pdf full-text indexing

I've installed Xpdf, and created the symlink to pdftotext-Macintel in my Zotero data directory.

Now what? ie, how can I tell if its working or not?
  • Aside from adding a PDF and searching for some text in it, you can also try starting Firefox via the command line with the Zotero debug pref enabled and looking at the first few lines of the debug output.

    The next version of Zotero will add an interface for fulltext indexing to make installation and usage must easier and clearer.
  • Starting from the command line gives me the following:

    loaded md5.js
    *** CLB *** Initializing Google Browser Sync...
    *** CLB *** Instanciating core objects...
    *** CLB *** Registering with XPCOM...
    *** CLB *** Adding categories...
    *** CLB *** Google Browser Sync initialized succesfully!

    No sign of anything Zotero-, or Xpdf-related. Hm.
  • Did you enable the Zotero debug pref?
  • I did this time...

    zotero(3): pdftotext registered at /Users/sth/Library/Application Support/Firefox/Profiles/kpyuc3gw.default/zotero/pdftotext-MacIntel

    I take it that's a Good Thing?
  • Yup—should be working. Note that in the current version there's no way to reindex existing documents, but the next version will allow you to do so.
  • Yep, still, it's nice to have. Thanks very much.
  • While on the topic of pdftotext, there's an odd bit of behaviour with zotero on Linux for me.

    If I symlink to the pdftotext binary (in my /usr/bin), it doesn't work, and gets the error: "zotero(3): pdftotext-Linux-i686 symlink target not found -- PDF indexing disabled"

    However if I copy the binary directly to my zotero data directory (and rename it to pdftotext-Linux-i686), the error goes away, and indexing happens. Strange.
  • If I symlink to the pdftotext binary (in my /usr/bin), it doesn't work, and gets the error: "zotero(3): pdftotext-Linux-i686 symlink target not found -- PDF indexing disabled"

    However if I copy the binary directly to my zotero data directory (and rename it to pdftotext-Linux-i686), the error goes away, and indexing happens.
    Fixed on the dev branch.
  • The FAQ says:
    It does not currently search PDFs or Word documents.
    But it does search PDFs if you index them. The FAQ should say that.
  • Thanks Mark, I have updated that text in the FAQ
Sign In or Register to comment.