Allow systems to add unique ids for records to allow for merge-back workflows

[vikhari]

Issue ported from old casework github repo (issue 28).
Original author: mike-parkhill

We often have workflows that involve exporting out data to a third-party tool and then after the user has done what they need to in the third-party tool (e.g. picture categorization, annotations, etc.) they want to bring the data back into our system. In order to aid merging back data it would be nice to have a way of including our own unique ids on the objects. Presumably other tools in the chain could want the same, so a collection of "system ids" is probably required.

[vikhari]

Reply by casework:

The ID's by default are generated by the library generating the RDF graph. Usually these libraries allow you to use your own unique ID if desired.

For example, the rdflib python library allows you set an ID when creating blank nodes (http://rdflib.readthedocs.io/en/stable/rdf_terms.html#bnodes) and even supports merging graphs (http://rdflib.readthedocs.io/en/stable/merging.html)

[vikhari]

Reply by sbarnum:

This should be covered by Issue 38 in the UCO tracker.