Unidata is building a demonstration JupyterLab server using Unidata resources in the NSF Jetstream cloud. The server is currently pre-loaded with Jupyter notebooks created as part of Unidata's Notebook Gallery, Online Python Training, and Python Training Workshops. NOTE: The JupyterLab server is in the early stages of its implementation; additional features and documentation will be available as the project progresses.

Take me to the JupyterLab server now

Logging in to the JupyterLab Server

To use the JupyterLab server, you will need to authenticate using your GitHub credentials. If you do not already have a GitHub account, you can sign up for a free account in a matter of minutes.

Note: To browse directly to the Unidata Science Gateway JupyterLab server, bookmark this address:

https://jupyterhub.unidata.ucar.edu

Using the JupyterLab Server

Once you have logged in to the JupyterLab server using your GitHub credentials, you will find numerous geoscientific notebooks. Use these to experiment or as templates for your own notebooks. Unidata developers have pre-configured all of the software dependencies for you, so you're ready to go. (BUT: see the note on kernels, below.) You can also create your own notebooks, folders, projects, and conda environments.

Kernels

If you try to execute one of the notebooks and get errors, it's possible the notebook is running under the wrong kernel. In JupyterLab, the kernel takes care of loading the correct code interpreter (a version of python, in the case of the Unidata notebooks) as well as all software modules required for the notebook to run. In general, you'll want to run a notebook in the kernel that corresponds to the top-level folder in shown on the left side of the JupyterLab screen. For example, if you're running one of the notebooks under the notebook-gallery folder, choose the kernel marked Python [conda env:notebook-gallery].

To change the kernel, select Kernel → Change Kernel... from the menu on the JupyterLab screen. The kernel that is currently in use is displayed in the upper right corner of the the JupyterLab screen; you can also click on that display to select a different kernel.

The Terminal

You can also interact with your JupyterLab server via a full-fledged Linux Terminal (Select File → New → Terminal from the menus). You may wish to use the terminal to:

use the git version control capability
create new conda environments.
(Be sure to load the nb_conda_kernels package in order to have these environments available from your Jupyter server via the Kernel menu.)

Issues