SolrCellService::detectLanguageFromFile() - Code Metrics - TYPO3-Solr/ext-tika - Measure and Improve Code Quality continuously with Scrutinizer

SolrCellService::detectLanguageFromFile() A
last analyzed 2024-05-03 16:22 UTC

↳ Parent: SolrCellService

Complexity

Conditions	1
Paths	1

Size

Total Lines	6
Code Lines	3

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	0
CRAP Score	2

Importance

Changes

Metric	Value
cc	1
eloc	3
nc	1
nop	1
dl	0
loc	6
ccs	0
cts	4
cp	0
crap	2
rs	10
c	0
b	0
f	0

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

TYPO3-Solr / ext-tika

SolrCellService::detectLanguageFromFile() A last analyzed 2024-05-03 16:22 UTC

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like

SolrCellService::detectLanguageFromFile() A
last analyzed 2024-05-03 16:22 UTC