"Find Available PDFs" Not Working with VPN

I'm working on a systematic review, and as part of the process, I've uploaded a reference list to Zotero that I need full text PDFs for to then upload to a separate software (Covidence). However, when I use "Find Available PDFs" to try to populate the collection with PDFs, it's not able to find many of the PDFs that I know I have institutional access to (and can access fine in my browser). I'm using a full tunnel institutional VPN, so I'm not sure why Zotero isn't recognizing so many of the PDFs. Any troubleshooting help you have would be greatly appreciated -- I've been through some of the older forums about this but still can't get it to work.

VPN: Cisco AnyConnect Secure Mobility Client v. 4.8.02045
Zotero: 5.0.90
  • Can you provide a Debug ID for an unsuccessful attempt to find a PDF for an item where you can access the PDF from the webpage?
  • For sure: D1664131587
  • ReferenceError: URL is not defined
    This is fixed in Zotero 5.0.91, available now via Help → Check for Updates.
  • Just updated Zotero and am still having the same problem, unfortunately!
  • Can you provide another Debug ID?
  • edited October 9, 2020
    If you open a private window in your browser and load this URL, do you see a PDF without logging into anything? Zotero is getting a webpage trying to download that, suggesting that you don't have direct system-wide access to that PDF.
  • When I went to that link, it automatically downloaded the PDF...then prompted me to log in to my university account...then downloaded it again.
  • I don't really understand what you mean by that, but if you're getting anything than a plain old PDF at that URL — e.g., what you see here — this won't work.
  • edited October 9, 2020
    The point of a system-wide VPN would be that it bypasses all authentication. If that's not happening, Zotero won't be able to download PDFs.

    It's possible that Springer is doing something weird where it sends a PDF separately from the containing webpage. I don't have a VPN with access to Springer, so that's not something I'm able to test, but one way or another, your Zotero is getting a webpage at that URL rather than a PDF.
  • edited October 9, 2020
    For what it's worth, via a web-based proxy I do get a direct PDF — with a download forced by the site, so it doesn't preview in the browser like other PDFs, but that doesn't matter — at that URL.

    I'm not sure why you would be getting any sort of webpage there.
  • When I went to the hyperlink from your earlier post in a Chrome browser, it didn't pull up a PDF (like your bitcoin example), it just automatically downloaded a PDF while showing a blank webpage.

    Incidentally in that example, after it downloaded the PDF the first time it THEN asked me to authenticate, and then downloaded it again. But I just disconnected and reconnected my VPN and tried again, and it no longer asks me to authenticate.

    If it is a problem with Springer in particular, is there any way around it? Or do I just need to re-find those citations and download them with the browser plug-in?
  • FWIW: I just tried again on the full listing of articles that I'm trying to find available PDFs for, and it seems like from the Debug Output that it's also happening with Wiley articles
  • When I went to the hyperlink from your earlier post in a Chrome browser, it didn't pull up a PDF (like your bitcoin example), it just automatically downloaded a PDF while showing a blank webpage.

    Incidentally in that example, after it downloaded the PDF the first time it THEN asked me to authenticate, and then downloaded it again.
    I just don't really understand this. Downloading while showing a blank webpage makes sense — the site triggers downloads for PDFs rather than letting them be shown in the browser — but if it's a blank webpage, then what's prompting for authentication? Can you take a screenshot, upload it somewhere (e.g., Dropbox), and provide a link here, or email it to support@zotero.org with a link to this thread?
    But I just disconnected and reconnected my VPN and tried again, and it no longer asks me to authenticate.
    Assuming this was a prompt within your browser, you would need to close the private window and try again in a new private window to test again. But I'm not understanding how this could be a prompt in your browser if it already downloaded the PDF.
  • edited October 9, 2020
    it seems like from the Debug Output that it's also happening with Wiley articles
    We can look at a Debug ID for one of those.
  • edited October 9, 2020
    I just emailed two screenshots!

    And a Debug ID for one of the Wiley articles is here: D496401982
  • (3)(+0000544): Downloading file from https://onlinelibrary.wiley.com/doi/pdfdirect/[…]

    (3)(+0000302): Getting contents of /Users/[…]/Zotero/storage/tmp-637BYK/[…].pdf

    (2)(+0000000): Downloaded PDF was not a PDF

    (3)(+0000000): <!DOCTYPE html>
    <html lang="en" class="pb-page" data-request-id="58ddcb26-a996-48d9-b533-1c6c53fd3edf"
    ><head data-pb-dropzone="head"><meta http-equiv="X-UA-Compatible" content="IE=edge">
    <meta name="pbContext"
    I'm afraid you're just not getting direct PDFs from these URLs when they're loaded from within Zotero. It's possible that Zotero isn't actually connecting through the VPN. One easy test would be to go to Tools → Developer → Translator Editor in Zotero, go to the Browser tab, and try to view a URL that you'd only have access to via the VPN. If you don't have access, Zotero isn't connected via the VPN. I don't know how the Cisco VPN client works, so I couldn't really advise how to fix that.

    Re: the redirects in your browser, that might be the Zotero Connector's web proxy redirection support doing that after the PDF download. If you always use your VPN to access gated resources, you can just disable the redirection feature. With a VPN, I wouldn't think you'd ever need to connect via a web-based proxy.
  • Thanks for the test advice -- from the Browser tab, it does seem that Zotero isn't connected via the VPN.
Sign In or Register to comment.