We should add vocabularies to cold and generate https://github.com/caltechlibrary/irdm_harvester/blob/main/options.yaml. That way we can centralize all out metadata