|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
namespace Onoi\Tesa\Tests; |
|
4
|
|
|
|
|
5
|
|
|
use Onoi\Tesa\Tokenizer\IcuWordBoundaryTokenizer; |
|
6
|
|
|
|
|
7
|
|
|
/** |
|
8
|
|
|
* @covers \Onoi\Tesa\Tokenizer\IcuWordBoundaryTokenizer |
|
9
|
|
|
* @group onoi-tesa |
|
10
|
|
|
* |
|
11
|
|
|
* @license GNU GPL v2+ |
|
12
|
|
|
* @since 0.1 |
|
13
|
|
|
* |
|
14
|
|
|
* @author mwjames |
|
15
|
|
|
*/ |
|
16
|
|
|
class IcuWordBoundaryTokenizerTest extends \PHPUnit_Framework_TestCase { |
|
17
|
|
|
|
|
18
|
|
|
protected function setUp() { |
|
19
|
|
|
$instance = new IcuWordBoundaryTokenizer(); |
|
20
|
|
|
|
|
21
|
|
|
if ( !$instance->isAvailable() || INTL_ICU_VERSION != '54.1' ) { |
|
22
|
|
|
$this->markTestSkipped( 'ICU extension is not available or does not match the expected version constraint.' ); |
|
23
|
|
|
} |
|
24
|
|
|
} |
|
25
|
|
|
|
|
26
|
|
|
public function testCanConstruct() { |
|
27
|
|
|
|
|
28
|
|
|
$this->assertInstanceOf( |
|
29
|
|
|
'\Onoi\Tesa\Tokenizer\IcuWordBoundaryTokenizer', |
|
30
|
|
|
new IcuWordBoundaryTokenizer() |
|
31
|
|
|
); |
|
32
|
|
|
} |
|
33
|
|
|
|
|
34
|
|
|
/** |
|
35
|
|
|
* @dataProvider stringProvider |
|
36
|
|
|
*/ |
|
37
|
|
|
public function testTokenize( $string, $expected ) { |
|
38
|
|
|
|
|
39
|
|
|
$instance = new IcuWordBoundaryTokenizer(); |
|
40
|
|
|
|
|
41
|
|
|
$this->assertEquals( |
|
42
|
|
|
$expected, |
|
43
|
|
|
$instance->tokenize( $string ) |
|
44
|
|
|
); |
|
45
|
|
|
} |
|
46
|
|
|
|
|
47
|
|
|
public function testSetOption() { |
|
48
|
|
|
|
|
49
|
|
|
$tokenizer = $this->getMockBuilder( '\Onoi\Tesa\Tokenizer\Tokenizer' ) |
|
50
|
|
|
->disableOriginalConstructor() |
|
51
|
|
|
->getMockForAbstractClass(); |
|
52
|
|
|
|
|
53
|
|
|
$tokenizer->expects( $this->once() ) |
|
54
|
|
|
->method( 'setOption' ); |
|
55
|
|
|
|
|
56
|
|
|
$instance = new IcuWordBoundaryTokenizer( |
|
57
|
|
|
$tokenizer |
|
58
|
|
|
); |
|
59
|
|
|
|
|
60
|
|
|
$instance->setOption( |
|
61
|
|
|
IcuWordBoundaryTokenizer::REGEX_EXEMPTION, |
|
62
|
|
|
array( 'Foo' ) |
|
63
|
|
|
); |
|
64
|
|
|
} |
|
65
|
|
|
|
|
66
|
|
|
public function testGeneralSetters() { |
|
67
|
|
|
|
|
68
|
|
|
$tokenizer = $this->getMockBuilder( '\Onoi\Tesa\Tokenizer\Tokenizer' ) |
|
69
|
|
|
->disableOriginalConstructor() |
|
70
|
|
|
->getMockForAbstractClass(); |
|
71
|
|
|
|
|
72
|
|
|
$instance = new IcuWordBoundaryTokenizer( |
|
73
|
|
|
$tokenizer |
|
74
|
|
|
); |
|
75
|
|
|
|
|
76
|
|
|
$instance->setLocale( 'en' ); |
|
77
|
|
|
$instance->setWordTokenizerAttribute( false ); |
|
78
|
|
|
|
|
79
|
|
|
$this->assertFalse( |
|
80
|
|
|
$instance->isWordTokenizer() |
|
81
|
|
|
); |
|
82
|
|
|
} |
|
83
|
|
|
|
|
84
|
|
|
public function stringProvider() { |
|
85
|
|
|
|
|
86
|
|
|
$provider[] = array( |
|
|
|
|
|
|
87
|
|
|
"安全テスト", |
|
88
|
|
|
array( '安全', 'テスト' ) |
|
89
|
|
|
); |
|
90
|
|
|
|
|
91
|
|
|
// Would expect 'すもも', 'も', 'もも', 'も', 'もも', 'の', 'うち', '。' |
|
92
|
|
|
$provider[] = array( |
|
93
|
|
|
"すもももももももものうち。", |
|
94
