Pinyin::format() - Code Metrics - Inspection of "update fork" - overtrue/pinyin - Measure and Improve Code Quality continuously with Scrutinizer

Test Failed

Pull Request — master (#136)

unknown

created 2019-04-22 09:45 UTC

Pinyin::format() A

↳ Parent: Pinyin

Complexity

Conditions	4
Paths	4

Size

Total Lines

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	9
CRAP Score	4

Importance

Changes

Metric	Value
dl	0
loc	18
ccs	9
cts	9
cp	1
rs	9.6666
c	0
b	0
f	0
cc	4
nc	4
nop	2
crap	4

<?php

/*
 * This file is part of the overtrue/pinyin.
 *
 * (c) overtrue <[email protected]>
 *
 * This source file is subject to the MIT license that is bundled
 * with this source code in the file LICENSE.
 */

namespace Overtrue\Pinyin;

use InvalidArgumentException;

/*
 * Chinese to pinyin translator.
 *
 * @author    overtrue <[email protected]>
 * @copyright 2015 overtrue <[email protected]>
 *
 * @link      https://github.com/overtrue/pinyin
 * @link      http://overtrue.me
 */

define('PINYIN_NONE', 'none');
define('PINYIN_ASCII', 'ascii');
define('PINYIN_UNICODE', 'unicode');

class Pinyin
{
    const NONE = 'none';

    const ASCII = 'ascii';

    const UNICODE = 'unicode';

    /**
     * Dict loader.
     *
     * @var \Overtrue\Pinyin\DictLoaderInterface
     */
    protected $loader;

    /**
     * Punctuations map.
     *
     * @var array
     */
    protected $punctuations = array(
        '，' => ',',
        '。' => '.',
        '！' => '!',
        '？' => '?',
        '：' => ':',
        '“' => '"',
        '”' => '"',
        '‘' => "'",
        '’' => "'",
    );

    /**
     * Constructor.
     *
     * @param string $loaderName
     */
    public function __construct($loaderName = null)
    {
        $this->loader = $loaderName ?: 'Overtrue\\Pinyin\\FileDictLoader';
    }

    /**
     * Convert string to pinyin.
     *
     * @param string $string
     * @param string $option
     *
     * @return array
     */
    public function convert($string, $option = self::NONE)
    {
        $pinyin = $this->romanize($string);

        return $this->splitWords($pinyin, $option);
    }

    /**
     * Convert string (person name) to pinyin.
     *
     * @param string $stringName
     * @param string $option
     *
     * @return array
     */
    public function name($stringName, $option = self::NONE)
    {
        $pinyin = $this->romanize($stringName, true);

        return $this->splitWords($pinyin, $option);
    }

    /**
     * Return a pinyin permalink from string.
     *
     * @param string $string
     * @param string $delimiter
     *
     * @return string
     */
    public function permalink($string, $delimiter = '-')
    {
        if (!in_array($delimiter, array('_', '-', '.', ''), true)) {
            throw new InvalidArgumentException("Delimiter must be one of: '_', '-', '', '.'.");
        }

        return implode($delimiter, $this->convert($string, false));
function acceptsInteger($int) { }

$x = '123'; // string "123"

// Instead of
acceptsInteger($x);

// we recommend to use
acceptsInteger((integer) $x);
    }

    /**
     * Return first letters.
     *
     * @param string $string
     * @param string $delimiter
     *
     * @return string
     */
    public function abbr($string, $delimiter = '')
    {
        return implode($delimiter, array_map(function ($pinyin) {
            return $pinyin[0];
        }, $this->convert($string, false)));
function acceptsInteger($int) { }

$x = '123'; // string "123"

// Instead of
acceptsInteger($x);

// we recommend to use
acceptsInteger((integer) $x);
    }

    /**
     * Chinese phrase to pinyin.
     *
     * @param string $string
     * @param string $delimiter
     * @param string $option
     *
     * @return string
     */
    public function phrase($string, $delimiter = ' ', $option = self::NONE)
    {
        return implode($delimiter, $this->convert($string, $option));
    }

