SHARPy Docker now releases to GitHub Packages

[ngoiz]

Release SHARPy Docker image automatically to GitHub Packages

This is done in a CI workflow through a new GitHub action. Master, release and old versions docker images of SHARPy should appear in the Packages page of the repo. The installation instructions have been amended to show how to download the image and build a container. The Dockerfile has been amended to install SHARPy through pip, as introduced in v2.0.

A minor update is the change of how we store SHARPy's version. I have added an argument such that you can query the code version through sharpy --version and stored the version in sharpy/version.py which should be the one version of truth and only file that needs modification once a new release is ready.

[ngoiz]

Added some modifications such that the Docker image is only published if it is a push to master, develop or an annotated version tag starting as v*. For pull requests, it builds the image but does not published. Once this is merged into develop, the job test_image_build should be added as a required check to the currently protected branches

[kccwing]

Hello @ngoiz , thanks for updating the docker image and the automatic github action to generate these! We aim plan to switch over to a much simpler environment.yml file as part of the upgrade from py3.7 to 3.10, in #222 from @ACea15 (environment_new.yml), and hence will merge #217 and #222, then this, then switching to master as the default branch, then a new version.