| 1 |
lars |
1 |
<?php
|
|
|
2 |
/**
|
|
|
3 |
* Zend Framework
|
|
|
4 |
*
|
|
|
5 |
* LICENSE
|
|
|
6 |
*
|
|
|
7 |
* This source file is subject to version 1.0 of the Zend Framework
|
|
|
8 |
* license, that is bundled with this package in the file LICENSE, and
|
|
|
9 |
* is available through the world-wide-web at the following URL:
|
|
|
10 |
* http://www.zend.com/license/framework/1_0.txt. If you did not receive
|
|
|
11 |
* a copy of the Zend Framework license and are unable to obtain it
|
|
|
12 |
* through the world-wide-web, please send a note to license@zend.com
|
|
|
13 |
* so we can mail you a copy immediately.
|
|
|
14 |
*
|
|
|
15 |
* @package Zend_Search_Lucene
|
|
|
16 |
* @subpackage Search
|
|
|
17 |
* @copyright Copyright (c) 2005-2006 Zend Technologies USA Inc. (http://www.zend.com)
|
|
|
18 |
* @license http://www.zend.com/license/framework/1_0.txt Zend Framework License version 1.0
|
|
|
19 |
*/
|
|
|
20 |
|
|
|
21 |
|
|
|
22 |
/** Zend_Search_Lucene_Search_QueryToken */
|
|
|
23 |
require_once 'Zend/Search/Lucene/Search/QueryToken.php';
|
|
|
24 |
|
|
|
25 |
/** Zend_Search_Lucene_Exception */
|
|
|
26 |
require_once 'Zend/Search/Lucene/Exception.php';
|
|
|
27 |
|
|
|
28 |
|
|
|
29 |
/**
|
|
|
30 |
* @package Zend_Search_Lucene
|
|
|
31 |
* @subpackage Search
|
|
|
32 |
* @copyright Copyright (c) 2005-2006 Zend Technologies USA Inc. (http://www.zend.com)
|
|
|
33 |
* @license http://www.zend.com/license/framework/1_0.txt Zend Framework License version 1.0
|
|
|
34 |
*/
|
|
|
35 |
class Zend_Search_Lucene_Search_QueryTokenizer implements Iterator
|
|
|
36 |
{
|
|
|
37 |
/**
|
|
|
38 |
* inputString tokens.
|
|
|
39 |
*
|
|
|
40 |
* @var array
|
|
|
41 |
*/
|
|
|
42 |
protected $_tokens = array();
|
|
|
43 |
|
|
|
44 |
/**
|
|
|
45 |
* tokens pointer.
|
|
|
46 |
*
|
|
|
47 |
* @var integer
|
|
|
48 |
*/
|
|
|
49 |
protected $_currToken = 0;
|
|
|
50 |
|
|
|
51 |
|
|
|
52 |
/**
|
|
|
53 |
* QueryTokenize constructor needs query string as a parameter.
|
|
|
54 |
*
|
|
|
55 |
* @param string $inputString
|
|
|
56 |
*/
|
|
|
57 |
public function __construct($inputString)
|
|
|
58 |
{
|
|
|
59 |
if (!strlen($inputString)) {
|
|
|
60 |
throw new Zend_Search_Lucene_Exception('Cannot tokenize empty query string.');
|
|
|
61 |
}
|
|
|
62 |
|
|
|
63 |
$currentToken = '';
|
|
|
64 |
for ($count = 0; $count < strlen($inputString); $count++) {
|
|
|
65 |
if (ctype_alnum( $inputString{$count} )) {
|
|
|
66 |
$currentToken .= $inputString{$count};
|
|
|
67 |
} else {
|
|
|
68 |
// Previous token is finished
|
|
|
69 |
if (strlen($currentToken)) {
|
|
|
70 |
$this->_tokens[] = new Zend_Search_Lucene_Search_QueryToken(Zend_Search_Lucene_Search_QueryToken::TOKTYPE_WORD,
|
|
|
71 |
$currentToken);
|
|
|
72 |
$currentToken = '';
|
|
|
73 |
}
|
|
|
74 |
|
|
|
75 |
