GarbageHandler::collectGarbage() - Code Metrics - Inspection of "[TASK:BP:11.5] Ensure recursive page update on pag..." - TYPO3-Solr/ext-solr - Measure and Improve Code Quality continuously with Scrutinizer

Passed

Push — release-11.5.x ( 1b5a31...ef0664 )

by Markus

created 2023-09-20 11:21 UTC

GarbageHandler::collectGarbage() A

↳ Parent: GarbageHandler

Complexity

Conditions	1
Paths	1

Size

Total Lines	4
Code Lines	2

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	3
CRAP Score	1

Importance

Changes	1
Bugs	0	Features	0

Metric	Value
eloc	2
dl	0
loc	4
ccs	3
cts	3
cp	1
rs	10
c	1
b	0
f	0
cc	1
nc	1
nop	2
crap	1

<?php

declare(strict_types=1);

/*
 * 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!
 */

namespace ApacheSolrForTypo3\Solr\Domain\Index\Queue\UpdateHandler;

use ApacheSolrForTypo3\Solr\Domain\Index\Queue\GarbageRemover\StrategyFactory;
use Doctrine\DBAL\Driver\Exception as DBALDriverException;
use PDO;
use Throwable;
use TYPO3\CMS\Backend\Utility\BackendUtility;
use TYPO3\CMS\Core\Utility\GeneralUtility;
use UnexpectedValueException;

/**
 * Garbage handler
 *
 * Handles updates on potential relevant records and
 * collects the garbage, e.g. a deletion might require
 * index and index queue updates.
 */
class GarbageHandler extends AbstractUpdateHandler
{
    /**
     * Configuration used to check if recursive updates are required
     *
     * Holds the configuration when a recursive page queuing should be triggered, while processing record
     * updates
     *
     * Note: The SQL transaction is already committed, so the current state covers only "non"-changed fields.
     *
     * @var array
     */
    protected array $updateSubPagesRecursiveTriggerConfiguration = [
        // the current page has the field "extendToSubpages" enabled and the field "hidden" was set to 1
        // covers following scenarios:
        //   'currentState' =>  ['hidden' => '0', 'extendToSubpages' => '0|1'], 'changeSet' => ['hidden' => '1', (optional)'extendToSubpages' => '1']
        'extendToSubpageEnabledAndHiddenFlagWasAdded' => [
            'currentState' =>  ['extendToSubpages' => '1'],
            'changeSet' => ['hidden' => '1'],
        ],
        // the current page has the field "hidden" enabled and the field "extendToSubpages" was set to 1
        // covers following scenarios:
        //   'currentState' =>  ['hidden' => '0|1', 'extendToSubpages' => '0'], 'changeSet' => [(optional)'hidden' => '1', 'extendToSubpages' => '1']
        'hiddenIsEnabledAndExtendToSubPagesWasAdded' => [
            'currentState' =>  ['hidden' => '1'],
            'changeSet' => ['extendToSubpages' => '1'],
        ],
        // the field "no_search_sub_entries" of current page was set to 1
        'no_search_sub_entriesFlagWasAdded' => [
            'changeSet' => ['no_search_sub_entries' => '1'],
        ],
    ];

    /**
     * Tracks down index documents belonging to a particular record or page and
     * removes them from the index and the Index Queue.
     *
     * @param string $table The record's table name.
     * @param int $uid The record's uid.
     * @throws UnexpectedValueException if a hook object does not implement interface {@linkt \ApacheSolrForTypo3\Solr\GarbageCollectorPostProcessor::}
     */
    public function collectGarbage(string $table, int $uid): void
    {
        $garbageRemoverStrategy = StrategyFactory::getByTable($table);
        $garbageRemoverStrategy->removeGarbageOf($table, $uid);
    }

    /**
     * Handles moved pages
     *
     * As rootline and page slug might have changed on page movement,
     * document have to be removed from Solr. Reindexing is taken
     * care of by the DataUpdateHandler.
     *
     * @param int $uid
     * @param int|null $previousParentId
     */
    public function handlePageMovement(int $uid, ?int $previousParentId = null): void
    {
        $this->collectGarbage('pages', $uid);

