Add data download test

[CodyCBakerPhD]

We need a minimal test of an easy-to-use workflow that downloads a particular file (not a full dataset) from EMBER

I'd also greatly prefer if this also serves as the first test for the style where we actually manage two things:

• the minimal GHCR image containing the DANDI package (could also conceptually just do it minimally via requests if Pythonic or pure curl, but that would involve 'figuring out' a handful of queries to the REST API)
  • let me know if you have any questions about how to build and host this
• the Pennsieve-style GitHub repo with a bare minimal Dockerfile that uses an exact version of that GHCR image

Will likely have to leverage environment variables (those pairwise key/value parameters Pennsieve allows) to communicate what file from what Dandiset is to be downloaded

Assigned to @rhingo, as always let me know if you have any questions or need any pointers

[rhingo]

Have a basic example of the 2nd version (Pennsieve-style GitHub repo) working:

• Downloading data from here: https://dandiarchive.org/dandiset/000027/0.210831.2033/files?location=sub-RAT123&page=1
• Pennsieve-style GitHub repo: https://github.com/aplbrain/dandi-cli-docker