Re-think how this module uses bibcite_entity.mapping.* configuration

Problem/Motivation

This issue comes from a discussion on #3584018: [RIS] Synonymous keys should be treated the same (Comment #14 and replies).

I feel like this is both opaque and not ideally flexible. If we wanted "ideal," this would probably involve reworking ReferenceNormalizerBase::getContributorKey to look for multiple keys that are considered by Bibcite to be authors/editors of some sort.

Perhaps the scope of this issue should be restricted to authors and editors, or perhaps it is worth considering all the entity mappings. What is the expected behavior if more than one key maps to the same field? Is a mapping of key to fields even the correct structure?

Should these be "simple" configuration objects or should the bibcite_entity module define a config entity type with validation?

Steps to reproduce

Proposed resolution

Remaining tasks

User interface changes

API changes

Data model changes