org.terrier.structures.indexing
Class BlockDocumentPostingList

java.lang.Object
  org.terrier.structures.indexing.DocumentPostingList
      org.terrier.structures.indexing.BlockDocumentPostingList

public class BlockDocumentPostingList
extends DocumentPostingList

Represents the postings of one document, and saves block (term position) information. Uses HashMaps internally.

Properties:

• indexing.avg.unique.terms.per.doc - number of unique terms per doc on average, used to tune the initial size of the haashmaps used in this class.

See Also:

DocumentPostingList

Nested Class Summary

Nested classes/interfaces inherited from class org.terrier.structures.indexing.DocumentPostingList

DocumentPostingList.postingIterator

Field Summary

protected int	blockCount number of blocks in this document.
protected gnu.trove.THashMap<java.lang.String,gnu.trove.TIntHashSet>	term_blocks mapping term to blockids in this document

Fields inherited from class org.terrier.structures.indexing.DocumentPostingList

AVG_DOCUMENT_UNIQUE_TERMS, documentLength, occurrences

Constructor Summary

BlockDocumentPostingList()
Instantiate a new block document posting list.

Method Summary

int[]	getBlocks(java.lang.String term) return blocks
int[][]	getPostings() returns the postings suitable to be written into the block direct index
void	insert(java.lang.String t, int blockId) Insert a term into this document, occurs at given block id
protected IterablePosting	makePostingIterator(java.lang.String[] _terms, int[] termIds)

Methods inherited from class org.terrier.structures.indexing.DocumentPostingList

clear, forEachTerm, getDocumentLength, getDocumentStatistics, getFrequency, getNumberOfPointers, getPostings2, getTermId, insert, insert, termSet

Methods inherited from class java.lang.Object

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

Field Detail

term_blocks

protected final gnu.trove.THashMap<java.lang.String,gnu.trove.TIntHashSet> term_blocks

mapping term to blockids in this document

blockCount

protected int blockCount

number of blocks in this document. usually equal to document length, but perhaps less

Constructor Detail

BlockDocumentPostingList

public BlockDocumentPostingList()

Instantiate a new block document posting list. Saves block information, but no fields

Method Detail

insert

public void insert(java.lang.String t,
                   int blockId)

Insert a term into this document, occurs at given block id

getBlocks

public int[] getBlocks(java.lang.String term)

return blocks

Parameters:

term -

Returns:

int[]

getPostings

public int[][] getPostings()

returns the postings suitable to be written into the block direct index

Overrides:

getPostings in class DocumentPostingList

makePostingIterator

protected IterablePosting makePostingIterator(java.lang.String[] _terms,
                                              int[] termIds)

Overrides:

makePostingIterator in class DocumentPostingList