L-electron-Rare/moodle
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

upgrades Zend engine to version 1.5.0

Browse Source
This commit is contained in:
diml
2008-04-07 21:23:13 +00:00
parent 185441df1a
commit 3702d88a6e
9 changed files with 1484 additions and 32 deletions
Download Patch File Download Diff File Expand all files Collapse all files
search/Zend/Search/Lucene/Analysis/Analyzer/Common/Utf8/CaseInsensitive.php
+48
View File
@@ -0,0 +1,48 @@
<?php
/**
* Zend Framework
*
* LICENSE
*
* This source file is subject to the new BSD license that is bundled
* with this package in the file LICENSE.txt.
* It is also available through the world-wide-web at this URL:
* http://framework.zend.com/license/new-bsd
* If you did not receive a copy of the license and are unable to
* obtain it through the world-wide-web, please send an email
* to [email protected] so we can send you a copy immediately.
*
* @category Zend
* @package Zend_Search_Lucene
* @subpackage Analysis
* @copyright Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
* @license http://framework.zend.com/license/new-bsd New BSD License
*/
/** Zend_Search_Lucene_Analysis_Analyzer_Common_Utf8 */
require_once 'Zend/Search/Lucene/Analysis/Analyzer/Common/Utf8.php';
/** Zend_Search_Lucene_Analysis_TokenFilter_LowerCaseUtf8 */
require_once 'Zend/Search/Lucene/Analysis/TokenFilter/LowerCaseUtf8.php';
/**
* @category Zend
* @package Zend_Search_Lucene
* @subpackage Analysis
* @copyright Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
* @license http://framework.zend.com/license/new-bsd New BSD License
*/
class Zend_Search_Lucene_Analysis_Analyzer_Common_Utf8_CaseInsensitive extends Zend_Search_Lucene_Analysis_Analyzer_Common_Utf8
{
public function __construct()
{
parent::__construct();
$this->addFilter(new Zend_Search_Lucene_Analysis_TokenFilter_LowerCaseUtf8());
}
}
search/Zend/Search/Lucene/Analysis/Analyzer/Common/Utf8Num/CaseInsensitive.php
+48
View File
@@ -0,0 +1,48 @@
<?php
/**
* Zend Framework
*
* LICENSE
*
* This source file is subject to the new BSD license that is bundled
* with this package in the file LICENSE.txt.
* It is also available through the world-wide-web at this URL:
* http://framework.zend.com/license/new-bsd
* If you did not receive a copy of the license and are unable to
* obtain it through the world-wide-web, please send an email
* to [email protected] so we can send you a copy immediately.
*
* @category Zend
* @package Zend_Search_Lucene
* @subpackage Analysis
* @copyright Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
* @license http://framework.zend.com/license/new-bsd New BSD License
*/
/** Zend_Search_Lucene_Analysis_Analyzer_Common_Utf8Num */
require_once 'Zend/Search/Lucene/Analysis/Analyzer/Common/Utf8Num.php';
/** Zend_Search_Lucene_Analysis_TokenFilter_LowerCaseUtf8 */
require_once 'Zend/Search/Lucene/Analysis/TokenFilter/LowerCaseUtf8.php';
/**
* @category Zend
* @package Zend_Search_Lucene
* @subpackage Analysis
* @copyright Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
* @license http://framework.zend.com/license/new-bsd New BSD License
*/
class Zend_Search_Lucene_Analysis_Analyzer_Common_Utf8Num_CaseInsensitive extends Zend_Search_Lucene_Analysis_Analyzer_Common_Utf8Num
{
public function __construct()
{
parent::__construct();
$this->addFilter(new Zend_Search_Lucene_Analysis_TokenFilter_LowerCaseUtf8());
}
}
search/Zend/Search/Lucene/Analysis/TokenFilter/LowerCaseUtf8.php
+69
View File
@@ -0,0 +1,69 @@
<?php
/**
* Zend Framework
*
* LICENSE
*
* This source file is subject to the new BSD license that is bundled
* with this package in the file LICENSE.txt.
* It is also available through the world-wide-web at this URL:
* http://framework.zend.com/license/new-bsd
* If you did not receive a copy of the license and are unable to
* obtain it through the world-wide-web, please send an email
* to [email protected] so we can send you a copy immediately.
*
* @category Zend
* @package Zend_Search_Lucene
* @subpackage Analysis
* @copyright Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
* @license http://framework.zend.com/license/new-bsd New BSD License
*/
/** Zend_Search_Lucene_Analysis_TokenFilter */
require_once 'Zend/Search/Lucene/Analysis/TokenFilter.php';
/**
* Lower case Token filter.
