Author: suat
Date: Tue Nov 27 10:36:32 2012
New Revision: 1414111
URL: http://svn.apache.org/viewvc?rev=1414111&view=rev
Log:
A missing file for the commit:
STANBOL-781:
Currently, when a ContentItem is submitted to an index which was created using
a custom LDPath, the LDPath is only executed on the recognized entities through
the Entityhub. This aims to gather additional information regarding to the
entities of the ContentItem.
However, we would need to execute custom LDPath programs on additional RDF
parts of ContentItems. Otherwise, it is not possible benefit from the
additional parts. With this commit, SemanticIndexManagerImpl will execute the
LDPath program on the RDF parts of ContentItem through the ContentItemBackend.
Modified:
stanbol/trunk/contenthub/ldpath/src/main/java/org/apache/stanbol/contenthub/ldpath/solr/SemanticIndexManagerImpl.java
stanbol/trunk/contenthub/store/solr/pom.xml
Modified:
stanbol/trunk/contenthub/ldpath/src/main/java/org/apache/stanbol/contenthub/ldpath/solr/SemanticIndexManagerImpl.java
URL:
http://svn.apache.org/viewvc/stanbol/trunk/contenthub/ldpath/src/main/java/org/apache/stanbol/contenthub/ldpath/solr/SemanticIndexManagerImpl.java?rev=1414111&r1=1414110&r2=1414111&view=diff
==============================================================================
---
stanbol/trunk/contenthub/ldpath/src/main/java/org/apache/stanbol/contenthub/ldpath/solr/SemanticIndexManagerImpl.java
(original)
+++
stanbol/trunk/contenthub/ldpath/src/main/java/org/apache/stanbol/contenthub/ldpath/solr/SemanticIndexManagerImpl.java
Tue Nov 27 10:36:32 2012
@@ -19,14 +19,17 @@ package org.apache.stanbol.contenthub.ld
import java.io.File;
import java.io.IOException;
import java.io.Reader;
+import java.io.StringReader;
import java.util.Collection;
import java.util.Collections;
import java.util.HashMap;
import java.util.HashSet;
import java.util.Iterator;
import java.util.Map;
+import java.util.Map.Entry;
import java.util.Set;
+import org.apache.clerezza.rdf.core.Resource;
import org.apache.commons.io.FileUtils;
import org.apache.commons.io.IOUtils;
import org.apache.felix.scr.annotations.Activate;
@@ -39,6 +42,9 @@ import org.apache.stanbol.contenthub.ser
import org.apache.stanbol.contenthub.servicesapi.ldpath.SemanticIndexManager;
import org.apache.stanbol.contenthub.servicesapi.store.StoreException;
import org.apache.stanbol.contenthub.store.solr.manager.SolrCoreManager;
+import org.apache.stanbol.enhancer.ldpath.EnhancerLDPath;
+import org.apache.stanbol.enhancer.ldpath.backend.ContentItemBackend;
+import org.apache.stanbol.enhancer.servicesapi.ContentItem;
import org.apache.stanbol.entityhub.core.model.InMemoryValueFactory;
import org.apache.stanbol.entityhub.ldpath.EntityhubLDPath;
import org.apache.stanbol.entityhub.ldpath.backend.SiteManagerBackend;
@@ -50,6 +56,7 @@ import org.osgi.service.component.Compon
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
+import at.newmedialab.ldpath.LDPath;
import at.newmedialab.ldpath.exception.LDPathParseException;
import at.newmedialab.ldpath.model.programs.Program;
@@ -136,7 +143,7 @@ public class SemanticIndexManagerImpl im
// Checks whether there is already a program with the same name
if (nameProgramMap.containsKey(programName)) {
- String msg = String.format("There is already an LDProgram with
name : " + programName);
+ String msg = String.format("There is already an LDProgram with
name : %s", programName);
logger.error(msg);
throw new LDPathException(msg);
}
@@ -193,23 +200,23 @@ public class SemanticIndexManagerImpl im
return nameProgramMap.get(programName);
}
-
+
@Override
public Program<Object> getParsedProgramByName(String programName) {
SiteManagerBackend backend = new
