[SCHEMA] Encode requirement levels for dataset_description and genetic_info

[effigies]

Two open questions I see:

1. How do we want to encode, e.g. NO_AUTHORS, where we effectively promote dataset_description.Authors from OPTIONAL to REQUIRED?
2. Do we want to specify nested fields, such as Genetics.Database directly? Or have some way of specifying a nested structure as its own object?

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 71.50%. Comparing base (1f6e4d2) to head (5e116de).

Additional details and impacted files

@@              Coverage Diff               @@
##           schema-sprint    #1064   +/-   ##
==============================================
  Coverage          71.50%   71.50%           
==============================================
  Files                  9        9           
  Lines                930      930           
==============================================
  Hits                 665      665           
  Misses               265      265           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[tsalo]

1. How do we want to encode, e.g. NO_AUTHORS, where we effectively promote dataset_description.Authors from OPTIONAL to REQUIRED?

I think since NO_AUTHORS is a warning, it would be upgrading from OPTIONAL to RECOMMENDED, right? In any case, what if we had a "publication" level for the validator? The checks in this level wouldn't have to be derived from the specification, but would be necessary for data sharing.