SanitizerTest::testSanitizeByStopwords() - Code Metrics - Inspection of "Add SanitizerFactory" - onoi/tesa - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Push — master ( 9401a4...2cc7a2 )

by mw

created 2016-08-02 14:22 UTC

SanitizerTest::testSanitizeByStopwords() A

↳ Parent: SanitizerTest

Complexity

Conditions	1
Paths	1

Size

Total Lines	12
Code Lines	7

Duplication

Lines	0
Ratio	0 %

Importance

Changes	2
Bugs	0	Features	1

Metric	Value
c	2
b	0
f	1
dl	0
loc	12
rs	9.4286
cc	1
eloc	7
nc	1
nop	0

<?php

namespace Onoi\Tesa\Tests;

use Onoi\Tesa\Sanitizer;
use Onoi\Tesa\SanitizerFactory;

/**
 * @covers \Onoi\Tesa\Sanitizer
 * @group onoi-tesa
 *
 * @license GNU GPL v2+
 * @since 0.1
 *
 * @author mwjames
 */
class SanitizerTest extends \PHPUnit_Framework_TestCase {

	private $sanitizerFactory;

	protected function setUp() {
		$this->sanitizerFactory = new SanitizerFactory();
	}

	public function testTransliteration() {

		$instance = new Sanitizer( 'ÀÁÂÃÄÅàáâãäåÒÓÔÕÕÖØòóôõöøÈÉÊËèéêëðÇçÐÌÍÎÏìíîïÙÚÛÜùúûüÑñŠšŸÿýŽž' );
		$instance->applyTransliteration();

		$this->assertEquals(
			'AAAAAEAaaaaaeaOOOOOOEOoooooeoEEEEeeeeðCcÐIIIIiiiiUUUUEuuuueNnSsYyyZz',
			$instance
		);
	}

	public function testToLowercase() {

		$instance = new Sanitizer( 'ÀÁÂÃÄÅ ABC 텍스트의 テスト часто הוא פשוט' );
		$instance->toLowercase();

		$this->assertEquals(
			'àáâãäå abc 텍스트의 テスト часто הוא פשוט',
			$instance
		);
	}

	public function testReduceLengthTo() {

		$instance = new Sanitizer( 'ABCDEF' );
		$instance->reduceLengthTo( 3 );

		$this->assertEquals(
			3,
			mb_strlen( $instance )
		);

		$instance->reduceLengthTo( 10 );

		$this->assertEquals(
			3,
			mb_strlen( $instance )
		);
	}

	public function testReduceLengthToNearestWholeWordForLatinString() {

		$instance = new Sanitizer( 'abc def gh in 123' );
		$instance->reduceLengthTo( 12 );

		$this->assertEquals(
			10,
			mb_strlen( $instance )
		);

		$this->assertEquals(
			'abc def gh',
			$instance
		);
	}

	public function testReduceLengthToNearestWholeWordForNonLatinString() {

		if ( version_compare( phpversion(), '5.4', '<' ) ) {
			$this->markTestSkipped(
				"Boo, PHP 5.3 returns with `Failed asserting that 9 matches expected 3`"
			);
		}

		$instance = new Sanitizer( '一　二　三' );
		$instance->reduceLengthTo( 3 );

		$this->assertEquals(
			3,
			mb_strlen( $instance )
		);

		$this->assertEquals(
			'一　二',
			$instance
		);
	}

	public function testReplace() {

		$instance = new Sanitizer( 'テスト' );
		$instance->replace( array( 'テスト' ), array( 'Test' ) );
function acceptsInteger($int) { }

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

// Instead of
acceptsInteger($x);

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

		$this->assertEquals(
			'Test',
			$instance
		);
	}

	public function testSanitizeWithSimpleStopwordList() {

		$text = 'Foo bar foobar';

		$tokenizer = $this->getMockBuilder( '\Onoi\Tesa\Tokenizer\Tokenizer' )
			->disableOriginalConstructor()
			->getMockForAbstractClass();

		$tokenizer->expects( $this->once() )
			->method( 'tokenize' )
			->with( $this->equalTo( $text ) )
			->will( $this->returnValue( array( 'Foo', 'bar', 'foobar' ) ) );