|
|
|
array( 'すもも', 'も', 'も', 'も', 'も', 'も', 'もの', 'うち', '。' ) |
|
95
|
|
|
); |
|
96
|
|
|
|
|
97
|
|
|
$provider[] = array( |
|
98
|
|
|
"李も桃も桃のうち。", |
|
99
|
|
|
array( '李', 'も', '桃', 'も', '桃', 'の', 'うち', '。' ) |
|
100
|
|
|
); |
|
101
|
|
|
|
|
102
|
|
|
$provider[] = array( |
|
103
|
|
|
"إسرائيل", |
|
104
|
|
|
array( 'إسرائيل' ) |
|
105
|
|
|
); |
|
106
|
|
|
|
|
107
|
|
|
$provider[] = array( |
|
108
|
|
|
"검색엔ㅇㅏ진", |
|
109
|
|
|
array( '검색엔', 'ㅇㅏ', '진' ) |
|
110
|
|
|
); |
|
111
|
|
|
|
|
112
|
|
|
$provider[] = array( |
|
113
|
|
|
"검색엔ㅇㅏ진1234abcdfrA", |
|
114
|
|
|
array( '검색엔', 'ㅇㅏ', '진', '1234abcdfrA' ) |
|
115
|
|
|
); |
|
116
|
|
|
|
|
117
|
|
|
$provider[] = array( |
|
118
|
|
|
"1234abcdfrA", |
|
119
|
|
|
array( '1234abcdfrA' ) |
|
120
|
|
|
); |
|
121
|
|
|
|
|
122
|
|
|
$provider[] = array( |
|
123
|
|
|
"公明執ようなＳＮＳもストーカー行為の対象に", |
|
124
|
|
|
array( |
|
125
|
|
|
'公明', '執よう','な','ＳＮＳ', 'も', |
|
126
|
|
|
'ストーカー', '行為', 'の', '対象', 'に' |
|
127
|
|
|
) |
|
128
|
|
|
); |
|
129
|
|
|
|
|
130
|
|
|
$provider[] = array( |
|
131
|
|
|
"公明執", |
|
132
|
|
|
array( '公明', '執' ) |
|
133
|
|
|
); |
|
134
|
|
|
|
|
135
|
|
|
$provider[] = array( |
|
136
|
|
|
"IQテスト", |
|
137
|
|
|
array( 'IQ', 'テスト' ) |
|
138
|
|
|
); |
|
139
|
|
|
|
|
140
|
|
|
$provider[] = array( |
|
141
|
|
|
"foo テスト bar", |
|
142
|
|
|
array( 'foo', 'テスト', 'bar' ) |
|
143
|
|
|
); |
|
144
|
|
|
|
|
145
|
|
|
$provider[] = array( |
|
146
|
|
|
"foo テスト bar 123abc ^&'", |
|
147
|
|
|
array( 'foo', 'テスト', 'bar', '123abc', '^', '&', "'" ) |
|
148
|
|
|
); |
|
149
|
|
|
|
|
150
|
|
|
$provider[] = array( |
|
151
|
|
|
"was discovered in 1957 and first sold as a medication in 1971", |
|
152
|
|
|
array( |
|
153
|
|
|
'was', 'discovered', 'in', '1957', 'and', |
|
154
|
|
|
'first', 'sold', 'as', 'a', 'medication', 'in', '1971' |
|
155
|
|
|
) |
|
156
|
|
|
); |
|
157
|
|
|
|
|
158
|
|
|
// See JaTinySegmenterTokenizerTest for comparison |
|
159
|
|
|
$provider[] = array( |
|
160
|
|
|
'日本語の新聞記事であれば文字単位で95%程度の精度で分かち書きが行えます。 ', |
|
161
|
|
|
array( |
|
162
|
|
|
'日本語', 'の', '新聞', '記事', 'で', |
|
163
|
|
|
'あれ', 'ば', '文字', '単位', |
|
164
|
|
|
'で', '95', '%', '程度', |
|
165
|
|
|
'の', '精度', 'で', '分かち書き', |
|
166
|
|
|
'が', '行', 'え', 'ます', '。' |
|
167
|
|
|
) |
|
168
|
|
|
); |
|
169
|
|
|
|
|
170
|
|
|
return $provider; |
|
171
|
|
|
} |
|
172
|
|
|
|
|
173
|
|
|
} |
|
174
|
|
|
|
onoi /
tesa
Adding an explicit array definition is generally preferable to implicit array definition as it guarantees a stable state of the code.
Let’s take a look at an example:
As you can see in this example, the array
$myArrayis initialized the first time when the foreach loop is entered. You can also see that the value of thebarkey is only written conditionally; thus, its value might result from a previous iteration.This might or might not be intended. To make your intention clear, your code more readible and to avoid accidental bugs, we recommend to add an explicit initialization $myArray = array() either outside or inside the foreach loop.