RisReader - 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 2021-03-02 13:38 UTC

RisReader B

↳ Parent: Project

Complexity

Total Complexity

Size/Duplication

Total Lines	288
Duplicated Lines	0 %

Importance

Changes	3
Bugs	0	Features	0

Metric	Value
eloc	187
dl	0
loc	288
rs	8.72
c	3
b	0
f	0
wmc	46

5 Methods

Rating	Name	Size	Complexity
A	tagToTagName()	3	1
D	readFile()	85	19
A	createRisRecords()	1	1
C	parseFile()	63	15
B	risRecordToXML()	29	10

How to fix Complexity

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

/*
 * 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 Symfony\Component\Serializer\Encoder\XmlEncoder;


/**
 * Class RisReader
 *
 * A reader for the Web of Science RIS-Format
 *
 * @package EWW\Dpf\Services\ImportExternalMetadata
 */
class RisReader
{
    public static $tagMap = [
        'FN' => 'File Name',
        'VR' => 'Version Number',
        'PT' => 'Publication Type',
        'AU' => 'Author',
        'AF' => 'Author Full Name',
        'BA' => 'Book Author',
        'BF' => 'Book Author Full Name',
        'CA' => 'Group Author',
        'GP' => 'Book Group Author',
        'BE' => 'Editor',
        'TI' => 'Document Title',
        'SO' => 'Publication Name',
        'SE' => 'Book Series Title',
        'BS' => 'Book Series Subtitle',
        'LA' => 'Language',
        'DT' => 'Document Type',
        'CT' => 'Conference Title',
        'CY' => 'Conference Date',
        'CL' => 'Conference Location',
        'SP' => 'Conference Sponsors',
        'HO' => 'Conference Host',
        'DE' => 'Author Keywords',
        'ID' => 'Keywords Plus',
        'AB' => 'Abstract',
        'C1' => 'Author Address',
        'RP' => 'Reprint Address',
        'EM' => 'E-mail Address',
        'RI' => 'ResearcherID Number',
        'OI' => 'ORCID Identifier',
        'FU' => 'Funding Agency and Grant Number',
        'FX' => 'Funding Text',
        'CR' => 'Cited References',
        'NR' => 'Cited Reference Count',
        'TC' => 'WoS Times Cited Count',
        'Z9' => 'Total Times Cited Count',
        'U1' => 'Usage Count las 180 days',
        'U2' => 'Usage Count since 2013',
        'PU' => 'Publisher',
        'PI' => 'Publisher City',
        'PA' => 'Publisher Address',
        'SN' => 'ISSN',
        'EI' => 'eISSN',
        'BN' => 'ISBN',
        'J9' => 'Character-29 Source Abbreviation',
        'JI' => 'ISO Source Abbreviation',
        'PD' => 'Publication Date',
        'PY' => 'Year Published',
        'VL' => 'Volume',
        'IS' => 'Issue',
        'SI' => 'Special Issue',
        'PN' => 'Part Number',
        'SU' => 'Supplement',
        'MA' => 'Meeting Abstract',
        'BP' => 'Beginning Page',
        'EP' => 'Ending Page',
        'AR' => 'Article Number',
        'DI' => 'DOI',
        'D2' => 'Book DOI',
        'EA' => 'Early access date',
        'EY' => 'Early access year',
        'PG' => 'Page Count',
        'P2' => 'Chapter Count',
        'WC' => 'WoS Categories',
        'SC' => 'Research Areas',
        'GA' => 'Document Delivery Number',
        'PM' => 'PubMed ID',
        'UT' => 'Accession Number',
        'OA' => 'Open Access Indicator',
        'HP' => 'ESI Hot Paper',
        'HC' => 'ESI Highly Cited Paper',
        'DA' => 'Date generated',
        'ER' => 'End of Record',
        'EF' => 'End of File'
    ];


    public static $publicationTypes = [
        'J' => 'Journal',
        'B' => 'Book',
        'S' => 'Series',
        'P' => 'Patent'
    ];

    /**
     * Gets the full tag name
     *
     * @param string $tag
     */
    public static function tagToTagName($tag)
    {
        return str_replace(" ", "-", strtolower(self::$tagMap[$tag]));
    }

    /**
     * @param string $filePath
     * @return array
     */
    protected function readFile($filePath, $contentOnly = false)
    {
        $separator = "\r\n";

        if ($contentOnly) {
            $line = strtok($filePath, $separator);
        } else {
            $flags = FILE_SKIP_EMPTY_LINES | FILE_TEXT;
            $lines = file($filePath, $flags);
        }

        $currentTag = '';
        $risRecords = [];
        $risRecord = [];
        $recordIndex = 0;

        if ($contentOnly) {

            while ($line !== false) {

                if (mb_detect_encoding($line) == 'UTF-8') {
                    $line = utf8_decode($line);
                    if (strpos($line, '?') === 0) {
                        $line = substr($line, 1);
                    }
                }

