DoctrineDatabase::buildWordIDArray() - Code Metrics - Inspection of "EZP-31287: Allowed indexation of strings including..." - ezsystems/ezpublish-kernel - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Push — EZP-31287 ( af9523...59859e )

unknown

created 2020-09-24 13:57 UTC

DoctrineDatabase::buildWordIDArray() B

↳ Parent: DoctrineDatabase

Complexity

Conditions	6
Paths	9

Size

Total Lines

Duplication

Lines	0
Ratio	0 %

Importance

Changes

Metric	Value
cc	6
nc	9
nop	1
dl	0
loc	43
rs	8.6097
c	0
b	0
f	0

<?php

/**
 * @copyright Copyright (C) eZ Systems AS. All rights reserved.
 * @license For full copyright and license information view LICENSE file distributed with this source code.
 */
namespace eZ\Publish\Core\Search\Legacy\Content\WordIndexer\Gateway;

use eZ\Publish\Core\Persistence\Legacy\Content\Language\MaskGenerator;
use eZ\Publish\Core\Search\Legacy\Content\WordIndexer\Gateway;
use eZ\Publish\Core\Persistence\Database\DatabaseHandler;
use eZ\Publish\Core\Persistence\TransformationProcessor;
use eZ\Publish\Core\Search\Legacy\Content\WordIndexer\Repository\SearchIndex;
use eZ\Publish\Core\Search\Legacy\Content\FullTextData;
use eZ\Publish\SPI\Persistence\Content;
use eZ\Publish\SPI\Persistence\Content\Type\Handler as SPITypeHandler;
use eZ\Publish\SPI\Search\Field;

/**
 * WordIndexer gateway implementation using the Doctrine database.
 */
class DoctrineDatabase extends Gateway
{
    /**
     * Max acceptable by any DBMS INT value.
     *
     * Note: 2^31-1 seems to be the most reasonable value that should work in any setup.
     */
    const DB_INT_MAX = 2147483647;

    /**
     * Database handler.
     *
     * @var \eZ\Publish\Core\Persistence\Database\DatabaseHandler
     *
     * @deprecated Start to use DBAL $connection instead.
     */
    protected $dbHandler;

    /**
     * SPI Content Type Handler.
     *
     * Need this for being able to pick fields that are searchable.
     *
     * @var \eZ\Publish\SPI\Persistence\Content\Type\Handler
     */
    protected $typeHandler;

    /**
     * Transformation processor.
     *
     * Need this for being able to transform text to searchable value
     *
     * @var \eZ\Publish\Core\Persistence\TransformationProcessor
     */
    protected $transformationProcessor;

    /**
     * LegacySearchService.
     *
     * Need this for queries on ezsearch* tables
     *
     * @var \eZ\Publish\Core\Search\Legacy\Content\WordIndexer\Repository\SearchIndex
     */
    protected $searchIndex;

    /** @var \eZ\Publish\Core\Persistence\Legacy\Content\Language\MaskGenerator */
    private $languageMaskGenerator;

    /**
     * Full text search configuration options.
     *
     * @var array
     */
    protected $fullTextSearchConfiguration;

    /**
     * Construct from handler handler.
     *
     * @param \eZ\Publish\Core\Persistence\Database\DatabaseHandler $dbHandler
     * @param \eZ\Publish\SPI\Persistence\Content\Type\Handler $typeHandler
     * @param \eZ\Publish\Core\Persistence\TransformationProcessor $transformationProcessor
     * @param \eZ\Publish\Core\Search\Legacy\Content\WordIndexer\Repository\SearchIndex $searchIndex
     * @param \eZ\Publish\Core\Persistence\Legacy\Content\Language\MaskGenerator $languageMaskGenerator
     * @param array $fullTextSearchConfiguration
     */
    public function __construct(
        DatabaseHandler $dbHandler,
        SPITypeHandler $typeHandler,
        TransformationProcessor $transformationProcessor,
        SearchIndex $searchIndex,
        MaskGenerator $languageMaskGenerator,
        array $fullTextSearchConfiguration
    ) {
        $this->dbHandler = $dbHandler;
        $this->typeHandler = $typeHandler;
        $this->transformationProcessor = $transformationProcessor;
        $this->searchIndex = $searchIndex;
        $this->fullTextSearchConfiguration = $fullTextSearchConfiguration;
        $this->languageMaskGenerator = $languageMaskGenerator;
    }

