ZotFile - Advanced PDF management for Zotero

1242527293063
  • This is now mapped correctly and will be in the next release. But I am not how long it takes until I submit it to mozilla. You can already get it on github.
  • I downloaded the file "install.rdf (version 3.0.4)" from github, but I couldn't install it. Could you please instruct me or point to where there is instructions to install Zotfile from github?
    Many thanks,
    Cadu
  • PDFXChange. I've posted elsewhere/previously, but in PDFXChange, if you highlight something, then right-click, the application creates a Pop-Up Note. (Everything I work with is PDF text as image, not text as text, BTW). Is there any possibility of making ZotFile "see" and import this kind of "annotation"? It would help enormously. Just checking. Thx!
  • @mbruffey

    I think it is already doing that; in my case, when I open a pdf with pdf exchange and I create a pop-up note, zotfile automatically recognise and extract it.

    I am using windows.
  • Ubuntu 12.04LTS with MLZ here. Only extracts "Sticky Notes."
  • Works for me, too.
    Which version of Zotefile? Which pdfXchange?

    In general, though, it's pretty much they work or they don't. If pdf.js extracts them they get extracted, if not then not.
  • Zotfile 3.0.3; PDFXchange 2.5 Build 212.
  • how odd. Works for me with 2.5 build 207 on Ubuntu. I have no idea what could make the difference.
  • It is interesting. I even booted to windows, installed the latest MLZ and PDFXchange, opened and edited a file, saved it, and still does not extract "pop-up" notes. I could send file (small). m
  • yeah, I'm curious, could you send me an annotated file?
    e-mail at the bottom of this post: http://www.zotero.org/blog/summer-zotero-workshops/
  • For future references, we figured this out:
    pop-up notes attached to text do extract, pop-up notes extracted to shapes (lines, rectangles etc.) don't.
    Depending on the document, using pdfXchange's free (even in the free version) OCR feature works very well and then allows you to insert pop-up notes that Zotfile extracts.
  • I have a very simple suggestion, which isn't a bug: Zotfile should be able to add a recently downloaded PDF (in the watched folder) as its own item rather than as an attachment. Of course, it's possible to do this manually by dragging the PDF into Zotero, just as it's possible to drag the PDF to an existing item to attach it, but the excellent ZotFile mechanism is just faster and easier. (If this already exists, please someone let me know how.)

    Thanks very much for the fantastic add-on to Zotero! ZotFile makes my life better.
  • The current development version on github implements that at least for the 'watch folder' stuff. So when you see the pop-up 'New file found. click here to add to zotero', it even works when no item is selected and automatically tries to create an item based on metadata. But there are still some problems...
  • edited November 13, 2013
    I'm eager to try the send-to-tablet function, but it doesn't appear on the right-click menu. i've installed the 2 plugins and created and assigned a tablet sync folder, but the "manage attachments" menu only shows me rename and extract options, and a greyed out selected items option.

    windows 7, firefox 25.0, zotero 4.0.15, zotfile 3.0.3.

    any advice would be most appreciated!

    thanks,
    Andy

    -----
    Edit: I figured out the problem--the items i was trying to send are in a group library--it works fine when i send an item from a personal library.
  • do you have the option "Use Zotfile to send..." checked in the "Tablet Settings" tab of the Zotfile preferences?
  • Thanks Joscha! I'll eagerly await the stable version.

    I have one more minor comment: Currently, when I click the "add to Zotero" icon in the URL bar in Chrome, Zotfile pops up and informs me that the new entry is being added to the currently selected collection in Zotero. In my work flow, at least, this isn't great; I move back and forth between collections often and it's rare that the currently selected collection has anything to do with what I'm browsing online. Better for me would be to simply put the entry in "My Library" and let me add it manually to a collection. (This is the default behavior for Zotero without Zotfile, right?) Maybe an option could be included in the Zotfile preferences to keep it from being automatically added to the current collection?
  • Better for me would be to simply put the entry in "My Library" and let me add it manually to a collection. (This is the default behavior for Zotero without Zotfile, right?)
    no, what you're seeing is Zotero behavior, it has nothing to do with Zotfile. Since ~98% of all users greatly appreciate the way this works, you'll have to live with this, pretty sure devs won't add a preference for this.
  • An update on the issue of the hanging 'Extract Annotations': I think this may be a pdf.js issue, because I can't properly view highlighted pdfs in pdf.js (which I understand is a modified version from Zotfile).

    Here is an example PDF
    https://www.dropbox.com/s/61v0t7hm84vefs9/Zou_2006_The adaptive lasso and its oracle properties.pdf?m

    The extracts used to work fine on it. Now, the extraction message hangs indefinitely, and the highlighting won't display in pdf.js (however, the comments seem to show up just fine in pdf.js, but are not extracted). There is highlighting on page 2.
    There has been a ticket created (thanks to Aurimas):https://github.com/mozilla/pdf.js/issues/3885
  • Thanks for posting this. It's a good case. pdf.js based extraction hangs for me too. poppler works (below is the extracted text from poppler).

