A xcube plugin that that allows generating data cubes from the CMEMS API.

In order to access the CMEMS API via the xcube-cmems plugin, you need to create a cmems account first, if you not do already have one.

This section describes three alternative methods you can use to install the xcube-cmems plugin.

conda can sometimes be inconveniently slow when resolving an environment. If this causes problems, consider using mamba as a much faster drop-in alternative to conda.

xcube-cmems and all necessary dependencies (including xcube itself) are available on conda-forge, and can be installed using the conda package manager. The conda package manager itself can be obtained in the miniconda distribution. Once conda is installed, xcube-cmems can be installed like this:

$ conda create --name xcube-cmems-environment --channel conda-forge xcube-cmems
$ conda activate xcube-cmems-environment

The name of the environment may be freely chosen.

This method assumes that you have an existing conda environment, and you want to install xcube-cmems into it.

xcube-cmems can also be installed into an existing conda environment. With the existing environment activated, execute this command:

$ conda install --channel conda-forge xcube-cmems

Once again, xcube and any other necessary dependencies will be installed automatically if they are not already installed.

If you want to install xcube-cmems directly from the git repository (for example if order to use an unreleased version or to modify the code), you can do so as follows:

$ conda create --name xcube-cmems-environment --channel conda-forge --only-deps xcube-cmems
$ conda activate xcube-cmems-environment
$ git clone https://github.com/dcs4cop/xcube-cmems.git
$ python -m pip install --no-deps --editable xcube-cmems/

This installs all the dependencies of xcube-cmems into a fresh conda environment, then installs xcube-cmems into this environment from the repository.

You can run the unit tests for xcube-cmems by executing

$ pytest

in the xcube-cmems repository. Note that, in order to successfully run the tests using the current repository version of xcube-cmems, you may also need to install the repository source version of xcube rather than its latest conda-forge release and have the cmems credentials set as environment variables CMEMS_USERNAME and CMEMS_PASSWORD.

To create a test coverage report, you can use

pytest --cov xcube_cmems --cov-report html

This will write a coverage report to htmlcov/index.html.

Jupyter notebooks demonstrating the use of the xcube-cmems plugin can be found in the examples/notebooks/ subdirectory of the repository.

To release xcube-cmems, please follow the steps outlined in the xcube Developer Guide.