SolrCellService::extractText() - Code Metrics - Inspection of "[TASK] TYPO3 11 LTS and PHP 8.1 compatibility" - TYPO3-Solr/ext-tika - Measure and Improve Code Quality continuously with Scrutinizer

Passed

Push — release-11.0.x ( 910681...ed160c )

by Rafael

created 2022-01-29 14:25 UTC

SolrCellService::extractText() A

↳ Parent: SolrCellService

Complexity

Conditions	1
Paths	1

Size

Total Lines	20
Code Lines	13

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	14
CRAP Score	1

Importance

Changes	2
Bugs	0	Features	0

Metric	Value
cc	1
eloc	13
c	2
b	0
f	0
nc	1
nop	1
dl	0
loc	20
ccs	14
cts	14
cp	1
crap	1
rs	9.8333

<?php

declare(strict_types=1);

namespace ApacheSolrForTypo3\Tika\Service\Tika;

/*
 * 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 ApacheSolrForTypo3\Solr\ConnectionManager;
use ApacheSolrForTypo3\Solr\System\Solr\SolrConnection;
use Solarium\QueryType\Extract\Query;
use TYPO3\CMS\Core\Resource\FileInterface;
use TYPO3\CMS\Core\Utility\GeneralUtility;

/**
 * A Tika service implementation using a Solr server
 *
 * @author Ingo Renner <[email protected]>
 */
class SolrCellService extends AbstractService
{

    /**
     * Solr connection
     *
     * @var SolrConnection
     */
    protected SolrConnection $solrConnection;

    /**
     * Service initialization
     */
    protected function initializeService(): void
    {
        // EM might define a different connection than already in use by
        // Index Queue
        /** @var ConnectionManager $connectionManager */
        $connectionManager =  GeneralUtility::makeInstance(ConnectionManager::class);

        $readNode = [
            'host' => $this->configuration['solrHost'],
            'port' => $this->configuration['solrPort'],
            'path' => $this->configuration['solrPath'],
            'scheme' => $this->configuration['solrScheme'],
        ];
        $writeNode = $readNode;
        $this->solrConnection = $connectionManager->getSolrConnectionForNodes($readNode, $writeNode);
    }

    /**
     * Retrieves a configuration value or a default value when not available.
     *
     * @param string $key
     * @param mixed $defaultValue
     * @return mixed
     */
    protected function getConfigurationOrDefaultValue(string $key, $defaultValue)
    {
        return $this->configuration[$key] ?? $defaultValue;
    }

    /**
     * Takes a file reference and extracts the text from it.
     *
     * @param FileInterface $file
     * @return string
     */
    public function extractText(FileInterface $file): string
    {
        $localTempFilePath = $file->getForLocalProcessing(false);
        /** @var Query $query */
        $query = GeneralUtility::makeInstance(Query::class);
        $query->setFile($localTempFilePath);
        $query->setExtractOnly(true);
        $query->addParam('extractFormat', 'text');

        $writer = $this->solrConnection->getWriteService();
        $response = $writer->extractByQuery($query);

        $this->log('Text Extraction using Solr', [
            'file' => $file,
            'solr connection' => (array)$writer,
            'query' => (array)$query,
            'response' => $response,
        ]);

        return $response[0] ?? '';
    }

    /**
     * Takes a file reference and extracts its meta-data.
     *
     * @param FileInterface $file
     * @return array
     */
    public function extractMetaData(FileInterface $file): array
    {
        $localTempFilePath = $file->getForLocalProcessing(false);
        /** @var Query $query */
        $query = GeneralUtility::makeInstance(Query::class);
        $query->setFile($localTempFilePath);
        $query->setExtractOnly(true);
        $query->addParam('extractFormat', 'text');

        $writer = $this->solrConnection->getWriteService();
        $response = $writer->extractByQuery($query);

        $metaData = [];
        if (isset($response[1]) && is_array($response[1])) {
            $metaData = $this->solrResponseToArray($response[1]);
        }

        $this->log('Meta Data Extraction using Solr', [
            'file' => $file,
            'solr connection' => (array)$writer,
            'query' => (array)$query,
            'response' => $response,
            'meta data' => $metaData,
        ]);

        return $metaData;
    }

    /**
     * Takes a file reference and detects its content's language.
     *
     * @param FileInterface $file
     * @return string Language ISO code
     */
    public function detectLanguageFromFile(FileInterface $file): string
    {
        // TODO check whether Solr supports text extraction now
        throw new UnsupportedOperationException(
            'The Tika Solr service does not support language detection',
            1423457153
        );
    }