    /**
     * Index search engine full text data corresponding to content object field values.
     *
     * Ported from the legacy code
     *
     * @see https://github.com/ezsystems/ezpublish-legacy/blob/master/kernel/search/plugins/ezsearchengine/ezsearchengine.php#L45
     *
     * @param \eZ\Publish\Core\Search\Legacy\Content\FullTextData $fullTextData
     */
    public function index(FullTextData $fullTextData)
    {
        $indexArray = [];
        $indexArrayOnlyWords = [];
        $wordCount = 0;
        $placement = 0;

        // Remove previously indexed content if exists to avoid keeping in index removed field values
        $this->remove($fullTextData->id);
        foreach ($fullTextData->values as $fullTextValue) {
            /** @var \eZ\Publish\Core\Search\Legacy\Content\FullTextValue $fullTextValue */
            if (is_numeric(trim($fullTextValue->value))) {
                $integerValue = (int)$fullTextValue->value;
                if ($integerValue > self::DB_INT_MAX) {
                    $integerValue = 0;
                }
            } else {
                $integerValue = 0;
            }
            $text = $this->transformationProcessor->transform(
                $fullTextValue->value,
                !empty($fullTextValue->transformationRules)
                    ? $fullTextValue->transformationRules
                    : $this->fullTextSearchConfiguration['commands']
            );
            // split by non-words
            $wordArray = $fullTextValue->splitFlag ? preg_split('/\W/u', $text, -1, PREG_SPLIT_NO_EMPTY) : [$text];
            foreach ($wordArray as $word) {
                if (trim($word) === '') {
                    continue;
                }
                // words stored in search index are limited to 150 characters
                if (mb_strlen($word) > 150) {
                    $word = mb_substr($word, 0, 150);
                }
                $indexArray[] = [
                    'Word' => $word,
                    'ContentClassAttributeID' => $fullTextValue->fieldDefinitionId,
                    'identifier' => $fullTextValue->fieldDefinitionIdentifier,
                    'integer_value' => $integerValue,
                    'language_code' => $fullTextValue->languageCode,
                    'is_main_and_always_available' => $fullTextValue->isMainAndAlwaysAvailable,
                ];
                $indexArrayOnlyWords[$word] = 1;
                ++$wordCount;
                // if we have "www." before word than
                // treat it as url and add additional entry to the index
                if (mb_strtolower(mb_substr($word, 0, 4)) === 'www.') {
                    $additionalUrlWord = substr($word, 4);
                    $indexArray[] = [
                        'Word' => $additionalUrlWord,
                        'ContentClassAttributeID' => $fullTextValue->fieldDefinitionId,
                        'identifier' => $fullTextValue->fieldDefinitionIdentifier,
                        'integer_value' => $integerValue,
                        'language_code' => $fullTextValue->languageCode,
                        'is_main_and_always_available' => $fullTextValue->isMainAndAlwaysAvailable,
                    ];
                    $indexArrayOnlyWords[$additionalUrlWord] = 1;
                    ++$wordCount;
                }
            }
        }

        $wordIDArray = $this->buildWordIDArray(array_keys($indexArrayOnlyWords));
        $this->dbHandler->beginTransaction();
        for ($arrayCount = 0; $arrayCount < $wordCount; $arrayCount += 1000) {
            $placement = $this->indexWords(
                $fullTextData,
                array_slice($indexArray, $arrayCount, 1000),
                $wordIDArray,
                $placement
            );
        }
        $this->dbHandler->commit();
    }

    /**
     * Indexes an array of FullTextData objects.
     *
     * Note: on large amounts of data make sure to iterate with several calls to this function with
     * a limited set of FullTextData objects. Amount you have memory for depends on server, size
     * of FullTextData objects & PHP version.
     *
     * @param \eZ\Publish\Core\Search\Legacy\Content\FullTextData[] $fullTextBulkData
     */
    public function bulkIndex(array $fullTextBulkData)
    {
        foreach ($fullTextBulkData as $fullTextData) {
            $this->index($fullTextData);
        }
    }