    Note that the modified version of pdf.js, which is included in zotfile, is not always up to date. Actually, it's updated pretty rarely because it's quit some work. But please post here when the issue is resolved in pdf.js so I know that the update might be worthwhile.

    Also, what do you mean with "used to work fine on it"? Do you know in which zotfile version? I only updated the underlying pdf.js version once and that was for verion 3 of zotfile, I think.





    Extracted Annotations (Thu Nov 21 09:16:54 2013)

    "oracle properties; namely, it performs as well as if the true underlying model were given in advance" (Einstein et al 1935:2)

    "the lasso shrinkage produces biased estimates for the large coefficients, and thus it could be suboptimal in terms of estimation risk" (Einstein et al 1935:2)

    "we can use ˆ (ols" (Einstein et al 1935:4)

    all this boils down to is using Weights on the penalized-coefficients B such that the weights are the inverse of the coefficients estimated during an initial OLS regression (note on p.4)
  • edited November 21, 2013
    Hi Joscha,
    Thanks for addressing the 'extraction' hanging message. Sorry, I don't know what zotfile or PDF.js version worked originally, all I know is that this issue started around July 2013. Annotations that I had extracted prior to that date can no longer be extracted. If I could and I knew how, I'd happily roll back my PDF.js to whatever version was released << July 2013. I would also settle for using poppler tools manually on Linux (too bad its only a mac-zotfile thing), but the outputs from my use of pdftotext are a bit daunting and don't seem to have highlights nor annotations. Are either option possible?
    Rob
  • edited November 22, 2013
    Just to be clear: Whether the extraction works or not only depends on your zotfile version, which comes packaged with a modified pdf.js version (the pdf.js version in firefox is irrelevant). So rolling back pdf.js doesn't make any sense. You can install old zotfile version from here: https://addons.mozilla.org/en-US/firefox/addon/zotfile/versions/
    I would do it in a test profile though and older version might not be compatible with the current zotero and/or the current FF. You can try version 2.3.5 but I would be surprised if it makes a difference. The poppler-based extraction isn't an option on linux.
  • (Joscha above should read:
    Whether the extraction works or not only depends on your zotfile version
    )
  • edited February 6, 2014
    Hi Joscha,
    I created a test Firefox profile and re-installed Zotfile to version 2.3.5. I was able to extract the annotations for the aforementioned test pdf at:

    https://www.dropbox.com/s/61v0t7hm84vefs9/Zou_2006_The adaptive lasso and its oracle properties.pdf?m

    ... so, that's an improvement! Overjoyed! The article still does not display properly in Firefox with pdf.js (the pages with highlights hang indefinitely), but that is minor compared to successfully being able to extract them.

    I've tested this on 6 old and new pdfs with annotations, and I have a success rate of 5/6. So the issue is still there for some pdfs, but with the most-uptodate Zotfile, my success rate was 0.

    Thanks for the help... any suggestions on moving forward?

    EDIT: Problem solved! I can now get both hightlights and comments, using Zotfile 3.1 updated 04/feb/2014 on Zotero 4.0.16 (both standalone and firefox addon). Once again this is an amazing addon! Thank you Joscha and the zotfile team.
  • That's good to hear but there isn't much you can do right now. It's something I am going to look at when I get back to the extraction of annotations but that might take a while. So don't expect a quick solution in the current version of zotfile. Thanks for reporting and feel free to post another pdf with the same problem just so that I have two cases when I look into it.
  • Thanks for a really fantastic tool!

    I use Zotfile on both my Windows and Mac computers and I use Dropbox to store my PDFs. However, because the Dropbox folder is in two different locations (e.g. /Users/... and C:/...), my Mac cannot find files that were added via Windows and vice versa.

    Is there a way to set the storage location for PDFs relative to Dropbox?
  • yes, that's built into Zotero: http://zoteromusings.wordpress.com/2013/04/11/whats-new-in-zotero-4-0-part-2/#relative-linked-attachment-filepaths
  • Fantastic, thanks -- that worked!
  • edited December 7, 2013
    Just wanted to thank Joscha for developing this plugin. It makes my work easier! Muchas gracias!
  • edited December 8, 2013
    Hi Joscha,
    I apologize if this question has already been raised, as I haven't read the entire thread to confirm that this hasn't been posted, but what is the suggested method for deleting the PDF files that have been "sent to tablet" from Zotero via Zotfiles? I really appreciate the forward synchronization of Zotfiles where I can add PDF files to my tablet directly from Zotero, but I haven't found a way to do the reverse, namely, to delete PDF files from the tablet and have the the change reflected in Zotero as well. It seems like I will need to get rid of the "_tablet" tag and delete the file manually if I wished to remove a file, which I find is rather cumbersome. Is there a more practical solution to this?
This discussion has been closed.