    /**
     * Takes a string as input and detects its language.
     *
     * @param string $input
     * @return string Language ISO code
     */
    public function detectLanguageFromString(string $input): string
    {
        // TODO check whether Solr supports text extraction now
        throw new UnsupportedOperationException(
            'The Tika Solr service does not support language detection',
            1423457153
        );
    }

    /**
     * Turns the nested Solr response into the same format as produced by a
     * local Tika jar call
     *
     * @param array $metaDataResponse The part of the Solr response containing the meta-data
     * @return array The cleaned meta-data, matching the Tika jar call format
     */
    protected function solrResponseToArray(array $metaDataResponse = []): array
    {
        $cleanedData = [];

        foreach ($metaDataResponse as $dataName => $dataArray) {
            if (!($dataName % 2) == 0) {
                continue;
            }
            $fieldName = $dataArray;
            $fieldValue = $metaDataResponse[$dataName + 1] ?? [''];

            $cleanedData[$fieldName] = $fieldValue[0];
        }

        return $cleanedData;
    }

    /**
     * Gets the Tika version
     *
     * @return string Apache Solr server version string
     */
    public function getTikaVersion(): string
    {
        // TODO add patch for endpoint on Apache Solr to return Tika version
        // for now returns the Solr version string f.e. "Apache Solr X.Y.Z"
        return $this->solrConnection->getAdminService()->getSolrServerVersion();
    }

    /**
     * Since solr cell does not allow to query the supported mimetypes, we return a list of known supported mimetypes here.
     *
     * @return array
     */
    public function getSupportedMimeTypes(): array
    {
        $mapping = [
            'application/epub+zip' => ['epub'],
            'application/gzip' => ['gz', 'tgz'],
            'application/msword' => ['doc'],
            'application/pdf' => ['pdf'],
            'application/rtf' => ['rtf'],
            'application/vnd.ms-excel' => ['xsl'],
            'application/vnd.ms-outlook' => ['msg'],
            'application/vnd.oasis.opendocument.formula' => ['odf'],
            'application/vnd.oasis.opendocument.text' => ['odt'],
            'application/vnd.openxmlformats-officedocument.presentationml.presentation' => ['pptx'],
            'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet' => ['xlsx'],
            'application/vnd.openxmlformats-officedocument.wordprocessingml.document' => ['docx'],
            'application/vnd.sun.xml.writer' => ['sxw'],
            'application/zip' => ['zip'],
            'application/x-midi' => ['mid'],
            'application/xml' => ['xml'],
            'audio/aiff' => ['aif', 'aiff'],
            'audio/basic' => ['au'],
            'audio/midi' => ['mid'],
            'audio/mpeg3' => ['mp3'],
            'audio/mpeg' => ['mp3'],
            'audio/wav' => ['wav'],
            'audio/x-mpeg-3' => ['mp3'],
            'audio/x-wav' => ['wav'],
            'image/bmp' => ['bmp'],
            'image/gif' => ['gif'],
            'image/jpeg' => ['jpg', 'jpeg'],
            'image/png' => ['png'],
            'image/svg+xml' => ['svg'],
            'image/tiff' => ['tif', 'tiff'],
            'text/html' => ['html', 'htm'],
            'text/plain' => ['txt'],
            'text/xml' => ['xml'],
            'video/mpeg' => ['mp3'],
            'video/x-mpeg' => ['mp3'],
        ];

        return array_keys($mapping);
    }

    /**
     * The service is available when the solr server is reachable.
     *
     * @return bool
     */
    public function isAvailable(): bool
    {
        return $this->solrConnection->getWriteService()->ping();
    }
}


