You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@ctakes.apache.org by st...@apache.org on 2012/10/23 22:36:09 UTC
svn commit: r1401453 - in
/incubator/ctakes/trunk/ctakes-relation-extractor/src/main/java/org/apache/ctakes/relationextractor:
eval/ pipelines/
Author: stevenbethard
Date: Tue Oct 23 20:36:09 2012
New Revision: 1401453
URL: http://svn.apache.org/viewvc?rev=1401453&view=rev
Log:
Fixes uses of deprecated ClearTK APIs in relation-extractor
Removed:
incubator/ctakes/trunk/ctakes-relation-extractor/src/main/java/org/apache/ctakes/relationextractor/eval/MultiClassLIBSVMDataWriterFactory.java
Modified:
incubator/ctakes/trunk/ctakes-relation-extractor/src/main/java/org/apache/ctakes/relationextractor/eval/ModifierExtractorEvaluation.java
incubator/ctakes/trunk/ctakes-relation-extractor/src/main/java/org/apache/ctakes/relationextractor/eval/RelationExtractorEvaluation.java
incubator/ctakes/trunk/ctakes-relation-extractor/src/main/java/org/apache/ctakes/relationextractor/pipelines/RelationExtractorTrain.java
Modified: incubator/ctakes/trunk/ctakes-relation-extractor/src/main/java/org/apache/ctakes/relationextractor/eval/ModifierExtractorEvaluation.java
URL: http://svn.apache.org/viewvc/incubator/ctakes/trunk/ctakes-relation-extractor/src/main/java/org/apache/ctakes/relationextractor/eval/ModifierExtractorEvaluation.java?rev=1401453&r1=1401452&r2=1401453&view=diff
==============================================================================
--- incubator/ctakes/trunk/ctakes-relation-extractor/src/main/java/org/apache/ctakes/relationextractor/eval/ModifierExtractorEvaluation.java (original)
+++ incubator/ctakes/trunk/ctakes-relation-extractor/src/main/java/org/apache/ctakes/relationextractor/eval/ModifierExtractorEvaluation.java Tue Oct 23 20:36:09 2012
@@ -28,10 +28,11 @@ import org.apache.uima.cas.CASException;
import org.apache.uima.collection.CollectionReader;
import org.apache.uima.jcas.JCas;
import org.apache.uima.resource.ResourceInitializationException;
-import org.cleartk.classifier.CleartkAnnotator;
+import org.cleartk.classifier.jar.DefaultDataWriterFactory;
import org.cleartk.classifier.jar.DirectoryDataWriterFactory;
import org.cleartk.classifier.jar.GenericJarClassifierFactory;
import org.cleartk.classifier.jar.JarClassifierBuilder;
+import org.cleartk.classifier.libsvm.LIBSVMStringOutcomeDataWriter;
import org.cleartk.eval.AnnotationStatistics;
import org.cleartk.eval.Evaluation_ImplBase;
import org.cleartk.util.Options_ImplBase;
@@ -103,8 +104,8 @@ public class ModifierExtractorEvaluation
collectionReader,
AnalysisEngineFactory.createPrimitiveDescription(OnlyGoldModifiers.class),
ModifierExtractorAnnotator.getDescription(
- CleartkAnnotator.PARAM_DATA_WRITER_FACTORY_CLASS_NAME,
- MultiClassLIBSVMDataWriterFactory.class.getName(),
+ DefaultDataWriterFactory.PARAM_DATA_WRITER_CLASS_NAME,
+ LIBSVMStringOutcomeDataWriter.class,
DirectoryDataWriterFactory.PARAM_OUTPUT_DIRECTORY,
directory.getPath()));
JarClassifierBuilder.trainAndPackage(directory, this.trainingArguments);
Modified: incubator/ctakes/trunk/ctakes-relation-extractor/src/main/java/org/apache/ctakes/relationextractor/eval/RelationExtractorEvaluation.java
URL: http://svn.apache.org/viewvc/incubator/ctakes/trunk/ctakes-relation-extractor/src/main/java/org/apache/ctakes/relationextractor/eval/RelationExtractorEvaluation.java?rev=1401453&r1=1401452&r2=1401453&view=diff
==============================================================================
--- incubator/ctakes/trunk/ctakes-relation-extractor/src/main/java/org/apache/ctakes/relationextractor/eval/RelationExtractorEvaluation.java (original)
+++ incubator/ctakes/trunk/ctakes-relation-extractor/src/main/java/org/apache/ctakes/relationextractor/eval/RelationExtractorEvaluation.java Tue Oct 23 20:36:09 2012
@@ -37,11 +37,12 @@ import org.apache.uima.jcas.JCas;
import org.apache.uima.jcas.tcas.Annotation;
import org.apache.uima.util.CasCopier;
import org.apache.uima.util.Level;
