get-dois

get-dois

A small module that takes a DOM tree and returns all Digital Object Identifiers (DOIs) located in that tree's metadata.

Unfortunately, academic publishers use a wide variety of metadata "standards". This package normalises between Dublin Core, Highwire Press tags, etc., and the different formatting rules used for DOIs.

To use this library outside of a browser context, you can use a library like jsdom.

This project also includes TypeScript type definitions.

Installation

npm install get-dois --save

Usage

import { getDois } from 'get-dois';

const head = document.getElementsByTagName('head').item(0);
const dois = getDois(head);
//=> [ '10.31222/osf.io/796tu' ]

API

getDois(domTree: Document | Element): string[]

Parameter: domTree

Either window.document or an HTML element whose children might include meta tags containing a DOI.

Returns

An array of DOIs. If no DOIs are found, this will return an empty array. Duplicate DOIs will only be listed once. Note that a page might list multiple DOIs.

Supported meta tags

At the time of writing, get-dois supports detecting DOIs embedded in a webpage as:

• Highwire Press tags (citation_doi)
• Eprints tags (eprints.id_number and/or eprints.official_url)
• bepress tags (bepress_citation_doi)
• PRISM tags (prism.doi)
• Dublin Core Schema markup (dc.identifier and/or dc.relation)

If you encounter a different format that is not yet supported, please file an issue.

Changelog

See CHANGELOG.md.

License

MIT © Vincent Tunru

Changelog

[2.0.0] - 2019-07-19

New features

• More ways of recognising DOI metadata are now supported! Specifically, the markup used by Pubpub is now recognised by get-dois as well. (Technical details: get-dois now also detects meta tags that label themselves as DOI metadata using RDFa (i.e. using the property attribute rather than name).)

Breaking changes

• If you're relying on a specific set of DOIs being detected, this release is breaking: more representations of DOIs can now be found. Altough I did not do that in the past, I'll be marking that as breaking from now on as well.