1		<?php
2
3		declare(strict_types=1);
4
5		namespace ApacheSolrForTypo3\Tika\Service\Tika;
6
7		/*
8		* This file is part of the TYPO3 CMS project.
9		*
10		* It is free software; you can redistribute it and/or modify it under
11		* the terms of the GNU General Public License, either version 2
12		* of the License, or any later version.
13		*
14		* For the full copyright and license information, please read the
15		* LICENSE.txt file that was distributed with this source code.
16		*
17		* The TYPO3 project - inspiring people to share!
18		*/
19
20		use ApacheSolrForTypo3\Solr\ConnectionManager;
21		use ApacheSolrForTypo3\Solr\System\Solr\SolrConnection;
22		use Solarium\QueryType\Extract\Query;
23		use TYPO3\CMS\Core\Resource\FileInterface;
24		use TYPO3\CMS\Core\Utility\GeneralUtility;
25
26		/**
27		* A Tika service implementation using a Solr server
28		*
29		* @author Ingo Renner <[email protected]>
30		*/
31		class SolrCellService extends AbstractService
32		{
33
34		/**
35		* Solr connection
36		*
37		* @var SolrConnection
38		*/
39		protected SolrConnection $solrConnection;
40
41		/**
42		* Service initialization
43		*/
44	6	protected function initializeService(): void
45		{
46		// EM might define a different connection than already in use by
47		// Index Queue
48		/** @var ConnectionManager $connectionManager */
49	6	$connectionManager = GeneralUtility::makeInstance(ConnectionManager::class);
50
51	6	$readNode = [
52	6	'host' => $this->configuration['solrHost'],
53	6	'port' => $this->configuration['solrPort'],
54	6	'path' => $this->configuration['solrPath'],
55	6	'scheme' => $this->configuration['solrScheme'],
56		];
57	6	$writeNode = $readNode;
58	6	$this->solrConnection = $connectionManager->getSolrConnectionForNodes($readNode, $writeNode);
59	6	}
60
61		/**
62		* Retrieves a configuration value or a default value when not available.
63		*
64		* @param string $key
65		* @param mixed $defaultValue
66		* @return mixed
67		*/
68		protected function getConfigurationOrDefaultValue(string $key, $defaultValue)
69		{
70		return $this->configuration[$key] ?? $defaultValue;
71		}
72
73		/**
74		* Takes a file reference and extracts the text from it.
75		*
76		* @param FileInterface $file
77		* @return string
78		*/
79	2	public function extractText(FileInterface $file): string
80		{
81	2	$localTempFilePath = $file->getForLocalProcessing(false);
82		/** @var Query $query */
83	2	$query = GeneralUtility::makeInstance(Query::class);
84	2	$query->setFile($localTempFilePath);
85	2	$query->setExtractOnly(true);
86	2	$query->addParam('extractFormat', 'text');
87
88	2	$writer = $this->solrConnection->getWriteService();
89	2	$response = $writer->extractByQuery($query);
90
91	2	$this->log('Text Extraction using Solr', [
92	2	'file' => $file,
93	2	'solr connection' => (array)$writer,
94	2	'query' => (array)$query,
95	2	'response' => $response,
96		]);
97
98	2	return $response[0] ?? '';
99		}
100
101		/**
102		* Takes a file reference and extracts its meta-data.
103		*
104		* @param FileInterface $file
105		* @return array
106		*/
107	2	public function extractMetaData(FileInterface $file): array
108		{
109	2	$localTempFilePath = $file->getForLocalProcessing(false);
110		/** @var Query $query */
111	2	$query = GeneralUtility::makeInstance(Query::class);
112	2	$query->setFile($localTempFilePath);
113	2	$query->setExtractOnly(true);
114	2	$query->addParam('extractFormat', 'text');
115
116	2	$writer = $this->solrConnection->getWriteService();
117	2	$response = $writer->extractByQuery($query);
118
119	2	$metaData = [];
120	2	if (isset($response[1]) && is_array($response[1])) {
121	2	$metaData = $this->solrResponseToArray($response[1]);
122		}
123
124	2	$this->log('Meta Data Extraction using Solr', [
125	2	'file' => $file,
126	2	'solr connection' => (array)$writer,
127	2	'query' => (array)$query,
128	2	'response' => $response,
129	2	'meta data' => $metaData,
130		]);
131
132	2	return $metaData;
133		}
134
135		/**
136		* Takes a file reference and detects its content's language.
137		*
