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 * @contributor Tatsuva Shirai 20090530
  10 * @date 2008/03/31
  11 * @license http://www.gnu.org/copyleft/gpl.html GNU Public License
  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 * @uses $CFG
  21 */
  22 function get_text_for_indexing_doc(&$resource, $directfile = ''){
  23     global $CFG;
  24
  25     // SECURITY : do not allow non admin execute anything on system !!
  26     if (!has_capability('moodle/site:doanything', get_context_instance(CONTEXT_SYSTEM))) return;
  27
  28     $moodleroot = (@$CFG->block_search_usemoodleroot) ? "{$CFG->dirroot}/" : '' ;
  29
  30     // just call antiword over stdout and capture the output
  31     if (!empty($CFG->block_search_word_to_text_cmd)){
  32         // we need to remove any line command options...
  33         preg_match("/^\S+/", $CFG->block_search_word_to_text_cmd, $matches);
  34         if (!file_exists("{$moodleroot}{$matches[0]}")){
  35             mtrace('Error with MSWord to text converter command : executable not found at '.$moodleroot.$CFG->block_search_word_to_text_cmd);
  36         } else {
  37             if ($directfile == ''){
  38                 $file = escapeshellarg("{$CFG->dataroot}/{$resource->course}/{$resource->reference}");
  39             } else {
  40                 $file = escapeshellarg("{$CFG->dataroot}/{$directfile}");
  41             }
  42             $command = trim($CFG->block_search_word_to_text_cmd);
  43             $text_converter_cmd = "{$moodleroot}{$command} -m UTF-8.txt $file";
  44             if ($CFG->block_search_word_to_text_env){
  45                 putenv($CFG->block_search_word_to_text_env);
  46             }
  47             mtrace("Executing : $text_converter_cmd");
  48             $result = shell_exec($text_converter_cmd);
  49             if ($result){
  50                 return mb_convert_encoding($result, 'UTF-8', 'auto');
  51             } else {
  52                 mtrace('Error with MSWord to text converter command : execution failed. ');
  53                 return '';
  54             }
  55         }
  56     } else {
  57         mtrace('Error with MSWord to text converter command : command not set up. Execute once search block configuration.');
  58         return '';
  59     }
  60 }
  61 ?>