*
* @category Zend
* @package Zend_Search_Lucene
* @subpackage Analysis
* @copyright Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
* @license http://framework.zend.com/license/new-bsd New BSD License
*/
class Zend_Search_Lucene_Analysis_TokenFilter_LowerCaseUtf8 extends Zend_Search_Lucene_Analysis_TokenFilter
{
/**
* Object constructor
*/
public function __construct()
{
if (!function_exists('mb_strtolower')) {
// mbstring extension is disabled
require_once 'Zend/Search/Lucene/Exception.php';
throw new Zend_Search_Lucene_Exception('Utf8 compatible lower case filter needs mbstring extension to be enabled.');
}
}
/**
* Normalize Token or remove it (if null is returned)
*
* @param Zend_Search_Lucene_Analysis_Token $srcToken
* @return Zend_Search_Lucene_Analysis_Token
*/
public function normalize(Zend_Search_Lucene_Analysis_Token $srcToken)
{
$newToken = new Zend_Search_Lucene_Analysis_Token(
mb_strtolower($srcToken->getTermText(), 'UTF-8'),
$srcToken->getStartOffset(),
$srcToken->getEndOffset());
$newToken->setPositionIncrement($srcToken->getPositionIncrement());
return $newToken;
}
}
search/Zend/Search/Lucene/EncodingConverter.php
-32
View File
@@ -1,32 +0,0 @@
<?php
class EncodingConverter {
private $last_error,
$in_encoding,
$out_encoding;
function __construct($in_encoding, $out_encoding) {
$this->in_encoding = $in_encoding;
$this->out_encoding = $out_encoding;
} //constructor
function handleError($err, $msg) {
$this->last_error = $msg;
} //handleError
function convert($str) {
$this->last_error = FALSE;
set_error_handler(array(&$this, 'handleError'));
$ret = iconv($this->in_encoding, $this->out_encoding, $str);
restore_error_handler();
return $ret;
} //convert
function getLastError() {
return $this->last_error;
} //getLastError
} //EncodingConverter
?>
search/Zend/Search/Lucene/LockManager.php
+161
View File
@@ -0,0 +1,161 @@
<?php
/**
* Zend Framework
*
* LICENSE
*
* This source file is subject to the new BSD license that is bundled
* with this package in the file LICENSE.txt.
* It is also available through the world-wide-web at this URL:
* http://framework.zend.com/license/new-bsd
* If you did not receive a copy of the license and are unable to
* obtain it through the world-wide-web, please send an email
* to [email protected] so we can send you a copy immediately.
*
* @category Zend
* @package Zend_Search_Lucene
* @copyright Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
* @license http://framework.zend.com/license/new-bsd New BSD License
*/
/** Zend_Search_Lucene_Exception */
require_once 'Zend/Search/Lucene/Exception.php';
/** Zend_Search_Lucene_Storage_Directory */
require_once 'Zend/Search/Lucene/Storage/Directory.php';
/** Zend_Search_Lucene_Storage_File */
require_once 'Zend/Search/Lucene/Storage/File.php';
/**
* This is an utility class which provides index locks processing functionality
*
* @category Zend
* @package Zend_Search_Lucene
* @copyright Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
* @license http://framework.zend.com/license/new-bsd New BSD License
*/
class Zend_Search_Lucene_LockManager
{
const WRITE_LOCK_FILE = 'write.lock.file';
const READ_LOCK_FILE = 'read.lock.file';
const OPTIMIZATION_LOCK_FILE = 'optimization.lock.file';
/**
* Obtain exclusive write lock on the index
*
* @param Zend_Search_Lucene_Storage_Directory $lockDirectory
* @return Zend_Search_Lucene_Storage_File
* @throws Zend_Search_Lucene_Exception
*/
public static function obtainWriteLock(Zend_Search_Lucene_Storage_Directory $lockDirectory)
{
$lock = $lockDirectory->createFile(self::WRITE_LOCK_FILE);
if (!$lock->lock(LOCK_EX)) {
throw new Zend_Search_Lucene_Exception('Can\'t obtain exclusive index lock');
}
return $lock;
}
/**
* Release exclusive write lock
*
* @param Zend_Search_Lucene_Storage_Directory $lockDirectory
*/
public static function releaseWriteLock(Zend_Search_Lucene_Storage_Directory $lockDirectory)
{
$lock = $lockDirectory->getFileObject(self::WRITE_LOCK_FILE);
$lock->unlock();
}
/**
* Obtain shared read lock on the index
*
* It doesn't block other read or update processes, but prevent index from the premature cleaning-up
*
* @param Zend_Search_Lucene_Storage_Directory $defaultLockDirectory
* @return Zend_Search_Lucene_Storage_File
* @throws Zend_Search_Lucene_Exception
*/
public static function obtainReadLock(Zend_Search_Lucene_Storage_Directory $lockDirectory)
{
$lock = $lockDirectory->createFile(self::READ_LOCK_FILE);
if (!$lock->lock(LOCK_SH)) {
throw new Zend_Search_Lucene_Exception('Can\'t obtain shared reading index lock');
}
return $lock;
}
/**
* Release shared read lock
*
* @param Zend_Search_Lucene_Storage_Directory $lockDirectory
*/
public static function releaseReadLock(Zend_Search_Lucene_Storage_Directory $lockDirectory)
{
$lock = $lockDirectory->getFileObject(self::READ_LOCK_FILE);
$lock->unlock();
}
/**
* Escalate Read lock to exclusive level
*
* @param Zend_Search_Lucene_Storage_Directory $lockDirectory
* @return boolean
*/
public static function escalateReadLock(Zend_Search_Lucene_Storage_Directory $lockDirectory)
{
$lock = $lockDirectory->getFileObject(self::READ_LOCK_FILE);
// Try to escalate read lock
if (!$lock->lock(LOCK_EX, true)) {
// Restore lock state
$lock->lock(LOCK_SH);
return false;
}
return true;
}
/**
* De-escalate Read lock to shared level
*
* @param Zend_Search_Lucene_Storage_Directory $lockDirectory
*/
public static function deEscalateReadLock(Zend_Search_Lucene_Storage_Directory $lockDirectory)
{
$lock = $lockDirectory->getFileObject(self::READ_LOCK_FILE);
$lock->lock(LOCK_SH);
}
/**
* Obtain exclusive optimization lock on the index
*
* Returns lock object on success and false otherwise (doesn't block execution)
*
* @param Zend_Search_Lucene_Storage_Directory $lockDirectory
* @return mixed
*/
public static function obtainOptimizationLock(Zend_Search_Lucene_Storage_Directory $lockDirectory)
{
$lock = $lockDirectory->createFile(self::OPTIMIZATION_LOCK_FILE);
if (!$lock->lock(LOCK_EX, true)) {
return false;
}
return $lock;
}
/**
* Release exclusive optimization lock
*
* @param Zend_Search_Lucene_Storage_Directory $lockDirectory
*/
public static function releaseOptimizationLock(Zend_Search_Lucene_Storage_Directory $lockDirectory)
{
$lock = $lockDirectory->getFileObject(self::OPTIMIZATION_LOCK_FILE);
$lock->unlock();
}
}
search/Zend/Search/Lucene/Search/Query/Fuzzy.php
+390
View File
@@ -0,0 +1,390 @@
<?php
/**
* Zend Framework
*
* LICENSE
*
* This source file is subject to the new BSD license that is bundled
* with this package in the file LICENSE.txt.