		$synonymizer = $this->getMockBuilder( '\Onoi\Tesa\Synonymizer\Synonymizer' )
			->disableOriginalConstructor()
			->getMockForAbstractClass();

		$synonymizer->expects( $this->any() )
			->method( 'synonymize' )
			->will($this->returnArgument( 0 ) );

		$instance = new Sanitizer( $text );

		$stopwordAnalyzer = $this->sanitizerFactory->newArrayStopwordAnalyzer(
			array( 'bar' )
		);

		$this->assertEquals(
			'Foo foobar',
			$instance->sanitizeWith( $tokenizer, $stopwordAnalyzer, $synonymizer )
		);
	}

	public function testSanitizeByStopwordsToIncludeExemptionWithMinLengthRestriction() {

		$text = 'Foo bar foobar';

		$tokenizer = $this->getMockBuilder( '\Onoi\Tesa\Tokenizer\Tokenizer' )
			->disableOriginalConstructor()
			->getMockForAbstractClass();

		$tokenizer->expects( $this->once() )
			->method( 'isWordTokenizer' )
			->will( $this->returnValue( true ) );

		$tokenizer->expects( $this->once() )
			->method( 'tokenize' )
			->with( $this->equalTo( $text ) )
			->will( $this->returnValue( array( 'Foo', 'bar', 'foobar' ) ) );

		$synonymizer = $this->getMockBuilder( '\Onoi\Tesa\Synonymizer\Synonymizer' )
			->disableOriginalConstructor()
			->getMockForAbstractClass();

		$synonymizer->expects( $this->any() )
			->method( 'synonymize' )
			->will($this->returnArgument( 0 ) );

		$instance = new Sanitizer( $text );

		$stopwordAnalyzer = $this->sanitizerFactory->newArrayStopwordAnalyzer(
			array( 'bar' )
		);

		$instance->setOption( Sanitizer::MIN_LENGTH, 4 );
		$instance->setOption( Sanitizer::WHITELIST, array( 'bar' ) );

		$this->assertEquals(
			'bar foobar',
			$instance->sanitizeWith( $tokenizer, $stopwordAnalyzer, $synonymizer )
		);
	}

	public function testTrySanitizeByStopwordsWithProximityCheck() {

		$text = 'foo foo テスト テスト';

		$tokenizer = $this->getMockBuilder( '\Onoi\Tesa\Tokenizer\Tokenizer' )
			->disableOriginalConstructor()
			->getMockForAbstractClass();

		$tokenizer->expects( $this->once() )
			->method( 'isWordTokenizer' )
			->will( $this->returnValue( true ) );

		$tokenizer->expects( $this->once() )
			->method( 'tokenize' )
			->with( $this->equalTo( $text ) )
			->will( $this->returnValue( array( 'foo', 'foo', 'テスト', 'テスト' ) ) );

		$synonymizer = $this->getMockBuilder( '\Onoi\Tesa\Synonymizer\Synonymizer' )
			->disableOriginalConstructor()
			->getMockForAbstractClass();

		$synonymizer->expects( $this->any() )
			->method( 'synonymize' )
			->will($this->returnArgument( 0 ) );

		$instance = new Sanitizer( $text );

		$stopwordAnalyzer = $this->sanitizerFactory->newArrayStopwordAnalyzer();

		$this->assertEquals(
			'foo テスト',
			$instance->sanitizeWith( $tokenizer, $stopwordAnalyzer, $synonymizer )
		);
	}