    /**
     * Remove whole content or a specific version from index.
     *
     * Ported from the legacy code
     *
     * @see https://github.com/ezsystems/ezpublish-legacy/blob/master/kernel/search/plugins/ezsearchengine/ezsearchengine.php#L386
     *
     * @param mixed $contentId
     * @param mixed|null $versionId
     *
     * @return bool
     */
    public function remove($contentId, $versionId = null)
    {
        $doDelete = false;
        $this->dbHandler->beginTransaction();
        // fetch all the words and decrease the object count on all the words
        $wordIDList = $this->searchIndex->getContentObjectWords($contentId);
        if (count($wordIDList) > 0) {
            $this->searchIndex->decrementWordObjectCount($wordIDList);
            $doDelete = true;
        }
        if ($doDelete) {
            $this->searchIndex->deleteWordsWithoutObjects();
            $this->searchIndex->deleteObjectWordsLink($contentId);
        }
        $this->dbHandler->commit();

        return true;
    }

    /**
     * Remove entire search index.
     */
    public function purgeIndex()
    {
        $this->searchIndex->purge();
    }

    /**
     * Index wordIndex.
     *
     * Ported from the legacy code
     *
     * @see https://github.com/ezsystems/ezpublish-legacy/blob/master/kernel/search/plugins/ezsearchengine/ezsearchengine.php#L255
     *
     * @param \eZ\Publish\Core\Search\Legacy\Content\FullTextData $fullTextData
     * @param array $indexArray
     * @param array $wordIDArray
     * @param int $placement
     *
     * @return int last placement
     */
    private function indexWords(FullTextData $fullTextData, array $indexArray, array $wordIDArray, $placement = 0)
    {
        $contentId = $fullTextData->id;

        $prevWordId = 0;

        for ($i = 0; $i < count($indexArray); ++$i) {
for ($i=0; $i<count($array); $i++) { // calls count() on each iteration
}

// Better
for ($i=0, $c=count($array); $i<$c; $i++) { // calls count() just once
}
            $indexWord = $indexArray[$i]['Word'];
            $contentFieldId = $indexArray[$i]['ContentClassAttributeID'];
            $identifier = $indexArray[$i]['identifier'];
            $integerValue = $indexArray[$i]['integer_value'];
            $languageCode = $indexArray[$i]['language_code'];
            $wordId = $wordIDArray[$indexWord];
            $isMainAndAlwaysAvailable = $indexArray[$i]['is_main_and_always_available'];
            $languageMask = $this->languageMaskGenerator->generateLanguageMaskFromLanguageCodes(
                [$languageCode],
                $isMainAndAlwaysAvailable
            );

            if (isset($indexArray[$i + 1])) {
                $nextIndexWord = $indexArray[$i + 1]['Word'];
                $nextWordId = $wordIDArray[$nextIndexWord];
            } else {
                $nextWordId = 0;
            }
            $frequency = 0;
            $this->searchIndex->addObjectWordLink(
                $wordId,
                $contentId,
                $frequency,
                $placement,
                $nextWordId,
                $prevWordId,
                $fullTextData->contentTypeId,
                $contentFieldId,
                $fullTextData->published,
                $fullTextData->sectionId,
                $identifier,
                $integerValue,
                $languageMask
            );
            $prevWordId = $wordId;
            ++$placement;
        }

        return $placement;
    }

    /**
     * Build WordIDArray and update ezsearch_word table.
     *
     * Ported from the legacy code
     *
     * @see https://github.com/ezsystems/ezpublish-legacy/blob/master/kernel/search/plugins/ezsearchengine/ezsearchengine.php#L155
     *
     * @param array $indexArrayOnlyWords words for object to add
     *
     * @return array wordIDArray
     */
    private function buildWordIDArray(array $indexArrayOnlyWords)
    {
        $wordCount = count($indexArrayOnlyWords);
        $wordIDArray = [];
        $wordArray = [];