        // collect garbage of subpages
        if ($previousParentId !== null) {
            $pageRecord = BackendUtility::getRecord('pages', $uid);
            if ($pageRecord !== null && (int)$pageRecord['pid'] !== $previousParentId) {
                $subPageIds = $this->getSubPageIds($uid);
                array_walk(
                    $subPageIds,
                    fn (int $subPageId) => $this->collectGarbage('pages', $subPageId)
                );
            }
        }
    }

    /**
     * Performs record garbage check
     *
     * @param int $uid
     * @param string $table
     * @param array $updatedFields
     * @param bool $frontendGroupsRemoved
     * @throws DBALDriverException
     */
    public function performRecordGarbageCheck(
        int $uid,
        string $table,
        array $updatedFields,
        bool $frontendGroupsRemoved
    ): void {
        $record = $this->getRecordWithFieldRelevantForGarbageCollection($table, $uid);

        // If no record could be found skip further processing
        if (empty($record)) {
            return;
        }

        if ($table === 'pages') {
            $this->deleteSubEntriesWhenRecursiveTriggerIsRecognized($table, $uid, $updatedFields);
        }

        $record = $this->tcaService->normalizeFrontendGroupField($table, $record);
        $isGarbage = $this->getIsGarbageRecord($table, $record, $frontendGroupsRemoved);
        if (!$isGarbage) {
            return;
        }

        $this->collectGarbage($table, $uid);
    }

    /**
     * @param string $table
     * @param int $uid
     * @param array $updatedFields
     * @throws DBALDriverException
     */
    protected function deleteSubEntriesWhenRecursiveTriggerIsRecognized(
        string $table,
        int $uid,
        array $updatedFields
    ): void {
        if (!$this->isRecursivePageUpdateRequired($uid, $updatedFields)) {
            return;
        }

        // get affected subpages when "extendToSubpages" flag was set
        $pagesToDelete = $this->getSubPageIds($uid);
        // we need to at least remove this page
        foreach ($pagesToDelete as $pageToDelete) {
            $this->collectGarbage($table, $pageToDelete);
        }
    }

    /**
     * Determines if a record is garbage and can be deleted.
     *
     * @param string $table
     * @param array $record
     * @param bool $frontendGroupsRemoved
     * @return bool
     * @throws DBALDriverException
     */
    protected function getIsGarbageRecord(string $table, array $record, bool $frontendGroupsRemoved): bool
    {
        return $frontendGroupsRemoved
            || $this->tcaService->isHidden($table, $record)
            || $this->isInvisibleByStartOrEndtime($table, $record)
            || ($table === 'pages' && $this->isPageExcludedFromSearch($record))
            || ($table === 'pages' && !$this->isIndexablePageType($record));
    }

    /**
     * Checks whether a page has a page type that can be indexed.
     * Currently, standard pages and mount pages can be indexed.
     *
     * @param array $record A page record
     * @return bool TRUE if the page can be indexed according to its page type, FALSE otherwise
     * @throws DBALDriverException
     */
    protected function isIndexablePageType(array $record): bool
    {
        return $this->frontendEnvironment->isAllowedPageType($record);
    }

    /**
     * Checks whether the page has been excluded from searching.
     *
     * @param array $record An array with record fields that may affect visibility.
     * @return bool True if the page has been excluded from searching, FALSE otherwise
     */
    protected function isPageExcludedFromSearch(array $record): bool
    {
        return (bool)$record['no_search'];
    }

    /**
     * Check if a record is getting invisible due to changes in start or endtime. In addition, it is checked that the related
     * queue item was marked as indexed.
     *
     * @param string $table
     * @param array $record
     * @return bool
     */
    protected function isInvisibleByStartOrEndtime(string $table, array $record): bool
    {
        return
            ($this->tcaService->isStartTimeInFuture($table, $record)
                || $this->tcaService->isEndTimeInPast($table, $record))
            && $this->isRelatedQueueRecordMarkedAsIndexed($table, $record)
        ;
    }

