You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@ctakes.apache.org by tm...@apache.org on 2014/01/28 16:08:54 UTC
svn commit: r1562089 -
/ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/
Author: tmill
Date: Tue Jan 28 15:08:54 2014
New Revision: 1562089
URL: http://svn.apache.org/r1562089
Log:
CTAKES-82: Refactored pipelines, including new pipeline that runs event-time relation annotator with placeholder for event-event relations.
Added:
ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/FullTemporalExtractionPipeline.java (with props)
ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/TemporalExtractionPipeline_ImplBase.java (with props)
Modified:
ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/EventAndTimeExtractionPipeline.java
ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/EventExtractionPipeline.java
ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/TimeExtractionPipeline.java
Modified: ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/EventAndTimeExtractionPipeline.java
URL: http://svn.apache.org/viewvc/ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/EventAndTimeExtractionPipeline.java?rev=1562089&r1=1562088&r2=1562089&view=diff
==============================================================================
--- ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/EventAndTimeExtractionPipeline.java (original)
+++ ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/EventAndTimeExtractionPipeline.java Tue Jan 28 15:08:54 2014
@@ -20,43 +20,16 @@ package org.apache.ctakes.temporal.pipel
import java.io.File;
-import org.apache.ctakes.chunker.ae.Chunker;
-import org.apache.ctakes.chunker.ae.DefaultChunkCreator;
-import org.apache.ctakes.chunker.ae.adjuster.ChunkAdjuster;
-import org.apache.ctakes.constituency.parser.ae.ConstituencyParser;
-import org.apache.ctakes.contexttokenizer.ae.ContextDependentTokenizerAnnotator;
-import org.apache.ctakes.core.ae.OverlapAnnotator;
-import org.apache.ctakes.core.ae.SentenceDetector;
-import org.apache.ctakes.core.ae.SimpleSegmentAnnotator;
-import org.apache.ctakes.core.ae.TokenizerAnnotatorPTB;
import org.apache.ctakes.core.cr.FilesInDirectoryCollectionReader;
-import org.apache.ctakes.core.resource.FileLocator;
-import org.apache.ctakes.core.resource.FileResourceImpl;
-import org.apache.ctakes.core.resource.JdbcConnectionResourceImpl;
-import org.apache.ctakes.core.resource.LuceneIndexReaderResourceImpl;
-import org.apache.ctakes.dependency.parser.ae.ClearNLPDependencyParserAE;
-import org.apache.ctakes.dependency.parser.ae.ClearNLPSemanticRoleLabelerAE;
-import org.apache.ctakes.dictionary.lookup.ae.UmlsDictionaryLookupAnnotator;
-import org.apache.ctakes.lvg.ae.LvgAnnotator;
-import org.apache.ctakes.lvg.resource.LvgCmdApiResourceImpl;
-import org.apache.ctakes.postagger.POSTagger;
+import org.apache.ctakes.temporal.ae.BackwardsTimeAnnotator;
import org.apache.ctakes.temporal.ae.EventAnnotator;
-import org.apache.ctakes.temporal.ae.TimeAnnotator;
-import org.apache.ctakes.temporal.eval.Evaluation_ImplBase.CopyNPChunksToLookupWindowAnnotations;
-import org.apache.ctakes.typesystem.type.syntax.BaseToken;
-import org.apache.ctakes.typesystem.type.textspan.LookupWindowAnnotation;
-import org.apache.ctakes.typesystem.type.textspan.Segment;
-import org.apache.ctakes.typesystem.type.textspan.Sentence;
+import org.apache.ctakes.temporal.ae.EventTimeRelationAnnotator;
import org.apache.uima.analysis_engine.AnalysisEngine;
-import org.apache.uima.analysis_engine.AnalysisEngineDescription;
import org.apache.uima.collection.CollectionReader;
import org.uimafit.component.xwriter.XWriter;
import org.uimafit.factory.AggregateBuilder;
import org.uimafit.factory.AnalysisEngineFactory;
import org.uimafit.factory.CollectionReaderFactory;
-import org.uimafit.factory.ExternalResourceFactory;
-import org.uimafit.factory.TypePrioritiesFactory;
-import org.uimafit.factory.TypeSystemDescriptionFactory;
import org.uimafit.pipeline.SimplePipeline;
import com.lexicalscope.jewel.cli.CliFactory;
@@ -68,32 +41,25 @@ import com.lexicalscope.jewel.cli.Option
*
* @author dmitriy dligach
*/
-public class EventAndTimeExtractionPipeline {
+public class EventAndTimeExtractionPipeline extends TemporalExtractionPipeline_ImplBase {
- static interface Options {
- // FYI: Command should say -i or --inputDirectory now
- @Option(shortName="i",
- description = "specify the path to the directory containing the clinical notes to be processed")
- public String getInputDirectory();
-
- @Option(shortName = "o",
- description = "specify the path to the directory where the output xmi files are to be saved")
- public String getOutputDirectory();
-
+ static interface ETOptions extends Options{
@Option(
+ shortName = "e",
description = "specify the path to the directory where the trained event model is located",
defaultValue="target/eval/event-spans/train_and_test/")
public String getEventModelDirectory();
@Option(
+ shortName = "t",
description = "specify the path to the directory where the trained event model is located",
- defaultValue="target/eval/time-spans/train_and_test/seq/")
+ defaultValue="target/eval/time-spans/train_and_test/BackwardsTimeAnnotator/")
public String getTimeModelDirectory();
}
public static void main(String[] args) throws Exception {
- Options options = CliFactory.parseArguments(Options.class, args);
+ ETOptions options = CliFactory.parseArguments(ETOptions.class, args);
CollectionReader collectionReader = CollectionReaderFactory.createCollectionReaderFromPath(
"../ctakes-core/desc/collection_reader/FilesInDirectoryCollectionReader.xml",
@@ -102,8 +68,7 @@ public class EventAndTimeExtractionPipel
AggregateBuilder aggregateBuilder = getPreprocessorAggregateBuilder();
aggregateBuilder.add(EventAnnotator.createAnnotatorDescription(new File(options.getEventModelDirectory())));
- aggregateBuilder.add(TimeAnnotator.createAnnotatorDescription(new File(options.getTimeModelDirectory())));
-
+ aggregateBuilder.add(BackwardsTimeAnnotator.createAnnotatorDescription(new File(options.getTimeModelDirectory())));
AnalysisEngine xWriter = AnalysisEngineFactory.createPrimitive(
XWriter.class,
XWriter.PARAM_OUTPUT_DIRECTORY_NAME,
@@ -114,184 +79,4 @@ public class EventAndTimeExtractionPipel
aggregateBuilder.createAggregate(),
xWriter);
}
-
- /**
- * Preprocessing needed for relation extraction.
