Create index terms / taxonomy / tag field for Merritt Object

[terrywbrady]

This idea originated from a proposal from UCB to track objects based on the originating DAMS object.

This is a feature that is common in many repositories. Merritt does not currently have such a concept.

What is the scope of an index term?

• repository-wide?
• owner-wide?
• collection-wide?
• namespace-wide?

What is the data type of an index term?

• String
• Number treated as a stream
• Identifier (namespace will identify the identifier)
• URL
• Lists of terms should be stored as 2 terms
• Paths (file path, taxonomy path)

When multiple terms are present, is there a hierarchy?

How and when are index terms applied?

• Is application of an index term considered to be a preservation action?

Should index terms for existing content be generated through database mining and/or application of AI?

Database storage

• TBD based on questions above

Which microservices would care about such a term?

• UI for search, presentation, object to object navigation
• INV for extraction to the database
• Ingest - if some sort of validation was enforced on index terms
  • Assumption: no update/replace decisions would be based on this field. That would be exclusive to the local id.
  • Assumption: objects would not be rejected from ingest because they duplicated an index term
• Unlikely to be accessed by other microservices

What system features could be replaced by this feature?

• creates an opportunity to have fewer collections
• not replacement per se, but existing search functionality could be refactored to use a search index rather than the database

Use Cases

• UCB - described above
• Palmu collection
• eScholarship
• Parsing sidecar metadata to surface index terms