Revision: 14445
http://gate.svn.sourceforge.net/gate/?rev=14445&view=rev
Author: valyt
Date: 2011-10-27 12:26:57 +0000 (Thu, 27 Oct 2011)
Log Message:
-----------
The mimir-core side of supporting the indexing of document metadata.
Modified Paths:
--------------
mimir/trunk/mimir-core/src/gate/mimir/AbstractSemanticAnnotationHelper.java
mimir/trunk/mimir-core/src/gate/mimir/SemanticAnnotationHelper.java
mimir/trunk/mimir-core/src/gate/mimir/index/mg4j/MentionsIndexBuilder.java
mimir/trunk/mimir-core/src/gate/mimir/util/DelegatingSemanticAnnotationHelper.java
Modified:
mimir/trunk/mimir-core/src/gate/mimir/AbstractSemanticAnnotationHelper.java
===================================================================
--- mimir/trunk/mimir-core/src/gate/mimir/AbstractSemanticAnnotationHelper.java
2011-10-27 12:26:19 UTC (rev 14444)
+++ mimir/trunk/mimir-core/src/gate/mimir/AbstractSemanticAnnotationHelper.java
2011-10-27 12:26:57 UTC (rev 14445)
@@ -174,6 +174,8 @@
*/
protected String annotationType;
+ private boolean isDocumentHelper = false;
+
public String getAnnotationType() {
return annotationType;
}
@@ -246,4 +248,20 @@
}
return concat;
}
+
+ @Override
+ public boolean isInDocumentMode() {
+ return isDocumentHelper;
+ }
+
+ /**
+ * Switches this helper in document helper mode.
+ * @return a reference to this helper, after it has been configured to work
+ * as a document helper.
+ */
+ public SemanticAnnotationHelper asDocumentHelper() {
+ isDocumentHelper = true;
+ return this;
+ }
+
}
Modified: mimir/trunk/mimir-core/src/gate/mimir/SemanticAnnotationHelper.java
===================================================================
--- mimir/trunk/mimir-core/src/gate/mimir/SemanticAnnotationHelper.java
2011-10-27 12:26:19 UTC (rev 14444)
+++ mimir/trunk/mimir-core/src/gate/mimir/SemanticAnnotationHelper.java
2011-10-27 12:26:57 UTC (rev 14445)
@@ -30,7 +30,24 @@
import java.util.Map;
/**
- * Interface for classes that convert annotations into semantic metadata.
+ * Interface for classes that convert annotations into semantic metadata. At
+ * indexing time, for each annotation to be indexed, the helper produces a set
+ * of mentions URIs (obtained by calling
+ * {@link #getMentionUris(Annotation, int, Indexer)}).
+ * At search time, given a set of constraints, the helper produces set of
+ * <mention URIs, mention length> pairs (obtained by calling on of the
+ * {@link #getMentions(String, List, QueryEngine)},
+ * {@link #getMentions(String, Map, QueryEngine)} methods).
+ *
+ * Each helper can function in annotation mode (the default) or in document
+ * mode. The current functioning mode can be obtained by calling
+ * {@link #isInDocumentMode()}. When in document mode, the helper should behave
+ * as if a single annotation covering the whole document span is being indexed
+ * (or searched for). In this mode, document features are used instead of
+ * annotation features. This has efficiency advantages over actually creating
a
+ * document-spanning annotation, as the implementations can avoid actually
+ * storing the annotation length in the index (as it is always the same as the
+ * document length).
*/
public interface SemanticAnnotationHelper extends Serializable{
@@ -110,5 +127,16 @@
* operations (such as closing connections to ORDI, etc) on this call.
*/
public void close(QueryEngine qEngine);
+
+ /**
+ * Checks whether this helper is configured to work in document helper mode.
+ * If that's the case, then the indexer will call its
+ * {@link #getMentionUris(Annotation, int, Indexer)} method only once for
+ * each input document, with a <code>null</code> value for the annotation.
+ *
+ * @return Returns <code>true</code> if this helper has been configured to
+ * work in document helper mode.