- */
- protected static AggregateBuilder getPreprocessorAggregateBuilder()
- throws Exception {
- AggregateBuilder aggregateBuilder = new AggregateBuilder();
-
- // identify segments; use simple segment annotator on non-mayo notes
- // aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(SegmentsFromBracketedSectionTagsAnnotator.class));
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(SimpleSegmentAnnotator.class));
-
- // identify sentences
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- SentenceDetector.class,
- SentenceDetector.SD_MODEL_FILE_PARAM,
- "org/apache/ctakes/core/sentdetect/sd-med-model.zip"));
- // identify tokens
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(TokenizerAnnotatorPTB.class));
- // merge some tokens
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(ContextDependentTokenizerAnnotator.class));
-
- // identify part-of-speech tags
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- POSTagger.class,
- TypeSystemDescriptionFactory.createTypeSystemDescription(),
- TypePrioritiesFactory.createTypePriorities(Segment.class, Sentence.class, BaseToken.class),
- POSTagger.POS_MODEL_FILE_PARAM,
- "org/apache/ctakes/postagger/models/mayo-pos.zip"));
-
- // identify chunks
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- Chunker.class,
- Chunker.CHUNKER_MODEL_FILE_PARAM,
- FileLocator.locateFile("org/apache/ctakes/chunker/models/chunker-model.zip"),
- Chunker.CHUNKER_CREATOR_CLASS_PARAM,
- DefaultChunkCreator.class));
-
- // identify UMLS named entities
-
- // adjust NP in NP NP to span both
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- ChunkAdjuster.class,
- ChunkAdjuster.PARAM_CHUNK_PATTERN,
- new String[] { "NP", "NP" },
- ChunkAdjuster.PARAM_EXTEND_TO_INCLUDE_TOKEN,
- 1));
- // adjust NP in NP PP NP to span all three
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- ChunkAdjuster.class,
- ChunkAdjuster.PARAM_CHUNK_PATTERN,
- new String[] { "NP", "PP", "NP" },
- ChunkAdjuster.PARAM_EXTEND_TO_INCLUDE_TOKEN,
- 2));
- // add lookup windows for each NP
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(CopyNPChunksToLookupWindowAnnotations.class));
- // maximize lookup windows
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- OverlapAnnotator.class,
- "A_ObjectClass",
- LookupWindowAnnotation.class,
- "B_ObjectClass",
- LookupWindowAnnotation.class,
- "OverlapType",
- "A_ENV_B",
- "ActionType",
- "DELETE",
- "DeleteAction",
- new String[] { "selector=B" }));
- // add UMLS on top of lookup windows
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- UmlsDictionaryLookupAnnotator.class,
- "ctakes.umlsaddr",
- "https://uts-ws.nlm.nih.gov/restful/isValidUMLSUser",
- "ctakes.umlsvendor",
- "NLM-6515182895",
- "LookupDescriptor",
- ExternalResourceFactory.createExternalResourceDescription(
- FileResourceImpl.class,
- new File("target/unpacked/org/apache/ctakes/dictionary/lookup/LookupDesc_Db.xml").getAbsoluteFile()),
- "DbConnection",
- ExternalResourceFactory.createExternalResourceDescription(
- JdbcConnectionResourceImpl.class,
- "",
- JdbcConnectionResourceImpl.PARAM_DRIVER_CLASS,
- "org.hsqldb.jdbcDriver",
- JdbcConnectionResourceImpl.PARAM_URL,
- // Should be the following but it's WAY too slow
- // "jdbc:hsqldb:res:/org/apache/ctakes/dictionary/lookup/umls2011ab/umls"),
- "jdbc:hsqldb:file:target/unpacked/org/apache/ctakes/dictionary/lookup/umls2011ab/umls"),
- "RxnormIndexReader",
- ExternalResourceFactory.createExternalResourceDescription(
- LuceneIndexReaderResourceImpl.class,
- "",
- "UseMemoryIndex",
- true,
- "IndexDirectory",
- new File("target/unpacked/org/apache/ctakes/dictionary/lookup/rxnorm_index").getAbsoluteFile()),
- "OrangeBookIndexReader",
- ExternalResourceFactory.createExternalResourceDescription(
- LuceneIndexReaderResourceImpl.class,
- "",
- "UseMemoryIndex",
- true,
- "IndexDirectory",
- new File("target/unpacked/org/apache/ctakes/dictionary/lookup/OrangeBook").getAbsoluteFile())));
-
- // add lvg annotator
- String[] XeroxTreebankMap = {
- "adj|JJ",
- "adv|RB",
- "aux|AUX",
- "compl|CS",
- "conj|CC",
- "det|DET",
- "modal|MD",
- "noun|NN",
- "prep|IN",
- "pron|PRP",
- "verb|VB" };
- String[] ExclusionSet = {
- "and",
- "And",
- "by",
- "By",
- "for",
- "For",
- "in",
- "In",
- "of",
- "Of",
- "on",
- "On",
- "the",
- "The",
- "to",
- "To",
- "with",
- "With" };
- AnalysisEngineDescription lvgAnnotator = AnalysisEngineFactory.createPrimitiveDescription(
- LvgAnnotator.class,
- "UseSegments",
- false,
- "SegmentsToSkip",
- new String[0],
- "UseCmdCache",
- false,
- "CmdCacheFileLocation",
- "/org/apache/ctakes/lvg/2005_norm.voc",
- "CmdCacheFrequencyCutoff",
- 20,
- "ExclusionSet",
- ExclusionSet,
- "XeroxTreebankMap",
- XeroxTreebankMap,
- "LemmaCacheFileLocation",
- "/org/apache/ctakes/lvg/2005_lemma.voc",
- "UseLemmaCache",
- false,
- "LemmaCacheFrequencyCutoff",
- 20,
- "PostLemmas",
- true,
- "LvgCmdApi",
