indexreader.cpp

clucene是c++版的全文检索引擎,完全移植于lucene,采用 stl 编写.

/*------------------------------------------------------------------------------
* Copyright (C) 2003-2006 Ben van Klinken and the CLucene Team
* 
* Distributable under the terms of either the Apache License (Version 2.0) or 
* the GNU Lesser General Public License, as specified in the COPYING file.
------------------------------------------------------------------------------*/
#include "CLucene/StdHeader.h"
#include "IndexReader.h"

#include "CLucene/store/Directory.h"
#include "CLucene/store/FSDirectory.h"
#include "CLucene/store/Lock.h"
#include "CLucene/document/Document.h"
#include "CLucene/search/Similarity.h"
#include "SegmentInfos.h"
#include "MultiReader.h"
#include "Terms.h"

CL_NS_USE(util)
CL_NS_USE(store)
CL_NS_DEF(index)

  IndexReader::IndexReader(Directory* dir):
   directory(_CL_POINTER(dir)){
  //Constructor.
  //Func - Creates an instance of IndexReader
  //Pre  - true
  //Post - An instance has been created with writeLock = NULL

      writeLock = NULL;
      segmentInfos = NULL;
      directoryOwner = false;
      closeDirectory = false;
      stale = false;
      hasChanges = false;
  }

   IndexReader::IndexReader(Directory* directory, SegmentInfos* segmentInfos, bool closeDirectory) {
    this->directory = _CL_POINTER(directory);
    this->segmentInfos = segmentInfos;
    directoryOwner = true;
    this->closeDirectory = closeDirectory;
    stale = false;
    hasChanges = false;
    writeLock = NULL;
  }

  IndexReader::~IndexReader(){
  //Func - Destructor
  //       Destroys the instance and releases the writeLock if needed
  //Pre  - true
  //Post - The instance has been destroyed if pre(writeLock) exists is has been released
      if (writeLock != NULL) {
		  //release writeLock
          writeLock->release();
		  _CLDELETE(writeLock);
      }
	  _CLDELETE(segmentInfos);
	  _CLDECDELETE(directory);
  }

  IndexReader* IndexReader::open(const char* path){
  //Func - Static method.
  //       Returns an IndexReader reading the index in an FSDirectory in the named path. 
  //Pre  - path != NULL and contains the path of the index for which an IndexReader must be 
  //       instantiated
  //       closeDir indicates if the directory needs to be closed
  //Post - An IndexReader has been returned that reads tnhe index located at path

	  CND_PRECONDITION(path != NULL, "path is NULL");

	  Directory* dir = FSDirectory::getDirectory(path,false);
     IndexReader* reader = open(dir,true);
     //because fsdirectory will now have a refcount of 1 more than
     //if the reader had been opened with a directory object,
     //we need to do a refdec
     _CLDECDELETE(dir);
     return reader;
  }

  IndexReader* IndexReader::open( Directory* directory, bool closeDirectory){
  //Func - Static method.
  //       Returns an IndexReader reading the index in an FSDirectory in the named path. 
  //Pre  - directory represents a directory 
  //       closeDir indicates if the directory needs to be closed
  //Post - An IndexReader has been returned that reads the index located at directory

	  // in- & inter-process sync
      SCOPED_LOCK_MUTEX(directory->THIS_LOCK)
      
     IndexReader* ret = NULL;     

	  LuceneLock* lock = directory->makeLock("commit.lock");

	  //Instantiate an IndexReaderLockWith which can produce an IndexReader
      IndexReaderLockWith with(lock,directory);

	  try{
	  //Create an IndexReader reading the index
		ret = (IndexReader*)with.run();
	  }_CLFINALLY(
        _CLDELETE( lock );
	  );

	  ret->closeDirectory = closeDirectory;

	   CND_CONDITION(ret != NULL,"ret is NULL");
	   //return reference 
       return ret;
  }

  void* IndexReader::IndexReaderLockWith::doBody() {
  //Func - Reads the segmentinfo file and depending on the number of segments found
  //       it returns a SegmentsReader or a SegmentReader
  //Pre  - directory != NULL
  //Post - Depending on the number of Segments present in directory this method
  //       returns an empty SegmentsReader when there are no segments, a SegmentReader when
  //       directory contains 1 segment and a nonempty SegmentsReader when directory
  //       contains multiple segements

	   CND_PRECONDITION(directory != NULL, "directory is NULL");

	   //Instantiate SegmentInfos
       SegmentInfos* infos = _CLNEW SegmentInfos;
	   try{
			//Have SegmentInfos read the segments file in directory
			infos->read(directory);
	   }catch(...){
	        //make sure infos is cleaned up
			_CLDELETE(infos);
			throw;
	   }