    /**
     * Checks if the related index queue item is indexed.
     *
     * * For tt_content the page from the pid is checked
     * * For all other records the table it's self is checked
     *
     * @param string $table The table name.
     * @param array $record An array with record fields that may affect visibility.
     * @return bool True if the record is marked as being indexed
     */
    protected function isRelatedQueueRecordMarkedAsIndexed(string $table, array $record): bool
    {
        if ($table === 'tt_content') {
            $table = 'pages';
            $uid = $record['pid'];
        } else {
            $uid = $record['uid'];
        }

        return $this->indexQueue->containsIndexedItem($table, $uid);
    }

    /**
     * Returns a record with all visibility affecting fields.
     *
     * @param string $table
     * @param int $uid
     * @return array|null
     */
    public function getRecordWithFieldRelevantForGarbageCollection(string $table, int $uid): ?array
    {
        $garbageCollectionRelevantFields = $this->tcaService->getVisibilityAffectingFieldsByTable($table);
        try {
            $queryBuilder = $this->getQueryBuilderForTable($table);
            $queryBuilder->getRestrictions()->removeAll();
            $row = $queryBuilder
                ->select(...GeneralUtility::trimExplode(',', $garbageCollectionRelevantFields, true))
                ->from($table)
                ->where($queryBuilder->expr()->eq('uid', $queryBuilder->createNamedParameter($uid, PDO::PARAM_INT)))
                ->executeQuery()
                ->fetchAssociative();
        } catch (Throwable $e) {
            $row = false;
        }

