Zenodo harvesting based on `sickle` is missing out on some information

The zenodo harvester currently makes use of the OAI-PMH interface of Zenodo and is based on the Python package sickle. However sickle is optimized for harvesting OAI-PMH resources based on the Dublin Core profile (metadataPrefix=oai_dc), which includes much less metadata then the datacite profile, see e.g.: https://zenodo.org/oai2d?verb=ListRecords&set=user-nfdi4earth&metadataPrefix=oai_dc https://zenodo.org/oai2d?verb=ListRecords&set=user-nfdi4earth&metadataPrefix=datacite

We want to harvest for example whether an article is part of the NFDI4Earth community, in the datacite profile this is: <relatedIdentifier relatedIdentifierType="URL" relationType="IsPartOf" >https://zenodo.org/communities/nfdi4earth</relatedIdentifier> while in oai_dc this is only represented as: <dc:relation>url:https://zenodo.org/communities/nfdi4earth</dc:relation> (the qualifier of the relation is missing so that it cannot be distinguished from other statements (e.g. isVersionOf) which also use dc:relation.

Furthermore we want to get the ORCID's of the authors not only the names, this is also not included in oai_dc.

I would therefore say that we move away from sickle and use another tool which harvests all metadata that zenodo offers. Any suggestions?