44] RECOVERY - check disk on ORES-web02.Experimental is OK: DISK OK [09:17:32] RECOVERY - check load on ORES-web02.Experimental is OK: OK - load average: 0.19, 0.17, 0.30 [09:17:32] RECOVERY - check users on ORES-web02.Experimental is OK: USERS OK - 0 users currently logged in [09:47:42] RECOVERY - puppet on ORES-web02.Experimental is OK: OK: Puppet is currently enabled, last run 1 minute ago with 0 failures [11:38:41] (03PS1) 10Awight: Drop endorsements from schema; add judgment.notes [extensions/JADE] - 10https://gerrit.wikimedia.org/r/442074 [11:39:12] 10Scoring-platform-team (Current), 10Documentation, 10Easy: Document JADE judgment structure - https://phabricator.wikimedia.org/T179301#4315019 (10awight) https://gerrit.wikimedia.org/r/#/c/mediawiki/extensions/JADE/+/442074 [11:41:01] 10Scoring-platform-team, 10ORES, 10Browser-Tests: ORES webdriver.io selenium test fail on CI due to lack of ORES server - https://phabricator.wikimedia.org/T198201#4315032 (10hashar) [11:48:58] 10Scoring-platform-team, 10Huggle, 10JADE: Use JADE as a repository for ORES counterexamples - https://phabricator.wikimedia.org/T197098#4315064 (10awight) /me crunches on popcorn. Thanks for the pointer! My query was actually against all wikis, but only for a 24-hour period. It's fine if this integration... [12:08:21] (03PS1) 10Hashar: selenium: skip test on CI [extensions/ORES] - 10https://gerrit.wikimedia.org/r/442085 [12:12:17] (03CR) 10Hashar: "check experimental" [extensions/ORES] - 10https://gerrit.wikimedia.org/r/442085 (owner: 10Hashar) [12:13:05] 10Scoring-platform-team, 10ORES, 10Browser-Tests: ORES webdriver.io selenium test fail on CI due to lack of ORES server - https://phabricator.wikimedia.org/T198201#4315032 (10hashar) [12:15:19] (03PS2) 10Hashar: selenium: skip test on CI [extensions/ORES] - 10https://gerrit.wikimedia.org/r/442085 [12:15:31] (03CR) 10Hashar: "check experimental" [extensions/ORES] - 10https://gerrit.wikimedia.org/r/442085 (owner: 10Hashar) [12:17:08] 10Scoring-platform-team, 10JADE, 10Design: Design conceptual prototype of JADE integration with MediaWiki - https://phabricator.wikimedia.org/T182829#4315145 (10awight) 05Open>03Resolved a:03awight [12:18:01] 10Scoring-platform-team, 10JADE: JADE: Validate endorsement user - https://phabricator.wikimedia.org/T196544#4315162 (10awight) 05Open>03Invalid [12:18:57] 10Scoring-platform-team, 10JADE, 10Design: JADE UI should provide an explanation of the target ORES score - https://phabricator.wikimedia.org/T181327#4315165 (10awight) See the experimental LIME work in T196475. [12:21:26] 10Scoring-platform-team (Current), 10ORES: Design JADE data storage schema - https://phabricator.wikimedia.org/T153152#4315173 (10awight) [12:21:28] 10Scoring-platform-team, 10JADE, 10ORES: Design a collaborative judgement pattern for JADE - https://phabricator.wikimedia.org/T171496#4315170 (10awight) 05Open>03Resolved a:03awight This looks good. For the first iteration, we're providing: * Collaboratively edited "preferred" judgment in case consen... [12:21:50] 10Scoring-platform-team, 10ORES: Meta ORES: API data storage and querying - https://phabricator.wikimedia.org/T153145#4315175 (10awight) 05Open>03Resolved a:03awight [12:21:52] 10Scoring-platform-team, 10JADE, 10ORES: JADE: UI/API for reviewing/refuting how ORES classifies you and your stuff - https://phabricator.wikimedia.org/T148700#4315177 (10awight) [12:23:20] 10Scoring-platform-team, 10JADE: Meta ORES: UI - https://phabricator.wikimedia.org/T153148#4315179 (10awight) [12:24:00] 10Scoring-platform-team, 10JADE: Meta ORES: Link from score'd edit to this tool feedback interface - https://phabricator.wikimedia.org/T153157#4315180 (10awight) [12:24:18] 10Scoring-platform-team, 10JADE, 10Design: Design curation tools for meta-ORES - https://phabricator.wikimedia.org/T153161#4315181 (10awight) [12:24:30] 10Scoring-platform-team, 10JADE, 10ORES: JADE: UI/API for reviewing/refuting how ORES classifies