org.apache.mahout.utils.vectors.lucene

Class LuceneIterator

java.lang.Object

com.google.common.collect.UnmodifiableIterator<T>

com.google.common.collect.AbstractIterator<Vector>

org.apache.mahout.utils.vectors.lucene.AbstractLuceneIterator

org.apache.mahout.utils.vectors.lucene.LuceneIterator

All Implemented Interfaces:

Iterator<Vector>

public class LuceneIterator
extends AbstractLuceneIterator

An Iterator over Vectors that uses a Lucene index as the source for creating the Vectors. The field used to create the vectors currently must have term vectors stored for it.

Field Summary

Fields

Modifier and Type	Field and Description
protected String	idField
protected Set<String>	idFieldSelector

Fields inherited from class org.apache.mahout.utils.vectors.lucene.AbstractLuceneIterator

bump, field, indexReader, maxErrorDocs, nextDocId, nextLogRecord, normPower, numErrorDocs, skippedErrorMessages, terminfo, weight

Constructor Summary

Constructors

Constructor and Description
LuceneIterator(org.apache.lucene.index.IndexReader indexReader, String idField, String field, TermInfo terminfo, Weight weight, double normPower) Produce a LuceneIterable that can create the Vector plus normalize it.
LuceneIterator(org.apache.lucene.index.IndexReader indexReader, String idField, String field, TermInfo terminfo, Weight weight, double normPower, double maxPercentErrorDocs)

Method Summary

Methods

Modifier and Type	Method and Description
protected String	getVectorName(int documentIndex) Given the document name, derive a name for the vector.

Methods inherited from class org.apache.mahout.utils.vectors.lucene.AbstractLuceneIterator

computeNext

Methods inherited from class com.google.common.collect.AbstractIterator

endOfData, hasNext, next, peek

Methods inherited from class com.google.common.collect.UnmodifiableIterator

remove

Methods inherited from class java.lang.Object

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

Field Detail

idFieldSelector

protected final Set<String> idFieldSelector

idField

protected final String idField

Constructor Detail

LuceneIterator

public LuceneIterator(org.apache.lucene.index.IndexReader indexReader,
              String idField,
              String field,
              TermInfo terminfo,
              Weight weight,
              double normPower)

Produce a LuceneIterable that can create the Vector plus normalize it.

Parameters:

indexReader - IndexReader to read the documents from.

idField - field containing the id. May be null.

field - field to use for the Vector

terminfo - terminfo

weight - weight

normPower - the normalization value. Must be nonnegative, or LuceneIterable.NO_NORMALIZING

LuceneIterator

public LuceneIterator(org.apache.lucene.index.IndexReader indexReader,
              String idField,
              String field,
              TermInfo terminfo,
              Weight weight,
              double normPower,
              double maxPercentErrorDocs)

Parameters:

indexReader - IndexReader to read the documents from.

idField - field containing the id. May be null.

field - field to use for the Vector

terminfo - terminfo

weight - weight

normPower - the normalization value. Must be nonnegative, or LuceneIterable.NO_NORMALIZING

maxPercentErrorDocs - most documents that will be tolerated without a term freq vector. In [0,1].

See Also:

LuceneIterator(org.apache.lucene.index.IndexReader, String, String, org.apache.mahout.utils.vectors.TermInfo, org.apache.mahout.vectorizer.Weight, double)

Method Detail

getVectorName

protected String getVectorName(int documentIndex)
                        throws IOException

Description copied from class: AbstractLuceneIterator

Given the document name, derive a name for the vector. This may involve reading the document from Lucene and setting up any other state that the subclass wants. This will be called once for each document that the iterator processes.

Specified by:

getVectorName in class AbstractLuceneIterator

Parameters:

documentIndex - the lucene document index.

Returns:

the name to store in the vector.

Throws:

IOException