If you do not have docker, you can setup your local development environment
with conda
. However, the Python version will be 3.9 and versions of some
Python packages will differ from those defined for Docker
and poetry
installations. This can cause issues where code that runs on Docker
does not
run in the conda
environment or produces different results. Any such issues
should be posted on GitHub.
A recent version of conda
must be installed. See:
The environment.yml
dependency specification is tested in eis_toolkit
with
the libmamba
solver instead of the default. If you encounter installation
issues following this guide further, especially on Windows, you can enable the
libmamba
solver globally(!) as follows:
conda install -n base conda-libmamba-solver
conda config --set solver libmamba
See https://conda.github.io/conda-libmamba-solver/getting-started/ for further info.
Run all commands in the repository root unless instructed otherwise
- Install dependencies and create a new
conda
environment using the providedenvironment.yml
file. The environment name is defined inenvironment.yml
(eis_toolkit).
conda env create -f environment.yml
# You can overwrite an existing environment named eis_toolkit with the --force flag
conda env create -f environment.yml --force
- Activate the environment.
conda activate eis_toolkit
-
With the environment active, the package and all its dependencies should be available for execution.
-
E.g. run
pytest
to verify that the test suite works with your installation. If not, you should first verify your installation and secondly make a GitHub issue if you cannot figure out the problem.
You can add your own packages to the environment as needed. E.g.
jupyterlab
:
# -c conda-forge specifies the conda-forge channel, which is recommended
conda install -n eis_toolkit -c conda-forge jupyterlab