|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
namespace StringObject\Normalize; |
|
4
|
|
|
|
|
5
|
|
|
use StringObject\AnyStrObj; |
|
6
|
|
|
|
|
7
|
|
|
class ASCII |
|
8
|
|
|
{ |
|
9
|
|
|
// STATIC PROPERTIES |
|
10
|
|
|
protected static $asciimap = [ |
|
11
|
|
|
'a' => ['à', 'á', 'ả', 'ã', 'ạ', 'ă', 'ắ', 'ằ', 'ẳ', 'ẵ', 'ặ', 'â', 'ấ', |
|
12
|
|
|
'ầ', 'ẩ', 'ẫ', 'ậ', 'ā', 'ą', 'å', 'α', 'ά', 'ἀ', 'ἁ', 'ἂ', 'ἃ', |
|
13
|
|
|
'ἄ', 'ἅ', 'ἆ', 'ἇ', 'ᾀ', 'ᾁ', 'ᾂ', 'ᾃ', 'ᾄ', 'ᾅ', 'ᾆ', 'ᾇ', 'ὰ', |
|
14
|
|
|
'ά', 'ᾰ', 'ᾱ', 'ᾲ', 'ᾳ', 'ᾴ', 'ᾶ', 'ᾷ', 'а', 'أ'], |
|
15
|
|
|
'b' => ['б', 'β', 'Ъ', 'Ь', 'ب'], |
|
16
|
|
|
'c' => ['ç', 'ć', 'č', 'ĉ', 'ċ'], |
|
17
|
|
|
'd' => ['ď', 'ð', 'đ', 'ƌ', 'ȡ', 'ɖ', 'ɗ', 'ᵭ', 'ᶁ', 'ᶑ', 'д', 'δ', 'د', 'ض'], |
|
18
|
|
|
'e' => ['é', 'è', 'ẻ', 'ẽ', 'ẹ', 'ê', 'ế', 'ề', 'ể', 'ễ', 'ệ', 'ë', 'ē', |
|
19
|
|
|
'ę', 'ě', 'ĕ', 'ė', 'ε', 'έ', 'ἐ', 'ἑ', 'ἒ', 'ἓ', 'ἔ', 'ἕ', 'ὲ', |
|
20
|
|
|
'έ', 'е', 'ё', 'э', 'є', 'ə'], |
|
21
|
|
|
'f' => ['ф', 'φ', 'ف'], |
|
22
|
|
|
'g' => ['ĝ', 'ğ', 'ġ', 'ģ', 'г', 'ґ', 'γ', 'ج'], |
|
23
|
|
|
'h' => ['ĥ', 'ħ', 'η', 'ή', 'ح', 'ه'], |
|
24
|
|
|
'i' => ['í', 'ì', 'ỉ', 'ĩ', 'ị', 'î', 'ï', 'ī', 'ĭ', 'į', 'ı', 'ι', 'ί', |
|
25
|
|
|
'ϊ', 'ΐ', 'ἰ', 'ἱ', 'ἲ', 'ἳ', 'ἴ', 'ἵ', 'ἶ', 'ἷ', 'ὶ', 'ί', 'ῐ', |
|
26
|
|
|
'ῑ', 'ῒ', 'ΐ', 'ῖ', 'ῗ', 'і', 'ї', 'и'], |
|
27
|
|
|
'j' => ['ĵ', 'ј', 'Ј'], |
|
28
|
|
|
'k' => ['ķ', 'ĸ', 'к', 'κ', 'Ķ', 'ق', 'ك'], |
|
29
|
|
|
'l' => ['ł', 'ľ', 'ĺ', 'ļ', 'ŀ', 'л', 'λ', 'ل'], |
|
30
|
|
|
'm' => ['м', 'μ', 'م'], |
|
31
|
|
|
'n' => ['ñ', 'ń', 'ň', 'ņ', 'ʼn', 'ŋ', 'ν', 'н', 'ن'], |
|
32
|
|
|
'o' => ['ó', 'ò', 'ỏ', 'õ', 'ọ', 'ô', 'ố', 'ồ', 'ổ', 'ỗ', 'ộ', 'ơ', 'ớ', |
|
33
|
|
|
'ờ', 'ở', 'ỡ', 'ợ', 'ø', 'ō', 'ő', 'ŏ', 'ο', 'ὀ', 'ὁ', 'ὂ', 'ὃ', |
|
34
|
|
|
'ὄ', 'ὅ', 'ὸ', 'ό', 'о', 'و', 'θ'], |
|
35
|
|
|
'p' => ['п', 'π'], |