                $tempTag = trim(substr($line, 0, 2));
                if ($tempTag == 'EF') {
                    // End of file
                    break;
                }

                if ($tempTag == 'ER') {
                    $risRecords[$recordIndex] = $risRecord;
                    $risRecord = [];
                    $recordIndex += 1;
                } else {
                    if ($tempTag) {
                        $currentTag = $tempTag;
                    }

                    $line = substr($line, 2);

                    if ($currentTag && array_key_exists($currentTag, self::$tagMap)) {
                        $risRecord[$currentTag][] = trim($line);
                    }
                }
                $line = strtok($separator);
            }

        } else {
            foreach($lines as $line) {


                if (mb_detect_encoding($line) == 'UTF-8') {
                    $line = utf8_decode($line);
                    if (strpos($line, '?') === 0) {
                        $line = substr($line, 1);
                    }
                }

                $tempTag = trim(substr($line, 0, 2));
                if ($tempTag == 'EF') {
                    // End of file
                    break;
                }

                if ($tempTag == 'ER') {
                    $risRecords[$recordIndex] = $risRecord;
                    $risRecord = [];
                    $recordIndex += 1;
                } else {
                    if ($tempTag) {
                        $currentTag = $tempTag;
                    }

                    $line = substr($line, 2);

                    if ($currentTag && array_key_exists($currentTag, self::$tagMap)) {
                        $risRecord[$currentTag][] = trim($line);
                    }
                }
            }
        }

        return $risRecords;
    }

    public function createRisRecords() {

    }

    public function parseFile($filePath, $contentOnly = false)
    {
        $risRecords = $this->readFile($filePath, $contentOnly);
        $risEntries = [];

        foreach ($risRecords as $risRecord) {

            $risEntry = [];

            foreach ($risRecord as $tag => $risFieldValues) {

                if (in_array($tag, ['AF','AU','BA','BF','CA','GP','BE'])) {
                    // Authors
                    foreach ($risFieldValues as $fieldValue) {

                        list($family, $given, $suffix) = array_map('trim', explode(',', $fieldValue));

                        $affiliations = [];
                        if ($tag == 'AF') {
                            if (array_key_exists('C1', $risRecord)) {
                                $c1 = $risRecord['C1'];
                                foreach ($c1 as $affiliation) {
                                    if (
                                        preg_match(
                                            "/^\[.*?(".$fieldValue.").*?\](.*)/u", trim($affiliation),
                                            $matches
                                        )
                                    ) {
                                        $affiliations[] = $matches[2];
                                    }
                                }
                            }
                        }

                        if ($family || $given || $suffix || $affiliations) {
                            $risEntry[$tag][] = [
                                'family' => $family,
                                'given' => $given,
                                'suffix' => $suffix,
                                'affiliation' => $affiliations
                            ];
                        }
                    }
                } else {
                    $value = implode(" ", $risFieldValues);

                    if ($tag == 'PT') {
                        if (array_key_exists($value, self::$publicationTypes)) {
                            $value = strtolower(self::$publicationTypes[$value]);
                        } else {
                            $value = 'unknown';
                        }
                    }

                    $risEntry[$tag] = $value;
                }

            }

            $risEntries[] = $risEntry;
        }

        return $risEntries;
    }

    /**
     * @param array $risRecord
     */
    public function risRecordToXML($risRecord)
    {
        $encoder = new XmlEncoder();
        $record = [];
        foreach ($risRecord as $tag => $fieldValues) {
            switch ($tag) {
                case 'AF':
                    $record[self::tagToTagName('AU')] = $fieldValues;
                    break;
                case 'BF':
                    $record[self::tagToTagName('BA')] = $fieldValues;
                    break;
                case 'AU':
                    if (!array_key_exists('AF', $risRecord) || empty($risRecord['AF'])) {
                        $record[self::tagToTagName($tag)] = $fieldValues;
                    }
                    break;
                case 'BA':
                    if (!array_key_exists('BF', $risRecord) || empty($risRecord['BF'])) {
                        $record[self::tagToTagName($tag)] = $fieldValues;
                    }
                    break;
                default:
                    $record[self::tagToTagName($tag)] = $fieldValues;
                    break;
            }
        }

