marytts.modules

Class ProsodyGeneric

java.lang.Object

marytts.modules.InternalModule

marytts.modules.ProsodyGeneric

All Implemented Interfaces:

MaryModule

Direct Known Subclasses:

Prosody, ProsodyGenericFST

public class ProsodyGeneric
extends InternalModule

The generic prosody module.

Author:

Stephanie Becker

Field Summary

Fields

Modifier and Type	Field and Description
protected boolean	accentedSyllables
protected String	accentPriorities
protected boolean	applyParagraphDeclination
protected HashMap<String,Object>	listMap
protected static Pattern	nextPlusXAttributesPattern
protected static Pattern	nextPlusXTextPattern
protected String	paragraphDeclination
protected static Pattern	previousMinusXAttributesPattern
protected static Pattern	previousMinusXTextPattern
protected Properties	priorities
protected String	syllableAccents
protected HashMap<String,String>	toBI2ContourMap
protected String	tobiPredFilename
protected HashMap<String,Element>	tobiPredMap

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
ProsodyGeneric()
ProsodyGeneric(Locale locale)
ProsodyGeneric(Locale locale, String propertyPrefix)
ProsodyGeneric(MaryDataType inputType, MaryDataType outputType, Locale locale, String tobipredFileName, String accentPriorities, String syllableAccents, String paragraphDeclination)
ProsodyGeneric(String locale)
ProsodyGeneric(String locale, String propertyPrefix)