       // If there is at least one segment (if infos.size() >= 1), the last
       // SegmentReader object will close the directory when the SegmentReader
       // object itself is closed (see SegmentReader::doClose).
       // If there are no segments, there will be no "last SegmentReader object"
       // to fulfill this responsibility, so we need to explicitly close the
       // directory in the segmentsreader.close
       
	   //Count the number segments in the directory
	   const uint32_t nSegs = infos->size();

       if (nSegs == 1 ) {
			// index is optimized 
            return _CLNEW SegmentReader(infos, infos->info(0));
	    }else{
			//Instantiate an array of pointers to SegmentReaders of size nSegs (The number of segments in the index)
			IndexReader** readers = NULL;

			if (nSegs > 0){
				uint32_t infosize=infos->size();
				readers = _CL_NEWARRAY(IndexReader*,infosize+1);
				for (uint32_t i = 0; i < infosize; i++) {
					//Instantiate a SegementReader responsible for reading the i-th segment and store it in
					//the readers array
					readers[i] = _CLNEW SegmentReader(infos->info(i));
				}
				readers[infosize] = NULL;
			}

			//return an instance of SegmentsReader which is a reader that manages all Segments
			return _CLNEW MultiReader(directory, infos, readers);
        }// end if
  }

  uint64_t IndexReader::lastModified(const char* directory) {
  //Func - Static method
  //       Returns the time the index in the named directory was last modified. 
  //Pre  - directory != NULL and contains the path name of the directory to check
  //Post - The last modified time of the index has been returned

      CND_PRECONDITION(directory != NULL, "directory is NULL");

	  return FSDirectory::fileModified(directory,"segments");
  }

  int64_t IndexReader::getCurrentVersion(Directory* directory) {
     return SegmentInfos::readCurrentVersion(directory);
  }


   int64_t IndexReader::getCurrentVersion(const char* directory){
      Directory* dir = FSDirectory::getDirectory(directory, false);
      int64_t version = getCurrentVersion(dir);
	  dir->close();
      _CLDECDELETE(dir);
      return version;
   }
    
  uint64_t IndexReader::lastModified(const Directory* directory) {
  //Func - Static method
  //       Returns the time the index in this directory was last modified. 
  //Pre  - directory contains a valid reference
  //Post - The last modified time of the index has been returned

      return directory->fileModified("segments");
  }


  bool IndexReader::indexExists(const char* directory){
  //Func - Static method
  //       Checks if an index exists in the named directory
  //Pre  - directory != NULL
  //Post - Returns true if an index exists at the specified directory->
  //       If the directory does not exist or if there is no index in it.
  //       false is returned.

       CND_PRECONDITION(directory != NULL, "directory is NULL");

	   //Create a buffer of length CL_MAXDIR
       char f[CL_MAX_PATH+10]; //add 10 in case that directory is already 260 long
	   //Copy the directory string to the buffer
       strcpy(f,directory);
	   //Cat the name of the segments to buffer
       strcat(f, "/segments");
	   //Check if the segments file exists
       return Misc::dir_Exists(f);
  }
    

  void IndexReader::setNorm(int32_t doc, const TCHAR* field, uint8_t value){
    SCOPED_LOCK_MUTEX(THIS_LOCK)
    if(directoryOwner)
      aquireWriteLock();
    doSetNorm(doc, field, value);
    hasChanges = true;
  }

 void IndexReader::aquireWriteLock() {
    if (stale)
      _CLTHROWA(CL_ERR_IO,"IndexReader out of date and no longer valid for delete, undelete, or setNorm operations");

    if (writeLock == NULL) {
      LuceneLock* writeLock = directory->makeLock("write.lock");
      if (!writeLock->obtain(LUCENE_WRITE_LOCK_TIMEOUT)) // obtain write lock
       _CLTHROWA(CL_ERR_IO,"Index locked for write"); // + writeLock
      this->writeLock = writeLock;

      // we have to check whether index has changed since this reader was opened.
      // if so, this reader is no longer valid for deletion
      if (SegmentInfos::readCurrentVersion(directory) > segmentInfos->getVersion()) {
        stale = true;
        this->writeLock->release();
        _CLDELETE(this->writeLock);
        _CLTHROWA(CL_ERR_IO,"IndexReader out of date and no longer valid for delete, undelete, or setNorm operations");
      }
    }
  }
  

  void IndexReader::setNorm(int32_t doc, const TCHAR* field, float_t value){
     setNorm(doc, field, CL_NS(search)::Similarity::encodeNorm(value));
  }

  bool IndexReader::indexExists(const Directory* directory){
  //Func - Static method
  //       Checks if an index exists in the directory
  //Pre  - directory is a valid reference
  //Post - Returns true if an index exists at the specified directory->
  //       If the directory does not exist or if there is no index in it.
  //       false is returned.