I guess the xmlrpc services are issuing stats from the apache log parsing.
I would like to propose a funny improvement of pypi with a time serie of dled stats à la http://pypi.python.org/pypi?:action=display&name=pypi-stat&version=1.2.1
What I would need is :
- some significant apache logs (to find an efficient way to store data)
- the number of packages hosted on pipy
- the functionnal domain in which I can experiment ;
I would give back :
- an estimation of data usage per package per year (in Ko) ;
- a prototype (parsing, storing, interacting, ploting) ;
and some metrics on CPU use.
Log in to post a comment.