ElasticSearch::index() - Code Metrics - Inspection of "Ap5 ap6 ap1 internal format" - kitodo/kitodo-publication - Measure and Improve Code Quality continuously with Scrutinizer

Passed

Pull Request — master (#195)

unknown

created 2021-02-08 17:01 UTC

ElasticSearch::index() F

↳ Parent: ElasticSearch

Complexity

Conditions	19
Paths	3841

Size

Total Lines	113
Code Lines	66

Duplication

Lines	0
Ratio	0 %

Importance

Changes	1
Bugs	0	Features	0

Metric	Value
cc	19
eloc	66
c	1
b	0
f	0
nc	3841
nop	1
dl	0
loc	113
rs	0.3499

How to fix Long Method Complexity

<?php
namespace EWW\Dpf\Services\ElasticSearch;

/*
 * 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 Elasticsearch\ClientBuilder;
use EWW\Dpf\Domain\Workflow\DocumentWorkflow;
use TYPO3\CMS\Extbase\Object\ObjectManager;
use EWW\Dpf\Configuration\ClientConfigurationManager;
use EWW\Dpf\Domain\Model\Document;
use TYPO3\CMS\Core\Utility\GeneralUtility;
use TYPO3\CMS\Core\Log\LogManager;

class ElasticSearch
{
    /**
     *
     * @var \TYPO3\CMS\Extbase\Configuration\ConfigurationManagerInterface
     * @inject
     */
    protected $configurationManager;

    /**
     * frontendUserRepository
     *
     * @var \EWW\Dpf\Domain\Repository\FrontendUserRepository
     * @inject
     */
    protected $frontendUserRepository = null;

    protected $client;

    protected $server = 'host.docker.internal'; //127.0.0.1';

    protected $port = '9200';

    protected $indexName = 'kitodo_publication';

    //protected $mapping = '';

    //protected $hits;

    protected $results;

    protected $elasticsearchMapper;


    /**
     * elasticsearch client constructor
     */
    public function __construct()
    {
        $objectManager = \TYPO3\CMS\Core\Utility\GeneralUtility::makeInstance(ObjectManager::class);

        $this->elasticsearchMapper = $objectManager->get(ElasticsearchMapper::class);

        $clientConfigurationManager = $objectManager->get(ClientConfigurationManager::class);

        $this->server = $clientConfigurationManager->getElasticSearchHost();
        $this->port = $clientConfigurationManager->getElasticSearchPort();

        $hosts = array(
            $this->server . ':' . $this->port,
        );

        $clientBuilder = ClientBuilder::create();
        $clientBuilder->setHosts($hosts);
        $this->client = $clientBuilder->build();

        $this->initializeIndex($this->indexName);

    }

    /**
     * Get typoscript settings
     *
     * @return mixed
     */
    public function getSettings()
    {
        $frameworkConfiguration = $this->configurationManager->getConfiguration(
            \TYPO3\CMS\Extbase\Configuration\ConfigurationManagerInterface::CONFIGURATION_TYPE_FRAMEWORK
        );
        return $frameworkConfiguration['settings'];
    }