- ExternalResourceFactory.createExternalResourceDescription(
- LvgCmdApiResourceImpl.class,
- new File(LvgCmdApiResourceImpl.class.getResource(
- "/org/apache/ctakes/lvg/data/config/lvg.properties").toURI())));
- aggregateBuilder.add(lvgAnnotator);
-
- // add dependency parser
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(ClearNLPDependencyParserAE.class));
-
- // add semantic role labeler
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(ClearNLPSemanticRoleLabelerAE.class));
-
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(ConstituencyParser.class));
-
- return aggregateBuilder;
- }
}
Modified: ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/EventExtractionPipeline.java
URL: http://svn.apache.org/viewvc/ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/EventExtractionPipeline.java?rev=1562089&r1=1562088&r2=1562089&view=diff
==============================================================================
--- ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/EventExtractionPipeline.java (original)
+++ ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/EventExtractionPipeline.java Tue Jan 28 15:08:54 2014
@@ -20,42 +20,14 @@ package org.apache.ctakes.temporal.pipel
import java.io.File;
-import org.apache.ctakes.chunker.ae.Chunker;
-import org.apache.ctakes.chunker.ae.DefaultChunkCreator;
-import org.apache.ctakes.chunker.ae.adjuster.ChunkAdjuster;
-import org.apache.ctakes.constituency.parser.ae.ConstituencyParser;
-import org.apache.ctakes.contexttokenizer.ae.ContextDependentTokenizerAnnotator;
-import org.apache.ctakes.core.ae.OverlapAnnotator;
-import org.apache.ctakes.core.ae.SentenceDetector;
-import org.apache.ctakes.core.ae.SimpleSegmentAnnotator;
-import org.apache.ctakes.core.ae.TokenizerAnnotatorPTB;
import org.apache.ctakes.core.cr.FilesInDirectoryCollectionReader;
-import org.apache.ctakes.core.resource.FileLocator;
-import org.apache.ctakes.core.resource.FileResourceImpl;
-import org.apache.ctakes.core.resource.JdbcConnectionResourceImpl;
-import org.apache.ctakes.core.resource.LuceneIndexReaderResourceImpl;
-import org.apache.ctakes.dependency.parser.ae.ClearNLPDependencyParserAE;
-import org.apache.ctakes.dependency.parser.ae.ClearNLPSemanticRoleLabelerAE;
-import org.apache.ctakes.dictionary.lookup.ae.UmlsDictionaryLookupAnnotator;
-import org.apache.ctakes.lvg.ae.LvgAnnotator;
-import org.apache.ctakes.lvg.resource.LvgCmdApiResourceImpl;
-import org.apache.ctakes.postagger.POSTagger;
import org.apache.ctakes.temporal.ae.EventAnnotator;
-import org.apache.ctakes.temporal.eval.Evaluation_ImplBase.CopyNPChunksToLookupWindowAnnotations;
-import org.apache.ctakes.typesystem.type.syntax.BaseToken;
-import org.apache.ctakes.typesystem.type.textspan.LookupWindowAnnotation;
-import org.apache.ctakes.typesystem.type.textspan.Segment;
-import org.apache.ctakes.typesystem.type.textspan.Sentence;
import org.apache.uima.analysis_engine.AnalysisEngine;
-import org.apache.uima.analysis_engine.AnalysisEngineDescription;
import org.apache.uima.collection.CollectionReader;
import org.uimafit.component.xwriter.XWriter;
import org.uimafit.factory.AggregateBuilder;
import org.uimafit.factory.AnalysisEngineFactory;
import org.uimafit.factory.CollectionReaderFactory;
-import org.uimafit.factory.ExternalResourceFactory;
-import org.uimafit.factory.TypePrioritiesFactory;
-import org.uimafit.factory.TypeSystemDescriptionFactory;
import org.uimafit.pipeline.SimplePipeline;
import com.lexicalscope.jewel.cli.CliFactory;
@@ -67,20 +39,9 @@ import com.lexicalscope.jewel.cli.Option
*
* @author dmitriy dligach
*/
-public class EventExtractionPipeline {
+public class EventExtractionPipeline extends TemporalExtractionPipeline_ImplBase {
- static interface Options {
- // FYI -- command line should use -i or --inputDirectory now
- @Option(
- shortName = "i",
- description = "specify the path to the directory containing the clinical notes to be processed")
- public String getInputDirectory();
-
- @Option(
- shortName = "o",
- description = "specify the path to the directory where the output xmi files are to be saved")
- public String getOutputDirectory();
-
+ static interface EventOptions extends Options{
@Option(
shortName = "m",
description = "specify the path to the directory where the trained model is located",
@@ -90,7 +51,7 @@ public class EventExtractionPipeline {
public static void main(String[] args) throws Exception {
- Options options = CliFactory.parseArguments(Options.class, args);
+ EventOptions options = CliFactory.parseArguments(EventOptions.class, args);
CollectionReader collectionReader = CollectionReaderFactory.createCollectionReaderFromPath(
"../ctakes-core/desc/collection_reader/FilesInDirectoryCollectionReader.xml",
@@ -110,185 +71,4 @@ public class EventExtractionPipeline {
aggregateBuilder.createAggregate(),
xWriter);
}
-
- /**
- * Preprocessing needed for relation extraction.
