search/documents/physical_pdf.php

   1 <?php
   2 /**
   3 * Global Search Engine for Moodle
   4 * add-on 1.8+ : Valery Fremaux [valery.fremaux@club-internet.fr]
   5 * 2007/08/02
   6 *
   7 * this is a format handler for getting text out of a proprietary binary format
   8 * so it can be indexed by Lucene search engine
   9 */
  10
  11 function get_text_for_indexing_pdf(&$resource){
  12     global $CFG, $USER;
  13
  14     // SECURITY : do not allow non admin execute anything on system !!
  15     if (!isadmin($USER->id)) return;
  16
  17     // just call pdftotext over stdout and capture the output
  18     if (!empty($CFG->block_search_pdf_to_text_cmd)){
  19         preg_match("/^\S+/", $CFG->block_search_pdf_to_text_cmd, $matches);
  20         if (!file_exists("{$CFG->dirroot}/{$matches[0]}")){
  21             mtrace('Error with pdf to text converter command : exectuable not found.');
  22         }
  23         else{
  24             $file = $CFG->dataroot.'/'.$resource->course.'/'.$resource->reference;
  25             $text_converter_cmd = "{$CFG->dirroot}/{$CFG->block_search_pdf_to_text_cmd} $file -";
  26             $result = shell_exec($text_converter_cmd);
  27             if ($result){
  28                 return $result;
  29             }
  30             else{
  31                 mtrace('Error with pdf to text converter command : execution failed.');
  32                 return '';
  33             }
  34         }
  35     }
  36     else {
  37         mtrace('Error with pdf to text converter command : command not set up. Execute once search block configuration.');
  38         return '';
  39     }
  40 }
  41 ?>