* It is also available through the world-wide-web at this URL:
* http://framework.zend.com/license/new-bsd
* If you did not receive a copy of the license and are unable to
* obtain it through the world-wide-web, please send an email
* to [email protected] so we can send you a copy immediately.
*
* @category Zend
* @package Zend_Search_Lucene
* @subpackage Search
* @copyright Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
* @license http://framework.zend.com/license/new-bsd New BSD License
*/
/** Zend_Search_Lucene_Search_Query */
require_once 'Zend/Search/Lucene/Search/Query.php';
/** Zend_Search_Lucene_Search_Query_MultiTerm */
require_once 'Zend/Search/Lucene/Search/Query/MultiTerm.php';
/**
* @category Zend
* @package Zend_Search_Lucene
* @subpackage Search
* @copyright Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
* @license http://framework.zend.com/license/new-bsd New BSD License
*/
class Zend_Search_Lucene_Search_Query_Fuzzy extends Zend_Search_Lucene_Search_Query
{
/** Default minimum similarity */
const DEFAULT_MIN_SIMILARITY = 0.5;
/**
* Maximum number of matched terms.
* Apache Lucene defines this limitation as boolean query maximum number of clauses:
* org.apache.lucene.search.BooleanQuery.getMaxClauseCount()
*/
const MAX_CLAUSE_COUNT = 1024;
/**
* Array of precalculated max distances
*
* keys are integers representing a word size
*/
private $_maxDistances = array();
/**
* Base searching term.
*
* @var Zend_Search_Lucene_Index_Term
*/
private $_term;
/**
* A value between 0 and 1 to set the required similarity
* between the query term and the matching terms. For example, for a
* _minimumSimilarity of 0.5 a term of the same length
* as the query term is considered similar to the query term if the edit distance
* between both terms is less than length(term)*0.5
*
* @var float
*/
private $_minimumSimilarity;
/**
* The length of common (non-fuzzy) prefix
*
* @var integer
*/
private $_prefixLength;
/**
* Matched terms.
*
* Matched terms list.
* It's filled during the search (rewrite operation) and may be used for search result
* post-processing
*
* Array of Zend_Search_Lucene_Index_Term objects
*
* @var array
*/
private $_matches = null;
/**
* Matched terms scores
*
* @var array
*/
private $_scores = null;
/**
* Array of the term keys.
* Used to sort terms in alphabetical order if terms have the same socres
*
* @var array
*/
private $_termKeys = null;
/**
* Zend_Search_Lucene_Search_Query_Wildcard constructor.
*
* @param Zend_Search_Lucene_Index_Term $pattern
* @throws Zend_Search_Lucene_Exception
*/
public function __construct(Zend_Search_Lucene_Index_Term $term, $minimumSimilarity = self::DEFAULT_MIN_SIMILARITY, $prefixLength = 0)
{
if ($minimumSimilarity < 0) {
throw new Zend_Search_Lucene_Exception('minimumSimilarity cannot be less than 0');
}
if ($minimumSimilarity >= 1) {
throw new Zend_Search_Lucene_Exception('minimumSimilarity cannot be greater than or equal to 1');
}
if ($prefixLength < 0) {
throw new Zend_Search_Lucene_Exception('prefixLength cannot be less than 0');
}
$this->_term = $term;
$this->_minimumSimilarity = $minimumSimilarity;
$this->_prefixLength = $prefixLength;
}
/**
* Calculate maximum distance for specified word length
*
* @param integer $prefixLength
* @param integer $termLength
* @param integer $length
* @return integer
*/
private function _calculateMaxDistance($prefixLength, $termLength, $length)
{
$this->_maxDistances[$length] = (int) ((1 - $this->_minimumSimilarity)*(min($termLength, $length) + $prefixLength));
return $this->_maxDistances[$length];
}
/**
* Re-write query into primitive queries in the context of specified index
*
* @param Zend_Search_Lucene_Interface $index
* @return Zend_Search_Lucene_Search_Query
*/
public function rewrite(Zend_Search_Lucene_Interface $index)
{
$this->_matches = array();
$this->_scores = array();
$this->_termKeys = array();
if ($this->_term->field === null) {
// Search through all fields
$fields = $index->getFieldNames(true /* indexed fields list */);
} else {
$fields = array($this->_term->field);
}
$prefix = Zend_Search_Lucene_Index_Term::getPrefix($this->_term->text, $this->_prefixLength);
$prefixByteLength = strlen($prefix);
$prefixUtf8Length = Zend_Search_Lucene_Index_Term::getLength($prefix);
$termLength = Zend_Search_Lucene_Index_Term::getLength($this->_term->text);
$termRest = substr($this->_term->text, $prefixByteLength);
// we calculate length of the rest in bytes since levenshtein() is not UTF-8 compatible
$termRestLength = strlen($termRest);
$scaleFactor = 1/(1 - $this->_minimumSimilarity);
foreach ($fields as $field) {
$index->resetTermsStream();
if ($prefix != '') {
$index->skipTo(new Zend_Search_Lucene_Index_Term($prefix, $field));
while ($index->currentTerm() !== null &&
$index->currentTerm()->field == $field &&
substr($index->currentTerm()->text, 0, $prefixByteLength) == $prefix) {
// Calculate similarity
$target = substr($index->currentTerm()->text, $prefixByteLength);
$maxDistance = isset($this->_maxDistances[strlen($target)])?
