Catalog schema updates

[ceblanton]

Proposal of "required" attribute for each column; either

• true: always required
• enhanced: optional but will be filled by canopy
• false: optional

Required vocabulary:

• experiment_id
• frequency
• realm
• variable_id
• time_range
• path (not exactly a normal vocabulary, but is required)

Optional vocabulary:

• chunk_freq
• activity_id
• institution_id
• source_id
• table_id
• member_id
• grid_label
• platform
• target
• version_id

Enhanced vocabulary:

• dimensions
• cell_methods
• standard_name

Unique vocabulary (can define a unique dataset):

• source_id
• experiment_id
• frequency
• table_id
• grid_label
• member_id
• chunk_freq

[ceblanton]

Let's make the tests pass; then merge

[ceblanton]

@aradhakrishnanGFDL and @Ciheim , the CI tests failures were real problems so they were good to see. After updating the "modeling_realm" mentions to "realm" the tests passed.

So as we agreed this is going in!