public function testAnalyzer() { $currentAnalyzer = Analyzer\Analyzer::getDefault(); $this->assertTrue($currentAnalyzer instanceof Analyzer); /** Zend_Search_Lucene_Analysis_Analyzer_Common_Utf8Num */ $newAnalyzer = new Common\Utf8Num(); Analyzer\Analyzer::setDefault($newAnalyzer); $this->assertTrue(Analyzer\Analyzer::getDefault() === $newAnalyzer); // Set analyzer to the default value (used in other tests) Analyzer\Analyzer::setDefault($currentAnalyzer); }
public function testFilteredTokensQueryParserProcessing() { $index = Lucene\Lucene::open(dirname(__FILE__) . '/_index23Sample/_files'); $this->assertEquals(count(\Zend\Search\Lucene\Analysis\Analyzer\Analyzer::getDefault()->tokenize('123456787654321')), 0); $hits = $index->find('"PEAR developers" AND Home AND 123456787654321'); $this->assertEquals(count($hits), 1); $expectedResultset = array(array(1, 0.16827, 'IndexSource/contributing.wishlist.html')); foreach ($hits as $resId => $hit) { $this->assertEquals($hit->id, $expectedResultset[$resId][0]); $this->assertTrue(abs($hit->score - $expectedResultset[$resId][1]) < 1.0E-6); $this->assertEquals($hit->path, $expectedResultset[$resId][2]); } }
/** * Query specific matches highlighting * * @param Highlighter $highlighter Highlighter object (also contains doc for highlighting) */ protected function _highlightMatches(Highlighter $highlighter) { $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'; } $docBody = $highlighter->getDocument()->getFieldUtf8Value('body'); $tokens = Analyzer::getDefault()->tokenize($docBody, 'UTF-8'); foreach ($tokens as $token) { if (preg_match($matchExpression, $token->getTermText()) === 1) { $words[] = $token->getTermText(); } } $highlighter->highlight($words); }
/** * Query specific matches highlighting * * @param \Zend\Search\Lucene\Search\Highlighter $highlighter Highlighter object (also contains doc for highlighting) */ protected function _highlightMatches(Highlighter $highlighter) { $words = array(); $docBody = $highlighter->getDocument()->getFieldUtf8Value('body'); $tokens = Lucene\Analysis\Analyzer\Analyzer::getDefault()->tokenize($docBody, 'UTF-8'); $lowerTermText = $this->_lowerTerm !== null ? $this->_lowerTerm->text : null; $upperTermText = $this->_upperTerm !== null ? $this->_upperTerm->text : null; if ($this->_inclusive) { foreach ($tokens as $token) { $termText = $token->getTermText(); if (($lowerTermText == null || $lowerTermText <= $termText) && ($upperTermText == null || $termText <= $upperTermText)) { $words[] = $termText; } } } else { foreach ($tokens as $token) { $termText = $token->getTermText(); if (($lowerTermText == null || $lowerTermText < $termText) && ($upperTermText == null || $termText < $upperTermText)) { $words[] = $termText; } } } $highlighter->highlight($words); }
/** * Highlight text using specified View helper or callback function. * * @param string|array $words Words to highlight. Words could be organized using the array or string. * @param callback $callback Callback method, used to transform (highlighting) text. * @param array $params Array of additionall callback parameters passed through into it * (first non-optional parameter is an HTML fragment for highlighting) * @throws \Zend\Search\Lucene\Exception\InvalidArgumentException * @return string */ public function highlightExtended($words, $callback, $params = array()) { if (!is_array($words)) { $words = array($words); } $wordsToHighlightList = array(); $analyzer = Analyzer\Analyzer::getDefault(); foreach ($words as $wordString) { $wordsToHighlightList[] = $analyzer->tokenize($wordString); } $wordsToHighlight = call_user_func_array('array_merge', $wordsToHighlightList); if (count($wordsToHighlight) == 0) { return $this->_doc->saveHTML(); } $wordsToHighlightFlipped = array(); foreach ($wordsToHighlight as $id => $token) { $wordsToHighlightFlipped[$token->getTermText()] = $id; } if (!is_callable($callback)) { throw new InvalidArgumentException('$viewHelper parameter mast be a View Helper name, View Helper object or callback.'); } $xpath = new \DOMXPath($this->_doc); $matchedNodes = $xpath->query("/html/body"); foreach ($matchedNodes as $matchedNode) { $this->_highlightNodeRecursive($matchedNode, $wordsToHighlightFlipped, $callback, $params); } }
