Task/DES-2618: Consistent user metadata across project types

[jarosenb]

Overview:

Our representation of project users/authors is inconsistent across project types, which will cause issues when we try to migrate things onto Tapis V3. This PR makes the following changes to project/entity metadata, which will be reflected when entities are saved:

• Add a users field to projects that contains the PI, co-PIs, and team members. This will drastically simplify name-based searches in the future.
• All "publishable" entities (entities that can be assigned a DOI) now have an "authors" field that contains the first name/last name/email/institution for all users
• Previously, the object key for storing usernames was sometimes user and sometimes name. Now the key is username universally.
• Arrays of authors/data collectors are now sorted by their "order" attribute. We've fixed the underlying bug that was causing arrays to lose their ordering on serialization, so during the V3 migration we can deprecate the use of JSON attributes for keeping track of array order.

The metadata schema changes in this PR are a strict superset of the existing schema, so the current curation/pipeline operations will just ignore them until we need to use them.

PR Status:

• Ready.
• Work in Progress.
• Hold.

Related Jira tickets:

• DES-2618
• Relies on task/DES-2616: Add TAS institution to user profile and update it on login #1130 to be merged/migrated so that user profiles contain institutions.

Summary of Changes:

Testing Steps:

UI Photos:

Notes: