[08:58:01] <dcausse>	 pfischer_: fyi, filed T322186, might be a good next step
[08:58:02] <stashbot>	 T322186: Consume revision based changes from the mediawiki.page-state stream - https://phabricator.wikimedia.org/T322186
[08:59:28] <dcausse>	 this stream is still in its infancy so it's a good time to provide feedback 
[11:02:38] <gehel>	 lunch
[12:14:46] <dcausse>	 lunch
[13:01:49] <gehel>	 dcausse: ping me when you have a few minutes...
[13:31:00] <dcausse>	 gehel: I'm around
[13:34:29] <gehel>	 dcausse: meet.google.com/bjq-uqix-uds
[14:32:10] <gehel>	 inflatador: around for our 1:1?
[14:41:52] <inflatador>	 gehel day off today, sorry for not declining
[14:42:14] <gehel>	 inflatador: yes, I see that now. Sorry for the ping. Please go away from IRC!
[17:35:37] <dcausse>	 dinner
[19:26:37] <ebernhardson>	 well, i guess this works, mostly. 467.38M pages are within 20% of the stored incoming_links value, in a dataset of 494.91M pages.  Will poke over the ones that are out just to make sure nothing crazy is happening
[19:26:50] <ebernhardson>	 (to be fair, more than half of the dataset has 1 incoming link :P)
[20:55:02] <ebernhardson>	 how fun...tiny bit of rain (potentially unrelated) and now i'm regularly dropping packets :S
[22:02:25] <ebernhardson>	 curious..but whatever.  The largest outliers for incoming_links in the the _source docs are all basically wrong.  Issuing fresh queries against elastic agrees with my batch computation, so for example https://en.wikipedia.org/wiki/Emergency_management has 17k in the _source doc, but batch and a fresh query report 1.4k. No clue how it's not getting updated, it was edited a couple days ago
[22:02:27] <ebernhardson>	 (after the dump was taken)
[22:03:18] <ebernhardson>	 still working out what thresholding would be appropriate, popularity_score has 188M pages in the latest dump, 99M of those overlap with the incoming_links computation