marytts.modules

Class JPhonemiser

java.lang.Object

marytts.modules.InternalModule

marytts.modules.JPhonemiser

All Implemented Interfaces:

MaryModule

Direct Known Subclasses:

JPhonemiser, Phonemiser

public class JPhonemiser
extends InternalModule

The phonemiser module -- java implementation.

Author:

Marc Schröder, Sathish, ingmar

Field Summary

Fields

Modifier and Type	Field and Description
protected AllophoneSet	allophoneSet
protected FSTLookup	lexicon
protected TrainedLTS	lts
protected Pattern	punctuationPosRegex
protected boolean	removeTrailingOneFromPhones
protected Pattern	unpronounceablePosRegex
protected Map<String,List<String>>	userdict

Fields inherited from class marytts.modules.InternalModule

logger, state

Fields inherited from interface marytts.modules.MaryModule

MODULE_OFFLINE, MODULE_RUNNING

Constructor Summary

Constructors

Constructor and Description
JPhonemiser(String propertyPrefix)
JPhonemiser(String componentName, MaryDataType inputType, MaryDataType outputType, String allophonesProperty, String userdictProperty, String lexiconProperty, String ltsProperty) Constructor providing the individual filenames of files that are required.
JPhonemiser(String componentName, MaryDataType inputType, MaryDataType outputType, String allophonesProperty, String userdictProperty, String lexiconProperty, String ltsProperty, String removetrailingonefromphonesProperty) Constructor providing the individual filenames of files that are required.

Method Summary

Methods

Modifier and Type	Method and Description
AllophoneSet	getAllophoneSet() Access the allophone set underlying this phonemiser.
boolean	isPosPunctuation(String pos) Based on the regex compiled in setPunctuationPosRegex(), determine whether a given POS string is classified as punctuation
boolean	isUnpronounceable(String pos)
String	lexiconLookup(String text, String pos) Look a given text up in the (standard) lexicon.
boolean	maybePronounceable(String text, String pos) Determine whether token should be pronounceable, based on text and POS tag.
String	phonemise(String text, String pos, StringBuilder g2pMethod) Phonemise the word text.
MaryData	process(MaryData d) Perform this module's processing on abstract "MaryData" input d.
protected Map<String,List<String>>	readLexicon(String lexiconFilename) Read a lexicon.
protected void	setPh(Element t, String ph)
protected void	setPunctuationPosRegex() Compile a regex pattern used to determine whether tokens are processed as punctuation or not, based on whether their pos attribute matches the pattern.
protected void	setUnpronounceablePosRegex() Compile a regex pattern used to determine whether tokens are processed as unprounounceable or not, based on whether their pos attribute matches the pattern.
void	startup() Allow the module to start up, performing whatever is necessary to become operational.
String	userdictLookup(String text, String pos) look a given text up in the userdict.

Methods inherited from class marytts.modules.InternalModule

getInputType, getLocale, getOutputType, getState, inputType, name, outputType, powerOnSelfTest, shutdown

Methods inherited from class java.lang.Object

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

Field Detail

userdict

protected Map<String,List<String>> userdict

lexicon

protected FSTLookup lexicon

lts

protected TrainedLTS lts

removeTrailingOneFromPhones

protected boolean removeTrailingOneFromPhones

allophoneSet

protected AllophoneSet allophoneSet

punctuationPosRegex

protected Pattern punctuationPosRegex

unpronounceablePosRegex

protected Pattern unpronounceablePosRegex

Constructor Detail

JPhonemiser

public JPhonemiser(String propertyPrefix)
            throws IOException,
                   MaryConfigurationException

Throws:

IOException

MaryConfigurationException

JPhonemiser

public JPhonemiser(String componentName,
           MaryDataType inputType,
           MaryDataType outputType,
           String allophonesProperty,
           String userdictProperty,
           String lexiconProperty,
           String ltsProperty)
            throws IOException,
                   MaryConfigurationException

Constructor providing the individual filenames of files that are required.

Parameters:

componentName - componentName

inputType - inputType

outputType - outputType

allophonesProperty - allophonesProperty

userdictProperty - userdictProperty

lexiconProperty - lexiconProperty

ltsProperty - ltsProperty

Throws:

IOException - IOException

MaryConfigurationException - MaryConfigurationException

JPhonemiser

public JPhonemiser(String componentName,
           MaryDataType inputType,
           MaryDataType outputType,
           String allophonesProperty,
           String userdictProperty,
           String lexiconProperty,
           String ltsProperty,
           String removetrailingonefromphonesProperty)
            throws IOException,
                   MaryConfigurationException

Constructor providing the individual filenames of files that are required.

