marytts.language.de.preprocess

Class NumberEP

java.lang.Object

marytts.language.de.preprocess.ExpansionPattern

marytts.language.de.preprocess.NumberEP

public class NumberEP
extends ExpansionPattern

An expansion pattern implementation for basic number patterns.

Author:

Marc Schröder

Field Summary

Fields

Modifier and Type	Field and Description
protected Pattern	reDigits
protected Pattern	reFloat
protected Pattern	reInteger
protected Pattern	reOrdinal
protected Pattern	reRoman
protected static String	sCardinal
protected static String	sDigits
protected static String	sFloat
protected static String	sInteger
protected static String	sOrdinal
protected static String	sRoman

Fields inherited from class marytts.language.de.preprocess.ExpansionPattern

abbrev, composite, currency, date, duration, measure, multiword, net, number, specialChar, telephone, time

Constructor Summary

Constructors

Constructor and Description
NumberEP()

Method Summary

Methods

Modifier and Type	Method and Description
protected boolean	allowMultipleTokens() Simple numbers are expected to be entire tokens.
protected int	canDealWith(String input, int typeCode) Decide whether we can expand a string according to type typeCode.
protected List<Element>	expand(List<Element> tokens, String s, int type) Subclasses do their expansion in this class.
protected List<Element>	expandDigits(Document doc, String s, boolean createMtu)
protected String	expandDigits(String digits)
protected List<Element>	expandFloat(Document doc, String s, boolean createMtu) This will correctly expand integers as well, although matchFloat() does not match them.
protected String	expandFloat(String number)
protected List<Element>	expandInteger(Document doc, long value, boolean createMtu, String orig)
protected List<Element>	expandInteger(Document doc, String s, boolean createMtu)
protected String	expandInteger(long value)
protected String	expandInteger(String s)
protected List<Element>	expandOrdinal(Document doc, long value, boolean createMtu, String orig)
protected List<Element>	expandOrdinal(Document doc, String s, boolean createMtu) For ordinals we put the expanded form in the sounds_like attribute and keep the surface form.
protected List<Element>	expandRoman(Document doc, String number, boolean createMtu)
protected List<Element>	expandRoman(Document doc, String number, boolean createMtu, boolean isOrdinal)
protected boolean	isCandidate(Element t)
List<String>	knownTypes() Returns the types known by this ExpansionPattern.
protected int	match(String s, int type) Subclasses do their matching in this class.
protected boolean	matchDigits(String s)
protected boolean	matchFloat(String s)
protected boolean	matchInteger(String s)
protected boolean	matchOrdinal(String s)
protected boolean	matchRoman(String s)
Pattern	reMatchingChars() Returns the regular expression object matching any of the chars occurring in the pattern.

Methods inherited from class marytts.language.de.preprocess.ExpansionPattern

allPatterns, doesFullExpansion, getPattern, getSplitAtChars, makeNewTokens, makeNewTokens, makeNewTokens, match, process, replaceTokens, reSplitAtChars, slowDown, slowDown

Methods inherited from class java.lang.Object

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

Field Detail

sFloat

protected static final String sFloat

See Also:

Constant Field Values

sInteger

protected static final String sInteger

See Also:

Constant Field Values

sOrdinal

protected static final String sOrdinal

See Also:

Constant Field Values

sRoman

protected static final String sRoman

See Also:

Constant Field Values

sDigits

protected static final String sDigits

See Also:

Constant Field Values

sCardinal

protected static final String sCardinal

See Also:

Constant Field Values

reFloat

protected final Pattern reFloat

reInteger

protected final Pattern reInteger

reOrdinal

protected final Pattern reOrdinal

reRoman

protected final Pattern reRoman

reDigits

protected final Pattern reDigits

Constructor Detail

NumberEP

public NumberEP()

Method Detail

knownTypes

public List<String> knownTypes()

Description copied from class: ExpansionPattern

Returns the types known by this ExpansionPattern. These are possible values of the type attribute to the say-as element, as defined in MaryXML.dtd. Each subclass needs to override this to return something meaningful.