- */
- protected static AggregateBuilder getPreprocessorAggregateBuilder()
- throws Exception {
- AggregateBuilder aggregateBuilder = new AggregateBuilder();
-
- // identify segments; use simple segment annotator on non-mayo notes
- // aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(SegmentsFromBracketedSectionTagsAnnotator.class));
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(SimpleSegmentAnnotator.class));
-
- // identify sentences
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- SentenceDetector.class,
- SentenceDetector.SD_MODEL_FILE_PARAM,
- "org/apache/ctakes/core/sentdetect/sd-med-model.zip"));
-
- // identify tokens
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(TokenizerAnnotatorPTB.class));
- // merge some tokens
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(ContextDependentTokenizerAnnotator.class));
-
- // identify part-of-speech tags
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- POSTagger.class,
- TypeSystemDescriptionFactory.createTypeSystemDescription(),
- TypePrioritiesFactory.createTypePriorities(Segment.class, Sentence.class, BaseToken.class),
- POSTagger.POS_MODEL_FILE_PARAM,
- "org/apache/ctakes/postagger/models/mayo-pos.zip"));
-
- // identify chunks
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- Chunker.class,
- Chunker.CHUNKER_MODEL_FILE_PARAM,
- FileLocator.locateFile("org/apache/ctakes/chunker/models/chunker-model.zip"),
- Chunker.CHUNKER_CREATOR_CLASS_PARAM,
- DefaultChunkCreator.class));
-
- // identify UMLS named entities
-
- // adjust NP in NP NP to span both
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- ChunkAdjuster.class,
- ChunkAdjuster.PARAM_CHUNK_PATTERN,
- new String[] { "NP", "NP" },
- ChunkAdjuster.PARAM_EXTEND_TO_INCLUDE_TOKEN,
- 1));
- // adjust NP in NP PP NP to span all three
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- ChunkAdjuster.class,
- ChunkAdjuster.PARAM_CHUNK_PATTERN,
- new String[] { "NP", "PP", "NP" },
- ChunkAdjuster.PARAM_EXTEND_TO_INCLUDE_TOKEN,
- 2));
- // add lookup windows for each NP
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(CopyNPChunksToLookupWindowAnnotations.class));
- // maximize lookup windows
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- OverlapAnnotator.class,
- "A_ObjectClass",
- LookupWindowAnnotation.class,
- "B_ObjectClass",
- LookupWindowAnnotation.class,
- "OverlapType",
- "A_ENV_B",
- "ActionType",
- "DELETE",
- "DeleteAction",
- new String[] { "selector=B" }));
- // add UMLS on top of lookup windows
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- UmlsDictionaryLookupAnnotator.class,
- "ctakes.umlsaddr",
- "https://uts-ws.nlm.nih.gov/restful/isValidUMLSUser",
- "ctakes.umlsvendor",
- "NLM-6515182895",
- "LookupDescriptor",
- ExternalResourceFactory.createExternalResourceDescription(
- FileResourceImpl.class,
- new File("target/unpacked/org/apache/ctakes/dictionary/lookup/LookupDesc_Db.xml").getAbsoluteFile()),
- "DbConnection",
- ExternalResourceFactory.createExternalResourceDescription(
- JdbcConnectionResourceImpl.class,
- "",
- JdbcConnectionResourceImpl.PARAM_DRIVER_CLASS,
- "org.hsqldb.jdbcDriver",
- JdbcConnectionResourceImpl.PARAM_URL,
- // Should be the following but it's WAY too slow
- // "jdbc:hsqldb:res:/org/apache/ctakes/dictionary/lookup/umls2011ab/umls"),
- "jdbc:hsqldb:file:target/unpacked/org/apache/ctakes/dictionary/lookup/umls2011ab/umls"),
- "RxnormIndexReader",
- ExternalResourceFactory.createExternalResourceDescription(
- LuceneIndexReaderResourceImpl.class,
- "",
- "UseMemoryIndex",
- true,
- "IndexDirectory",
- new File("target/unpacked/org/apache/ctakes/dictionary/lookup/rxnorm_index").getAbsoluteFile()),
- "OrangeBookIndexReader",
- ExternalResourceFactory.createExternalResourceDescription(
- LuceneIndexReaderResourceImpl.class,
- "",
- "UseMemoryIndex",
- true,
- "IndexDirectory",
- new File("target/unpacked/org/apache/ctakes/dictionary/lookup/OrangeBook").getAbsoluteFile())));
-
- // add lvg annotator
- String[] XeroxTreebankMap = {
- "adj|JJ",
- "adv|RB",
- "aux|AUX",
- "compl|CS",
- "conj|CC",
- "det|DET",
- "modal|MD",
- "noun|NN",
- "prep|IN",
- "pron|PRP",
- "verb|VB" };
- String[] ExclusionSet = {
- "and",
- "And",
- "by",
- "By",
- "for",
- "For",
- "in",
- "In",
- "of",
- "Of",
- "on",
- "On",
- "the",
- "The",
- "to",
- "To",
- "with",
- "With" };
- AnalysisEngineDescription lvgAnnotator = AnalysisEngineFactory.createPrimitiveDescription(
- LvgAnnotator.class,
- "UseSegments",
- false,
- "SegmentsToSkip",
- new String[0],
- "UseCmdCache",
- false,
- "CmdCacheFileLocation",
- "/org/apache/ctakes/lvg/2005_norm.voc",
- "CmdCacheFrequencyCutoff",
- 20,
- "ExclusionSet",
- ExclusionSet,
- "XeroxTreebankMap",
- XeroxTreebankMap,
- "LemmaCacheFileLocation",
- "/org/apache/ctakes/lvg/2005_lemma.voc",
- "UseLemmaCache",
- false,
- "LemmaCacheFrequencyCutoff",
- 20,
- "PostLemmas",
- true,
- "LvgCmdApi",
- ExternalResourceFactory.createExternalResourceDescription(
- LvgCmdApiResourceImpl.class,
- new File(LvgCmdApiResourceImpl.class.getResource(
- "/org/apache/ctakes/lvg/data/config/lvg.properties").toURI())));
- aggregateBuilder.add(lvgAnnotator);
-
- // add dependency parser
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(ClearNLPDependencyParserAE.class));
-
- // add semantic role labeler
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(ClearNLPSemanticRoleLabelerAE.class));
-
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(ConstituencyParser.class));
-
- return aggregateBuilder;
- }
}
Added: ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/FullTemporalExtractionPipeline.java
URL: http://svn.apache.org/viewvc/ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/FullTemporalExtractionPipeline.java?rev=1562089&view=auto
==============================================================================
--- ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/FullTemporalExtractionPipeline.java (added)
+++ ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/FullTemporalExtractionPipeline.java Tue Jan 28 15:08:54 2014
@@ -0,0 +1,80 @@
+package org.apache.ctakes.temporal.pipelines;
+
+import java.io.File;
+
+import org.apache.ctakes.core.cr.FilesInDirectoryCollectionReader;
+import org.apache.ctakes.temporal.ae.BackwardsTimeAnnotator;
+import org.apache.ctakes.temporal.ae.EventAnnotator;