Parameters:

componentName - componentName

inputType - inputType

outputType - outputType

allophonesProperty - allophonesProperty

userdictProperty - userdictProperty

lexiconProperty - lexiconProperty

ltsProperty - ltsProperty

removetrailingonefromphonesProperty - removetrailingonefromphonesProperty

Throws:

IOException - IOException

MaryConfigurationException - MaryConfigurationException

Method Detail

startup

public void startup()
             throws Exception

Description copied from interface: MaryModule

Allow the module to start up, performing whatever is necessary to become operational. After successful completion, getState() should return MODULE_RUNNING.

Specified by:

startup in interface MaryModule

Overrides:

startup in class InternalModule

Throws:

Exception - Exception

process

public MaryData process(MaryData d)
                 throws Exception

Description copied from class: InternalModule

Perform this module's processing on abstract "MaryData" input d. Subclasses need to make sure that the process() method is thread-safe, because in server-mode, it will be called from different threads at the same time. A sensible way to do this seems to be not to use any global or static variables, or to use them read-only.

Specified by:

process in interface MaryModule

Overrides:

process in class InternalModule

Parameters:

d - d

Returns:

A MaryData object of type outputType() encapsulating the processing result.

This method just returns its input. Subclasses should override this.

Throws:

Exception - Exception

phonemise

public String phonemise(String text,
               String pos,
               StringBuilder g2pMethod)

Phonemise the word text. This starts with a simple lexicon lookup, followed by some heuristics, and finally applies letter-to-sound rules if nothing else was successful.

Parameters:

text - the textual (graphemic) form of a word.

pos - the part-of-speech of the word

g2pMethod - This is an awkward way to return a second String parameter via a StringBuilder. If a phonemisation of the text is found, this parameter will be filled with the method of phonemisation ("lexicon", ... "rules").

Returns:

a phonemisation of the text if one can be generated, or null if no phonemisation method was successful.

lexiconLookup

public String lexiconLookup(String text,
                   String pos)

Look a given text up in the (standard) lexicon. part-of-speech is used in case of ambiguity.

Parameters:

text - text

pos - pos

Returns:

null if text == null or text.length is 0, null if entries.length is 0, entries[0] otherwise

userdictLookup

public String userdictLookup(String text,
                    String pos)

look a given text up in the userdict. part-of-speech is used in case of ambiguity.

Parameters:

text - text

pos - pos

Returns:

null if userdict is null or text is null or text.length is 0, null if entries is null, transcr otherwise

getAllophoneSet

public AllophoneSet getAllophoneSet()

Access the allophone set underlying this phonemiser.

Returns:

allophoneSet

readLexicon

protected Map<String,List<String>> readLexicon(String lexiconFilename)
                                        throws IOException

Read a lexicon. Lines must have the format graphemestring | phonestring | optional-parts-of-speech The pos-item is optional. Different pos's belonging to one grapheme chain may be separated by whitespace

Parameters:

lexiconFilename - lexiconFilename

Returns:

fLexicon

Throws:

IOException - IOException

setPh

protected void setPh(Element t,
         String ph)

setPunctuationPosRegex

protected void setPunctuationPosRegex()

Compile a regex pattern used to determine whether tokens are processed as punctuation or not, based on whether their pos attribute matches the pattern.

setUnpronounceablePosRegex

protected void setUnpronounceablePosRegex()

Compile a regex pattern used to determine whether tokens are processed as unprounounceable or not, based on whether their pos attribute matches the pattern.

isPosPunctuation

public boolean isPosPunctuation(String pos)

Based on the regex compiled in setPunctuationPosRegex(), determine whether a given POS string is classified as punctuation

Parameters:

pos - the POS tag

Returns:

true if the POS tag matches the regex pattern; false otherwise

Throws:

NullPointerException - if the regex pattern is null (because it hasn't been set during module startup)

isUnpronounceable

public boolean isUnpronounceable(String pos)

maybePronounceable

public boolean maybePronounceable(String text,
                         String pos)

Determine whether token should be pronounceable, based on text and POS tag.

Parameters:

text - the text of the token

pos - the POS tag of the token

Returns:

false if the text is empty, or if it contains no word characters and the POS tag indicates punctuation; true otherwise