Xpdf Integration

Warning: This Extension requires the use of exec() and also requires you to install Xpdf command line tools yourself.

SearchWP offers the unique feature of extracting plain text from PDF files uploaded to your WordPress website. Out of the box, SearchWP attempts to do this using only PHP, but due to the complexity and variation of the PDF format that sometimes results in content not being accurately extracted. Enter Xpdf.

Xpdf has a set of command line tools that must be installed on your server in order for this Extension to work. Instructions are included here.

Using the Xpdf Integration Extension you can offload all the work PHP has to do in processing your PDF files to Xpdf’s command line tools, which are extremely fast and accurate when extracting content from your PDFs. After activating the Extension, you will need to follow the installation instructions. Once installed, SearchWP will offload the PDF content extraction process to Xpdf.

Installing Xpdf command line tools

Using this extension you can utilize Xpdf to extract the content from your PDFs.

IMPORTANT: Xpdf command line tools are not provided in this Extension download. You must follow these instructions to download the command line tools and upload them to a non-public (outside your Web root) location.

You can download the Xpdf command line tools for both Windows and Linux at http://www.xpdfreader.com/download.html.

Installation

Once you have downloaded the command line tools for your server type:

Extract xpdf-tools-linux-4.03.tar.gz (the version number may be different)
Upload the pdftotext binary (found in either the bin32 or bin64 directory after extracting, depending on your server architecture) to a non-public location, outside your Web root
Upload the pdfinfo binary (found in either the bin32 or bin64 directory after extracting, depending on your server architecture) to a non-public location, outside your Web root
Ensure that both pdftotext and pdfinfo have execute permissions for the PHP user on your server

The last step is to tell SearchWP Xpdf Integration where you installed pdftotext and pdfinfo. To do this:

Add the following to your SearchWP Customizations plugin, replacing /path/to/pdftotext with the actual path to the pdftotext and pdfinfo binaries (not the folder) on your server.

// Tell SearchWP the location of the pdftotext binary.
add_filter( 'searchwp_xpdf_path', function() {
    return '/home/johndoe/pdftotext'; // Full absolute path to the binary NOT A FOLDER, NOT A URL.
} );

// Tell SearchWP the location of the pdfinfo binary.
add_filter( 'searchwp_pdfinfo_path', function() {
    return '/home/johndoe/pdfinfo'; // Full absolute path to the binary NOT A FOLDER, NOT A URL.
} );

That’s it!

Adding PDF password support in Xpdf Integration

Xpdf does support parsing of password protected (read: not encrypted) PDFs using the searchwp_xpdf_command filter. This filter allows you to directly manipulate the command being executed to fire Xpdf, and since Xpdf supports an option to include a password, you can go ahead and do that like so:

	<?php

	function my_searchwp_xpdf_command( $cmd, $filename ) {
	return $cmd . ' -upw password';
	}

	add_filter( 'searchwp_xpdf_command', 'my_searchwp_xpdf_command', 10, 2 );

view raw searchwp-customizations.php hosted with ❤ by GitHub

Manually Testing Xpdf Integration

After uploading and activating the Xpdf Integration Extension and defining your path to pdftotext, you can manually confirm that Xpdf text extraction is working as expected on specific PDFs uploaded to your Media library. Begin by going to the SearchWP Settings screen (Settings > SearchWP) and find the Xpdf Integration link in the Extensions on the SearchWP settings screen.

On the Xpdf Integration Testing screen, you can enter in the ID of the PDF you’d like to test:

Screen Shot 2013-12-09 at 11.34.24 AM

The ID can be found by navigating to your Media section and then clicking the Edit link of your PDF, the ID will be in the URL, followed by post=

After submitting a valid ID you will be given a detailed log of the steps taken by the Xpdf Integration Extension as well as any failure points that may have occurred. You’re also shown the exact content Xpdf extracted from the PDF:

Screen Shot 2013-12-09 at 11.36.16 AM

If the log displays a point of failure, please include that in any support requests you submit.