public function testGetPrefixUtf8()
{
// UTF-8 string with non-ascii symbols (Russian alphabet)
$this->assertEquals(Index\Term::getPrefix('абвгдеёжзийклмнопрстуфхцчшщьыъэюя', 64), 'абвгдеёжзийклмнопрстуфхцчшщьыъэюя');
$this->assertEquals(Index\Term::getPrefix('абвгдеёжзийклмнопрстуфхцчшщьыъэюя', 33), 'абвгдеёжзийклмнопрстуфхцчшщьыъэюя');
$this->assertEquals(Index\Term::getPrefix('абвгдеёжзийклмнопрстуфхцчшщьыъэюя', 4), 'абвг');
$this->assertEquals(Index\Term::getPrefix('абвгдеёжзийклмнопрстуфхцчшщьыъэюя', 0), '');
}
/**
* Query specific matches highlighting
*
* @param Highlighter $highlighter Highlighter object (also contains doc for highlighting)
*/
protected function _highlightMatches(Highlighter $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) / $prefixUtf8Length;
} elseif (strlen($target) == 0) {
$similarity = $prefixUtf8Length == 0 ? 0 : 1 - $termRestLength / $prefixUtf8Length;
} elseif ($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);
}
/**
* Adds a document to this segment.
*
* @param \ZendSearch\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][$this->_docCount] = array();
} elseif (!isset($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 * $field->boost));
}
} elseif (($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][$this->_docCount] = array();
} elseif (!isset($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);
}
/**
* Scans terms dictionary and returns next term
*
* @return \ZendSearch\Lucene\Index\Term|null
*/
public function nextTerm()
{
if ($this->_tisFile === null || $this->_termCount == 0) {
$this->_lastTerm = null;
$this->_lastTermInfo = null;
$this->_lastTermPositions = null;
$this->_docMap = null;
// may be necessary for "empty" segment
$this->_tisFile = null;
$this->_frqFile = null;
$this->_prxFile = null;
return null;
}
$termPrefixLength = $this->_tisFile->readVInt();
$termSuffix = $this->_tisFile->readString();
$termFieldNum = $this->_tisFile->readVInt();
$termValue = Term::getPrefix($this->_lastTerm->text, $termPrefixLength) . $termSuffix;
$this->_lastTerm = new Term($termValue, $this->_fields[$termFieldNum]->name);
$docFreq = $this->_tisFile->readVInt();
$freqPointer = $this->_lastTermInfo->freqPointer + $this->_tisFile->readVInt();
$proxPointer = $this->_lastTermInfo->proxPointer + $this->_tisFile->readVInt();
if ($docFreq >= $this->_skipInterval) {
$skipOffset = $this->_tisFile->readVInt();
} else {
$skipOffset = 0;
}
$this->_lastTermInfo = new TermInfo($docFreq, $freqPointer, $proxPointer, $skipOffset);
if ($this->_termsScanMode == self::SM_FULL_INFO || $this->_termsScanMode == self::SM_MERGE_INFO) {
$this->_lastTermPositions = array();
$this->_frqFile->seek($this->_lastTermInfo->freqPointer + $this->_frqFileOffset, SEEK_SET);
$freqs = array();
$docId = 0;
for ($count = 0; $count < $this->_lastTermInfo->docFreq; $count++) {
$docDelta = $this->_frqFile->readVInt();
if ($docDelta % 2 == 1) {
$docId += ($docDelta - 1) / 2;
$freqs[$docId] = 1;
} else {
$docId += $docDelta / 2;
$freqs[$docId] = $this->_frqFile->readVInt();
}
}
$this->_prxFile->seek($this->_lastTermInfo->proxPointer + $this->_prxFileOffset, SEEK_SET);
foreach ($freqs as $docId => $freq) {
$termPosition = 0;
$positions = array();
for ($count = 0; $count < $freq; $count++) {
$termPosition += $this->_prxFile->readVInt();
$positions[] = $termPosition;
}
if (isset($this->_docMap[$docId])) {
$this->_lastTermPositions[$this->_docMap[$docId]] = $positions;
}
}
}
$this->_termCount--;
if ($this->_termCount == 0) {
$this->_tisFile = null;
$this->_frqFile = null;
$this->_prxFile = null;
}
return $this->_lastTerm;
}