Easier handling of Docker images for computing environment vs. included data

[poplarShift]

The template presently defines two Docker images: env for only the computing environment and full for one including all data (potentially much bigger). This is needlessly complicated. A better idea would be to keep and up-to-date computing environment image, and only add in data and directory structure once the project has reached a milestone (basing it off the computing environment image using FROM and COPY)

[poplarShift]

A better option may be to base it off https://gitlab.com/achimr/datasci