    /**
     * Creates an index named by $indexName if it doesn't exist.
     *
     * @param $indexName
     */
    protected function initializeIndex($indexName)
    {
        $paramsIndex = [
            'index' => $indexName,
            'body' => [
                'settings' => [
                    //'index.requests.cache.enable' => false,
                    'analysis' => [
                        'filter' => [
                            'ngram' => [
                                'type' => 'ngram',
                                'min_gram' => 3,
                                'max_gram' => 3,
                                'token_chars' => [
                                    'letter',
                                    'digit'
                                ],
                            ]
                        ],
                        'analyzer' => [
                            'keyword_lowercase' => [
                                'tokenizer' => 'keyword',
                                'filter' => ['lowercase']
                            ]
                        ],
                        'normalizer' => [
                            'lowercase_normalizer' => [
                                'type' => 'custom',
                                'char_filter' => [],
                                'filter' => [
                                    'lowercase',
                                    'asciifolding'
                                ]
                            ]
                        ]
                    ]
                ],
                'mappings' => [
                    '_source' => [
                        'enabled' => true
                    ],
                    //'dynamic' => 'strict',
                    'properties' => [
                        'title' => [
                            'type' => 'text',
                            'fields' => [
                                'keyword' => [
                                    'type' => 'keyword',
                                    'normalizer' => 'lowercase_normalizer'
                                ]
                            ]
                        ],
                        'state' => [
                            'type' => 'keyword'
                        ],
                        'aliasState' => [
                            'type' => 'keyword'
                        ],
                        'year' => [
                            'type' => 'integer'
                        ],
                        'persons' => [
                            'type' => 'keyword'
                        ],
                        'personsSort' => [
                            'type' => 'text',
                            'fields' => [
                                'keyword' => [
                                    'type' => 'keyword',
                                    'normalizer' => 'lowercase_normalizer'
                                ]
                            ]
                        ],
                        'doctype' => [
                            'type' => 'keyword'
                        ],
                        'collections' => [
                            'type' => 'keyword'
                        ],
                        'hasFiles' => [
                            'type' => 'keyword'
                        ],
                        'creator' => [
                            'type' => 'keyword'
                        ],
                        'creatorRole' => [
                            'type' => 'keyword'
                        ],
                        'source' => [
                            'type' => 'text'
                        ],
                        'fobIdentifiers' => [
                            'type' => 'keyword'
                        ],
                        'personData' => [
                            //'enabled' => false,
                            'properties' => [
                                'name' => [
                                    'type' => 'keyword'
                                ],
                                'fobId' => [
                                    //'type' => 'keyword'
                                    'enabled' => false
                                ],
                                'index' => [
                                    //'type' => 'integer'
                                    'enabled' => false
                                ]
                            ]
                        ],
                        'affiliation' => [
                            'type' => 'keyword'
                        ],
                        'process_number' => [
                            'type' => 'keyword'
                        ],
                        'creationDate' => [
                            'type' =>  'date',
                            'format'=>  "yyyy-MM-dd"
                        ],
                        'embargoDate' => [
                            'type' =>  'date',
                            'format'=>  "yyyy-MM-dd"
                        ]
                    ]
                ]
            ]
        ];

        if (!$this->client->indices()->exists(['index' => $indexName])) {
            $this->client->indices()->create($paramsIndex);
        }

    }

    /**
     * Adds an document to the index.
     *
     * @param Document $document
     */
    public function index($document)
    {
        $data = json_decode($this->elasticsearchMapper->getElasticsearchJson($document));

        if ($data) {

            $data->state = $document->getState();
            $data->aliasState = DocumentWorkflow::STATE_TO_ALIASSTATE_MAPPING[$document->getState()];

            $data->objectIdentifier = $document->getObjectIdentifier();

            if (!$data->identifier || !is_array($data->identifier)) {
                $data->identifier = [];
            }
            $data->identifier[] = $document->getObjectIdentifier();
            $data->identifier[] = $document->getProcessNumber();

            if ($document->getCreator()) {
                $data->creator = $document->getCreator();
            } else {
                $data->creator = null;
            }


            if ($document->getCreator()) {
                /** @var \EWW\Dpf\Domain\Model\FrontendUser $creatorFeUser */
                $creatorFeUser = $this->frontendUserRepository->findByUid($document->getCreator());
                $data->creatorRole = $creatorFeUser->getUserRole();
            } else {
                $data->creatorRole = '';
            }

            $creationDate = new \DateTime($document->getCreationDate());

            $data->creationDate = $creationDate->format('Y-m-d');

            $data->year = $document->getPublicationYear();

            $notes = $document->getNotes();

            if ($notes && is_array($notes)) {
                $data->notes = $notes;
            } else {
                $data->notes = array();
            }

            $files = $document->getFile();
            if ($files->count() > 0) {
                $data->hasFiles = true;
            } else {
                $data->hasFiles = false;
            }


            $internalFormat = new \EWW\Dpf\Helper\InternalFormat($document->getXmlData());

            //$persons = array_merge($internalFormat->getAuthors(), $internalFormat->getPublishers());
            $persons = $internalFormat->getPersons();

