[FEAT] Data release versions vs. data reduction pipeline versions

[adrn]

(For context: This idea came up during @blanton144's tutorial during the science Friday call today, as Mike was showing off sdss_access)

I raised the point that often the data releases have a well defined reduction pipeline version associated with them. For example, for DR19 MWM, the relevant mapping is to "astra v0.6.0". Or, for this other example in the sdss_access docs:

from sdss_access import HttpAccess
http_access = HttpAccess(release='DR17', verbose=True)
http_access.get('mangacube', drpver='v3_1_1', plate='8485', ifu='1901', wave='LOG')

DR17 MANGA -> drpver v3_1_1

It would be useful if there was a way for sdss_access to do this mapping for the user. For example, so doing:

from sdss_access.path import Path
sdss_path = Path(release='DR19')
sdss_path.full("mwmAllStar") # knows that, by default: v_astra="0.6.0"

[havok2063]

This mapping is available in the datamodel product. To implement this, we would need to make the datamodel product a dependency of sdss_access to use as a lookup. However, before we can do that, datamodel needs some re-working to accommodate its use in other packages. e.g. make it a pypi package, and create a light-weight install of the python package.