Viewing file:  SegmentWriter.php (19.97 KB) -rw-rw-rw-
Select action/file-type:
 (+) |  (+) |  (+) | Code (+) | Session (+) |  (+) | SDB (+) |  (+) |  (+) |  (+) |  (+) |  (+) |
<?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 license@zend.com so we can send you a copy immediately.
*
* @category Zend
* @package Zend_Search_Lucene
* @subpackage Index
* @copyright Copyright (c) 2005-2009 Zend Technologies USA Inc. (http://www.zend.com)
* @license http://framework.zend.com/license/new-bsd New BSD License
* @version $Id: SegmentWriter.php 18947 2009-11-12 11:57:17Z alexander $
*/
/** Zend_Search_Lucene_Index_FieldInfo */
require_once 'Zend/Search/Lucene/Index/FieldInfo.php';
/** Zend_Search_Lucene_Index_Term */
require_once 'Zend/Search/Lucene/Index/Term.php';
/** Zend_Search_Lucene_Index_TermInfo */
require_once 'Zend/Search/Lucene/Index/TermInfo.php';
/**
* @category Zend
* @package Zend_Search_Lucene
* @subpackage Index
* @copyright Copyright (c) 2005-2009 Zend Technologies USA Inc. (http://www.zend.com)
* @license http://framework.zend.com/license/new-bsd New BSD License
*/
abstract class Zend_Search_Lucene_Index_SegmentWriter
{
/**
* Expert: The fraction of terms in the "dictionary" which should be stored
* in RAM. Smaller values use more memory, but make searching slightly
* faster, while larger values use less memory and make searching slightly
* slower. Searching is typically not dominated by dictionary lookup, so
* tweaking this is rarely useful.
*
* @var integer
*/
public static $indexInterval = 128;
/**
* Expert: The fraction of TermDocs entries stored in skip tables.
* Larger values result in smaller indexes, greater acceleration, but fewer
* accelerable cases, while smaller values result in bigger indexes,
* less acceleration and more
* accelerable cases. More detailed experiments would be useful here.
*
* 0x7FFFFFFF indicates that we don't use skip data
*
* Note: not used in current implementation
*
* @var integer
*/
public static $skipInterval = 0x7FFFFFFF;
/**
* Expert: The maximum number of skip levels. Smaller values result in
* slightly smaller indexes, but slower skipping in big posting lists.
*
* 0 indicates that we don't use skip data
*
* Note: not used in current implementation
*
* @var integer
*/
public static $maxSkipLevels = 0;
/**
* Number of docs in a segment
*
* @var integer
*/
protected $_docCount = 0;
/**
* Segment name
*
* @var string
*/
protected $_name;
/**
* File system adapter.
*
* @var Zend_Search_Lucene_Storage_Directory
*/
protected $_directory;
/**
* List of the index files.
* Used for automatic compound file generation
*
* @var unknown_type
*/
protected $_files = array();
/**
* Segment fields. Array of Zend_Search_Lucene_Index_FieldInfo objects for this segment
*
* @var array
*/
protected $_fields = array();
/**
* Normalization factors.
* An array fieldName => normVector
* normVector is a binary string.
* Each byte corresponds to an indexed document in a segment and
* encodes normalization factor (float value, encoded by
* Zend_Search_Lucene_Search_Similarity::encodeNorm())
*
* @var array
*/
protected $_norms = array();
/**
* '.fdx' file - Stored Fields, the field index.
*
* @var Zend_Search_Lucene_Storage_File
*/
protected $_fdxFile = null;
/**
* '.fdt' file - Stored Fields, the field data.
*
* @var Zend_Search_Lucene_Storage_File
*/
protected $_fdtFile = null;
/**
* Object constructor.