Specified by:

knownTypes in class ExpansionPattern

Returns:

known types

reMatchingChars

public Pattern reMatchingChars()

Description copied from class: ExpansionPattern

Returns the regular expression object matching any of the chars occurring in the pattern. Each subclass needs to override this to return something meaningful.

Specified by:

reMatchingChars in class ExpansionPattern

Returns:

reMatchingChars

allowMultipleTokens

protected boolean allowMultipleTokens()

Simple numbers are expected to be entire tokens. They should not be joined together out of several tokens.

Overrides:

allowMultipleTokens in class ExpansionPattern

Returns:

false

isCandidate

protected boolean isCandidate(Element t)

Overrides:

isCandidate in class ExpansionPattern

match

protected int match(String s,
        int type)

Description copied from class: ExpansionPattern

Subclasses do their matching in this class.

Specified by:

match in class ExpansionPattern

Parameters:

s - is the String to be matched,

type - is the index in knownTypes to match with.

Returns:

type actually matched on successful match with this type (if typeCode is a general type ( typeCode == 0), it may have matched with a more specific subtype). On failure, -1 is returned.

canDealWith

protected int canDealWith(String input,
              int typeCode)

Description copied from class: ExpansionPattern

Decide whether we can expand a string according to type typeCode. This is important in cases where a particular expansion is requested via a say-as element. As a default, reply that a string can be expanded if it would be matched by the pattern recogniser. Subclasses may wish to override this with less strict requirements. Returns the type as which it can be expanded, or -1 if expansion is not possible.

Specified by:

canDealWith in class ExpansionPattern

Parameters:

input - input

typeCode - typeCode

Returns:

true if it can deal with input and typecode

expand

protected List<Element> expand(List<Element> tokens,
                   String s,
                   int type)

Description copied from class: ExpansionPattern

Subclasses do their expansion in this class.

Specified by:

expand in class ExpansionPattern

Parameters:

tokens - is a list of token Elements to be replaced with their expanded form. The expanded forms are inserted into the DOM tree at the same positions as the tokens in List tokens. If there are more new tokens than old tokens, the rest are inserted as siblings at the position of the last old token.

s - is the String to be expanded,

type - is the index in knownTypes this string has matched with before.

Returns:

the list of expanded (=new) tokens.

matchFloat

protected boolean matchFloat(String s)

matchInteger

protected boolean matchInteger(String s)

matchOrdinal

protected boolean matchOrdinal(String s)

matchRoman

protected boolean matchRoman(String s)

matchDigits

protected boolean matchDigits(String s)

expandInteger

protected List<Element> expandInteger(Document doc,
                          String s,
                          boolean createMtu)

expandInteger

protected List<Element> expandInteger(Document doc,
                          long value,
                          boolean createMtu,
                          String orig)

expandInteger

protected String expandInteger(String s)

expandInteger

protected String expandInteger(long value)

expandFloat

protected List<Element> expandFloat(Document doc,
                        String s,
                        boolean createMtu)

This will correctly expand integers as well, although matchFloat() does not match them. This seems to be convenient in cases where "some number", i.e. integer or float, was matched, and needs to be expanded.

Parameters:

doc - doc

s - s

createMtu - createMtu

Returns:

makeNewTokens(doc, expString, createMtu, s)

expandFloat

protected String expandFloat(String number)

expandDigits

protected List<Element> expandDigits(Document doc,
                         String s,
                         boolean createMtu)

expandDigits

protected String expandDigits(String digits)

expandOrdinal

protected List<Element> expandOrdinal(Document doc,
                          String s,
                          boolean createMtu)

For ordinals we put the expanded form in the sounds_like attribute and keep the surface form. This is for the POS tagger to tell a later module whether the ordinal is adverbial or adjectival.

Parameters:

doc - doc

s - s

createMtu - createMtu

Returns:

expandOrdinal(doc, value, createMtu, s)

expandOrdinal

protected List<Element> expandOrdinal(Document doc,
                          long value,
                          boolean createMtu,
                          String orig)

expandRoman

protected List<Element> expandRoman(Document doc,
                        String number,
                        boolean createMtu)

expandRoman

protected List<Element> expandRoman(Document doc,
                        String number,
                        boolean createMtu,
                        boolean isOrdinal)