When using the "Add by identifier" feature to import entries from Zenodo (or other Invenio RDM repos) by DOI, the author names end up fully in the last name field. As Zenodo recommends writing authors in the "last, first" format, this results in a Zotero entries having "Doe, John" as last name, followed by an empty first name.

As far as I can see, the problem is that the DataCite JSON translator only uses the familyName and givenName fields if the `nameType="Personal"` attribute has been declared for the creator field. As a fallback it stores the entire string as last name, but unfortunately fails to switch to single field names (it does in fact declare `"fieldMode": true`, so I suspect an API regression here). Zenodo/Invenio RDM don't provide the attribute, but do provide familyName/givenName fields, as well as OrcIDs.

I don't know whether ignoring familyName/givenName was a conscious decision because it might come from auto-splitting strings, but I think it should be trusted (at least) when accompanied by OrcIDs, probably always. I expect the number of thus properly imported person names should vastly outstrip the number of organisation names that get incorrectly split due to having a single comma in their name.

Tested on Zotero 6.0.26, OSX.