de.up.ling.irtg

Class TemplateInterpretedTreeAutomaton

java.lang.Object

de.up.ling.irtg.TemplateInterpretedTreeAutomaton

public class TemplateInterpretedTreeAutomaton
extends Object

A template IRTG. A template IRTG is an IRTG in which rules may additionally be prefixed by quantification over variables using the foreach keyword, like so:

  foreach { x,y | left-of(x,y) and target(x) and distractor(y)}:
  NP_$x -> leftof_$x(N_$x, NP_$y)
  [sem] project_1(intersect_2(intersect_1(left-of, ?1), ?2))
  [string] *(?1, *("left of", ?2))

A template IRTG can be instantiated to an ordinary IRTG based on the contents of a FirstOrderModel using the instantiate(de.up.ling.irtg.util.FirstOrderModel) method. This will enumerate all values (a,b) for the variables (x,y) that satisfy all the conditions in the foreach clause, and will replace the occurrences of $x and $y in the IRTG rule by a and b respectively.

Template IRTGs are only rarely constructed programmatically. The common use-case is to read a template IRTG from a file using TemplateIrtgInputCodec. Most methods in this class are meant to be called from that input codec.

Author:

koller

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	TemplateInterpretedTreeAutomaton.AtomicGuard
static class	TemplateInterpretedTreeAutomaton.ConjGuard
static interface	TemplateInterpretedTreeAutomaton.Guard
static class	TemplateInterpretedTreeAutomaton.TemplateRule This class represents a template for a rule, from which multiple concrete rules can be instantiated.
static class	TemplateInterpretedTreeAutomaton.TopGuard

Constructor Summary

Constructors

Constructor and Description
TemplateInterpretedTreeAutomaton() Constructs an empty template IRTG.

Method Summary

Modifier and Type	Method and Description
void	addAlgebraClass(String interpretation, String className) Declares an interpretation.
void	addConstructorFeatureDeclaration(String id, String featureClass, List<String> arguments) Adds a declaration of a constructure feature to the template IRTG.
void	addRuleTemplate(TemplateInterpretedTreeAutomaton.TemplateRule trule) Adds a rule template to the template IRTG.
void	addStaticFeatureDeclaration(String id, String featureClass, String featureMethod, List<String> arguments) Adds a declaration of a static-method feature to the template IRTG.
Map<String,String>	getAlgebraClasses() Returns a string to string map of algebra classes.
List<String>	getFeatureIds() Returns a list of ids of the feature declarations this IRTG contains.
double	getFeatureWeight(int i) Returns an specific entry for the feature weights.
double[]	getFeatureWeights() Returns a double array which contains the current settings of the feature weights.
int	getNumFeatures() Returns the number of feature declarations this IRTG contains.
InterpretedTreeAutomaton	instantiate(FirstOrderModel model) Instantiates this template IRTG into a concrete IRTG.
void	setFeatureWeight(int index, double weight) Allows for setting a single entry for the feature weights.
void	setFeatureWeights(double[] featureWeights) Allows for the feature weights to be set externally.
String	toString()

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, wait, wait, wait

Constructor Detail

TemplateInterpretedTreeAutomaton

public TemplateInterpretedTreeAutomaton()

Constructs an empty template IRTG.

Method Detail

addRuleTemplate

public void addRuleTemplate(TemplateInterpretedTreeAutomaton.TemplateRule trule)

Adds a rule template to the template IRTG.

Parameters:

trule -

addConstructorFeatureDeclaration

public void addConstructorFeatureDeclaration(String id,
                                             String featureClass,
                                             List<String> arguments)

Adds a declaration of a constructure feature to the template IRTG.

Parameters:

id -

featureClass -

arguments -

addStaticFeatureDeclaration

public void addStaticFeatureDeclaration(String id,
                                        String featureClass,
                                        String featureMethod,
                                        List<String> arguments)

Adds a declaration of a static-method feature to the template IRTG.

Parameters:

id -

featureClass -

featureMethod -

arguments -

addAlgebraClass

public void addAlgebraClass(String interpretation,
                            String className)

Declares an interpretation.

Parameters:

interpretation -

className -

instantiate

public InterpretedTreeAutomaton instantiate(FirstOrderModel model)
                                     throws ClassNotFoundException,
                                            InstantiationException,
                                            IllegalAccessException,
                                            CodecParseException

Instantiates this template IRTG into a concrete IRTG. This is done by enumerating valid instances of each foreach-rule in the given first-order model (see the documentation for the class as a whole).

Parameters:

model -

Returns:

Throws:

ClassNotFoundException

InstantiationException

IllegalAccessException

CodecParseException

getNumFeatures

public int getNumFeatures()

Returns the number of feature declarations this IRTG contains.

Returns:

getFeatureIds

public List<String> getFeatureIds()

Returns a list of ids of the feature declarations this IRTG contains.

Returns:

getAlgebraClasses

public Map<String,String> getAlgebraClasses()

Returns a string to string map of algebra classes. Returns the actual map, not a defensive copy.

Returns:

the algebraClasses

getFeatureWeights

public double[] getFeatureWeights()

Returns a double array which contains the current settings of the feature weights. Returns actual array used by the class, not a defensive copy.

Returns:

the featureWeights

setFeatureWeights

public void setFeatureWeights(double[] featureWeights)

Allows for the feature weights to be set externally.

Parameters:

featureWeights - the featureWeights to set

getFeatureWeight

public double getFeatureWeight(int i)

Returns an specific entry for the feature weights.

Parameters:

i -

Returns:

setFeatureWeight

public void setFeatureWeight(int index,
                             double weight)

Allows for setting a single entry for the feature weights.

Parameters:

index -

weight -

toString

public String toString()

Overrides:

toString in class Object