$this->_maxDistances[strlen($target)] :
$this->_calculateMaxDistance($prefixUtf8Length, $termRestLength, strlen($target));
if ($termRestLength == 0) {
// we don't have anything to compare. That means if we just add
// the letters for current term we get the new word
$similarity = (($prefixUtf8Length == 0)? 0 : 1 - strlen($target)/$prefixUtf8Length);
} else if (strlen($target) == 0) {
$similarity = (($prefixUtf8Length == 0)? 0 : 1 - $termRestLength/$prefixUtf8Length);
} else if ($maxDistance < abs($termRestLength - strlen($target))){
//just adding the characters of term to target or vice-versa results in too many edits
//for example "pre" length is 3 and "prefixes" length is 8. We can see that
//given this optimal circumstance, the edit distance cannot be less than 5.
//which is 8-3 or more precisesly abs(3-8).
//if our maximum edit distance is 4, then we can discard this word
//without looking at it.
$similarity = 0;
} else {
$similarity = 1 - levenshtein($termRest, $target)/($prefixUtf8Length + min($termRestLength, strlen($target)));
}
if ($similarity > $this->_minimumSimilarity) {
$this->_matches[] = $index->currentTerm();
$this->_termKeys[] = $index->currentTerm()->key();
$this->_scores[] = ($similarity - $this->_minimumSimilarity)*$scaleFactor;
}
$index->nextTerm();
}
} else {
$index->skipTo(new Zend_Search_Lucene_Index_Term('', $field));
while ($index->currentTerm() !== null && $index->currentTerm()->field == $field) {
// Calculate similarity
$target = $index->currentTerm()->text;
$maxDistance = isset($this->_maxDistances[strlen($target)])?
$this->_maxDistances[strlen($target)] :
$this->_calculateMaxDistance(0, $termRestLength, strlen($target));
if ($maxDistance < abs($termRestLength - strlen($target))){
//just adding the characters of term to target or vice-versa results in too many edits
//for example "pre" length is 3 and "prefixes" length is 8. We can see that
//given this optimal circumstance, the edit distance cannot be less than 5.
//which is 8-3 or more precisesly abs(3-8).
//if our maximum edit distance is 4, then we can discard this word
//without looking at it.
$similarity = 0;
} else {
$similarity = 1 - levenshtein($termRest, $target)/min($termRestLength, strlen($target));
}
if ($similarity > $this->_minimumSimilarity) {
$this->_matches[] = $index->currentTerm();
$this->_termKeys[] = $index->currentTerm()->key();
$this->_scores[] = ($similarity - $this->_minimumSimilarity)*$scaleFactor;
}
$index->nextTerm();
}
}
$index->closeTermsStream();
}
if (count($this->_matches) == 0) {
return new Zend_Search_Lucene_Search_Query_Empty();
} else if (count($this->_matches) == 1) {
return new Zend_Search_Lucene_Search_Query_Term(reset($this->_matches));
} else {
$rewrittenQuery = new Zend_Search_Lucene_Search_Query_Boolean();
array_multisort($this->_scores, SORT_DESC, SORT_NUMERIC,
$this->_termKeys, SORT_ASC, SORT_STRING,
$this->_matches);
$termCount = 0;
foreach ($this->_matches as $id => $matchedTerm) {
$subquery = new Zend_Search_Lucene_Search_Query_Term($matchedTerm);
$subquery->setBoost($this->_scores[$id]);
$rewrittenQuery->addSubquery($subquery);
$termCount++;
if ($termCount >= self::MAX_CLAUSE_COUNT) {
break;
}
}
return $rewrittenQuery;
}
}
/**
* Optimize query in the context of specified index
*
* @param Zend_Search_Lucene_Interface $index
* @return Zend_Search_Lucene_Search_Query
*/
public function optimize(Zend_Search_Lucene_Interface $index)
{
throw new Zend_Search_Lucene_Exception('Wildcard query should not be directly used for search. Use $query->rewrite($index)');
}
/**
* Return query terms
*
* @return array
* @throws Zend_Search_Lucene_Exception
*/
public function getQueryTerms()
{
if ($this->_matches === null) {
throw new Zend_Search_Lucene_Exception('Search has to be performed first to get matched terms');
}
return $this->_matches;
}
/**
* Constructs an appropriate Weight implementation for this query.
