| 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 Analysis
|
|
|
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_Analysis_Analyzer_Common */
|
|
|
23 |
require_once 'Zend/Search/Lucene/Analysis/Analyzer/Common.php';
|
|
|
24 |
|
|
|
25 |
|
|
|
26 |
/**
|
|
|
27 |
* @package Zend_Search_Lucene
|
|
|
28 |
* @subpackage Analysis
|
|
|
29 |
* @copyright Copyright (c) 2005-2006 Zend Technologies USA Inc. (http://www.zend.com)
|
|
|
30 |
* @license http://www.zend.com/license/framework/1_0.txt Zend Framework License version 1.0
|
|
|
31 |
*/
|
|
|
32 |
|
|
|
33 |
class Zend_Search_Lucene_Analysis_Analyzer_Common_Text extends Zend_Search_Lucene_Analysis_Analyzer_Common
|
|
|
34 |
{
|
|
|
35 |
/**
|
|
|
36 |
* Tokenize text to a terms
|
|
|
37 |
* Returns array of Zend_Search_Lucene_Analysis_Token objects
|
|
|
38 |
*
|
|
|
39 |
* @param string $data
|
|
|
40 |
* @return array
|
|
|
41 |
*/
|
|
|
42 |
public function tokenize($data)
|
|
|
43 |
{
|
|
|
44 |
$tokenStream = array();
|
|
|
45 |
|
|
|
46 |
$position = 0;
|
|
|
47 |
while ($position < strlen($data)) {
|
|
|
48 |
// skip white space
|
|
|
49 |
while ($position < strlen($data) && !ctype_alpha( $data{$position} )) {
|
|
|
50 |
$position++;
|
|
|
51 |
}
|
|
|
52 |
|
|
|
53 |
$termStartPosition = $position;
|
|
|
54 |
|
|
|
55 |
// read token
|
|
|
56 |
while ($position < strlen($data) && ctype_alpha( $data{$position} )) {
|
|
|
57 |
$position++;
|
|
|
58 |
}
|
|
|
59 |
|
|
|
60 |
// Empty token, end of stream.
|
|
|
61 |
if ($position == $termStartPosition) {
|
|
|
62 |
break;
|
|
|
63 |
}
|
|
|
64 |
|
|
|
65 |
$token = new Zend_Search_Lucene_Analysis_Token(substr($data,
|
|
|
66 |
$termStartPosition,
|
|
|
67 |
$position-$termStartPosition),
|
|
|
68 |
$termStartPosition,
|
|
|
69 |
$position);
|
|
|
70 |
$tokenStream[] = $this->normalize($token);
|
|
|
71 |
}
|
|
|
72 |
|
|
|
73 |
return $tokenStream;
|
|
|
74 |
}
|
|
|
75 |
}
|
|
|
76 |
|