names: split name datastream during ORCiD dump loading

[slint]

When running the ORCiD dumps (or even in the schema?), if no family_name and given_name are present we should either:

• Split the name field by comma (,). E.g. name: "Doe, John" -> {family_name: "Doe", given_name: "John"}
• Don't allow only name to be accepted. This also requires updating the documentation at https://inveniordm.docs.cern.ch/customize/vocabularies/names/