    /**
     * Chinese to pinyin sentense.
     *
     * @param string $sentence
     * @param bool   $withTone
     *
     * @return string
     */
    public function sentence($sentence, $withTone = false)
    {
        $marks = array_keys($this->punctuations);
        $punctuationsRegex = preg_quote(implode(array_merge($marks, $this->punctuations)), '/');
        $regex = '/[^üāēīōūǖáéíóúǘǎěǐǒǔǚàèìòùǜɑa-z0-9'.$punctuationsRegex.'\s_]+/iu';

        $pinyin = preg_replace($regex, '', $this->romanize($sentence));

        $punctuations = array_merge($this->punctuations, array("\t" => ' ', '  ' => ' '));
        $pinyin = trim(str_replace(array_keys($punctuations), $punctuations, $pinyin));

        return $withTone ? $pinyin : $this->format($pinyin, false);
    }

    /**
     * Loader setter.
     *
     * @param \Overtrue\Pinyin\DictLoaderInterface $loader
     *
     * @return $this
     */
    public function setLoader(DictLoaderInterface $loader)
    {
        $this->loader = $loader;

        return $this;
    }

    /**
     * Return dict loader,.
     *
     * @return \Overtrue\Pinyin\DictLoaderInterface
     */
    public function getLoader()
    {
        if (!($this->loader instanceof DictLoaderInterface)) {
            $dataDir = dirname(__DIR__).'/data/';

            $loaderName = $this->loader;
            $this->loader = new $loaderName($dataDir);
        }

        return $this->loader;
    }

    /**
     * Preprocess.
     *
     * @param string $string
     *
     * @return string
     */
    protected function prepare($string)
    {
        $string = preg_replace_callback('~[a-z0-9_-]+~i', function ($matches) {
            return "\t".$matches[0];
        }, $string);

        return preg_replace("~[^\p{Han}\p{P}\p{Z}\p{M}\p{N}\p{L}\t]~u", '', $string);
    }

    /**
     * Convert Chinese to pinyin.
     *
     * @param string $string
     * @param bool   $isName
     *
     * @return string
     */
    protected function romanize($string, $isName = false)
    {
        $string = $this->prepare($string);

        $dictLoader = $this->getLoader();

        if ($isName) {
            $string = $this->convertSurname($string, $dictLoader);
        }

        $dictLoader->map(function ($dictionary) use (&$string) {
            $string = strtr($string, $dictionary);
        });

        return $string;
    }

    /**
     * Convert Chinese Surname to pinyin.
     *
     * @param string                               $string
     * @param \Overtrue\Pinyin\DictLoaderInterface $dictLoader
     *
     * @return string
     */
    protected function convertSurname($string, $dictLoader)
    {
        $dictLoader->mapSurname(function ($dictionary) use (&$string) {
            foreach ($dictionary as $surname => $pinyin) {
                if (0 === strpos($string, $surname)) {
                    $string = $pinyin.mb_substr($string, mb_strlen($surname, 'UTF-8'), mb_strlen($string, 'UTF-8') - 1, 'UTF-8');

                    break;
                }
            }
        });

        return $string;
    }

    /**
     * Split pinyin string to words.
     *
     * @param string $pinyin
     * @param string $option
     *
     * @return array
     */
    public function splitWords($pinyin, $option)
    {
        $split = array_filter(preg_split('/[^üāēīōūǖáéíóúǘǎěǐǒǔǚàèìòùǜɑa-z\d]+/iu', $pinyin));

        if (self::UNICODE !== $option) {
            foreach ($split as $index => $pinyin) {
                $split[$index] = $this->format($pinyin, self::ASCII === $option);
            }
        }