-import org.cleartk.classifier.CleartkAnnotator;
-import org.cleartk.classifier.DataWriterFactory;
+import org.cleartk.classifier.DataWriter;
+import org.cleartk.classifier.jar.DefaultDataWriterFactory;
import org.cleartk.classifier.jar.DirectoryDataWriterFactory;
import org.cleartk.classifier.jar.GenericJarClassifierFactory;
import org.cleartk.classifier.jar.JarClassifierBuilder;
+import org.cleartk.classifier.libsvm.LIBSVMStringOutcomeDataWriter;
import org.cleartk.eval.AnnotationStatistics;
import org.cleartk.eval.Evaluation_ImplBase;
import org.cleartk.util.Options_ImplBase;
@@ -118,7 +119,7 @@ public class RelationExtractorEvaluation
: EntityMentionPairRelationExtractorAnnotator.class;
// determine the type of classifier to be trained
- Class<? extends DataWriterFactory<String>> dataWriterFactoryClass = MultiClassLIBSVMDataWriterFactory.class;
+ Class<? extends DataWriter<String>> dataWriterClass = LIBSVMStringOutcomeDataWriter.class;
// define the set of possible training parameters
List<ParameterSettings> possibleParams = options.runDegreeOf
@@ -153,7 +154,7 @@ public class RelationExtractorEvaluation
RelationExtractorEvaluation evaluation = new RelationExtractorEvaluation(
modelsDir,
annotatorClass,
- dataWriterFactoryClass,
+ dataWriterClass,
additionalParameters,
trainingArguments);
@@ -215,8 +216,8 @@ public class RelationExtractorEvaluation
* The directory where models, etc. should be written
* @param classifierAnnotatorClass
* The CleartkAnnotator class that learns a relation extractor model
- * @param dataWriterFactoryClass
- * The DataWriterFactory defining what type of classifier to train
+ * @param dataWriterClass
+ * The DataWriter defining what type of classifier to train
* @param additionalParameters
* Additional parameters that should be supplied when creating the CleartkAnnotator
* @param trainingArguments
@@ -225,19 +226,19 @@ public class RelationExtractorEvaluation
public RelationExtractorEvaluation(
File baseDirectory,
Class<? extends RelationExtractorAnnotator> classifierAnnotatorClass,
- Class<? extends DataWriterFactory<String>> dataWriterFactoryClass,
+ Class<? extends DataWriter<String>> dataWriterClass,
Object[] additionalParameters,
String[] trainingArguments) {
super(baseDirectory);
this.classifierAnnotatorClass = classifierAnnotatorClass;
- this.dataWriterFactoryClass = dataWriterFactoryClass;
+ this.dataWriterClass = dataWriterClass;
this.additionalParameters = additionalParameters;
this.trainingArguments = trainingArguments;
}
private Class<? extends RelationExtractorAnnotator> classifierAnnotatorClass;
- private Class<? extends DataWriterFactory<String>> dataWriterFactoryClass;
+ private Class<? extends DataWriter<String>> dataWriterClass;
private Object[] additionalParameters;
@@ -272,8 +273,8 @@ public class RelationExtractorEvaluation
classifierAnnotator,
RelationExtractorAnnotator.PARAM_GOLD_VIEW_NAME,
RelationExtractorEvaluation.GOLD_VIEW_NAME,
- CleartkAnnotator.PARAM_DATA_WRITER_FACTORY_CLASS_NAME,
- this.dataWriterFactoryClass.getName(),
+ DefaultDataWriterFactory.PARAM_DATA_WRITER_CLASS_NAME,
+ this.dataWriterClass,
DirectoryDataWriterFactory.PARAM_OUTPUT_DIRECTORY,
directory.getPath());
builder.add(classifierAnnotator);
Modified: incubator/ctakes/trunk/ctakes-relation-extractor/src/main/java/org/apache/ctakes/relationextractor/pipelines/RelationExtractorTrain.java
URL: http://svn.apache.org/viewvc/incubator/ctakes/trunk/ctakes-relation-extractor/src/main/java/org/apache/ctakes/relationextractor/pipelines/RelationExtractorTrain.java?rev=1401453&r1=1401452&r2=1401453&view=diff
==============================================================================
--- incubator/ctakes/trunk/ctakes-relation-extractor/src/main/java/org/apache/ctakes/relationextractor/pipelines/RelationExtractorTrain.java (original)
+++ incubator/ctakes/trunk/ctakes-relation-extractor/src/main/java/org/apache/ctakes/relationextractor/pipelines/RelationExtractorTrain.java Tue Oct 23 20:36:09 2012