        return is_array($row) ? $row : null;
    }
}


1		<?php
2
3		declare(strict_types=1);
4
5		/*
6		* This file is part of the TYPO3 CMS project.
7		*
8		* It is free software; you can redistribute it and/or modify it under
9		* the terms of the GNU General Public License, either version 2
10		* of the License, or any later version.
11		*
12		* For the full copyright and license information, please read the
13		* LICENSE.txt file that was distributed with this source code.
14		*
15		* The TYPO3 project - inspiring people to share!
16		*/
17
18		namespace ApacheSolrForTypo3\Solr\Domain\Index\Queue\UpdateHandler;
19
20		use ApacheSolrForTypo3\Solr\Domain\Index\Queue\GarbageRemover\StrategyFactory;
21		use Doctrine\DBAL\Driver\Exception as DBALDriverException;
22		use PDO;
23		use Throwable;
24		use TYPO3\CMS\Backend\Utility\BackendUtility;
25		use TYPO3\CMS\Core\Utility\GeneralUtility;
26		use UnexpectedValueException;
27
28		/**
29		* Garbage handler
30		*
31		* Handles updates on potential relevant records and
32		* collects the garbage, e.g. a deletion might require
33		* index and index queue updates.
34		*/
35		class GarbageHandler extends AbstractUpdateHandler
36		{
37		/**
38		* Configuration used to check if recursive updates are required
39		*
40		* Holds the configuration when a recursive page queuing should be triggered, while processing record
41		* updates
42		*
43		* Note: The SQL transaction is already committed, so the current state covers only "non"-changed fields.
44		*
45		* @var array
46		*/
47		protected array $updateSubPagesRecursiveTriggerConfiguration = [
48		// the current page has the field "extendToSubpages" enabled and the field "hidden" was set to 1
49		// covers following scenarios:
50		// 'currentState' => ['hidden' => '0', 'extendToSubpages' => '0\|1'], 'changeSet' => ['hidden' => '1', (optional)'extendToSubpages' => '1']
51		'extendToSubpageEnabledAndHiddenFlagWasAdded' => [
52		'currentState' => ['extendToSubpages' => '1'],
53		'changeSet' => ['hidden' => '1'],
54		],
55		// the current page has the field "hidden" enabled and the field "extendToSubpages" was set to 1
56		// covers following scenarios:
57		// 'currentState' => ['hidden' => '0\|1', 'extendToSubpages' => '0'], 'changeSet' => [(optional)'hidden' => '1', 'extendToSubpages' => '1']
58		'hiddenIsEnabledAndExtendToSubPagesWasAdded' => [
59		'currentState' => ['hidden' => '1'],
60		'changeSet' => ['extendToSubpages' => '1'],
61		],
62		// the field "no_search_sub_entries" of current page was set to 1
63		'no_search_sub_entriesFlagWasAdded' => [
64		'changeSet' => ['no_search_sub_entries' => '1'],
65		],
66		];
67
68		/**
69		* Tracks down index documents belonging to a particular record or page and
70		* removes them from the index and the Index Queue.
71		*
72		* @param string $table The record's table name.
73		* @param int $uid The record's uid.
74		* @throws UnexpectedValueException if a hook object does not implement interface {@linkt \ApacheSolrForTypo3\Solr\GarbageCollectorPostProcessor::}
75		*/
76	36	public function collectGarbage(string $table, int $uid): void
77		{
78	36	$garbageRemoverStrategy = StrategyFactory::getByTable($table);
79	36	$garbageRemoverStrategy->removeGarbageOf($table, $uid);
80		}
81
82		/**
83		* Handles moved pages
84		*
85		* As rootline and page slug might have changed on page movement,
86		* document have to be removed from Solr. Reindexing is taken
87		* care of by the DataUpdateHandler.
88		*
89		* @param int $uid
90		* @param int\|null $previousParentId
91		*/
92	7	public function handlePageMovement(int $uid, ?int $previousParentId = null): void
93		{
94	7	$this->collectGarbage('pages', $uid);
95
96		// collect garbage of subpages
97	7	if ($previousParentId !== null) {
98	6	$pageRecord = BackendUtility::getRecord('pages', $uid);
99	6	if ($pageRecord !== null && (int)$pageRecord['pid'] !== $previousParentId) {
100	5	$subPageIds = $this->getSubPageIds($uid);
101	5	array_walk(
102	5	$subPageIds,
103	5	fn (int $subPageId) => $this->collectGarbage('pages', $subPageId)
104	5	);
105		}
106		}
107		}
108
109		/**
110		* Performs record garbage check
111		*
112		* @param int $uid
113		* @param string $table
114		* @param array $updatedFields
115		* @param bool $frontendGroupsRemoved
116		* @throws DBALDriverException
117		*/
118	19	public function performRecordGarbageCheck(
119		int $uid,
120		string $table,
121		array $updatedFields,
122		bool $frontendGroupsRemoved
123		): void {
124	19	$record = $this->getRecordWithFieldRelevantForGarbageCollection($table, $uid);
125
126		// If no record could be found skip further processing
127	19	if (empty($record)) {
128		return;
129		}
130
131	19	if ($table === 'pages') {
132	10	$this->deleteSubEntriesWhenRecursiveTriggerIsRecognized($table, $uid, $updatedFields);
133		}
134