+ */
+ public boolean isInDocumentMode();
}
Modified:
mimir/trunk/mimir-core/src/gate/mimir/index/mg4j/MentionsIndexBuilder.java
===================================================================
--- mimir/trunk/mimir-core/src/gate/mimir/index/mg4j/MentionsIndexBuilder.java
2011-10-27 12:26:19 UTC (rev 14444)
+++ mimir/trunk/mimir-core/src/gate/mimir/index/mg4j/MentionsIndexBuilder.java
2011-10-27 12:26:57 UTC (rev 14445)
@@ -24,11 +24,14 @@
import gate.Document;
import gate.mimir.SemanticAnnotationHelper;
import gate.mimir.IndexConfig.SemanticIndexerConfig;
+import gate.mimir.index.IndexException;
import gate.mimir.index.Indexer;
import gate.util.OffsetComparator;
import java.util.Arrays;
import java.util.HashMap;
+import java.util.LinkedList;
+import java.util.List;
import java.util.Map;
import java.util.concurrent.BlockingQueue;
@@ -47,8 +50,10 @@
/**
* Helpers for each semantic annotation type.
*/
- protected Map<String, SemanticAnnotationHelper> helpers;
+ protected Map<String, SemanticAnnotationHelper> annotationHelpers;
+ protected List<SemanticAnnotationHelper> documentHelpers;
+
/**
* An {@link OffsetComparator} used to sort the annotations by offset before
* indexing.
@@ -60,10 +65,17 @@
Indexer indexer, String baseName, SemanticIndexerConfig config){
super(inputQueue, outputQueue, indexer, baseName);
//get the helpers
- helpers = new HashMap<String, SemanticAnnotationHelper>(
+ annotationHelpers = new HashMap<String, SemanticAnnotationHelper>(
config.getAnnotationTypes().length);
+ documentHelpers = new LinkedList<SemanticAnnotationHelper>();
for(int i = 0; i < config.getAnnotationTypes().length; i++){
- helpers.put(config.getAnnotationTypes()[i], config.getHelpers()[i]);
+ SemanticAnnotationHelper theHelper = config.getHelpers()[i];
+ if(theHelper.isInDocumentMode()) {
+ documentHelpers.add(theHelper);
+ } else {
+ annotationHelpers.put(config.getAnnotationTypes()[i], theHelper);
+ }
+
}
offsetComparator = new OffsetComparator();
}
@@ -73,18 +85,31 @@
* Inform the helpers that a new document is about to start.
*/
protected void documentStarting(GATEDocument gateDocument) {
- for(SemanticAnnotationHelper aHelper : helpers.values()){
+ for(SemanticAnnotationHelper aHelper : annotationHelpers.values()){
aHelper.documentStart(gateDocument.getDocument());
}
+ for(SemanticAnnotationHelper aHelper : documentHelpers){
+ aHelper.documentStart(gateDocument.getDocument());
+ }
}
/**
* Inform the helpers that we finished the current document.
+ * @throws IndexException
*/
- protected void documentEnding(GATEDocument gateDocument) {
- for(SemanticAnnotationHelper aHelper : helpers.values()){
+ protected void documentEnding(GATEDocument gateDocument) throws
IndexException {
+ for(SemanticAnnotationHelper aHelper : annotationHelpers.values()){
aHelper.documentEnd();
}
+
+ if(!documentHelpers.isEmpty()) {
+ processAnnotation(null, gateDocument);
+ }
+
+ for(SemanticAnnotationHelper aHelper : documentHelpers){
+ aHelper.documentEnd();
+ }
+
}
/**
@@ -103,7 +128,7 @@
if(semAnnSet.size() > 0){
AnnotationSet semAnns = null;
synchronized(semAnnSet) {
- semAnns = semAnnSet.get(helpers.keySet());
+ semAnns = semAnnSet.get(annotationHelpers.keySet());
}
semanticAnnots = semAnns.toArray(new Annotation[semAnns.size()]);
Arrays.sort(semanticAnnots, offsetComparator);
@@ -126,24 +151,29 @@
*/
protected void calculateStartPositionForAnnotation(Annotation ann,
GATEDocument gateDocument) {
- //calculate the term position for the current semantic annotation
- while(tokenPosition < gateDocument.getTokenAnnots().length &&
- gateDocument.getTokenAnnots()[tokenPosition].
- getEndNode().getOffset().longValue() <=
- ann.getStartNode().getOffset().longValue()){
- tokenPosition++;
+ if(ann == null) {
+ // we're supposed index the document metadata
+ tokenPosition = 0;
+ } else {
+ //calculate the term position for the current semantic annotation
+ while(tokenPosition < gateDocument.getTokenAnnots().length &&
+ gateDocument.getTokenAnnots()[tokenPosition].
