DOI identifiers

Archive-reader recognizes DOI as a known identifier scheme when a publication states it explicitly. A publication can also carry a DOI as a general URL identifier, which is how the DOI resolver addresses one.

Publication metadata	Resolved identifier
EPUB 3 identifier explicitly typed DOI	{ value: name, scheme: "DOI" }
Absolute doi.org URL in an EPUB identifier	{ value: url, scheme: "URL" }
Absolute doi.org URL in ComicInfo Web	{ value: url, scheme: "URL" }

A DOI name is a registrant prefix and a suffix the registrant assigns — 10.1000/182. Suffixes are opaque and case-insensitive, and carry no length limit.

Explicitly typed EPUB identifier

<dc:identifier id="doi">10.1000/182</dc:identifier>
<meta refines="#doi" property="identifier-type">DOI</meta>

Resolves to:

{
  value: "10.1000/182",
  scheme: "DOI",
}

Reference URL

In EPUB metadata:

<dc:identifier>https://doi.org/10.1000/182</dc:identifier>

In a CBZ or CBR ComicInfo.xml:

<ComicInfo>
  <Web>https://doi.org/10.1000/182</Web>
</ComicInfo>

Both resolve to the literal URL:

{
  value: "https://doi.org/10.1000/182",
  scheme: "URL",
}

Reading the name

Because the same identifier arrives either explicitly typed or as a resolver URL, read it with identifierValue rather than filtering on the scheme:

import { identifierValue } from "@prose-reader/archive-reader"

const doi = identifierValue(metadata.identifiers, "DOI")
// "10.1000/182" | undefined

The doi: prefix is dropped and a percent-encoded suffix is decoded, so doi:10.1000/182 and https://dx.doi.org/10.1038%2Fnphys1170 both answer with the bare name.

See Publication identifiers for the shared scheme resolution rules, custom schemes, URLs, and the Unknown fallback.