Dockerized containers for running Jupyter, as well as a git filter appropriate for scrubbing notebook output cells so they don't end up in your history.
These images use pip for package management.
Add the filter to your git project by running configure.sh.
Build docker containers of your choice using docker/docker-compose commands. Before running any docker-compose commands, set up the environment by exporting the variables listed in docker/env.sh. Reasonable defaults can be set by sourcing docker/env.sh.
Alternatively, you can use ./scripts/jupyter-here.sh as a convenience function for serving notebooks on your system.
The Dockerfile's aren't complicated, but do document nuances in building various libraries and executables needed by each package.