135	19	$record = $this->tcaService->normalizeFrontendGroupField($table, $record);
136	19	$isGarbage = $this->getIsGarbageRecord($table, $record, $frontendGroupsRemoved);
137	19	if (!$isGarbage) {
138	4	return;
139		}
140
141	15	$this->collectGarbage($table, $uid);
142		}
143
144		/**
145		* @param string $table
146		* @param int $uid
147		* @param array $updatedFields
148		* @throws DBALDriverException
149		*/
150	10	protected function deleteSubEntriesWhenRecursiveTriggerIsRecognized(
151		string $table,
152		int $uid,
153		array $updatedFields
154		): void {
155	10	if (!$this->isRecursivePageUpdateRequired($uid, $updatedFields)) {
156	5	return;
157		}
158
159		// get affected subpages when "extendToSubpages" flag was set
160	5	$pagesToDelete = $this->getSubPageIds($uid);
161		// we need to at least remove this page
162	5	foreach ($pagesToDelete as $pageToDelete) {
163	5	$this->collectGarbage($table, $pageToDelete);
164		}
165		}
166
167		/**
168		* Determines if a record is garbage and can be deleted.
169		*
170		* @param string $table
171		* @param array $record
172		* @param bool $frontendGroupsRemoved
173		* @return bool
174		* @throws DBALDriverException
175		*/
176	19	protected function getIsGarbageRecord(string $table, array $record, bool $frontendGroupsRemoved): bool
177		{
178	19	return $frontendGroupsRemoved
179	19	\|\| $this->tcaService->isHidden($table, $record)
180	19	\|\| $this->isInvisibleByStartOrEndtime($table, $record)
181	19	\|\| ($table === 'pages' && $this->isPageExcludedFromSearch($record))
182	19	\|\| ($table === 'pages' && !$this->isIndexablePageType($record));
183		}
184
185		/**
186		* Checks whether a page has a page type that can be indexed.
187		* Currently, standard pages and mount pages can be indexed.
188		*
189		* @param array $record A page record
190		* @return bool TRUE if the page can be indexed according to its page type, FALSE otherwise
191		* @throws DBALDriverException
192		*/
193	3	protected function isIndexablePageType(array $record): bool
194		{
195	3	return $this->frontendEnvironment->isAllowedPageType($record);
196		}
197
198		/**
199		* Checks whether the page has been excluded from searching.
200		*
201		* @param array $record An array with record fields that may affect visibility.
202		* @return bool True if the page has been excluded from searching, FALSE otherwise
203		*/
204	3	protected function isPageExcludedFromSearch(array $record): bool
205		{
206	3	return (bool)$record['no_search'];
207		}
208
209		/**
210		* Check if a record is getting invisible due to changes in start or endtime. In addition, it is checked that the related
211		* queue item was marked as indexed.
212		*
213		* @param string $table
214		* @param array $record
215		* @return bool
216		*/
217	9	protected function isInvisibleByStartOrEndtime(string $table, array $record): bool
218		{
219	9	return
220	9	($this->tcaService->isStartTimeInFuture($table, $record)
221	9	\|\| $this->tcaService->isEndTimeInPast($table, $record))
222	9	&& $this->isRelatedQueueRecordMarkedAsIndexed($table, $record)
223	9	;
224		}
225
226		/**
227		* Checks if the related index queue item is indexed.
228		*
229		* * For tt_content the page from the pid is checked
230		* * For all other records the table it's self is checked
231		*
232		* @param string $table The table name.
233		* @param array $record An array with record fields that may affect visibility.
234		* @return bool True if the record is marked as being indexed
235		*/
236	4	protected function isRelatedQueueRecordMarkedAsIndexed(string $table, array $record): bool
237		{
238	4	if ($table === 'tt_content') {
239	4	$table = 'pages';
240	4	$uid = $record['pid'];
241		} else {
242		$uid = $record['uid'];
243		}
244
245	4	return $this->indexQueue->containsIndexedItem($table, $uid);
246		}
247
248		/**
249		* Returns a record with all visibility affecting fields.
250		*
251		* @param string $table
252		* @param int $uid
253		* @return array\|null
254		*/
255	20	public function getRecordWithFieldRelevantForGarbageCollection(string $table, int $uid): ?array
256		{
257	20	$garbageCollectionRelevantFields = $this->tcaService->getVisibilityAffectingFieldsByTable($table);
258		try {
259	20	$queryBuilder = $this->getQueryBuilderForTable($table);
260	20	$queryBuilder->getRestrictions()->removeAll();
261	20	$row = $queryBuilder
262	20	->select(...GeneralUtility::trimExplode(',', $garbageCollectionRelevantFields, true))
263	20	->from($table)
264	20	->where($queryBuilder->expr()->eq('uid', $queryBuilder->createNamedParameter($uid, PDO::PARAM_INT)))
265	20	->executeQuery()
266	20	->fetchAssociative();
267		} catch (Throwable $e) {
268		$row = false;
269		}
270
271	20	return is_array($row) ? $row : null;
272		}
273		}
274

TYPO3-Solr / ext-solr

Push — release-11.5.x ( 1b5a31...ef0664 )

GarbageHandler::collectGarbage() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like