How to Use Perl to Search a PDF Doc

Perl's mascot is a camel.

Perl scripts "crawl" across the web, sifting through the ocean of HTML pages for information, and inevitably encounter some of the millions of Adobe Acrobat PDF files sprinkled across the Internet. Though they are ready to print and attractive, from the point of view of a web crawler, PDFs are a slightly harder nut to crack than simple, text-based HTML pages. Luckily, as with many web tasks, there is a Perl module that can help make life easier: CAM::PDF. This programming interface, though mostly intended to create and manipulate PDF files, has some utilities that enables scripts to search their content.

Things You'll Need

• Perl scripting environment
• CAM::PDF Perl module
• Text or code editor
• PDF file

Instructions

1. Install CAM::PDF. The cpan utility provides the easiest way to do this--start cpan at the command line and at the prompt, type "install CAM::PDF" (without quotes).

2. Open an editor and start the script, entering the following lines to start the Perl interpreter and import the necessary module:

   #!/usr/bin/perl
   use CAM::PDF;

   Add the next two lines to process the command line arguments that the user will pass in:

   my $file = shift;
   my $search = shift;

   The first argument passed to the script will be the name of a PDF file, and the second, the search string.

3. Create a new CAM::PDF object by adding the following line to the script:

   my $doc = CAM::PDF->new($file);

   Using the imported module's numPages method to define the upper limit, create a loop to process each page of the document:

   foreach my $p ((1 .. $doc->numPages()))
   {

4. Within the loop, add this line to get each page of text from the PDF file:

   my $str = $doc->getPageText($p);

   Add the next script statement to split the page's text up into an array of separate lines:

   @lines = split(/\n/ , $str);

   Finish the loop statement by entering a closing bracket:

   }

5. Finally, add another loop to the script to process each line of the page and seek a match for the user's search string as a regular expression. If the regular expression returns a match, this example prints the line and page number to stdout. In place of these print statements, you should implement code to process the results as needed.

   my $i = 0;
   foreach $line (@lines)
   {
   ++$i;
   if($line =~ /$search/)
   {
   print "\"$search\" found in line $i of page $p\n";
   print "$line\n\n"
   }
   }

Tips & Warnings

• Because the text in PDF files is not hierarchically organized like that of HTML files, you will most likely need to design scripts to search certain types of PDF files (e.g. forms, bulletins, schedules). It may not be possible to write a robust Perl script that can effectively search any type of PDF file.

References

• The CPAN CAM::PDF website
• "Perl & LWP"; Sean M. Burke; 2002