        return $encoder->encode($record, 'xml');
    }

}

1			<?php
2			namespace EWW\Dpf\Services\ImportExternalMetadata;
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 Symfony\Component\Serializer\Encoder\XmlEncoder;
18
19
20			/**
21			* Class RisReader
22			*
23			* A reader for the Web of Science RIS-Format
24			*
25			* @package EWW\Dpf\Services\ImportExternalMetadata
26			*/
27			class RisReader
28			{
29			public static $tagMap = [
30			'FN' => 'File Name',
31			'VR' => 'Version Number',
32			'PT' => 'Publication Type',
33			'AU' => 'Author',
34			'AF' => 'Author Full Name',
35			'BA' => 'Book Author',
36			'BF' => 'Book Author Full Name',
37			'CA' => 'Group Author',
38			'GP' => 'Book Group Author',
39			'BE' => 'Editor',
40			'TI' => 'Document Title',
41			'SO' => 'Publication Name',
42			'SE' => 'Book Series Title',
43			'BS' => 'Book Series Subtitle',
44			'LA' => 'Language',
45			'DT' => 'Document Type',
46			'CT' => 'Conference Title',
47			'CY' => 'Conference Date',
48			'CL' => 'Conference Location',
49			'SP' => 'Conference Sponsors',
50			'HO' => 'Conference Host',
51			'DE' => 'Author Keywords',
52			'ID' => 'Keywords Plus',
53			'AB' => 'Abstract',
54			'C1' => 'Author Address',
55			'RP' => 'Reprint Address',
56			'EM' => 'E-mail Address',
57			'RI' => 'ResearcherID Number',
58			'OI' => 'ORCID Identifier',
59			'FU' => 'Funding Agency and Grant Number',
60			'FX' => 'Funding Text',
61			'CR' => 'Cited References',
62			'NR' => 'Cited Reference Count',
63			'TC' => 'WoS Times Cited Count',
64			'Z9' => 'Total Times Cited Count',
65			'U1' => 'Usage Count las 180 days',
66			'U2' => 'Usage Count since 2013',
67			'PU' => 'Publisher',
68			'PI' => 'Publisher City',
69			'PA' => 'Publisher Address',
70			'SN' => 'ISSN',
71			'EI' => 'eISSN',
72			'BN' => 'ISBN',
73			'J9' => 'Character-29 Source Abbreviation',
74			'JI' => 'ISO Source Abbreviation',
75			'PD' => 'Publication Date',
76			'PY' => 'Year Published',
77			'VL' => 'Volume',
78			'IS' => 'Issue',
79			'SI' => 'Special Issue',
80			'PN' => 'Part Number',
81			'SU' => 'Supplement',
82			'MA' => 'Meeting Abstract',
83			'BP' => 'Beginning Page',
84			'EP' => 'Ending Page',
85			'AR' => 'Article Number',
86			'DI' => 'DOI',
87			'D2' => 'Book DOI',
88			'EA' => 'Early access date',
89			'EY' => 'Early access year',
90			'PG' => 'Page Count',
91			'P2' => 'Chapter Count',
92			'WC' => 'WoS Categories',
93			'SC' => 'Research Areas',
94			'GA' => 'Document Delivery Number',
95			'PM' => 'PubMed ID',
96			'UT' => 'Accession Number',
97			'OA' => 'Open Access Indicator',
98			'HP' => 'ESI Hot Paper',
99			'HC' => 'ESI Highly Cited Paper',
100			'DA' => 'Date generated',
101			'ER' => 'End of Record',
102			'EF' => 'End of File'
103			];
104
105
106			public static $publicationTypes = [
107			'J' => 'Journal',
108			'B' => 'Book',
109			'S' => 'Series',
110			'P' => 'Patent'
111			];
112
113			/**
114			* Gets the full tag name
115			*
116			* @param string $tag
117			*/
118			public static function tagToTagName($tag)
119			{
120			return str_replace(" ", "-", strtolower(self::$tagMap[$tag]));
121			}
122
123			/**
124			* @param string $filePath
125			* @return array
126			*/
127			protected function readFile($filePath, $contentOnly = false)
128			{
129			$separator = "\r\n";
130
131			if ($contentOnly) {
132			$line = strtok($filePath, $separator);
133			} else {
134			$flags = FILE_SKIP_EMPTY_LINES \| FILE_TEXT;
135			$lines = file($filePath, $flags);
136			}
137
138			$currentTag = '';
139			$risRecords = [];
140			$risRecord = [];
141			$recordIndex = 0;
142
143			if ($contentOnly) {
144
145			while ($line !== false) {
			0 ignored issues – show Comprehensibility Best Practice introduced 2020-12-15 14:55 UTC by Report Bug Copy Issue Report The variable `$line` does not seem to be defined for all execution paths leading up to this point. Loading history...
146			if (mb_detect_encoding($line) == 'UTF-8') {
147			$line = utf8_decode($line);
148			if (strpos($line, '?') === 0) {
149			$line = substr($line, 1);
150			}
151			}
152
153			$tempTag = trim(substr($line, 0, 2));
154			if ($tempTag == 'EF') {
155			// End of file
156			break;