*
* @param Zend_Search_Lucene_Interface $reader
* @return Zend_Search_Lucene_Search_Weight
* @throws Zend_Search_Lucene_Exception
*/
public function createWeight(Zend_Search_Lucene_Interface $reader)
{
throw new Zend_Search_Lucene_Exception('Wildcard query should not be directly used for search. Use $query->rewrite($index)');
}
/**
* Execute query in context of index reader
* It also initializes necessary internal structures
*
* @param Zend_Search_Lucene_Interface $reader
* @throws Zend_Search_Lucene_Exception
*/
public function execute(Zend_Search_Lucene_Interface $reader)
{
throw new Zend_Search_Lucene_Exception('Wildcard query should not be directly used for search. Use $query->rewrite($index)');
}
/**
* Get document ids likely matching the query
*
* It's an array with document ids as keys (performance considerations)
*
* @return array
* @throws Zend_Search_Lucene_Exception
*/
public function matchedDocs()
{
throw new Zend_Search_Lucene_Exception('Wildcard query should not be directly used for search. Use $query->rewrite($index)');
}
/**
* Score specified document
*
* @param integer $docId
* @param Zend_Search_Lucene_Interface $reader
* @return float
* @throws Zend_Search_Lucene_Exception
*/
public function score($docId, Zend_Search_Lucene_Interface $reader)
{
throw new Zend_Search_Lucene_Exception('Wildcard query should not be directly used for search. Use $query->rewrite($index)');
}
/**
* Highlight query terms
*
* @param integer &$colorIndex
* @param Zend_Search_Lucene_Document_Html $doc
*/
public function highlightMatchesDOM(Zend_Search_Lucene_Document_Html $doc, &$colorIndex)
{
$words = array();
foreach ($this->_matches as $term) {
$words[] = $term->text;
}
$doc->highlight($words, $this->_getHighlightColor($colorIndex));
}
/**
* Print a query
*
* @return string
*/
public function __toString()
{
// It's used only for query visualisation, so we don't care about characters escaping
return (($this->_term->field === null)? '' : $this->_term->field . ':')
. $this->_term->text . '~'
. (($this->_minimumSimilarity != self::DEFAULT_MIN_SIMILARITY)? round($this->_minimumSimilarity, 4) : '');
}
}
search/Zend/Search/Lucene/Search/Query/Insignificant.php
+140
View File
@@ -0,0 +1,140 @@
<?php
/**
* Zend Framework
*
* LICENSE
*
* This source file is subject to the new BSD license that is bundled
* with this package in the file LICENSE.txt.
* It is also available through the world-wide-web at this URL:
* http://framework.zend.com/license/new-bsd
* If you did not receive a copy of the license and are unable to
* obtain it through the world-wide-web, please send an email
* to [email protected] so we can send you a copy immediately.
*
* @category Zend
* @package Zend_Search_Lucene
* @subpackage Search
* @copyright Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
* @license http://framework.zend.com/license/new-bsd New BSD License
*/
/** Zend_Search_Lucene_Search_Query */
require_once 'Zend/Search/Lucene/Search/Query.php';
/** Zend_Search_Lucene_Search_Weight_Empty */
require_once 'Zend/Search/Lucene/Search/Weight/Empty.php';
/**
* The insignificant query returns empty result, but doesn't limit result set as a part of other queries
*
* @category Zend
* @package Zend_Search_Lucene
* @subpackage Search
* @copyright Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
* @license http://framework.zend.com/license/new-bsd New BSD License
*/
class Zend_Search_Lucene_Search_Query_Insignificant extends Zend_Search_Lucene_Search_Query
{
/**
* Re-write query into primitive queries in the context of specified index
*
* @param Zend_Search_Lucene_Interface $index
* @return Zend_Search_Lucene_Search_Query
*/
public function rewrite(Zend_Search_Lucene_Interface $index)
{
return $this;
}
/**
* Optimize query in the context of specified index
*
* @param Zend_Search_Lucene_Interface $index
* @return Zend_Search_Lucene_Search_Query
*/
public function optimize(Zend_Search_Lucene_Interface $index)
{
return $this;
}
/**
* Constructs an appropriate Weight implementation for this query.
*
* @param Zend_Search_Lucene_Interface $reader
* @return Zend_Search_Lucene_Search_Weight
*/
public function createWeight(Zend_Search_Lucene_Interface $reader)
{
return new Zend_Search_Lucene_Search_Weight_Empty();
}
/**
* Execute query in context of index reader
* It also initializes necessary internal structures
*
* @param Zend_Search_Lucene_Interface $reader
*/
public function execute(Zend_Search_Lucene_Interface $reader)
{
// Do nothing
}
/**
* Get document ids likely matching the query
*
* It's an array with document ids as keys (performance considerations)
*
* @return array
*/
public function matchedDocs()
{
return array();
}
/**
* Score specified document
*
* @param integer $docId
* @param Zend_Search_Lucene_Interface $reader
* @return float
*/
public function score($docId, Zend_Search_Lucene_Interface $reader)
{
return 0;
}
/**
* Return query terms
*
* @return array
*/
public function getQueryTerms()
{
return array();
}
/**
* Highlight query terms
*
* @param integer &$colorIndex
* @param Zend_Search_Lucene_Document_Html $doc
*/
public function highlightMatchesDOM(Zend_Search_Lucene_Document_Html $doc, &$colorIndex)
{
// Do nothing
}
/**
* Print a query
*
* @return string
*/
public function __toString()
{
return '<InsignificantQuery>';
}
}
search/Zend/Search/Lucene/Search/Query/Range.php
+331
View File
@@ -0,0 +1,331 @@
<?php
/**
* Zend Framework
*
* LICENSE
*
* This source file is subject to the new BSD license that is bundled
* with this package in the file LICENSE.txt.
* It is also available through the world-wide-web at this URL:
* http://framework.zend.com/license/new-bsd
* If you did not receive a copy of the license and are unable to
* obtain it through the world-wide-web, please send an email
* to [email protected] so we can send you a copy immediately.