|
36
|
|
|
'r' => ['ŕ', 'ř', 'ŗ', 'р', 'ρ', 'ر'], |
|
37
|
|
|
's' => ['ś', 'š', 'ş', 'с', 'σ', 'ș', 'ς', 'س', 'ص'], |
|
38
|
|
|
't' => ['ť', 'ţ', 'т', 'τ', 'ț', 'ت', 'ط'], |
|
39
|
|
|
'u' => ['ú', 'ù', 'ủ', 'ũ', 'ụ', 'ư', 'ứ', 'ừ', 'ử', 'ữ', 'ự', 'û', 'ū', |
|
40
|
|
|
'ů', 'ű', 'ŭ', 'ų', 'µ', 'у'], |
|
41
|
|
|
'v' => ['в'], |
|
42
|
|
|
'w' => ['ŵ', 'ω', 'ώ'], |
|
43
|
|
|
'x' => ['χ'], |
|
44
|
|
|
'y' => ['ý', 'ỳ', 'ỷ', 'ỹ', 'ỵ', 'ÿ', 'ŷ', 'й', 'ы', 'υ', 'ϋ', 'ύ', 'ΰ', 'ي'], |
|
45
|
|
|
'z' => ['ź', 'ž', 'ż', 'з', 'ζ', 'ز'], |
|
46
|
|
|
'aa' => ['ع'], |
|
47
|
|
|
'ae' => ['ä', 'æ'], |
|
48
|
|
|
'ch' => ['ч'], |
|
49
|
|
|
'dj' => ['ђ', 'đ'], |
|
50
|
|
|
'dz' => ['џ'], |
|
51
|
|
|
'gh' => ['غ'], |
|
52
|
|
|
'kh' => ['х', 'خ'], |
|
53
|
|
|
'lj' => ['љ'], |
|
54
|
|
|
'nj' => ['њ'], |
|
55
|
|
|
'oe' => ['ö', 'œ'], |
|
56
|
|
|
'ps' => ['ψ'], |
|
57
|
|
|
'sh' => ['ш'], |
|
58
|
|
|
'shch' => ['щ'], |
|
59
|
|
|
'ss' => ['ß'], |
|
60
|
|
|
'th' => ['þ', 'ث', 'ذ', 'ظ'], |
|
61
|
|
|
'ts' => ['ц'], |
|
62
|
|
|
'ue' => ['ü'], |
|
63
|
|
|
'ya' => ['я'], |
|
64
|
|
|
'yu' => ['ю'], |
|
65
|
|
|
'zh' => ['ж'], |
|
66
|
|
|
'(c)' => ['©'], |
|
67
|
|
|
'A' => ['Á', 'À', 'Ả', 'Ã', 'Ạ', 'Ă', 'Ắ', 'Ằ', 'Ẳ', 'Ẵ', 'Ặ', 'Â', 'Ấ', |
|
68
|
|
|
'Ầ', 'Ẩ', 'Ẫ', 'Ậ', 'Å', 'Ā', 'Ą', 'Α', 'Ά', 'Ἀ', 'Ἁ', 'Ἂ', 'Ἃ', |
|
69
|
|
|
'Ἄ', 'Ἅ', 'Ἆ', 'Ἇ', 'ᾈ', 'ᾉ', 'ᾊ', 'ᾋ', 'ᾌ', 'ᾍ', 'ᾎ', 'ᾏ', 'Ᾰ', |
|
70
|
|
|
'Ᾱ', 'Ὰ', 'Ά', 'ᾼ', 'А'], |
|
71
|
|
|
'B' => ['Б', 'Β'], |
|
72
|
|
|
'C' => ['Ç', 'Ć', 'Č', 'Ĉ', 'Ċ'], |
|
73
|
|
|
'D' => ['Ď', 'Ð', 'Đ', 'Ɖ', 'Ɗ', 'Ƌ', 'ᴅ', 'ᴆ', 'Д', 'Δ'], |
|
74
|
|
|
'E' => ['É', 'È', 'Ẻ', 'Ẽ', 'Ẹ', 'Ê', 'Ế', 'Ề', 'Ể', 'Ễ', 'Ệ', 'Ë', 'Ē', |
|
75
|
|
|
'Ę', 'Ě', 'Ĕ', 'Ė', 'Ε', 'Έ', 'Ἐ', 'Ἑ', 'Ἒ', 'Ἓ', 'Ἔ', 'Ἕ', 'Έ', |
|
76
|
|
|
'Ὲ', 'Е', 'Ё', 'Э', 'Є', 'Ə'], |
|
77
|
|
|
'F' => ['Ф', 'Φ'], |
|
78