Method Summary

Methods

Modifier and Type	Method and Description
protected boolean	applyRules(Node n) Verify whether this Node has a parent preventing the application of intonation rules.
protected void	buildListMap()
protected boolean	checkAttributes(Element currentRulePart, Element token) checks rule part with tag "attributes"; checks if the MaryXML attributes and values of current token are the same as in the rule
protected boolean	checkAttributesOfOtherToken(String tag, Element currentRulePart, int position, NodeList tokens) checks rule part with tag "nextAttributes","previousAttributes","nextPlusXAttributes","previousMinusXAttributes"; checks if the MaryXML attributes and values of other token than the current one are the same as in rule (f.e.
protected boolean	checkFolTokens(Element currentRulePart, int position, NodeList tokens) checks rule part with tag "folTokens"; there is only the "num" attribute right now; checks if the number of the following tokens after the current token is the same as the value of the num attribute; f.e.
protected boolean	checkFolWords(Element currentRulePart, int position, NodeList tokens) checks rule part with tag "folWords"; there is only the "num" attribute right now; checks if the number of the following words after the current token is the same as the value of the num attribute; f.e.
protected boolean	checkList(String currentVal, String tokenValue) Checks if tokenValue is contained in list.
protected boolean	checkPrevTokens(Element currentRulePart, int position, NodeList tokens) checks rule part with tag "prevTokens"; there is only the "num" attribute right now; checks if the number of the tokens preceding the current token is the same as the value of the num attribute; f.e.
protected boolean	checkPrevWords(Element currentRulePart, int position, NodeList tokens) checks rule part with tag "prevWords"; there is only the "num" attribute right now; checks if the number of the words preceding the current token is the same as the value of the num attribute; f.e.
protected boolean	checkProsodicPosition(Element currentRulePart, String prosodicPositionType) checks rule part with tag "prosodicPosition"; there is only the "type" attribute right now: checks if prosodic position of a token is the same as the value of the type attribute in the rule; values: prenuclear, nuclearParagraphFinal, nuclearParagraphNonFinal, postnuclear
protected boolean	checkRulePart(Element currentRulePart, Element token, NodeList tokens, int position, String sentenceType, String specialPositionType, String tokenText) checks condition of a rule part, f.e.
protected boolean	checkSentence(Element currentRulePart, String sentenceType) checks rule part with tag "sentence"; there is only the "type" attribute right now: checks if sentence type of a token is the same as the value of the type attribute in the rule
protected boolean	checkSpecialPosition(Element currentRulePart, String specialPositionType) checks rule part with tag "specialPosition"; there is only the "type" attribute right now: checks if specialPosition value of a token is the same as the value of the type attribute in the rule; values: endofvorfeld, endofpar (end of paragraph)
protected boolean	checkText(Element currentRulePart, String tokenText) checks rule part with tag "text"; there is only the "word" attribute right now: checks if text of a token is the same as the value of the word attribute in the rule
protected boolean	checkTextOfOtherToken(String tag, Element currentRulePart, int position, NodeList tokens) checks rule part with tag "nextText","previousText","nextPlusXText" or "previousMinusXText"; there is only the "word" attribute right now: checks if text of a token is the same as the value of the word attribute in the rule
protected void	copyAccentsToSyllables(Document doc) Go through all tokens in a document, and copy any accents to the first accented syllable.
protected void	getAccentPosition(Element token, NodeList tokens, int position, String sentenceType, String specialPositionType) checks if token receives an accent or not the information is contained in the accentposition part of rules in xml file the token attribute "accent" receives the value "tone","force"(force accent(Druckakzent)) or ""(no accent)
protected boolean	getAccentShape(Element token, NodeList tokens, int position, String sentenceType, String specialPositionType, boolean nucleusAssigned) determines accent types; tokens with accent="tone" will receive an accent type (f.e."L+H*"), accent="force" becomes "*" the relevant information is contained in the accentshape part of rules in xml file
protected Element	getBoundary(Element token, NodeList tokens, int position, String sentenceType, String specialPositionType, boolean invalidXML, Element firstTokenInPhrase) checks if a boundary is to be inserted after the current token the information is contained in the boundaries part of rules in xml file
protected String	getForceAccent(Element token) Check whether token is enclosed by a <prosody> element containing an attribute force-accent.
protected String	getSentenceType(NodeList tokens) determination of sentence type values: decl, excl, interrog, interrogYN or interrogW
protected Element	insertBoundary(Element token, String tone, int bi) Insert a boundary after token, with the given tone and breakindex.
protected Element	insertMajorBoundary(NodeList tokens, int i, Element firstToken, String tone, int breakindex) Insert a major boundary after token number i in tokens.
protected boolean	insertPhraseNode(Element first, Element last) Inserte a phrase element, enclosing the first and last element, into the tree.
protected boolean	isPunctuation(Element token) Verify whether a given token is a punctuation.
protected void	loadTobiPredRules()
MaryData	process(MaryData d) Perform this module's processing on abstract "MaryData" input d.
protected void	processSentence(Element sentence)
protected Object	readListFromResource(String resourceName) Read a list from an external file.
protected void	setAccent(Element token, String accent) Assign an accent to the given token.
void	startup() Allow the module to start up, performing whatever is necessary to become operational.

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

paragraphDeclination

protected String paragraphDeclination

applyParagraphDeclination

protected boolean applyParagraphDeclination

syllableAccents

protected String syllableAccents

accentedSyllables

protected boolean accentedSyllables

accentPriorities

protected String accentPriorities

priorities

protected Properties priorities

tobiPredFilename

protected String tobiPredFilename

tobiPredMap

protected HashMap<String,Element> tobiPredMap

listMap

protected HashMap<String,Object> listMap

toBI2ContourMap

protected HashMap<String,String> toBI2ContourMap

nextPlusXTextPattern

protected static final Pattern nextPlusXTextPattern

previousMinusXTextPattern

protected static final Pattern previousMinusXTextPattern

nextPlusXAttributesPattern

protected static final Pattern nextPlusXAttributesPattern

previousMinusXAttributesPattern

protected static final Pattern previousMinusXAttributesPattern

Constructor Detail

ProsodyGeneric

public ProsodyGeneric()

ProsodyGeneric

public ProsodyGeneric(MaryDataType inputType,
              MaryDataType outputType,
              Locale locale,
              String tobipredFileName,
              String accentPriorities,
              String syllableAccents,
              String paragraphDeclination)