+import org.apache.ctakes.temporal.ae.EventEventRelationAnnotator;
+import org.apache.ctakes.temporal.ae.EventTimeRelationAnnotator;
+import org.apache.uima.analysis_engine.AnalysisEngine;
+import org.apache.uima.collection.CollectionReader;
+import org.uimafit.component.xwriter.XWriter;
+import org.uimafit.factory.AggregateBuilder;
+import org.uimafit.factory.AnalysisEngineFactory;
+import org.uimafit.factory.CollectionReaderFactory;
+import org.uimafit.pipeline.SimplePipeline;
+
+import com.lexicalscope.jewel.cli.CliFactory;
+import com.lexicalscope.jewel.cli.Option;
+
+public class FullTemporalExtractionPipeline extends
+ TemporalExtractionPipeline_ImplBase {
+
+ static interface FullOptions extends Options {
+ @Option(
+ shortName = "e",
+ description = "specify the path to the directory where the trained event model is located",
+ defaultValue="target/eval/event-spans/train_and_test/")
+ public String getEventModelDirectory();
+
+ @Option(
+ shortName = "t",
+ description = "specify the path to the directory where the trained event model is located",
+ defaultValue="target/eval/time-spans/train_and_test/BackwardsTimeAnnotator/")
+ public String getTimeModelDirectory();
+
+ @Option(
+ shortName = "r",
+ description = "Specify the path to the directory where the trained event-time relation model is located",
+ defaultValue="target/eval/temporal-relations/event-time/train_and_test/")
+ public String getEventTimeRelationModelDirectory();
+
+ @Option(
+ shortName = "s",
+ description = "Specify the path to the directory where the trained event-event relation model is located",
+ defaultToNull=true) // add in default value once we have a satisfying trained model
+ public String getEventEventRelationModelDirectory();
+ }
+
+ /**
+ * @param args
+ * @throws Exception
+ */
+ public static void main(String[] args) throws Exception {
+ FullOptions options = CliFactory.parseArguments(FullOptions.class, args);
+
+ CollectionReader collectionReader = CollectionReaderFactory.createCollectionReaderFromPath(
+ "../ctakes-core/desc/collection_reader/FilesInDirectoryCollectionReader.xml",
+ FilesInDirectoryCollectionReader.PARAM_INPUTDIR,
+ options.getInputDirectory());
+
+ AggregateBuilder aggregateBuilder = getPreprocessorAggregateBuilder();
+ aggregateBuilder.add(EventAnnotator.createAnnotatorDescription(new File(options.getEventModelDirectory())));
+ aggregateBuilder.add(BackwardsTimeAnnotator.createAnnotatorDescription(new File(options.getTimeModelDirectory())));
+ aggregateBuilder.add(EventTimeRelationAnnotator.createAnnotatorDescription(new File(options.getEventTimeRelationModelDirectory())));
+ if(options.getEventEventRelationModelDirectory()!=null){
+ aggregateBuilder.add(EventEventRelationAnnotator.createAnnotatorDescription(new File(options.getEventEventRelationModelDirectory())));
+ }
+ AnalysisEngine xWriter = AnalysisEngineFactory.createPrimitive(
+ XWriter.class,
+ XWriter.PARAM_OUTPUT_DIRECTORY_NAME,
+ options.getOutputDirectory());
+
+ SimplePipeline.runPipeline(
+ collectionReader,
+ aggregateBuilder.createAggregate(),
+ xWriter);
+ }
+
+}
Propchange: ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/FullTemporalExtractionPipeline.java
------------------------------------------------------------------------------
svn:mime-type = text/plain
Added: ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/TemporalExtractionPipeline_ImplBase.java
URL: http://svn.apache.org/viewvc/ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/TemporalExtractionPipeline_ImplBase.java?rev=1562089&view=auto
==============================================================================
--- ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/TemporalExtractionPipeline_ImplBase.java (added)
+++ ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/TemporalExtractionPipeline_ImplBase.java Tue Jan 28 15:08:54 2014
@@ -0,0 +1,231 @@
+package org.apache.ctakes.temporal.pipelines;
+
+import java.io.File;
+
+import org.apache.ctakes.chunker.ae.Chunker;
+import org.apache.ctakes.chunker.ae.DefaultChunkCreator;
+import org.apache.ctakes.chunker.ae.adjuster.ChunkAdjuster;
+import org.apache.ctakes.constituency.parser.ae.ConstituencyParser;
+import org.apache.ctakes.contexttokenizer.ae.ContextDependentTokenizerAnnotator;
+import org.apache.ctakes.core.ae.OverlapAnnotator;
+import org.apache.ctakes.core.ae.SentenceDetector;
+import org.apache.ctakes.core.ae.SimpleSegmentAnnotator;
+import org.apache.ctakes.core.ae.TokenizerAnnotatorPTB;
+import org.apache.ctakes.core.resource.FileLocator;
+import org.apache.ctakes.core.resource.FileResourceImpl;
+import org.apache.ctakes.core.resource.JdbcConnectionResourceImpl;
+import org.apache.ctakes.core.resource.LuceneIndexReaderResourceImpl;
+import org.apache.ctakes.dependency.parser.ae.ClearNLPDependencyParserAE;
+import org.apache.ctakes.dependency.parser.ae.ClearNLPSemanticRoleLabelerAE;
+import org.apache.ctakes.dictionary.lookup.ae.UmlsDictionaryLookupAnnotator;
+import org.apache.ctakes.lvg.ae.LvgAnnotator;
+import org.apache.ctakes.lvg.resource.LvgCmdApiResourceImpl;
+import org.apache.ctakes.postagger.POSTagger;
+import org.apache.ctakes.temporal.eval.Evaluation_ImplBase.CopyNPChunksToLookupWindowAnnotations;
+import org.apache.ctakes.typesystem.type.syntax.BaseToken;
+import org.apache.ctakes.typesystem.type.textspan.LookupWindowAnnotation;
+import org.apache.ctakes.typesystem.type.textspan.Segment;
+import org.apache.ctakes.typesystem.type.textspan.Sentence;
+import org.apache.uima.analysis_engine.AnalysisEngineDescription;
+import org.uimafit.factory.AggregateBuilder;
+import org.uimafit.factory.AnalysisEngineFactory;
+import org.uimafit.factory.ExternalResourceFactory;
+import org.uimafit.factory.TypePrioritiesFactory;
+import org.uimafit.factory.TypeSystemDescriptionFactory;
+
+import com.lexicalscope.jewel.cli.Option;
+
+public abstract class TemporalExtractionPipeline_ImplBase {
+ static interface Options {
+
+ @Option(
+ shortName = "i",
+ description = "specify the path to the directory containing the clinical notes to be processed")
+ public String getInputDirectory();
+
+ @Option(
+ shortName = "o",
+ description = "specify the path to the directory where the output xmi files are to be saved")
+ public String getOutputDirectory();
+ }
+
+ /**
+ * Preprocessing needed for relation extraction.
