EUCP Jupyterlab environment specification

Dockerfile for the EUCP project - based on the Jupyter datascience notebook
View on dockerhub

Branches

• master: for regular development and updates. Merging via approved pull requests. Linked to 'latest' tag on dockerhub.
• feature branches: make a separate branch for each PR/feature that you want to add.

Releases

• We'll configure the login environment such that you can select which release you want to use for your compute environment.

Update process

1. Checkout, build and run the docker image locally
2. Make desired changes to the environment
3. Build and run updated docker image locally for testing
4. Use conda list --export > package_list_for_change_diffs.txt, download and commit the result
5. Summarize changes made to the environment in the changelog below
6. Make a pull request to the master branch of this repository
7. Once the PR is merged, test changed in the 'live' environment by choosing 'latest'.
8. Consider making a new release and adding it to the ansible setup

Changelog

May 20, 2020:

• Inherit from newer jupyter/datascience-notebook
• Updated jupyter hub & jupyter lab versions
• Added explicit dependency of matplotlib
• Removed some packages as they are now part of base- or scipy-nb
• Removed some packages as they were installed as dependencies of others
• Moved all pip-installs to conda
• Add esmvalcore and esmvaltool-python
• Upgrade several packages to newer versions
• Disable not-working jupyter-thredds plugin

May 19, 2020:

• Pin versions to current environment configuration
• Add package_list_for_change_diffs.txt for keeping track of environment changes.

May 18, 2020:

• Start changelog
• Enable autobuild on dockerhub
• Create master and stable branches