ProsodyGeneric

public ProsodyGeneric(String locale,
              String propertyPrefix)

ProsodyGeneric

public ProsodyGeneric(Locale locale,
              String propertyPrefix)

ProsodyGeneric

public ProsodyGeneric(String locale)

ProsodyGeneric

public ProsodyGeneric(Locale locale)

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

loadTobiPredRules

protected void loadTobiPredRules()
                          throws FactoryConfigurationError,
                                 ParserConfigurationException,
                                 SAXException,
                                 IOException,
                                 NoSuchPropertyException,
                                 MaryConfigurationException

Throws:

FactoryConfigurationError

ParserConfigurationException

SAXException

IOException

NoSuchPropertyException

MaryConfigurationException

buildListMap

protected void buildListMap()
                     throws IOException

Throws:

IOException

readListFromResource

protected Object readListFromResource(String resourceName)
                               throws IOException

Read a list from an external file. This generic implementation can read from text files (filenames ending in .txt). Subclasses may override this class to provide additional file formats. They must make sure that checkList() can deal with all list formats.

Parameters:

resourceName - resource file in classpath from which to read the list; suffix identifies list format.

Returns:

An Object representing the list; checkList() must be able to make sense of this. This base implementation returns a Set<String>.

Throws:

IllegalArgumentException - if the fileName suffix cannot be identified as a list file format.

IOException - if the file given in fileName cannot be found or read from

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

processSentence

protected void processSentence(Element sentence)

getAccentPosition

protected void getAccentPosition(Element token,
                     NodeList tokens,
                     int position,
                     String sentenceType,
                     String specialPositionType)

checks if token receives an accent or not the information is contained in the accentposition part of rules in xml file the token attribute "accent" receives the value "tone","force"(force accent(Druckakzent)) or ""(no accent)

Parameters:

token - (current token)

tokens - (list of all tokens in sentence)

position - (position in token list)

sentenceType - (declarative, exclamative or interrogative)

specialPositionType - (end of vorfeld or end of paragraph)

getAccentShape

protected boolean getAccentShape(Element token,
                     NodeList tokens,
                     int position,
                     String sentenceType,
                     String specialPositionType,
                     boolean nucleusAssigned)

determines accent types; tokens with accent="tone" will receive an accent type (f.e."L+H*"), accent="force" becomes "*" the relevant information is contained in the accentshape part of rules in xml file

Parameters:

token - (current token)

tokens - (list of all tokens in sentence)

position - position

sentenceType - (declarative, exclamative or interrogative)

specialPositionType - (position in sentence)

nucleusAssigned - (test, if nuclear accent is already assigned)

Returns:

nucleusAssigned

getBoundary

protected Element getBoundary(Element token,
                  NodeList tokens,
                  int position,
                  String sentenceType,
                  String specialPositionType,
                  boolean invalidXML,
                  Element firstTokenInPhrase)

checks if a boundary is to be inserted after the current token the information is contained in the boundaries part of rules in xml file

Parameters:

token - (current token)

tokens - (list of tokens in sentence)

position - (position in token list)

sentenceType - (declarative, exclamative or interrogative)

specialPositionType - (endofvorfeld if sentence has vorfeld and the next token is a finite verb or end of paragraph)

invalidXML - (true if xml structure allows boundary insertion)

firstTokenInPhrase - (begin of intonation phrase)

Returns:

firstTokenInPhrase (if a boundary was inserted, firstTokenInPhrase gets null)

checkRulePart

protected boolean checkRulePart(Element currentRulePart,
                    Element token,
                    NodeList tokens,
                    int position,
                    String sentenceType,
                    String specialPositionType,
                    String tokenText)

checks condition of a rule part, f.e. attributes pos="NN"

Parameters:

currentRulePart - currentRulePart

token - (current token)

tokens - (list of all tokens)

position - (position in token list)

sentenceType - (declarative, exclamative or interrogative)

specialPositionType - (special position in sentence(end of vorfeld) or text(end of paragraph))

tokenText - (text of token)