            $fobIdentifiers = [];
            $personData = [];
            foreach ($persons as $person) {
                $fobIdentifiers[] = $person['fobId'];
                $personData[] = $person;
                //$data->persons[] = $person['name'];
                $data->persons[] = $person['fobId'];

                foreach ($person['affiliations'] as $affiliation) {
                    $data->affiliation[] = $affiliation;
                }

                foreach ($person['affiliationIdentifiers'] as $affiliationIdentifier) {
                    $data->affiliation[] = $affiliationIdentifier;
                }
            }

            $data->fobIdentifiers = $fobIdentifiers;
            $data->personData = $personData;

            if (sizeof($persons) > 0) {
                if (array_key_exists('family', $persons[0])) {
                    $data->personsSort = $persons[0]['family'];
                }
            }

            $data->source = $document->getSourceDetails();

            $data->universityCollection = false;
            if ($data->collections && is_array($data->collections)) {
                foreach ($data->collections as $collection) {
                    if ($collection == $this->getSettings()['universityCollection']) {
                        $data->universityCollection = true;
                        break;
                    }
                }
            }

            //$embargoDate = $document->getEmbargoDate();
            if ($embargoDate instanceof \DateTime) {

                $data->embargoDate = $embargoDate->format("Y-m-d");
            } else {
                $data->embargoDate = null;
            }

            $data->originalSourceTitle = $internalFormat->getOriginalSourceTitle();

            $data->fobIdentifiers = $internalFormat->getPersonFisIdentifiers();

            $this->client->index([
                'refresh' => 'wait_for',
                'index' => $this->indexName,
                'id' => $document->getDocumentIdentifier(),
                'body' => $data
            ]);

        }

    }


    /**
     * Deletes a document from the index
     *
     * @param string $identifier
     */
    public function delete($identifier)
    {
        try {

            $params = [
                'refresh' => 'wait_for',
                'index' => $this->indexName,
                'id' => $identifier
            ];

            $this->client->delete($params);

        } catch (\Exception $e) {
            /** @var $logger \TYPO3\CMS\Core\Log\Logger */
            $logger = GeneralUtility::makeInstance(LogManager::class)->getLogger(__CLASS__);
            $logger->warning('Document could not be deleted from the index.',
                [
                    'Document identifier' => $identifier
                ]
            );
        }
    }


    /**
     * @param $identifier
     */
    public function getDocument($identifier)
    {
        $params = [
            'index' => $this->indexName,
            'id'    => $identifier
        ];

        return $this->client->get($params);
    }


    /**
     * performs the
     * @param  array $query search query
     * @return array        result list
     */
    public function search($query, $type = null)
    {
        try {
            // define type and index
            if (empty($query['index'])) {
                $query['index'] = $this->indexName;
            }
            if (!empty($type)) {
                //$query['type'] = $type;
                // $query['type'] = $this->type;
            }

            // Search request
            $results = $this->client->search($query);

            //$this->hits = $results['hits']['total'];

            //$this->resultList = $results['hits'];

            $this->results = $results;

            return $this->results;
        } catch ( \Elasticsearch\Common\Exceptions\Curl\CouldNotConnectToHost $exception) {
            throw new \EWW\Dpf\Exceptions\ElasticSearchConnectionErrorException("Could not connect to repository server.");
        } catch (\Elasticsearch\Common\Exceptions\Curl\CouldNotResolveHostException $exception) {
            throw new \EWW\Dpf\Exceptions\ElasticSearchConnectionErrorException("Could not connect to repository server.");
        }
    }

    /**
     * Get the results
     * @return mixed
     */
    public function getResults()
    {
        // return results from the last search request
        return $this->results;
    }
}


