H4dr1en / jupyterflame

Bringing flamegraphs into jupyter notebooks for performance diagnosis

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Jupyterflame

Flamegraphs are a visualization of profiled software, allowing the most frequent code-paths to be identified quickly and accurately.

Jupyterflame will let you create flamegraphs inside jupyter notebooks. It will performs a %prun command in your cell and use the output to create a flamegraph.

Installation


The latest version of this package can currently be installed from this repo with the following command:

pip install git+https://github.com/H4dr1en/jupyterflame.git

Dependencies:

  • Perl, in order to support flamegraph.pl. Perl can be installed in a conda environnment with:
conda install perl
  • Windows users: as long as this PR is not merged, you will have to install this package for compatibility:
pip install git+https://github.com/H4dr1en/flameplot.git

Usage


To generate flamegraphs inside jupyter notebooks, simple enable it using:

%load_ext jupyterflame

This will register the %flame and %%flame ipython magic commands that you can now use to create flamegraphs in any cell of your notebook.

API


%flame and %%flame ipython magic commands are wrappers around %prun. Therefore all parameters of %prun are compatible with %flame and %%flame.

Flamegraph.pl parameters are also available in order to customize the flamegraph.

Example: demo

Current limitations


  • Limitations relative to the %prun command
  • The generated flamegraphs use javascript and html unique ids to be interactive, therefore only one flamegraph can be created at a time.

Contributions


Contributions are welcomed!

About

Bringing flamegraphs into jupyter notebooks for performance diagnosis

License:MIT License


Languages

Language:Perl 83.9%Language:Python 16.1%