[08:34:19] <wikibugs>	 10Scoring-platform-team-Backlog, 10Wikilabels, 10editquality-modeling, 10artificial-intelligence: Complete Albanian Wikipedia editquality campaign - https://phabricator.wikimedia.org/T163010#3358136 (10Arianit) Done. Thanks!
[14:58:11] <glorian_wd>	 halfak: o/
[15:01:43] <icinga2-wm>	 CUSTOM - Host ores-05 is UP: PING OK - Packet loss = 0%, RTA = 0.46 ms test
[15:10:09] <icinga2-wm>	 CUSTOM - Host ores-redis.01 is UP: PING OK - Packet loss = 0%, RTA = 0.58 ms test
[15:16:51] <icinga2-wm>	 PROBLEM - mem-check on ores-redis.01 is UNKNOWN: NRPE: Unable to read output
[15:16:51] <icinga2-wm>	 PROBLEM - mem-check on ores-web-05 is UNKNOWN: NRPE: Unable to read output
[15:16:52] <icinga2-wm>	 PROBLEM - mem-check on ores-worker-05 is UNKNOWN: NRPE: Unable to read output
[15:16:52] <icinga2-wm>	 PROBLEM - mem-check on ores-lb-02 is UNKNOWN: NRPE: Unable to read output
[15:17:04] <icinga2-wm>	 PROBLEM - mem-check on ores.wmflabs.org is CRITICAL: CHECK_NRPE: Socket timeout after 10 seconds.
[15:17:12] <paladox>	 Oh woops
[15:17:29] <icinga2-wm>	 PROBLEM - mem-check on ores-redis.01 is UNKNOWN: NRPE: Unable to read output
[15:17:30] <icinga2-wm>	 PROBLEM - mem-check on ores-web-05 is UNKNOWN: NRPE: Unable to read output
[15:17:31] <icinga2-wm>	 PROBLEM - mem-check on ores-worker-05 is UNKNOWN: NRPE: Unable to read output
[15:17:32] <icinga2-wm>	 PROBLEM - mem-check on ores-lb-02 is UNKNOWN: NRPE: Unable to read output
[15:17:50] <icinga2-wm>	 PROBLEM - mem-check on ores.wmflabs.org is CRITICAL: CHECK_NRPE: Socket timeout after 10 seconds.
[15:18:07] <icinga2-wm>	 PROBLEM - mem-check on ores-redis.01 is UNKNOWN: NRPE: Unable to read output
[15:18:08] <icinga2-wm>	 PROBLEM - mem-check on ores-web-05 is UNKNOWN: NRPE: Unable to read output
[15:18:10] <icinga2-wm>	 PROBLEM - mem-check on ores-worker-05 is UNKNOWN: NRPE: Unable to read output
[15:18:11] <icinga2-wm>	 PROBLEM - mem-check on ores-lb-02 is UNKNOWN: NRPE: Unable to read output
[15:18:38] <icinga2-wm>	 PROBLEM - mem-check on ores.wmflabs.org is CRITICAL: CHECK_NRPE: Socket timeout after 10 seconds.
[15:18:46] <icinga2-wm>	 PROBLEM - mem-check on ores-redis.01 is UNKNOWN: NRPE: Unable to read output
[15:18:47] <icinga2-wm>	 PROBLEM - mem-check on ores-web-05 is UNKNOWN: NRPE: Unable to read output
[15:18:50] <icinga2-wm>	 PROBLEM - mem-check on ores-worker-05 is UNKNOWN: NRPE: Unable to read output
[15:18:50] <icinga2-wm>	 PROBLEM - mem-check on ores-lb-02 is UNKNOWN: NRPE: Unable to read output
[15:19:14] <paladox>	 fixed it now
[15:23:01] <icinga2-wm>	 PROBLEM - ssh on ores.wmflabs.org is CRITICAL: CRITICAL - Socket timeout after 10 seconds
[15:28:08] <paladox>	 ignore ssh on ores.wmflabs.org. it's now monotoring the site but ssh should not work for the site anyways :)
[15:53:07] <wikibugs>	 10Scoring-platform-team-Backlog, 10Wikilabels, 10editquality-modeling, 10artificial-intelligence: Complete Albanian Wikipedia editquality campaign - https://phabricator.wikimedia.org/T163010#3358441 (10Halfak) Great!  We'll get a model built! Thanks for your work.  We'll ping you when we're ready to deploy...
