[Sommaire] [Imprimer]

   1  <?php
   2  /***************************************************************
   3  *  Copyright notice
   4  *
   5  *  (c) 2001-2005 Kasper Skaarhoj (kasperYYYY@typo3.com)
   6  *  All rights reserved
   7  *
   8  *  This script is part of the TYPO3 project. The TYPO3 project is
   9  *  free software; you can redistribute it and/or modify
  10  *  it under the terms of the GNU General Public License as published by
  11  *  the Free Software Foundation; either version 2 of the License, or
  12  *  (at your option) any later version.
  13  *
  14  *  The GNU General Public License can be found at
  15  *  http://www.gnu.org/copyleft/gpl.html.
  16  *  A copy is found in the textfile GPL.txt and important notices to the license
  17  *  from the author is found in LICENSE.txt distributed with these scripts.
  18  *
  19  *
  20  *  This script is distributed in the hope that it will be useful,
  21  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  22  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  23  *  GNU General Public License for more details.
  24  *
  25  *  This copyright notice MUST APPEAR in all copies of the script!
  26  ***************************************************************/
  27  /**
  28   * Index search crawler hook example
  29   *
  30   * $Id: class.crawlerhook.php 1421 2006-04-10 09:27:15Z mundaun $
  31   *
  32   * @author    Kasper Skårhøj <kasperYYYY@typo3.com>
  33   */
  34  /**
  35   * [CLASS/FUNCTION INDEX of SCRIPT]
  36   *
  37   *
  38   *
  39   *   57: class tx_indexedsearch_crawlerhook
  40   *   64:     function initMessage()
  41   *   80:     function indexOperation($cfgRec,&$session_data,$params,&$pObj)
  42   *
  43   * TOTAL FUNCTIONS: 2
  44   * (This index is automatically created/updated by the extension "extdeveval")
  45   *
  46   */
  47  
  48  
  49  
  50  /**
  51   * Index search crawler hook example
  52   *
  53   * @package TYPO3
  54   * @subpackage tx_indexedsearch
  55   * @author    Kasper Skårhøj <kasperYYYY@typo3.com>
  56   */
  57  class tx_indexedsearch_crawlerhook {
  58  
  59      /**
  60       * Function is called when an indexing session starts according to the time intervals set for the indexing configuration.
  61       *
  62       * @return    string        Return a text string for the first, initiating queue entry for the crawler.
  63       */
  64  	function initMessage()    {
  65          return 'Start of Custom Example Indexing session!';
  66      }
  67  
  68      /**
  69       * This will do two things:
  70       * 1) Carry out actual indexing of content (one or more items)
  71       * 2) Add one or more new entries into the crawlers queue so we are called again (another instance) for further indexing in the session (optional of course, if all indexing is done, we add no new entries)
  72       *
  73       * @param    array        Indexing Configuration Record (the record which holds the information that lead to this indexing session...)
  74       * @param    array        Session data variable. Passed by reference. Changed content is saved and passed back upon next instance in the session.
  75       * @param    array        Params array from the queue entry.
  76       * @param    object        Grant Parent Object (from "crawler" extension)
  77       * @param    object        Parent Object (from "indexed_search" extension)
  78       * @return    void
  79       */
  80  	function indexOperation($cfgRec,&$session_data,$params,&$pObj)    {
  81  
  82              // Init session data array if not already:
  83          if (!is_array($session_data))    {
  84              $session_data = array(
  85                  'step' => 0
  86              );
  87          }
  88  
  89              // Increase step counter (this is just an example of how the session data can be used - to track how many instances of indexing is left)
  90          $session_data['step']++;
  91  
  92  
  93          switch((int)$session_data['step'])    {
  94               case 1:    // Indexing Example: Content accessed with GET parameters added to URL:
  95  
  96                      // Load indexer if not yet [DON'T CHANGE]:
  97                  $pObj->loadIndexerClass();
  98  
  99                      // Get rootline from the Indexing Record (needed because the indexer relates all search results to a position in the page tree!) [DON'T CHANGE]:
 100                  $rl = $pObj->getUidRootLineForClosestTemplate($cfgRec['pid']);
 101  
 102                      // Set up language uid, if any:
 103                  $sys_language_uid = 0;
 104  
 105                      // Set up 2 example items to index:
 106                  $exampleItems = array(
 107                      array(
 108                          'ID' => '123',
 109                          'title' => 'Title of Example 1',
 110                          'content' => 'Vestibulum leo turpis, fringilla sit amet, semper eget, vestibulum ut, arcu. Vestibulum mauris orci, vulputate quis, congue eget, nonummy'
 111                      ),
 112                      array(
 113                          'ID' => 'example2',
 114                          'title' => 'Title of Example 2',
 115                          'content' => 'Cras tortor turpis, vulputate non, accumsan a, pretium in, magna. Cras turpis turpis, pretium pulvinar, pretium vel, nonummy eu.'
