Installation

Prerequisites

Python >= 3.8
packaging >= 17.0
pybind11 >= 3.0.0
numpy >= 1.12.0
Sphinx >= 2.0.0 (optional, to build documentation)
sphinx-rtd-theme (optional, to build documentation)
myst-parser (optional, to build documentation)

Installation from sources

Installing pycommute from sources is as easy as running

LIBCOMMUTE_INCLUDEDIR="<path_to_libcommute>/include" pip install .

in the unpacked source directory. <path_to_libcommute> must be either installation or source directory of libcommute 1.0.0 or newer.

If you need to build documentation locally, you should additionally run

sphinx-build docs/ <path_to_sphinx_output_dir>

Automated installation of the PyPI package

pip install pycommute

Automated installation of the conda package (Anaconda.org)

Installation via ‘conda’ or ‘mamba’ package manager:

$ conda install krivenko::pycommute
$ mamba install krivenko::pycommute

The packages have been provided through the ‘krivenko’ Anaconda.org channel since version 1.0.0.

Docker images

Docker images of tagged releases of pycommute are available from Docker Hub.

docker run --rm -p 8888:8888 ikrivenko/pycommute:latest

This command will pull the most recent image and launch a Jupyter notebook server accessible at http://127.0.0.1:8888/. The server is run in a directory with a few interactive example notebooks.