[00:46:24] have fun halfak :) [00:57:24] More emails, but I have a few tasks made :) [01:27:04] halfak: one thing: Your code loads user data no matter what, I want to change it to load only when it's asked in arguments [01:27:11] would it be okay with you? [01:28:20] +1 [01:28:30] So one of the problems I have worked out is this: [01:28:46] When we are doing damaging revert detection, we only want to load user data on reverted edits. [01:29:27] When we are doing maybe-damaging-edit detection (for wiki labels), we want to check if all users are blocked and in a trusted group before checking if they were reverted. [01:29:28] L/ [01:29:30] *:/ [01:30:11] Our cue if the "reverted-only" flag [01:30:24] *cue is [01:30:40] Which is poorly named maybe. [01:30:49] But not that poorly! [01:31:45] Yay! I get to code. [01:31:52] I have... 25 minutes. [01:32:08] I'm going to kill the sphinx [01:32:11] Or fix it [02:15:06] halfak: I pushed my changes [02:18:23] halfak: Oh I forgot, I'm working on the KDD paper, about writing references, should I use BibTeX or simple? [02:20:11] BibTeX [02:20:21] Assuming you are willing to tolerate latex. [02:20:56] Do you have a summary of methodological strategies from the papers you have read? [02:27:19] not yet [02:27:22] I just started [02:28:19] I'm latex tolerant, lactose intolerant but I can bear some latex :D [02:28:48] I know those methods in my mind [02:28:48] lol [02:29:08] haven't wrritten it yet [02:45:10] Sphinx is 100% insane [02:46:38] :))))) [02:46:46] afk for dishes [03:13:15] * halfak curses more at sphinx [03:49:23] back! [04:04:43] sphinx is still crazy [04:05:07] I have a replica of one branch of our directory tree and I'm sanity checking everything that I can with sphinx. [04:05:16] I still get two different results. [04:08:05] * halfak clears all __pycache__ [04:10:52] * halfak copies config files around [04:13:25] Alright. I give up. [04:13:35] This is the second time I have bashed my head against this. [04:13:43] Have a good night folks! [21:29:00] hey that's me [21:30:37] * YuviPanda waves at arlolra [21:30:44] \o [21:32:25] arlolra: halfak does allll the things [21:33:23] o/ al [21:33:31] * arlolra, [21:33:38] Thanks for your contribs. [21:33:52] I'm looking into that weird redis issue now. I know when it came up. It's a tricky one. [21:34:45] np. I'll hang here in case you prefer not to have your inbox flooded [21:35:02] :D [21:35:07] * halfak <3 PR pings [21:35:30] :D [21:37:08] arlolra, got it. Rebase plz :) [21:38:16] ok [21:39:34] Now it should work just fine without redis :) [21:39:46] This was a good bug to catch. :) [21:41:58] dropped the redis commit from https://github.com/wiki-ai/ores/pull/114 [21:42:23] and https://github.com/wiki-ai/ores/pull/116 fixes the travis complaints there [21:43:43] halfak: and yes, confirmed it works for me [21:43:55] w/o redis [21:46:09] {{merged}} 114 [21:46:10] 10[1] 04https://meta.wikimedia.org/wiki/Template:merged [21:46:22] Thanks [[User:AsimovBot]] [21:46:24] 10[2] 04https://meta.wikimedia.org/wiki/User:AsimovBot [21:47:54] {{merged}} again [21:47:54] 10[3] 04https://meta.wikimedia.org/wiki/Template:merged [21:47:56] Thanks arlolra [21:48:24] :) [21:53:23] arlolra, I think I'll need to talk to sage to think better re. params vs. JSON docs. [21:53:40] Since he's the user that initiated the request, he'll probably have a good perspective [21:57:32] halfak: sure, no rush. can you point me to what do with the params though? https://github.com/wiki-ai/ores/pull/115/files#diff-a77c187c926782c58ac0062c451265beR145 [21:58:21] {{done}} [21:58:21] How efficient, halfak! [21:58:28] Thanks AsimovBot [21:59:59] wow [22:00:06] and now I'm late for my meeting with reading [22:00:09] see ya [22:57:25] halfak should we give the submission priority? [22:58:35] ToAruShiroiNeko, I think that we should make sure that final report is done. [22:58:43] But maybe we can do both. [22:58:50] What kind of panel would you like to organize. [22:58:51] ? [22:59:02] well [22:59:05] Also, how would it fit into wikimania's presentation types? [22:59:07] it would be the revision scoring team [22:59:20] we have had several of thoise [22:59:29] half presentation of our progress and half Q&A [23:00:05] Yeah. Seems like they are moving away from that format this year. [23:00:06] or 3/4 presentation v 1/4 q&a [23:00:24] I think such feedback would be useful [23:00:51] where we get to interract with interested people and they get to help discuss matters [23:01:57] ToAruShiroiNeko, still, there's no panel format. There's 42 slots for "critical issues" presentations. [23:02:17] do we qualify for that? [23:02:29] are we critical? [23:02:53] I do feel we are an afterthought at times and yet we are the best thing since sliced bread at other times :/ [23:02:56] Well, we need to talk about an issue. Not progress. [23:03:14] wikipedia does have plenty of issues [23:03:38] say gender equality [23:03:45] there is an AI for that [23:09:05] So, I'm not sure that I see the vision of this panel, but I'm willing to hear more. [23:09:34] I want to brainstorm actually [23:10:02] I dont think we have enough attention despite everything [23:10:56] Oh! Why is that? [23:12:53] hmm? [23:12:59] we have been covered on the media [23:13:09] despite that we didnt get an overflow of interest [23:13:19] not that I expected that [23:13:30] most people also still dont know what AI is about [23:13:40] We have 20 wikis. We receive about 5 requests per second for scores. [23:13:56] I want us to be a group where people bring complicated problems for us to solve [23:13:58] new problems [23:14:01] not just vandalism [23:14:19] we do not have that kind of attention yet [23:14:34] we are the counter vandalism guys, which is not bad but we can do more as we proceed [23:15:10] https://en.wikipedia.org/wiki/Wikipedia:Labels/Edit_types [23:17:18] yes [23:17:30] what else can we do, I want to ask that question [23:17:42] what is a problem community is seeking to be fixed [23:19:09] ToAruShiroiNeko, we barely have enough resources to address the modeling problems we already have though. [23:19:23] yes [23:19:28] FWIW, I added a bunch of request from the dev summit to our workboard. [23:19:38] demand would promote more resources [23:19:54] this is exactly what I mean by not enough attention [23:25:45] I'm not sure that more items in our backlog translates to more resources. [23:28:09] backlog isnt the thing [23:28:25] we can discuss our lack of resources maybe [23:28:37] explain why we need more resources [23:28:44] Now that's starting to sound like a critical issue. [23:28:52] But I don't think we can plead for our project. [23:28:58] and what would be the resource cost in helping with a certain problem or two [23:29:09] We should maybe talk about how these kind of projects don't have an easy route toward support. [23:29:18] if we can get the interest of a few chapters we may get some resources [23:29:31] we could talk about our experiences [23:29:37] and how we managed and how things can be better [23:29:41] we could do that too [23:30:21] and/or we could talk about a future plan [23:30:42] where we would hope to be and how we can support wikipedia in the future [23:32:42] that's sounding less like a critical issue. I think it would be helpful if you could read through their docs re submissions and figure out what format we could fit into. [23:34:41] sure [23:34:59] I want to give the submission a priority because this sunday is a hard deadline [23:35:03] where as it is not for the report [23:35:10] last report was a bit delayed [23:35:18] is this a problem for us? [23:35:21] OK. Let's figure out what we might like to submit first. [23:35:24] Not a problem [23:35:29] are we aiming for another renewal? [23:35:59] I think a breater would be nice but this is just my perspective [23:36:15] We're too late for another renewal, I think. [23:36:19] I am not implying we stop working, rather slow progress a little [23:36:25] Also, we were explicitly told not to do that. [23:36:37] I won't be slowing progress :) [23:36:46] No I mean with reporting and all [23:36:50] focus on the actual work :p [23:37:08] I want to play with ecoc and even implement the algorithm I used in my thesis [23:37:18] I still need to deliver you that :p [23:38:05] renewals dont have a deadline IIRC but I dont think a second renewal has happened before [23:39:02] even you seem to be having a hard time keeping up with friday resysncs :p [23:39:16] Travel makes that hard, yes. [23:39:17] to be fair it had taken the WMF all hands to sever you from us :D\ [23:39:22] yeah that too [23:39:34] thats what I mean by a breather because I fear we may be burning you out [23:40:17] we could do bi-weekly resyncs when you have a heavy travel schedule [23:40:59] or maybe move it to saturday [23:41:06] that would work best probably [23:41:18] na. The stuff that is burning me out isn't revscoring exactly. [23:41:27] I have been doing a lot of engineering work recently though [23:46:17] Do you have a plan in mind? [23:46:43] Yeah. A roadmap. [23:46:59] I'm trying to find a product manager for Wiki labels. [23:47:11] Amir1 is going to do product management for the ORES extension. [23:47:20] And I'm looking at the roadmap for revscoring/ORES [23:47:28] I thought that was you :p [23:47:43] Can't manage it all [23:47:45] :/ [23:47:50] hehe [23:47:59] everyone wants a peice of halfak, and amir too [23:48:39] do you have any person in mind or will you be asking the market? [23:50:08] or maybe you cannot discuss this [23:52:02] If there was funding, there could be a new person. [23:52:11] Right now, I'm trying to saddle it onto someone else >:) [23:52:40] * halfak begins debugging the sphinx documentation for the 3rd time. [23:52:41] I can volunteer for a while until you find someone [23:52:48] Really gonna need to rubber-duck this one. [23:53:30] halfak: new engineeeer! [23:54:47] The only part that's not covered now is working on design and engineering concerns around the extension. [23:55:10] YuviPanda, good point. [23:56:52] ToAruShiroiNeko, re. design and engineering concerns, violetto is looking at the design, but needs input on what to look at. [23:57:10] for labels? [23:57:11] bmansurov_gone, is working on the extension engineering, but I'm not sure where he is blocked right now. [23:57:14] Yeah. [23:57:56] Right now, all I can handle is comms and some engineering work, but I'm blocked on this damned sphinx nonsense. [23:57:57] do we have proposals? [23:58:17] I seem to work well with design people [23:58:21] :# [23:58:23] :3 [23:59:03] I have a sketch of an admin UI page somewhere.