Returns:

true if condition is satisfied

checkText

protected boolean checkText(Element currentRulePart,
                String tokenText)

checks rule part with tag "text"; there is only the "word" attribute right now: checks if text of a token is the same as the value of the word attribute in the rule

Parameters:

currentRulePart - currentRulePart

tokenText - tokenText

Returns:

checkList(currentVal, tokenText)

checkTextOfOtherToken

protected boolean checkTextOfOtherToken(String tag,
                            Element currentRulePart,
                            int position,
                            NodeList tokens)

checks rule part with tag "nextText","previousText","nextPlusXText" or "previousMinusXText"; there is only the "word" attribute right now: checks if text of a token is the same as the value of the word attribute in the rule

Parameters:

tag - tag

currentRulePart - currentRulePart

position - position

tokens - tokens

Returns:

checkText(currentRulePart, otherTokenText)

checkFolTokens

protected boolean checkFolTokens(Element currentRulePart,
                     int position,
                     NodeList tokens)

checks rule part with tag "folTokens"; there is only the "num" attribute right now; checks if the number of the following tokens after the current token is the same as the value of the num attribute; f.e. the value "3+" means: at least 3 following tokens, "3-": not more than 3, "3": exactly 3

Parameters:

currentRulePart - currentRulePart

position - position

tokens - tokens

Returns:

true if everything is fine

checkPrevTokens

protected boolean checkPrevTokens(Element currentRulePart,
                      int position,
                      NodeList tokens)

checks rule part with tag "prevTokens"; there is only the "num" attribute right now; checks if the number of the tokens preceding the current token is the same as the value of the num attribute; f.e. the value "3+" means: at least 3 preceding tokens, "3-": not more than 3, "3": exactly 3

Parameters:

currentRulePart - currentRulePart

position - position

tokens - tokens

Returns:

true if everything passes

checkFolWords

protected boolean checkFolWords(Element currentRulePart,
                    int position,
                    NodeList tokens)

checks rule part with tag "folWords"; there is only the "num" attribute right now; checks if the number of the following words after the current token is the same as the value of the num attribute; f.e. the value "3+" means: at least 3 following tokens, "3-": not more than 3, "3": exactly 3

Parameters:

currentRulePart - currentRulePart

position - position

tokens - tokens

Returns:

true if everything passes

checkPrevWords

protected boolean checkPrevWords(Element currentRulePart,
                     int position,
                     NodeList tokens)

checks rule part with tag "prevWords"; there is only the "num" attribute right now; checks if the number of the words preceding the current token is the same as the value of the num attribute; f.e. the value "3+" means: at least 3 preceding tokens, "3-": not more than 3, "3": exactly 3

Parameters:

currentRulePart - currentRulePart

position - position

tokens - tokens

Returns:

true if everything passes

checkSentence

protected boolean checkSentence(Element currentRulePart,
                    String sentenceType)

checks rule part with tag "sentence"; there is only the "type" attribute right now: checks if sentence type of a token is the same as the value of the type attribute in the rule

Parameters:

currentRulePart - currentRulePart

sentenceType - sentenceType

Returns:

true if everything passes

checkSpecialPosition

protected boolean checkSpecialPosition(Element currentRulePart,
                           String specialPositionType)

checks rule part with tag "specialPosition"; there is only the "type" attribute right now: checks if specialPosition value of a token is the same as the value of the type attribute in the rule; values: endofvorfeld, endofpar (end of paragraph)

Parameters:

currentRulePart - currentRulePart

specialPositionType - specialPositionType

Returns:

true if everything passes

checkProsodicPosition

protected boolean checkProsodicPosition(Element currentRulePart,
                            String prosodicPositionType)

checks rule part with tag "prosodicPosition"; there is only the "type" attribute right now: checks if prosodic position of a token is the same as the value of the type attribute in the rule; values: prenuclear, nuclearParagraphFinal, nuclearParagraphNonFinal, postnuclear

