CJKSimpleCharacterRegExTokenizer::tokenize() - 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

CJKSimpleCharacterRegExTokenizer::tokenize() A

↳ Parent: CJKSimpleCharacterRegExTokenizer

Complexity

Conditions	3
Paths	4

Size

Total Lines	21
Code Lines	11

Duplication

Lines	0
Ratio	0 %

Importance

Changes	1
Bugs	0	Features	1

Metric	Value
c	1
b	0
f	1
dl	0
loc	21
rs	9.3142
cc	3
eloc	11
nc	4
nop	1

<?php

namespace Onoi\Tesa\Tokenizer;

/**
 * @license GNU GPL v2+
 * @since 0.1
 *
 * @author mwjames
 */
class CJKSimpleCharacterRegExTokenizer implements Tokenizer {

	/**
	 * @var Tokenizer
	 */
	private $tokenizer;

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

	/**
	 * @since 0.1
	 *
	 * @param Tokenizer $tokenizer
	 */
	public function __construct( Tokenizer $tokenizer = null ) {
		$this->tokenizer = $tokenizer;
	}

	/**
	 * @since 0.1
	 *
	 * {@inheritDoc}
	 */
	public function setOption( $name, $value ) {

		if ( $this->tokenizer !== null ) {
			$this->tokenizer->setOption( $name, $value );
		}

		if ( $name === self::REGEX_EXEMPTION ) {
			$this->patternExemption = $value;
		}
	}

	/**
	 * @since 0.1
	 *
	 * {@inheritDoc}
	 */
	public function isWordTokenizer() {
		return false;
	}

	/**
	 * @since 0.1
	 *
	 * {@inheritDoc}
	 */
	public function tokenize( $string ) {

		if ( $this->tokenizer !== null ) {
			$string = implode( " ", $this->tokenizer->tokenize( $string ) );
		}

		// Filter is based on https://github.com/kitech/cms-drupal/blob/master/modules/csplitter/filter.txt
		$pattern = str_replace(
			$this->patternExemption,
			'',
			'([\s\、，,。／？《》〈〉；：“”＂〃＇｀［］｛｝＼｜～！－＝＿＋）（()＊…—─％￥…◆★◇□■【】＃·啊吧把并被才从的得当对但到地而该过个给还和叫将就可来了啦里没你您哪那呢去却让使是时省随他我为现县向像象要由矣已以也又与于在之这则最乃\/\(\)\[\]{}<>\r\n"]|(?<!\d)\.(?!\d))'
		);

		$result = preg_split( '/' . $pattern . '/u', $string, null, PREG_SPLIT_NO_EMPTY );

		if ( $result !== false ) {
			return $result;
		}

		return array();
	}

}


1			<?php
2
3			namespace Onoi\Tesa\Tokenizer;
4
5			/**
6			* @license GNU GPL v2+
7			* @since 0.1
8			*
9			* @author mwjames
10			*/
11			class CJKSimpleCharacterRegExTokenizer implements Tokenizer {
12
13			/**
14			* @var Tokenizer
15			*/
16			private $tokenizer;
17
18			/**
19			* @var string
20			*/
21			private $patternExemption = '';
22
23			/**
24			* @since 0.1
25			*
26			* @param Tokenizer $tokenizer
27			*/
28			public function __construct( Tokenizer $tokenizer = null ) {
29			$this->tokenizer = $tokenizer;
30			}
31
32			/**
33			* @since 0.1
34			*
35			* {@inheritDoc}
36			*/
37			public function setOption( $name, $value ) {
38
39			if ( $this->tokenizer !== null ) {
40			$this->tokenizer->setOption( $name, $value );
41			}
42
43			if ( $name === self::REGEX_EXEMPTION ) {
44			$this->patternExemption = $value;
45			}
46			}
47
48			/**
49			* @since 0.1
50			*
51			* {@inheritDoc}
52			*/
53			public function isWordTokenizer() {
54			return false;
55			}
56
57			/**
58			* @since 0.1
59			*
60			* {@inheritDoc}
61			*/
62			public function tokenize( $string ) {
63
64			if ( $this->tokenizer !== null ) {
65			$string = implode( " ", $this->tokenizer->tokenize( $string ) );
66			}
67
68			// Filter is based on https://github.com/kitech/cms-drupal/blob/master/modules/csplitter/filter.txt
69			$pattern = str_replace(
70			$this->patternExemption,
71			'',
72			'([\s\、，,。／？《》〈〉；：“”＂〃＇｀［］｛｝＼｜～！－＝＿＋）（()＊…—─％￥…◆★◇□■【】＃·啊吧把并被才从的得当对但到地而该过个给还和叫将就可来了啦里没你您哪那呢去却让使是时省随他我为现县向像象要由矣已以也又与于在之这则最乃\/\(\)\[\]{}<>\r\n"]\|(?<!\d)\.(?!\d))'
73			);
74
75			$result = preg_split( '/' . $pattern . '/u', $string, null, PREG_SPLIT_NO_EMPTY );
76
77			if ( $result !== false ) {
78			return $result;
79			}
80
81			return array();
82			}
83
84			}
85

onoi / tesa

Push — master ( 9401a4...2cc7a2 )

CJKSimpleCharacterRegExTokenizer::tokenize() A

Complexity

Size

Duplication

Importance

Duplication Side-by-Side

Filter issues like