marytts.cart.io

Class WagonCARTReader

java.lang.Object

marytts.cart.io.WagonCARTReader

public class WagonCARTReader
extends Object

IO functions for CARTs in WagonCART format

Author:

Anna Hunecke, Marc Schröder, Marcela Charfuelan

Constructor Summary

Constructors

Constructor and Description
WagonCARTReader(LeafNode.LeafType leafType) When creating a WagonCARTReader provide a tree type

Method Summary

Modifier and Type	Method and Description
protected LeafNode	createFeatureVectorLeafNode(String line)
protected LeafNode	createFloatLeafNode(String line)
protected LeafNode	createIntAndFloatArrayLeafNode(String line)
protected LeafNode	createIntArrayLeafNode(String line)
protected LeafNode	createLeafNode(String line) For a line representing a leaf in Wagon format, create a leaf.
protected LeafNode	createStringAndFloatLeafNode(String line)
void	fillLeafs(Node root, FeatureVector[] featureVectors) Fill the FeatureVector leafs of a tree with the given feature vectors.
Node	load(BufferedReader reader, FeatureDefinition featDefinition) This loads a cart from a wagon tree in textual format, from a reader.
Node	load(String fileName, FeatureDefinition featDefinition, String[] dummy) Load the cart from the given file

Methods inherited from class java.lang.Object

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

Constructor Detail

WagonCARTReader

public WagonCARTReader(LeafNode.LeafType leafType)

When creating a WagonCARTReader provide a tree type

Parameters:

leafType - ClasificationTree, ExtendedClassificationTree, RegressionTree, or TopLevelTree.

ClasificationTree → IntArrayLeafNode

ExtendedClassificationTree → IntAndFloatArrayLeafNode

RegressionTree → FloatLeafNode

TopLevelTree → FeatureVectorLeafNode

StringCART → StringAndFloatLeafNode

Method Detail

createLeafNode

protected LeafNode createLeafNode(String line)

For a line representing a leaf in Wagon format, create a leaf. This method decides which implementation of LeafNode is used, i.e. which data format is appropriate. Lines are of the form ((index1 float1)...(indexN floatN)) 0))

Parameters:

line - a line from a wagon cart file, representing a leaf

Returns:

a leaf node representing the line.

load

public Node load(BufferedReader reader,
                 FeatureDefinition featDefinition)
          throws IOException

This loads a cart from a wagon tree in textual format, from a reader.

Parameters:

reader - the Reader providing the wagon tree

featDefinition - featDefinition

Returns:

rootNode

Throws:

IOException - IOException

load

public Node load(String fileName,
                 FeatureDefinition featDefinition,
                 String[] dummy)
          throws IOException,
                 MaryConfigurationException

Load the cart from the given file

Parameters:

fileName - the file to load the cart from

featDefinition - the feature definition

dummy - unused, just here for compatibility with the FeatureFileIndexer.

Returns:

rootNode

Throws:

IOException - if a problem occurs while loading

MaryConfigurationException - MaryConfigurationException

createIntArrayLeafNode

protected LeafNode createIntArrayLeafNode(String line)

createIntAndFloatArrayLeafNode

protected LeafNode createIntAndFloatArrayLeafNode(String line)

createFloatLeafNode

protected LeafNode createFloatLeafNode(String line)

createFeatureVectorLeafNode

protected LeafNode createFeatureVectorLeafNode(String line)

fillLeafs

public void fillLeafs(Node root,
                      FeatureVector[] featureVectors)

Fill the FeatureVector leafs of a tree with the given feature vectors. This function is only used in TopLeavelTree.

Parameters:

root - node of the tree.

featureVectors - the feature vectors.

createStringAndFloatLeafNode

protected LeafNode createStringAndFloatLeafNode(String line)