コード例 #1
0
ファイル: Fuzzy.php プロジェクト: Mendim/gtv-resources
 /**
  * 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);
     } 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;
         }
     }
 }
コード例 #2
0
ファイル: Fuzzy.php プロジェクト: netconstructor/Centurion
 /**
  * Query specific matches highlighting
  *
  * @param Zend_Search_Lucene_Search_Highlighter_Interface $highlighter  Highlighter object (also contains doc for highlighting)
  */
 protected function _highlightMatches(Zend_Search_Lucene_Search_Highlighter_Interface $highlighter)
 {
     $words = array();
     //$1 'Zend/Search/Lucene/Index/Term.php';
     $prefix = Zend_Search_Lucene_Index_Term::getPrefix($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);
     $docBody = $highlighter->getDocument()->getFieldUtf8Value('body');
     //$1 'Zend/Search/Lucene/Analysis/Analyzer.php';
     $tokens = Zend_Search_Lucene_Analysis_Analyzer::getDefault()->tokenize($docBody, 'UTF-8');
     foreach ($tokens as $token) {
         $termText = $token->getTermText();
         if (substr($termText, 0, $prefixByteLength) == $prefix) {
             // Calculate similarity
             $target = substr($termText, $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) {
                 $words[] = $termText;
             }
         }
     }
     $highlighter->highlight($words);
 }
コード例 #3
0
ファイル: Zend_Search_Lucene.php プロジェクト: Blu2z/implsk
 protected function _highlightMatches(Zend_Search_Lucene_Search_Highlighter_Interface $highlighter)
 {
     $words = array();
     $prefix = Zend_Search_Lucene_Index_Term::getPrefix($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);
     $termRestLength = strlen($termRest);
     $scaleFactor = 1 / (1 - $this->_minimumSimilarity);
     $docBody = $highlighter->getDocument()->getFieldUtf8Value('body');
     $tokens = Zend_Search_Lucene_Analysis_Analyzer::getDefault()->tokenize($docBody, 'UTF-8');
     foreach ($tokens as $token) {
         $termText = $token->getTermText();
         if (substr($termText, 0, $prefixByteLength) == $prefix) {
             $target = substr($termText, $prefixByteLength);
             $maxDistance = isset($this->_maxDistances[strlen($target)]) ? $this->_maxDistances[strlen($target)] : $this->_calculateMaxDistance($prefixUtf8Length, $termRestLength, strlen($target));
             if ($termRestLength == 0) {
                 $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))) {
                         $similarity = 0;
                     } else {
                         $similarity = 1 - levenshtein($termRest, $target) / ($prefixUtf8Length + min($termRestLength, strlen($target)));
                     }
                 }
             }
             if ($similarity > $this->_minimumSimilarity) {
                 $words[] = $termText;
             }
         }
     }
     $highlighter->highlight($words);
 }