        // store the words in the index and remember the ID
        $this->dbHandler->beginTransaction();
        for ($arrayCount = 0; $arrayCount < $wordCount; $arrayCount += 500) {
            // Fetch already indexed words from database
            $wordArrayChuck = array_slice($indexArrayOnlyWords, $arrayCount, 500);
            $wordRes = $this->searchIndex->getWords($wordArrayChuck);

            // Build a has of the existing words
            $wordResCount = count($wordRes);
            $existingWordArray = [];
            for ($i = 0; $i < $wordResCount; ++$i) {
                $wordIDArray[] = $wordRes[$i]['id'];
                $existingWordArray[] = $wordRes[$i]['word'];
                $wordArray[$wordRes[$i]['word']] = $wordRes[$i]['id'];
            }

            // Update the object count of existing words by one
            if (count($wordIDArray) > 0) {
                $this->searchIndex->incrementWordObjectCount($wordIDArray);
            }

            // Insert if there is any news words
            $newWordArray = array_diff($wordArrayChuck, $existingWordArray);
            if (count($newWordArray) > 0) {
                $this->searchIndex->addWords($newWordArray);
                $newWordRes = $this->searchIndex->getWords($newWordArray);
                $newWordCount = count($newWordRes);
                for ($i = 0; $i < $newWordCount; ++$i) {
                    $wordLowercase = $this->transformationProcessor->transformByGroup($newWordRes[$i]['word'], 'lowercase');
                    $wordArray[$wordLowercase] = $newWordRes[$i]['id'];
                }
            }
        }
        $this->dbHandler->commit();