you and your stuff - https://phabricator.wikimedia.org/T148700#4315184 (10awight) [12:24:32] 10Scoring-platform-team, 10AbuseFilter, 10ORES: [Spec] Suppression system for JADE freeform text fields - https://phabricator.wikimedia.org/T153142#4315182 (10awight) 05Open>03Invalid [12:25:40] 10Scoring-platform-team, 10JADE, 10Wikilabels: [Discuss] Could Wiki Labels backend be replaced with Meta ORES? - https://phabricator.wikimedia.org/T153150#4315189 (10awight) We're looking at similar integrations with patrolling, PageTriage, and FlaggedRevs. Wiki labels should fit into this pattern nicely. [12:25:54] 10Scoring-platform-team, 10JADE, 10ORES: Spike: Design how Meta ORES feedback plugs back into the ORES models - https://phabricator.wikimedia.org/T153162#4315191 (10awight) [12:27:27] 10Scoring-platform-team, 10ORES: [Spike] Should we integrate JADE with Flow? - https://phabricator.wikimedia.org/T153147#4315196 (10awight) 05Open>03declined We've decided to use talk pages and leave the coupling very loose, just that a Jade page is 1:1 with a Jade_talk page, with no additional structure.... [12:27:30] 10Scoring-platform-team, 10JADE, 10ORES: JADE: UI/API for reviewing/refuting how ORES classifies you and your stuff - https://phabricator.wikimedia.org/T148700#4315198 (10awight) [12:29:47] 10Scoring-platform-team, 10JADE, 10ORES: [Spec] ORES query results should include Meta-ORES refutations - https://phabricator.wikimedia.org/T153143#4315201 (10awight) [12:30:29] 10Scoring-platform-team (Current), 10ORES: Design JADE data storage schema - https://phabricator.wikimedia.org/T153152#4315205 (10awight) [12:30:31] 10Scoring-platform-team, 10JADE, 10ORES: JADE: UI/API for reviewing/refuting how ORES classifies you and your stuff - https://phabricator.wikimedia.org/T148700#4315206 (10awight) [12:30:33] 10Scoring-platform-team, 10ORES: [Spec] Design meta ORES general wiki entity matching schema - https://phabricator.wikimedia.org/T153149#4315202 (10awight) 05Open>03Resolved a:03awight We can target Diff, Revision, and Page, which is sufficient for the moment. [12:30:53] 10Scoring-platform-team, 10ORES: Meta ORES: API data storage and querying - https://phabricator.wikimedia.org/T153145#4315216 (10awight) [12:30:55] 10Scoring-platform-team, 10JADE, 10Spike: [Spike] potential technical implementations of JADE - https://phabricator.wikimedia.org/T166053#4315212 (10awight) 05Open>03Resolved a:03awight [12:32:07] 10Scoring-platform-team, 10JADE: Meta ORES: UI - https://phabricator.wikimedia.org/T153148#4315220 (10awight) [12:32:08] 10Scoring-platform-team, 10JADE, 10Design: Design curation tools for meta-ORES - https://phabricator.wikimedia.org/T153161#4315217 (10awight) 05Open>03Resolved a:03awight >>! In T153161#2895624, @Nemo_bis wrote: > Please don't reinvent the wheel. Store your data in actual pages (or in Special:Log) and... [12:39:10] 10Scoring-platform-team, 10JADE: Write JADE internal APIs to simplify integrations - https://phabricator.wikimedia.org/T198207#4315245 (10awight) [12:39:37] 10Scoring-platform-team, 10JADE: Write JADE internal APIs to simplify integrations - https://phabricator.wikimedia.org/T198207#4315258 (10awight) [12:39:39] 10Scoring-platform-team, 10Huggle, 10JADE: Use JADE as a repository for ORES counterexamples - https://phabricator.wikimedia.org/T197098#4315261 (10awight) [12:40:47] (03PS3) 10Hashar: selenium: skip test on CI [extensions/ORES] - 10https://gerrit.wikimedia.org/r/442085 (https://phabricator.wikimedia.org/T198201) [12:41:01] (03CR) 10Hashar: "check experimental" [extensions/ORES] - 10https://gerrit.wikimedia.org/r/442085 (https://phabricator.wikimedia.org/T198201) (owner: 10Hashar) [12:45:51] (03CR) 10Hashar: [C: 04-1] selenium: skip test on CI [extensions/ORES] - 10https://gerrit.wikimedia.org/r/442085 (https://phabricator.wikimedia.org/T198201) (owner: 10Hashar) [12:55:18] (03PS4) 10Hashar: selenium: skip test on CI [extensions/ORES] - 