 116                      )
 117                  );
 118  
 119                      // For each item, index it (this is what you might like to do in batches of like 100 items if all your content spans thousands of items!)
 120                  foreach($exampleItems as $item)    {
 121  
 122                          // Prepare the GET variables array that must be added to the page URL in order to view result:
 123                      parse_str('&itemID='.rawurlencode($item['ID']), $GETparams);
 124  
 125                          // Prepare indexer (make instance, initialize it, set special features for indexing parameterized content - probably none of this should be changed by you) [DON'T CHANGE]:
 126                      $indexerObj = &t3lib_div::makeInstance('tx_indexedsearch_indexer');
 127                      $indexerObj->backend_initIndexer($cfgRec['pid'], 0, $sys_language_uid, '', $rl, $GETparams, FALSE);
 128                      $indexerObj->backend_setFreeIndexUid($cfgRec['uid'], $cfgRec['set_id']);
 129                      $indexerObj->forceIndexing = TRUE;
 130  
 131                          // Indexing the content of the item (see tx_indexedsearch_indexer::backend_indexAsTYPO3Page() for options)
 132                      $indexerObj->backend_indexAsTYPO3Page(
 133                          $item['title'],
 134                          '',
 135                          '',
 136                          $item['content'],
 137                          $GLOBALS['LANG']->charSet,    // Charset of content - MUST be set.
 138                          $item['tstamp'],            // Last-modified date
 139                          $item['create_date'],        // Created date
 140                          $item['ID']
 141                      );
 142                  }
 143               break;
 144               case 2: // Indexing Example: Content accessed directly in file system:
 145  
 146                      // Load indexer if not yet [DON'T CHANGE]:
 147                  $pObj->loadIndexerClass();
 148  
 149                      // Get rootline from the Indexing Record (needed because the indexer relates all search results to a position in the page tree!) [DON'T CHANGE]:
 150                  $rl = $pObj->getUidRootLineForClosestTemplate($cfgRec['pid']);
 151  
 152                      // Set up language uid, if any:
 153                  $sys_language_uid = 0;
 154  
 155                      // Prepare indexer (make instance, initialize it, set special features for indexing parameterized content - probably none of this should be changed by you) [DON'T CHANGE]:
 156                  $indexerObj = &t3lib_div::makeInstance('tx_indexedsearch_indexer');
 157                  $indexerObj->backend_initIndexer($cfgRec['pid'], 0, $sys_language_uid, '', $rl);
 158                  $indexerObj->backend_setFreeIndexUid($cfgRec['uid'], $cfgRec['set_id']);
 159                  $indexerObj->hash['phash'] = -1;    // To avoid phash_t3 being written to file sections (otherwise they are removed when page is reindexed!!!)
 160  
 161                      // Index document:
 162                  $indexerObj->indexRegularDocument('fileadmin/templates/index.html', TRUE);
 163               break;
 164               case 3: // Indexing Example: Content accessed on External URLs:
 165  
 166                      // Load indexer if not yet.
 167                  $pObj->loadIndexerClass();
 168  
 169                      // Index external URL:
 170                  $indexerObj = &t3lib_div::makeInstance('tx_indexedsearch_indexer');
 171                  $indexerObj->backend_initIndexer($cfgRec['pid'], 0, $sys_language_uid, '', $rl);
 172                  $indexerObj->backend_setFreeIndexUid($cfgRec['uid'], $cfgRec['set_id']);
 173                  $indexerObj->hash['phash'] = -1;    // To avoid phash_t3 being written to file sections (otherwise they are removed when page is reindexed!!!)
 174  
 175                      // Index external URL (HTML only):
 176                  $indexerObj->indexExternalUrl('http://www.google.com/');
 177               break;
 178          }
 179  
 180              // Finally, set entry for next indexing instance (if all steps are not completed)
 181          if ($session_data['step']<=3)    {
 182              $title = 'Step #'.$session_data['step'].' of 3';    // Just information field. Never mind that the field is called "url" - this is what will be shown in the "crawler" log. Could be a URL - or whatever else tells what that indexing instance will do.
 183              $pObj->addQueueEntryForHook($cfgRec, $title);
 184          }
 185      }
 186  }
 187  
 188  
 189  if (defined('TYPO3_MODE') && $TYPO3_CONF_VARS[TYPO3_MODE]['XCLASS']['ext/indexed_search/example/class.crawlerhook.php'])    {
 190      include_once($TYPO3_CONF_VARS[TYPO3_MODE]['XCLASS']['ext/indexed_search/example/class.crawlerhook.php']);
 191  }
 192  ?>