[15:53:26] <wikibugs>	 10Scoring-platform-team, 10Wikilabels, 10editquality-modeling, 10artificial-intelligence: Complete Albanian Wikipedia editquality campaign - https://phabricator.wikimedia.org/T163010#3358442 (10Halfak)
[15:53:38] <wikibugs>	 10Scoring-platform-team, 10editquality-modeling, 10artificial-intelligence: Train/test damaging & goodfaith models for Albaian Wikipedia - https://phabricator.wikimedia.org/T163009#3358443 (10Halfak)
[15:57:01] <glorian_wd>	 halfak: o/
[15:57:34] <halfak>	 o/ glorian_wd 
[15:57:46] <halfak>	 back in a bit.  getting breakfast together. 
[16:00:15] <glorian_wd>	 halfak: ok. 
[16:00:15] <glorian_wd>	 I am now working on the PR. As you said yesterday I should write: external_sources_ratio = div(all_external_sources, sources). However, sources can be 0 because there could be items that have no sources at all. In this case, I am thinking to use the try-except. 
[16:00:15] <glorian_wd>	 But I want to ask, if you could propose a better approach for dealing with this.
[16:00:50] <glorian_wd>	 I'd prefer to ask you in advance before I got a comment from you again  in PR xD
[16:15:36] <glorian_wd>	 BTW, I am talking about your feedback: "This processing should depend on [source, external_sources] so that work isn't duplicated."
[16:34:49] <halfak>	 glorian.  Look at all the other uses of "/" in features. 
[16:35:23] <halfak>	 they will do <some feature> / max(other_feature, 1)
[16:36:27] <glorian_wd>	 halfak: alright
[16:36:31] <glorian_wd>	 I'll have a look
[16:37:16] <glorian_wd>	 halfak: BTW, I will try to re-submit the PR before I go to sleep today
[16:38:31] <halfak>	 kk  I should be able to review.  I need a day off, but I'll be around for a few more hours. 
[16:39:02] <glorian_wd>	 halfak: Okay :)
[17:23:04] <glorian_wd>	 halfak: PR submitted
[18:03:39] <halfak>	 glorian_wd, I have serious concerns with the PR's performance. 
[18:04:02] <halfak>	 You're really opening a file and reading through it sequentially multiple times per revision scored!? 
[18:04:08] <halfak>	 I'm reading that right, right? 
[18:04:35] <halfak>	 Are you familiar with big O notation for computational complexity?
[18:04:42] <glorian_wd>	 halfak: hmm yes. 
[18:05:09] <glorian_wd>	 I'll think some better way for it
[18:05:17] <halfak>	 What's the complexity of "_process_wikimedia_sources"?
[18:07:01] <halfak>	 Currently, you're running O(M * N) where M is the number of sources and N is the number of exclusions. 
[18:07:07] <halfak>	 You could instead just be O(M) 
[18:07:24] <halfak>	 By putting the internal sources in a `set`
[18:07:36] <halfak>	 You should expect (practically) constant-time lookups. 
[18:07:52] <halfak>	 Also you won't have to incur the (SUPER DUPER SLOW!!) IO of opening a file and reading it into memory. 
[18:10:17] <glorian_wd>	 Hmm. Your concern is more to the opening file right?
[18:10:37] <glorian_wd>	 not about when I unpack the item.claims
[18:11:44] <halfak>	 Let me ask this.  What's your assessment?
[18:12:08] <icinga2-wm>	 hi all lol
[18:12:20] <halfak>	 ha
[18:13:04] <paladox>	 I just registered the bot for an irc cloak :)
[18:13:18] <paladox>	 Also ores.wmflabs.org is now monotored by icinga2 too :)
[18:13:30] <halfak>	 Saw the memory pings a few hours ago
[18:13:33] <halfak>	 Testing?