        return array_values($split);
    }

    /**
     * Format.
     *
     * @param string $pinyin
     * @param bool   $tone
     *
     * @return string
     */
    protected function format($pinyin, $tone = false)
    {
        $replacements = array(
            'üē' => array('ue', 1), 'üé' => array('ue', 2), 'üě' => array('ue', 3), 'üè' => array('ue', 4),
            'ā' => array('a', 1), 'ē' => array('e', 1), 'ī' => array('i', 1), 'ō' => array('o', 1), 'ū' => array('u', 1), 'ǖ' => array('v', 1),
            'á' => array('a', 2), 'é' => array('e', 2), 'í' => array('i', 2), 'ó' => array('o', 2), 'ú' => array('u', 2), 'ǘ' => array('v', 2),
            'ǎ' => array('a', 3), 'ě' => array('e', 3), 'ǐ' => array('i', 3), 'ǒ' => array('o', 3), 'ǔ' => array('u', 3), 'ǚ' => array('v', 3),
            'à' => array('a', 4), 'è' => array('e', 4), 'ì' => array('i', 4), 'ò' => array('o', 4), 'ù' => array('u', 4), 'ǜ' => array('v', 4),
        );

        foreach ($replacements as $unicde => $replacement) {
            if (false !== strpos($pinyin, $unicde)) {
                $pinyin = str_replace($unicde, $replacement[0], $pinyin).($tone ? $replacement[1] : '');
            }
        }

