Example/CrawlerHook.php

Go to the documentation of this file.

<?php
namespace TYPO3\CMS\IndexedSearch\Example;

/*
 * This file is part of the TYPO3 CMS project.
 *
 * It is free software; you can redistribute it and/or modify it under
 * the terms of the GNU General Public License, either version 2
 * of the License, or any later version.
 *
 * For the full copyright and license information, please read the
 * LICENSE.txt file that was distributed with this source code.
 *
 * The TYPO3 project - inspiring people to share!
 */

use TYPO3\CMS\Core\Utility\GeneralUtility;
use TYPO3\CMS\IndexedSearch\Indexer;

class CrawlerHook
{
    public function initMessage()
    {
        return 'Start of Custom Example Indexing session!';
    }

    public function indexOperation($cfgRec, &$session_data, $params, &$pObj)
    {
        // Set up language uid, if any:
        $sys_language_uid = 0;

        // Init session data array if not already:
        if (!is_array($session_data)) {
            $session_data = array(
                'step' => 0
            );
        }
        // 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)
        $session_data['step']++;
        switch ((int)$session_data['step']) {
            case 1:
                // Indexing Example: Content accessed with GET parameters added to URL:
                // Get rootline from the Indexing Record (needed because the indexer relates all search results to a position in the page tree!) [DON'T CHANGE]:
                $rl = $pObj->getUidRootLineForClosestTemplate($cfgRec['pid']);
                // Set up 2 example items to index:
                $exampleItems = array(
                    array(
                        'ID' => '123',
                        'title' => 'Title of Example 1',
                        'content' => 'Vestibulum leo turpis, fringilla sit amet, semper eget, vestibulum ut, arcu. Vestibulum mauris orci, vulputate quis, congue eget, nonummy'
                    ),
                    array(
                        'ID' => 'example2',
                        'title' => 'Title of Example 2',
                        'content' => 'Cras tortor turpis, vulputate non, accumsan a, pretium in, magna. Cras turpis turpis, pretium pulvinar, pretium vel, nonummy eu.'
                    )
                );
                // For each item, index it (this is what you might like to do in batches of like 100 items if all your content spans thousands of items!)
                foreach ($exampleItems as $item) {
                    // Prepare the GET variables array that must be added to the page URL in order to view result:
                    parse_str('&itemID=' . rawurlencode($item['ID']), $GETparams);
                    // 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]:
                    $indexerObj = GeneralUtility::makeInstance(Indexer::class);
                    $indexerObj->backend_initIndexer($cfgRec['pid'], 0, $sys_language_uid, '', $rl, $GETparams, false);
                    $indexerObj->backend_setFreeIndexUid($cfgRec['uid'], $cfgRec['set_id']);
                    $indexerObj->forceIndexing = true;
                    // Indexing the content of the item (see \TYPO3\CMS\IndexedSearch\Indexer::backend_indexAsTYPO3Page() for options)
                    $indexerObj->backend_indexAsTYPO3Page($item['title'], '', '', $item['content'], $GLOBALS['LANG']->charSet, $item['tstamp'], $item['create_date'], $item['ID']);
                }
                break;
            case 2:
                // Indexing Example: Content accessed directly in file system:
                // Get rootline from the Indexing Record (needed because the indexer relates all search results to a position in the page tree!) [DON'T CHANGE]:
                $rl = $pObj->getUidRootLineForClosestTemplate($cfgRec['pid']);
                // 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]:
                $indexerObj = GeneralUtility::makeInstance(Indexer::class);
                $indexerObj->backend_initIndexer($cfgRec['pid'], 0, $sys_language_uid, '', $rl);
                $indexerObj->backend_setFreeIndexUid($cfgRec['uid'], $cfgRec['set_id']);
                $indexerObj->hash['phash'] = -1;
                // To avoid phash_t3 being written to file sections (otherwise they are removed when page is reindexed!!!)
                // Index document:
                $indexerObj->indexRegularDocument('fileadmin/templates/index.html', true);
                break;
            case 3:
                // Indexing Example: Content accessed on External URLs:
                // Index external URL:
                $indexerObj = GeneralUtility::makeInstance(Indexer::class);
                $indexerObj->backend_initIndexer($cfgRec['pid'], 0, $sys_language_uid, '', null);
                $indexerObj->backend_setFreeIndexUid($cfgRec['uid'], $cfgRec['set_id']);
                $indexerObj->hash['phash'] = -1;
                // To avoid phash_t3 being written to file sections (otherwise they are removed when page is reindexed!!!)
                // Index external URL (HTML only):
                $indexerObj->indexExternalUrl('http://www.google.com/');
                break;
        }
        // Finally, set entry for next indexing instance (if all steps are not completed)
        if ($session_data['step'] <= 3) {
            $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.
            $pObj->addQueueEntryForHook($cfgRec, $title);
        }
    }
}