Jupyterhub plugin for Moodle
This repository contains 2 plugins made for Moodle 4.3. Those plugins are designed to make the communication from Moodle to a JupyterHub installation possible, allowing Moodle users to access their Jupyter files from Moodle - and more.
This plugin adds a new type of Assignments: Jupyter notebooks.
- Teachers can select the assignment's material (notebooks, data, images, etc.) from their own Jupyter workspace
- Students can upload the assignment's material to their own Jupyter workspace, and submit their work
This plugin is still under development ; the current version allows teachers to download all students' submissions in one click into the teacher's Jupyter workspace.
Plugins' content need to be copied over to:
[moodle_root]/mod/assign/submission/noto
and
[moodle_root]/mod/assign/feedback/noto
respectively, on the Moodle server.
On the JupyterHub side, an API needs to be deployed on a server that has access to all user's files - typically the file server of the JupyterHub installation. In Kubernetes, the API is deployed in the same namespace as the JupyterHub.
See this repository for the API: epfl-cede/jupyterhub-fileserver-api
Go to Site administration and set some global options first:
Select the option ETHZ Installation to switch to the settings suitable for Kubernetes.
At ETHZ, we run one JupyterHub per Moodle course, isolated in different namespaces. The key for the different instances is the Moodle course ID. The configuration field allows to use a place holder for the Moodle course ID:
https://your-api-url-base-[courseid].example.com
Authentication to the API is secured by a shared key/secret pair. Python 3.6+ example to create a key or secret:
import secrets
secrets.token_hex(32)
Use the same credentials to configure the AUTH_USER and AUTH_KEY variables for the API
deployment.
Set to idnumber and leave the prefix empty.
To configure an assignment activity, your Jupyter Hub home directory must already exists. In Kubernetes, it gets created upon your first login into Jupyter.
In the Submission types section, activate the Jupyter notebooks option. The contents of your home directory will be displayed. Browse your directory tree and select the source folder of your assignment.
Note: the path to the source folder is immutable and cannot be changed later.
Students assigned to the activity will see a similar dialog where they can select a folder in their own home directory to fetch a copy of your source.