[00:46:37] <halfak>	 hey ottomata, are you around?
[09:06:46] <grrrit-wm>	 (CR) Nuria: [C: 2] Make coalesced report format more efficient [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/146521 (https://bugzilla.wikimedia.org/67822) (owner: Milimetric)
[09:14:56] <grrrit-wm>	 (CR) Nuria: [V: 2] Make coalesced report format more efficient [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/146521 (https://bugzilla.wikimedia.org/67822) (owner: Milimetric)
[09:15:38] <grrrit-wm>	 (CR) QChris: Document 'refinery-' prefix for bin directory (1 comment) [analytics/refinery] - https://gerrit.wikimedia.org/r/145976 (owner: QChris)
[09:35:16] <grrrit-wm>	 (PS1) QChris: Document '.hql' filename ending for HiveQL files [analytics/refinery] - https://gerrit.wikimedia.org/r/147402
[09:37:54] <grrrit-wm>	 (CR) QChris: Document '.hql' filename ending for HiveQL files (1 comment) [analytics/refinery] - https://gerrit.wikimedia.org/r/147402 (owner: QChris)
[09:40:11] <grrrit-wm>	 (CR) Nuria: [C: 2 V: 2] "Vety handy. Tested on my machine and works as expected. Merging." [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/147225 (owner: QChris)
[09:55:37] <wikibugs>	 Analytics / General/Unknown: The page http://myMediawiki.dgfip/wt/index.php/Special:Pages_sp eciales crashes - https://bugzilla.wikimedia.org/68212 (DBoussant) UNCO p:Unprio s:major a:None This page is write and we read in the Apache's log:   PHP Fatal error:  Possible integer overflow in m...
[10:06:35] <nuria>	 qchris, hola
[10:06:44] <qchris>	 Hi nuria
[10:07:09] <nuria>	 for the "rolling active editors" (http://meta.wikimedia.org/wiki/Research:Rolling_active_editor)
[10:07:19] <nuria>	 we agreed registration date sdo not matter, right?
[10:07:24] <nuria>	 *do not matter
[10:07:40] <qchris>	 No clue. Let me check the tasking etherpad.
[10:09:36] <qchris>	 Mhmm. The etherpad says:
[10:09:44] <qchris>	 Assumption:    * "Story: WikimetricsUser runs 'Rolling Monthly Active Editors' report" is done
[10:09:58] <qchris>	 but does not mention registration date.
[10:10:54] <qchris>	 Also the like an bug https://bugzilla.wikimedia.org/show_bug.cgi?id=67458 goes to 'Rolling_active_editor' :-)
[10:11:12] <nuria>	 I think that bit is uneeded (from a meeting conversation we had, as this metric is supposed to measure "community size")
[10:11:17] <nuria>	 will note that on CR
[10:11:59] <qchris>	 So registration dates come into play here? Let me look at the code a bit.
[10:14:14] <qchris>	 nuria this is messed up ... The file is 'rolling_active_editor' but the commit message talks about 'monthly active editor' and 'rolling monthly active editors'
[10:14:27] <nuria>	 ya, i noted that on CR
[10:15:08] <qchris>	 Oh. Cool. (Your CR is not yet published, so I do not see it)
[10:15:14] <qchris>	 But does my confusion about
[10:15:21] <qchris>	 different metrics help with your question?
[10:15:25] <qchris>	 I guess not :-/
[10:15:25] <nuria>	 The sql matches "rolling (monthly) active editor"
[10:16:12] <nuria>	 I think not. I  will just publish CR noting the 'registration' stuff which oi think does not apply.
[10:16:34] <qchris>	 Ok. Sorry.
[10:24:34] <qchris>	 nuria: Looking through the various defs and sqls, I also think the the registration date should not be relevant for 'rolling active editor'.
[10:24:43] <nuria>	 ya, cut & paste
[10:24:46] <nuria>	 seems to me
[10:24:52] <nuria>	 ok
[10:24:53] <qchris>	 Full ACK.
[10:25:16] <qchris>	 The join on the logging table got stripped, but the registration part looks like an oversight.
[10:28:04] <grrrit-wm>	 (CR) Nuria: "There is some confusion on code & naming as to which one of the metrics is this one. there are two similar metrics:" [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/147312 (https://bugzilla.wikimedia.org/67458) (owner: Milimetric)
[10:40:17] <grrrit-wm>	 (CR) QChris: [C: -1] "Test is still failing for me." (4 comments) [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/146297 (https://bugzilla.wikimedia.org/67823) (owner: Milimetric)
[11:59:44] <grrrit-wm>	 (PS1) QChris: Remove outdated camus jar [analytics/refinery] - https://gerrit.wikimedia.org/r/147421
[12:00:46] <grrrit-wm>	 (PS2) QChris: Remove outdated camus jar [analytics/refinery] - https://gerrit.wikimedia.org/r/147421
[12:23:52] <grrrit-wm>	 (PS5) Nuria: Avoid pickle max recursion while serializing chain [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/146297 (https://bugzilla.wikimedia.org/67823) (owner: Milimetric)
[12:23:54] <grrrit-wm>	 (CR) jenkins-bot: [V: -1] Avoid pickle max recursion while serializing chain [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/146297 (https://bugzilla.wikimedia.org/67823) (owner: Milimetric)
[12:26:39] <grrrit-wm>	 (CR) Nuria: Avoid pickle max recursion while serializing chain (1 comment) [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/146297 (https://bugzilla.wikimedia.org/67823) (owner: Milimetric)
[12:32:19] <wikibugs>	 Analytics / Wikistats: Please update language families classification - https://bugzilla.wikimedia.org/33115#c9 (Andre Klapper) NEW>RESO/INV Unfortunately closing this report as no further information has been provided.  Ilya: Please feel free to reopen this report if you can provide the URL that...