        return $wordArray;
    }
}


1			<?php
2
3			/**
4			* @copyright Copyright (C) eZ Systems AS. All rights reserved.
5			* @license For full copyright and license information view LICENSE file distributed with this source code.
6			*/
7			namespace eZ\Publish\Core\Search\Legacy\Content\WordIndexer\Gateway;
8
9			use eZ\Publish\Core\Persistence\Legacy\Content\Language\MaskGenerator;
10			use eZ\Publish\Core\Search\Legacy\Content\WordIndexer\Gateway;
11			use eZ\Publish\Core\Persistence\Database\DatabaseHandler;
12			use eZ\Publish\Core\Persistence\TransformationProcessor;
13			use eZ\Publish\Core\Search\Legacy\Content\WordIndexer\Repository\SearchIndex;
14			use eZ\Publish\Core\Search\Legacy\Content\FullTextData;
15			use eZ\Publish\SPI\Persistence\Content;
16			use eZ\Publish\SPI\Persistence\Content\Type\Handler as SPITypeHandler;
17			use eZ\Publish\SPI\Search\Field;
18
19			/**
20			* WordIndexer gateway implementation using the Doctrine database.
21			*/
22			class DoctrineDatabase extends Gateway
23			{
24			/**
25			* Max acceptable by any DBMS INT value.
26			*
27			* Note: 2^31-1 seems to be the most reasonable value that should work in any setup.
28			*/
29			const DB_INT_MAX = 2147483647;
30
31			/**
32			* Database handler.
33			*
34			* @var \eZ\Publish\Core\Persistence\Database\DatabaseHandler
35			*
36			* @deprecated Start to use DBAL $connection instead.
37			*/
38			protected $dbHandler;
39
40			/**
41			* SPI Content Type Handler.
42			*
43			* Need this for being able to pick fields that are searchable.
44			*
45			* @var \eZ\Publish\SPI\Persistence\Content\Type\Handler
46			*/
47			protected $typeHandler;
48
49			/**
50			* Transformation processor.
51			*
52			* Need this for being able to transform text to searchable value
53			*
54			* @var \eZ\Publish\Core\Persistence\TransformationProcessor
55			*/
56			protected $transformationProcessor;
57
58			/**
59			* LegacySearchService.
60			*
61			* Need this for queries on ezsearch* tables
62			*
63			* @var \eZ\Publish\Core\Search\Legacy\Content\WordIndexer\Repository\SearchIndex
64			*/
65			protected $searchIndex;
66
67			/** @var \eZ\Publish\Core\Persistence\Legacy\Content\Language\MaskGenerator */
68			private $languageMaskGenerator;
69
70			/**
71			* Full text search configuration options.
72			*
73			* @var array
74			*/
75			protected $fullTextSearchConfiguration;
76
77			/**
78			* Construct from handler handler.
79			*
80			* @param \eZ\Publish\Core\Persistence\Database\DatabaseHandler $dbHandler
81			* @param \eZ\Publish\SPI\Persistence\Content\Type\Handler $typeHandler
82			* @param \eZ\Publish\Core\Persistence\TransformationProcessor $transformationProcessor
83			* @param \eZ\Publish\Core\Search\Legacy\Content\WordIndexer\Repository\SearchIndex $searchIndex
84			* @param \eZ\Publish\Core\Persistence\Legacy\Content\Language\MaskGenerator $languageMaskGenerator
85			* @param array $fullTextSearchConfiguration
86			*/
87			public function __construct(
88			DatabaseHandler $dbHandler,
89			SPITypeHandler $typeHandler,
90			TransformationProcessor $transformationProcessor,
91			SearchIndex $searchIndex,
92			MaskGenerator $languageMaskGenerator,
93			array $fullTextSearchConfiguration
94			) {
95			$this->dbHandler = $dbHandler;
96			$this->typeHandler = $typeHandler;
97			$this->transformationProcessor = $transformationProcessor;
98			$this->searchIndex = $searchIndex;
99			$this->fullTextSearchConfiguration = $fullTextSearchConfiguration;
100			$this->languageMaskGenerator = $languageMaskGenerator;
101			}
102
103			/**
104			* Index search engine full text data corresponding to content object field values.
105			*
106			* Ported from the legacy code
107			*
108			* @see https://github.com/ezsystems/ezpublish-legacy/blob/master/kernel/search/plugins/ezsearchengine/ezsearchengine.php#L45
109			*
110			* @param \eZ\Publish\Core\Search\Legacy\Content\FullTextData $fullTextData
111			*/
112			public function index(FullTextData $fullTextData)
113			{
114			$indexArray = [];
115			$indexArrayOnlyWords = [];
116			$wordCount = 0;
117			$placement = 0;
118
119			// Remove previously indexed content if exists to avoid keeping in index removed field values
120			$this->remove($fullTextData->id);
121			foreach ($fullTextData->values as $fullTextValue) {
122			/** @var \eZ\Publish\Core\Search\Legacy\Content\FullTextValue $fullTextValue */
123			if (is_numeric(trim($fullTextValue->value))) {
