search/documents/physical_htm.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 $resource
  18 * @uses CFG, USER
  19 */
  20 function get_text_for_indexing_htm(&$resource){
  21     global $CFG, $USER;
  22
  23     // SECURITY : do not allow non admin execute anything on system !!
  24     if (!isadmin($USER->id)) return;
  25
  26     // just get text
  27     $text = implode('', file("{$CFG->dataroot}/{$resource->course}/{$resource->reference}"));
  28
  29     // extract keywords and other interesting meta information and put it back as real content for indexing
  30     if (preg_match('/(.*)<meta ([^>]*)>(.*)/is', $text, $matches)){
  31         $prefix = $matches[1];
  32         $meta_attributes = $matches[2];
  33         $suffix = $matches[3];
  34         if (preg_match('/name="(keywords|description)"/i', $meta_attributes)){
  35             preg_match('/content="([^"]+)"/i', $meta_attributes, $matches);
  36             $text = $prefix.' '.$matches[1].' '.$suffix;
  37         }
  38     }
  39     // brutally filters all html tags
  40     $text = preg_replace("/<[^>]*>/", '', $text);
  41     $text = preg_replace("/<!--[^>]*-->/", '', $text);
  42     $text = html_entity_decode($text, ENT_COMPAT, 'UTF-8');
  43     $text = mb_convert_encoding($text, 'UTF-8', 'AUTO');
  44
  45     /*
  46     * debug code for tracing input
  47     echo "<hr/>";
  48     $FILE = fopen("filetrace.log", 'w');
  49     fwrite($FILE, $text);
  50     fclose($FILE);
  51     echo "<hr/>";
  52     */
  53
  54     if (!empty($CFG->block_search_limit_index_body)){
  55         $text = shorten($text, $CFG->block_search_limit_index_body);
  56     }
  57     return $text;
  58 }
  59 ?>