}


1			<?php
2
3			namespace Onoi\Tesa\Tests;
4
5			use Onoi\Tesa\Sanitizer;
6			use Onoi\Tesa\SanitizerFactory;
7
8			/**
9			* @covers \Onoi\Tesa\Sanitizer
10			* @group onoi-tesa
11			*
12			* @license GNU GPL v2+
13			* @since 0.1
14			*
15			* @author mwjames
16			*/
17			class SanitizerTest extends \PHPUnit_Framework_TestCase {
18
19			private $sanitizerFactory;
20
21			protected function setUp() {
22			$this->sanitizerFactory = new SanitizerFactory();
23			}
24
25			public function testTransliteration() {
26
27			$instance = new Sanitizer( 'ÀÁÂÃÄÅàáâãäåÒÓÔÕÕÖØòóôõöøÈÉÊËèéêëðÇçÐÌÍÎÏìíîïÙÚÛÜùúûüÑñŠšŸÿýŽž' );
28			$instance->applyTransliteration();
29
30			$this->assertEquals(
31			'AAAAAEAaaaaaeaOOOOOOEOoooooeoEEEEeeeeðCcÐIIIIiiiiUUUUEuuuueNnSsYyyZz',
32			$instance
33			);
34			}
35
36			public function testToLowercase() {
37
38			$instance = new Sanitizer( 'ÀÁÂÃÄÅ ABC 텍스트의 テスト часто הוא פשוט' );
39			$instance->toLowercase();
40
41			$this->assertEquals(
42			'àáâãäå abc 텍스트의 テスト часто הוא פשוט',
43			$instance
44			);
45			}
46
47			public function testReduceLengthTo() {
48
49			$instance = new Sanitizer( 'ABCDEF' );
50			$instance->reduceLengthTo( 3 );
51
52			$this->assertEquals(
53			3,
54			mb_strlen( $instance )
55			);
56
57			$instance->reduceLengthTo( 10 );
58
59			$this->assertEquals(
60			3,
61			mb_strlen( $instance )
62			);
63			}
64
65			public function testReduceLengthToNearestWholeWordForLatinString() {
66
67			$instance = new Sanitizer( 'abc def gh in 123' );
68			$instance->reduceLengthTo( 12 );
69
70			$this->assertEquals(
71			10,
72			mb_strlen( $instance )
73			);
74
75			$this->assertEquals(
76			'abc def gh',
77			$instance
78			);
79			}
80
81			public function testReduceLengthToNearestWholeWordForNonLatinString() {
82
83			if ( version_compare( phpversion(), '5.4', '<' ) ) {
84			$this->markTestSkipped(
85			"Boo, PHP 5.3 returns with `Failed asserting that 9 matches expected 3`"
86			);
87			}
88
89			$instance = new Sanitizer( '一　二　三' );
90			$instance->reduceLengthTo( 3 );
91
92			$this->assertEquals(
93			3,
94			mb_strlen( $instance )
95			);
96
97			$this->assertEquals(
98			'一　二',
99			$instance
100			);
101			}
102
103			public function testReplace() {
104
105			$instance = new Sanitizer( 'テスト' );
106			$instance->replace( array( 'テスト' ), array( 'Test' ) );
			0 ignored issues – show Documentation introduced 2015-11-27 19:42 UTC by Report Bug Copy Issue Report `array('テスト')` is of type `array<integer,string,{"0":"string"}>`, but the function expects a `string`. It seems like the type of the argument is not accepted by the function/method which you are calling. In some cases, in particular if PHP’s automatic type-juggling kicks in this might be fine. In other cases, however this might be a bug. We suggest to add an explicit type cast like in the following example: function acceptsInteger($int) { } $x = '123'; // string "123" // Instead of acceptsInteger($x); // we recommend to use acceptsInteger((integer) $x); Loading history... Documentation introduced 2015-11-27 19:42 UTC by Report Bug Copy Issue Report `array('Test')` is of type `array<integer,string,{"0":"string"}>`, but the function expects a `string`. It seems like the type of the argument is not accepted by the function/method which you are calling. In some cases, in particular if PHP’s automatic type-juggling kicks in this might be fine. In other cases, however this might be a bug. We suggest to add an explicit type cast like in the following example: function acceptsInteger($int) { } $x = '123'; // string "123" // Instead of acceptsInteger($x); // we recommend to use acceptsInteger((integer) $x); Loading history...
107
108			$this->assertEquals(