124			$integerValue = (int)$fullTextValue->value;
125			if ($integerValue > self::DB_INT_MAX) {
126			$integerValue = 0;
127			}
128			} else {
129			$integerValue = 0;
130			}
131			$text = $this->transformationProcessor->transform(
132			$fullTextValue->value,
133			!empty($fullTextValue->transformationRules)
134			? $fullTextValue->transformationRules
135			: $this->fullTextSearchConfiguration['commands']
136			);
137			// split by non-words
138			$wordArray = $fullTextValue->splitFlag ? preg_split('/\W/u', $text, -1, PREG_SPLIT_NO_EMPTY) : [$text];
139			foreach ($wordArray as $word) {
140			if (trim($word) === '') {
141			continue;
142			}
143			// words stored in search index are limited to 150 characters
144			if (mb_strlen($word) > 150) {
145			$word = mb_substr($word, 0, 150);
146			}
147			$indexArray[] = [
148			'Word' => $word,
149			'ContentClassAttributeID' => $fullTextValue->fieldDefinitionId,
150			'identifier' => $fullTextValue->fieldDefinitionIdentifier,
151			'integer_value' => $integerValue,
152			'language_code' => $fullTextValue->languageCode,
153			'is_main_and_always_available' => $fullTextValue->isMainAndAlwaysAvailable,
154			];
155			$indexArrayOnlyWords[$word] = 1;
156			++$wordCount;
157			// if we have "www." before word than
158			// treat it as url and add additional entry to the index
159			if (mb_strtolower(mb_substr($word, 0, 4)) === 'www.') {
160			$additionalUrlWord = substr($word, 4);
161			$indexArray[] = [
162			'Word' => $additionalUrlWord,
163			'ContentClassAttributeID' => $fullTextValue->fieldDefinitionId,
164			'identifier' => $fullTextValue->fieldDefinitionIdentifier,
165			'integer_value' => $integerValue,
166			'language_code' => $fullTextValue->languageCode,
167			'is_main_and_always_available' => $fullTextValue->isMainAndAlwaysAvailable,
168			];
169			$indexArrayOnlyWords[$additionalUrlWord] = 1;
170			++$wordCount;
171			}
172			}
173			}
174
175			$wordIDArray = $this->buildWordIDArray(array_keys($indexArrayOnlyWords));
176			$this->dbHandler->beginTransaction();
177			for ($arrayCount = 0; $arrayCount < $wordCount; $arrayCount += 1000) {
178			$placement = $this->indexWords(
179			$fullTextData,
180			array_slice($indexArray, $arrayCount, 1000),
181			$wordIDArray,
182			$placement
183			);
184			}
185			$this->dbHandler->commit();
186			}
187
188			/**
189			* Indexes an array of FullTextData objects.
190			*
191			* Note: on large amounts of data make sure to iterate with several calls to this function with
192			* a limited set of FullTextData objects. Amount you have memory for depends on server, size
193			* of FullTextData objects & PHP version.
194			*
195			* @param \eZ\Publish\Core\Search\Legacy\Content\FullTextData[] $fullTextBulkData
196			*/
197			public function bulkIndex(array $fullTextBulkData)
198			{
199			foreach ($fullTextBulkData as $fullTextData) {
200			$this->index($fullTextData);
201			}
202			}
203
204			/**
205			* Remove whole content or a specific version from index.
206			*
207			* Ported from the legacy code
208			*
209			* @see https://github.com/ezsystems/ezpublish-legacy/blob/master/kernel/search/plugins/ezsearchengine/ezsearchengine.php#L386
210			*
211			* @param mixed $contentId
212			* @param mixed\|null $versionId
213			*
214			* @return bool
215			*/
216			public function remove($contentId, $versionId = null)
217			{
218			$doDelete = false;
219			$this->dbHandler->beginTransaction();
220			// fetch all the words and decrease the object count on all the words
221			$wordIDList = $this->searchIndex->getContentObjectWords($contentId);
222			if (count($wordIDList) > 0) {
223			$this->searchIndex->decrementWordObjectCount($wordIDList);
224			$doDelete = true;
225			}
226			if ($doDelete) {
227			$this->searchIndex->deleteWordsWithoutObjects();
228			$this->searchIndex->deleteObjectWordsLink($contentId);
229			}
230			$this->dbHandler->commit();
231
232			return true;
233			}
234
235			/**
236			* Remove entire search index.
237			*/
238			public function purgeIndex()
239			{
240			$this->searchIndex->purge();
241			}
242
243			/**
244			* Index wordIndex.
245			*
246			* Ported from the legacy code
247			*
248			* @see https://github.com/ezsystems/ezpublish-legacy/blob/master/kernel/search/plugins/ezsearchengine/ezsearchengine.php#L255
249			*
250			* @param \eZ\Publish\Core\Search\Legacy\Content\FullTextData $fullTextData
251			* @param array $indexArray
