gate.creole.kea
Class Kea

java.lang.Object
  gate.util.AbstractFeatureBearer
      gate.creole.AbstractResource
          gate.creole.AbstractProcessingResource
              gate.creole.AbstractLanguageAnalyser
                  gate.creole.kea.Kea

All Implemented Interfaces:

ANNIEConstants, Executable, ActionsPublisher, LanguageAnalyser, ProcessingResource, Resource, FeatureBearer, NameBearer, Serializable

public class Kea
extends AbstractLanguageAnalyser
implements ActionsPublisher

This is wrapper for using the KEA Keyphrase extractor (http://www.nzdl.org/Kea/) within the GATE Language Engineering architecture (http://gate.ac.uk). It exposes KEA as a GATE Processing Resource that has two functioning modes:

• Training mode: when keyphrases (marked as annotations on documents) are collected and a model is built.
• Application mode: when a built model is applied on documents to the end of extracting keyphrases.

See Also:

Serialized Form

Nested Class Summary

protected class	Kea.LoadModelAction Action for loading a saved model.
protected class	Kea.SaveModelAction Action used to save a trained model.

Nested classes/interfaces inherited from class gate.creole.AbstractProcessingResource

AbstractProcessingResource.InternalStatusListener, AbstractProcessingResource.IntervalProgressListener

Field Summary

protected List	actions The list of GUI actions available from this PR on popup menus.
protected weka.core.FastVector	atts Data structure used internally to define the dataset.
protected weka.core.Instances	data The dataset.
protected kea.KEAFilter	keaFilter The KEA filter object which incorporates the actual model.
protected boolean	trainingFinished This flag is used to determine whether the model has been constructed or not.

Fields inherited from class gate.creole.AbstractLanguageAnalyser

corpus, document

Fields inherited from class gate.creole.AbstractProcessingResource

interrupted

Fields inherited from class gate.creole.AbstractResource

name

Fields inherited from class gate.util.AbstractFeatureBearer

features

Fields inherited from interface gate.creole.ANNIEConstants

ANNOTATION_COREF_FEATURE_NAME, DATE_ANNOTATION_TYPE, DATE_POSTED_ANNOTATION_TYPE, DOCUMENT_COREF_FEATURE_NAME, JOB_ID_ANNOTATION_TYPE, LOCATION_ANNOTATION_TYPE, LOOKUP_ANNOTATION_TYPE, LOOKUP_CLASS_FEATURE_NAME, LOOKUP_MAJOR_TYPE_FEATURE_NAME, LOOKUP_MINOR_TYPE_FEATURE_NAME, LOOKUP_ONTOLOGY_FEATURE_NAME, MONEY_ANNOTATION_TYPE, ORGANIZATION_ANNOTATION_TYPE, PERSON_ANNOTATION_TYPE, PERSON_GENDER_FEATURE_NAME, PR_NAMES, SENTENCE_ANNOTATION_TYPE, SPACE_TOKEN_ANNOTATION_TYPE, TOKEN_ANNOTATION_TYPE, TOKEN_CATEGORY_FEATURE_NAME, TOKEN_KIND_FEATURE_NAME, TOKEN_LENGTH_FEATURE_NAME, TOKEN_ORTH_FEATURE_NAME, TOKEN_STRING_FEATURE_NAME

Constructor Summary

Kea()
Anonymous constructor, required by GATE.

Method Summary

protected void	annotateKeyPhrases(List phrases) Annnotates the document with all the occurences of keyphrases from a List.
void	execute() Executes this PR.
protected void	finishTraining() Stops the training phase and builds the actual model.
List	getActions() Gets the list of GUI actions available from this PR.
Boolean	getDisallowInternalPeriods()
String	getInputAS() Gets the name for the input annotation set.
String	getKeyphraseAnnotationType() Sets the annotation type to be used for keyphrases.
Integer	getMaxPhraseLength()
Integer	getMinNumOccur()
Integer	getMinPhraseLength()
String	getOutputAS() Gets the name for the output annotation set.
Integer	getPhrasesToExtract()
Boolean	getTrainingMode()
Boolean	getUseKFrequency()
Resource	init() Initialises this KEA Processing Resource.
protected void	initModel() Initialises the KEA model.
void	setDisallowInternalPeriods(Boolean dissallowInternalPeriods)
void	setInputAS(String inputAS) Sets the name for the input annotation set.
void	setKeyphraseAnnotationType(String keyphraseAnnotationType) Sets the annotation type to be used for keyphrases.
void	setMaxPhraseLength(Integer maxPhraseLength)
void	setMinNumOccur(Integer minNumOccur)
void	setMinPhraseLength(Integer minPhraseLength)
void	setOutputAS(String outputAS) Sets the name for the output annotation set.
void	setPhrasesToExtract(Integer phrasesToExtract)
void	setTrainingMode(Boolean trainingMode)
void	setUseKFrequency(Boolean useKFrequency)

Methods inherited from class gate.creole.AbstractLanguageAnalyser

getCorpus, getDocument, setCorpus, setDocument

Methods inherited from class gate.creole.AbstractProcessingResource

addProgressListener, addStatusListener, cleanup, fireProcessFinished, fireProgressChanged, fireStatusChanged, interrupt, isInterrupted, reInit, removeProgressListener, removeStatusListener