*
* @param Zend_Search_Lucene_Storage_Directory $directory
* @param string $name
*/
public function __construct(Zend_Search_Lucene_Storage_Directory $directory, $name)
{
$this->_directory = $directory;
$this->_name = $name;
}
/**
* Add field to the segment
*
* Returns actual field number
*
* @param Zend_Search_Lucene_Field $field
* @return integer
*/
public function addField(Zend_Search_Lucene_Field $field)
{
if (!isset($this->_fields[$field->name])) {
$fieldNumber = count($this->_fields);
$this->_fields[$field->name] =
new Zend_Search_Lucene_Index_FieldInfo($field->name,
$field->isIndexed,
$fieldNumber,
$field->storeTermVector);
return $fieldNumber;
} else {
$this->_fields[$field->name]->isIndexed |= $field->isIndexed;
$this->_fields[$field->name]->storeTermVector |= $field->storeTermVector;
return $this->_fields[$field->name]->number;
}
}
/**
* Add fieldInfo to the segment
*
* Returns actual field number
*
* @param Zend_Search_Lucene_Index_FieldInfo $fieldInfo
* @return integer
*/
public function addFieldInfo(Zend_Search_Lucene_Index_FieldInfo $fieldInfo)
{
if (!isset($this->_fields[$fieldInfo->name])) {
$fieldNumber = count($this->_fields);
$this->_fields[$fieldInfo->name] =
new Zend_Search_Lucene_Index_FieldInfo($fieldInfo->name,
$fieldInfo->isIndexed,
$fieldNumber,
$fieldInfo->storeTermVector);
return $fieldNumber;
} else {
$this->_fields[$fieldInfo->name]->isIndexed |= $fieldInfo->isIndexed;
$this->_fields[$fieldInfo->name]->storeTermVector |= $fieldInfo->storeTermVector;
return $this->_fields[$fieldInfo->name]->number;
}
}
/**
* Returns array of FieldInfo objects.
*
* @return array
*/
public function getFieldInfos()
{
return $this->_fields;
}
/**
* Add stored fields information
*
* @param array $storedFields array of Zend_Search_Lucene_Field objects
*/
public function addStoredFields($storedFields)
{
if (!isset($this->_fdxFile)) {
$this->_fdxFile = $this->_directory->createFile($this->_name . '.fdx');
$this->_fdtFile = $this->_directory->createFile($this->_name . '.fdt');
$this->_files[] = $this->_name . '.fdx';
$this->_files[] = $this->_name . '.fdt';
}
$this->_fdxFile->writeLong($this->_fdtFile->tell());
$this->_fdtFile->writeVInt(count($storedFields));
foreach ($storedFields as $field) {
$this->_fdtFile->writeVInt($this->_fields[$field->name]->number);
$fieldBits = ($field->isTokenized ? 0x01 : 0x00) |
($field->isBinary ? 0x02 : 0x00) |
0x00; /* 0x04 - third bit, compressed (ZLIB) */
$this->_fdtFile->writeByte($fieldBits);
if ($field->isBinary) {
$this->_fdtFile->writeVInt(strlen($field->value));
$this->_fdtFile->writeBytes($field->value);
} else {
$this->_fdtFile->writeString($field->getUtf8Value());
}
}
$this->_docCount++;
}
/**
* Returns the total number of documents in this segment.
*
* @return integer
*/
public function count()
{
return $this->_docCount;
}
/**
* Return segment name
*
* @return string
*/
public function getName()
{
return $this->_name;
}
/**
* Dump Field Info (.fnm) segment file
*/
protected function _dumpFNM()
{
$fnmFile = $this->_directory->createFile($this->_name . '.fnm');
$fnmFile->writeVInt(count($this->_fields));
$nrmFile = $this->_directory->createFile($this->_name . '.nrm');
// Write header
$nrmFile->writeBytes('NRM');
// Write format specifier
$nrmFile->writeByte((int)0xFF);
foreach ($this->_fields as $field) {
$fnmFile->writeString($field->name);
$fnmFile->writeByte(($field->isIndexed ? 0x01 : 0x00) |
($field->storeTermVector ? 0x02 : 0x00)
// not supported yet 0x04 /* term positions are stored with the term vectors */ |
// not supported yet 0x08 /* term offsets are stored with the term vectors */ |
);
if ($field->isIndexed) {
// pre-2.1 index mode (not used now)
// $normFileName = $this->_name . '.f' . $field->number;
// $fFile = $this->_directory->createFile($normFileName);
// $fFile->writeBytes($this->_norms[$field->name]);
// $this->_files[] = $normFileName;
$nrmFile->writeBytes($this->_norms[$field->name]);
}
}
$this->_files[] = $this->_name . '.fnm';
$this->_files[] = $this->_name . '.nrm';
}
/**
* Term Dictionary file
*
* @var Zend_Search_Lucene_Storage_File
*/
private $_tisFile = null;
/**
* Term Dictionary index file
*
* @var Zend_Search_Lucene_Storage_File
*/
private $_tiiFile = null;
/**
* Frequencies file
*
* @var Zend_Search_Lucene_Storage_File
*/
private $_frqFile = null;
/**
* Positions file
*