        return $pinyin;
    }
}


1		<?php
2
3		/*
4		* This file is part of the overtrue/pinyin.
5		*
6		* (c) overtrue <[email protected]>
7		*
8		* This source file is subject to the MIT license that is bundled
9		* with this source code in the file LICENSE.
10		*/
11
12		namespace Overtrue\Pinyin;
13
14		use InvalidArgumentException;
15
16		/*
17		* Chinese to pinyin translator.
18		*
19		* @author overtrue <[email protected]>
20		* @copyright 2015 overtrue <[email protected]>
21		*
22		* @link https://github.com/overtrue/pinyin
23	1	* @link http://overtrue.me
24	1	*/
25	1
26		define('PINYIN_NONE', 'none');
27		define('PINYIN_ASCII', 'ascii');
28		define('PINYIN_UNICODE', 'unicode');
29
30		class Pinyin
31		{
32		const NONE = 'none';
33
34		const ASCII = 'ascii';
35
36		const UNICODE = 'unicode';
37
38		/**
39		* Dict loader.
40		*
41		* @var \Overtrue\Pinyin\DictLoaderInterface
42		*/
43		protected $loader;
44
45		/**
46		* Punctuations map.
47		*
48		* @var array
49		*/
50		protected $punctuations = array(
51		'，' => ',',
52		'。' => '.',
53		'！' => '!',
54		'？' => '?',
55		'：' => ':',
56		'“' => '"',
57		'”' => '"',
58		'‘' => "'",
59		'’' => "'",
60		);
61
62	21	/**
63		* Constructor.
64	21	*
65	21	* @param string $loaderName
66		*/
67		public function __construct($loaderName = null)
68		{
69		$this->loader = $loaderName ?: 'Overtrue\\Pinyin\\FileDictLoader';
70		}
71
72		/**
73		* Convert string to pinyin.
74		*
75	10	* @param string $string
76		* @param string $option
77	10	*
78		* @return array
79	10	*/
80		public function convert($string, $option = self::NONE)
81		{
82		$pinyin = $this->romanize($string);
83
84		return $this->splitWords($pinyin, $option);
85		}
86
87		/**
88		* Convert string (person name) to pinyin.
89		*
90	2	* @param string $stringName
91		* @param string $option
92	2	*
93		* @return array
94	2	*/
95		public function name($stringName, $option = self::NONE)
96		{
97		$pinyin = $this->romanize($stringName, true);
98
99		return $this->splitWords($pinyin, $option);
100		}
101
102		/**
103		* Return a pinyin permalink from string.
104		*
105	2	* @param string $string
106		* @param string $delimiter
107	2	*
108	2	* @return string
109		*/
110		public function permalink($string, $delimiter = '-')
111	2	{
112		if (!in_array($delimiter, array('_', '-', '.', ''), true)) {
113		throw new InvalidArgumentException("Delimiter must be one of: '_', '-', '', '.'.");
114		}
115
116		return implode($delimiter, $this->convert($string, false));
		0 ignored issues – show Documentation introduced 2016-04-25 13:14 UTC by Report Bug Copy Issue Report `false` is of type `boolean`, but the function expects a `string`. It seems like the type of the argument is not accepted by the function/method which you are calling. In some cases, in particular if PHP’s automatic type-juggling kicks in this might be fine. In other cases, however this might be a bug. We suggest to add an explicit type cast like in the following example: function acceptsInteger($int) { } $x = '123'; // string "123" // Instead of acceptsInteger($x); // we recommend to use acceptsInteger((integer) $x); Loading history...
117		}
118
119		/**
120		* Return first letters.
121		*
122	2	* @param string $string
123		* @param string $delimiter
124		*
125	2	* @return string
126	2	*/
127		public function abbr($string, $delimiter = '')
128		{
129		return implode($delimiter, array_map(function ($pinyin) {
130		return $pinyin[0];
131		}, $this->convert($string, false)));
		0 ignored issues – show Documentation introduced 2016-04-25 13:14 UTC by Report Bug Copy Issue Report `false` is of type `boolean`, but the function expects a `string`. It seems like the type of the argument is not accepted by the function/method which you are calling. In some cases, in particular if PHP’s automatic type-juggling kicks in this might be fine. In other cases, however this might be a bug. We suggest to add an explicit type cast like in the following example: function acceptsInteger($int) { } $x = '123'; // string "123" // Instead of acceptsInteger($x); // we recommend to use acceptsInteger((integer) $x); Loading history...
132		}
133
134		/**
135		* Chinese phrase to pinyin.
136		*
137		* @param string $string
138	2	* @param string $delimiter
139		* @param string $option
140	2	*
141		* @return string
142		*/
143		public function phrase($string, $delimiter = ' ', $option = self::NONE)
144		{
145		return implode($delimiter, $this->convert($string, $option));
146		}
147
148		/**
149		* Chinese to pinyin sentense.
150		*
151	9	* @param string $sentence
152		* @param bool $withTone
153	9	*
154	9	* @return string
155	9	*/