138		* @param FileInterface $file
139		* @return string Language ISO code
140		*/
141		public function detectLanguageFromFile(FileInterface $file): string
142		{
143		// TODO check whether Solr supports text extraction now
144		throw new UnsupportedOperationException(
145		'The Tika Solr service does not support language detection',
146		1423457153
147		);
148		}
149
150		/**
151		* Takes a string as input and detects its language.
152		*
153		* @param string $input
154		* @return string Language ISO code
155		*/
156		public function detectLanguageFromString(string $input): string
157		{
158		// TODO check whether Solr supports text extraction now
159		throw new UnsupportedOperationException(
160		'The Tika Solr service does not support language detection',
161		1423457153
162		);
163		}
164
165		/**
166		* Turns the nested Solr response into the same format as produced by a
167		* local Tika jar call
168		*
169		* @param array $metaDataResponse The part of the Solr response containing the meta-data
170		* @return array The cleaned meta-data, matching the Tika jar call format
171		*/
172	2	protected function solrResponseToArray(array $metaDataResponse = []): array
173		{
174	2	$cleanedData = [];
175
176	2	foreach ($metaDataResponse as $dataName => $dataArray) {
177	2	if (!($dataName % 2) == 0) {
178	1	continue;
179		}
180	2	$fieldName = $dataArray;
181	2	$fieldValue = $metaDataResponse[$dataName + 1] ?? [''];
182
183	2	$cleanedData[$fieldName] = $fieldValue[0];
184		}
185
186	2	return $cleanedData;
187		}
188
189		/**
190		* Gets the Tika version
191		*
192		* @return string Apache Solr server version string
193		*/
194		public function getTikaVersion(): string
195		{
196		// TODO add patch for endpoint on Apache Solr to return Tika version
197		// for now returns the Solr version string f.e. "Apache Solr X.Y.Z"
198		return $this->solrConnection->getAdminService()->getSolrServerVersion();
199		}
200
201		/**
202		* Since solr cell does not allow to query the supported mimetypes, we return a list of known supported mimetypes here.
203		*
204		* @return array
205		*/
206	1	public function getSupportedMimeTypes(): array
207		{
208	1	$mapping = [
209		'application/epub+zip' => ['epub'],
210		'application/gzip' => ['gz', 'tgz'],
211		'application/msword' => ['doc'],
212		'application/pdf' => ['pdf'],
213		'application/rtf' => ['rtf'],
214		'application/vnd.ms-excel' => ['xsl'],
215		'application/vnd.ms-outlook' => ['msg'],
216		'application/vnd.oasis.opendocument.formula' => ['odf'],
217		'application/vnd.oasis.opendocument.text' => ['odt'],
218		'application/vnd.openxmlformats-officedocument.presentationml.presentation' => ['pptx'],
219		'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet' => ['xlsx'],
220		'application/vnd.openxmlformats-officedocument.wordprocessingml.document' => ['docx'],
221		'application/vnd.sun.xml.writer' => ['sxw'],
222		'application/zip' => ['zip'],
223		'application/x-midi' => ['mid'],
224		'application/xml' => ['xml'],
225		'audio/aiff' => ['aif', 'aiff'],
226		'audio/basic' => ['au'],
227		'audio/midi' => ['mid'],
228		'audio/mpeg3' => ['mp3'],
229		'audio/mpeg' => ['mp3'],
230		'audio/wav' => ['wav'],
231		'audio/x-mpeg-3' => ['mp3'],
232		'audio/x-wav' => ['wav'],
233		'image/bmp' => ['bmp'],
234		'image/gif' => ['gif'],
235		'image/jpeg' => ['jpg', 'jpeg'],
236		'image/png' => ['png'],
237		'image/svg+xml' => ['svg'],
238		'image/tiff' => ['tif', 'tiff'],
239		'text/html' => ['html', 'htm'],
240		'text/plain' => ['txt'],
241		'text/xml' => ['xml'],
242		'video/mpeg' => ['mp3'],
243		'video/x-mpeg' => ['mp3'],
244		];
245
246	1	return array_keys($mapping);
247		}
248
249		/**
250		* The service is available when the solr server is reachable.
251		*
252		* @return bool
253		*/
254	1	public function isAvailable(): bool
255		{
256	1	return $this->solrConnection->getWriteService()->ping();
257		}
258		}
259

TYPO3-Solr / ext-tika

Push — release-11.0.x ( 910681...ed160c )

SolrCellService::extractText() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like