This is an old revision of the document!


PDF Full-Text Indexing

Zotero uses the pdfinfo and pdftotext utilities from the open-source Xpdf project for PDF indexing.

If installation via the Search tab of the Zotero preferences window fails, the easiest solution is to install the tools from another computer running the same OS and copy the pdfinfo* and pdftotext* files over from the Zotero data directory on that computer. (On macOS and Linux, be sure to preserve the executable bit on the executables and script files.)

Manual Installation

If you wish to build the Xpdf tools manually or use preinstalled system versions (e.g., from the Poppler fork), you must create a number of files in your Zotero data directory:

  1. The different executables can be placed directly in the data directory or linked to from there. Either way, platform-specific files must be created, conforming to the format pdfinfo-{platform} and pdftotext-{platform}, where {platform} is “Win32”, “MacIntel”, “Linux-i686”, etc. To determine your current platform, open the Firefox Web Console and enter navigator.platform.replace(' ', '-'). The Windows executables require the .exe extension (e.g., pdftotext-Win32.exe).
  2. One or two scripts are required for proper functioning of the PDF tools within Zotero.
    1. On OS X and Linux, redirect.sh should be saved as pdfinfo.sh, with the executable bit set.
    2. On Windows systems, redirect.vbs should be saved as pdfinfo.vbs and hide.vbs should be saved as pdftotext.vbs. [Note: These scripts are not currently usable on Windows, and Zotero instead uses custom versions of the binaries. If you use different versions of the tools, PDF page counts will not be available, and console windows will appear briefly while the tools run.]
  3. Text files containing the installed version number should be created with filenames in the format {executable-name}.version (e.g., pdftotext-MacIntel.version).
pdf_fulltext_indexing.1512392506.txt.gz · Last modified: 2017/12/04 08:01 by bwiernik