Notebook Execution

There are three options to work with this course: via Binder, a dedicated Cloud Hub, or locally on your machine.

Binder¶

The whole course runs in your browser on the Project Pythia Binder — no installation needed. Click the badge to launch JupyterLab with the full course environment (Py-ART, wradlib, xradar, BALTRAD, LROSE):

The first launch can take a few minutes while the course image is fetched. Note that Binder sessions are temporary: download anything you want to keep before closing the tab.

Cloud Hub¶

For this course we will use a jupyterlab instance on a dedicated Cloud Hub. The needed instructions will be added on short notice.

Locally on your own premises¶

The easiest is to make use of our pre-built docker image which includes all the needed bits and pieces and is actually also used as base image for the Project Pythia Hub instance.

The only thing you need is a running docker service on your local machine.

Then it’s just a one-liner:

$ docker run -ti --name erad2026 -p 8888:8888 -e LOCAL_USER_ID=$UID ghcr.io/openradar/erad2026:latest /srv/conda/envs/notebook/bin/jupyter lab --ip='*' --port=8888

Some of the notebooks are utilizing quite some large cloud data files. So a decent network connection (LAN connection) is preferred to make use of those notebooks. Don’t try this at the course as bandwidth will be limited.

If you need a connection to the outside world, you can mount local folders and environment variables into the running container. Please check out the docker documentation for more details

$ docker run -ti --name erad2026 -p 8888:8888 -v /host/path/to/your/folder:/home/your/folder -e LOCAL_USER_ID=$UID -e YOUR_ENV_VAR=your_env_var ghcr.io/openradar/erad2026:latest /srv/conda/envs/notebook/bin/jupyter lab --ip='*' --port=8888

Known issues¶

Slideshows not working¶

Slideshows do not work when @pyviz/jupyterlab_pyviz is enabled. A temporary fix is to disable it in the running jupyterlab instance (eg in a terminal):

$ jupyter labextension disable @pyviz/jupyterlab_pyviz

Then, the rise reveal.js slideshow can be run. Afterwards, you might enable it again:

$ jupyter labextension enable @pyviz/jupyterlab_pyviz