PDFs from Safari on iOS imported as Web page instead of PDF

edited December 16, 2022
When I try to import a PDF from Safari into Zotero using the share sheet, Zotero will initially view it as a PDF (PDF icon shows with file name), but after a few seconds will change to a web snapshot (web icon with URL). If I continue to import it, the resulting item type in Zotero is a web page and the PDF is not imported.

It doesn't happen every time. Sometimes I can add the PDF file as expected.

It seems to happen frequently on PMC, FWIW.

Example: https://www.sciencedirect.com/science/article/pii/S1556370722000256
PDF: https://pdf.sciencedirectassets.com/274165/1-s2.0-S1556370722X00072/1-s2.0-S1556370722000256/main.pdf?X-Amz-Security-Token=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&X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Date=20221216T234341Z&X-Amz-SignedHeaders=host&X-Amz-Expires=300&X-Amz-Credential=ASIAQ3PHCVTY352NNKUP/20221216/us-east-1/s3/aws4_request&X-Amz-Signature=6016f227531409ddc227fde881f8520276e20b22764a24beb658340f4ea824be&hash=cb8230863210299039294de0227513f22dfa936ffb5869e31deca2aa5a2a1709&host=68042c943591013ac2b2430a89b270f6af2c76d8dfd086a07176afe7c76c2c61&pii=S1556370722000256&tid=spdf-24e8bf65-f4c3-4286-9fa2-5fea1297c690&sid=f8891dc757bad14d993bf9f9db2951c30e02gxrqa&type=client&download=true&ua=515404035a575003540b&rr=77ab485d7e9e8e63

Any ideas?
  • Where exactly are you saving from? If you're saving on a supported site like ScienceDirect, you'd always want to save from the article page, not the PDF itself. Zotero should save high-quality metadata and also automatically download the PDF.

    On the ScienceDirect URL above, the "PDF" isn't actually the PDF — it's a custom web-based reader — so saving from there wouldn't download the PDF itself. (The iOS app also doesn't currently support retrieving metadata for PDFs, so while in the desktop app saving a PDF itself would often result in a parent item being created, that won't happen on iOS anyway. But it's not necessary because you can just save from the article page to get the metadata and PDF.)
  • Unfortunately, ScienceDirect in particular is a tricky case, since Elsevier has some zealous anti-bot behavior that makes saving PDFs harder than on any other site, and I am currently seeing it hang trying to attach the PDF on that article page. We'll look into that. But this should be unique to ScienceDirect.
  • And I'm also seeing the webpage behavior you describe when using the download button from the web-based reader and trying to save that PDF directly. That may also be a result of their anti-bot efforts, but we'll see what we can do. I think you'll find that saving PDFs directly on other sites generally works fine, even if it's still not the recommended approach.
  • I always try to save from the publisher’s article page to create an article item with metadata and the retrievable PDF as an attachment.

    But in many of these cases, automatically retrieving the PDF fails. I can often click “Download PDF” on the article page to open in Safari, then I can import using the Zotero share extension and merge the PDF to the item in Zotero.

    This post is about the extreme case where even that doesn’t work. I did notice the web-based reader behavior you describe on desktop. But on iPad it *looks* like a standard PDF, so I wasn’t sure. In these cases, I save the PDF to Files using the share menu and then manually import with “Add File” in the Zotero app.

    ScienceDirect is frequently a problem, but I’m fairly certain it is not the only offender. I’ll try to keep track of any other sires and report back.
  • Thanks for reporting @nathancashion, we were able to resolve the issue and you should be able to use the share extension in above links when new app update is available.
  • Thanks, @michalrentka. Glad to hear it.

    Here’s another example with slightly different behavior.
    https://journals.lww.com/jopte/fulltext/2015/29020/the_use_of_motivational_interviewing_in_physical.9.aspx

    PDF wouldn’t import automatically using the Zotero share sheet extension (iPadOS). When I opened the PDF from the article page, and tried to import to Zotero I got the error seen here: https://www.dropbox.com/s/lqhek41eu72msax/ZoteroPDFerror.jpg?dl=0

    Unlike above, where the Zotero extension appears to attempt to download it but can’t find an available PDF, this seemed to fail to try.

    Let me know if I can explain that better.
  • Thanks for your report @nathancashion. The link you posted seems to work for me. We'll push the fix to beta so that you can check it out and see if it resolves your issues.
  • @nathancashion: OK, if you don't mind, could you install the beta version and confirm that that fixes the problem for you? (You can switch back to the App Store version after if you want, or switch when 1.0.9 comes out. There are no other changes in the current beta.)
Sign In or Register to comment.