Sanitizer::setOption() - Code Metrics - Inspection of "Tidy" - onoi/tesa - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Push — master ( 83b1b2...0f3bb4 )

by mw

created 2015-12-08 21:18 UTC

Sanitizer::setOption() A

↳ Parent: Sanitizer

Complexity

Conditions	3
Paths	2

Size

Total Lines	8
Code Lines	4

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	6
CRAP Score	3

Importance

Changes	1
Bugs	0	Features	0

Metric	Value
c	1
b	0
f	0
dl	0
loc	8
ccs	6
cts	6
cp	1
rs	9.4286
cc	3
eloc	4
nc	2
nop	2
crap	3

<?php

namespace Onoi\Tesa;

/**
 * @since 0.1
 *
 * @{
 */
// @codeCoverageIgnoreStart
define( 'ONOI_TESA_CHARACTER_MIN_LENGTH', 'ONOI_TESA_CHARACTER_MIN_LENGTH' );
define( 'ONOI_TESA_WORD_WHITELIST', 'ONOI_TESA_WORD_WHITELIST' );
// @codeCoverageIgnoreEnd
/**@}
 */

/**
 * @license GNU GPL v2+
 * @since 0.1
 *
 * @author mwjames
 */
class Sanitizer {

	/**
	 * Any change to the content of its data files should be reflected in a
	 * version change (the version number does not necessarily correlate with
	 * the library version)
	 */
	const VERSION = '0.1.1';

	/**
	 * @var string
	 */
	private $string = '';

	/**
	 * @var null
	 */
	private $encoding = null;

	/**
	 * @var array
	 */
	private $options = array();

	/**
	 * @since 0.1
	 *
	 * @param string $string
	 */
	public function __construct( $string ) {
		$this->string = $string;
		$this->encoding = $this->detectEncoding( $string );

		$this->setOption( ONOI_TESA_CHARACTER_MIN_LENGTH, 3 );
		$this->setOption( ONOI_TESA_WORD_WHITELIST, array() );
	}

	/**
	 * @since 1.0
	 *
	 * @param string $name
	 * @param mixed $value
	 */
	public function setOption( $name, $value ) {

		if ( $name === ONOI_TESA_WORD_WHITELIST && $value !== array() ) {
			$value = array_fill_keys( $value, true );
		}

		$this->options[$name] = $value;
	}

	/**
	 * @since 0.1
	 *
	 * @param integer $flag
	 */
	public function applyTransliteration( $flag = Transliterator::DIACRITICS ) {
		$this->string = Transliterator::transliterate( $this->string, $flag );
	}

	/**
	 * @since 0.1
	 *
	 * @param integer $flag
	 *
	 * @return array
	 */
	public function getTokens( $flag = Tokenizer::STRICT ) {
		return Tokenizer::tokenize( $this->string, $flag );
	}

	/**
	 * @since 0.1
	 *
	 * @param StopwordAnalyzer $stopwordAnalyzer
	 *
	 * @return string
	 */
	public function sanitizeBy( StopwordAnalyzer $stopwordAnalyzer ) {

		$words = $this->getTokens();
		$wordWhitelist = $this->options['ONOI_TESA_WORD_WHITELIST'];
		$minLength = (int)$this->options['ONOI_TESA_CHARACTER_MIN_LENGTH'];

		$index = array();
		$pos = 0;

		if ( !$words || !is_array( $words ) ) {
			return $this->string;
		}

		foreach ( $words as $key => $word ) {

			// If it is not an exemption and less than the required minimum length
			// or identified as stop word it is removed
			if ( !isset( $wordWhitelist[$word] ) && ( mb_strlen( $word ) < $minLength || $stopwordAnalyzer->isStopWord( $word ) ) ) {
				continue;
			}

			// Simple proximity checker for same words appearing next to each other
			if ( isset( $index[$pos-1] ) && $index[$pos-1] === $word ) {
				continue;
			}

			$index[] = $word;
			$pos++;
		}

		return implode( ' ' , $index );
	}

	/**
	 * @since 0.1
	 */
	public function toLowercase() {
		$this->string = mb_strtolower( $this->string, $this->encoding );
	}

