marytts.cart.impose

Class FeatureArrayIndexer

java.lang.Object

marytts.cart.impose.FeatureArrayIndexer

public class FeatureArrayIndexer
extends Object

A class branched from FeatureFileIndexer which works directly on a feature array, rather than extending FeatureFileReader.

Author:

Marc Schröder

Field Summary

Fields

Modifier and Type	Field and Description
static int	MAXDEPTH
static int	MAXLEVEL
static int	MINUNITS

Constructor Summary

Constructors

Constructor and Description
FeatureArrayIndexer(FeatureVector[] featureVectors, FeatureDefinition featureDefinition) Constructor which loads the feature vector array but does not launch an indexing operation.
FeatureArrayIndexer(FeatureVector[] featureVectors, FeatureDefinition featureDefinition, int[] setFeatureSequence) Constructor which takes an array of feature vectors and launches an indexing operation according to a feature sequence constraint.
FeatureArrayIndexer(FeatureVector[] featureVectors, FeatureDefinition featureDefinition, String[] setFeatureSequence) Constructor which takes an array of feature vectors and launches an indexing operation according to a feature sequence constraint.

Method Summary

Methods

Modifier and Type	Method and Description
void	deepFill(MaryNode specTree) Fill a tree which specifies a feature hierarchy but no corresponding units.
void	deepSort(int[] setFeatureSequence) Launches a deep sort on the array of feature vectors.
void	deepSort(String[] setFeatureSequence) Launches a deep sort on the array of feature vectors.
FeatureDefinition	getFeatureDefinition()
int[]	getFeatureSequence() Get the feature sequence, as an information about the underlying tree structure.
FeatureVector[]	getFeatureVectors(int from, int to) Get the feature vectors from the big array according to the given indices
long	getNumberOfLeaves() Get the number of leaves.
long	getTheoreticalNumberOfLeaves(int[] feaSeq) Get the theoretical number of leaves, given a feature sequence.
MaryNode	getTree() Get the tree
FeatureFileIndexingResult	retrieve(FeatureVector v) Retrieve an array of unit features which complies with a specific target specification, according to an underlying tree.
FeatureFileIndexingResult	retrieve(FeatureVector v, int condition, int parameter) Retrieve an array of unit features which complies with a specific target specification, according to an underlying tree, and given a stopping condition.

Methods inherited from class java.lang.Object

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

Field Detail

MAXDEPTH

public static final int MAXDEPTH

See Also:

Constant Field Values

MAXLEVEL

public static final int MAXLEVEL

See Also:

Constant Field Values

MINUNITS

public static final int MINUNITS

See Also:

Constant Field Values

Constructor Detail

FeatureArrayIndexer

public FeatureArrayIndexer(FeatureVector[] featureVectors,
                   FeatureDefinition featureDefinition,
                   int[] setFeatureSequence)

Constructor which takes an array of feature vectors and launches an indexing operation according to a feature sequence constraint.

Parameters:

featureVectors - an array of feature vectors

featureDefinition - a feature definition to make sense of the feature vectors

setFeatureSequence - An array of indexes indicating the hierarchical order (or, equivalently, the sequence) of the features to use for the indexing.

FeatureArrayIndexer

public FeatureArrayIndexer(FeatureVector[] featureVectors,
                   FeatureDefinition featureDefinition,
                   String[] setFeatureSequence)

Constructor which takes an array of feature vectors and launches an indexing operation according to a feature sequence constraint.

Parameters:

featureVectors - an array of feature vectors

featureDefinition - a feature definition to make sense of the feature vectors

setFeatureSequence - An array of feature names indicating the hierarchical order (or, equivalently, the sequence) of the features to use for the indexing.

FeatureArrayIndexer

public FeatureArrayIndexer(FeatureVector[] featureVectors,
                   FeatureDefinition featureDefinition)

Constructor which loads the feature vector array but does not launch an indexing operation.

Parameters:

featureVectors - an array of feature vectors

featureDefinition - a feature definition to make sense of the feature vectors

Method Detail

deepSort

public void deepSort(int[] setFeatureSequence)

Launches a deep sort on the array of feature vectors. This is public because it can be used to re-index the previously read feature file.

Parameters:

setFeatureSequence - An array of feature indexes, indicating the sequence of features according to which the sorting should be performed.

deepSort

public void deepSort(String[] setFeatureSequence)

Launches a deep sort on the array of feature vectors. This is public because it can be used to re-index the previously read feature file.

Parameters:

setFeatureSequence - An array of feature names, indicating the sequence of features according to which the sorting should be performed.

deepFill

public void deepFill(MaryNode specTree)

Fill a tree which specifies a feature hierarchy but no corresponding units.

Parameters:

specTree - A specific tree

retrieve

public FeatureFileIndexingResult retrieve(FeatureVector v)

Retrieve an array of unit features which complies with a specific target specification, according to an underlying tree.

Parameters:

v - A feature vector for which to send back an array of complying unit indexes.

Returns:

A query result, comprising an array of feature vectors and the depth level which was actually reached.

See Also:

deepSort(int[]), deepFill(MaryNode)

retrieve

public FeatureFileIndexingResult retrieve(FeatureVector v,
                                 int condition,
                                 int parameter)

Retrieve an array of unit features which complies with a specific target specification, according to an underlying tree, and given a stopping condition.

Parameters:

v - A feature vector for which to send back an array of complying unit indexes.

condition - A constant indicating a stopping criterion, among: FeatureFileIndexer.MAXDEPTH : walk the tree until its leaves (maximum depth); FeatureFileIndexer.MAXLEVEL : walk the tree until a certain depth level; FeatureFileIndexer.MINUNITS : walk the tree until a certain number of units is reached.

parameter - A parameter interpreted according to the above condition: MAXDEPTH → parameter is ignored; MAXLEVEL → parameter = maximum level to reach; MINUNITS → parameter = lower bound on the number of units to return.

Returns:

A query result, comprising an array of feature vectors and the depth level which was actually reached.

See Also:

deepSort(int[]), deepFill(MaryNode)

getFeatureSequence

public int[] getFeatureSequence()

Get the feature sequence, as an information about the underlying tree structure.

Returns:

the feature sequence

getTree

public MaryNode getTree()

Get the tree

Returns:

the tree

getFeatureVectors

public FeatureVector[] getFeatureVectors(int from,
                                int to)

Get the feature vectors from the big array according to the given indices

Parameters:

from - the start index

to - the end index

Returns:

the feature vectors

getFeatureDefinition

public FeatureDefinition getFeatureDefinition()

getNumberOfLeaves

public long getNumberOfLeaves()

Get the number of leaves.

Returns:

The number of leaves, or -1 if the tree has not been computed.

getTheoreticalNumberOfLeaves

public long getTheoreticalNumberOfLeaves(int[] feaSeq)

Get the theoretical number of leaves, given a feature sequence.

Parameters:

feaSeq - feaSeq

Returns:

The number of leaves, or -1 if the capacity of the long integer was blown.