if ($inputString{$count} == '+' || $inputString{$count} == '-') {
|
|
|
76 |
$this->_tokens[] = new Zend_Search_Lucene_Search_QueryToken(Zend_Search_Lucene_Search_QueryToken::TOKTYPE_SIGN,
|
|
|
77 |
$inputString{$count});
|
|
|
78 |
} elseif ($inputString{$count} == '(' || $inputString{$count} == ')') {
|
|
|
79 |
$this->_tokens[] = new Zend_Search_Lucene_Search_QueryToken(Zend_Search_Lucene_Search_QueryToken::TOKTYPE_BRACKET,
|
|
|
80 |
$inputString{$count});
|
|
|
81 |
} elseif ($inputString{$count} == ':' && $this->count()) {
|
|
|
82 |
if ($this->_tokens[count($this->_tokens)-1]->type == Zend_Search_Lucene_Search_QueryToken::TOKTYPE_WORD) {
|
|
|
83 |
$this->_tokens[count($this->_tokens)-1]->type = Zend_Search_Lucene_Search_QueryToken::TOKTYPE_FIELD;
|
|
|
84 |
}
|
|
|
85 |
}
|
|
|
86 |
}
|
|
|
87 |
}
|
|
|
88 |
|
|
|
89 |
if (strlen($currentToken)) {
|
|
|
90 |
$this->_tokens[] = new Zend_Search_Lucene_Search_QueryToken(Zend_Search_Lucene_Search_QueryToken::TOKTYPE_WORD, $currentToken);
|
|
|
91 |
}
|
|
|
92 |
|
|
|
93 |
if (empty($this->_tokens))
|
|
|
94 |
$this->_tokens[] = $inputString;
|
|
|
95 |
}
|
|
|
96 |
|
|
|
97 |
|
|
|
98 |
/**
|
|
|
99 |
* Returns number of tokens
|
|
|
100 |
*
|
|
|
101 |
* @return integer
|
|
|
102 |
*/
|
|
|
103 |
public function count()
|
|
|
104 |
{
|
|
|
105 |
return count($this->_tokens);
|
|
|
106 |
}
|
|
|
107 |
|
|
|
108 |
|
|
|
109 |
/**
|
|
|
110 |
* Returns TRUE if a token exists at the current position.
|
|
|
111 |
*
|
|
|
112 |
* @return boolean
|
|
|
113 |
*/
|
|
|
114 |
public function valid()
|
|
|
115 |
{
|
|
|
116 |
return $this->_currToken < $this->count();
|
|
|
117 |
}
|
|
|
118 |
|
|
|
119 |
|
|
|
120 |
/**
|
|
|
121 |
* Resets token stream.
|
|
|
122 |
*
|
|
|
123 |
* @return integer
|
|
|
124 |
*/
|
|
|
125 |
public function rewind()
|
|
|
126 |
{
|
|
|
127 |
$this->_currToken = 0;
|
|
|
128 |
}
|
|
|
129 |
|
|
|
130 |
|
|
|
131 |
/**
|
|
|
132 |
* Returns the token at the current position or FALSE if
|
|
|
133 |
* the position does not contain a valid token.
|
|
|
134 |
*
|
|
|
135 |
* @return mixed
|
|
|
136 |
*/
|
|
|
137 |
public function current()
|
|
|
138 |
{
|
|
|
139 |
return $this->valid() ? $this->_tokens[$this->_currToken] : false;
|
|
|
140 |
}
|
|
|
141 |
|
|
|
142 |
|
|
|
143 |
/**
|
|
|
144 |
* Returns next token
|
|
|
145 |
*
|
|
|
146 |
* @return Zend_Search_Lucene_Search_QueryToken
|
|
|
147 |
*/
|
|
|
148 |
public function next()
|
|
|
149 |
{
|
|
|
150 |
return ++$this->_currToken;
|
|
|
151 |
}
|
|
|
152 |
|
|
|
153 |
|
|
|
154 |
/**
|
|
|
155 |
* Return the position of the current token.
|
|
|
156 |
*
|
|
|
157 |
* @return integer
|
|
|
158 |
*/
|
|
|
159 |
public function key()
|
|
|
160 |
{
|
|
|
161 |
return $this->_currToken;
|
|
|
162 |
}
|
|
|
163 |
|
|
|
164 |
}
|
|
|
165 |
|