marytts.tools.dbselection

Class FeatureMakerMaryServer

java.lang.Object

marytts.tools.dbselection.FeatureMakerMaryServer

public class FeatureMakerMaryServer
extends Object

Takes text and converts to features Needs a running Mary server

Author:

Anna Hunecke

Field Summary

Fields

Modifier and Type	Field and Description
protected static FeatureDefinition	featDef
protected static MaryHttpClient	mary
protected static String	maryHost
protected static String	maryPort
protected static int	numSentences
protected static int	numUnreliableSentences
protected static Vector<String>	selectionFeature
protected static int[]	selectionFeatureIndex
protected static boolean	strangeSymbols
protected static boolean	strictReliability
protected static boolean	unknownWords
protected static boolean	usefulSentence
protected static DBHandler	wikiToDB

Constructor Summary

Constructors

Constructor and Description
FeatureMakerMaryServer()

Method Summary

Methods

Modifier and Type	Method and Description
protected static int	checkReliability(Element t) Phonemise the given document with the help of JPhonemiser g2p_method "contains-unknown-words" or "contains-strange-symbols",
protected static StringBuffer	collectTokens(Node nextToken, StringBuffer sentence) Collect the tokens of a sentence
protected static byte[]	getFeatures(MaryData d) Process the target features and print them to the given file
protected static void	getXMLAsString(Node motherNode, StringBuilder ppText) Convert the given xml-node and its subnodes to Strings and collect them in the given StringBuilder
static void	main(String[] args)
protected static Document	phonemiseText(String textString, int id) Process the given text with the MaryClient from Text to Chunked
protected static void	printUsage() Print usage of this program
protected static MaryData	processSentence(String nextSentence, int textId, String feas) Process one sentences from text to target features
protected static boolean	readArgs(String[] args) Read and parse the command line args
protected static Vector<String>	splitIntoSentences(String text, int id, boolean test) Split the text into separate sentences

Methods inherited from class java.lang.Object

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

Field Detail

mary

protected static MaryHttpClient mary

usefulSentence

protected static boolean usefulSentence

unknownWords

protected static boolean unknownWords

strangeSymbols

protected static boolean strangeSymbols

featDef

protected static FeatureDefinition featDef

selectionFeature

protected static Vector<String> selectionFeature

selectionFeatureIndex

protected static int[] selectionFeatureIndex

maryHost

protected static String maryHost

maryPort

protected static String maryPort

strictReliability

protected static boolean strictReliability

numSentences

protected static int numSentences

numUnreliableSentences

protected static int numUnreliableSentences

wikiToDB

protected static DBHandler wikiToDB

Constructor Detail

FeatureMakerMaryServer

public FeatureMakerMaryServer()

Method Detail

main

public static void main(String[] args)
                 throws Exception

Throws:

Exception

printUsage

protected static void printUsage()

Print usage of this program

readArgs

protected static boolean readArgs(String[] args)

Read and parse the command line args

Parameters:

args - the args

Returns:

true, if successful, false otherwise

processSentence

protected static MaryData processSentence(String nextSentence,
                       int textId,
                       String feas)

Process one sentences from text to target features

Parameters:

nextSentence - the sentence

textId - the file containing the sentence

feas - target features names separated by space (ex. "phone next_phone selection_prosody")

Returns:

the result of the processing as MaryData object

phonemiseText

protected static Document phonemiseText(String textString,
                     int id)
                                 throws Exception

Process the given text with the MaryClient from Text to Chunked

Parameters:

textString - the text to process

id - id

Returns:

the resulting XML-Document

Throws:

Exception - Exception

getFeatures

protected static byte[] getFeatures(MaryData d)
                             throws Exception

Process the target features and print them to the given file

Parameters:

d - the target features as Mary Data object

Returns:

feasVector

Throws:

Exception - Exception

splitIntoSentences

protected static Vector<String> splitIntoSentences(String text,
                                int id,
                                boolean test)
                                            throws Exception

Split the text into separate sentences

Parameters:

text - the file

id - id

test - test

Returns:

true, if successful

Throws:

Exception - Exception

collectTokens

protected static StringBuffer collectTokens(Node nextToken,
                         StringBuffer sentence)

Collect the tokens of a sentence

Parameters:

nextToken - the Node to start from checkCredibility returns 0 if the sentence is useful 1 if the sentence contains unknownWords (so the sentence is not useful) 2 if the sentence contains strangeSymbols (so the sentence is not useful)

sentence - sentence

Returns:

sentence

checkReliability

protected static int checkReliability(Element t)

Phonemise the given document with the help of JPhonemiser g2p_method "contains-unknown-words" or "contains-strange-symbols",

Parameters:

t - t

Returns:

0 if the sentence is useful 1 if the sentence contains unknownWords 2 if the sentence contains strangeSymbols

getXMLAsString

protected static void getXMLAsString(Node motherNode,
                  StringBuilder ppText)

Convert the given xml-node and its subnodes to Strings and collect them in the given StringBuilder

Parameters:

motherNode - the xml-node

ppText - the StringBuilder