*
* @category Zend
* @package Zend_Search_Lucene
* @subpackage Search
* @copyright Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
* @license http://framework.zend.com/license/new-bsd New BSD License
*/
/** Zend_Search_Lucene_Search_Query */
require_once 'Zend/Search/Lucene/Search/Query.php';
/** Zend_Search_Lucene_Search_Query_MultiTerm */
require_once 'Zend/Search/Lucene/Search/Query/MultiTerm.php';
/**
* @category Zend
* @package Zend_Search_Lucene
* @subpackage Search
* @copyright Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
* @license http://framework.zend.com/license/new-bsd New BSD License
*/
class Zend_Search_Lucene_Search_Query_Range extends Zend_Search_Lucene_Search_Query
{
/**
* Lower term.
*
* @var Zend_Search_Lucene_Index_Term
*/
private $_lowerTerm;
/**
* Upper term.
*
* @var Zend_Search_Lucene_Index_Term
*/
private $_upperTerm;
/**
* Search field
*
* @var string
*/
private $_field;
/**
* Inclusive
*
* @var boolean
*/
private $_inclusive;
/**
* Matched terms.
*
* Matched terms list.
* It's filled during the search (rewrite operation) and may be used for search result
* post-processing
*
* Array of Zend_Search_Lucene_Index_Term objects
*
* @var array
*/
private $_matches;
/**
* Zend_Search_Lucene_Search_Query_Range constructor.
*
* @param Zend_Search_Lucene_Index_Term|null $lowerTerm
* @param Zend_Search_Lucene_Index_Term|null $upperTerm
* @param boolean $inclusive
* @throws Zend_Search_Lucene_Exception
*/
public function __construct($lowerTerm, $upperTerm, $inclusive)
{
if ($lowerTerm === null && $upperTerm === null) {
throw new Zend_Search_Lucene_Exception('At least one term must be non-null');
}
if ($lowerTerm !== null && $upperTerm !== null && $lowerTerm->field != $upperTerm->field) {
throw new Zend_Search_Lucene_Exception('Both terms must be for the same field');
}
$this->_field = ($lowerTerm !== null)? $lowerTerm->field : $upperTerm->field;
$this->_lowerTerm = $lowerTerm;
$this->_upperTerm = $upperTerm;
$this->_inclusive = $inclusive;
}
/**
* Get query field name
*
* @return string|null
*/
public function getField()
{
return $this->_field;
}
/**
* Get lower term
*
* @return Zend_Search_Lucene_Index_Term|null
*/
public function getLowerTerm()
{
return $this->_lowerTerm;
}
/**
* Get upper term
*
* @return Zend_Search_Lucene_Index_Term|null
*/
public function getUpperTerm()
{
return $this->_upperTerm;
}
/**
* Get upper term
*
* @return boolean
*/
public function isInclusive()
{
return $this->_inclusive;
}
/**
* Re-write query into primitive queries in the context of specified index
*
* @param Zend_Search_Lucene_Interface $index
* @return Zend_Search_Lucene_Search_Query
*/
public function rewrite(Zend_Search_Lucene_Interface $index)
{
$this->_matches = array();
if ($this->_field === null) {
// Search through all fields
$fields = $index->getFieldNames(true /* indexed fields list */);
} else {
$fields = array($this->_field);
}
foreach ($fields as $field) {
$index->resetTermsStream();
if ($this->_lowerTerm !== null) {
$lowerTerm = new Zend_Search_Lucene_Index_Term($this->_lowerTerm->text, $field);
$index->skipTo($lowerTerm);
if (!$this->_inclusive &&
$index->currentTerm() == $lowerTerm) {
// Skip lower term
$index->nextTerm();
}
} else {
$index->skipTo(new Zend_Search_Lucene_Index_Term('', $field));
}
if ($this->_upperTerm !== null) {
// Walk up to the upper term
$upperTerm = new Zend_Search_Lucene_Index_Term($this->_upperTerm->text, $field);
while ($index->currentTerm() !== null &&
$index->currentTerm()->field == $field &&
$index->currentTerm()->text < $upperTerm->text) {
$this->_matches[] = $index->currentTerm();
$index->nextTerm();
}
if ($this->_inclusive && $index->currentTerm() == $upperTerm) {
// Include upper term into result
$this->_matches[] = $upperTerm;
}
} else {
// Walk up to the end of field data
while ($index->currentTerm() !== null && $index->currentTerm()->field == $field) {
$this->_matches[] = $index->currentTerm();
$index->nextTerm();
}
}
$index->closeTermsStream();
}
if (count($this->_matches) == 0) {
return new Zend_Search_Lucene_Search_Query_Empty();
} else if (count($this->_matches) == 1) {
return new Zend_Search_Lucene_Search_Query_Term(reset($this->_matches));
} else {
$rewrittenQuery = new Zend_Search_Lucene_Search_Query_MultiTerm();
foreach ($this->_matches as $matchedTerm) {
$rewrittenQuery->addTerm($matchedTerm);
}
return $rewrittenQuery;
}
}
/**
* Optimize query in the context of specified index
*
* @param Zend_Search_Lucene_Interface $index
* @return Zend_Search_Lucene_Search_Query
*/
public function optimize(Zend_Search_Lucene_Interface $index)
{
throw new Zend_Search_Lucene_Exception('Range query should not be directly used for search. Use $query->rewrite($index)');
}
/**
* Return query terms
*
* @return array
* @throws Zend_Search_Lucene_Exception
*/
public function getQueryTerms()
{
if ($this->_matches === null) {
throw new Zend_Search_Lucene_Exception('Search has to be performed first to get matched terms');
}
return $this->_matches;
}
/**
* Constructs an appropriate Weight implementation for this query.
