Pinyin::convert() - Code Metrics - Inspection of "add phrase" - overtrue/pinyin - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Pull Request — master (#79)

by xiaobei

created 2017-01-22 09:05 UTC

Pinyin::convert() A

↳ Parent: Pinyin

Complexity

Conditions	1
Paths	1

Size

Total Lines	6
Code Lines	3

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	3
CRAP Score	1

Importance

Changes

Metric	Value
dl	0
loc	6
c	0
b	0
f	0
ccs	3
cts	3
cp	1
rs	9.4285
cc	1
eloc	3
nc	1
nop	2
crap	1

<?php

/*
 * This file is part of the overtrue/pinyin.
 *
 * (c) 2016 overtrue <[email protected]>
 */

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, array_map(function ($pinyin) {
            return $pinyin;
        }, $this->convert($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)), '/');
        $regex = '/[^üāēīōūǖáéíóúǘǎěǐǒǔǚàèìòùǜa-z0-9'.$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 => $pinyin) {
                if (strpos($string, $surname) === 0) {
                    $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('/[^üāēīōūǖáéíóúǘǎěǐǒǔǚàèìòùǜa-z\d]+/iu', $pinyin));

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

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

        return $pinyin;
    }
}


1		<?php
2
3		/*
4		* This file is part of the overtrue/pinyin.
5		*
6		* (c) 2016 overtrue <[email protected]>
7		*/
8
9		namespace Overtrue\Pinyin;
10
11		use InvalidArgumentException;
12
13		/**
14		* Chinese to pinyin translator.
15		*
16		* @author overtrue <[email protected]>
17		* @copyright 2015 overtrue <[email protected]>
18		*
19		* @link https://github.com/overtrue/pinyin
20		* @link http://overtrue.me
21		*/
22
23	1	define('PINYIN_NONE', 'none');
24	1	define('PINYIN_ASCII', 'ascii');
25	1	define('PINYIN_UNICODE', 'unicode');
26
27		class Pinyin
28		{
29		const NONE = 'none';
30		const ASCII = 'ascii';
31		const UNICODE = 'unicode';
32
33		/**
34		* Dict loader.
35		*
36		* @var \Overtrue\Pinyin\DictLoaderInterface
37		*/
38		protected $loader;
39
40		/**
41		* Punctuations map.
42		*
43		* @var array
44		*/
45		protected $punctuations = array(
46		'，' => ',',
47		'。' => '.',
48		'！' => '!',
49		'？' => '?',
50		'：' => ':',
51		'“' => '"',
52		'”' => '"',
53		'‘' => "'",
54		'’' => "'",
55		);
56
57		/**
58		* Constructor.
59		*
60		* @param string $loaderName
61		*/
62	19	public function __construct($loaderName = null)
63		{
64	19	$this->loader = $loaderName ?: 'Overtrue\\Pinyin\\FileDictLoader';
		0 ignored issues – show Documentation Bug introduced 2016-11-17 05:39 UTC by Report Bug Copy Issue Report It seems like `$loaderName ?: 'Overtrue\\Pinyin\\FileDictLoader'` of type `string` is incompatible with the declared type `object<Overtrue\Pinyin\DictLoaderInterface>` of property `$loader`. Our type inference engine has found an assignment to a property that is incompatible with the declared type of that property. Either this assignment is in error or the assigned type should be added to the documentation/type hint for that property.. Loading history...
65	19	}
66
67		/**
68		* Convert string to pinyin.
69		*
70		* @param string $string
71		* @param string $option
72		*
73		* @return array
74		*/
75	8	public function convert($string, $option = self::NONE)
76		{
77	8	$pinyin = $this->romanize($string);
78
79	8	return $this->splitWords($pinyin, $option);
80		}
81
82		/**
83		* Convert string (person name) to pinyin.
84		*
85		* @param string $stringName
86		* @param string $option
87		*
88		* @return array
89		*/
90	2	public function name($stringName, $option = self::NONE)
91		{
92	2	$pinyin = $this->romanize($stringName, true);
93
94	2	return $this->splitWords($pinyin, $option);
95		}
96
97		/**
98		* Return a pinyin permalink from string.
99		*
100		* @param string $string
101		* @param string $delimiter
102		*
103		* @return string
104		*/
105	2	public function permalink($string, $delimiter = '-')
106		{
107	2	if (!in_array($delimiter, array('_', '-', '.', ''), true)) {
108	2	throw new InvalidArgumentException("Delimiter must be one of: '_', '-', '', '.'.");
109		}
110
111	2	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...
112		}
113
114		/**
115		* Return first letters.
116		*
117		* @param string $string
118		* @param string $delimiter
119		*
120		* @return string
121		*/
122	2	public function abbr($string, $delimiter = '')
123		{
124		return implode($delimiter, array_map(function ($pinyin) {
125	2	return $pinyin[0];
126	2	}, $this->convert($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...
127		}
128
129		/**
130		* Chinese phrase to pinyin.
131		*
132		* @param string $string
133		* @param string $delimiter
134		* @param string $option
135		*
136		* @return string
137	9	*/
138		public function phrase($string, $delimiter = ' ', $option = self::NONE) {
139	9	return implode($delimiter, array_map(function ($pinyin) {
140	9	return $pinyin;
