[16:43:06] * lzia waves to the channel [16:50:28] o/ leila [16:55:24] bmansurov: hi. I'm reading T209882 and I'm thinking: was the survey going to run on WD to begin with? [16:55:24] T209882: Quicksurvey for reader trust - https://phabricator.wikimedia.org/T209882 [16:55:39] lzia: no, that's just for labs [16:55:51] bmansurov: oooo. got you. thanks. [16:57:49] miriam: it would be really great if you create a master task for organizing Wiki Workshop similar to the 2018 version (T171118), and create subtasks for it. You all will be doing a lot of work and it's good to track those in phab. ;) [16:57:49] T171118: [Objective 4.4.1.] Wiki Workshop 2018 - https://phabricator.wikimedia.org/T171118 [18:07:33] lzia: o/ [18:07:42] your ops nightmare is back [18:07:43] elukey: hiii. what's up? [18:07:46] :D [18:07:51] elukey: I'm ready. :D [18:07:55] I added some notes to https://phabricator.wikimedia.org/T172410#4803991 about the future of dbstore1002 [18:08:12] OW NO! ... (just kidding. /me reads) [18:08:23] just wanted to let you guys know and see if there are issues with the plan [18:08:37] elukey: what's the timeline to get back to you? [18:08:57] elukey: we need at least 2-3 days to discuss this within the team async. [18:09:24] oh yes Monday/Tue is fine [18:11:30] elukey: on it. thanks for the extra heads up [18:12:48] super :) [18:12:58] I'd rather wait a bit more and be sure that people are onboard [18:13:13] :) [19:40:00] hey bmansurov [19:40:07] (still doing emails and lunch stuff, gonna run home soon) [19:40:12] but, i have a q: [19:40:13] is this dataset you created [19:40:18] does it have PII? [19:40:21] could it be made public? [19:47:53] ottomata: yeah, it can be made public [19:48:00] nothing private in there [19:50:28] hm [19:50:33] and how big is it again? [19:51:04] ottomata: uncompressed, about 300MB [19:53:25] ok awesome [19:53:34] bmansurov: let's put it in /srv/published-datasets on stat1007 [19:53:46] that will be synced to a public location on analytics.wikimedia.org/datasets [19:53:52] and we can download it from there [19:54:10] ottomata: ok [19:54:16] if you put the data files there now somewhere, (maybe in the one-off dir for now) [19:54:27] then i can get back to you when i get home and we can dl it and run it on the mysql host [19:55:40] ottomata: do you want me to compress the files first? [19:55:55] sure why not! :) [19:58:35] ottomata: copied a 54MB tar.gz file to /srv/published-datasets/one-off/article-recommender/. It's the only file there. [19:59:08] great ok, that should sync sometime soon, if it doesn't by when I need it i'll manually do it [19:59:12] gonna run home now, be back in a bgit [20:26:00] thoughts / opinions on license for a github repo for analyzing reader behavior surveys? do we have a default that we use? i saw Apache and GPL 3.0 [20:31:02] isaacj: I usually go with GPL. [20:31:28] Either one is fine for your use case imo. [20:34:32] bmansurov: thanks! what i was leaning towards as well [20:36:37] isaacj: have a look at second bullet point under B in https://foundation.wikimedia.org/wiki/Open_access_policy [20:37:16] which matches the recommendation to go with GPL. [20:37:39] ahh - excellent! thanks for the link too [20:37:46] should have known something like this would exist [20:37:47] uhu [20:37:53] not really. ;) [20:38:48] isaacj: what should really exist is a landing page under Research that tells you what the expectations/recommendations are in terms of code and data release. [20:39:08] for data release, we even have documentation thanks to DarTar. but people won't know it without a proper pathway to it. [20:39:56] akhakh. now I see that we have Hangout and IRC discussions about it. :D [20:40:12] ok bmansurov let's do this! [20:40:27] ottomata: let's! [20:40:54] where's your script again? on stat1007 somewhere [20:40:54] ? [20:41:00] elukey: which team is data persistence team? (I feel awful not knowing the answer.) [20:42:01] ottomata: it's here https://gerrit.wikimedia.org/r/#/admin/projects/research/article-recommender/deploy [20:43:02] bmansurov: paste some commands to run somewhere for me? [20:43:14] ottomata: ok [20:49:48] oh but we don't have virtualenv or pip there [20:49:48] haha [23:50:07] lzia: o/ what's your favorite fawiki article? [23:51:04] * lzia thinks [23:51:16] :D [23:53:20] bmansurov: this is a vague question. I hope you appreciate it. [23:53:36] bmansurov: let's go with https://fa.wikipedia.org/wiki/محمود_دولت‌آبادی [23:54:27] I was going to show you the new API in farsi, but alas, it's not working. Try spanish: https://es.wikipedia.org/api/rest_v1/data/recommendation/article/morelike/translation/Hey_Jude [23:54:39] It's working on uzwiki and eswiki [23:54:53] It was supposed to work on fawiki, but nooo [23:54:57] lzia: ^ [23:55:44] something has to do with the arabic letters imo [23:56:39] bmansurov: let me give you some messy letters for the article. how about if you try https://fa.wikipedia.org/wiki/%D9%85%D8%AD%D9%85%D9%88%D8%AF_%D8%AF%D9%88%D9%84%D8%AA%E2%80%8C%D8%A2%D8%A8%D8%A7%D8%AF%DB%8C ? [23:56:43] * lzia checks the other link [23:57:10] no luck with messy letters either [23:57:20] ow no! :/ [23:57:37] bmansurov: what is the score? [23:57:50] what is the range of it, I mean? [23:57:57] the higher the better ;) [23:58:04] i forget the details [23:58:15] ;) [23:58:24] so Petr found why farsi isn't working, hopefully, it'll be deployed on Monday [23:58:25] it's good to know what's the range. [23:58:41] * lzia is thankful to Petr and all those involved. [23:58:57] it's the normalized score [23:59:21] i can look up the range when I'm little less busy ;) [23:59:22] bmansurov: so, are these Wikidata items that are closest to Hey_Jude and are missing from Spanish Wikipedia? [23:59:32] yes [23:59:43] bmansurov: sounds good re later for range.