+ */
+ protected static AggregateBuilder getPreprocessorAggregateBuilder()
+ throws Exception {
+ AggregateBuilder aggregateBuilder = new AggregateBuilder();
+
+ // identify segments; use simple segment annotator on non-mayo notes
+ // aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(SegmentsFromBracketedSectionTagsAnnotator.class));
+ aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(SimpleSegmentAnnotator.class));
+
+ // identify sentences
+ aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
+ SentenceDetector.class,
+ SentenceDetector.SD_MODEL_FILE_PARAM,
+ "org/apache/ctakes/core/sentdetect/sd-med-model.zip"));
+ // identify tokens
+ aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(TokenizerAnnotatorPTB.class));
+ // merge some tokens
+ aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(ContextDependentTokenizerAnnotator.class));
+
+ // identify part-of-speech tags
+ aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
+ POSTagger.class,
+ TypeSystemDescriptionFactory.createTypeSystemDescription(),
+ TypePrioritiesFactory.createTypePriorities(Segment.class, Sentence.class, BaseToken.class),
+ POSTagger.POS_MODEL_FILE_PARAM,
+ "org/apache/ctakes/postagger/models/mayo-pos.zip"));
+
+ // identify chunks
+ aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
+ Chunker.class,
+ Chunker.CHUNKER_MODEL_FILE_PARAM,
+ FileLocator.locateFile("org/apache/ctakes/chunker/models/chunker-model.zip"),
+ Chunker.CHUNKER_CREATOR_CLASS_PARAM,
+ DefaultChunkCreator.class));
+
+ // identify UMLS named entities
+
+ // adjust NP in NP NP to span both
+ aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
+ ChunkAdjuster.class,
+ ChunkAdjuster.PARAM_CHUNK_PATTERN,
+ new String[] { "NP", "NP" },
+ ChunkAdjuster.PARAM_EXTEND_TO_INCLUDE_TOKEN,
+ 1));
+ // adjust NP in NP PP NP to span all three
+ aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
+ ChunkAdjuster.class,
+ ChunkAdjuster.PARAM_CHUNK_PATTERN,
+ new String[] { "NP", "PP", "NP" },
+ ChunkAdjuster.PARAM_EXTEND_TO_INCLUDE_TOKEN,
+ 2));
+ // add lookup windows for each NP
+ aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(CopyNPChunksToLookupWindowAnnotations.class));
+ // maximize lookup windows
+ aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
+ OverlapAnnotator.class,
+ "A_ObjectClass",
+ LookupWindowAnnotation.class,
+ "B_ObjectClass",
+ LookupWindowAnnotation.class,
+ "OverlapType",
+ "A_ENV_B",
+ "ActionType",
+ "DELETE",
+ "DeleteAction",
+ new String[] { "selector=B" }));
+ // add UMLS on top of lookup windows
+ aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
+ UmlsDictionaryLookupAnnotator.class,
+ "ctakes.umlsaddr",
+ "https://uts-ws.nlm.nih.gov/restful/isValidUMLSUser",
+ "ctakes.umlsvendor",
+ "NLM-6515182895",
+ "LookupDescriptor",
+ ExternalResourceFactory.createExternalResourceDescription(
+ FileResourceImpl.class,
+ FileLocator.locateFile("org/apache/ctakes/dictionary/lookup/LookupDesc_Db.xml")),
+ "DbConnection",
+ ExternalResourceFactory.createExternalResourceDescription(
+ JdbcConnectionResourceImpl.class,
+ "",
+ JdbcConnectionResourceImpl.PARAM_DRIVER_CLASS,
+ "org.hsqldb.jdbcDriver",
+ JdbcConnectionResourceImpl.PARAM_URL,
+ // Should be the following but it's WAY too slow
+ // "jdbc:hsqldb:res:/org/apache/ctakes/dictionary/lookup/umls2011ab/umls"),
+ "jdbc:hsqldb:file:target/unpacked/org/apache/ctakes/dictionary/lookup/umls2011ab/umls"),
+ "RxnormIndexReader",
+ ExternalResourceFactory.createExternalResourceDescription(
+ LuceneIndexReaderResourceImpl.class,
+ "",
+ "UseMemoryIndex",
+ true,
+ "IndexDirectory",
+ new File("target/unpacked/org/apache/ctakes/dictionary/lookup/rxnorm_index").getAbsoluteFile()),
+ "OrangeBookIndexReader",
+ ExternalResourceFactory.createExternalResourceDescription(
+ LuceneIndexReaderResourceImpl.class,
+ "",
+ "UseMemoryIndex",
+ true,
+ "IndexDirectory",
+ "org/apache/ctakes/dictionary/lookup/OrangeBook")));
+
+ // add lvg annotator
+ String[] XeroxTreebankMap = {
+ "adj|JJ",
+ "adv|RB",
+ "aux|AUX",
+ "compl|CS",
+ "conj|CC",
+ "det|DET",
+ "modal|MD",
+ "noun|NN",
+ "prep|IN",
+ "pron|PRP",
+ "verb|VB" };
+ String[] ExclusionSet = {
+ "and",
+ "And",
+ "by",
+ "By",
+ "for",
+ "For",
+ "in",
+ "In",
+ "of",
+ "Of",
+ "on",
+ "On",
+ "the",
+ "The",
+ "to",
+ "To",
+ "with",
+ "With" };
+ AnalysisEngineDescription lvgAnnotator = AnalysisEngineFactory.createPrimitiveDescription(
+ LvgAnnotator.class,
+ "UseSegments",
+ false,
+ "SegmentsToSkip",
+ new String[0],
+ "UseCmdCache",
+ false,
+ "CmdCacheFileLocation",
+ "/org/apache/ctakes/lvg/2005_norm.voc",
+ "CmdCacheFrequencyCutoff",
+ 20,
+ "ExclusionSet",
+ ExclusionSet,
+ "XeroxTreebankMap",
+ XeroxTreebankMap,
+ "LemmaCacheFileLocation",
+ "/org/apache/ctakes/lvg/2005_lemma.voc",
+ "UseLemmaCache",
+ false,
+ "LemmaCacheFrequencyCutoff",
+ 20,
+ "PostLemmas",
+ true,
+ "LvgCmdApi",
+ ExternalResourceFactory.createExternalResourceDescription(
+ LvgCmdApiResourceImpl.class,
+ new File(LvgCmdApiResourceImpl.class.getResource(