* @var Zend_Search_Lucene_Storage_File
*/
private $_prxFile = null;
/**
* Number of written terms
*
* @var integer
*/
private $_termCount;
/**
* Last saved term
*
* @var Zend_Search_Lucene_Index_Term
*/
private $_prevTerm;
/**
* Last saved term info
*
* @var Zend_Search_Lucene_Index_TermInfo
*/
private $_prevTermInfo;
/**
* Last saved index term
*
* @var Zend_Search_Lucene_Index_Term
*/
private $_prevIndexTerm;
/**
* Last saved index term info
*
* @var Zend_Search_Lucene_Index_TermInfo
*/
private $_prevIndexTermInfo;
/**
* Last term dictionary file position
*
* @var integer
*/
private $_lastIndexPosition;
/**
* Create dicrionary, frequency and positions files and write necessary headers
*/
public function initializeDictionaryFiles()
{
$this->_tisFile = $this->_directory->createFile($this->_name . '.tis');
$this->_tisFile->writeInt((int)0xFFFFFFFD);
$this->_tisFile->writeLong(0 /* dummy data for terms count */);
$this->_tisFile->writeInt(self::$indexInterval);
$this->_tisFile->writeInt(self::$skipInterval);
$this->_tisFile->writeInt(self::$maxSkipLevels);
$this->_tiiFile = $this->_directory->createFile($this->_name . '.tii');
$this->_tiiFile->writeInt((int)0xFFFFFFFD);
$this->_tiiFile->writeLong(0 /* dummy data for terms count */);
$this->_tiiFile->writeInt(self::$indexInterval);
$this->_tiiFile->writeInt(self::$skipInterval);
$this->_tiiFile->writeInt(self::$maxSkipLevels);
/** Dump dictionary header */
$this->_tiiFile->writeVInt(0); // preffix length
$this->_tiiFile->writeString(''); // suffix
$this->_tiiFile->writeInt((int)0xFFFFFFFF); // field number
$this->_tiiFile->writeByte((int)0x0F);
$this->_tiiFile->writeVInt(0); // DocFreq
$this->_tiiFile->writeVInt(0); // FreqDelta
$this->_tiiFile->writeVInt(0); // ProxDelta
$this->_tiiFile->writeVInt(24); // IndexDelta
$this->_frqFile = $this->_directory->createFile($this->_name . '.frq');
$this->_prxFile = $this->_directory->createFile($this->_name . '.prx');
$this->_files[] = $this->_name . '.tis';
$this->_files[] = $this->_name . '.tii';
$this->_files[] = $this->_name . '.frq';
$this->_files[] = $this->_name . '.prx';
$this->_prevTerm = null;
$this->_prevTermInfo = null;
$this->_prevIndexTerm = null;
$this->_prevIndexTermInfo = null;
$this->_lastIndexPosition = 24;
$this->_termCount = 0;
}
/**
* Add term
*
* Term positions is an array( docId => array(pos1, pos2, pos3, ...), ... )
*
* @param Zend_Search_Lucene_Index_Term $termEntry
* @param array $termDocs
*/
public function addTerm($termEntry, $termDocs)
{
$freqPointer = $this->_frqFile->tell();
$proxPointer = $this->_prxFile->tell();
$prevDoc = 0;
foreach ($termDocs as $docId => $termPositions) {
$docDelta = ($docId - $prevDoc)*2;
$prevDoc = $docId;
if (count($termPositions) > 1) {
$this->_frqFile->writeVInt($docDelta);
$this->_frqFile->writeVInt(count($termPositions));
} else {
$this->_frqFile->writeVInt($docDelta + 1);
}
$prevPosition = 0;
foreach ($termPositions as $position) {
$this->_prxFile->writeVInt($position - $prevPosition);
$prevPosition = $position;
}
}
if (count($termDocs) >= self::$skipInterval) {
/**
* @todo Write Skip Data to a freq file.
* It's not used now, but make index more optimal
*/
$skipOffset = $this->_frqFile->tell() - $freqPointer;
} else {
$skipOffset = 0;
}
$term = new Zend_Search_Lucene_Index_Term($termEntry->text,
$this->_fields[$termEntry->field]->number);
$termInfo = new Zend_Search_Lucene_Index_TermInfo(count($termDocs),
$freqPointer, $proxPointer, $skipOffset);
$this->_dumpTermDictEntry($this->_tisFile, $this->_prevTerm, $term, $this->_prevTermInfo, $termInfo);
if (($this->_termCount + 1) % self::$indexInterval == 0) {
$this->_dumpTermDictEntry($this->_tiiFile, $this->_prevIndexTerm, $term, $this->_prevIndexTermInfo, $termInfo);
$indexPosition = $this->_tisFile->tell();
$this->_tiiFile->writeVInt($indexPosition - $this->_lastIndexPosition);
$this->_lastIndexPosition = $indexPosition;
}
$this->_termCount++;
}
/**
* Close dictionary
*/
public function closeDictionaryFiles()
{
$this->_tisFile->seek(4);
$this->_tisFile->writeLong($this->_termCount);
$this->_tiiFile->seek(4);
// + 1 is used to count an additional special index entry (empty term at the start of the list)
$this->_tiiFile->writeLong(($this->_termCount - $this->_termCount % self::$indexInterval)/self::$indexInterval + 1);
}
/**
* Dump Term Dictionary segment file entry.