Parameters:

currentRulePart - currentRulePart

prosodicPositionType - prosodicPositionType

Returns:

true if everything passes

checkAttributes

protected boolean checkAttributes(Element currentRulePart,
                      Element token)

checks rule part with tag "attributes"; checks if the MaryXML attributes and values of current token are the same as in the rule

Parameters:

currentRulePart - currentRulePart

token - token

Returns:

checkList(currentVal, token.getAttribute(currentAtt))

checkAttributesOfOtherToken

protected boolean checkAttributesOfOtherToken(String tag,
                                  Element currentRulePart,
                                  int position,
                                  NodeList tokens)

checks rule part with tag "nextAttributes","previousAttributes","nextPlusXAttributes","previousMinusXAttributes"; checks if the MaryXML attributes and values of other token than the current one are the same as in rule (f.e. the 3th token after current token)

Parameters:

tag - tag

currentRulePart - currentRulePart

position - position

tokens - tokens

Returns:

checkAttributes(currentRulePart, otherToken)

checkList

protected boolean checkList(String currentVal,
                String tokenValue)

Checks if tokenValue is contained in list. This base implementation is able to deal with list types represented as Sets; subclasses may override this method to be able to deal with different list representations.

Parameters:

currentVal - the condition to check; can be either INLIST: or !INLIST: followed by the list name to check.

tokenValue - value to look up in the list

Returns:

whether or not tokenValue is contained in the list.

getSentenceType

protected String getSentenceType(NodeList tokens)

determination of sentence type values: decl, excl, interrog, interrogYN or interrogW

Parameters:

tokens - tokens

Returns:

sentenceType

setAccent

protected void setAccent(Element token,
             String accent)

Assign an accent to the given token.

Parameters:

token - a token element

accent - the accent string to assign.

insertBoundary

protected Element insertBoundary(Element token,
                     String tone,
                     int bi)

Insert a boundary after token, with the given tone and breakindex. If a boundary element already exists after token (but before the following token), it is reused, if both token and boundary have the same parent node. In addition, if token is punctuation, a boundary preceding token can be reused, if both have the same parent node. When choosing between the values already given in the existing element and the ones passed as arguments to this function, the higher / more concrete values are taken: Only if bi is higher than an already existing breakindex, the old value is replaced with bi. Only if tone is a concrete tone (like "h-") and the previous tone was "unknown" or not specified at all, tone is taken into account.

Parameters:

token - token

tone - tone

bi - bi

Returns:

the boundary element on success, null on failure.

insertMajorBoundary

protected Element insertMajorBoundary(NodeList tokens,
                          int i,
                          Element firstToken,
                          String tone,
                          int breakindex)

Insert a major boundary after token number i in tokens.

Also inserts a phrase tag at the appropriate position.

Parameters:

tokens - tokens

i - i

firstToken - firstToken

tone - tone

breakindex - breakindex

Returns:

The boundary element.

insertPhraseNode

protected boolean insertPhraseNode(Element first,
                       Element last)

Inserte a phrase element, enclosing the first and last element, into the tree. Typically first element would be a token, last element a boundary.

Parameters:

first - first

last - last

Returns:

true on success, false on failure.

applyRules

protected boolean applyRules(Node n)

Verify whether this Node has a parent preventing the application of intonation rules.

Parameters:

n - n

Returns:

true if rules are to be applied, false otherwise.

copyAccentsToSyllables

protected void copyAccentsToSyllables(Document doc)

Go through all tokens in a document, and copy any accents to the first accented syllable.

Parameters:

doc - doc

getForceAccent

protected String getForceAccent(Element token)

Check whether token is enclosed by a <prosody> element containing an attribute force-accent.

Parameters:

token - token

Returns:

the value of the force-accent attribute, if one exists, or the empty string otherwise.

isPunctuation

protected boolean isPunctuation(Element token)

Verify whether a given token is a punctuation.

Parameters:

token - the t element to be tested.

Returns:

true if token is a punctuation, false otherwise.