I've been in an optimization phase for obnam lately. This means running it under the Python profiler, and examining the results.

It is easy enough to generate a text file with the results, but it might be even better if one could have an interactive tool to examine the profiling results.

It might, for example, be more efficient to sort the report in various ways interactively, or to drill down into the results to see where time is actually spent.

Anyone want to write this for me? :)

On the perl side, I'm generally impressed with Devel::NYTProf. Generates static html files (with javascript for sorting) that allow drilling down from global hotspots t per-line profiling statistics.
Comment by joey [kitenet.net] Mon Jun 28 16:06:21 2010
There is http://pypi.python.org/pypi/pyprof2calltree which makes it possible to view profiling data in kcachegrind and there is http://www.vrplumber.com/programming/runsnakerun/.
Comment by Trundle Mon Jun 28 17:55:18 2010