[14:35:57] <grrrit-wm>	 (CR) Milimetric: "The code is consistently referring to it as "Rolling Active Editor", could you point out where this is not true? I used the SQL provided " [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/147312 (https://bugzilla.wikimedia.org/67458) (owner: Milimetric)
[15:01:39] <nuria>	 hello
[15:01:47] <nuria>	 hello milimetric
[15:02:04] <milimetric>	 hi!
[15:02:39] <nuria>	 regarding teh metric, I think reserach might have cut & paste the sql  for rolling new active editor
[15:02:48] <milimetric>	 nuria: I was wondering if you tested the metric anyway, despite the potential naming problem
[15:02:55] <milimetric>	 I don't think so
[15:03:09] <milimetric>	 because the "rolling new active editor" has the "newly registered" definition nested
[15:03:13] <nuria>	 no, cause i was not sure which one was it.
[15:03:25] <milimetric>	 and the "rolling active editor" doesn't care about the logging table, it's just looking at the user table
[15:03:33] <milimetric>	 ok, that's fair
[15:03:38] <milimetric>	 I'm waiting for halfak to show up
[15:03:48] <milimetric>	 meanwhile I get to play with vega :)
[15:03:59] * milimetric 's secret plan is to always play with vega
[15:05:03] <nuria>	 ok, I think just removing the "registration" from rolling active editor should be sufficient but research can verify.
[15:05:24] <milimetric>	 yeah, that'd be easy enough either way
[15:05:43] <milimetric>	 did you see what I meant about the timeseries implementation though?
[15:06:05] <nuria>	 no, i did not looked up at that part in detail
[15:06:25] <milimetric>	 k, it's just a comment in the commit message
[15:31:57] <nuria>	 milimetric, did you see my comments about the custom binding?
[15:32:06] <nuria>	 the field alredy has a binding
[15:32:13] <nuria>	 *already
[15:34:49] <milimetric>	 k, i'll take a look
[15:35:59] <milimetric>	 hm, nuria I don't think that's a problem; would you like me to give it a shot?
[16:02:54] <nuria>	 no milimetric, i will work on it a bit but having two bindings on the same field sounds very confusing
[16:03:34] <milimetric>	 that could be, I'm too far in ko world to have perspective :)
[16:03:43] <milimetric>	 but there are plenty of examples around the code
[16:04:18] <milimetric>	 multiple bindings don't conflict at all, they just control different things, like for example style and value:
[16:04:46] <milimetric>	 <input data-bind="css: {'someclass': boolProperty}, value: textProperty"/>
[16:05:26] <milimetric>	 in this case, the typeahead binding would control the same stuff that the collection of data-source, data-provide, etc. control
[16:06:09] <milimetric>	 nuria: if it's confusing though, it could be that my advice was misphrased, feel free to push non-working code and I can hlep
[16:06:12] <milimetric>	 *help
[16:51:48] <nuria>	 milimetric , ping me when you talk to halfak to take a look at CR gain, plis
[16:52:10] <milimetric>	 yeah, will do nuria
[18:23:05] <grrrit-wm>	 (PS10) Nuria: Add autcomplete to tags on tag insertion [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/145039 (owner: Terrrydactyl)
[18:23:07] <grrrit-wm>	 (CR) jenkins-bot: [V: -1] Add autcomplete to tags on tag insertion [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/145039 (owner: Terrrydactyl)
[18:36:38] <grrrit-wm>	 (CR) Qgil: "Can someone review this old changeset? Or should it be abandoned?" [analytics/webstatscollector] - https://gerrit.wikimedia.org/r/103241 (https://bugzilla.wikimedia.org/58316) (owner: QChris)
[19:48:08] <milimetric>	 NOTE: Jibo vs. Christian should be made into an epic ninja movie
[19:48:41] <lzia>	 totally milimetric :D
[19:51:47] <milimetric>	 halfak: hey you got a sec to talk metrics?
[19:54:07] <halfak>	 Yup.
[19:54:10] <halfak>	 milimetric, ^
[19:54:16] <milimetric>	 k, so rolling new active editor
[19:54:19] <milimetric>	 and rolling active editor
[19:54:24] <milimetric>	 nuria and I were a bit confused
[19:54:46] <milimetric>	 because the sql for rolling active editor includes a "this user registered in the past <u> days"
[19:55:06] <milimetric>	 so, is that how it's supposed to be?  and if so, there seems to be a very small difference between the two
[19:55:48] <halfak>	 Can you link me to the SQL you are looking at?