1			<?php
2			namespace EWW\Dpf\Services\ElasticSearch;
3
4			/*
5			* This file is part of the TYPO3 CMS project.
6			*
7			* It is free software; you can redistribute it and/or modify it under
8			* the terms of the GNU General Public License, either version 2
9			* of the License, or any later version.
10			*
11			* For the full copyright and license information, please read the
12			* LICENSE.txt file that was distributed with this source code.
13			*
14			* The TYPO3 project - inspiring people to share!
15			*/
16
17			use Elasticsearch\ClientBuilder;
18			use EWW\Dpf\Domain\Workflow\DocumentWorkflow;
19			use TYPO3\CMS\Extbase\Object\ObjectManager;
20			use EWW\Dpf\Configuration\ClientConfigurationManager;
21			use EWW\Dpf\Domain\Model\Document;
22			use TYPO3\CMS\Core\Utility\GeneralUtility;
23			use TYPO3\CMS\Core\Log\LogManager;
24
25			class ElasticSearch
26			{
27			/**
28			*
29			* @var \TYPO3\CMS\Extbase\Configuration\ConfigurationManagerInterface
30			* @inject
31			*/
32			protected $configurationManager;
33
34			/**
35			* frontendUserRepository
36			*
37			* @var \EWW\Dpf\Domain\Repository\FrontendUserRepository
38			* @inject
39			*/
40			protected $frontendUserRepository = null;
41
42			protected $client;
43
44			protected $server = 'host.docker.internal'; //127.0.0.1';
45
46			protected $port = '9200';
47
48			protected $indexName = 'kitodo_publication';
49
50			//protected $mapping = '';
51
52			//protected $hits;
53
54			protected $results;
55
56			protected $elasticsearchMapper;
57
58
59			/**
60			* elasticsearch client constructor
61			*/
62			public function __construct()
63			{
64			$objectManager = \TYPO3\CMS\Core\Utility\GeneralUtility::makeInstance(ObjectManager::class);
65
66			$this->elasticsearchMapper = $objectManager->get(ElasticsearchMapper::class);
67
68			$clientConfigurationManager = $objectManager->get(ClientConfigurationManager::class);
69
70			$this->server = $clientConfigurationManager->getElasticSearchHost();
71			$this->port = $clientConfigurationManager->getElasticSearchPort();
72
73			$hosts = array(
74			$this->server . ':' . $this->port,
75			);
76
77			$clientBuilder = ClientBuilder::create();
78			$clientBuilder->setHosts($hosts);
79			$this->client = $clientBuilder->build();
80
81			$this->initializeIndex($this->indexName);
82
83			}
84
85			/**
86			* Get typoscript settings
87			*
88			* @return mixed
89			*/
90			public function getSettings()
91			{
92			$frameworkConfiguration = $this->configurationManager->getConfiguration(
93			\TYPO3\CMS\Extbase\Configuration\ConfigurationManagerInterface::CONFIGURATION_TYPE_FRAMEWORK
94			);
95			return $frameworkConfiguration['settings'];
96			}
97
98
99			/**
100			* Creates an index named by $indexName if it doesn't exist.
101			*
102			* @param $indexName
103			*/
104			protected function initializeIndex($indexName)
105			{
106			$paramsIndex = [
107			'index' => $indexName,
108			'body' => [
109			'settings' => [
110			//'index.requests.cache.enable' => false,
111			'analysis' => [
112			'filter' => [
113			'ngram' => [
114			'type' => 'ngram',
115			'min_gram' => 3,
116			'max_gram' => 3,
117			'token_chars' => [
118			'letter',
119			'digit'
120			],
121			]
122			],
123			'analyzer' => [
124			'keyword_lowercase' => [
125			'tokenizer' => 'keyword',
126			'filter' => ['lowercase']
127			]
128			],
129			'normalizer' => [
130			'lowercase_normalizer' => [
131			'type' => 'custom',
132			'char_filter' => [],
133			'filter' => [
134			'lowercase',
135			'asciifolding'
136			]
137			]
138			]
139			]
140			],
141			'mappings' => [
142			'_source' => [
143			'enabled' => true
144			],
145			//'dynamic' => 'strict',
146			'properties' => [
147			'title' => [
148			'type' => 'text',
149			'fields' => [
150			'keyword' => [
151			'type' => 'keyword',
152			'normalizer' => 'lowercase_normalizer'
153			]
154			]
155			],
156			'state' => [
157			'type' => 'keyword'
