org.terrier.structures.postings.integer

Class BasicIntegerCodingIterablePosting

java.lang.Object

org.terrier.structures.postings.integer.BasicIntegerCodingIterablePosting

All Implemented Interfaces:

Closeable, AutoCloseable, IterablePosting, Posting

Direct Known Subclasses:

BlockIntegerCodingIterablePosting, FieldIntegerCodingIterablePosting

public class BasicIntegerCodingIterablePosting
extends Object
implements IterablePosting

An IterablePosting implementation, which can optionally support fields and/or blocks. Its content is compressed using some integer coding technique. Note: this class read the input in chunks of posting. The chunk size is specified in chunkSize

Since:

4.0

Author:

Matteo Catena

Field Summary

Fields

Modifier and Type	Field and Description
protected int	blkCnt
protected int	chunkSize Number of postings in a chunk
protected int	currentPosting
protected boolean	decompressed
protected DocumentIndex	documentIndex
protected int	id
protected int[]	ids
protected IntegerCodec	idsCodec
protected ByteIn	input
protected static org.apache.log4j.Logger	logger The logger used
protected int	numberOfEntries
protected int	tf
protected int[]	tfs
protected IntegerCodec	tfsCodec

Fields inherited from interface org.terrier.structures.postings.IterablePosting

EOL

Constructor Summary

Constructors

Constructor and Description
BasicIntegerCodingIterablePosting(ByteIn input, int numberOfEntries, DocumentIndex documentIndex, int chunkSize, IntegerCodec idsCodec, IntegerCodec tfsCodec)

Method Summary

Methods

Modifier and Type	Method and Description
WritablePosting	asWritablePosting() Copies this posting to one free of an iterator.
void	close()
protected void	decompress() This load the a chunk (tfs) and decompress it
boolean	endOfPostings() Status method to see if this posting list iterator has been finished
protected void	get(int pos) Read the posting components from the chunk
long	getCurrentAddress()
int	getDocumentLength() Return the length of the document for this posting.
int	getFrequency() Return the frequency of the term in the current document
int	getId() Return the id of the current posting.
protected void	load() This load a chunk (just the document ids!) and decompress it
int	next() Move this iterator to the next posting.
int	next(int targetId) Move this iterator to the posting with specified id, or next posting after that.
void	setId(int id) It may be of benefit to update the frequency of the id in some applications.
protected void	skip() If tfs, fields frequencies and blocks are not required, skip that part of the chunk
String	toString() Makes a human readable form of this posting

Methods inherited from class java.lang.Object

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

Field Detail

logger

protected static final org.apache.log4j.Logger logger

The logger used

documentIndex

protected DocumentIndex documentIndex

chunkSize

protected int chunkSize

Number of postings in a chunk

numberOfEntries

protected int numberOfEntries

currentPosting

protected int currentPosting

id

protected int id

tf

protected int tf

ids

protected final int[] ids

tfs

protected final int[] tfs

input

protected final ByteIn input

idsCodec

protected final IntegerCodec idsCodec

tfsCodec

protected final IntegerCodec tfsCodec

blkCnt

protected int blkCnt

decompressed

protected boolean decompressed

Constructor Detail

BasicIntegerCodingIterablePosting

public BasicIntegerCodingIterablePosting(ByteIn input,
                                 int numberOfEntries,
                                 DocumentIndex documentIndex,
                                 int chunkSize,
                                 IntegerCodec idsCodec,
                                 IntegerCodec tfsCodec)
                                  throws IOException

Parameters:

input - the input channel

numberOfEntries - the number of postings

documentIndex - the document index

chunkSize - the size of the chunk

idsCodec - the IntegerCodec to use to decode docIds

tfsCodec - the IntegerCodec to use to decode term frequencies

Throws:

IOException

Method Detail

getId

public int getId()

Description copied from interface: Posting

Return the id of the current posting. For the inverted index, this is the docid; for the direct index it corresponds to the term id

Specified by:

getId in interface Posting

Returns:

id of the posting

getFrequency

public int getFrequency()

Description copied from interface: Posting

Return the frequency of the term in the current document

Specified by:

getFrequency in interface Posting

Returns:

frequency of posting, in tokens

getDocumentLength

public int getDocumentLength()

Description copied from interface: Posting

Return the length of the document for this posting. Usually uses the DocumentIndex, may do otherwise if document length statistics are in posting list

Specified by:

getDocumentLength in interface Posting

Returns:

length of the document of the current posting in tokens.

setId

public void setId(int id)

Description copied from interface: Posting

It may be of benefit to update the frequency of the id in some applications. This should not be used lightly, as it may break the decompression of other postings in iterators.

Specified by:

setId in interface Posting

asWritablePosting

public WritablePosting asWritablePosting()

Description copied from interface: Posting

Copies this posting to one free of an iterator. Kind of like a clone.

Specified by:

asWritablePosting in interface Posting

Returns:

an identical posting, but which can be maniulated free of this iterator

close

public void close()
           throws IOException

Specified by:

close in interface Closeable

Specified by:

close in interface AutoCloseable

Throws:

IOException

load

protected final void load()
                   throws IOException

This load a chunk (just the document ids!) and decompress it

Throws:

IOException

decompress

protected void decompress()
                   throws IOException

This load the a chunk (tfs) and decompress it

Throws:

IOException

skip

protected void skip()
             throws IOException

If tfs, fields frequencies and blocks are not required, skip that part of the chunk

Throws:

IOException

get

protected void get(int pos)
            throws IOException

Read the posting components from the chunk

Parameters:

pos - the posting to load (as index in the internal arrays)

Throws:

IOException

next

public int next()
         throws IOException

Description copied from interface: IterablePosting

Move this iterator to the next posting.

Specified by:

next in interface IterablePosting

Returns:

id of next posting, or EOL if end of posting list

Throws:

IOException

next

public int next(int targetId)
         throws IOException

Description copied from interface: IterablePosting

Move this iterator to the posting with specified id, or next posting after that. This is usually implemented internally by next(), but more efficient implementations can override this behaviour.

Specified by:

next in interface IterablePosting

Parameters:

targetId - id of the posting to find in this postingn list.

Returns:

id of the posting found, or EOL if end of posting list.

Throws:

IOException

endOfPostings

public boolean endOfPostings()

Description copied from interface: IterablePosting

Status method to see if this posting list iterator has been finished

Specified by:

endOfPostings in interface IterablePosting

Returns:

true if next() or next(int) would return EOL or have returned EOL.

toString

public String toString()

Makes a human readable form of this posting

Overrides:

toString in class Object

getCurrentAddress

public long getCurrentAddress()