|
|
|
'G' => ['Ğ', 'Ġ', 'Ģ', 'Г', 'Ґ', 'Γ'], |
|
79
|
|
|
'H' => ['Η', 'Ή'], |
|
80
|
|
|
'I' => ['Í', 'Ì', 'Ỉ', 'Ĩ', 'Ị', 'Î', 'Ï', 'Ī', 'Ĭ', 'Į', 'İ', 'Ι', 'Ί', |
|
81
|
|
|
'Ϊ', 'Ἰ', 'Ἱ', 'Ἳ', 'Ἴ', 'Ἵ', 'Ἶ', 'Ἷ', 'Ῐ', 'Ῑ', 'Ὶ', 'Ί', 'И', |
|
82
|
|
|
'І', 'Ї'], |
|
83
|
|
|
'K' => ['К', 'Κ'], |
|
84
|
|
|
'L' => ['Ĺ', 'Ł', 'Л', 'Λ', 'Ļ'], |
|
85
|
|
|
'M' => ['М', 'Μ'], |
|
86
|
|
|
'N' => ['Ń', 'Ñ', 'Ň', 'Ņ', 'Ŋ', 'Н', 'Ν'], |
|
87
|
|
|
'O' => ['Ó', 'Ò', 'Ỏ', 'Õ', 'Ọ', 'Ô', 'Ố', 'Ồ', 'Ổ', 'Ỗ', 'Ộ', 'Ơ', 'Ớ', |
|
88
|
|
|
'Ờ', 'Ở', 'Ỡ', 'Ợ', 'Ø', 'Ō', 'Ő', 'Ŏ', 'Ο', 'Ό', 'Ὀ', 'Ὁ', 'Ὂ', |
|
89
|
|
|
'Ὃ', 'Ὄ', 'Ὅ', 'Ὸ', 'Ό', 'О', 'Θ', 'Ө'], |
|
90
|
|
|
'P' => ['П', 'Π'], |
|
91
|
|
|
'R' => ['Ř', 'Ŕ', 'Р', 'Ρ'], |
|
92
|
|
|
'S' => ['Ş', 'Ŝ', 'Ș', 'Š', 'Ś', 'С', 'Σ'], |
|
93
|
|
|
'T' => ['Ť', 'Ţ', 'Ŧ', 'Ț', 'Т', 'Τ'], |
|
94
|
|
|
'U' => ['Ú', 'Ù', 'Ủ', 'Ũ', 'Ụ', 'Ư', 'Ứ', 'Ừ', 'Ử', 'Ữ', 'Ự', 'Û', 'Ū', |
|
95
|
|
|
'Ů', 'Ű', 'Ŭ', 'Ų', 'У'], |
|
96
|
|
|
'V' => ['В'], |
|
97
|
|
|
'W' => ['Ω', 'Ώ'], |
|
98
|
|
|
'X' => ['Χ'], |
|
99
|
|
|
'Y' => ['Ý', 'Ỳ', 'Ỷ', 'Ỹ', 'Ỵ', 'Ÿ', 'Ῠ', 'Ῡ', 'Ὺ', 'Ύ', 'Ы', 'Й', 'Υ', 'Ϋ'], |
|
100
|
|
|
'Z' => ['Ź', 'Ž', 'Ż', 'З', 'Ζ'], |
|
101
|
|
|
'AE' => ['Ä', 'Æ'], |
|
102
|
|
|
'CH' => ['Ч'], |
|
103
|
|
|
'DJ' => ['Ђ'], |
|
104
|
|
|
'DZ' => ['Џ'], |
|
105
|
|
|
'KH' => ['Х'], |
|
106
|
|
|
'LJ' => ['Љ'], |
|
107
|
|
|
'NJ' => ['Њ'], |
|
108
|
|
|
'OE' => ['Ö'], |
|
109
|
|
|
'PS' => ['Ψ'], |
|
110
|
|
|
'SH' => ['Ш'], |
|
111
|
|
|
'SHCH' => ['Щ'], |
|
112
|
|
|
'SS' => ['ẞ'], |
|
113
|
|
|
'TH' => ['Þ'], |
|
114
|
|
|
'TS' => ['Ц'], |
|
115
|
|
|
'UE' => ['Ü'], |
|
116
|
|
|
'YA' => ['Я'], |
|
117
|
|
|
'YU' => ['Ю'], |
|
118
|
|
|
'ZH' => ['Ж'], |
|
119
|
|
|
' ' => ["\xC2\xA0", "\xE2\x80\x80", "\xE2\x80\x81", "\xE2\x80\x82", |
|
120
|
|
|