141	9	}, $this->convert($string, $option)));
142		}
143	9
144		/**
145	9	* Chinese to pinyin sentense.
146	9	*
147		* @param string $sentence
148	9	* @param bool $withTone
149		*
150		* @return string
151		*/
152		public function sentence($sentence, $withTone = false)
153		{
154		$marks = array_keys($this->punctuations);
155		$punctuationsRegex = preg_quote(implode(array_merge($marks, $this->punctuations)), '/');
156		$regex = '/[^üāēīōūǖáéíóúǘǎěǐǒǔǚàèìòùǜa-z0-9'.$punctuationsRegex.'\s_]+/iu';
157
158	1	$pinyin = preg_replace($regex, '', $this->romanize($sentence));
159
160	1	$punctuations = array_merge($this->punctuations, array("\t" => ' ', ' ' => ' '));
161		$pinyin = trim(str_replace(array_keys($punctuations), $punctuations, $pinyin));
162	1
163		return $withTone ? $pinyin : $this->format($pinyin, false);
164		}
165
166		/**
167		* Loader setter.
168		*
169		* @param \Overtrue\Pinyin\DictLoaderInterface $loader
170	19	*
171		* @return $this
172	19	*/
173	18	public function setLoader(DictLoaderInterface $loader)
174		{
175	18	$this->loader = $loader;
176	18
177	18	return $this;
178		}
179	19
180		/**
181		* Return dict loader,.
182		*
183		* @return \Overtrue\Pinyin\DictLoaderInterface
184		*/
185		public function getLoader()
186		{
187		if (!($this->loader instanceof DictLoaderInterface)) {
188		$dataDir = dirname(__DIR__).'/data/';
189	19
190		$loaderName = $this->loader;
191		$this->loader = new $loaderName($dataDir);
192	8	}
193	19
194		return $this->loader;
195	19	}
196
197		/**
198		* Preprocess.
199		*
200		* @param string $string
201		*
202		* @return string
203		*/
204		protected function prepare($string)
205		{
206	19	$string = preg_replace_callback('~[a-z0-9_-]+~i', function ($matches) {
207		return "\t".$matches[0];
208	19	}, $string);
209
210	19	return preg_replace("~[^\p{Han}\p{P}\p{Z}\p{M}\p{N}\p{L}\t]~u", '', $string);
211		}
212	19
213	2	/**
214	2	* Convert Chinese to pinyin.
215		*
216		* @param string $string
217	19	* @param bool $isName
218	19	*
219		* @return string
220	19	*/
221		protected function romanize($string, $isName = false)
222		{
223		$string = $this->prepare($string);
224
225		$dictLoader = $this->getLoader();
226
227		if ($isName) {
228		$string = $this->convertSurname($string, $dictLoader);
229		}
230
231		$dictLoader->map(function ($dictionary) use (&$string) {
232		$string = strtr($string, $dictionary);
233	2	});
234	2
235	2	return $string;
236	2	}
237	2
238		/**
239	2	* Convert Chinese Surname to pinyin.
240	2	*
241		* @param string $string
242	2	* @param \Overtrue\Pinyin\DictLoaderInterface $dictLoader
243		*
244		* @return string
245		*/
246		protected function convertSurname($string, $dictLoader)
247		{
248		$dictLoader->mapSurname(function ($dictionary) use (&$string) {
249		foreach ($dictionary as $surname => $pinyin) {
250		if (strpos($string, $surname) === 0) {
251		$string = $pinyin.mb_substr($string, mb_strlen($surname, 'UTF-8'), mb_strlen($string, 'UTF-8') - 1, 'UTF-8');
252		break;
253	10	}
254		}
255	10	});
256
257	10	return $string;
258	8	}
259	8
260	8	/**
261	8	* Split pinyin string to words.
262		*
263	10	* @param string $pinyin
264		* @param string $option
265		*
266		* @return array
267		*/
268		public function splitWords($pinyin, $option)
269		{
270		$split = array_filter(preg_split('/[^üāēīōūǖáéíóúǘǎěǐǒǔǚàèìòùǜa-z\d]+/iu', $pinyin));
271
272		if ($option !== self::UNICODE) {
273		foreach ($split as $index => $pinyin) {
274	15	$split[$index] = $this->format($pinyin, $option === self::ASCII);
275		}
276		}
277	15
278	15	return array_values($split);
279	15	}
280	15
281	15	/**
282	15	* Format.
283		*
284	15	* @param string $pinyin
285	15	* @param bool $tone
286	14	*
287	14	* @return string
288	15	*/
289		protected function format($pinyin, $tone = false)
290	15	{
291		$replacements = array(
292		'üē' => array('ue', 1), 'üé' => array('ue', 2), 'üě' => array('ue', 3), 'üè' => array('ue', 4),
293		'ā' => array('a', 1), 'ē' => array('e', 1), 'ī' => array('i', 1), 'ō' => array('o', 1), 'ū' => array('u', 1), 'ǖ' => array('v', 1),
294		'á' => array('a', 2), 'é' => array('e', 2), 'í' => array('i', 2), 'ó' => array('o', 2), 'ú' => array('u', 2), 'ǘ' => array('v', 2),
295		'ǎ' => array('a', 3), 'ě' => array('e', 3), 'ǐ' => array('i', 3), 'ǒ' => array('o', 3), 'ǔ' => array('u', 3), 'ǚ' => array('v', 3),
296		'à' => array('a', 4), 'è' => array('e', 4), 'ì' => array('i', 4), 'ò' => array('o', 4), 'ù' => array('u', 4), 'ǜ' => array('v', 4),
297		);
298
299		foreach ($replacements as $unicde => $replacements) {
300		if (false !== strpos($pinyin, $unicde)) {
301		$pinyin = str_replace($unicde, $replacements[0], $pinyin).($tone ? $replacements[1] : '');
302		}
303		}
304
305		return $pinyin;
306		}
307		}
308

overtrue / pinyin

GitHub Access Token became invalid

Pull Request — master (#79)

Pinyin::convert() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like