Python and associated packages for science are freely available. Convenient science python package setups are available from, Anaconda, and other locations (some platform specific). Anaconda also includes a developer environment that works well with pysat. Core science packages such as numpy, scipy, matplotlib, pandas and many others may also be installed directly via the python package installer “pip” or your favorite package manager.

For maximum safety, you can install pysat into its own virtual environment. This ensures there are no conflicts with any other installed Python distributions.

To use Anaconda’s tools for creating a suitable virtual environment,

conda create -n virt_env_name python=3
conda activate virt_env_name
conda install numpy -c conda

Standard installation

pysat itself may be installed from a terminal command line via:

pip install pysat

There are a few packages that pysat depends on that will be installed as needed by the installer:

  1. dask

  2. netCDF4

  3. numpy

  4. pandas

  5. portalocker

  6. pytest

  7. scipy

  8. toolz

  9. xarray

Development Installation

pysat may also be installed directly from the source repository on github:

git clone
cd pysat
python -m build .
pip install --user .

An advantage to installing through github is access to the development branches. The latest bugfixes can be found in the develop branch. However, this branch is not stable (as the name implies). We recommend using this branch in a virtual environment and using:

git clone
cd pysat
git checkout develop
python -m build .
pip install -e .

The use of -e in the setup command installs the code ‘in-place’, so any changes to the software do not have to be reinstalled to take effect. It is not related to changing the pysat working branch from main to develop in the preceeding line.