retrieve pdf metadata -- no progress [Err 1875800455]

conanm
Hi I love zotero and now trying to use it as my one reference manager tool. As such, I'm dealing with a lot of "naked" pdfs and trying to retrieve their metadata for good indexing. However, the function isn't working & the problem isn't exactly the same as ones I've seen in the forums. Basically, I can right click and get the "retrieve metadata" function, but when I run it, the Progress window appears with the name of the pdf in the window but the progress bar never makes any progress (stuck on "Retrieving Metadata...")

Using Zotero 1.5 Beta, with pdftotext version 3.02, pdfinfo 3.02. All the files I've tried are indexed, including some of the files in other programs known to work. The error # I reported is 1875800455, the error window reads:

[JavaScript Error: "t.replace is not a function" {file: "chrome://browser/content/browser.js" line: 2412}]

If I could actually get this to work (even for the % that are in Google Scholar), it would be awesome...
tx!
  • rhockey
    Same here get error :
    [JavaScript Error: "uncaught exception: Zotero.Translate only takes one parameter"]

    R
  • dstillman
    conanm's issue was different, but it looks like Retrieve PDF Metadata is broken in 2.0b3, unfortunately. It's now fixed on the trunk. We'll likely get out a new 2.0 build with a bunch of fixes, including this one, within the next few days. Thanks.
  • conanm
    Downoladed and installed 2.0b4 but still same problem. In addition to the error message above there's another one in the recent history:

    [JavaScript Error: "this._items[i].getField is not a function" {file: "chrome://zotero/content/recognizePDF.js" line: 124}]

    Not sure if that's related or not.
  • CHRYSOR_D
    Just the same problem for me since b3 ; Retrieve Metadata rolls on forever and then fails but without any JavaScript Error message.
    In 2.0 b4 I do not even get the 'Retrieve Metadata' option in the pop-up menu - seems to have disappeared.
    I checked in the Preferences and the pdftotext/pdfinfo files are installed and up to date though.
  • tomeames
    I'm having the same problems with PDF metadata retrieval. b3 worked fine on my home and work computers. After updating to b4 on my work computer I noticed that the PDF metadata option disppeared from the right click menu, but only on newly added PDFs. It was still there on PDFs I had added prior to b4.
    The same thing just happened on my home computer after upgrade to b4.

    Suggests there may be a problem with way PDFs are added to the library?
  • dstillman
    Suggests there may be a problem with way PDFs are added to the library?
    http://forums.zotero.org/discussion/7116/zotero-doesnt-recognise-pdf-as-pdf/#Item_8
  • tomeames
    Thanks for the link.
    Note to self --> do a search before pointing out the same thing again.

    Glad to know it's tagged for the next build.
    Congrats on creating an excellent piece of software BTW.
  • conanm
    Just to be clear, the problem I'm experiencing isn't apparently all that similar to what's being described above: I can access the "Retrieve metadata for PDF" function. It's just that the window that pops-up never shows any progress.

    Same story after upgrading to 2b05. Not sure if this error message is more relevant or not:
    2058382737
  • mcswell
    I have Z v2.0b6.4, and "retrieve metadata" works on some PDFs I have, but it hangs "Retrieving Metadata..." on others. An example of a PDF where it hangs is James Blevins "Word-based Morphology", which can be found through Google Scholar (it's the first hit when you search for "blevins word-based morphology). Here's the page GS points to:
    http://journals.cambridge.org/action/displayAbstract?fromPage=online&aid=521468
    The PDF appears normal (it's not an image). AFAICT, everything necessary for extracting the metadata is at the GS-linked web page (the above link) so I'm not sure why Z's metadata retrieval hangs.
  • dstillman
    mcswell: That PDF works fine for me, so the problem is somewhere else (assuming you were trying with the version available from journals.cambridge.org). Does that PDF reliably fail for you, or did that one just happen to hang when you tried it along with a bunch of others?

    Also, if you get a hang, send in an error report using Report Errors under the gear menu and post the Report ID here.
  • mcswell
    Dan Stillman wrote:
    > That PDF works fine for me, so the problem is somewhere else
    > (assuming you were trying with the version available from
    > journals.cambridge.org).

    I'm not sure where I got the PDF from; I suspect it was there, but I'm not sure.

    What version of Z are you using? I have 2.0b6.4. I just tried the Firefox update for plugins, and it says there are no updates for Z.

    > Also, if you get a hang, send in an error report using Report
    > Errors under the gear menu and post the Report ID here.

    OK, I did that for the Blevins article; there report ID is 1508490536; do I need to report this in a separate thread, or is this good enough? Something about an uncaught Java exception. There doesn't appear to be a way to attach the offending PDF file, although maybe it doesn't matter.

    I have around a dozen PDFs thus far for which Z hangs.
  • dstillman
    Need an answer to this: Does that PDF reliably fail for you, or did that one just happen to hang when you tried it along with a bunch of others?

    If the former, you can send it to support@zot....org, and we'll take a look.
  • mcswell
    Sorry, I missed that. Yes, it hangs when I try to retrieve the metadata for that PDF by itself. I just emailed the PDF with the error # (1508490536) to support. Thanks!
  • dstillman
    OK, with that version of the file I can reproduce the hang. We'll look into it. Thanks.
  • dstillman
    Most of the hangs on PDFs should now be fixed via an updated translator. Your copy of Zotero should auto-update within 24 hours, or you can update manually by clicking Update Now in the General pane of the Zotero prefs.

    If you're still having trouble, post a Report ID from after the hang.

    mcswell: The version of the paper you're using isn't actually the version on Google Scholar (which is a revised, published version from two years later). Currently, at least, Zotero looks only for exact text matches, and so it doesn't find results for that PDF even though a non-quoted search on the same text returns that paper as the first match. While we might be able to refine it so that it could find near matches, doing so wouldn't necessarily be a good idea, since the metadata from Google Scholar wouldn't actually be correct.
  • mcswell
    Thanks, works for the old Blevins paper; I'll test it on the others where it hung.

    And thanks for pointing out that I had an old version of that paper! I wonder how many of the others that hung were old versions...
  • mcswell
    I re-ran all the ones that had hung it before. I got one hang, but couldn't reproduce it. Other than that, smooth sailing.

    I now have a list of PDFs for which Z couldn't find metadata for any of several reasons: no OCRed text, "could not read text from PDF" (not sure what the diff is), no matching papers found. I presume it's searching in Google Scholar (I don't have a JSTOR account, and yesterday I hit GS's limit several times). Is there a setting somewhere for where it looks (GS, JSTOR,...)?

    And for all these left-over PDFs, does Z keep a record of what the problem was? I'm guessing that in many of the "no match" cases, I have old versions of the papers, and I should download newer versions. But the only way I seem to have of finding what the cause of the "no metadata" is to re-run each failure individually. (The retrieval applet's window seems to go away soon after it's done searching.) If I do that, it's going to be laborious, and I suspect I will soon run into GS's query limit.
Sign In or Register to comment.