IndexService::indexItem() - Code Metrics - Inspection of "[BUGFIX] Download of stopwords and synonyms not wo..." - timohund/ext-solr - Measure and Improve Code Quality continuously with Scrutinizer

Passed

Push — master ( a5d0a9...8581cd )

by Timo

created 2018-06-26 12:07 UTC

IndexService::indexItem() B

↳ Parent: IndexService

Complexity

Conditions	6
Paths	16

Size

Total Lines

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	15
CRAP Score	6.0585

Importance

Changes

Metric	Value
dl	0
loc	33
ccs	15
cts	17
cp	0.8824
rs	8.7697
c	0
b	0
f	0
cc	6
nc	16
nop	2
crap	6.0585

<?php

namespace ApacheSolrForTypo3\Solr\Domain\Index;

/***************************************************************
 *  Copyright notice
 *
 *  (c) 2015-2016 Timo Hund <[email protected]>
 *  All rights reserved
 *
 *  This script is part of the TYPO3 project. The TYPO3 project is
 *  free software; you can redistribute it and/or modify
 *  it under the terms of the GNU General Public License as published by
 *  the Free Software Foundation; either version 3 of the License, or
 *  (at your option) any later version.
 *
 *  The GNU General Public License can be found at
 *  http://www.gnu.org/copyleft/gpl.html.
 *
 *  This script is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *  GNU General Public License for more details.
 *
 *  This copyright notice MUST APPEAR in all copies of the script!
 ***************************************************************/

use ApacheSolrForTypo3\Solr\ConnectionManager;
use ApacheSolrForTypo3\Solr\IndexQueue\Indexer;
use ApacheSolrForTypo3\Solr\IndexQueue\Item;
use ApacheSolrForTypo3\Solr\IndexQueue\Queue;
use ApacheSolrForTypo3\Solr\Site;
use ApacheSolrForTypo3\Solr\System\Configuration\TypoScriptConfiguration;
use ApacheSolrForTypo3\Solr\System\Logging\SolrLogManager;
use ApacheSolrForTypo3\Solr\Task\IndexQueueWorkerTask;
use TYPO3\CMS\Backend\Utility\BackendUtility;
use TYPO3\CMS\Core\Utility\GeneralUtility;
use TYPO3\CMS\Extbase\SignalSlot\Dispatcher;

/**
 * Service to perform indexing operations
 *
 * @author Timo Hund <[email protected]>
 */
class IndexService
{
    /**
     * @var TypoScriptConfiguration
     */
    protected $configuration;

    /**
     * @var Site
     */
    protected $site;

    /**
     * @var IndexQueueWorkerTask
     */
    protected $contextTask;

    /**
     * @var Queue
     */
    protected $indexQueue;

    /**
     * @var Dispatcher
     */
    protected $signalSlotDispatcher;

    /**
     * @var \ApacheSolrForTypo3\Solr\System\Logging\SolrLogManager
     */
    protected $logger = null;

    /**
     * IndexService constructor.
     * @param Site $site
     * @param Queue|null $queue
     * @param Dispatcher|null $dispatcher
     * @param SolrLogManager|null $solrLogManager
     */
    public function __construct(Site $site, Queue $queue = null, Dispatcher $dispatcher = null, SolrLogManager $solrLogManager = null)
    {
        $this->site = $site;
        $this->indexQueue = $queue ?? GeneralUtility::makeInstance(Queue::class);
        $this->signalSlotDispatcher = $dispatcher ?? GeneralUtility::makeInstance(Dispatcher::class);
        $this->logger = $solrLogManager ?? GeneralUtility::makeInstance(SolrLogManager::class, /** @scrutinizer ignore-type */ __CLASS__);
        define('EXT_SOLR_INDEXING_CONTEXT', true);
    }

    /**
     * @param \ApacheSolrForTypo3\Solr\Task\IndexQueueWorkerTask $contextTask
     */
    public function setContextTask($contextTask)
    {
        $this->contextTask = $contextTask;
    }

    /**
     * @return \ApacheSolrForTypo3\Solr\Task\IndexQueueWorkerTask
     */
    public function getContextTask()
    {
        return $this->contextTask;
    }

    /**
     * Indexes items from the Index Queue.
     *
     * @param int $limit
     * @return bool
     */
    public function indexItems($limit)
    {
        $errors     = 0;
        $indexRunId = uniqid();
        $configurationToUse = $this->site->getSolrConfiguration();
        $enableCommitsSetting = $configurationToUse->getEnableCommits();

        // get items to index
        $itemsToIndex = $this->indexQueue->getItemsToIndex($this->site, $limit);

