Allow custom PDF resolver to use downloadPDFViaBrowser

yuanqi
As implemented in #2248. This would allow use of libkey.io, which uses Javascript to generate the page that eventually redirects to the PDF file, as suggested in this post (https://forums.zotero.org/discussion/91646/libkey-and-custom-pdf-resolver).

It would also be nice if Zotero Connector can share the browser's EZproxy cookie so that this can work with EZproxy, instead of requiring IP authentication or VPN.
  • dstillman
    But as I say on that page, Find Available PDF already follows the DOI redirect to the canonical URL and runs translation on the page to see if there's a PDF. What would this accomplish that that doesn't already provide?
    It would also be nice if Zotero Connector can share the browser's EZproxy cookie so that this can work with EZproxy
    Unfortunately it's a bit more complicated than that, but we're hoping to address this.
  • yuanqi
    Libkey links directly to the PDF, so it can help when the translators fail to identify a PDF file on the article page.
  • dstillman
    If a translator fails to identify a PDF, the translator should be fixed so that the PDF is saved for everyone saving from the site.
  • yuanqi
    Libkey in fact offers an endpoint that redirects to the PDF without using client-side javascript, but they don't return that URL through their public API. It is however possible to construct this URL manually using the ID field returned by their API. I just don't know how I can interface this with the custom PDF resolver mechanism
  • dstillman
    Right, I understand that. I'm saying this is already functionality that Zotero provides. If Zotero isn't saving a PDF from a publisher page you have access to, that's a bug that you should report so that it can be fixed for all Zotero users.
  • yuanqi
    I prefer to use PubMed as the source of my citation data, does the DOI article retrieval function work if I save from the PubMed page?
  • dstillman
    PubMed is actually a bit of special case. When saving from the connector and a PDF isn't available, we don't follow the DOI URL to look for the PDF, because in the vast majority of cases people will already be on the publisher site, so it would just be a redundant request. But PubMed is different — that will never be the canonical URL, and they don't have PDFs.

    We'll look into adding special handling for PubMed and a few similar sites to better deal with this. For now, you can obviously just click the DOI link and save from there to get the PDF.

    (A manual Find Available PDF after the fact would check the DOI URL, but that's of course limited by the IP/VPN restriction for the moment.)
Sign In or Register to comment.