@@ -29,7 +29,6 @@ import org.apache.ctakes.relationextract
import org.apache.ctakes.relationextractor.ae.ModifierExtractorAnnotator;
import org.apache.ctakes.relationextractor.ae.RelationExtractorAnnotator;
import org.apache.ctakes.relationextractor.eval.ModifierExtractorEvaluation;
-import org.apache.ctakes.relationextractor.eval.MultiClassLIBSVMDataWriterFactory;
import org.apache.ctakes.relationextractor.eval.RelationExtractorEvaluation;
import org.apache.ctakes.relationextractor.eval.RelationExtractorEvaluation.ParameterSettings;
import org.apache.uima.UIMAFramework;
@@ -37,8 +36,9 @@ import org.apache.uima.analysis_engine.A
import org.apache.uima.collection.CollectionReader;
import org.apache.uima.util.XMLInputSource;
import org.apache.uima.util.XMLParser;
-import org.cleartk.classifier.DataWriterFactory;
+import org.cleartk.classifier.DataWriter;
import org.cleartk.classifier.jar.GenericJarClassifierFactory;
+import org.cleartk.classifier.libsvm.LIBSVMStringOutcomeDataWriter;
import org.cleartk.util.Options_ImplBase;
import org.kohsuke.args4j.Option;
import org.uimafit.factory.AggregateBuilder;
@@ -73,8 +73,7 @@ public class RelationExtractorTrain {
public static AnalysisEngineDescription trainModifierExtractor(
File modelsDir,
- List<File> trainFiles,
- Class<? extends DataWriterFactory<String>> dataWriterFactoryClass) throws Exception {
+ List<File> trainFiles) throws Exception {
ModifierExtractorEvaluation evaluation = new ModifierExtractorEvaluation(
modelsDir,
@@ -97,7 +96,7 @@ public class RelationExtractorTrain {
File modelsDir,
List<File> trainFiles,
Class<? extends RelationExtractorAnnotator> annotatorClass,
- Class<? extends DataWriterFactory<String>> dataWriterFactoryClass,
+ Class<? extends DataWriter<String>> dataWriterClass,
ParameterSettings params) throws Exception {
// define additional configuration parameters for the annotator
@@ -121,7 +120,7 @@ public class RelationExtractorTrain {
RelationExtractorEvaluation evaluation = new RelationExtractorEvaluation(
modelsDir,
annotatorClass,
- dataWriterFactoryClass,
+ dataWriterClass,
additionalParameters,
trainingArguments);
@@ -162,14 +161,14 @@ public class RelationExtractorTrain {
ParameterSettings emPairParams = new ParameterSettings(false, 0.5f, "linear", 0.05, 1.0);
// For now all three components use MultiClass SVMs for classification
- Class<? extends DataWriterFactory<String>> dataWriterFactoryClass = MultiClassLIBSVMDataWriterFactory.class;
+ Class<? extends DataWriter<String>> dataWriterClass = LIBSVMStringOutcomeDataWriter.class;
// Train and write models
- AnalysisEngineDescription modifierExtractorDesc = trainModifierExtractor(modelsDirModExtractor, trainFiles, dataWriterFactoryClass);
+ AnalysisEngineDescription modifierExtractorDesc = trainModifierExtractor(modelsDirModExtractor, trainFiles);
writeDesc(options.descDir, "ModifierExtractorAnnotator", modifierExtractorDesc);
- AnalysisEngineDescription degreeOfRelationExtractorDesc = trainRelationExtractor(modelsDirDegreeOf, trainFiles, DegreeOfRelationExtractorAnnotator.class, dataWriterFactoryClass, degreeOfParams);
+ AnalysisEngineDescription degreeOfRelationExtractorDesc = trainRelationExtractor(modelsDirDegreeOf, trainFiles, DegreeOfRelationExtractorAnnotator.class, dataWriterClass, degreeOfParams);
writeDesc(options.descDir, "DegreeOfRelationExtractorAnnotator", degreeOfRelationExtractorDesc);
- AnalysisEngineDescription emPairRelationExtractorDesc = trainRelationExtractor(modelsDirEMPair, trainFiles, EntityMentionPairRelationExtractorAnnotator.class, dataWriterFactoryClass, emPairParams);
+ AnalysisEngineDescription emPairRelationExtractorDesc = trainRelationExtractor(modelsDirEMPair, trainFiles, EntityMentionPairRelationExtractorAnnotator.class, dataWriterClass, emPairParams);
writeDesc(options.descDir, "EntityMentionPairRelationExtractorAnnotator", emPairRelationExtractorDesc);
// create the aggregate description