/** * Query specific matches highlighting * * @param Highlighter $highlighter Highlighter object (also contains doc for highlighting) */ protected function _highlightMatches(Highlighter $highlighter) { /** Skip fields detection. We don't need it, since we expect all fields presented in the HTML body and don't differentiate them */ /** Skip exact term matching recognition, keyword fields highlighting is not supported */ /** Skip wildcard queries recognition. Supported wildcards are removed by text analyzer */ // tokenize phrase using current analyzer and process it as a phrase query $tokens = Analyzer::getDefault()->tokenize($this->_phrase, $this->_phraseEncoding); if (count($tokens) == 0) { // Do nothing return; } if (count($tokens) == 1) { $highlighter->highlight($tokens[0]->getTermText()); return; } //It's non-trivial phrase query $words = array(); foreach ($tokens as $token) { $words[] = $token->getTermText(); } $highlighter->highlight($words); }
/** * Process last range query term (closed interval) * * @throws \Zend\Search\Lucene\Search\Exception\QueryParserException */ public function closedRQLastTerm() { $tokens = Analyzer\Analyzer::getDefault()->tokenize($this->_rqFirstTerm, $this->_encoding); if (count($tokens) > 1) { throw new QueryParserException('Range query boundary terms must be non-multiple word terms'); } else { if (count($tokens) == 1) { $from = new Index\Term(reset($tokens)->getTermText(), $this->_context->getField()); } else { $from = null; } } $tokens = Analyzer\Analyzer::getDefault()->tokenize($this->_currentToken->text, $this->_encoding); if (count($tokens) > 1) { throw new QueryParserException('Range query boundary terms must be non-multiple word terms'); } else { if (count($tokens) == 1) { $to = new Index\Term(reset($tokens)->getTermText(), $this->_context->getField()); } else { $to = null; } } if ($from === null && $to === null) { throw new QueryParserException('At least one range query boundary term must be non-empty term'); } $rangeQuery = new Query\Range($from, $to, true); $entry = new QueryEntry\Subquery($rangeQuery); $this->_context->addEntry($entry); }
/** * Adds a document to this segment. * * @param \Zend\Search\Lucene\Document $document * @throws LuceneException\UnsupportedMethodCallException */ public function addDocument(Document $document) { $storedFields = array(); $docNorms = array(); $similarity = AbstractSimilarity::getDefault(); foreach ($document->getFieldNames() as $fieldName) { $field = $document->getField($fieldName); if ($field->storeTermVector) { /** * @todo term vector storing support */ throw new LuceneException\UnsupportedMethodCallException('Store term vector functionality is not supported yet.'); } if ($field->isIndexed) { if ($field->isTokenized) { $analyzer = Analyzer\Analyzer::getDefault(); $analyzer->setInput($field->value, $field->encoding); $position = 0; $tokenCounter = 0; while (($token = $analyzer->nextToken()) !== null) { $tokenCounter++; $term = new Index\Term($token->getTermText(), $field->name); $termKey = $term->key(); if (!isset($this->_termDictionary[$termKey])) { // New term $this->_termDictionary[$termKey] = $term; $this->_termDocs[$termKey] = array(); $this->_termDocs[$termKey][$this->_docCount] = array(); } else { if (!isset($this->_termDocs[$termKey][$this->_docCount])) { // Existing term, but new term entry $this->_termDocs[$termKey][$this->_docCount] = array(); } } $position += $token->getPositionIncrement(); $this->_termDocs[$termKey][$this->_docCount][] = $position; } if ($tokenCounter == 0) { // Field contains empty value. Treat it as non-indexed and non-tokenized $field = clone $field; $field->isIndexed = $field->isTokenized = false; } else { $docNorms[$field->name] = chr($similarity->encodeNorm($similarity->lengthNorm($field->name, $tokenCounter) * $document->boost * $field->boost)); } } else { if (($fieldUtf8Value = $field->getUtf8Value()) == '') { // Field contains empty value. Treat it as non-indexed and non-tokenized $field = clone $field; $field->isIndexed = $field->isTokenized = false; } else { $term = new Index\Term($fieldUtf8Value, $field->name); $termKey = $term->key(); if (!isset($this->_termDictionary[$termKey])) { // New term $this->_termDictionary[$termKey] = $term; $this->_termDocs[$termKey] = array(); $this->_termDocs[$termKey][$this->_docCount] = array(); } else { if (!isset($this->_termDocs[$termKey][$this->_docCount])) { // Existing term, but new term entry $this->_termDocs[$termKey][$this->_docCount] = array(); } } $this->_termDocs[$termKey][$this->_docCount][] = 0; // position $docNorms[$field->name] = chr($similarity->encodeNorm($similarity->lengthNorm($field->name, 1) * $document->boost * $field->boost)); } } } if ($field->isStored) { $storedFields[] = $field; } $this->addField($field); } foreach ($this->_fields as $fieldName => $field) { if (!$field->isIndexed) { continue; } if (!isset($this->_norms[$fieldName])) { $this->_norms[$fieldName] = str_repeat(chr($similarity->encodeNorm($similarity->lengthNorm($fieldName, 0))), $this->_docCount); } if (isset($docNorms[$fieldName])) { $this->_norms[$fieldName] .= $docNorms[$fieldName]; } else { $this->_norms[$fieldName] .= chr($similarity->encodeNorm($similarity->lengthNorm($fieldName, 0))); } } $this->addStoredFields($storedFields); }