SiteManagerBackend(referencedSiteManager);
String ldPathProgram = getProgramByName(programName);
ValueFactory vf = InMemoryValueFactory.getInstance();
EntityhubLDPath ldPath = new EntityhubLDPath(backend, vf);
- Program<Object> program = null;
+ Program<Object> program = null;
try {
program =
ldPath.parseProgram(LDPathUtils.constructReader(ldPathProgram));
} catch (LDPathParseException e) {
- String msg = "Should never happen!!!!! Cannot parse the already
stored LDPath program.";
+ String msg = "Should never happen!!!!! Cannot parse the already
stored LDPath program.";
logger.error(msg, e);
} catch (LDPathException e) {
- String msg = "Should never happen!!!!! Cannot parse the already
stored LDPath program.";
+ String msg = "Should never happen!!!!! Cannot parse the already
stored LDPath program.";
logger.error(msg, e);
- }
+ }
return program;
}
@@ -239,7 +246,7 @@ public class SemanticIndexManagerImpl im
}
@Override
- public Map<String,Collection<?>> executeProgram(String programName,
Set<String> contexts) throws LDPathException {
+ public Map<String,Collection<?>> executeProgram(String programName,
Set<String> contexts, ContentItem ci) throws LDPathException {
Map<String,Collection<?>> results = new
HashMap<String,Collection<?>>();
SiteManagerBackend backend = new
SiteManagerBackend(referencedSiteManager);
String ldPathProgram = getProgramByName(programName);
@@ -260,6 +267,7 @@ public class SemanticIndexManagerImpl im
while (fieldNames.hasNext()) {
String fieldName = fieldNames.next();
Iterator<Object> valueIterator = representation.get(fieldName);
+ if (!valueIterator.hasNext()) continue;
Set<Object> values = new HashSet<Object>();
while (valueIterator.hasNext()) {
values.add(valueIterator.next());
@@ -276,6 +284,30 @@ public class SemanticIndexManagerImpl im
}
}
}
+
+ // execute the LDPath on the given ContentItem
+ ContentItemBackend contentItemBackend = new ContentItemBackend(ci,
true);
+ LDPath<Resource> resourceLDPath = new
LDPath<Resource>(contentItemBackend, EnhancerLDPath.getConfig());
+ Program<Resource> resourceProgram;
+ try {
+ resourceProgram = resourceLDPath.parseProgram(new
StringReader(ldPathProgram));
+ Map<String,Collection<?>> ciBackendResults =
resourceProgram.execute(contentItemBackend,
+ ci.getUri());
+ for (Entry<String,Collection<?>> result :
ciBackendResults.entrySet()) {
+ if (result.getValue() == null || result.getValue().isEmpty())
continue;
+ if (results.containsKey(result.getKey())) {
+ @SuppressWarnings("unchecked")
+ Collection<Object> resultsValue = (Collection<Object>)
results.get(result.getKey());
+ resultsValue.addAll(result.getValue());
+ } else {
+ results.put(result.getKey(), result.getValue());
+ }
+
+ }
+ } catch (LDPathParseException e) {
+ logger.error("Failed to create Program<Resource> from the LDPath
program", e);
+ }
+
return results;
}
}
Modified: stanbol/trunk/contenthub/store/solr/pom.xml
URL:
http://svn.apache.org/viewvc/stanbol/trunk/contenthub/store/solr/pom.xml?rev=1414111&r1=1414110&r2=1414111&view=diff
==============================================================================
--- stanbol/trunk/contenthub/store/solr/pom.xml (original)
+++ stanbol/trunk/contenthub/store/solr/pom.xml Tue Nov 27 10:36:32 2012
@@ -71,12 +71,6 @@
<artifactId>org.apache.stanbol.contenthub.servicesapi</artifactId>
<version>0.10.0-SNAPSHOT</version>
</dependency>
- <!-- Needed 0.10.0-SNAPSHOT version for InMemoryBlob -->
- <dependency>
- <groupId>org.apache.stanbol</groupId>
- <artifactId>org.apache.stanbol.enhancer.core</artifactId>
- <version>0.10.0-SNAPSHOT</version>
- </dependency>
<dependency>
<groupId>org.apache.stanbol</groupId>
<artifactId>org.apache.stanbol.commons.solr.core</artifactId>
