tutor-jupyter 0.1.0

Jupyter Notebook plugin for Tutor

This is a plugin for Tutor that makes it easy to integrate Jupyter notebooks in Open edX. It achieves the following:

1. Install the official jupyter-xblock in the Open edX LMS and Studio.

2. Run a Docker-based JupyterHub instance with a Docker spawner.

In pratice, it means that students will be allocated Docker containers with limited CPU and memory to run their custom notebooks.

⚠️ This plugin is not compatible with Kubernetes. If you wish to run JupyterHub on Kubernetes, you are encouraged to check the documentation of the Zero to JupyterHub with Kubernetes project.

Installation

pip install tutor-jupyter

Usage

Launch a local platform

tutor plugins enable jupyter mfe
tutor config save
tutor images build openedx
tutor local launch

TODO Create LTI passport in Open edX:

echo "$(tutor config printvalue JUPYTER_DEFAULT_PASSPORT_ID):$(tutor config printvalue JUPYTER_LTI_CLIENT_KEY):$(tutor config printvalue JUPYTER_LTI_CLIENT_SECRET)"

Configuration

Settings

TODO

• JUPYTER_DOCKER_IMAGE_HUB (default: "{{ DOCKER_REGISTRY }}overhangio/jupyterhub:{{ JUPYTER_VERSION }}")

• JUPYTER_DOCKER_IMAGE_LAB (default: "{{ DOCKER_REGISTRY }}overhangio/jupyterlab:{{ JUPYTER_VERSION }}")

• JUPYTER_HOST (default: "jupyter.{{ LMS_HOST }}")

• JUPYTER_DEFAULT_PASSPORT_ID (default: "jupyterhub")

• JUPYTER_LTI_CLIENT_KEY (default: "openedx")

• JUPYTER_HUB_MYSQL_DATABASE (default: "jupyterhub")

• JUPYTER_HUB_MYSQL_USERNAME (default: "jupyterhub")

• JUPYTER_LAB_CPU_LIMIT”, None

• JUPYTER_LAB_MEMORY_LIMIT (default: "200M")

• JUPYTER_HUB_COOKIE_SECRET (default: "{{ 32|jupyterhub_crypt_key }}")

• JUPYTER_HUB_CRYPT_KEY (default: "{{ 32|jupyterhub_crypt_key }}")

• JUPYTER_HUB_MYSQL_PASSWORD (default: "{{ 24|random_string }}")

• JUPYTER_LTI_CLIENT_SECRET (default: "{{ 24|random_string }}")

Modifying the default lab environment

TODO implement “jupyter-lab-dockerfile” patch:

from tutor import hooks

hooks.Filters.ENV_PATCHES.add_item(
    (
        "jupyter-lab-dockerfile",
        """
# Install extra Python packages
RUN pip install matplotlib scipy seaborn
"""
    )
)

Then build the lab image again:

tutor config save
tutor images build jupyterlab

License

This software is licensed under the terms of the AGPLv3.