158			],
159			'aliasState' => [
160			'type' => 'keyword'
161			],
162			'year' => [
163			'type' => 'integer'
164			],
165			'persons' => [
166			'type' => 'keyword'
167			],
168			'personsSort' => [
169			'type' => 'text',
170			'fields' => [
171			'keyword' => [
172			'type' => 'keyword',
173			'normalizer' => 'lowercase_normalizer'
174			]
175			]
176			],
177			'doctype' => [
178			'type' => 'keyword'
179			],
180			'collections' => [
181			'type' => 'keyword'
182			],
183			'hasFiles' => [
184			'type' => 'keyword'
185			],
186			'creator' => [
187			'type' => 'keyword'
188			],
189			'creatorRole' => [
190			'type' => 'keyword'
191			],
192			'source' => [
193			'type' => 'text'
194			],
195			'fobIdentifiers' => [
196			'type' => 'keyword'
197			],
198			'personData' => [
199			//'enabled' => false,
200			'properties' => [
201			'name' => [
202			'type' => 'keyword'
203			],
204			'fobId' => [
205			//'type' => 'keyword'
206			'enabled' => false
207			],
208			'index' => [
209			//'type' => 'integer'
210			'enabled' => false
211			]
212			]
213			],
214			'affiliation' => [
215			'type' => 'keyword'
216			],
217			'process_number' => [
218			'type' => 'keyword'
219			],
220			'creationDate' => [
221			'type' => 'date',
222			'format'=> "yyyy-MM-dd"
223			],
224			'embargoDate' => [
225			'type' => 'date',
226			'format'=> "yyyy-MM-dd"
227			]
228			]
229			]
230			]
231			];
232
233			if (!$this->client->indices()->exists(['index' => $indexName])) {
234			$this->client->indices()->create($paramsIndex);
235			}
236
237			}
238
239			/**
240			* Adds an document to the index.
241			*
242			* @param Document $document
243			*/
244			public function index($document)
245			{
246			$data = json_decode($this->elasticsearchMapper->getElasticsearchJson($document));
247
248			if ($data) {
249
250			$data->state = $document->getState();
251			$data->aliasState = DocumentWorkflow::STATE_TO_ALIASSTATE_MAPPING[$document->getState()];
252
253			$data->objectIdentifier = $document->getObjectIdentifier();
254
255			if (!$data->identifier \|\| !is_array($data->identifier)) {
256			$data->identifier = [];
257			}
258			$data->identifier[] = $document->getObjectIdentifier();
259			$data->identifier[] = $document->getProcessNumber();
260
261			if ($document->getCreator()) {
262			$data->creator = $document->getCreator();
263			} else {
264			$data->creator = null;
265			}
266
267
268			if ($document->getCreator()) {
269			/** @var \EWW\Dpf\Domain\Model\FrontendUser $creatorFeUser */
270			$creatorFeUser = $this->frontendUserRepository->findByUid($document->getCreator());
271			$data->creatorRole = $creatorFeUser->getUserRole();
272			} else {
273			$data->creatorRole = '';
274			}
275
276			$creationDate = new \DateTime($document->getCreationDate());
277
278			$data->creationDate = $creationDate->format('Y-m-d');
279
280			$data->year = $document->getPublicationYear();
281
282			$notes = $document->getNotes();
283
284			if ($notes && is_array($notes)) {
285			$data->notes = $notes;
286			} else {
287			$data->notes = array();
288			}
289
290			$files = $document->getFile();
291			if ($files->count() > 0) {
292			$data->hasFiles = true;
293			} else {
294			$data->hasFiles = false;
295			}
296
297
298			$internalFormat = new \EWW\Dpf\Helper\InternalFormat($document->getXmlData());
299
300			//$persons = array_merge($internalFormat->getAuthors(), $internalFormat->getPublishers());
301			$persons = $internalFormat->getPersons();
302
303			$fobIdentifiers = [];
304			$personData = [];
305			foreach ($persons as $person) {
306			$fobIdentifiers[] = $person['fobId'];
307			$personData[] = $person;
308			//$data->persons[] = $person['name'];
309			$data->persons[] = $person['fobId'];
310
311			foreach ($person['affiliations'] as $affiliation) {
312			$data->affiliation[] = $affiliation;
313			}
314
315			foreach ($person['affiliationIdentifiers'] as $affiliationIdentifier) {
