com.bigdata.rdf.lexicon

Class BigdataValueCentricFullTextIndex

java.lang.Object

com.bigdata.relation.AbstractResource<IRelation<E>>

com.bigdata.relation.AbstractRelation

com.bigdata.search.FullTextIndex

com.bigdata.rdf.lexicon.BigdataValueCentricFullTextIndex

All Implemented Interfaces:

ITextIndexer<Hit>, IValueCentricTextIndexer<Hit>, IMutableRelation, IMutableResource, IRelation, ILocatableResource

public class BigdataValueCentricFullTextIndex
extends FullTextIndex
implements IValueCentricTextIndexer<Hit>

Implementation based on the built-in keyword search capabilities for bigdata.

Version:

$Id$

Author:

Bryan Thompson

Nested Class Summary

Nested classes/interfaces inherited from class com.bigdata.search.FullTextIndex

FullTextIndex.Options

Nested classes/interfaces inherited from interface com.bigdata.rdf.lexicon.ITextIndexer

ITextIndexer.FullTextQuery

Field Summary

Fields inherited from class com.bigdata.search.FullTextIndex

NAME_SEARCH

Fields inherited from class com.bigdata.relation.AbstractResource

indexManager

Constructor Summary

Constructors

Constructor and Description
BigdataValueCentricFullTextIndex(IIndexManager indexManager, String namespace, Long timestamp, Properties properties)

Method Summary

Methods

Modifier and Type	Method and Description
protected Hit[]	applyRegex(Hit[] hits, Pattern regex) Subclasses can override this method to do regex post-processing.
void	create() Conditionally registers the necessary index(s).
void	destroy() The full text index is currently located in the same namespace as the lexicon relation.
boolean	getIndexDatatypeLiterals() Return true iff datatype literals are being indexed.
static BigdataValueCentricFullTextIndex	getInstance(IIndexManager indexManager, String namespace, Long timestamp, Properties properties)
LexiconRelation	getLexiconRelation()
void	index(int capacity, Iterator<BigdataValue> valuesIterator) Add the terms to the full text index so that we can do fast lookup of the corresponding term identifiers.
protected Hit[]	matchExact(Hit[] hits, String query) Subclasses can override this method to do exact match processing.

Methods inherited from class com.bigdata.search.FullTextIndex

_search, count, delete, executeQuery, getAnalyzer, getElementClass, getIndex, getIndexNames, getKeyBuilder, getKeyOrder, getKeyOrders, getPrimaryKeyOrder, getTokenStream, index, index, insert, isOverwrite, isReadOnly, newElement, search, slice, tokenize

Methods inherited from class com.bigdata.relation.AbstractRelation

getAccessPath, getAccessPath, getAccessPath, getFQN, getFQN, getFQN, getIndex, getIndex, getIndex, newAccessPath, newIndexMetadata

Methods inherited from class com.bigdata.relation.AbstractResource

acquireExclusiveLock, assertWritable, getBareProperties, getChunkCapacity, getChunkOfChunksCapacity, getChunkTimeout, getCommitTime, getContainer, getContainerNamespace, getExecutorService, getFullyBufferedReadThreshold, getIndexManager, getMaxParallelSubqueries, getNamespace, getProperties, getProperty, getProperty, getTimestamp, init, isForceSerialExecution, toString, unlock

Methods inherited from class java.lang.Object

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

Methods inherited from interface com.bigdata.rdf.lexicon.ITextIndexer

count, search

Methods inherited from interface com.bigdata.relation.IRelation

getExecutorService, getIndexManager

Methods inherited from interface com.bigdata.relation.locator.ILocatableResource

getContainerNamespace, getNamespace, getTimestamp, init

Constructor Detail

BigdataValueCentricFullTextIndex

public BigdataValueCentricFullTextIndex(IIndexManager indexManager,
                                String namespace,
                                Long timestamp,
                                Properties properties)

Parameters:

indexManager -

namespace -

timestamp -

properties -

Method Detail

getInstance

public static BigdataValueCentricFullTextIndex getInstance(IIndexManager indexManager,
                                           String namespace,
                                           Long timestamp,
                                           Properties properties)

getIndexDatatypeLiterals

public boolean getIndexDatatypeLiterals()

Description copied from interface: ITextIndexer

Return true iff datatype literals are being indexed.

Specified by:

getIndexDatatypeLiterals in interface ITextIndexer<Hit>

create

public void create()

Conditionally registers the necessary index(s).

Specified by:

create in interface ITextIndexer<Hit>

Specified by:

create in interface IMutableResource

Overrides:

create in class FullTextIndex

Throws:

IllegalStateException - if the client does not have write access.

destroy

public void destroy()

The full text index is currently located in the same namespace as the lexicon relation. However, the distributed zookeeper locks (ZLocks) are not reentrant. Therefore this method is overridden to NOT acquire the ZLock for the namespace of the relation when destroying the full text index -- that lock is already held for the same namespace by the LexiconRelation.

Specified by:

destroy in interface ITextIndexer<Hit>

Specified by:

destroy in interface IMutableResource

Overrides:

destroy in class FullTextIndex

index

public void index(int capacity,
         Iterator<BigdataValue> valuesIterator)

Description copied from interface: IValueCentricTextIndexer

Add the terms to the full text index so that we can do fast lookup of the corresponding term identifiers. Only literals are tokenized. Literals that have a language code property are parsed using a tokenizer appropriate for the specified language family. Other literals and URIs are tokenized using the default Locale.

Specified by:

index in interface IValueCentricTextIndexer<Hit>

Parameters:

capacity - A hint to the underlying layer about the buffer size before an incremental flush of the index.

getLexiconRelation

public final LexiconRelation getLexiconRelation()

matchExact

protected Hit[] matchExact(Hit[] hits,
               String query)

Description copied from class: FullTextIndex

Subclasses can override this method to do exact match processing. This involves materializing the hits into their original text values and checking for the query string in the materialized value. Not possible from the base class. The value-centric RDF version can use the lexicon to materialize the hits and check them for exact match.

Overrides:

matchExact in class FullTextIndex

applyRegex

protected Hit[] applyRegex(Hit[] hits,
               Pattern regex)

Description copied from class: FullTextIndex

Subclasses can override this method to do regex post-processing. This involves materializing the hits into their original text values and checking against the regex string in the materialized value. Not possible from the base class. The value-centric RDF version can use the lexicon to materialize the hits and check them against the regex.

Overrides:

applyRegex in class FullTextIndex