Add support for Croissant

[DonRichards]

Overview of feature request
Add support for Croissant, a high-level format for machine learning datasets that integrates four rich layers, allowing the export of metadata and resources as machine learning-friendly datasets. This feature would process content into a format that combines metadata, resource file descriptions, data structures, and default ML semantics into a single file.

What kind of user is the feature intended for?
Collections Manager, Systems Administrator, end User

What inspired the request?
The need for a feature to export metadata as datasets for machine learning purposes using Croissant. Croissant is a format designed to merge metadata, resource file descriptions, data structures, and machine learning semantics. While I am not a metadata librarian, this feature seems potentially beneficial for ML-driven projects.

What existing behavior do you want changed?
Allow existing records to be exported as datasets that are compatible with machine learning.

Any brand new behavior do you want to add to Islandora?
Add an export endpoint to enable this functionality.

Any related open or closed issues to this feature request?
N/a