Installing SWEpy¶

SWEpy is published and maintained on both pip and conda-forge. It is reccomended that you stick with conda-forge as dependency issues can be a problem in pip.

You can create a conda environemnt using the provided yml file, swepy_env.yml to get up and running right away.

In your terminal, navigate to the directory containing your yml file and run the following command:

conda env create -f swepy_env.yml

conda activate swepy_env

or

source activate swepy_env

Installing ipykernel for Jupyter¶

In order for jupyter to find your kernel, you need to install ipykernel. First activate your new environment, then enter the following:

python -m ipykernel install --user --name <env name> --display-name "<display name>"

Make sure that you do not include the brackets “<>” in your environment name and display name!

Importing SWEpy in Python¶

Now in Python you will be able to import SWEpy and its submodules.

>>> import swepy
>>> import swepy.pipeline as pipeline
>>> import swepy.process as process
>>> import swepy.analysis as analysis
>>> import swepy.nsidcDownloader as nD

Installation Troubleshooting¶

Many of the issues that can arise when importing SWEpy for the first time are related to dependency conflicts in your conda environment. GDAL in particular tends to produce dependency issues. Since open source python libraries are always evolving, it is common for the “links” between the to be broken after updates occur.

To solve this, activate your conda environment:

conda activate swepy_env

Then update all of the libraries in your environment:

conda update --all

Generally this will solve most dependency issues, but if it does not, try checking if any critical dependencies have been updated recently. If something was recently updated, rolling it back to the second latest release may fix issues in your environment.