[18:14:08] <paladox>	 halfak those were a mistake
[18:14:13] <halfak>	 gotcha
[18:14:23] <paladox>	 those wont work on your instances due to the command not being on the instance (it's in puppet but i had to manually copy the file and command)
[18:14:47] <paladox>	 I was migrating things from icinga2 director to public view in the repo
[18:14:58] <paladox>	 https://phabricator.wikimedia.org/diffusion/LICT/
[18:17:18] <mavrikant>	 halfak:  https://phabricator.wikimedia.org/T164671
[18:17:18] <mavrikant>	 I class is missing here "başlanğıç"
[18:17:58] <mavrikant>	 https://github.com/wiki-ai/wikiclass/blob/master/wikiclass/extractors/trwiki.py
[18:17:58] <mavrikant>	 > ("baslağıç", re.compile(r"\bbaşlanğıç\b", re.I)), # start class
[18:17:58] <mavrikant>	 Its here but not in results.
[18:20:04] <halfak>	 Oh!  Gotcha.  I'll look into it.  Thanks for pointing it out. 
[18:20:33] <halfak>	 paladox, didn't we have a patchset for enabling it in the base ORES role?
[18:20:50] <paladox>	 halfak think so but i think that was rejected
[18:21:00] <halfak>	 mavrikant, can you find me a page in that class? 
[18:21:16] <paladox>	 halfak we should just manually add it since they rejected your change.
[18:21:37] <halfak>	 paladox, what?  It was rejected?  
[18:21:38] <mavrikant>	 halfak:  https://tr.wikipedia.org/w/index.php?title=Kategori:Ba%C5%9Flang%C4%B1%C3%A7-s%C4%B1n%C4%B1f_Vikiproje_sayfalar%C4%B1&action=edit&redlink=1
[18:21:39] <mavrikant>	 all pages in here
[18:21:39] <halfak>	 link?
[18:21:48] <halfak>	 thanks mavrikant 
[18:22:08] <paladox>	 yep
[18:22:09] <paladox>	 https://gerrit.wikimedia.org/r/#/c/358240/
[18:22:58] <halfak>	 paladox, do we have a task for this?
[18:23:08] <paladox>	 uep
[18:23:09] <paladox>	 ep
[18:23:10] <paladox>	 yep
[18:23:10] <halfak>	 I want to ask akosiaris what he suggests in a place where we can have a good conversation
[18:23:11] <paladox>	 https://phabricator.wikimedia.org/T167602
[18:23:46] <paladox>	 i think they should just stick those checks in base for everyone to have access including labs.
[18:24:26] <wikibugs>	 10Scoring-platform-team, 10Patch-For-Review, 10User-Zppix: Create memory checks for instances - https://phabricator.wikimedia.org/T167602#3358540 (10Halfak) @akosiaris, I saw that you -2'd https://gerrit.wikimedia.org/r/#/c/358240/.  It seems we want to set up these checks.  So.  What's a good alternative fr...
[18:24:42] <halfak>	 paladox, I replied there.  We'll see. 
[18:24:46] <paladox>	 ok
[18:24:48] <paladox>	 thanks
[18:25:22] <paladox>	 halfak do you use a puppet master?
[18:25:33] <halfak>	 same as everyone else
[18:25:38] <halfak>	 No custom one
[18:26:07] <paladox>	 oh
[18:43:18] <halfak>	 mavrikant, looks like "Başlangıç" is the right string. 
[18:43:55] <halfak>	 But we were matching "Başlanğıç"
[18:44:07] <halfak>	 So I just replaced "ğ" with "[gğ
[18:44:10] <halfak>	 ']"
[18:44:12] <halfak>	 Woops
[18:44:17] <halfak>	 Does that make sense?
[18:44:38] <halfak>	 Could it be that "Başlanğıç" never occurs even once?
[18:45:10] <mavrikant>	 sorry my bad. "Başlangıç"  is true one.
[18:45:18] <halfak>	 Oh!  OK.  
[18:47:25] <halfak>	 OK re-extracting.  This will take a while, but with any luck, I'll have it ready before I need to travel tomorrow. 
[18:47:28] <halfak>	 mavrikant, ^ 
[18:47:40] <halfak>	 I'll become kind of unavailable next week as a do a management offsite thing. 
[18:48:40] <mavrikant>	 okey. thanks. we can use next to examine results. no problem.