Zotero downloads then deletes PDF

andrew.weisman
For about a year now Zotero appeared to not download any PDFs on the page I was reading, those that have working translators, like those in acs.org. I ignored the problem, but I just noticed something: Zotero appears to download and then immediately delete the PDF! That is, when I click the download icon in the address bar, both the snapshot and PDF appear to be downloading and I see the icons/lines for these in Zotero, but immediately the PDF disappears and I'm left with just the snapshot. Any ideas? I apologize if I missed this issue in a previous post.

Thanks,
Andrew
  • adamsmith
    which site? What you're seeing happens when the file that Zotero downloads is not, in fact, a pdf.
  • andrew.weisman
    Hi Adam, thanks for your fast reply.

    Here is an example of a NOT working article (PDF does not download):
    http://pubs.acs.org/doi/abs/10.1021/jp310110r

    Here is an example of a WORKING article (PDF downloads as expected):
    http://pra.aps.org/abstract/PRA/v87/i2/e020302

    Please let me know what other information you need from me!

    Thank you,
    Andrew
  • adamsmith
    ACS PDFs download fine for me.

    - Can you open the PDF w/Links from the URL above?
    - Are you using Zotero in Firefox or with another browser and connector?
    - Are those the URLs as you see them or is there an institutional proxy in the URL (something like: http://pubs.acs.org.ezproxy.yourinstitution.edu//doi/abs/10.1021/jp310110r )
  • olunet
    Confirm. See also http://forums.zotero.org/discussion/9494/bug-pdf-is-added-and-then-deleted/

    That is a bit annoying. I can download the article and add it manually, but I would love to rely on automatic solution.

    I've tried to use VPN and proxy from different institutions without success.
  • adamsmith
    confirm for ACS or another site?
    The disappearing of the PDF itself is not a bug, but Zotero doing what it's supposed to do when something isn't a PDF. That's exactly what happened in the thread you link to.

    The problem is always site/translator specific, so we need sample URLs. Also, see again the question on whether this is using Firefox or Standalone.
  • olunet
    It is standalone version.
    I'm using firefox and chrome with connector (with proxy, with VPN).
    I've tested two sites above.
    But also for my searched articles I could not get pdfs from APS and ACS sites. This is systematic.
    Have a feeling that this is related to zotero and how it parse the pages.
    Please write what kind of feedback or information would you need.
  • dstillman
    edited March 2, 2013
    Have a feeling that this is related to zotero and how it parse the pages.
    Unless you're familiar with the code, don't make statements like this if you actually want help—you're just going to annoy the people who would otherwise help you. You have no idea what the problem is.

    As adamsmith said, we need sample URLs where this happens—not just site names—in order to debug this further. A Debug ID for a save attempt would also be helpful.
  • adamsmith
    edited March 2, 2013
    this is a different problem then - the users above _does_ get PDF downloads from APS.
    Start a new thread, say whether you have ever gotten any PDF to attach, e.g. on open access sites such as
    http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2738645/
    also, include information on 5,6, and 11 here: http://www.zotero.org/support/troubleshooting_translator_issues

    edit: like Dan says, specific sample URLs and Debug would be helpful as well, but do start a new thread please.
  • olunet
    Dear Dan, thank you for pointing to Debug_ID.
    Zotero standalone 3.0.14, ZotFile 2.3.1, three journals:
    1) http://pubs.rsc.org/en/content/articlelanding/2013/ra/c3ra23239e
    "Full Text PDF" is added and then dissapears
    2) http://jcp.aip.org/resource/1/jcpsa6/v138/i3/p034703_s1
    Webpage snapshot added, which is not expected, because in general preferences the corresponding field are untick
    3) http://pubs.acs.org/doi/abs/10.1021/ja310640b
    "Full Text PDF" is added and then dissapears

    This is the same as reported by Frisbee. The Debug ID is D1106735880.
  • olunet
    Until now, I've tested Zotero Standalone on Linux.
    Just updated translators and turned off addons.
    1) http://pubs.rsc.org/en/content/articlelanding/2013/ra/c3ra23239e
    "Full Text PDF" is added and then dissapears
    2) http://jcp.aip.org/resource/1/jcpsa6/v138/i3/p034703_s1
    Webpage snapshot added, which is still unexpected and strange
    3) http://pubs.acs.org/doi/abs/10.1021/ja310640b
    PDF attached!!!

