pandas-msgpack is an interface to msgpack for pandas. It allows for convenient IO of common pandas objects to the msgpack
format,
including performant compression like blosc
.
This package is a fork of the original pandas-msgpack
, which has not been maintained since
its inception in 2017 by the PyData team. However, it is to date an extremely performant dataformat.
We are not the original authors of this package, and we do not claim any ownership of the original code. We are simply maintaining and updating it for use in our main project ISF (https://github.com/mpinb/in_silico_framework). We have no interest in adding new features, or making this broadly applicable. PRs are however very welcome.
Install latest release version via pip
pip install isf-pandas-msgpack
API tests are run on a single OS (ubuntu-latest
) for all pandas
versions (1.1
through 2.2
at the time of writing).
Python versions are not explicitly tested for, and we let the environment solver determine which Python version is appropriate for the pandas
version (these turn out to be Python 3.9
through 3.12
)
Building the package is tested for Python versions >=3.8
on three major platforms: Ubuntu, macOS and Windows (all 64 bit).
Publication to PyPI is triggered upon creating a github release.
For macOS and Linux, only the source distribution is published. setup.py
and pyproject.toml
have been properly configured with the setuptools
build backend, so this should not give too much issue.
For Windows, we provide wheels for Python >=3.8
.