        $this->emitSignal('beforeIndexItems', [$itemsToIndex, $this->getContextTask(), $indexRunId]);

        foreach ($itemsToIndex as $itemToIndex) {
            try {
                // try indexing
                $this->emitSignal('beforeIndexItem', [$itemToIndex, $this->getContextTask(), $indexRunId]);
                $this->indexItem($itemToIndex, $configurationToUse);
                $this->emitSignal('afterIndexItem', [$itemToIndex, $this->getContextTask(), $indexRunId]);
            } catch (\Exception $e) {
                $errors++;
                $this->indexQueue->markItemAsFailed($itemToIndex, $e->getCode() . ': ' . $e->__toString());
                $this->generateIndexingErrorLog($itemToIndex, $e);
            }
        }

        $this->emitSignal('afterIndexItems', [$itemsToIndex, $this->getContextTask(), $indexRunId]);

        if ($enableCommitsSetting && count($itemsToIndex) > 0) {
            $solrServers = GeneralUtility::makeInstance(ConnectionManager::class)->getConnectionsBySite($this->site);
            foreach ($solrServers as $solrServer) {
                $solrServer->getWriteService()->commit(false, false, false);
            }
        }

        return ($errors === 0);
    }

    /**
     * Generates a message in the error log when an error occured.
     *
     * @param Item $itemToIndex
     * @param \Exception  $e
     */
    protected function generateIndexingErrorLog(Item $itemToIndex, \Exception $e)
    {
        $message = 'Failed indexing Index Queue item ' . $itemToIndex->getIndexQueueUid();
        $data = ['code' => $e->getCode(), 'message' => $e->getMessage(), 'trace' => $e->getTraceAsString(), 'item' => (array)$itemToIndex];

        $this->logger->log(
            SolrLogManager::ERROR,
            $message,
            $data
        );
    }

    /**
     * Builds an emits a singal for the IndexService.
     *
     * @param string $name
     * @param array $arguments
     * @return mixed
     */
    protected function emitSignal($name, $arguments)
    {
        return $this->signalSlotDispatcher->dispatch(__CLASS__, $name, $arguments);
    }

    /**
     * Indexes an item from the Index Queue.
     *
     * @param Item $item An index queue item to index
     * @param TypoScriptConfiguration $configuration
     * @return bool TRUE if the item was successfully indexed, FALSE otherwise
     */
    protected function indexItem(Item $item, TypoScriptConfiguration $configuration)
    {
        $indexer = $this->getIndexerByItem($item->getIndexingConfigurationName(), $configuration);

        // Remember original http host value
        $originalHttpHost = isset($_SERVER['HTTP_HOST']) ? $_SERVER['HTTP_HOST'] : null;

        $itemChangedDate = $item->getChanged();
        $itemChangedDateAfterIndex = 0;

        $this->initializeHttpServerEnvironment($item);
        $itemIndexed = $indexer->index($item);

        // update IQ item so that the IQ can determine what's been indexed already
        if ($itemIndexed) {
            $this->indexQueue->updateIndexTimeByItem($item);
            $itemChangedDateAfterIndex = $item->getChanged();
        }

        if ($itemChangedDateAfterIndex > $itemChangedDate && $itemChangedDateAfterIndex > time()) {
            $this->indexQueue->setForcedChangeTimeByItem($item, $itemChangedDateAfterIndex);
        }

        if (!is_null($originalHttpHost)) {
            $_SERVER['HTTP_HOST'] = $originalHttpHost;
        } else {
            unset($_SERVER['HTTP_HOST']);
        }

        // needed since TYPO3 7.5
        GeneralUtility::flushInternalRuntimeCaches();

        return $itemIndexed;
    }

    /**
     * A factory method to get an indexer depending on an item's configuration.
     *
     * By default all items are indexed using the default indexer
     * (ApacheSolrForTypo3\Solr\IndexQueue\Indexer) coming with EXT:solr. Pages by default are
     * configured to be indexed through a dedicated indexer
     * (ApacheSolrForTypo3\Solr\IndexQueue\PageIndexer). In all other cases a dedicated indexer
     * can be specified through TypoScript if needed.
     *
     * @param string $indexingConfigurationName Indexing configuration name.
     * @param TypoScriptConfiguration $configuration
     * @return Indexer
     */
    protected function getIndexerByItem($indexingConfigurationName, TypoScriptConfiguration $configuration)
    {
        $indexerClass = $configuration->getIndexQueueIndexerByConfigurationName($indexingConfigurationName);
        $indexerConfiguration = $configuration->getIndexQueueIndexerConfigurationByConfigurationName($indexingConfigurationName);