	/**
	 * @since 0.1
	 *
	 * @param integer $length
	 */
	public function reduceLengthTo( $length ) {

		if ( mb_strlen( $this->string ) <= $length ) {
			return;
		}

		if ( strpos( $this->string, ' ' ) !== false ) {
			$length = strrpos( mb_substr( $this->string, 0, $length, $this->encoding ), ' ' ); // last whole word
		}

		$this->string = mb_substr( $this->string, 0, $length, $this->encoding );
	}

	/**
	 * @see http://www.phpwact.org/php/i18n/utf-8#str_replace
	 * @since 0.1
	 *
	 * @param string $search
	 * @param string $replace
	 */
	public function replace( $search, $replace ) {
		$this->string = str_replace( $search, $replace, $this->string );
	}

	/**
	 * @since 0.1
	 *
	 * @return boolean
	 */
	public function containsJapaneseCharacters() {
		return preg_match('/[\x{3040}-\x{309F}]/u', $this->string ) > 0 || preg_match('/[\x{30A0}-\x{30FF}]/u', $this->string ) > 0; // isHiragana || isKatakana
	}

	/**
	 * @since 0.1
	 *
	 * @return boolean
	 */
	public function containsKoreanCharacters() {
		return preg_match('/[\x{3130}-\x{318F}]/u', $this->string ) > 0 || preg_match('/[\x{AC00}-\x{D7AF}]/u', $this->string ) > 0;
	}

	/**
	 * @since 0.1
	 *
	 * @return boolean
	 */
	public function containsChineseCharacters() {
		return preg_match('/[\x{4e00}-\x{9fa5}]/u', $this->string ) > 0;
	}

	/**
	 * @since 0.1
	 *
	 * @return string
	 */
	public function __toString() {
		return $this->string;
	}

	private function detectEncoding( $string) {
		return mb_detect_encoding( $string );
	}

