Revision: 17769
http://sourceforge.net/p/gate/code/17769
Author: ian_roberts
Date: 2014-04-03 15:03:12 +0000 (Thu, 03 Apr 2014)
Log Message:
-----------
Ignore (with a warning) entity annotations that are not covered by a context
annotation, rather than throwing an impenetrable error.
Modified Paths:
--------------
gate/trunk/plugins/Crowd_Sourcing/src/gate/crowdsource/classification/EntityClassificationJobBuilder.java
Modified:
gate/trunk/plugins/Crowd_Sourcing/src/gate/crowdsource/classification/EntityClassificationJobBuilder.java
===================================================================
---
gate/trunk/plugins/Crowd_Sourcing/src/gate/crowdsource/classification/EntityClassificationJobBuilder.java
2014-04-03 14:37:54 UTC (rev 17768)
+++
gate/trunk/plugins/Crowd_Sourcing/src/gate/crowdsource/classification/EntityClassificationJobBuilder.java
2014-04-03 15:03:12 UTC (rev 17769)
@@ -36,10 +36,14 @@
import gate.crowdsource.rest.CrowdFlowerClient;
import gate.gui.ActionsPublisher;
+import org.apache.log4j.Logger;
+
@CreoleResource(name = "Entity Classification Job Builder",
comment = "Build a CrowdFlower job asking users to select the right label
for entities")
public class EntityClassificationJobBuilder extends AbstractLanguageAnalyser
implements ActionsPublisher {
+ private static final Logger log =
Logger.getLogger(EntityClassificationJobBuilder.class);
+
private static final long serialVersionUID = -1584716901194104888L;
private String apiKey;
@@ -148,13 +152,20 @@
fireProgressChanged((100 * entityIdx++) / allEntities.size());
if(isInterrupted()) throw new ExecutionInterruptedException();
AnnotationSet thisEntityContext =
Utils.getCoveringAnnotations(contextAnnotations, entity);
- // get the "closest" context, i.e. the shortest annotation in the
covering set.
- // usually we'd expect this set to contain just one annotation
- Annotation context = Collections.min(thisEntityContext,
ANNOTATION_LENGTH_COMPARATOR);
- crowdFlowerClient.createClassificationUnit(jobId, getDocument(),
entityASName, context, entity);
+ if(thisEntityContext.isEmpty()) {
+ log.warn(entityAnnotationType + " with ID " + entity.getId() +
+ " at offsets (" + Utils.start(entity) + ":" + Utils.end(entity) +
+ ") in document " + getDocument().getName() +
+ " has no surrounding " + contextAnnotationType + " - ignored");
+ } else {
+ // get the "closest" context, i.e. the shortest annotation in the
covering set.
+ // usually we'd expect this set to contain just one annotation
+ Annotation context = Collections.min(thisEntityContext,
ANNOTATION_LENGTH_COMPARATOR);
+ crowdFlowerClient.createClassificationUnit(jobId, getDocument(),
entityASName, context, entity);
+ }
}
fireProcessFinished();
- fireStatusChanged(allEntities.size() + " units created");
+ fireStatusChanged(entityIdx + " units created");
} finally {
interrupted = false;
}
This was sent by the SourceForge.net collaborative development platform, the
world's largest Open Source development site.
------------------------------------------------------------------------------
_______________________________________________
GATE-cvs mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/gate-cvs