Methods inherited from class gate.creole.AbstractResource

checkParameterValues, getBeanInfo, getName, getParameterValue, getParameterValue, removeResourceListeners, setName, setParameterValue, setParameterValue, setParameterValues, setParameterValues, setResourceListeners

Methods inherited from class gate.util.AbstractFeatureBearer

getFeatures, setFeatures

Methods inherited from class java.lang.Object

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

Methods inherited from interface gate.ProcessingResource

reInit

Methods inherited from interface gate.Resource

cleanup, getParameterValue, setParameterValue, setParameterValues

Methods inherited from interface gate.util.FeatureBearer

getFeatures, setFeatures

Methods inherited from interface gate.util.NameBearer

getName, setName

Methods inherited from interface gate.Executable

interrupt, isInterrupted

Field Detail

trainingFinished

protected boolean trainingFinished

This flag is used to determine whether the model has been constructed or not. During training mode the training data is simply collected and this flag is set to false. The first time when the traied model is required (which could be either the first time the application mode is started or when the model is being saved) the model is built from the collected instances and this flag is set to true.
If this flag is found to be true during training phase (i.e. there is an attempt to train an already triend model) then the current model will be discarded and a new one will be created. The traininig will be performed using the newly created model.

keaFilter

protected kea.KEAFilter keaFilter

The KEA filter object which incorporates the actual model.

atts

protected weka.core.FastVector atts

Data structure used internally to define the dataset.

data

protected weka.core.Instances data

The dataset.

actions

protected List actions

The list of GUI actions available from this PR on popup menus.

Constructor Detail

Kea

public Kea()

Anonymous constructor, required by GATE. Does nothing.

Method Detail

getActions

public List getActions()

Gets the list of GUI actions available from this PR. Currently Load and Save model.

Specified by:

getActions in interface ActionsPublisher

Returns:

execute

public void execute()
             throws ExecutionException

Executes this PR. Depeding on the state of the trainingMode switch it will either train a model or apply it over the documents.
Trainig consists of collecting keyphrase annotations from the input annotation set of the input documents. The first time a trained model is required (either application mode has started or the model is being saved) the actual model ({link @ #keaModel}) will be constructed.
The application mode consists of using a trained model to generate keyphrase annotations on the output annotation set of the input documents.

Specified by:

execute in interface Executable

Overrides:

execute in class AbstractProcessingResource

Throws:

ExecutionException

annotateKeyPhrases

protected void annotateKeyPhrases(List phrases)
                           throws Exception

Annnotates the document with all the occurences of keyphrases from a List. Uses the java.util.regex package to search for ocurences of keyphrases.

Parameters:

phrases - the list of keyphrases.

Throws:

Exception

setKeyphraseAnnotationType

public void setKeyphraseAnnotationType(String keyphraseAnnotationType)

Sets the annotation type to be used for keyphrases.

Parameters:

keyphraseAnnotationType -

getKeyphraseAnnotationType

public String getKeyphraseAnnotationType()

Sets the annotation type to be used for keyphrases.

Returns:

setInputAS

public void setInputAS(String inputAS)

Sets the name for the input annotation set.

Parameters:

inputAS -

getInputAS

public String getInputAS()

Gets the name for the input annotation set.

Returns:

setOutputAS

public void setOutputAS(String outputAS)

Sets the name for the output annotation set.

Parameters:

outputAS -

getOutputAS

public String getOutputAS()

Gets the name for the output annotation set.

Returns:

init

public Resource init()
              throws ResourceInstantiationException

Initialises this KEA Processing Resource.

Specified by:

init in interface Resource

Overrides:

init in class AbstractProcessingResource

Returns:

Throws:

ResourceInstantiationException

initModel

protected void initModel()
                  throws Exception

Initialises the KEA model.

Throws:

Exception

finishTraining

protected void finishTraining()
                       throws ExecutionException

Stops the training phase and builds the actual model.

Throws:

ExecutionException

setMaxPhraseLength

public void setMaxPhraseLength(Integer maxPhraseLength)

getMaxPhraseLength

public Integer getMaxPhraseLength()

setMinPhraseLength

public void setMinPhraseLength(Integer minPhraseLength)

getMinPhraseLength

public Integer getMinPhraseLength()

setDisallowInternalPeriods

public void setDisallowInternalPeriods(Boolean dissallowInternalPeriods)

getDisallowInternalPeriods

public Boolean getDisallowInternalPeriods()

setUseKFrequency

public void setUseKFrequency(Boolean useKFrequency)

getUseKFrequency

public Boolean getUseKFrequency()

setMinNumOccur

public void setMinNumOccur(Integer minNumOccur)

getMinNumOccur

public Integer getMinNumOccur()

getTrainingMode

public Boolean getTrainingMode()

setTrainingMode

public void setTrainingMode(Boolean trainingMode)

setPhrasesToExtract

public void setPhrasesToExtract(Integer phrasesToExtract)

getPhrasesToExtract

public Integer getPhrasesToExtract()