[19:55:50] <nuria>	 Seems that "Rolling Active Editors" should not need registration info
[19:55:59] <halfak>	 +1
[19:56:07] <halfak>	 Must be a mistake
[19:56:09] <milimetric>	 https://meta.wikimedia.org/wiki/Research:Rolling_active_editor
[19:56:45] <nuria>	 k
[19:58:50] <nuria>	 milimetric, did you looked ta the ko stuff?
[19:59:00] <nuria>	 *at the ko stuff
[19:59:10] <milimetric>	 not yet, wanted to take care of this metric first
[19:59:18] <milimetric>	 (halfak I pasted the URL above)
[19:59:22] <halfak>	 UI think I have it fixed.
[19:59:26] <halfak>	 -U
[19:59:28] <milimetric>	 oh ok
[19:59:43] <milimetric>	 so halfak, the other question was about the ar_rev_id / ar_id edit that I made and you reverted
[19:59:49] <halfak>	 Sorry for the trouble guys.
[20:00:09] <milimetric>	 no prob halfak, easy fix in sqlalch
[20:00:15] <halfak>	 Oh yes.  So, it's not totally necessary to even check for NOT NULL, but I like to do it since it helps me catch errors.
[20:00:29] <halfak>	 So, many revisions in the archive table will have NULL ar_rev_id.
[20:00:33] <milimetric>	 right
[20:00:37] <halfak>	 But they are still legit revisions.
[20:00:43] <milimetric>	 oh!
[20:00:48] <milimetric>	 i get what your comment means now
[20:00:55] <milimetric>	 gotcha, ok
[20:01:02] <halfak>	 :)
[20:03:34] <milimetric>	 so halfak, had you found problems with something like:
[20:03:34] <milimetric>	 select count(*) from user inner join archive
[20:03:34] <milimetric>	 instead of
[20:03:34] <milimetric>	 select sum(ar_id is not null) from user left join archive
[20:03:34] <milimetric>	 ?
[20:03:48] <milimetric>	 (pseudosql of course)
[20:09:49] <halfak>	 It's not really that there are problems so much as it was a defensive programming strategy.  I'd rather have zero counts and know I did something wrong than have a wrong, but positive count.
[20:10:02] <halfak>	 It's not something that needs to go to production.
[20:10:12] <halfak>	 I guess I did it without thinking.
[20:11:25] <halfak>	 milimetric, ^
[20:11:34] <milimetric>	 thanks I was pondering what you mean
[20:12:00] <milimetric>	 I think the unit tests on the translated query should take care of any problems, so I'll go with the less safe approach
[20:12:10] <milimetric>	 but interesting, I never thought about that
[20:12:35] <milimetric>	 so really, we can ignore the user table completely now right?
[20:12:42] <halfak>	 Yes.
[20:12:53] <milimetric>	 ok, I won't even join to it, this thing should scream :)
[20:12:59] <halfak>	 Yup.
[20:13:06] <halfak>	 Only need to read the user_timestamp index.
[20:13:18] <halfak>	 FYI, archive doesn't have a user_timestamp index on labs.
[20:13:32] <halfak>	 So, it'll be a bit slower.
[20:14:11] <milimetric>	 ar_timestamp?
[20:14:30] <milimetric>	 i noticed that, but was somewhat consoled that archive is a bit smaller of a table
[21:18:13] <grrrit-wm>	 (PS5) Milimetric: Add monthly active editor [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/147312 (https://bugzilla.wikimedia.org/67458)
[21:18:51] <grrrit-wm>	 (CR) Milimetric: "After talking with Aaron, nuria was right. He updated the SQL on-wiki and this latest patchset fixes the metric. Should be much faster n" [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/147312 (https://bugzilla.wikimedia.org/67458) (owner: Milimetric)
[21:19:04] <milimetric>	 looking at ko thin gnow
[21:19:30] <nuria>	 milimetric, it's fine, i did look at it and fixed it i think, will send patch
[21:19:48] <milimetric>	 oh ok, i'll look monday then?
[21:33:05] <grrrit-wm>	 (PS11) Nuria: Add autcomplete to tags on tag insertion [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/145039 (owner: Terrrydactyl)
[21:33:07] <grrrit-wm>	 (CR) jenkins-bot: [V: -1] Add autcomplete to tags on tag insertion [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/145039 (owner: Terrrydactyl)
[21:33:25] <kevinator>	 milimetric: I got dashiki running, but it’s not showing anything.  what am I supposed to type in the textfield?
[21:33:49] <milimetric>	 did you change /etc/hosts kevinator?
[21:33:56] <kevinator>	 yes
[21:33:56] <nuria>	 sure milimetric Monday is good.
[21:34:07] <milimetric>	 cool, I'll do that nuria - have a great weekend
[21:34:07] <kevinator>	 I loaded the web page fine
[21:34:12] <milimetric>	 kevinator: batcave?
[21:34:16] <kevinator>	 ok
[21:35:04] <nuria>	 I will work on autocompletes for dashboard next with knockout components