Draft manuscript PDFs automatically imported instead of available full-text in browser

edited November 18, 2020
Zotero 5.0.93 on MacOS Mojave
Connector 5.0.76
Google Chrome

I've had this happen multiple times:
Navigated to a full-text resource in a journal, hit the connector button to save PDF. The entry is created but Zotero automatically finds an open access PDF from some other source, sometimes a draft of a paper manuscript with different formatting. The full-text PDF is directly available from the journal, I have to manually click on the link download it and drag it to the Library and replace the automatically imported PDF or use the Zutilo attach new file using the watch folder. This is cumbersome and defies the point of having the connector import PDFs.

Screenshots:
Full-text access: https://ibb.co/r0FQr3b
Zotero (ideally) should import: https://ibb.co/qjVYwcd
Zotero imports: https://ibb.co/tMy0qW7

Suspect:
Might have something to do with institutional access through EZproxy library/Single Sign On (SSO) access; full text becomes available only when journal detects appropriate credentials. Somehow Zotero just does not 'see' the direct PDF resource available in the browser. However, if I am not mistaken this might have happened in cases of particular open-access PDFs as well not requiring institutional access - Zotero picks up its own PDF version from crawling the web rather than the version available in the browser.


  • Zotero searches for an open access copy when it can't download the "original" PDF, so this is site specific. We'd need a URL where this happens to say more.
  • Thanks @adamsmith
    Here's the URL for the example given in the screenshots
    https://www.nature.com/articles/nclimate2765

    It has happened with other journal sites as well.
  • That should certainly work. Is that the URL exactly as you see it or is there some proxy information included that you left out? When you hover over the Save to Zotero icon, what does it say exactly?
  • There is an institutional validation after credentials are entered but the only visible URL that remains is the link I posted. It says 'Save to Zotero (Nature Publishing Group) has access to this site.'
  • I'm guessing that's the problem then. Most institutions authenticate using web-proxies and those work fine, but Zotero may not be picking up/be able to pick up your authentication. Do you happen to know which service is used for that (like OpenAthens etc.)?
  • It doesn't explicitly say, but when I hover over (my) institution re-direct link from the drop-down for credentials and full-text access, shibboleth.xx.xx.xx is part of the link URL so I suppose Shibboleth is the service in this case.
  • Can you provide Debug IDs from the Zotero Connector and Zotero for a save attempt that doesn't save the published PDF?
Sign In or Register to comment.