Add vendorization of the Publisher for to_datacite

[yarikoptic]

ATM we have it hardcoded

    attributes["publisher"] = {
        "name": "DANDI Archive",
        "schemeUri": "https://scicrunch.org/resolver/",
        "publisherIdentifier": "https://scicrunch.org/resolver/RRID:SCR_017571",
        "publisherIdentifierScheme": "RRID",
        "lang": "en",
    }

[yarikoptic]

publisher is described at https://datacite-metadata-schema.readthedocs.io/en/4.6/properties/publisher/

should be addressed as a part of the

• Initial Multi-DANDI instance support: import time - generic regexes by default + instance config #294

by adding an env variable [DJANGO_]DANDI_DOI_PUBLISHER which for be a URI where we add for now support only for RRID. So that env var would be RRID:SCR_017571. Later we could add others there. In Config make it required to be a URI like RRID:\S+.

[candleindark]

@yarikoptic Are you sure you want to name the env var to [DJANGO_]DANDI_DOI_PUBLISHER instead of just [DJANGO_]DANDI_PUBLISHER? After reading the doc, https://datacite-metadata-schema.readthedocs.io/en/4.6/properties/publisher/, I think this env var should point to a publisher of the resource, and I don't think the resource in the current context is a DOI.

[yarikoptic]

I don't really care, whatever works ;-)

[yarikoptic]

note that "Publisher" without "DOI" might mislead since we do "Publish" dandisets as "release them".

[candleindark]

Isn't the publisher we are providing here the entity that's publishing the dandisets in a DANDI archive instance? That's how I see it as you currently have "publisherIdentifier": "https://scicrunch.org/resolver/RRID:SCR_017571" hardcoded in the to_datacite() function. With that understanding, I don't see why "doi" needs to be in the name of the field.

[yarikoptic]

ok... I think we better be off if we get away from the "Publisher" here. What about DANDI_INSTANCE_IDENTIFIER ? then we use that IDENTIFIER for the "Publisher" which is the archive instance!