}


1		<?php
2
3		namespace Onoi\Tesa;
4
5		/**
6		* @since 0.1
7		*
8		* @{
9		*/
10		// @codeCoverageIgnoreStart
11		define( 'ONOI_TESA_CHARACTER_MIN_LENGTH', 'ONOI_TESA_CHARACTER_MIN_LENGTH' );
12		define( 'ONOI_TESA_WORD_WHITELIST', 'ONOI_TESA_WORD_WHITELIST' );
13		// @codeCoverageIgnoreEnd
14		/**@}
15		*/
16
17		/**
18		* @license GNU GPL v2+
19		* @since 0.1
20		*
21		* @author mwjames
22		*/
23		class Sanitizer {
24
25		/**
26		* Any change to the content of its data files should be reflected in a
27		* version change (the version number does not necessarily correlate with
28		* the library version)
29		*/
30		const VERSION = '0.1.1';
31
32		/**
33		* @var string
34		*/
35		private $string = '';
36
37		/**
38		* @var null
39		*/
40		private $encoding = null;
41
42		/**
43		* @var array
44		*/
45		private $options = array();
46
47		/**
48		* @since 0.1
49		*
50		* @param string $string
51		*/
52	20	public function __construct( $string ) {
53	20	$this->string = $string;
54	20	$this->encoding = $this->detectEncoding( $string );
		0 ignored issues – show Documentation Bug introduced 2015-11-27 19:42 UTC by Report Bug Copy Issue Report It seems like `$this->detectEncoding($string)` of type `string` is incompatible with the declared type `null` of property `$encoding`. 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...
55	20	$this->setOption( ONOI_TESA_CHARACTER_MIN_LENGTH, 3 );
56	20	$this->setOption( ONOI_TESA_WORD_WHITELIST, array() );
57	20	}
58
59		/**
60		* @since 1.0
61		*
62		* @param string $name
63		* @param mixed $value
64		*/
65	20	public function setOption( $name, $value ) {
66
67	20	if ( $name === ONOI_TESA_WORD_WHITELIST && $value !== array() ) {
68	1	$value = array_fill_keys( $value, true );
69	1	}
70
71	20	$this->options[$name] = $value;
72	20	}
73
74		/**
75		* @since 0.1
76		*
77		* @param integer $flag
78		*/
79	1	public function applyTransliteration( $flag = Transliterator::DIACRITICS ) {
80	1	$this->string = Transliterator::transliterate( $this->string, $flag );
81	1	}
82
83		/**
84		* @since 0.1
85		*
86		* @param integer $flag
87		*
88		* @return array
89		*/
90	11	public function getTokens( $flag = Tokenizer::STRICT ) {
91	11	return Tokenizer::tokenize( $this->string, $flag );
92		}
93
94		/**
95		* @since 0.1
96		*
97		* @param StopwordAnalyzer $stopwordAnalyzer
98		*
99		* @return string
100		*/
101	8	public function sanitizeBy( StopwordAnalyzer $stopwordAnalyzer ) {
102
103	8	$words = $this->getTokens();
104	8	$wordWhitelist = $this->options['ONOI_TESA_WORD_WHITELIST'];
105	8	$minLength = (int)$this->options['ONOI_TESA_CHARACTER_MIN_LENGTH'];
106
107	8	$index = array();
108	8	$pos = 0;
109
110	8	if ( !$words \|\| !is_array( $words ) ) {
111	1	return $this->string;
112		}
113
114	7	foreach ( $words as $key => $word ) {
115
116		// If it is not an exemption and less than the required minimum length
117		// or identified as stop word it is removed
118	7	if ( !isset( $wordWhitelist[$word] ) && ( mb_strlen( $word ) < $minLength \|\| $stopwordAnalyzer->isStopWord( $word ) ) ) {
119	6	continue;
120		}
121
122		// Simple proximity checker for same words appearing next to each other
123	7	if ( isset( $index[$pos-1] ) && $index[$pos-1] === $word ) {
124	2	continue;
125		}
126
127	7	$index[] = $word;
128	7	$pos++;
129	7	}
130
131	7	return implode( ' ' , $index );
132		}
133
134		/**
135		* @since 0.1
136		*/
137	5	public function toLowercase() {
138	5	$this->string = mb_strtolower( $this->string, $this->encoding );
139	5	}
140
141		/**
142		* @since 0.1
143		*
144		* @param integer $length
145		*/
146	3	public function reduceLengthTo( $length ) {
147
148	3	if ( mb_strlen( $this->string ) <= $length ) {
149	1	return;
150		}
151
152	3	if ( strpos( $this->string, ' ' ) !== false ) {
153	1	$length = strrpos( mb_substr( $this->string, 0, $length, $this->encoding ), ' ' ); // last whole word
154	1	}
155
156	3	$this->string = mb_substr( $this->string, 0, $length, $this->encoding );
157	3	}
158
159		/**
160		* @see http://www.phpwact.org/php/i18n/utf-8#str_replace
161		* @since 0.1
162		*
163		* @param string $search
164		* @param string $replace
165		*/
166	1	public function replace( $search, $replace ) {
167	1	$this->string = str_replace( $search, $replace, $this->string );
168	1	}
169
170		/**
171		* @since 0.1
172		*
173		* @return boolean
174		*/
175	1	public function containsJapaneseCharacters() {
176	1	return preg_match('/[\x{3040}-\x{309F}]/u', $this->string ) > 0 \|\| preg_match('/[\x{30A0}-\x{30FF}]/u', $this->string ) > 0; // isHiragana \|\| isKatakana
177		}
178
179		/**
180		* @since 0.1
181		*
182		* @return boolean
183		*/
184	1	public function containsKoreanCharacters() {
185	1	return preg_match('/[\x{3130}-\x{318F}]/u', $this->string ) > 0 \|\| preg_match('/[\x{AC00}-\x{D7AF}]/u', $this->string ) > 0;
186		}
187
188		/**
189		* @since 0.1
190		*
191		* @return boolean
192		*/
193	1	public function containsChineseCharacters() {
194	1	return preg_match('/[\x{4e00}-\x{9fa5}]/u', $this->string ) > 0;
195		}
196
197		/**
198		* @since 0.1
199		*
200		* @return string
201		*/
202	6	public function __toString() {
203	6	return $this->string;
204		}
205
206	20	private function detectEncoding( $string) {
207	20	return mb_detect_encoding( $string );
208		}
209
210		}
211

onoi / tesa

Push — master ( 83b1b2...0f3bb4 )

Sanitizer::setOption() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like