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