252			* @param array $wordIDArray
253			* @param int $placement
254			*
255			* @return int last placement
256			*/
257			private function indexWords(FullTextData $fullTextData, array $indexArray, array $wordIDArray, $placement = 0)
258			{
259			$contentId = $fullTextData->id;
260
261			$prevWordId = 0;
262
263			for ($i = 0; $i < count($indexArray); ++$i) {
			0 ignored issues – show Performance Best Practice introduced 2016-07-26 08:11 UTC by Report Bug Copy Issue Report It seems like you are calling the size function `count()` as part of the test condition. You might want to compute the size beforehand, and not on each iteration. If the size of the collection does not change during the iteration, it is generally a good practice to compute it beforehand, and not on each iteration: for ($i=0; $i<count($array); $i++) { // calls count() on each iteration } // Better for ($i=0, $c=count($array); $i<$c; $i++) { // calls count() just once } Loading history...
264			$indexWord = $indexArray[$i]['Word'];
265			$contentFieldId = $indexArray[$i]['ContentClassAttributeID'];
266			$identifier = $indexArray[$i]['identifier'];
267			$integerValue = $indexArray[$i]['integer_value'];
268			$languageCode = $indexArray[$i]['language_code'];
269			$wordId = $wordIDArray[$indexWord];
270			$isMainAndAlwaysAvailable = $indexArray[$i]['is_main_and_always_available'];
271			$languageMask = $this->languageMaskGenerator->generateLanguageMaskFromLanguageCodes(
272			[$languageCode],
273			$isMainAndAlwaysAvailable
274			);
275
276			if (isset($indexArray[$i + 1])) {
277			$nextIndexWord = $indexArray[$i + 1]['Word'];
278			$nextWordId = $wordIDArray[$nextIndexWord];
279			} else {
280			$nextWordId = 0;
281			}
282			$frequency = 0;
283			$this->searchIndex->addObjectWordLink(
284			$wordId,
285			$contentId,
286			$frequency,
287			$placement,
288			$nextWordId,
289			$prevWordId,
290			$fullTextData->contentTypeId,
291			$contentFieldId,
292			$fullTextData->published,
293			$fullTextData->sectionId,
294			$identifier,
295			$integerValue,
296			$languageMask
297			);
298			$prevWordId = $wordId;
299			++$placement;
300			}
301
302			return $placement;
303			}
304
305			/**
306			* Build WordIDArray and update ezsearch_word table.
307			*
308			* Ported from the legacy code
309			*
310			* @see https://github.com/ezsystems/ezpublish-legacy/blob/master/kernel/search/plugins/ezsearchengine/ezsearchengine.php#L155
311			*
312			* @param array $indexArrayOnlyWords words for object to add
313			*
314			* @return array wordIDArray
315			*/
316			private function buildWordIDArray(array $indexArrayOnlyWords)
317			{
318			$wordCount = count($indexArrayOnlyWords);
319			$wordIDArray = [];
320			$wordArray = [];
321
322			// store the words in the index and remember the ID
323			$this->dbHandler->beginTransaction();
324			for ($arrayCount = 0; $arrayCount < $wordCount; $arrayCount += 500) {
325			// Fetch already indexed words from database
326			$wordArrayChuck = array_slice($indexArrayOnlyWords, $arrayCount, 500);
327			$wordRes = $this->searchIndex->getWords($wordArrayChuck);
328
329			// Build a has of the existing words
330			$wordResCount = count($wordRes);
331			$existingWordArray = [];
332			for ($i = 0; $i < $wordResCount; ++$i) {
333			$wordIDArray[] = $wordRes[$i]['id'];
334			$existingWordArray[] = $wordRes[$i]['word'];
335			$wordArray[$wordRes[$i]['word']] = $wordRes[$i]['id'];
336			}
337
338			// Update the object count of existing words by one
339			if (count($wordIDArray) > 0) {
340			$this->searchIndex->incrementWordObjectCount($wordIDArray);
341			}
342
343			// Insert if there is any news words
344			$newWordArray = array_diff($wordArrayChuck, $existingWordArray);
345			if (count($newWordArray) > 0) {
346			$this->searchIndex->addWords($newWordArray);
347			$newWordRes = $this->searchIndex->getWords($newWordArray);
348			$newWordCount = count($newWordRes);
349			for ($i = 0; $i < $newWordCount; ++$i) {
350			$wordLowercase = $this->transformationProcessor->transformByGroup($newWordRes[$i]['word'], 'lowercase');
351			$wordArray[$wordLowercase] = $newWordRes[$i]['id'];
352			}
353			}
354			}
355			$this->dbHandler->commit();
356
357			return $wordArray;
358			}
359			}
360

ezsystems / ezpublish-kernel

Push — EZP-31287 ( af9523...59859e )

DoctrineDatabase::buildWordIDArray() B

Complexity

Size

Duplication

Importance

Duplication Side-by-Side

Filter issues like