10https://gerrit.wikimedia.org/r/442085 (https://phabricator.wikimedia.org/T198201) [12:56:53] (03Abandoned) 10Hashar: selenium: skip test on CI [extensions/ORES] - 10https://gerrit.wikimedia.org/r/442085 (https://phabricator.wikimedia.org/T198201) (owner: 10Hashar) [13:00:12] 10Scoring-platform-team, 10ORES, 10Browser-Tests, 10Patch-For-Review: ORES webdriver.io selenium test fail on CI due to lack of ORES server - https://phabricator.wikimedia.org/T198201#4315316 (10hashar) The patch would skip the ORES selenium tests just fine however mediawiki/core tests end up failing.... [13:05:13] 10Scoring-platform-team (Current), 10ORES, 10Browser-Tests, 10Patch-For-Review: ORES webdriver.io selenium test fail on CI due to lack of ORES server - https://phabricator.wikimedia.org/T198201#4315331 (10awight) @hashar Thanks for the investigation! Can you assign this an urgency? I don't understand whe... [13:06:39] 10Scoring-platform-team (Current), 10ORES, 10Browser-Tests, 10Patch-For-Review: ORES webdriver.io selenium test fail on CI due to lack of ORES server - https://phabricator.wikimedia.org/T198201#4315334 (10awight) See also our task to integrate a working ORES service under docker: T195073. If that would be... [13:29:43] o/ [13:41:59] 10Scoring-platform-team (Current), 10ORES, 10Browser-Tests, 10Patch-For-Review: ORES webdriver.io selenium test fail on CI due to lack of ORES server - https://phabricator.wikimedia.org/T198201#4315504 (10hashar) Right now patches sent to mediawiki/extensions/ORES do not run the Selenium tests. I am in a p... [14:22:40] 10Scoring-platform-team (Current), 10ORES, 10Browser-Tests, 10Patch-For-Review: ORES webdriver.io selenium test fail on CI due to lack of ORES server - https://phabricator.wikimedia.org/T198201#4315627 (10awight) @hashar Okay I see, we're blocking the migration, maybe we have a few weeks to act? So does i... [14:24:16] Amir1: halfak: Some patches to review, ending in https://gerrit.wikimedia.org/r/#/c/mediawiki/extensions/JADE/+/442074 [14:30:57] srrodlund: o/ doxxing now if you're around [14:42:03] halfak: note for later--I want to chat about that AggregatorScalar thing, noticed some weird stuff. [15:44:43] 10Scoring-platform-team, 10JADE: Spec version indicator for JADE content - https://phabricator.wikimedia.org/T198224#4315881 (10awight) [15:46:53] 10Scoring-platform-team, 10JADE: Spec version indicator for JADE content - https://phabricator.wikimedia.org/T198224#4315900 (10awight) [15:49:32] halfak: I'm happy with our decision to keep both "notes" and the talk page. It isn't necessarily data duplication, because we're leaving the user free to annotate in either location. [15:50:19] Also, I was gonna give a shot at defining an internal API for e-z integration. [15:52:11] 10Scoring-platform-team, 10articlequality-modeling, 10artificial-intelligence: Add wp10 model in PetScan - https://phabricator.wikimedia.org/T197871#4315935 (10Halfak) @Ladsgroup, it looks like we're blocking here. Should we create a new task for enabling ORES ext and wp10 tracking to euwiki? [15:52:54] +1 awight [15:53:27] awight, re. the aggregatorscalar thing. I think I worked it out. It's related to making aggregators work on both single values and vectors. [15:53:46] Preping for tech management meeting, but I should have a PR soon afterwards. [15:53:57] ah. Okay so what I noticed is that the Python and REST APIs return different features for calculations. [15:55:26] py, cache: ('feature.log((wikitext.revision.parent.chars + 1))', 5.71042701737487), [15:55:29] API: "feature.wikitext.revision.parent.chars": 301.0, [15:56:44] halfak: I think I see what you're saying. The generator breaks .name inheritance [15:57:04] right. [15:57:27] one of those is log'd [15:58:12] halfak: Unrelatedly, I'm thinking about a new URL scheme: Jade:Diff/12345678/damaging, but reading py, cache: ('feature.log((wikitext.revision.parent.chars + 1))', 5.71042701737487), [15:58:16] ugh [15:58:29] reading