316			$data->affiliation[] = $affiliationIdentifier;
317			}
318			}
319
320			$data->fobIdentifiers = $fobIdentifiers;
321			$data->personData = $personData;
322
323			if (sizeof($persons) > 0) {
324			if (array_key_exists('family', $persons[0])) {
325			$data->personsSort = $persons[0]['family'];
326			}
327			}
328
329			$data->source = $document->getSourceDetails();
330
331			$data->universityCollection = false;
332			if ($data->collections && is_array($data->collections)) {
333			foreach ($data->collections as $collection) {
334			if ($collection == $this->getSettings()['universityCollection']) {
335			$data->universityCollection = true;
336			break;
337			}
338			}
339			}
340
341			//$embargoDate = $document->getEmbargoDate();
342			if ($embargoDate instanceof \DateTime) {
			0 ignored issues – show Comprehensibility Best Practice introduced 2020-12-15 14:55 UTC by Report Bug Copy Issue Report The variable `$embargoDate` seems to be never defined. Loading history...
343			$data->embargoDate = $embargoDate->format("Y-m-d");
344			} else {
345			$data->embargoDate = null;
346			}
347
348			$data->originalSourceTitle = $internalFormat->getOriginalSourceTitle();
349
350			$data->fobIdentifiers = $internalFormat->getPersonFisIdentifiers();
351
352			$this->client->index([
353			'refresh' => 'wait_for',
354			'index' => $this->indexName,
355			'id' => $document->getDocumentIdentifier(),
356			'body' => $data
357			]);
358
359			}
360
361			}
362
363
364			/**
365			* Deletes a document from the index
366			*
367			* @param string $identifier
368			*/
369			public function delete($identifier)
370			{
371			try {
372
373			$params = [
374			'refresh' => 'wait_for',
375			'index' => $this->indexName,
376			'id' => $identifier
377			];
378
379			$this->client->delete($params);
380
381			} catch (\Exception $e) {
382			/** @var $logger \TYPO3\CMS\Core\Log\Logger */
383			$logger = GeneralUtility::makeInstance(LogManager::class)->getLogger(__CLASS__);
384			$logger->warning('Document could not be deleted from the index.',
385			[
386			'Document identifier' => $identifier
387			]
388			);
389			}
390			}
391
392
393			/**
394			* @param $identifier
395			*/
396			public function getDocument($identifier)
397			{
398			$params = [
399			'index' => $this->indexName,
400			'id' => $identifier
401			];
402
403			return $this->client->get($params);
404			}
405
406
407			/**
408			* performs the
409			* @param array $query search query
410			* @return array result list
411			*/
412			public function search($query, $type = null)
413			{
414			try {
415			// define type and index
416			if (empty($query['index'])) {
417			$query['index'] = $this->indexName;
418			}
419			if (!empty($type)) {
420			//$query['type'] = $type;
421			// $query['type'] = $this->type;
422			}
423
424			// Search request
425			$results = $this->client->search($query);
426
427			//$this->hits = $results['hits']['total'];
428
429			//$this->resultList = $results['hits'];
430
431			$this->results = $results;
432
433			return $this->results;
434			} catch ( \Elasticsearch\Common\Exceptions\Curl\CouldNotConnectToHost $exception) {
435			throw new \EWW\Dpf\Exceptions\ElasticSearchConnectionErrorException("Could not connect to repository server.");
436			} catch (\Elasticsearch\Common\Exceptions\Curl\CouldNotResolveHostException $exception) {
437			throw new \EWW\Dpf\Exceptions\ElasticSearchConnectionErrorException("Could not connect to repository server.");
438			}
439			}
440
441			/**
442			* Get the results
443			* @return mixed
444			*/
445			public function getResults()
446			{
447			// return results from the last search request
448			return $this->results;
449			}
450			}
451

kitodo / kitodo-publication

Pull Request — master (#195)

ElasticSearch::index() F

Complexity

Size

Duplication

Importance

How to fix Long Method Complexity

Long Method

Duplication Side-by-Side

Filter issues like