157			}
158
159			if ($tempTag == 'ER') {
160			$risRecords[$recordIndex] = $risRecord;
161			$risRecord = [];
162			$recordIndex += 1;
163			} else {
164			if ($tempTag) {
165			$currentTag = $tempTag;
166			}
167
168			$line = substr($line, 2);
169
170			if ($currentTag && array_key_exists($currentTag, self::$tagMap)) {
171			$risRecord[$currentTag][] = trim($line);
172			}
173			}
174			$line = strtok($separator);
175			}
176
177			} else {
178			foreach($lines as $line) {
			0 ignored issues – show Comprehensibility Best Practice introduced 2020-12-15 14:55 UTC by Report Bug Copy Issue Report The variable `$lines` does not seem to be defined for all execution paths leading up to this point. Loading history...
179
180			if (mb_detect_encoding($line) == 'UTF-8') {
181			$line = utf8_decode($line);
182			if (strpos($line, '?') === 0) {
183			$line = substr($line, 1);
184			}
185			}
186
187			$tempTag = trim(substr($line, 0, 2));
188			if ($tempTag == 'EF') {
189			// End of file
190			break;
191			}
192
193			if ($tempTag == 'ER') {
194			$risRecords[$recordIndex] = $risRecord;
195			$risRecord = [];
196			$recordIndex += 1;
197			} else {
198			if ($tempTag) {
199			$currentTag = $tempTag;
200			}
201
202			$line = substr($line, 2);
203
204			if ($currentTag && array_key_exists($currentTag, self::$tagMap)) {
205			$risRecord[$currentTag][] = trim($line);
206			}
207			}
208			}
209			}
210
211			return $risRecords;
212			}
213
214			public function createRisRecords() {
215
216			}
217
218			public function parseFile($filePath, $contentOnly = false)
219			{
220			$risRecords = $this->readFile($filePath, $contentOnly);
221			$risEntries = [];
222
223			foreach ($risRecords as $risRecord) {
224
225			$risEntry = [];
226
227			foreach ($risRecord as $tag => $risFieldValues) {
228
229			if (in_array($tag, ['AF','AU','BA','BF','CA','GP','BE'])) {
230			// Authors
231			foreach ($risFieldValues as $fieldValue) {
232
233			list($family, $given, $suffix) = array_map('trim', explode(',', $fieldValue));
234
235			$affiliations = [];
236			if ($tag == 'AF') {
237			if (array_key_exists('C1', $risRecord)) {
238			$c1 = $risRecord['C1'];
239			foreach ($c1 as $affiliation) {
240			if (
241			preg_match(
242			"/^\[.?(".$fieldValue.").?\](.*)/u", trim($affiliation),
243			$matches
244			)
245			) {
246			$affiliations[] = $matches[2];
247			}
248			}
249			}
250			}
251
252			if ($family \|\| $given \|\| $suffix \|\| $affiliations) {
253			$risEntry[$tag][] = [
254			'family' => $family,
255			'given' => $given,
256			'suffix' => $suffix,
257			'affiliation' => $affiliations
258			];
259			}
260			}
261			} else {
262			$value = implode(" ", $risFieldValues);
263
264			if ($tag == 'PT') {
265			if (array_key_exists($value, self::$publicationTypes)) {
266			$value = strtolower(self::$publicationTypes[$value]);
267			} else {
268			$value = 'unknown';
269			}
270			}
271
272			$risEntry[$tag] = $value;
273			}
274
275			}
276
277			$risEntries[] = $risEntry;
278			}
279
280			return $risEntries;
281			}
282
283			/**
284			* @param array $risRecord
285			*/
286			public function risRecordToXML($risRecord)
287			{
288			$encoder = new XmlEncoder();
289			$record = [];
290			foreach ($risRecord as $tag => $fieldValues) {
291			switch ($tag) {
292			case 'AF':
293			$record[self::tagToTagName('AU')] = $fieldValues;
294			break;
295			case 'BF':
296			$record[self::tagToTagName('BA')] = $fieldValues;
297			break;
298			case 'AU':
299			if (!array_key_exists('AF', $risRecord) \|\| empty($risRecord['AF'])) {
300			$record[self::tagToTagName($tag)] = $fieldValues;
301			}
302			break;
303			case 'BA':
304			if (!array_key_exists('BF', $risRecord) \|\| empty($risRecord['BF'])) {
305			$record[self::tagToTagName($tag)] = $fieldValues;
306			}
307			break;
308			default:
309			$record[self::tagToTagName($tag)] = $fieldValues;
310			break;
311			}
312			}
313
314			return $encoder->encode($record, 'xml');
315			}
316
317			}

kitodo / kitodo-publication

Pull Request — master (#195)

RisReader B

Complexity

Size/Duplication

Importance

5 Methods

How to fix Complexity

Complex Class

Duplication Side-by-Side

Filter issues like