+ "/org/apache/ctakes/lvg/data/config/lvg.properties").toURI())));
+ aggregateBuilder.add(lvgAnnotator);
+
+ // add dependency parser
+ aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(ClearNLPDependencyParserAE.class));
+
+ // add semantic role labeler
+ aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(ClearNLPSemanticRoleLabelerAE.class));
+
+ aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(ConstituencyParser.class));
+
+ return aggregateBuilder;
+ }
+}
Propchange: ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/TemporalExtractionPipeline_ImplBase.java
------------------------------------------------------------------------------
svn:mime-type = text/plain
Modified: ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/TimeExtractionPipeline.java
URL: http://svn.apache.org/viewvc/ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/TimeExtractionPipeline.java?rev=1562089&r1=1562088&r2=1562089&view=diff
==============================================================================
--- ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/TimeExtractionPipeline.java (original)
+++ ctakes/trunk/ctakes-temporal/src/main/java/org/apache/ctakes/temporal/pipelines/TimeExtractionPipeline.java Tue Jan 28 15:08:54 2014
@@ -20,42 +20,14 @@ package org.apache.ctakes.temporal.pipel
import java.io.File;
-import org.apache.ctakes.chunker.ae.Chunker;
-import org.apache.ctakes.chunker.ae.DefaultChunkCreator;
-import org.apache.ctakes.chunker.ae.adjuster.ChunkAdjuster;
-import org.apache.ctakes.constituency.parser.ae.ConstituencyParser;
-import org.apache.ctakes.contexttokenizer.ae.ContextDependentTokenizerAnnotator;
-import org.apache.ctakes.core.ae.OverlapAnnotator;
-import org.apache.ctakes.core.ae.SentenceDetector;
-import org.apache.ctakes.core.ae.SimpleSegmentAnnotator;
-import org.apache.ctakes.core.ae.TokenizerAnnotatorPTB;
import org.apache.ctakes.core.cr.FilesInDirectoryCollectionReader;
-import org.apache.ctakes.core.resource.FileLocator;
-import org.apache.ctakes.core.resource.FileResourceImpl;
-import org.apache.ctakes.core.resource.JdbcConnectionResourceImpl;
-import org.apache.ctakes.core.resource.LuceneIndexReaderResourceImpl;
-import org.apache.ctakes.dependency.parser.ae.ClearNLPDependencyParserAE;
-import org.apache.ctakes.dependency.parser.ae.ClearNLPSemanticRoleLabelerAE;
-import org.apache.ctakes.dictionary.lookup.ae.UmlsDictionaryLookupAnnotator;
-import org.apache.ctakes.lvg.ae.LvgAnnotator;
-import org.apache.ctakes.lvg.resource.LvgCmdApiResourceImpl;
-import org.apache.ctakes.postagger.POSTagger;
import org.apache.ctakes.temporal.ae.BackwardsTimeAnnotator;
-import org.apache.ctakes.temporal.eval.Evaluation_ImplBase.CopyNPChunksToLookupWindowAnnotations;
-import org.apache.ctakes.typesystem.type.syntax.BaseToken;
-import org.apache.ctakes.typesystem.type.textspan.LookupWindowAnnotation;
-import org.apache.ctakes.typesystem.type.textspan.Segment;
-import org.apache.ctakes.typesystem.type.textspan.Sentence;
import org.apache.uima.analysis_engine.AnalysisEngine;
-import org.apache.uima.analysis_engine.AnalysisEngineDescription;
import org.apache.uima.collection.CollectionReader;
import org.uimafit.component.xwriter.XWriter;
import org.uimafit.factory.AggregateBuilder;
import org.uimafit.factory.AnalysisEngineFactory;
import org.uimafit.factory.CollectionReaderFactory;
-import org.uimafit.factory.ExternalResourceFactory;
-import org.uimafit.factory.TypePrioritiesFactory;
-import org.uimafit.factory.TypeSystemDescriptionFactory;
import org.uimafit.pipeline.SimplePipeline;
import com.lexicalscope.jewel.cli.CliFactory;
@@ -67,21 +39,11 @@ import com.lexicalscope.jewel.cli.Option
*
* @author dmitriy dligach
*/
-public class TimeExtractionPipeline {
+public class TimeExtractionPipeline extends TemporalExtractionPipeline_ImplBase {
- static interface Options {
+ static interface TimexOptions extends Options {
@Option(
- shortName = "i",
- description = "specify the path to the directory containing the clinical notes to be processed")
- public String getInputDirectory();
-
- @Option(
- shortName = "o",
- description = "specify the path to the directory where the output xmi files are to be saved")
- public String getOutputDirectory();
-
- @Option(
shortName = "m",
description = "specify the path to the directory where the temporal expression model is located",
defaultValue="target/eval/time-spans/train_and_test/BackwardsTimeAnnotator/")
@@ -90,7 +52,7 @@ public class TimeExtractionPipeline {
public static void main(String[] args) throws Exception {
- Options options = CliFactory.parseArguments(Options.class, args);
+ TimexOptions options = CliFactory.parseArguments(TimexOptions.class, args);
CollectionReader collectionReader = CollectionReaderFactory.createCollectionReaderFromPath(
"../ctakes-core/desc/collection_reader/FilesInDirectoryCollectionReader.xml",
@@ -110,184 +72,4 @@ public class TimeExtractionPipeline {
aggregateBuilder.createAggregate(),
xWriter);
}
-
- /**
- * Preprocessing needed for relation extraction.