*
* @param Zend_Search_Lucene_Interface $reader
* @return Zend_Search_Lucene_Search_Weight
* @throws Zend_Search_Lucene_Exception
*/
public function createWeight(Zend_Search_Lucene_Interface $reader)
{
throw new Zend_Search_Lucene_Exception('Range query should not be directly used for search. Use $query->rewrite($index)');
}
/**
* Execute query in context of index reader
* It also initializes necessary internal structures
*
* @param Zend_Search_Lucene_Interface $reader
* @throws Zend_Search_Lucene_Exception
*/
public function execute(Zend_Search_Lucene_Interface $reader)
{
throw new Zend_Search_Lucene_Exception('Range query should not be directly used for search. Use $query->rewrite($index)');
}
/**
* Get document ids likely matching the query
*
* It's an array with document ids as keys (performance considerations)
*
* @return array
* @throws Zend_Search_Lucene_Exception
*/
public function matchedDocs()
{
throw new Zend_Search_Lucene_Exception('Range query should not be directly used for search. Use $query->rewrite($index)');
}
/**
* Score specified document
*
* @param integer $docId
* @param Zend_Search_Lucene_Interface $reader
* @return float
* @throws Zend_Search_Lucene_Exception
*/
public function score($docId, Zend_Search_Lucene_Interface $reader)
{
throw new Zend_Search_Lucene_Exception('Range query should not be directly used for search. Use $query->rewrite($index)');
}
/**
* Highlight query terms
*
* @param integer &$colorIndex
* @param Zend_Search_Lucene_Document_Html $doc
*/
public function highlightMatchesDOM(Zend_Search_Lucene_Document_Html $doc, &$colorIndex)
{
$words = array();
foreach ($this->_matches as $term) {
$words[] = $term->text;
}
$doc->highlight($words, $this->_getHighlightColor($colorIndex));
}
/**
* Print a query
*
* @return string
*/
public function __toString()
{
// It's used only for query visualisation, so we don't care about characters escaping
return (($this->_field === null)? '' : $this->_field . ':')
. (($this->_inclusive)? '[' : '{')
. (($this->_lowerTerm !== null)? $this->_lowerTerm->text : 'null')
. ' TO '
. (($this->_upperTerm !== null)? $this->_upperTerm->text : 'null')
. (($this->_inclusive)? ']' : '}');
}
}
search/Zend/Search/Lucene/Search/Query/Wildcard.php
+297
View File
@@ -0,0 +1,297 @@
<?php
/**
* Zend Framework
*
* LICENSE
*
* This source file is subject to the new BSD license that is bundled
* with this package in the file LICENSE.txt.
* It is also available through the world-wide-web at this URL:
* http://framework.zend.com/license/new-bsd
* If you did not receive a copy of the license and are unable to
* obtain it through the world-wide-web, please send an email
* to [email protected] so we can send you a copy immediately.
*
* @category Zend
* @package Zend_Search_Lucene
* @subpackage Search
* @copyright Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
* @license http://framework.zend.com/license/new-bsd New BSD License
*/
/** Zend_Search_Lucene_Search_Query */
require_once 'Zend/Search/Lucene/Search/Query.php';
/** Zend_Search_Lucene_Search_Query_MultiTerm */
require_once 'Zend/Search/Lucene/Search/Query/MultiTerm.php';
/**
* @category Zend
* @package Zend_Search_Lucene
* @subpackage Search
* @copyright Copyright (c) 2005-2008 Zend Technologies USA Inc. (http://www.zend.com)
* @license http://framework.zend.com/license/new-bsd New BSD License
*/
class Zend_Search_Lucene_Search_Query_Wildcard extends Zend_Search_Lucene_Search_Query
{
/**
* Search pattern.
*
* Field has to be fully specified or has to be null
* Text may contain '*' or '?' symbols
*
* @var Zend_Search_Lucene_Index_Term
*/
private $_pattern;
/**
* Matched terms.
*
* Matched terms list.
* It's filled during the search (rewrite operation) and may be used for search result
* post-processing
*
* Array of Zend_Search_Lucene_Index_Term objects
*
* @var array
*/
private $_matches = null;
/**
* Zend_Search_Lucene_Search_Query_Wildcard constructor.
