Tells TLJH to use DockerSpawner to spin up jupyter/datascience-notebook containers for each user. This makes it easy to include R and Julia Kernels for your JupyterHub users. Read more about what's included in the jupyter/datascience-notebook here.
The plugin also sets jupyterlab as the default interface.
Include --plugin tljh-datascience in your TLJH install script. For example, here user kschuler with password password installs TLJH with tljh-datascience:
#!/bin/bash
curl https://raw.githubusercontent.com/jupyterhub/the-littlest-jupyterhub/master/bootstrap/bootstrap.py \
| sudo python3 - \
--admin kschuler:password --plugin git+https://github.com/pennchildlanglab/tljh-datascience
To add other or different images for your users to select, edit the dockerspawner config file by SSH-ing into your server and running
sudo nano /opt/tljh/config/jupyterhub_config.d/dockerspawner_tljh_config.py
The current list of available images is in c.DockerSpawner.image_whitelist = ['jupyter/datascience-notebook:r-4.0.3', 'jupyter/datascience-notebook:r-3.6.6']. You can edit this list to include any docker images you want to make available to your users. Then reload the hub.
sudo tljh-config reload
The plugin currently pulls only one docker image, so other images will take a while to spin up the first time. If you want to pre-pull the images, you can also run the following, where <tag> is the specific tag for the image.
sudo docker pull jupyter/datascience-notebook:<tag>"
This plugin simply uses docker spawner to start user servers in the docker containers you make available in a list. For more complex use-cases, check out tljh-repo2docker plugin.
This plugin was inspired by this Ideonate post and the Rxns stack plugin
- figure out how to include jupyterlab plugins (probably just a docker image based on the datascience-notebook is the easiest)
- we could prob do this without using subprocesses -- maybe require dockerspawner in the setup.py and then import it; and then just install docker.io via additional apt packages.