        $indexer = GeneralUtility::makeInstance($indexerClass, /** @scrutinizer ignore-type */ $indexerConfiguration);
        if (!($indexer instanceof Indexer)) {
            throw new \RuntimeException(
                'The indexer class "' . $indexerClass . '" for indexing configuration "' . $indexingConfigurationName . '" is not a valid indexer. Must be a subclass of ApacheSolrForTypo3\Solr\IndexQueue\Indexer.',
                1260463206
            );
        }

        return $indexer;
    }

    /**
     * Gets the indexing progress.
     *
     * @return float Indexing progress as a two decimal precision float. f.e. 44.87
     */
    public function getProgress()
    {
        return $this->indexQueue->getStatisticsBySite($this->site)->getSuccessPercentage();
    }

    /**
     * Returns the amount of failed queue items for the current site.
     *
     * @return int
     */
    public function getFailCount()
    {
        return $this->indexQueue->getStatisticsBySite($this->site)->getFailedCount();
    }

    /**
     * Initializes the $_SERVER['HTTP_HOST'] environment variable in CLI
     * environments dependent on the Index Queue item's root page.
     *
     * When the Index Queue Worker task is executed by a cron job there is no
     * HTTP_HOST since we are in a CLI environment. RealURL needs the host
     * information to generate a proper URL though. Using the Index Queue item's
     * root page information we can determine the correct host although being
     * in a CLI environment.
     *
     * @param Item $item Index Queue item to use to determine the host.
     * @param
     */
    protected function initializeHttpServerEnvironment(Item $item)
    {
        static $hosts = [];
        $rootpageId = $item->getRootPageUid();
        $hostFound = !empty($hosts[$rootpageId]);

        if (!$hostFound) {
            $rootline = BackendUtility::BEgetRootLine($rootpageId);
            $host = BackendUtility::firstDomainRecord($rootline);
            $hosts[$rootpageId] = $host;
        }

        $_SERVER['HTTP_HOST'] = $hosts[$rootpageId];

        // needed since TYPO3 7.5
        GeneralUtility::flushInternalRuntimeCaches();
    }
}