109			'Test',
110			$instance
111			);
112			}
113
114			public function testSanitizeWithSimpleStopwordList() {
115
116			$text = 'Foo bar foobar';
117
118			$tokenizer = $this->getMockBuilder( '\Onoi\Tesa\Tokenizer\Tokenizer' )
119			->disableOriginalConstructor()
120			->getMockForAbstractClass();
121
122			$tokenizer->expects( $this->once() )
123			->method( 'tokenize' )
124			->with( $this->equalTo( $text ) )
125			->will( $this->returnValue( array( 'Foo', 'bar', 'foobar' ) ) );
126
127			$synonymizer = $this->getMockBuilder( '\Onoi\Tesa\Synonymizer\Synonymizer' )
128			->disableOriginalConstructor()
129			->getMockForAbstractClass();
130
131			$synonymizer->expects( $this->any() )
132			->method( 'synonymize' )
133			->will($this->returnArgument( 0 ) );
134
135			$instance = new Sanitizer( $text );
136
137			$stopwordAnalyzer = $this->sanitizerFactory->newArrayStopwordAnalyzer(
138			array( 'bar' )
139			);
140
141			$this->assertEquals(
142			'Foo foobar',
143			$instance->sanitizeWith( $tokenizer, $stopwordAnalyzer, $synonymizer )
144			);
145			}
146
147			public function testSanitizeByStopwordsToIncludeExemptionWithMinLengthRestriction() {
148
149			$text = 'Foo bar foobar';
150
151			$tokenizer = $this->getMockBuilder( '\Onoi\Tesa\Tokenizer\Tokenizer' )
152			->disableOriginalConstructor()
153			->getMockForAbstractClass();
154
155			$tokenizer->expects( $this->once() )
156			->method( 'isWordTokenizer' )
157			->will( $this->returnValue( true ) );
158
159			$tokenizer->expects( $this->once() )
160			->method( 'tokenize' )
161			->with( $this->equalTo( $text ) )
162			->will( $this->returnValue( array( 'Foo', 'bar', 'foobar' ) ) );
163
164			$synonymizer = $this->getMockBuilder( '\Onoi\Tesa\Synonymizer\Synonymizer' )
165			->disableOriginalConstructor()
166			->getMockForAbstractClass();
167
168			$synonymizer->expects( $this->any() )
169			->method( 'synonymize' )
170			->will($this->returnArgument( 0 ) );
171
172			$instance = new Sanitizer( $text );
173
174			$stopwordAnalyzer = $this->sanitizerFactory->newArrayStopwordAnalyzer(
175			array( 'bar' )
176			);
177
178			$instance->setOption( Sanitizer::MIN_LENGTH, 4 );
179			$instance->setOption( Sanitizer::WHITELIST, array( 'bar' ) );
180
181			$this->assertEquals(
182			'bar foobar',
183			$instance->sanitizeWith( $tokenizer, $stopwordAnalyzer, $synonymizer )
184			);
185			}
186
187			public function testTrySanitizeByStopwordsWithProximityCheck() {
188
189			$text = 'foo foo テストテスト';
190
191			$tokenizer = $this->getMockBuilder( '\Onoi\Tesa\Tokenizer\Tokenizer' )
192			->disableOriginalConstructor()
193			->getMockForAbstractClass();
194
195			$tokenizer->expects( $this->once() )
196			->method( 'isWordTokenizer' )
197			->will( $this->returnValue( true ) );
198
199			$tokenizer->expects( $this->once() )
200			->method( 'tokenize' )
201			->with( $this->equalTo( $text ) )
202			->will( $this->returnValue( array( 'foo', 'foo', 'テスト', 'テスト' ) ) );
203
204			$synonymizer = $this->getMockBuilder( '\Onoi\Tesa\Synonymizer\Synonymizer' )
205			->disableOriginalConstructor()
206			->getMockForAbstractClass();
207
208			$synonymizer->expects( $this->any() )
209			->method( 'synonymize' )
210			->will($this->returnArgument( 0 ) );
211
212			$instance = new Sanitizer( $text );
213
214			$stopwordAnalyzer = $this->sanitizerFactory->newArrayStopwordAnalyzer();
215
216			$this->assertEquals(
217			'foo テスト',
218			$instance->sanitizeWith( $tokenizer, $stopwordAnalyzer, $synonymizer )
219			);
220			}
221
222			}
223

onoi / tesa

Push — master ( 9401a4...2cc7a2 )

SanitizerTest::testSanitizeByStopwords() A

Complexity

Size

Duplication

Importance

Duplication Side-by-Side

Filter issues like