[13:28:25] o/ [13:29:38] early morning meeting. ;) [13:32:44] Hi [13:36:38] am_xlia, hi [13:37:19] How is everyone [13:41:05] I'm feeling scattered while I get my coffee into me :) [17:58:21] hiya, is there someone here who knows all the different mysql / eventlogging databases and hosts that are used for research? [17:58:26] trying to get a handle of what is there and why [18:22:18] ottomata, [18:22:21] I might be able to help [18:22:38] I know about analytics-store and analytics-slave [18:22:53] oh and db1047 [18:23:05] halfak: i just posted my q: https://phabricator.wikimedia.org/T156844#3107357 [19:43:13] halfak: leila question (very short) [19:43:21] o/ [19:43:52] i need to put in meta the design document about AB testing that ellery timo, brandon and myself work on [19:43:55] *worked on [19:44:10] is there a specific location where that would go? [19:44:24] ^ halfak leila [19:44:56] https://meta.wikimedia.org/wiki/Research:New_project [19:44:57] Probably [19:45:14] You're already done with step 3 :) [19:46:50] halfak: oohhh [19:46:54] i am glad i asked [19:47:24] :) [19:47:31] just arrived back with my lunch. [19:47:53] Noo run away. Eat lunch away from PC (if you can!) :D [19:48:28] it's tough man! how do you do it when you're working from home, halfak? [19:49:06] Oh mostly I don't succeed in not eating in front of my desk. [19:49:11] (eating is not enough of an activity on its own for me. I need some entertainment. ;) [19:49:15] In the summer though, I take yardwork breaks and they are AMAZING [19:49:30] yeah. /that/ I can do [20:09:29] what are these “breaks” that you people talk of? :P [20:18:16] Their like when you put the brakes on work for a little bit. [20:18:36] Put the brakes on and have a break man. ;D [20:18:47] * halfak joins documentation time and sees noone [20:18:49] :( [20:51:10] schana: do you see more than one "missing" recommended in AR.json? [20:51:28] or is your question about ordering of sections about what comes in "current"? [20:53:59] leila: ordering for "current" [20:54:33] I see. The ordering doesn't matter for current. [20:55:28] schana: the thinking is that if this set of sections is available in an article, then section x should also be available with probability equal to confidence. [20:57:09] did the research showcase happen yesterday? it's in the calendar but no mention in the wiki or youtube [21:00:26] tgr: we have missed it. :( [21:04:16] leila: could someone remove the youtube video then to avoid confusing people? [21:04:43] yes, tgr. /me checks the calendar. [21:04:48] btw we seem to have multiple youtube channels called "Wikimedia Foundation", it would be nice if they cross-referenced each other [21:05:34] tgr: which calendar are you seeing Research Showcase in? I don't see it in fun and learning and staff [21:05:53] sorry, take it back. it's in staff, tgr [21:08:56] yup, in staff and there is a prescheduled youtube broadcast thingie which comes up on top when you google research showcase but then just says it has not started yet [21:09:31] (while I am here, these showcases tend to be great and thanks to whoever is organizing them!) [21:09:45] tgr: I asked Sarah to remove the event altogether. I don't have access to be able to do it. Sorry about the confusion we've caused. [21:10:33] and tgr: thanks for the word of motivation. We really need them on this front, as you can see. :) [21:11:40] halfak: I just saw that you had a speaker confirmed for the research showcase in March: https://phabricator.wikimedia.org/T156138 [21:12:00] Yeah. that was a sad mixup :( [21:12:09] so, what shall we do? [21:12:29] It's too late now. Might re-book Yan for a later showcase. [21:13:34] ok. I wanted to reach out to Markus K. to see if he or someone from his team would be interested to talk about https://meta.wikimedia.org/wiki/Research:Understanding_Wikidata_Queries Shall I ask him for April? [21:13:38] halfak, ^ [21:15:00] Yeah +1 [21:17:33] ok, I'll do that, halfak, then. Please make sure you communicate with Yan (I'm assuming you are in contact with her). If you need my help there, let me know. [21:17:51] Yeah. Already did. Will tell you a story over a beer at some point :S [22:14:34] halfak: If it is interesting to you, I've created a Wikidata item on the doi.org query endpoint. https://www.wikidata.org/wiki/Q28946522 [22:18:57] heh nice [22:19:14] harej, are you coming to wikicite? [22:19:19] yes [22:19:23] woot :) [22:19:32] Coming to the hackathon too then? [22:19:34] yes [22:19:38] woot again [23:36:10] J-Mo: the connection died. [23:50:00] schana: let's review the fields in category_templates file together. [23:50:39] sure, hangout? [23:50:46] schana: "category" is clear. what is "total"? [23:50:51] schana: here's fine if it works for you. [23:51:25] from what I can tell, total is the total number of times a specific category appears [23:51:59] with the 'template' set of sections [23:52:11] appearing 'counter' times [23:52:36] and the probability is 'counter' / 'total' [23:53:28] correct. [23:53:34] that's my understanding, too, schana. [23:54:05] then the heuristic you suggested in https://phabricator.wikimedia.org/T159705 should be changed or it will contain too many potentially low quality recommendations. [23:55:40] I'm not sure how it could be improved for this dataset, since there rarely appears to be entries with 'counter' > 1 [23:56:19] can you give a percentage, schana? [23:58:03] ...running query...