Jade:Diff/12345678 still works as before (magically) [15:58:57] That page could be created by default as a lua macro to transclude subpages [16:01:48] awight, interesting. This would end up creating more pages. [16:02:00] We'd considered the sub-page pattern briefly in the past. [16:02:10] What's got you interested in it again? [16:03:45] halfak: Another topic: I assume that title blocking already applies to every namespace, nothing we have to do? [16:03:57] "title blocking"? [16:04:08] "title protection" [16:05:45] Sorry. Still not sure I know what you are referring to. [16:05:54] halfak: re. sub-pages, I like it now because we can turn the talk page into "notes" and drop from the formal schema. And it only requires linearly more resources, so not too painful. [16:06:08] awight, I think notes are content data [16:06:20] Getting them right should be iterative. [16:06:26] Okay I buy the content argument too. [16:06:34] You can improve upon the notes by adding details. [16:06:48] halfak: title protection is how certain article titles which are too offensive to mention, but irresistable to vandals, are impossible to create for mortal users. [16:07:00] Oh! [16:07:01] I see. [16:07:05] I don't know anything about it. [16:07:11] hopefully we get it for free. [16:07:38] So that no one creates Jade:Foobar? [16:07:43] yes [16:08:53] Got it. [16:09:14] +1 that notes belong in both content and talk, for now. My motivation to delay as much content as possible at least until we have a UI or trial integration is that moderation will have to be done in twitchy JSON, which may be unpleasant to editors. [16:09:41] Rollback should be more than sufficient tho [16:12:26] awight, fair point. Maybe we can implement a custom diff of some sort. [16:12:44] E.g. just removing cruft -- showing changes to notes, data, guid only. [16:14:27] I just realized the talk page thing wouldn't work as notes unless we get down to ///, otherwise the talk page still applies to multiple judgments. [16:14:46] seems like an anti-pattern, to make a REST tree out of wiki pages. [16:14:56] +1 for the custom diff, ty [16:16:05] 10Scoring-platform-team, 10JADE, 10Design: Custom diff for JADE content - https://phabricator.wikimedia.org/T198226#4316056 (10awight) [16:16:28] halfak: jfyi I'm working under the assumption that we dropped guid [16:16:38] Oh yeah. That makes sense. [16:17:10] Yeah. I don't think the talk pages will be structured in any way. [16:18:16] remember when all Strategy talk pages were split into three parts, for each major fork of the strategy? [16:18:50] It was cute at times, but generally I think it may have diluted conversation, working against critical mass. [16:19:04] hmm... I do not. But I can see what you're saying. [16:19:20] I do agree that one talk page per diff/revision/page is right. [16:19:45] Often damaging/goodfaith/edittype/etc. can be discussed all at once. [16:23:55] harej: I'm wondering how a wiki community would, relatively, view the Jade namespace if it's only attracting a few legit edits and some spam, vs. if we can quickly populate it with data streams coming from existing workflows. Your thoughts? [16:24:21] halfak: great point > Often damaging/goodfaith/edittype/etc. can be discussed all at once. [16:42:59] Interesting. I'm not sure if tool authors would prefer to pass JADE judgments as flat parameters; or as an array with the same structure as page content; or as mw-core native objects like Revision. [16:43:10] halfak: harej: ^ no rush [16:43:10] awight: it’s worth talking to Community members about but in their shoes I would be leery of a random new namespace popping up with all sorts of activity that I didn’t understand. I would be more inclined to favor a gradual ramp up or if community members really want this have them go to bat for it. [16:43:30] +1 okay that sounds wise [16:44:10] Not sure off-hand how to ramp up, it would be odd to take 1% of patrolling, for example. [16:45:54] re: