search/documents/physical_xml.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 *
  12 * this is a format handler for getting text out of a proprietary binary format
  13 * so it can be indexed by Lucene search engine
  14 */
  15
  16 /**
  17 * @param object $resource
  18 * @uses $CFG
  19 */
  20 function get_text_for_indexing_xml(&$resource, $directfile = ''){
  21     global $CFG;
  22
  23     // SECURITY : do not allow non admin execute anything on system !!
  24     if (!has_capability('moodle/site:doanything', get_context_instance(CONTEXT_SYSTEM))) return;
  25
  26     // just get text
  27     if ($directfile == ''){
  28         $text = implode('', file("{$CFG->dataroot}/{$resource->course}/{$resource->reference}"));
  29     } else {
  30         $text = implode('', file("{$CFG->dataroot}/{$directfile}"));
  31     }
  32
  33     // filter out all xml tags
  34     $text = preg_replace("/<[^>]*>/", ' ', $text);
  35
  36     if (!empty($CFG->block_search_limit_index_body)){
  37         $text = shorten_text($text, $CFG->block_search_limit_index_body);
  38     }
  39     return $text;
  40 }
  41 ?>