Information for Publishers
How to ensure Mendeley Web Importer can read documents from your website
Mendeley Web Importer tries to get document metadata from websites in four ways.
1. DOI lookup (Recommended)
If the DOI (Digital Object Identifier) of the document on the web page is tagged correctly, Mendeley Web Importer will automatically read it and retrieve metadata for it.
We recommended specifying DOIs on your web pages using the following META syntax:
<meta name="citation_doi" content="10.1038/nature09108">
2. HTML Meta Elements (Supported)
Sometimes a document does not have a DOI, so Mendeley Web Importer also supports reading other META elements directly.
For catalog pages, HTML meta elements should be specified in the HTML HEAD section.
There are 5 supported "standards" for describing articles in the META elements:
Highwire Press tags (citation_*)
bepress tags (bepress_citation_*)
Eprints tags (eprints.*)
PRISM tags (prism.*)
Dublin Core (dc.*/dcterms.*)
Presenting more than one set of tags is not a problem. Use Dublin Core tags only as a last resort because they lack unambiguous fields for journal articles.
Here is an example of how an HTML page may be marked up with META elements:
<meta name="citation_title" content="Global and local fMRI signals driven by neurons defined optogenetically by type and wiring.">
3. Metadata extraction (Fallback)
It is sometimes not possible or feasible to implement HTML META elements for standalone PDF content.
In this case, Mendeley Web Importer will upload the PDF file to its PDF metadata extraction service to identify metadata heuristically. (The uploaded PDF is discarded after processing and not stored on the server.)
It is not possible for this service to operate with 100% accuracy. Providing DOI or HTML META elements specifying metadata is preferred.
4. COinS – ContextObjects in Spans (Legacy)
COinS was a method of making metadata available within <span> elements in HTML, but is now largely unsupported. We are no longer officially supporting COinS, but we retain our legacy COinS parsers within the application as a last resort for legacy compatibility.
Please note that we only currently support the following OpenURL metadata description formats:
Both 0.1 and 1.0 OpenURL versions should work, but no assurances is made regarding compatibility.
How Mendeley Web Importer retrieves full text PDFs from websites
Mendeley Web Importer retrieves full text PDFs on behalf of the user in order to save them time and energy in “clicking through”.
Following is a brief overview of how this works. This is the technical approach behind both the “View PDF” button injection and full text PDF retrieval from Mendeley Web Importer upon import to library.
Elsevier (Mendeley) is part of the coalition for responsible sharing. Mendeley does not source full text PDFs from illicit or grey area sources, such as Sci-Hub.
1. Mendeley Web Importer automatically runs a script on allowlisted web pages
Mendeley Web Importer automatically runs a script on a selected allowlist of webpages to detect article identifiers.
It also runs scripts to detect article identifiers any time that the user manually invokes the Mendeley Web Importer window from the browser toolbar.
The detected identifiers are passed to the next step.
2. The extension uses proprietary mechanisms to locate full text PDF links in real time
If one or more articles is detected and the site is allowlisted, Mendeley Web Importer attempts to locate a valid full text PDF for the top-most or main article only.
The extension draws upon Mendeley’s Open Access links catalogue and a custom library of link-follow and page-parsing strategies in order to perform this function.
If a PDF link is successfully located, this is indicated to the user through the user interface. Mendeley Web Importer displays either a red “View PDF” button on the web page, or displays a red “PDF” icon adjacent to the detected reference within the Mendeley Web Importer window.
3. Requests for full text are made at the point of need
If the user chooses to click the “View PDF” button or import a reference along with a full text PDF, then the browser extension at that time makes a request for the PDF on behalf of the user.
This is functionally equivalent to the user clicking the download link on your website by themselves and ensures full counter compliant.