a560089aed
also push to gitea |
||
---|---|---|
.github/workflows | ||
Dockerfile | ||
LICENSE | ||
README.md | ||
conf.py | ||
entrypoint.sh |
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
. On each startup, the entrypoint script checks for the existence of the file /data/requirements
and runs pip install -Ur /data/requirements
if present - an easy way to make sure everything you're using is up-to-date is to simply restart the container. This is also how you are expected to install plugins.
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.