pwaller/pyprof2calltree

Overview

Script to help visualize profiling data collected with the cProfile Python module with the kcachegrind (screenshots) graphical calltree analyser.

This is a rebranding of the venerable http://www.gnome.org/~johan/lsprofcalltree.py script by David Allouche et Al. It aims at making it easier to distribute (e.g. through PyPI) and behave more like the scripts of the debian kcachegrind-converters package. The final goal is to make it part of the official upstream kdesdk package.

Installation

On Debian ≥ 11, or derivatives such as Ubuntu ≥ 20.04, sudo apt install kcachegrind pyprof2calltree.

Command line usage

Upon installation you should have a pyprof2calltree script in your path:

$ pyprof2calltree --help
usage: pyprof2calltree [-h] [-o output_file_path] [-i input_file_path] [-k]
                       [-r scriptfile [args ...]]

optional arguments:
  -h, --help            show this help message and exit
  -o output_file_path, --outfile output_file_path
                        Save calltree stats to <outfile>
  -i input_file_path, --infile input_file_path
                        Read Python stats from <infile>
  -k, --kcachegrind     Run the kcachegrind tool on the converted data
  -r scriptfile [args ...], --run-script scriptfile [args ...]
                        Name of the Python script to run to collect profiling
                        data
  -s {s,ms,us,ns}, --scale {s,ms,us,ns}
                        Time scale

Python shell usage

pyprof2calltree is also best used from an interactive Python shell such as the default shell. For instance let us profile XML parsing:

>>> from xml.etree import ElementTree
>>> from cProfile import Profile
>>> xml_content = '<a>\n' + '\t<b/><c><d>text</d></c>\n' * 100 + '</a>'
>>> profiler = Profile()
>>> profiler.runctx(
...     "ElementTree.fromstring(xml_content)",
...     locals(), globals())

>>> from pyprof2calltree import convert, visualize
>>> visualize(profiler.getstats())                            # run kcachegrind
>>> convert(profiler.getstats(), 'profiling_results.kgrind')  # save for later

or with the ipython:

In [1]: %doctest_mode
Exception reporting mode: Plain
Doctest mode is: ON

>>> from xml.etree import ElementTree
>>> xml_content = '<a>\n' + '\t<b/><c><d>text</d></c>\n' * 100 + '</a>'
>>> %prun -D out.stats ElementTree.fromstring(xml_content)

*** Profile stats marshalled to file 'out.stats'

>>> from pyprof2calltree import convert, visualize
>>> visualize('out.stats')
>>> convert('out.stats', 'out.kgrind')

>>> results = %prun -r ElementTree.fromstring(xml_content)
>>> visualize(results)

Change log

• 1.4.4 - 2018-10-19: Numerous small improvements, drop support for EOL python versions
• 1.4.3 - 2017-07-28: Windows support (fixed is_installed check - #21)
• 1.4.2 - 2017-07-19: No feature or bug fixes, just license clarification (#20)
• 1.4.1 - 2017-05-20: No feature or bug fixes, just test distribution (#17)
• 1.4.0 - 2016-09-03: Support multiple functions with the same name, tick unit from millis to nanos, tests added (#15)
• 1.3.2 - 2014-07-05: Bugfix: correct source file paths (#12)
• 1.3.1 - 2013-11-27: Bugfix for broken output writing on Python 3 (#8)
• 1.3.0 - 2013-11-19: qcachegrind support
• 1.2.0 - 2013-11-09: Python 3 support
• 1.1.1 - 2013-09-25: Miscellaneous bugfixes
• 1.1.0 - 2008-12-21: integrate fix in conversion by David Glick
• 1.0.3 - 2008-10-16: fix typos in 1.0 release
• 1.0 - 2008-10-16: initial release under the pyprof2calltree name