org.terrier.structures

Class IndexOnDisk

java.lang.Object

org.terrier.structures.Index

org.terrier.structures.IndexOnDisk

All Implemented Interfaces:

Closeable, Flushable, AutoCloseable

public class IndexOnDisk
extends Index

The replacement for what was Index in earlier Terrier versions. Represents the most common type of index, i.e. one which is stored on disk.

Since:

4.0

Author:

Stuart Mackie, Craig Macdonald, Richard McCreadie

Nested Class Summary

Nested classes/interfaces inherited from class org.terrier.structures.Index

Index.UpdatingCollectionStatistics

Field Summary

Fields

Modifier and Type	Field and Description
protected String	loadError
protected boolean	loadSuccess Set to true if loading an index succeeds
protected String	path path component of this index's location
protected String	prefix prefix component of this index's location
protected HashMap<String,Object>	structureCache Cache of all opened index structures, but not input streams

Fields inherited from class org.terrier.structures.Index

dirtyProperties, EMPTY_CLASS_ARRAY, lastLoadError, logger, MINIMUM_INDEX_TERRIER_VERSION, properties, PROPERTIES_SUFFIX, RETRIEVAL_LOADING_PROFILE

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	IndexOnDisk() A default constructor that creates an instance of the index.
	IndexOnDisk(long l, long m, long n)
protected	IndexOnDisk(String _path, String _prefix, boolean isNew) Constructs a new Index object.

Method Summary

Methods

Modifier and Type	Method and Description
void	close() Closes the data structures of the index.
void	flush() Write any dirty data structures down to disk
protected void	flushProperties() Write any dirty properties down to disk
CollectionStatistics	getCollectionStatistics() Get the collection statistics
PostingIndex<Pointer>	getDirectIndex() Return the DirectIndex associated with this index
DocumentIndex	getDocumentIndex() Return the DocumentIndex associated with this index
Object	getIndexStructure(String structureName) Obtains the named index structure, using an already loaded one if possible.
Object	getIndexStructureInputStream(String structureName) Return the input stream associated with the specified structure of this index
PostingIndex<?>	getInvertedIndex() Returns the InvertedIndex to use for this index
Lexicon<String>	getLexicon() Return the Lexicon associated with this index
MetaIndex	getMetaIndex() Get the Meta Index structure
String	getPath() Returns the path of this index
String	getPrefix() Returns the prefix of this index
protected Object	loadIndexStructure(String structureName) Load a new instance of the named index structure.
protected void	loadIndices() load all index structures.
protected boolean	loadProperties() loads in the properties file, falling back to the Terrier 1.xx log file if no properties exist.
protected void	loadStatistics() for an immutable index, use a normal collection statistics, never changes
protected void	loadUpdatingStatistics() for an index that is not yet built, use an UpdatingCollectionStatistics, which is slower but can support updates of the index statistics
String	toString() Returns a String representation of this index

Methods inherited from class org.terrier.structures.Index

addIndexStructure, addIndexStructure, addIndexStructure, addIndexStructure, addIndexStructureInputStream, addIndexStructureInputStream, addIndexStructureInputStream, addIndexStructureInputStream, allExists, createIndex, createIndex, createNewIndex, existsIndex, getEnd, getIndexLoadingProfileAsRetrieval, getIndexProperty, getIntIndexProperty, getLastIndexLoadError, getProperties, getStart, hasIndexStructure, hasIndexStructureInputStream, join, main, setIndexLoadingProfileAsRetrieval, setIndexProperty

Methods inherited from class java.lang.Object

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

Field Detail

path

protected String path

path component of this index's location

prefix

protected String prefix

prefix component of this index's location

structureCache

protected final HashMap<String,Object> structureCache

Cache of all opened index structures, but not input streams

loadSuccess

protected boolean loadSuccess

Set to true if loading an index succeeds

loadError

protected String loadError

Constructor Detail

IndexOnDisk

protected IndexOnDisk()

A default constructor that creates an instance of the index.

IndexOnDisk

protected IndexOnDisk(String _path,
           String _prefix,
           boolean isNew)

Constructs a new Index object. Don't call this method, call the createIndex(String) factory method to construct an Index object.

Parameters:

_path - String the path in which the data structures will be created.

_prefix - String the prefix of the files to be created.

isNew - where a new Index should be created if there is no index at the specified location

IndexOnDisk

public IndexOnDisk(long l,
           long m,
           long n)

Method Detail

close

public void close()
           throws IOException

Description copied from class: Index

Closes the data structures of the index.

Specified by:

close in interface Closeable

Specified by:

close in interface AutoCloseable

Specified by:

close in class Index

Throws:

IOException

flush

public void flush()
           throws IOException

Description copied from class: Index

Write any dirty data structures down to disk

Specified by:

flush in interface Flushable

Specified by:

flush in class Index

Throws:

IOException

flushProperties

protected void flushProperties()
                        throws IOException

Write any dirty properties down to disk

Throws:

IOException

getCollectionStatistics

public CollectionStatistics getCollectionStatistics()

Description copied from class: Index

Get the collection statistics

Specified by:

getCollectionStatistics in class Index

getDirectIndex

public PostingIndex<Pointer> getDirectIndex()

Description copied from class: Index

Return the DirectIndex associated with this index

Specified by:

getDirectIndex in class Index

getDocumentIndex

public DocumentIndex getDocumentIndex()

Description copied from class: Index

Return the DocumentIndex associated with this index

Specified by:

getDocumentIndex in class Index

getIndexStructure

public Object getIndexStructure(String structureName)

Obtains the named index structure, using an already loaded one if possible.

Specified by:

getIndexStructure in class Index

Parameters:

structureName - name of the required structure

Returns:

desired object or null if not found

getIndexStructureInputStream

public Object getIndexStructureInputStream(String structureName)

Description copied from class: Index

Return the input stream associated with the specified structure of this index

Specified by:

getIndexStructureInputStream in class Index

Parameters:

structureName - The name of the structure of which you want the inputstream. Eg "lexicon"

Returns:

Required structure, or null if not found

getInvertedIndex

public PostingIndex<?> getInvertedIndex()

Description copied from class: Index

Returns the InvertedIndex to use for this index

Specified by:

getInvertedIndex in class Index

getLexicon

public Lexicon<String> getLexicon()

Description copied from class: Index

Return the Lexicon associated with this index

Specified by:

getLexicon in class Index

getMetaIndex

public MetaIndex getMetaIndex()

Description copied from class: Index

Get the Meta Index structure

Specified by:

getMetaIndex in class Index

getPath

public String getPath()

Returns the path of this index

getPrefix

public String getPrefix()

Returns the prefix of this index

loadIndexStructure

protected Object loadIndexStructure(String structureName)

Load a new instance of the named index structure.

Parameters:

structureName - name of the required structure

Returns:

desired object or null if not found

loadIndices

protected void loadIndices()

load all index structures. Is disabled if index property index.preloadIndices.disabled is set to true. It is false by default, which means that all non-inputstream indices are loaded on initialisation of the index. When the property is true, indices are loaded as required.

loadProperties

protected boolean loadProperties()

loads in the properties file, falling back to the Terrier 1.xx log file if no properties exist.

loadStatistics

protected void loadStatistics()

for an immutable index, use a normal collection statistics, never changes

loadUpdatingStatistics

protected void loadUpdatingStatistics()

for an index that is not yet built, use an UpdatingCollectionStatistics, which is slower but can support updates of the index statistics

toString

public String toString()

Description copied from class: Index

Returns a String representation of this index

Specified by:

toString in class Index