This project does not seem to handle request data directly as such no vulnerable execution paths were found.
include
, or for example
via PHP's auto-loading mechanism.
These results are based on our legacy PHP analysis, consider migrating to our new PHP analysis engine instead. Learn more
1 | <?php |
||
2 | |||
3 | namespace Onoi\Tesa; |
||
4 | |||
5 | use Onoi\Tesa\Tokenizer\Tokenizer; |
||
6 | use Onoi\Tesa\Synonymizer\Synonymizer; |
||
7 | use Onoi\Tesa\StopwordAnalyzer\StopwordAnalyzer; |
||
8 | use RuntimeException; |
||
9 | |||
10 | /** |
||
11 | * @license GNU GPL v2+ |
||
12 | * @since 0.1 |
||
13 | * |
||
14 | * @author mwjames |
||
15 | */ |
||
16 | class Sanitizer { |
||
17 | |||
18 | const WHITELIST = 'WHITELIST'; |
||
19 | const MIN_LENGTH = 'MIN_LENGTH'; |
||
20 | |||
21 | /** |
||
22 | * Any change to the content of its data files should be reflected in a |
||
23 | * version change (the version number does not necessarily correlate with |
||
24 | * the library version) |
||
25 | */ |
||
26 | const VERSION = '0.2'; |
||
27 | |||
28 | /** |
||
29 | * @var string |
||
30 | */ |
||
31 | private $string = ''; |
||
32 | |||
33 | /** |
||
34 | * @var array |
||
35 | */ |
||
36 | private $whiteList = array(); |
||
37 | |||
38 | /** |
||
39 | * @var array |
||
40 | */ |
||
41 | private $minLength = 3; |
||
42 | |||
43 | /** |
||
44 | * @since 0.1 |
||
45 | * |
||
46 | * @param string $string |
||
47 | */ |
||
48 | 13 | public function __construct( $string = '' ) { |
|
49 | 13 | $this->setText( $string ); |
|
50 | 13 | } |
|
51 | |||
52 | /** |
||
53 | * @since 0.1 |
||
54 | * |
||
55 | * @param string $name |
||
56 | * @param mixed $value |
||
57 | */ |
||
58 | 1 | public function setOption( $name, $value ) { |
|
59 | |||
60 | 1 | if ( $name === self::WHITELIST && is_array( $value ) && $value !== array() ) { |
|
61 | 1 | $this->whiteList = array_fill_keys( $value, true ); |
|
62 | 1 | } |
|
63 | |||
64 | 1 | if ( $name === self::MIN_LENGTH ) { |
|
65 | 1 | $this->minLength = (int)$value; |
|
0 ignored issues
–
show
|
|||
66 | 1 | } |
|
67 | 1 | } |
|
68 | |||
69 | /** |
||
70 | * @since 0.1 |
||
71 | * |
||
72 | * @param string $string |
||
73 | */ |
||
74 | 13 | public function setText( $string ) { |
|
75 | 13 | $this->string = $string; |
|
76 | 13 | } |
|
77 | |||
78 | /** |
||
79 | * @since 0.1 |
||
80 | * |
||
81 | * @param integer $flag |
||
82 | */ |
||
83 | 1 | public function applyTransliteration( $flag = Transliterator::DIACRITICS ) { |
|
84 | 1 | $this->string = Normalizer::applyTransliteration( $this->string, $flag ); |
|
85 | 1 | } |
|
86 | |||
87 | /** |
||
88 | * @see Localizer::convertDoubleWidth |
||
89 | * |
||
90 | * @since 0.1 |
||
91 | * |
||
92 | * @param integer $flag |
||
93 | */ |
||
94 | public function convertDoubleWidth() { |
||
95 | $this->string = Normalizer::convertDoubleWidth( $this->string ); |
||
96 | } |
||
97 | |||
98 | /** |
||
99 | * @since 0.1 |
||
100 | * |
||
101 | * @param Tokenizer $tokenizer |
||
102 | * @param StopwordAnalyzer $stopwordAnalyzer |
||
103 | * |
||
104 | * @return string |
||
105 | */ |
||
106 | 7 | public function sanitizeWith( Tokenizer $tokenizer, StopwordAnalyzer $stopwordAnalyzer, Synonymizer $synonymizer ) { |
|
107 | |||
108 | // Treat non-words tokenizers (Ja,Zh*) differently |
||
109 | 7 | $minLength = $tokenizer->isWordTokenizer() ? $this->minLength : 1; |
|
110 | |||
111 | 7 | $words = $tokenizer->tokenize( $this->string ); |
|
112 | |||
113 | 7 | if ( !$words || !is_array( $words ) ) { |
|
114 | return $this->string; |
||
115 | } |
||
116 | |||
117 | 7 | $index = array(); |
|
118 | 7 | $pos = 0; |
|
119 | |||
120 | 7 | foreach ( $words as $key => $word ) { |
|
121 | |||
122 | 7 | $word = $synonymizer->synonymize( $word ); |
|
123 | |||
124 | // If it is not an exemption and less than the required minimum length |
||
125 | // or identified as stop word it is removed |
||
126 | 7 | if ( !isset( $this->whiteList[$word] ) && ( mb_strlen( $word ) < $minLength || $stopwordAnalyzer->isStopWord( $word ) ) ) { |
|
127 | 6 | continue; |
|
128 | } |
||
129 | |||
130 | // Simple proximity, check for same words appearing next to each other |
||
131 | 7 | if ( isset( $index[$pos-1] ) && $index[$pos-1] === $word ) { |
|
132 | 2 | continue; |
|
133 | } |
||
134 | |||
135 | 7 | $index[] = trim( $word ); |
|
136 | 7 | $pos++; |
|
137 | 7 | } |
|
138 | |||
139 | 7 | return implode( ' ' , $index ); |
|
140 | } |
||
141 | |||
142 | /** |
||
143 | * @since 0.1 |
||
144 | */ |
||
145 | 5 | public function toLowercase() { |
|
146 | 5 | $this->string = Normalizer::toLowercase( $this->string ); |
|
147 | 5 | } |
|
148 | |||
149 | /** |
||
150 | * @since 0.1 |
||
151 | * |
||
152 | * @param integer $length |
||
153 | */ |
||
154 | 3 | public function reduceLengthTo( $length ) { |
|
155 | 3 | $this->string = Normalizer::reduceLengthTo( $this->string, $length ); |
|
156 | 3 | } |
|
157 | |||
158 | /** |
||
159 | * @see http://www.phpwact.org/php/i18n/utf-8#str_replace |
||
160 | * @since 0.1 |
||
161 | * |
||
162 | * @param string $search |
||
163 | * @param string $replace |
||
164 | */ |
||
165 | 1 | public function replace( $search, $replace ) { |
|
166 | 1 | $this->string = str_replace( $search, $replace, $this->string ); |
|
167 | 1 | } |
|
168 | |||
169 | /** |
||
170 | * @since 0.1 |
||
171 | * |
||
172 | * @return string |
||
173 | */ |
||
174 | 6 | public function __toString() { |
|
175 | 6 | return $this->string; |
|
176 | } |
||
177 | |||
178 | } |
||
179 |
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..