Very simple JupyterLab container
Go to file
Matthew Connelly 48c5fda6f3 entrypoint: Build before starting
Because /v/share/jupyter/lab is now a volume, there is the potential
that it will be empty or out-of-date. We should always kick off a build
before starting jupyter-lab.
2022-08-17 16:44:14 +01:00
.github/workflows Update and rename docker-publish-ghcr.yml to docker-publish.yml 2022-08-17 14:38:36 +01:00
Dockerfile build: Expose all volumes 2022-08-17 16:43:44 +01:00
LICENSE initial commit 2022-08-17 13:53:57 +01:00
README.md doc: Update readme 2022-08-17 14:02:23 +01:00
conf.py initial commit 2022-08-17 13:53:57 +01:00
entrypoint.sh entrypoint: Build before starting 2022-08-17 16:44:14 +01:00

README.md

docker-jupyter-simple

Very, very simple container for running JupyterLab (+Notebook, +Voila).

Runs whatever the latest stable Python is, exclusively runs inside a virtualenv, has NodeJS installed using nodeenv inside the virtualenv.

Using

docker volume create jupyter-data
docker volume create jupyter-conf
docker run --rm -p 8888:8888 -v jupyter-data:/data -v jupyter-conf:/root/.jupyter maffsie/jupyter-simple:latest

The JupyterLab web interface is exposed on port 8888. No packages are installed by default, but can easily be installed using pip.

Volumes

  • /data - where all user data is stored
  • /root/.jupyter - where all config and state is stored

Security

This container deliberately does not have any security considerations. It is built as root, runs as root, and I have no interest in changing that. This is not because I believe root is fine, or because I don't care about security, but instead is because this container was created in order to have a lowest-friction Jupyter setup. The existing docker containers do not have a clearly defined usage, for whatever reason.