[15:17:49] <guillom>	 hello hello
[15:19:44] <Nettrom>	 morning!
[15:23:15] <guillom>	 hello Nettrom
[17:00:22] <halfak>	 Hey hey science people :) 
[17:32:12] <harej>	 hello aaron halfaker
[17:32:19] <harej>	 have you considered porting your utilities to pywikibot?
[17:34:17] <halfak>	 harej, sure.  When I first started working with mediawiki's API, I looked at pywikibot.  I didn't like it's constraints, so I designed a different interaction strategy. 
[17:34:52] <halfak>	 I'm certainly not alone in disliking the way that pywikibot structures your use of the API. 
[17:35:08] <halfak>	 But I can understand your desire for consistency. 
[17:35:12] <harej>	 Well, we'll reconcile this all out during the hack-a-thon.
[17:35:16] <halfak>	 Why not just use both?
[17:35:21] <harej>	 I will use both.
[17:35:23] <halfak>	 If you'd like to. 
[17:35:25] <harej>	 I also used a third API wrapper.
[17:39:18] <halfak>	 ha. 
[17:40:02] <harej>	 lego's super simple mediawiki for straightforward get queries; your API for research; and pywikibot because it can save pages
[18:16:53] <harej>	 Ironholds: I want you to port pywikibot to R.
[18:16:54] <harej>	 pyrbot.
[18:16:57] <harej>	 rwikibot?
[19:57:36] <milimetric>	 Ironholds: HaeB and I are talking pageviews.  What definition did you load up in Pentaho for him last time?
[19:58:12] <milimetric>	 feel free to point me to docs
[19:59:28] <HaeB>	 context: https://phabricator.wikimedia.org/T88844
[20:04:35] <HaeB>	 ...i don't think anything was loaded specifically for this; i used what was already in Pentaho, with both spiders and automata excluded 
[20:06:51] <milimetric>	 right, but Pentaho is loaded manually in a process that I was not involved in.  So to get to the same data source, we need to know the specifics.
[20:09:28] <Ironholds>	 milimetric, the definition as it stood two weeks ago
[20:09:46] <milimetric>	 Ironholds: so the new pageview definition run on sample logs?
[20:09:58] <Ironholds>	 yes
[20:10:03] <Ironholds>	 I'd never use the old definition for anything
[20:10:13] <Ironholds>	 up to and including printing it off and using it as cheap insulation
[20:10:13] <milimetric>	 :) neither would any sensible person.
[20:17:18] <milimetric>	 Not even verbose enough for insulation, huh?  Maybe if it used denser letters like B instead of all those weak i letters.  New code metric: R-value
[20:28:34] <Ironholds>	 milimetric, heh
[21:12:49] <milimetric>	 HaeB: it looks to me like Pentaho has data partway through April
[21:12:56] <milimetric>	 so that's at least through March which is what you need
[21:13:22] <milimetric>	 Ironholds: just checking when it says "2015-04-01" that really just means "2015-04", right?
[21:13:28] <Ironholds>	 yup
[21:13:31] <HaeB>	 that should work fine, thanks
[21:13:39] <Ironholds>	 but POSIX dates aren't storable without days
[21:13:41] <milimetric>	 k, then we're good HaeB.  Sorry if Kevin confused you :)
[21:13:54] <milimetric>	 Ironholds: I am sympathetic to the problem, just making sure :)
[21:14:02] <Ironholds>	 heh
[21:16:18] <HaeB>	 is "new-pv-monthly.saiku" the correct dataset to use?
[21:18:03] <HaeB>	 ...it gives me "Error Loading Query", trying "pv-demo.saiku" instead
[21:18:45] <HaeB>	 ...but "pv-demo.saiku" only has data until aug 2014
[21:20:47] <Ironholds>	 HaeB, hangon, lemme check
[21:21:27] <Ironholds>	 HaeB, "create new" "saiku analytics" and then select pageviews 0.5 from the dropdown
[21:25:13] <HaeB>	 Ironholds: got it, thanks!
[21:27:55] <Ironholds>	 HaeB, awesome!
[22:06:39] <HaeB>	 Ironholds: hmm, i'm getting different (i.e. somewhat higher) numbers now, compared to the query i did in february...
[22:06:54] <HaeB>	 ...here is the MDX query that pentaho generated, does that look OK to you?
[22:07:00] <HaeB>	 WITH SET [~ROWS] AS     {[Year and Month].[Year and Month].Members} SELECT NON EMPTY {[Measures].[Pageviews]} ON COLUMNS, NON EMPTY [~ROWS] ON ROWS FROM [Monthly Pageviews Cube v0.5 (2013-March 2015)] WHERE CrossJoin(Except({[is_automata].[is_automata].Members}, {[is_automata].[true]}), Except({[is_spider].[is_spider].Members}, {[is_spider].[true]}))
[22:08:13] <Ironholds>	 I don't know, I've never dealt with MDX before
[22:08:37] <Ironholds>	 but I'd expect higher numbers because the dataset was updated in February/March-ish and one of the updates was to include some previously excluded requests
[22:08:46] <Ironholds>	 (e.g. Wikidata requests)
[22:09:56] <HaeB>	 me neither .. (it's what pentaho generates in the background when one clicks together a query, i obtained this with the "Show MDX" button on top)
[22:10:03] <HaeB>	 ah ok
[22:10:47] <HaeB>	 e.g. for december 2014: 16324187000 (as obtained in february) vs. 17,628,852,000 (now)
[22:11:05] <HaeB>	 does that look plausible?
[22:11:51] <HaeB>	 if so, i'll just go ahead with this and include a little note in the scorecard that the scope was updated compared to the last report
[22:14:24] <Ironholds>	 it does, yep
[22:15:03] <HaeB>	 ok thanks
[23:30:57] <yuvipanda>	 halfak: hey! have a few mins to chat at some point? I’ve to leave at 5:30 tho :(
[23:31:18] <halfak>	 Yes.  Now is good.  Want to swing by?
[23:31:24] <yuvipanda>	 halfak: yes
[23:31:25] <halfak>	 COuld give the bike a ride :)