    The Debug ID is D1314614383.
  • olunet
    to adamsmith, points 5 and 6 checked -- all fine: http://www.zotero.org/support/troubleshooting_translator_issues
  • olunet
    Similar on Windows PC: Zotero Standalone latest, almost clear installation, no plugins, translators up to date:
    1) http://pubs.rsc.org/en/content/articlelanding/2013/ra/c3ra23239e
    "Full Text PDF" is added and then dissapears
    2) http://jcp.aip.org/resource/1/jcpsa6/v138/i3/p034703_s1
    Webpage snapshot added, but should not be
    3) http://pubs.acs.org/doi/abs/10.1021/ja310640b
    "Full Text PDF" is added and then dissapears

    The Debug ID is D283072362.
  • adamsmith
    double check RSC with add-ons disabled in Chrome - that should really work.
    AIP PDFs won't attach with connectors, so that's as expected.
  • dstillman
    From Standalone:

    (5)(+0000063): CookieSandbox: Managing cookies for
    http://pubs.rsc.org/en/content/articlepdf/2013/ra/c3ra23239e

    (5)(+0000000): CookieSandbox: Added cookies for request to
    http://pubs.rsc.org/en/content/articlepdf/2013/ra/c3ra23239e

    (5)(+0000179): CookieSandbox: Managing cookies for
    http://pubs.rsc.org/en/content/articlepdf/2013/ra/c3ra23239e

    (5)(+0000000): CookieSandbox: Slurped cookies from
    http://pubs.rsc.org/en/content/articlepdf/2013/ra/c3ra23239e

    (5)(+0000002): CookieSandbox: Managing cookies for
    http://pubs.rsc.org/en/content/articlelanding/2013/ra/c3ra23239e

    (5)(+0000000): CookieSandbox: Added cookies for request to
    http://pubs.rsc.org/en/content/articlelanding/2013/ra/c3ra23239e

    (5)(+0000128): CookieSandbox: Managing cookies for
    http://pubs.rsc.org/en/content/articlelanding/2013/ra/c3ra23239e

    (2)(+0000206): Downloaded PDF did not have MIME type
    'application/pdf' in Attachments.importFromURL()

    (3)(+0000000):
    <!DOCTYPE html SYSTEM "http://pubs.rsc.org/Content/dtd/custom.dtd">;


    <html xmlns="http://www.w3.org/1999/xhtml"; xml:lang="

    (3)(+0000000): Deleting item 2475
  • dstillman
    edited March 2, 2013
    And same deal with ACS. A lot of whitespace, and then this:

    <!DOCTYPE HTML>
    <html>

    <script
    (We only include snippets in the debug output, so I can't tell you what the pages say, but clearly it's getting HTML pages back.)
  • dstillman
    Also, in that second debug output, it doesn't look like ACS actually worked—I see the same problem.
  • adamsmith
    I'm trying both RSC and ACS with Chrome (actually - Chromium) connector on Linux and getting PDF attachments, this is why I suspect something else is going on. You have no other Chrome extensions?
  • andrew.weisman
    I apologize for not following up with this, but thanks olunet for doing so!

    Any progress with this? I was using the Firefox version on Linux, but the problem continues to occur on Windows. I am indeed using a proxy: ezproxy.cul.columbia.edu.

    URLs for which it occurs (in Windows I get a red X for the PDF):

    http://apl.aip.org.ezproxy.cul.columbia.edu/resource/1/applab/v102/i8/p082109_s1
    http://apl.aip.org.ezproxy.cul.columbia.edu/resource/1/applab/v102/i8/p083305_s1
    http://jcp.aip.org.ezproxy.cul.columbia.edu/resource/1/jcpsa6/v138/i3/p034703_s1
    http://pubs.rsc.org.ezproxy.cul.columbia.edu/en/content/articlelanding/2013/ra/c3ra23239e
    http://pubs.acs.org.ezproxy.cul.columbia.edu/doi/abs/10.1021/ja310640b

    Working URLs:

    http://pra.aps.org.ezproxy.cul.columbia.edu/abstract/PRA/v87/i2/e020302
    http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2738645/
    http://pubs.acs.org.ezproxy.cul.columbia.edu/doi/abs/10.1021/jp310110r

    Note that the last links in each section above are nearly identical but only one works!

    Through my proxy I certainly have permissions to get the PDF, and clicking on the PDF link (e.g., PDF w/links) definitely always brings up the PDF in Firefox.

    Thanks to everybody,
    Andrew
  • adamsmith
    do you happen to use privacy mode or "Do note Remember History" in Firefox? There are problems using that with Zotero & a proxy currently.

    If not, we'd like a debug ID for one of the failing downloads from you, too.
    http://www.zotero.org/support/debug_output
  • gianlucabertaina
    edited July 2, 2013
    Hi,
    I have the same problem with APS online journals with the Zotero plugin on Ubuntu. I have access to the journals through my University.
    For example, the pdf link in:
    http://prl.aps.org/abstract/PRL/v110/i16/e165302
    Maybe the problem relates in the check that is performed by APS before accessing the pdf (one has to click on a picture of A. Einstein).
    Actually for me a workaround is to select the pdf link in the browser and click the image to get access to the pdf, then going back to the initial page and Saving to Zotero as usual. For some minutes Zotero has no problems in importing APS pdfs, because the "Einstein check" is not performed anymore.

    Hope this helps,
    Gianluca
  • adamsmith
    we won't be able to fix that. If a site requires clicking something to show that you're human or to agree to terms of service before getting a PDF, Zotero won't get the PDF (it deletes whatever it downloads because it is not, in fact, a PDF).
Sign In or Register to comment.