- */
- protected static AggregateBuilder getPreprocessorAggregateBuilder()
- throws Exception {
- AggregateBuilder aggregateBuilder = new AggregateBuilder();
-
- // identify segments; use simple segment annotator on non-mayo notes
- // aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(SegmentsFromBracketedSectionTagsAnnotator.class));
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(SimpleSegmentAnnotator.class));
-
- // identify sentences
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- SentenceDetector.class,
- SentenceDetector.SD_MODEL_FILE_PARAM,
- "org/apache/ctakes/core/sentdetect/sd-med-model.zip"));
- // identify tokens
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(TokenizerAnnotatorPTB.class));
- // merge some tokens
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(ContextDependentTokenizerAnnotator.class));
-
- // identify part-of-speech tags
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- POSTagger.class,
- TypeSystemDescriptionFactory.createTypeSystemDescription(),
- TypePrioritiesFactory.createTypePriorities(Segment.class, Sentence.class, BaseToken.class),
- POSTagger.POS_MODEL_FILE_PARAM,
- "org/apache/ctakes/postagger/models/mayo-pos.zip"));
-
- // identify chunks
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- Chunker.class,
- Chunker.CHUNKER_MODEL_FILE_PARAM,
- FileLocator.locateFile("org/apache/ctakes/chunker/models/chunker-model.zip"),
- Chunker.CHUNKER_CREATOR_CLASS_PARAM,
- DefaultChunkCreator.class));
-
- // identify UMLS named entities
-
- // adjust NP in NP NP to span both
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- ChunkAdjuster.class,
- ChunkAdjuster.PARAM_CHUNK_PATTERN,
- new String[] { "NP", "NP" },
- ChunkAdjuster.PARAM_EXTEND_TO_INCLUDE_TOKEN,
- 1));
- // adjust NP in NP PP NP to span all three
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- ChunkAdjuster.class,
- ChunkAdjuster.PARAM_CHUNK_PATTERN,
- new String[] { "NP", "PP", "NP" },
- ChunkAdjuster.PARAM_EXTEND_TO_INCLUDE_TOKEN,
- 2));
- // add lookup windows for each NP
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(CopyNPChunksToLookupWindowAnnotations.class));
- // maximize lookup windows
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- OverlapAnnotator.class,
- "A_ObjectClass",
- LookupWindowAnnotation.class,
- "B_ObjectClass",
- LookupWindowAnnotation.class,
- "OverlapType",
- "A_ENV_B",
- "ActionType",
- "DELETE",
- "DeleteAction",
- new String[] { "selector=B" }));
- // add UMLS on top of lookup windows
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(
- UmlsDictionaryLookupAnnotator.class,
- "ctakes.umlsaddr",
- "https://uts-ws.nlm.nih.gov/restful/isValidUMLSUser",
- "ctakes.umlsvendor",
- "NLM-6515182895",
- "LookupDescriptor",
- ExternalResourceFactory.createExternalResourceDescription(
- FileResourceImpl.class,
- FileLocator.locateFile("org/apache/ctakes/dictionary/lookup/LookupDesc_Db.xml")),
- "DbConnection",
- ExternalResourceFactory.createExternalResourceDescription(
- JdbcConnectionResourceImpl.class,
- "",
- JdbcConnectionResourceImpl.PARAM_DRIVER_CLASS,
- "org.hsqldb.jdbcDriver",
- JdbcConnectionResourceImpl.PARAM_URL,
- // Should be the following but it's WAY too slow
- // "jdbc:hsqldb:res:/org/apache/ctakes/dictionary/lookup/umls2011ab/umls"),
- "jdbc:hsqldb:file:target/unpacked/org/apache/ctakes/dictionary/lookup/umls2011ab/umls"),
- "RxnormIndexReader",
- ExternalResourceFactory.createExternalResourceDescription(
- LuceneIndexReaderResourceImpl.class,
- "",
- "UseMemoryIndex",
- true,
- "IndexDirectory",
- new File("target/unpacked/org/apache/ctakes/dictionary/lookup/rxnorm_index").getAbsoluteFile()),
- "OrangeBookIndexReader",
- ExternalResourceFactory.createExternalResourceDescription(
- LuceneIndexReaderResourceImpl.class,
- "",
- "UseMemoryIndex",
- true,
- "IndexDirectory",
- "org/apache/ctakes/dictionary/lookup/OrangeBook")));
-
- // add lvg annotator
- String[] XeroxTreebankMap = {
- "adj|JJ",
- "adv|RB",
- "aux|AUX",
- "compl|CS",
- "conj|CC",
- "det|DET",
- "modal|MD",
- "noun|NN",
- "prep|IN",
- "pron|PRP",
- "verb|VB" };
- String[] ExclusionSet = {
- "and",
- "And",
- "by",
- "By",
- "for",
- "For",
- "in",
- "In",
- "of",
- "Of",
- "on",
- "On",
- "the",
- "The",
- "to",
- "To",
- "with",
- "With" };
- AnalysisEngineDescription lvgAnnotator = AnalysisEngineFactory.createPrimitiveDescription(
- LvgAnnotator.class,
- "UseSegments",
- false,
- "SegmentsToSkip",
- new String[0],
- "UseCmdCache",
- false,
- "CmdCacheFileLocation",
- "/org/apache/ctakes/lvg/2005_norm.voc",
- "CmdCacheFrequencyCutoff",
- 20,
- "ExclusionSet",
- ExclusionSet,
- "XeroxTreebankMap",
- XeroxTreebankMap,
- "LemmaCacheFileLocation",
- "/org/apache/ctakes/lvg/2005_lemma.voc",
- "UseLemmaCache",
- false,
- "LemmaCacheFrequencyCutoff",
- 20,
- "PostLemmas",
- true,
- "LvgCmdApi",
- ExternalResourceFactory.createExternalResourceDescription(
- LvgCmdApiResourceImpl.class,
- new File(LvgCmdApiResourceImpl.class.getResource(
- "/org/apache/ctakes/lvg/data/config/lvg.properties").toURI())));
- aggregateBuilder.add(lvgAnnotator);
-
- // add dependency parser
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(ClearNLPDependencyParserAE.class));
-
- // add semantic role labeler
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(ClearNLPSemanticRoleLabelerAE.class));
-
- aggregateBuilder.add(AnalysisEngineFactory.createPrimitiveDescription(ConstituencyParser.class));
-
- return aggregateBuilder;
- }
}