1		<?php
2
3		namespace ApacheSolrForTypo3\Solr\Domain\Index;
4
5		/***************************************************************
6		* Copyright notice
7		*
8		* (c) 2015-2016 Timo Hund <[email protected]>
9		* All rights reserved
10		*
11		* This script is part of the TYPO3 project. The TYPO3 project is
12		* free software; you can redistribute it and/or modify
13		* it under the terms of the GNU General Public License as published by
14		* the Free Software Foundation; either version 3 of the License, or
15		* (at your option) any later version.
16		*
17		* The GNU General Public License can be found at
18		* http://www.gnu.org/copyleft/gpl.html.
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		use ApacheSolrForTypo3\Solr\ConnectionManager;
29		use ApacheSolrForTypo3\Solr\IndexQueue\Indexer;
30		use ApacheSolrForTypo3\Solr\IndexQueue\Item;
31		use ApacheSolrForTypo3\Solr\IndexQueue\Queue;
32		use ApacheSolrForTypo3\Solr\Site;
33		use ApacheSolrForTypo3\Solr\System\Configuration\TypoScriptConfiguration;
34		use ApacheSolrForTypo3\Solr\System\Logging\SolrLogManager;
35		use ApacheSolrForTypo3\Solr\Task\IndexQueueWorkerTask;
36		use TYPO3\CMS\Backend\Utility\BackendUtility;
37		use TYPO3\CMS\Core\Utility\GeneralUtility;
38		use TYPO3\CMS\Extbase\SignalSlot\Dispatcher;
39
40		/**
41		* Service to perform indexing operations
42		*
43		* @author Timo Hund <[email protected]>
44		*/
45		class IndexService
46		{
47		/**
48		* @var TypoScriptConfiguration
49		*/
50		protected $configuration;
51
52		/**
53		* @var Site
54		*/
55		protected $site;
56
57		/**
58		* @var IndexQueueWorkerTask
59		*/
60		protected $contextTask;
61
62		/**
63		* @var Queue
64		*/
65		protected $indexQueue;
66
67		/**
68		* @var Dispatcher
69		*/
70		protected $signalSlotDispatcher;
71
72		/**
73		* @var \ApacheSolrForTypo3\Solr\System\Logging\SolrLogManager
74		*/
75		protected $logger = null;
76
77		/**
78		* IndexService constructor.
79		* @param Site $site
80		* @param Queue\|null $queue
81		* @param Dispatcher\|null $dispatcher
82		* @param SolrLogManager\|null $solrLogManager
83		*/
84	7	public function __construct(Site $site, Queue $queue = null, Dispatcher $dispatcher = null, SolrLogManager $solrLogManager = null)
85		{
86	7	$this->site = $site;
87	7	$this->indexQueue = $queue ?? GeneralUtility::makeInstance(Queue::class);
88	7	$this->signalSlotDispatcher = $dispatcher ?? GeneralUtility::makeInstance(Dispatcher::class);
89	7	$this->logger = $solrLogManager ?? GeneralUtility::makeInstance(SolrLogManager::class, /** @scrutinizer ignore-type */ __CLASS__);
90	7	define('EXT_SOLR_INDEXING_CONTEXT', true);
91	7	}
92
93		/**
94		* @param \ApacheSolrForTypo3\Solr\Task\IndexQueueWorkerTask $contextTask
95		*/
96	1	public function setContextTask($contextTask)
97		{
98	1	$this->contextTask = $contextTask;
99	1	}
100
101		/**
102		* @return \ApacheSolrForTypo3\Solr\Task\IndexQueueWorkerTask
103		*/
104	5	public function getContextTask()
105		{
106	5	return $this->contextTask;
107		}
108
109		/**
110		* Indexes items from the Index Queue.
111		*
112		* @param int $limit
113		* @return bool
114		*/
115	5	public function indexItems($limit)
116		{
117	5	$errors = 0;
118	5	$indexRunId = uniqid();
119	5	$configurationToUse = $this->site->getSolrConfiguration();
120	5	$enableCommitsSetting = $configurationToUse->getEnableCommits();
121
122		// get items to index
123	5	$itemsToIndex = $this->indexQueue->getItemsToIndex($this->site, $limit);
124
125	5	$this->emitSignal('beforeIndexItems', [$itemsToIndex, $this->getContextTask(), $indexRunId]);
126
127	5	foreach ($itemsToIndex as $itemToIndex) {
128		try {
129		// try indexing
130	5	$this->emitSignal('beforeIndexItem', [$itemToIndex, $this->getContextTask(), $indexRunId]);
131	5	$this->indexItem($itemToIndex, $configurationToUse);
132	5	$this->emitSignal('afterIndexItem', [$itemToIndex, $this->getContextTask(), $indexRunId]);
133		} catch (\Exception $e) {
134		$errors++;
135		$this->indexQueue->markItemAsFailed($itemToIndex, $e->getCode() . ': ' . $e->__toString());
136	5	$this->generateIndexingErrorLog($itemToIndex, $e);
137		}
138		}
139
140	5	$this->emitSignal('afterIndexItems', [$itemsToIndex, $this->getContextTask(), $indexRunId]);
141
142	5	if ($enableCommitsSetting && count($itemsToIndex) > 0) {
143	4	$solrServers = GeneralUtility::makeInstance(ConnectionManager::class)->getConnectionsBySite($this->site);
144	4	foreach ($solrServers as $solrServer) {
145	4	$solrServer->getWriteService()->commit(false, false, false);
146		}
147		}
148
149	5	return ($errors === 0);
150		}
151
152		/**
153		* Generates a message in the error log when an error occured.
154		*
155		* @param Item $itemToIndex
156		* @param \Exception $e
157		*/
158		protected function generateIndexingErrorLog(Item $itemToIndex, \Exception $e)
159		{
160		$message = 'Failed indexing Index Queue item ' . $itemToIndex->getIndexQueueUid();
161		$data = ['code' => $e->getCode(), 'message' => $e->getMessage(), 'trace' => $e->getTraceAsString(), 'item' => (array)$itemToIndex];
162
163		$this->logger->log(
164		SolrLogManager::ERROR,