"\xE2\x80\x83", "\xE2\x80\x84", "\xE2\x80\x85", "\xE2\x80\x86", |
|
121
|
|
|
"\xE2\x80\x87", "\xE2\x80\x88", "\xE2\x80\x89", "\xE2\x80\x8A", |
|
122
|
|
|
"\xE2\x80\xAF", "\xE2\x81\x9F", "\xE3\x80\x80"], |
|
123
|
|
|
]; |
|
124
|
|
|
|
|
125
|
|
|
// PROPERTIES |
|
126
|
|
|
protected $strobj; |
|
127
|
|
|
|
|
128
|
|
|
protected function __construct(AnyStrObj $strobj) |
|
129
|
|
|
{ |
|
130
|
|
|
$this->strobj = $strobj; |
|
131
|
|
|
|
|
132
|
|
|
} |
|
133
|
|
|
|
|
134
|
|
|
public function __call($name, $args) |
|
|
|
|
|
|
135
|
|
|
{ |
|
136
|
|
|
return $this->strobj->$name($args); |
|
137
|
|
|
} |
|
138
|
|
|
|
|
139
|
|
|
public function __toString() |
|
140
|
|
|
{ |
|
141
|
|
|
return $this->strobj->__toString(); |
|
142
|
|
|
} |
|
143
|
|
|
|
|
144
|
|
|
public static function convert(AnyStrObj $strobj) |
|
145
|
|
|
{ |
|
146
|
|
|
$str = $strobj->raw; |
|
|
|
|
|
|
147
|
|
|
foreach (self::$asciimap as $key => $value) { |
|
148
|
|
|
$str = \str_replace($value, $key, $str); |
|
149
|
|
|
} |
|
150
|
|
|
return self::duplicate($strobj, $str); |
|
151
|
|
|
} |
|
152
|
|
|
|
|
153
|
|
|
public static function convertPure(AnyStrObj $strobj) |
|
154
|
|
|
{ |
|
155
|
|
|
$asciiobj = self::convert($strobj); |
|
156
|
|
|
$str = \preg_replace('/[^\x20-\x7E]/u', '', $asciiobj->raw); |
|
|
|
|
|
|
157
|
|
|
return self::duplicate($strobj, $str); |
|
158
|
|
|
} |
|
159
|
|
|
|
|
160
|
|
|
protected static function duplicate(AnyStrObj $strobj, $str) |
|
161
|
|
|
{ |
|
162
|
|
|
$classname = \get_class($strobj); |
|
163
|
|
|
return new self(new $classname($str)); |
|
164
|
|
|
} |
|
165
|
|
|
} |
|
166
|
|
|
|
garrettw /
stringobject
Loading history...
Our type inference engine in quite powerful, but sometimes the code does not provide enough clues to go by. In these cases we request you to add a
@returnannotation as described here.