*
* @param Zend_Search_Lucene_Index_Term $pattern
*/
public function __construct(Zend_Search_Lucene_Index_Term $pattern)
{
$this->_pattern = $pattern;
}
/**
* Get terms prefix
*
* @param string $word
* @return string
*/
private static function _getPrefix($word)
{
$questionMarkPosition = strpos($word, '?');
$astrericPosition = strpos($word, '*');
if ($questionMarkPosition !== false) {
if ($astrericPosition !== false) {
return substr($word, 0, min($questionMarkPosition, $astrericPosition));
}
return substr($word, 0, $questionMarkPosition);
} else if ($astrericPosition !== false) {
return substr($word, 0, $astrericPosition);
}
return $word;
}
/**
* Re-write query into primitive queries in the context of specified index
*
* @param Zend_Search_Lucene_Interface $index
* @return Zend_Search_Lucene_Search_Query
*/
public function rewrite(Zend_Search_Lucene_Interface $index)
{
$this->_matches = array();
if ($this->_pattern->field === null) {
// Search through all fields
$fields = $index->getFieldNames(true /* indexed fields list */);
} else {
$fields = array($this->_pattern->field);
}
$prefix = self::_getPrefix($this->_pattern->text);
$prefixLength = strlen($prefix);
$matchExpression = '/^' . str_replace(array('\\?', '\\*'), array('.', '.*') , preg_quote($this->_pattern->text, '/')) . '$/';
/** @todo check for PCRE unicode support may be performed through Zend_Environment in some future */
if (@preg_match('/\pL/u', 'a') == 1) {
// PCRE unicode support is turned on
// add Unicode modifier to the match expression
$matchExpression .= 'u';
}
foreach ($fields as $field) {
$index->resetTermsStream();
if ($prefix != '') {
$index->skipTo(new Zend_Search_Lucene_Index_Term($prefix, $field));
while ($index->currentTerm() !== null &&
$index->currentTerm()->field == $field &&
substr($index->currentTerm()->text, 0, $prefixLength) == $prefix) {
if (preg_match($matchExpression, $index->currentTerm()->text) === 1) {
$this->_matches[] = $index->currentTerm();
}
$index->nextTerm();
}
} else {
$index->skipTo(new Zend_Search_Lucene_Index_Term('', $field));
while ($index->currentTerm() !== null && $index->currentTerm()->field == $field) {
if (preg_match($matchExpression, $index->currentTerm()->text) === 1) {
$this->_matches[] = $index->currentTerm();
}
$index->nextTerm();
}
}
$index->closeTermsStream();
}
if (count($this->_matches) == 0) {
return new Zend_Search_Lucene_Search_Query_Empty();
} else if (count($this->_matches) == 1) {
return new Zend_Search_Lucene_Search_Query_Term(reset($this->_matches));
} else {
$rewrittenQuery = new Zend_Search_Lucene_Search_Query_MultiTerm();
foreach ($this->_matches as $matchedTerm) {
$rewrittenQuery->addTerm($matchedTerm);
}
return $rewrittenQuery;
}
}
/**
* Optimize query in the context of specified index
*
* @param Zend_Search_Lucene_Interface $index
* @return Zend_Search_Lucene_Search_Query
*/
public function optimize(Zend_Search_Lucene_Interface $index)
{
throw new Zend_Search_Lucene_Exception('Wildcard query should not be directly used for search. Use $query->rewrite($index)');
}
/**
* Returns query pattern
*
* @return Zend_Search_Lucene_Index_Term
*/
public function getPattern()
{
return $this->_pattern;
}
/**
* Return query terms
*
* @return array
* @throws Zend_Search_Lucene_Exception
*/
public function getQueryTerms()
{
if ($this->_matches === null) {
throw new Zend_Search_Lucene_Exception('Search has to be performed first to get matched terms');
}
return $this->_matches;
}
/**
* Constructs an appropriate Weight implementation for this query.
*
* @param Zend_Search_Lucene_Interface $reader
* @return Zend_Search_Lucene_Search_Weight
* @throws Zend_Search_Lucene_Exception
*/
public function createWeight(Zend_Search_Lucene_Interface $reader)
{
throw new Zend_Search_Lucene_Exception('Wildcard query should not be directly used for search. Use $query->rewrite($index)');
}
/**
* Execute query in context of index reader
* It also initializes necessary internal structures
*
* @param Zend_Search_Lucene_Interface $reader
* @throws Zend_Search_Lucene_Exception
*/
public function execute(Zend_Search_Lucene_Interface $reader)
{
throw new Zend_Search_Lucene_Exception('Wildcard query should not be directly used for search. Use $query->rewrite($index)');
}
/**
* Get document ids likely matching the query
*
* It's an array with document ids as keys (performance considerations)
*
* @return array
* @throws Zend_Search_Lucene_Exception
*/
public function matchedDocs()
{
throw new Zend_Search_Lucene_Exception('Wildcard query should not be directly used for search. Use $query->rewrite($index)');
}
/**
* Score specified document
*
* @param integer $docId
* @param Zend_Search_Lucene_Interface $reader
* @return float
* @throws Zend_Search_Lucene_Exception
*/
public function score($docId, Zend_Search_Lucene_Interface $reader)
{
throw new Zend_Search_Lucene_Exception('Wildcard query should not be directly used for search. Use $query->rewrite($index)');
}
/**
* Highlight query terms
*
* @param integer &$colorIndex
* @param Zend_Search_Lucene_Document_Html $doc
*/
public function highlightMatchesDOM(Zend_Search_Lucene_Document_Html $doc, &$colorIndex)
{
$words = array();
$matchExpression = '/^' . str_replace(array('\\?', '\\*'), array('.', '.*') , preg_quote($this->_pattern->text, '/')) . '$/';
if (@preg_match('/\pL/u', 'a') == 1) {
// PCRE unicode support is turned on
// add Unicode modifier to the match expression
$matchExpression .= 'u';
}
$tokens = Zend_Search_Lucene_Analysis_Analyzer::getDefault()->tokenize($doc->getFieldUtf8Value('body'), 'UTF-8');
foreach ($tokens as $token) {
if (preg_match($matchExpression, $token->getTermText()) === 1) {
$words[] = $token->getTermText();
}
}
$doc->highlight($words, $this->_getHighlightColor($colorIndex));
}
/**
* Print a query
*
* @return string
*/
public function __toString()
{
// It's used only for query visualisation, so we don't care about characters escaping
return (($this->_pattern->field === null)? '' : $this->_pattern->field . ':') . $this->_pattern->text;
}
}