search/documents/physical_doc.php

   1 <?php
   2 /**
   3 * Global Search Engine for Moodle
   4 *
   5 * @package search
   6 * @category core
   7 * @subpackage document_wrappers
   8 * @author Valery Fremaux [valery.fremaux@club-internet.fr] > 1.8
   9 * @date 2008/03/31
  10 * @license http://www.gnu.org/copyleft/gpl.html GNU Public License
  11 * @version revised for Moodle 2.0
  12 *
  13 * this is a format handler for getting text out of a proprietary binary format
  14 * so it can be indexed by Lucene search engine
  15 */
  16
  17 /**
  18 * MS Word extractor
  19 * @param object $resource
  20 * @param string $directfile if the resource is given as a direct file path, use it as reference to the file
  21 * @uses $CFG
  22 */
  23 function get_text_for_indexing_doc(&$resource, $directfile = ''){
  24     global $CFG;
  25
  26     // SECURITY : do not allow non admin execute anything on system !!
  27     if (!has_capability('moodle/site:config', get_context_instance(CONTEXT_SYSTEM))) return;
  28
  29     // adds moodle root switch if none was defined
  30     if (!isset($CFG->block_search_usemoodleroot)){
  31         set_config('block_search_usemoodleroot', 1);
  32     }
  33
  34     $moodleroot = ($CFG->block_search_usemoodleroot) ? "{$CFG->dirroot}/" : '' ;
  35
  36     // just call pdftotext over stdout and capture the output
  37     if (!empty($CFG->block_search_word_to_text_cmd)){
  38         if (!file_exists("{$moodleroot}{$CFG->block_search_word_to_text_cmd}")){
  39             mtrace('Error with MSWord to text converter command : executable not found at '.$moodleroot.$CFG->block_search_word_to_text_cmd);
  40         } else {
  41             if ($directfile == ''){
  42                 $file = escapeshellarg("{$CFG->dataroot}/{$resource->course}/{$resource->reference}");
  43             } else {
  44                 $file = escapeshellarg("{$CFG->dataroot}/{$directfile}");
  45             }
  46             $command = trim($CFG->block_search_word_to_text_cmd);
  47             $text_converter_cmd = "{$moodleroot}{$command} -m UTF-8.txt $file";
  48             if ($CFG->block_search_word_to_text_env){
  49                 putenv($CFG->block_search_word_to_text_env);
  50             }
  51             mtrace("Executing : $text_converter_cmd");
  52             $result = shell_exec($text_converter_cmd);
  53             if ($result){
  54                 return mb_convert_encoding($result, 'UTF-8', 'auto');
  55             } else {
  56                 mtrace('Error with MSWord to text converter command : execution failed. ');
  57                 return '';
  58             }
  59         }
  60     } else {
  61         mtrace('Error with MSWord to text converter command : command not set up. Execute once search block configuration.');
  62         return '';
  63     }
  64 }
  65 ?>