community members wanting JADE, we should strategize about that. What would obviously deliver value to them? [16:47:02] I was thinking about providing a means to share information between patrolling, FlaggedRevs, and PageTriage, but they might * not want that and * not overlap much on a given wiki, which probably chooses one or the other. [16:48:56] to my other question, I think I'll go with flat parameters. Should be easier even when the client has loaded a judgment, updated, and is re-saving. [16:49:43] The only drawback of that interface is that conflicts are harder to detect correctly, so we'll have to be more conservative when merging automatically. [16:50:56] i.e. updating by passing the full content page makes section granularity easier. [16:56:41] 10Scoring-platform-team (Current), 10ORES, 10Documentation: Draft of ORES threshold optimization documentation - https://phabricator.wikimedia.org/T198232#4316359 (10awight) [16:57:27] 10Scoring-platform-team (Current), 10ORES, 10Documentation: Draft of ORES threshold optimization documentation - https://phabricator.wikimedia.org/T198232#4316374 (10awight) [17:00:30] * halfak --> lunch [17:09:42] nasty. I'm considering an internal API to "upsert judgment", and it looks like we need to always create a new judgment rather than replace an existing one. [17:10:24] i.e., I think we have to decide between the multiple-judgment or single-judgment model as the default behavior. [17:12:10] Might as well choose to insert and have multiple judgments when integrating existing workflows, and newer tools which are aware of existing judgments can choose to update or insert. [17:21:36] harej: https://etherpad.wikimedia.org/p/JADE_API [17:22:08] gtg for now! [17:23:09] 10Scoring-platform-team, 10JADE: Write JADE internal APIs to simplify integrations - https://phabricator.wikimedia.org/T198207#4316540 (10awight) [17:27:12] 10Scoring-platform-team, 10Huggle, 10JADE: Use JADE as a repository for ORES counterexamples - https://phabricator.wikimedia.org/T197098#4278749 (10Framawiki) >>! In T197098#4315064, @awight wrote: > I assume you're sharing some type of data with other Huggle users to prevent them from redundant review of th... [17:42:54] (03PS1) 10Umherirrender: Fix types of ThresholdLookup::prepareThresholdRequestParam [extensions/ORES] - 10https://gerrit.wikimedia.org/r/442155 [18:08:35] Damn. Wanted to ask awight what he's talking about. [18:08:45] I can't imagine a scenario where we want multiple judgements. [18:08:59] Unless we have endorsements and the preference bit. [18:34:58] It sounds like a plausible scenario to me. People often disagree on the value of edits. [18:35:56] harej, yes. That would be what endorsement would be fore. [18:35:57] *for [18:36:19] Then we have multiple judgments with some information about what they mean. [18:37:38] Multiple judgments with no structured information about what the differences are seems counter-productive. [19:29:51] * halfak heads to doc. Back in ~ an hour [20:17:32] (03PS1) 10Awight: Invalid extra property at the top level [extensions/JADE] - 10https://gerrit.wikimedia.org/r/442195 [20:17:34] (03PS1) 10Awight: Fix tests; schema prevents extra top-level properties [extensions/JADE] - 10https://gerrit.wikimedia.org/r/442196 [20:21:42] (03Abandoned) 10Awight: Fix tests; schema prevents extra top-level properties [extensions/JADE] - 10https://gerrit.wikimedia.org/r/442196 (owner: 10Awight) [20:24:52] (03Restored) 10Awight: Fix tests; schema prevents extra top-level properties [extensions/JADE] - 10https://gerrit.wikimedia.org/r/442196 (owner: 10Awight) [20:24:57] (03Abandoned) 10Awight: Invalid extra property at the top level [extensions/JADE] - 10https://gerrit.wikimedia.org/r/442195 (owner: 10Awight) [20:25:08] (03PS2) 10Awight: Fix tests; schema prevents extra top-level properties [extensions/JADE] - 10https://gerrit.wikimedia.org/r/442196 [21:56:05] well that took longer than expected!