Skip to content

Installation

Hint

If it is your first time using Python, we recommend conda, mamba, pip or uv as easy-to-use package managers. They are available for Windows, macOS, and GNU/Linux. It is always helpful to use dedicated environments.

You can install PyPSA via all common package managers:

pip install pypsa

conda install -c conda-forge pypsa

uv add pypsa

PyPSA is written and tested to be compatible with Python 3.11 and above. We recommend to use the latest version with active support (see endoflife.date).

Solvers

PyPSA passes optimisation problems (see overview) to an external solver and is deeply integrated with the optimisation framework linopy to do so. Some examples of available solvers:

Free & open source Commercial & proprietary
HiGHS Gurobi
Cbc CPLEX
GLPK FICO Xpress
SCIP MOSEK
COPT

PyPSA ships with the open-source solver HiGHS by default. For installation instructions of further solvers for your operating system, follow the links above.

Note

Commercial solvers currently significantly outperform open-source solvers for large-scale problems. It might be the case that you can only retrieve solutions by using a commercial solver. Many commercial solvers provide free academic licenses.

Upgrading

We recommend always keeping your PyPSA installation up-to-date, since bugs get fixed and new features are added. PyPSA is also only tested with the latest stable versions of all the dependent packages for the respective Python versions.

To upgrade PyPSA, run:

pip install --upgrade pypsa
# Or upgrade to a specific version:
pip install pypsa==0.35.2

conda update pypsa
# Or upgrade to a specific version:
conda install -c conda-forge pypsa==0.35.2

uv add --upgrade pypsa
# Or upgrade to a specific version:
uv add pypsa==0.35.2

Check the release notes for API changes that may require you to update your code. PyPSA releases new versions according to the semantic versioning scheme. Any breaking changes are always announced via deprecation warnings in the code and in the release notes, including a version when they are going to be removed (always the next major version, e.g. v2.0.0). That way you can be sure that your code will continue to work at least until the next major version. But this does not include bug fixes, which you only get when upgrading to the latest version. If you are upgrading from a pre , we recommend you upgrade in small steps and fix any deprecation warnings before upgrading to v1.0.0.

Dependencies

PyPSA relies heavily on other open-source Python packages. Some of them are:

  • pandas for storing data about components and time series
  • numpy and scipy for calculations, such as linear algebra and sparse matrix calculations
  • linopy for preparing optimisation problems (LP, QP, MILP)
  • matplotlib, seaborn and plotly for static and interactive plotting
  • networkx for some network calculations
  • pytest for unit testing

Find the full list of dependencies in the pyproject.toml file.

Optional dependencies

Besides the mandatory dependencies that are installed by default, PyPSA has a number of optional dependencies that are not installed by default and are only needed for certain features. You can install them by executing the command:

pip install "pypsa[<feature>]"

conda install -c conda-forge "pypsa[<feature>]"

uv add "pypsa[<feature>]"

where <feature> can be one of the following:

IO

  • pypsa[hdf5]: for reading and writing HDF5 files
  • pypsa[excel]: for reading and writing Excel files
  • pypsa[cloudpath]: for reading and writing files from cloud storage

Plots

  • pypsa[cartopy]: for plotting geographical maps

Configuration

  • pypsa[dotenv]: for loading PyPSA options from .env files

Solvers

  • pypsa[gurobipy]: for installing the Gurobi Python API

Development

  • pypsa[dev]: for installing all development dependencies, including pytest and pre-commit
  • pypsa[docs]: for installing all dependencies needed to build the documentation