---

InvPushIndex Class Reference

#include <InvPushIndex.hpp>

Inheritance diagram for InvPushIndex:

List of all members.

Public Methods
	InvPushIndex (char *prefix="DefaultIndex",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)
	returns the internal id of given docmgr if not already registered, mgr will be added.
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 ( char * prefix = "DefaultIndex", 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. Reimplemented from PushIndex. Reimplemented in InvFPPushIndex.

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

the beginning of a new document, returns true if initiation was successful. Reimplemented from PushIndex.

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 InvFPPushIndex.

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. Reimplemented from 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. Reimplemented from PushIndex.

void InvPushIndex::lastWriteCache ( ) [protected]

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

set the document manager to use for succeeding documents. Reimplemented from 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::writeDTIDs ( ) [protected]

void InvPushIndex::writeDocIDs ( ) [protected]

void InvPushIndex::writeDocMgrIDs ( ) [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

---