General QuickStart

The HyTEST workflows are built largely on the Pangeo software stack. For general advice, you should start with Pangeo’s Deployment Setup Guides and select an option which resembles your compute environment.

The usual sticking points are to do with

• Getting the virtual environment correct

• Launching a Jupyter server

Virtual Environment

We recommend conda (or its functional equivalent, mamba) as the environment configuration tool. The software stack we use is defined in this environment file.

Use conda to create an environment called hytest with this command:

conda env create -f ./HyTEST.yml

You will then need to select the hytest environment as your kernel when you try to run the notebook.

While is is possible to run most of these workflows on a desktop system, they have been developed to run on a Linux variant. This will affect which versions and which architectures of the various software packages are installed.

Jupyter Server

If you are working in an environment that already has a pathway to run Jupyter notebooks, you will not need to worry about this section. You will simply launch the Jupyter notebook in whatever way you know how.

If you do need to set up your own Jupyter server, this is a harder problem to solve – primarily because there are so many variables and options to consider when setting it up. The Pangeo docs cover a few of the options, and the Jupyter docs also give some help for setting up. These deployments, generally, are meant for a single user to run notebooks on a semi-private host.

If multiple users will be using the same server, consider deploying a Jupyter Hub on a centralized computing host.