Closed jacobthill closed 2 years ago
I think we discovered in huddle yesterday that if metadataPrefix is used, then hard coding the resumption token should not be needed?
https://api.qdl.qa/api/oaipmh?verb=ListRecords&metadataPrefix=mods_no_ocr
@jacobthill do you think the metadataPrefix is sufficient here since so far QNL is the only provider we have with that resumptionToken concern?
Yes, I think it makes more sense from an OAI-PMH perspective for us to configure metadataPrefix
in the catalog (using metadata_prefix
key in the YAML) rather than hard coding a resumption token.
To get QNL to work we will need to adjust the dlme_airflow.drivers.OaiXmlSource
to handle the mods xml, but we have #175 for that (which I picked up). The QNL catalog entry already is set to metadata_prefix: mods_no_ocr
so I think this issue can be closed?
Ooops, I just noticed the question was asked of @jacobthill :-) Sorry for jumping in. The resumption tokens are meant to be opaque, and possible transient identifiers for result sets. So I think we would be better off using the metadata prefix instead as a pattern here.
Sorry for the delay, I agree with @edsu
@edsu thank you, great info. closing. If we find we may need this, we can revisit later.
At least one collection uses resumption tokens which are configurable in Sickle like
records = sickle.ListRecords(resumptionToken='0mods_no_ocr')
. We need to pull this field from the config when available and, if not in the config, continue behaving as is.