The identifier a portal knows a dataset by, read off its own URL

A fallback, never an override: a curated id in metadata/distribution.csv always wins. Returns NA_character_ when the URL says nothing — a portal home, a search page — because a guessed identifier is worse than none.

Usage

derive_registration_id(url)

Arguments

url

character vector of URLs

Value

character vector of the same length

Details

Recognised: EDI (packageid=, or scope=&identifier=&revision=), NCEI (id=), OBIS (/dataset/{uuid}), a GBIF/OBIS IPT resource (?r=), any DOI (doi.org/10.…), CalOOS (#module-metadata/{uuid}), any ERDDAP (a tabledap/griddap page, an info page, or an ISO 19115 / FGDC document), DataZoo (/datasets/{n}) and an NCBI BioProject.

Examples

derive_registration_id("https://obis.org/dataset/0e223f55-c826-4513-ae9a-b04cbf2e189c")
#> [1] "0e223f55-c826-4513-ae9a-b04cbf2e189c"
derive_registration_id("https://portal.edirepository.org/nis/mapbrowse?packageid=edi.109.4")
#> [1] "edi.109.4"