eprints2rdm: related urls

[tmorrell]

For each related url:

• if type is doi, run the value through idutils normalize_doi function. If DOI is already in the DOI field, discard.

• if type is pmc, pull out the identifier at the end of the url. If the identifier is already in the PMC_ID field, discard. Otherwise transfer with scheme pmcid

• if type is arxiv, run the value through idutils normalize_arxiv function. If arixiv identifier is in DOI field, remove arxiv link.

• if type is ads, run the value through idutils normalize_ads and transfer

• if type pub, whitelist all urls that match any of the following:

• [ ] rdcu.be

• [ ] geoscienceworld

• [ ] http://ieeexplore.ieee.org/ and DOI prefix 10.1364

• if type is pub and description is "Article" and DOI is already in the DOI field (and not whitelisted), then discard

Assemble non-discarded urls as related identifiers with type url and relation type Describes. Run through all identifiers and remove any duplicates.

[rsdoiel]

Note to self, we want to do this in the Python fixup code, not the Go based cli.

[rsdoiel]

This is fixed in upcoming release v0.0.27.