profiling big project
Andrea Crotti
andrea.crotti.0 at gmail.com
Tue Oct 11 16:15:15 EDT 2011
Hi everyone,
I am in the situation that I'll have to profile huge python programs,
huge because they use many external libraries like numpy / pyqt / ETS
etc etc...
The applications are very slow and we wanted to understand what is
actually going on.
I like to use normally pycallgraph, but in this case is quite useless,
because the graph would be huge, so I had the following thought, if I'm
able to serialize the generate a text representation of the graph I
should be then able to actually understand something.
So I forked pycallgraph (https://github.com/AndreaCrotti/pycallgraph)
and my idea is to create an org-mode or RST file instead of a graph,
adding also links to the source code maybe.
For example (in org mode)
* function1 50%
** function2 25%
** function3 22%
...
and so on, in RST maybe I could even use autodoc to get also the doc
from the functions.
Any idea / suggestions / comment? Maybe there is something like this
already?
More information about the Python-list
mailing list