/** * Query specific matches highlighting * * @param \Zend\Search\Lucene\Search\Highlighter $highlighter Highlighter object (also contains doc for highlighting) */ protected function _highlightMatches(Highlighter $highlighter) { /** Skip fields detection. We don't need it, since we expect all fields presented in the HTML body and don't differentiate them */ /** Skip exact term matching recognition, keyword fields highlighting is not supported */ // ------------------------------------- // Recognize wildcard queries /** @todo check for PCRE unicode support may be performed through Zend_Environment in some future */ if (@preg_match('/\\pL/u', 'a') == 1) { $word = iconv($this->_encoding, 'UTF-8', $this->_word); $wildcardsPattern = '/[*?]/u'; $subPatternsEncoding = 'UTF-8'; } else { $word = $this->_word; $wildcardsPattern = '/[*?]/'; $subPatternsEncoding = $this->_encoding; } $subPatterns = preg_split($wildcardsPattern, $word, -1, PREG_SPLIT_OFFSET_CAPTURE); if (count($subPatterns) > 1) { // Wildcard query is recognized $pattern = ''; foreach ($subPatterns as $id => $subPattern) { // Append corresponding wildcard character to the pattern before each sub-pattern (except first) if ($id != 0) { $pattern .= $word[$subPattern[1] - 1]; } // Check if each subputtern is a single word in terms of current analyzer $tokens = Analyzer\Analyzer::getDefault()->tokenize($subPattern[0], $subPatternsEncoding); if (count($tokens) > 1) { // Do nothing (nothing is highlighted) return; } foreach ($tokens as $token) { $pattern .= $token->getTermText(); } } $term = new Index\Term($pattern, $this->_field); $query = new Query\Wildcard($term); $query->_highlightMatches($highlighter); return; } // ------------------------------------- // Recognize one-term multi-term and "insignificant" queries $tokens = Analyzer\Analyzer::getDefault()->tokenize($this->_word, $this->_encoding); if (count($tokens) == 0) { // Do nothing return; } if (count($tokens) == 1) { $highlighter->highlight($tokens[0]->getTermText()); return; } //It's not insignificant or one term query $words = array(); foreach ($tokens as $token) { $words[] = $token->getTermText(); } $highlighter->highlight($words); }
/** * Query specific matches highlighting * * @param \Zend\Search\Lucene\Search\Highlighter\HighlighterInterface $highlighter Highlighter object (also contains doc for highlighting) */ protected function _highlightMatches(Highlighter\HighlighterInterface $highlighter) { $words = array(); $prefix = Index\Term::getPrefix($this->_term->text, $this->_prefixLength); $prefixByteLength = strlen($prefix); $prefixUtf8Length = Index\Term::getLength($prefix); $termLength = Index\Term::getLength($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'); $tokens = Lucene\Analysis\Analyzer\Analyzer::getDefault()->tokenize($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); }
/** * Query specific matches highlighting * * @param \Zend\Search\Lucene\Search\Highlighter $highlighter Highlighter object (also contains doc for highlighting) */ protected function _highlightMatches(Highlighter $highlighter) { /** Skip fields detection. We don't need it, since we expect all fields presented in the HTML body and don't differentiate them */ /** Skip exact term matching recognition, keyword fields highlighting is not supported */ // ------------------------------------- // Recognize wildcard queries /** @todo check for PCRE unicode support may be performed through Zend_Environment in some future */ if (@preg_match('/\\pL/u', 'a') == 1) { $subPatterns = preg_split('/[*?]/u', iconv($this->_encoding, 'UTF-8', $this->_word)); } else { $subPatterns = preg_split('/[*?]/', $this->_word); } if (count($subPatterns) > 1) { // Do nothing return; } // ------------------------------------- // Recognize one-term multi-term and "insignificant" queries $tokens = Analyzer\Analyzer::getDefault()->tokenize($this->_word, $this->_encoding); if (count($tokens) == 0) { // Do nothing return; } if (count($tokens) == 1) { $term = new Index\Term($tokens[0]->getTermText(), $this->_field); $query = new Query\Fuzzy($term, $this->_minimumSimilarity); $query->_highlightMatches($highlighter); return; } // Word is tokenized into several tokens // But fuzzy search is supported only for non-multiple word terms // Do nothing }