Jupyter Lab

This is an implementation of JupyterLab that runs in a container using software called Apptainer. Since the app runs in a container, Slurm commands are not available from within the Jupyter Lab interface, although they will still be available through the default Open OnDemand terminal app.

Running the app

When you start the app, you will see the standard options to set the duration of your session and the number of CPUs that you will need, as described on the interactive apps page.

In addition, you will see a dropdown menu allowing you to select a container. There are two different default options (scipy-notebook and datascience-notebook), as well as other course-specific images that may have been prepared for particular use cases.

We use two of the Jupyter Docker stacks created by Project Jupyter. Specifically, we have the scipy-notebook and datascience-notebook containers configured for use.

The scipy-notebook container is a general purpose, Python-only Jupyter Lab app with commonly used Python libraries such as numpy, pandas, matplotlib, and of course scipy. For a full overview of what's included, see the documentation for the docker container linked above.

The datascience-notebook container has everything in the scipy-notebook image with the addition of kernels for the R and Julia programming languages. For a full overview of what's included, see the documentation for the docker container linked above.

In addition, we can configure custom software configurations, and this app is one of the ways that we can do that. If you see an image with your course's name on it, it's likely for you. Ask your instructor if you aren't sure which container to use.