---

InvPushIndex Class Reference

#include <InvPushIndex.hpp>

Inheritance diagram for InvPushIndex:

List of all members.

Public Methods
	InvPushIndex ()
	InvPushIndex (char *prefix, int cachesize=128000000, long maxfilesize=2100000000, DOCID_T startdocid=1)
	~InvPushIndex ()
void	setName (char *prefix)
	sets the name for this index. the name will be the prefix for all files related to this index
bool	beginDoc (DocumentProps *dp)
	the beginning of a new document, returns true if initiation was successful
bool	addTerm (Term &t)
	adding a term to the current document, returns true if term was added successfully.
void	endDoc (DocumentProps *dp)
	signify the end of current document
virtual void	endDoc (DocumentProps *dp, const char *mgr)
	signify the end of current document and associate with certain document manager. this doesn't change the mgr that was previously set.
void	endCollection (CollectionProps *cp)
	signify the end of this collection. properties passed at the beginning of a collection should be handled by the constructor.
void	setDocManager (const char *mgrID)
	set the document manager to use for succeeding documents
Protected Methods
void	writeTOC (int numinv)
void	writeDocIDs ()
void	writeCache ()
void	lastWriteCache ()
void	writeDTIDs ()
void	writeDocMgrIDs ()
int	docMgrID (const char *mgr)
virtual void	doendDoc (DocumentProps *dp, int mgrid)
Protected Attributes
long	maxfile
MemCache *	cache
	the biggest our file size can be
vector< char * >	docIDs
	the main memory handler for building
vector< char * >	termIDs
	list of external docids in internal docid order
vector< char * >	tempfiles
	list of terms in termid order
vector< char * >	dtfiles
	list of tempfiles we've written to flush cache
vector< char * >	docmgrs
	list of dt index files
FILE *	writetlookup
ofstream	writetlist
	filestream for writing the lookup table to the docterm db
int	tcount
	filestream for writing the list of located terms for each document
int	tidcount
	count of total terms
int	dtidcount
	count of unique terms
char *	name
	count of unique terms in a current doc
int	namelen
	the prefix name
TABLE_T	wordtable
	the length of the name (avoid many calls to strlen)
map< int, int >	termlist
	table of all terms and their doclists
int *	membuf
	maps of terms and freqs
int	membufsize
	memory to use for cache and buffers
int	curdocmgr

---

Constructor & Destructor Documentation

InvPushIndex::InvPushIndex (    )  [inline]

InvPushIndex::InvPushIndex (  char *    prefix, int    cachesize = 128000000, long    maxfilesize = 2100000000, DOCID_T    startdocid = 1 )

InvPushIndex::~InvPushIndex (    )

---

Member Function Documentation

bool InvPushIndex::addTerm (  Term &    t )  [virtual]

adding a term to the current document, returns true if term was added successfully. Implements PushIndex. Reimplemented in IncPassagePushIndex, InvFPPushIndex, and InvPassagePushIndex.

bool InvPushIndex::beginDoc (  DocumentProps *    dp )  [virtual]

the beginning of a new document, returns true if initiation was successful Implements PushIndex. Reimplemented in IncPassagePushIndex, and InvPassagePushIndex.

int InvPushIndex::docMgrID (  const char *    mgr )  [protected]

returns the internal id of given docmgr if not already registered, mgr will be added

void InvPushIndex::doendDoc (  DocumentProps *    dp, int    mgrid )  [protected, virtual]

Reimplemented in IncPassagePushIndex, InvFPPushIndex, and InvPassagePushIndex.

void InvPushIndex::endCollection (  CollectionProps *    cp )  [virtual]

signify the end of this collection. properties passed at the beginning of a collection should be handled by the constructor. Implements PushIndex. Reimplemented in InvFPPushIndex.

void InvPushIndex::endDoc (  DocumentProps *    dp, const char *    mgr )  [virtual]

signify the end of current document and associate with certain document manager. this doesn't change the mgr that was previously set.

void InvPushIndex::endDoc (  DocumentProps *    dp )  [virtual]

signify the end of current document Implements PushIndex.

void InvPushIndex::lastWriteCache (    )  [protected]

void InvPushIndex::setDocManager (  const char *    mgrID )  [virtual]

set the document manager to use for succeeding documents Implements PushIndex.

void InvPushIndex::setName (  char *    prefix )

sets the name for this index. the name will be the prefix for all files related to this index

void InvPushIndex::writeCache (    )  [protected]

void InvPushIndex::writeDocIDs (    )  [protected]

void InvPushIndex::writeDocMgrIDs (    )  [protected]

void InvPushIndex::writeDTIDs (    )  [protected]

void InvPushIndex::writeTOC (  int    numinv )  [protected]

Reimplemented in InvFPPushIndex.

---

Member Data Documentation

MemCache* InvPushIndex::cache [protected]

the biggest our file size can be

int InvPushIndex::curdocmgr [protected]

vector<char*> InvPushIndex::docIDs [protected]

the main memory handler for building

vector<char*> InvPushIndex::docmgrs [protected]

list of dt index files

vector<char*> InvPushIndex::dtfiles [protected]

list of tempfiles we've written to flush cache

int InvPushIndex::dtidcount [protected]

count of unique terms

long InvPushIndex::maxfile [protected]

int* InvPushIndex::membuf [protected]

maps of terms and freqs

int InvPushIndex::membufsize [protected]

memory to use for cache and buffers

char* InvPushIndex::name [protected]

count of unique terms in a current doc

int InvPushIndex::namelen [protected]

the prefix name

int InvPushIndex::tcount [protected]

filestream for writing the list of located terms for each document

vector<char*> InvPushIndex::tempfiles [protected]

list of terms in termid order

vector<char*> InvPushIndex::termIDs [protected]

list of external docids in internal docid order

map<int, int> InvPushIndex::termlist [protected]

table of all terms and their doclists Reimplemented in InvFPPushIndex.

int InvPushIndex::tidcount [protected]

count of total terms

TABLE_T InvPushIndex::wordtable [protected]

the length of the name (avoid many calls to strlen)

ofstream InvPushIndex::writetlist [protected]

filestream for writing the lookup table to the docterm db

FILE* InvPushIndex::writetlookup [protected]

---

The documentation for this class was generated from the following files:

• InvPushIndex.hpp
• InvPushIndex.cpp

---