* Used to write entry to .tis or .tii files
*
* @param Zend_Search_Lucene_Storage_File $dicFile
* @param Zend_Search_Lucene_Index_Term $prevTerm
* @param Zend_Search_Lucene_Index_Term $term
* @param Zend_Search_Lucene_Index_TermInfo $prevTermInfo
* @param Zend_Search_Lucene_Index_TermInfo $termInfo
*/
protected function _dumpTermDictEntry(Zend_Search_Lucene_Storage_File $dicFile,
&$prevTerm, Zend_Search_Lucene_Index_Term $term,
&$prevTermInfo, Zend_Search_Lucene_Index_TermInfo $termInfo)
{
if (isset($prevTerm) && $prevTerm->field == $term->field) {
$matchedBytes = 0;
$maxBytes = min(strlen($prevTerm->text), strlen($term->text));
while ($matchedBytes < $maxBytes &&
$prevTerm->text[$matchedBytes] == $term->text[$matchedBytes]) {
$matchedBytes++;
}
// Calculate actual matched UTF-8 pattern
$prefixBytes = 0;
$prefixChars = 0;
while ($prefixBytes < $matchedBytes) {
$charBytes = 1;
if ((ord($term->text[$prefixBytes]) & 0xC0) == 0xC0) {
$charBytes++;
if (ord($term->text[$prefixBytes]) & 0x20 ) {
$charBytes++;
if (ord($term->text[$prefixBytes]) & 0x10 ) {
$charBytes++;
}
}
}
if ($prefixBytes + $charBytes > $matchedBytes) {
// char crosses matched bytes boundary
// skip char
break;
}
$prefixChars++;
$prefixBytes += $charBytes;
}
// Write preffix length
$dicFile->writeVInt($prefixChars);
// Write suffix
$dicFile->writeString(substr($term->text, $prefixBytes));
} else {
// Write preffix length
$dicFile->writeVInt(0);
// Write suffix
$dicFile->writeString($term->text);
}
// Write field number
$dicFile->writeVInt($term->field);
// DocFreq (the count of documents which contain the term)
$dicFile->writeVInt($termInfo->docFreq);
$prevTerm = $term;
if (!isset($prevTermInfo)) {
// Write FreqDelta
$dicFile->writeVInt($termInfo->freqPointer);
// Write ProxDelta
$dicFile->writeVInt($termInfo->proxPointer);
} else {
// Write FreqDelta
$dicFile->writeVInt($termInfo->freqPointer - $prevTermInfo->freqPointer);
// Write ProxDelta
$dicFile->writeVInt($termInfo->proxPointer - $prevTermInfo->proxPointer);
}
// Write SkipOffset - it's not 0 when $termInfo->docFreq > self::$skipInterval
if ($termInfo->skipOffset != 0) {
$dicFile->writeVInt($termInfo->skipOffset);
}
$prevTermInfo = $termInfo;
}
/**
* Generate compound index file
*/
protected function _generateCFS()
{
$cfsFile = $this->_directory->createFile($this->_name . '.cfs');
$cfsFile->writeVInt(count($this->_files));
$dataOffsetPointers = array();
foreach ($this->_files as $fileName) {
$dataOffsetPointers[$fileName] = $cfsFile->tell();
$cfsFile->writeLong(0); // write dummy data
$cfsFile->writeString($fileName);
}
foreach ($this->_files as $fileName) {
// Get actual data offset
$dataOffset = $cfsFile->tell();
// Seek to the data offset pointer
$cfsFile->seek($dataOffsetPointers[$fileName]);
// Write actual data offset value
$cfsFile->writeLong($dataOffset);
// Seek back to the end of file
$cfsFile->seek($dataOffset);
$dataFile = $this->_directory->getFileObject($fileName);
$byteCount = $this->_directory->fileLength($fileName);
while ($byteCount > 0) {
$data = $dataFile->readBytes(min($byteCount, 131072 /*128Kb*/));
$byteCount -= strlen($data);
$cfsFile->writeBytes($data);
}
$this->_directory->deleteFile($fileName);
}
}
/**
* Close segment, write it to disk and return segment info
*
* @return Zend_Search_Lucene_Index_SegmentInfo
*/
abstract public function close();
}
|