156		public function sentence($sentence, $withTone = false)
157	9	{
158		$marks = array_keys($this->punctuations);
159	9	$punctuationsRegex = preg_quote(implode(array_merge($marks, $this->punctuations)), '/');
160	9	$regex = '/[^üāēīōūǖáéíóúǘǎěǐǒǔǚàèìòùǜɑa-z0-9'.$punctuationsRegex.'\s_]+/iu';
161
162	9	$pinyin = preg_replace($regex, '', $this->romanize($sentence));
163
164		$punctuations = array_merge($this->punctuations, array("\t" => ' ', ' ' => ' '));
165		$pinyin = trim(str_replace(array_keys($punctuations), $punctuations, $pinyin));
166
167		return $withTone ? $pinyin : $this->format($pinyin, false);
168		}
169
170		/**
171		* Loader setter.
172	1	*
173		* @param \Overtrue\Pinyin\DictLoaderInterface $loader
174	1	*
175		* @return $this
176	1	*/
177		public function setLoader(DictLoaderInterface $loader)
178		{
179		$this->loader = $loader;
180
181		return $this;
182		}
183
184	21	/**
185		* Return dict loader,.
186	21	*
187	20	* @return \Overtrue\Pinyin\DictLoaderInterface
188		*/
189	20	public function getLoader()
190	20	{
191	20	if (!($this->loader instanceof DictLoaderInterface)) {
192		$dataDir = dirname(__DIR__).'/data/';
193	21
194		$loaderName = $this->loader;
195		$this->loader = new $loaderName($dataDir);
196		}
197
198		return $this->loader;
199		}
200
201		/**
202		* Preprocess.
203	21	*
204		* @param string $string
205		*
206	8	* @return string
207	21	*/
208		protected function prepare($string)
209	21	{
210		$string = preg_replace_callback('~[a-z0-9_-]+~i', function ($matches) {
211		return "\t".$matches[0];
212		}, $string);
213
214		return preg_replace("~[^\p{Han}\p{P}\p{Z}\p{M}\p{N}\p{L}\t]~u", '', $string);
215		}
216
217		/**
218		* Convert Chinese to pinyin.
219		*
220	21	* @param string $string
221		* @param bool $isName
222	21	*
223		* @return string
224	21	*/
225		protected function romanize($string, $isName = false)
226	21	{
227	2	$string = $this->prepare($string);
228	2
229		$dictLoader = $this->getLoader();
230
231	21	if ($isName) {
232	21	$string = $this->convertSurname($string, $dictLoader);
233		}
234	21
235		$dictLoader->map(function ($dictionary) use (&$string) {
236		$string = strtr($string, $dictionary);
237		});
238
239		return $string;
240		}
241
242		/**
243		* Convert Chinese Surname to pinyin.
244		*
245		* @param string $string
246		* @param \Overtrue\Pinyin\DictLoaderInterface $dictLoader
247	2	*
248	2	* @return string
249	2	*/
250	2	protected function convertSurname($string, $dictLoader)
251	2	{
252		$dictLoader->mapSurname(function ($dictionary) use (&$string) {
253	2	foreach ($dictionary as $surname => $pinyin) {
254	2	if (0 === strpos($string, $surname)) {
255		$string = $pinyin.mb_substr($string, mb_strlen($surname, 'UTF-8'), mb_strlen($string, 'UTF-8') - 1, 'UTF-8');
256	2
257		break;
258		}
259		}
260		});
261
262		return $string;
263		}
264
265		/**
266		* Split pinyin string to words.
267	12	*
268		* @param string $pinyin
269	12	* @param string $option
270		*
271	12	* @return array
272	10	*/
273	10	public function splitWords($pinyin, $option)
274	10	{
275	10	$split = array_filter(preg_split('/[^üāēīōūǖáéíóúǘǎěǐǒǔǚàèìòùǜɑa-z\d]+/iu', $pinyin));
276
277	12	if (self::UNICODE !== $option) {
278		foreach ($split as $index => $pinyin) {
279		$split[$index] = $this->format($pinyin, self::ASCII === $option);
280		}
281		}
282
283		return array_values($split);
284		}
285
286		/**
287		* Format.
288	17	*
289		* @param string $pinyin
290		* @param bool $tone
291	17	*
292	17	* @return string
293	17	*/
294	17	protected function format($pinyin, $tone = false)
295	17	{
296	17	$replacements = array(
297		'üē' => array('ue', 1), 'üé' => array('ue', 2), 'üě' => array('ue', 3), 'üè' => array('ue', 4),
298	17	'ā' => array('a', 1), 'ē' => array('e', 1), 'ī' => array('i', 1), 'ō' => array('o', 1), 'ū' => array('u', 1), 'ǖ' => array('v', 1),
299	17	'á' => array('a', 2), 'é' => array('e', 2), 'í' => array('i', 2), 'ó' => array('o', 2), 'ú' => array('u', 2), 'ǘ' => array('v', 2),
300	16	'ǎ' => array('a', 3), 'ě' => array('e', 3), 'ǐ' => array('i', 3), 'ǒ' => array('o', 3), 'ǔ' => array('u', 3), 'ǚ' => array('v', 3),
301	16	'à' => array('a', 4), 'è' => array('e', 4), 'ì' => array('i', 4), 'ò' => array('o', 4), 'ù' => array('u', 4), 'ǜ' => array('v', 4),
302	17	);
303
304	17	foreach ($replacements as $unicde => $replacement) {
305		if (false !== strpos($pinyin, $unicde)) {
306		$pinyin = str_replace($unicde, $replacement[0], $pinyin).($tone ? $replacement[1] : '');
307		}
308		}
309
310		return $pinyin;
311		}
312		}
313

overtrue / pinyin

GitHub Access Token became invalid

Pull Request — master (#136)

Pinyin::format() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like