+ getEndNode().getOffset().longValue() <=
+ ann.getStartNode().getOffset().longValue()){
+ tokenPosition++;
+ }
+ //check if lastTokenposition is valid
+ if(tokenPosition >= gateDocument.getTokenAnnots().length){
+ //malfunction
+ logger.error(
+ "Semantic annotation [Type:" + ann.getType() +
+ ", start: " + ann.getStartNode().getOffset().toString() +
+ ", end: " + ann.getEndNode().getOffset().toString() +
+ "] outside of the tokens area in document" +
+ " URI: " + gateDocument.uri() +
+ " Title: " + gateDocument.title());
+ }
}
- //check if lastTokenposition is valid
- if(tokenPosition >= gateDocument.getTokenAnnots().length){
- //malfunction
- logger.error(
- "Semantic annotation [Type:" + ann.getType() +
- ", start: " + ann.getStartNode().getOffset().toString() +
- ", end: " + ann.getEndNode().getOffset().toString() +
- "] outside of the tokens area in document" +
- " URI: " + gateDocument.uri() +
- " Title: " + gateDocument.title());
- }
}
/**
@@ -155,17 +185,30 @@
*/
protected String[] calculateTermStringForAnnotation(Annotation ann,
GATEDocument gateDocument) {
- //calculate the annotation length (as number of terms)
- SemanticAnnotationHelper helper = helpers.get(ann.getType());
- int length = 1;
- while(tokenPosition + length < gateDocument.getTokenAnnots().length &&
- gateDocument.getTokenAnnots()[tokenPosition + length].
- getStartNode().getOffset().longValue() <
- ann.getEndNode().getOffset().longValue()){
- length++;
+ if(ann == null) {
+ // obtain the URIs to be indexed for the *document* metadata
+ List<String> terms = new LinkedList<String>();
+ for(SemanticAnnotationHelper aHelper : documentHelpers) {
+ String[] someTerms = aHelper.getMentionUris(null, -1, indexer);
+ if(someTerms != null) {
+ for(String aTerm : someTerms) {
+ terms.add(aTerm);
+ }
+ }
}
- //get the annotation URI
- return helper.getMentionUris(ann, length, indexer);
-// currentTerm.replace(helper.getMentionUri(ann, length, indexer));
+ return terms.toArray(new String[terms.size()]);
+ } else {
+ //calculate the annotation length (as number of terms)
+ SemanticAnnotationHelper helper = annotationHelpers.get(ann.getType());
+ int length = 1;
+ while(tokenPosition + length < gateDocument.getTokenAnnots().length &&
+ gateDocument.getTokenAnnots()[tokenPosition + length].
+ getStartNode().getOffset().longValue() <
+ ann.getEndNode().getOffset().longValue()){
+ length++;
+ }
+ //get the annotation URI
+ return helper.getMentionUris(ann, length, indexer);
+ }
}
}
Modified:
mimir/trunk/mimir-core/src/gate/mimir/util/DelegatingSemanticAnnotationHelper.java
===================================================================
---
mimir/trunk/mimir-core/src/gate/mimir/util/DelegatingSemanticAnnotationHelper.java
2011-10-27 12:26:19 UTC (rev 14444)
+++
mimir/trunk/mimir-core/src/gate/mimir/util/DelegatingSemanticAnnotationHelper.java
2011-10-27 12:26:57 UTC (rev 14445)
@@ -110,7 +110,22 @@
delegate.documentEnd();
}
+
@Override
+ public boolean isInDocumentMode() {
+ return delegate.isInDocumentMode();
+ }
+
+
+ @Override
+ public SemanticAnnotationHelper asDocumentHelper() {
+ if(delegate.isInDocumentMode()) return this;
+ else throw new UnsupportedOperationException(
+ "The delegate provided is not configured to work as a document helper.");
+ }
+
+
+ @Override
public void close(Indexer indexer) {
delegate.close(indexer);
}
This was sent by the SourceForge.net collaborative development platform, the
world's largest Open Source development site.
------------------------------------------------------------------------------
The demand for IT networking professionals continues to grow, and the
demand for specialized networking skills is growing even more rapidly.
Take a complimentary Learning@Cisco Self-Assessment and learn
about Cisco certifications, training, and career opportunities.
http://p.sf.net/sfu/cisco-dev2dev
_______________________________________________
GATE-cvs mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/gate-cvs