165		$message,
166		$data
167		);
168		}
169
170		/**
171		* Builds an emits a singal for the IndexService.
172		*
173		* @param string $name
174		* @param array $arguments
175		* @return mixed
176		*/
177	5	protected function emitSignal($name, $arguments)
178		{
179	5	return $this->signalSlotDispatcher->dispatch(__CLASS__, $name, $arguments);
180		}
181
182		/**
183		* Indexes an item from the Index Queue.
184		*
185		* @param Item $item An index queue item to index
186		* @param TypoScriptConfiguration $configuration
187		* @return bool TRUE if the item was successfully indexed, FALSE otherwise
188		*/
189	4	protected function indexItem(Item $item, TypoScriptConfiguration $configuration)
190		{
191	4	$indexer = $this->getIndexerByItem($item->getIndexingConfigurationName(), $configuration);
192
193		// Remember original http host value
194	4	$originalHttpHost = isset($_SERVER['HTTP_HOST']) ? $_SERVER['HTTP_HOST'] : null;
195
196	4	$itemChangedDate = $item->getChanged();
197	4	$itemChangedDateAfterIndex = 0;
198
199	4	$this->initializeHttpServerEnvironment($item);
200	4	$itemIndexed = $indexer->index($item);
201
202		// update IQ item so that the IQ can determine what's been indexed already
203	4	if ($itemIndexed) {
204	4	$this->indexQueue->updateIndexTimeByItem($item);
205	4	$itemChangedDateAfterIndex = $item->getChanged();
206		}
207
208	4	if ($itemChangedDateAfterIndex > $itemChangedDate && $itemChangedDateAfterIndex > time()) {
209		$this->indexQueue->setForcedChangeTimeByItem($item, $itemChangedDateAfterIndex);
210		}
211
212	4	if (!is_null($originalHttpHost)) {
213		$_SERVER['HTTP_HOST'] = $originalHttpHost;
214		} else {
215	4	unset($_SERVER['HTTP_HOST']);
216		}
217
218		// needed since TYPO3 7.5
219	4	GeneralUtility::flushInternalRuntimeCaches();
220
221	4	return $itemIndexed;
222		}
223
224		/**
225		* A factory method to get an indexer depending on an item's configuration.
226		*
227		* By default all items are indexed using the default indexer
228		* (ApacheSolrForTypo3\Solr\IndexQueue\Indexer) coming with EXT:solr. Pages by default are
229		* configured to be indexed through a dedicated indexer
230		* (ApacheSolrForTypo3\Solr\IndexQueue\PageIndexer). In all other cases a dedicated indexer
231		* can be specified through TypoScript if needed.
232		*
233		* @param string $indexingConfigurationName Indexing configuration name.
234		* @param TypoScriptConfiguration $configuration
235		* @return Indexer
236		*/
237	4	protected function getIndexerByItem($indexingConfigurationName, TypoScriptConfiguration $configuration)
238		{
239	4	$indexerClass = $configuration->getIndexQueueIndexerByConfigurationName($indexingConfigurationName);
240	4	$indexerConfiguration = $configuration->getIndexQueueIndexerConfigurationByConfigurationName($indexingConfigurationName);
241
242	4	$indexer = GeneralUtility::makeInstance($indexerClass, /** @scrutinizer ignore-type */ $indexerConfiguration);
243	4	if (!($indexer instanceof Indexer)) {
244		throw new \RuntimeException(
245		'The indexer class "' . $indexerClass . '" for indexing configuration "' . $indexingConfigurationName . '" is not a valid indexer. Must be a subclass of ApacheSolrForTypo3\Solr\IndexQueue\Indexer.',
246		1260463206
247		);
248		}
249
250	4	return $indexer;
251		}
252
253		/**
254		* Gets the indexing progress.
255		*
256		* @return float Indexing progress as a two decimal precision float. f.e. 44.87
257		*/
258	1	public function getProgress()
259		{
260	1	return $this->indexQueue->getStatisticsBySite($this->site)->getSuccessPercentage();
261		}
262
263		/**
264		* Returns the amount of failed queue items for the current site.
265		*
266		* @return int
267		*/
268	1	public function getFailCount()
269		{
270	1	return $this->indexQueue->getStatisticsBySite($this->site)->getFailedCount();
271		}
272
273		/**
274		* Initializes the $_SERVER['HTTP_HOST'] environment variable in CLI
275		* environments dependent on the Index Queue item's root page.
276		*
277		* When the Index Queue Worker task is executed by a cron job there is no
278		* HTTP_HOST since we are in a CLI environment. RealURL needs the host
279		* information to generate a proper URL though. Using the Index Queue item's
280		* root page information we can determine the correct host although being
281		* in a CLI environment.
282		*
283		* @param Item $item Index Queue item to use to determine the host.
284		* @param
285		*/
286	4	protected function initializeHttpServerEnvironment(Item $item)
287		{
288	4	static $hosts = [];
289	4	$rootpageId = $item->getRootPageUid();
290	4	$hostFound = !empty($hosts[$rootpageId]);
291
292	4	if (!$hostFound) {
293	4	$rootline = BackendUtility::BEgetRootLine($rootpageId);
294	4	$host = BackendUtility::firstDomainRecord($rootline);
295	4	$hosts[$rootpageId] = $host;
296		}
297
298	4	$_SERVER['HTTP_HOST'] = $hosts[$rootpageId];
299
300		// needed since TYPO3 7.5
301	4	GeneralUtility::flushInternalRuntimeCaches();
302	4	}
303		}
304

timohund / ext-solr

Push — master ( a5d0a9...8581cd )

IndexService::indexItem() B

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like