org.terrier.structures.indexing

Class LexiconMap

java.lang.Object

org.terrier.structures.indexing.LexiconMap

Direct Known Subclasses:

FieldLexiconMap

public class LexiconMap
extends Object

This class keeps track of the total counts of terms within a bundle of documents being indexed. Internally, uses hashmaps. This class replaces the LexiconTree etc.

Properties

• indexing.avg.unique.terms.per.bundle - the unique number of terms expected to be indexed in a bundle of documents. Not a limit, just a hint for the sizing of the hashmaps.Default to 120.

Field Summary

Fields

Modifier and Type	Field and Description
protected static int	BUNDLE_AVG_UNIQUE_TERMS Number of unique terms expected to be indexed in a bundle of documents.
protected gnu.trove.TObjectIntHashMap<String>	maxtfs mapping: term to max tf
protected gnu.trove.TObjectIntHashMap<String>	nts mapping: term to document frequency
protected int	numberOfNodes number of different terms
protected int	numberOfPointers number of different entries there will be in the inverted index
protected gnu.trove.TObjectIntHashMap<String>	tfs mapping: term to term frequency in the collection

Constructor Summary

Constructors

Constructor and Description
LexiconMap()

Method Summary

Modifier and Type	Method and Description
void	clear() Clear the lexicon map
int	getNumberOfNodes() Returns the numbe of nodes in the tree.
int	getNumberOfPointers() Returns the number of pointers in the tree.
void	insert(DocumentPostingList doc) Inserts all the terms from a document posting into the lexicon map
void	insert(String term, int tf) Inserts a new term in the lexicon map.
void	storeToStream(LexiconOutputStream<String> lexiconStream, TermCodes termCodes) Stores the lexicon tree to a lexicon stream as a sequence of entries.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

BUNDLE_AVG_UNIQUE_TERMS

protected static final int BUNDLE_AVG_UNIQUE_TERMS

Number of unique terms expected to be indexed in a bundle of documents.

numberOfNodes

protected int numberOfNodes

number of different terms

numberOfPointers

protected int numberOfPointers

number of different entries there will be in the inverted index

tfs

protected final gnu.trove.TObjectIntHashMap<String> tfs

mapping: term to term frequency in the collection

nts

protected final gnu.trove.TObjectIntHashMap<String> nts

mapping: term to document frequency

maxtfs

protected final gnu.trove.TObjectIntHashMap<String> maxtfs

mapping: term to max tf

Constructor Detail

LexiconMap

public LexiconMap()

Method Detail

clear

public void clear()

Clear the lexicon map

insert

public void insert(String term,
                   int tf)

Inserts a new term in the lexicon map.

Parameters:

term - The term to be inserted.

tf - The id of the term.

insert

public void insert(DocumentPostingList doc)

Inserts all the terms from a document posting into the lexicon map

Parameters:

doc - The postinglist for that document

storeToStream

public void storeToStream(LexiconOutputStream<String> lexiconStream,
                          TermCodes termCodes)
                   throws IOException

Stores the lexicon tree to a lexicon stream as a sequence of entries. The binary tree is traversed in order, by called the method traverseAndStoreToStream.

Parameters:

lexiconStream - The lexicon output stream to store to.

Throws:

IOException

getNumberOfNodes

public int getNumberOfNodes()

Returns the numbe of nodes in the tree.

Returns:

int the number of nodes in the tree.

getNumberOfPointers

public int getNumberOfPointers()

Returns the number of pointers in the tree.

Returns:

int the number of pointers in the tree.