Improving the logged in home page experience in JupyterHub with `jupyterhub-fancy-profiles`
On most research oriented JupyterHub installations, users would like to customize their server (the environment, resources available, etc) after logging in. In Kubernetes based JupyterHub environments, a profile list provides this functionality.
The profile list is the de-facto “logged in homepage” for these users, as that is what they see after they have logged in.
In collaboration with
Development Seed, funded by our
earlier grant from
GESIS as well as the
NASA VEDA project, we have been building the
jupyterhub-fancy-profiles project to improve this experience.
jupyterhub-fancy-profiles)
Last week, we rolled this new experience out to all 2i2c managed JupyterHubs! Here’s a quick rundown of what this enables:
Descriptions for choices in the dropdowns, making it much easier for users to know what they are getting with each environment (or resource selection).
Fully backwards compatible with the existing KubeSpawner profile list implementation. In our PR to roll this out to all hubs, you notice that we didn’t have to change the structure of any profile lists! So you can safely roll this out to your hubs too without needing to fundamentally change how your profiles are set up.
It is a modern web app (built with react), just like the JupyterHub admin panel. This allows us to evolve and satisfy user needs much faster, as well as expanding the pool of people who can contribute to the project!
Support for dynamically building images using mybinder.org style repositories! It talks to the binderhub API so users can build reproducible environments as they wish without admin involvement nor needing to fully understand how docker and containers work. Our earlier blog post has more information.
This is just the start, and thanks to ongoing funding from the NASA VEDA project, we are going to continue making improvements to this experience.
Use this in your JupyterHub #
As with everything we build at 2i2c (per our right to replicate policy), this project can be used with any JupyterHub installation that uses Kubernetes. There are instructions in the README. Please try it out on yours and let us know what you think!
Credit #
- The project was initiated with funding generously provided by GESIS in cooperation with NFDI4DS (project number: 460234259) and CESSDA (see our earlier blog post).
- Sanjay Bhangar and Oliver Roick from Development Seed for advocating for this project and contributing heavily to it.
- The NASA VEDA project (in particular, Brian Freitag and Alex Mandel), for continued funding (in the form of engineering time) plus being early adopters!
