On Thu, Feb 27, 2014 at 02:49:28PM -0800, Gergo Tisza wrote:
As I understand, Mark has some scripts running from a
cronjob to aggregate
data from the event logs (which cannot be public for privacy reasons) into
CSV files, and those CSV files are fed to Limn. Building those CSVs for
every wiki and making them downloadable should be simple, I imagine?
My favorite answer: "Yes".
Once I have a first-pass script that does this for one wiki, it should be
"trivial" to replace the places that call a specific wiki with variables
that get replaced and saved into multiple scripts, then run all the scripts.
But, this is a wishlist thing that I'm not super interested in worrying
about right now - I want to do the metrics we need, not the ones we might
later.
But, I'll leave it open to do later. :)
You can see the scripts I'm running at Gitorious [0]. There are some leftovers
from my first shitty implementation and the cronjobs aren't there, but
you can get a good picture.
[0]
https://gitorious.org/analytics/multimedia
--
Mark Holmquist
Software Engineer, Multimedia
Wikimedia Foundation
mtraceur(a)member.fsf.org
https://wikimediafoundation.org/wiki/User:MHolmquist