You are viewing a plain text version of this content. The canonical link for it is here.
Posted to slide-dev@jakarta.apache.org by un...@apache.org on 2004/09/29 17:28:06 UTC
cvs commit: jakarta-slide/src/stores/org/apache/slide/index TextContentIndexer.java
unico 2004/09/29 08:28:06
Modified: src/share/org/apache/slide/extractor MSWordExtractor.java
AbstractPropertyExtractor.java
ExtractorManager.java AbstractContentExtractor.java
OfficeExtractor.java Extractor.java
MSPowerPointExtractor.java PDFExtractor.java
SimpleXmlExtractor.java
PropertyExtractorTrigger.java MSExcelExtractor.java
src/share/org/apache/slide/common Namespace.java Domain.java
src/stores/org/apache/slide/index TextContentIndexer.java
Log:
per-namespace extractor configuration:
extractors can now be declared for each namespace separately
in addition to the global extractor configuration for the whole domain.
Revision Changes Path
1.2 +6 -6 jakarta-slide/src/share/org/apache/slide/extractor/MSWordExtractor.java
Index: MSWordExtractor.java
===================================================================
RCS file: /home/cvs/jakarta-slide/src/share/org/apache/slide/extractor/MSWordExtractor.java,v
retrieving revision 1.1
retrieving revision 1.2
diff -u -r1.1 -r1.2
--- MSWordExtractor.java 29 Jun 2004 08:10:57 -0000 1.1
+++ MSWordExtractor.java 29 Sep 2004 15:28:06 -0000 1.2
@@ -35,8 +35,8 @@
public class MSWordExtractor extends AbstractContentExtractor {
- public MSWordExtractor(String uri, String contentType) {
- super(uri, contentType);
+ public MSWordExtractor(String uri, String contentType, String namespace) {
+ super(uri, contentType, namespace);
}
public Reader extract(InputStream content) throws ExtractorException {
@@ -57,7 +57,7 @@
{
FileInputStream in = new FileInputStream(args[0]);
- MSWordExtractor ex = new MSWordExtractor(null, null);
+ MSWordExtractor ex = new MSWordExtractor(null, null, null);
Reader reader = ex.extract(in);
1.3 +16 -5 jakarta-slide/src/share/org/apache/slide/extractor/AbstractPropertyExtractor.java
Index: AbstractPropertyExtractor.java
===================================================================
RCS file: /home/cvs/jakarta-slide/src/share/org/apache/slide/extractor/AbstractPropertyExtractor.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -u -r1.2 -r1.3
--- AbstractPropertyExtractor.java 28 Jul 2004 09:36:01 -0000 1.2
+++ AbstractPropertyExtractor.java 29 Sep 2004 15:28:06 -0000 1.3
@@ -28,15 +28,21 @@
/**
* The AbstractPropertyExtractor class
- *
*/
public abstract class AbstractPropertyExtractor implements PropertyExtractor {
+
private String contentType;
private String uri;
+ private String namespace;
public AbstractPropertyExtractor(String uri, String contentType) {
+ this(uri, contentType, null);
+ }
+
+ public AbstractPropertyExtractor(String uri, String contentType, String namespace) {
this.contentType = contentType;
this.uri = uri;
+ this.namespace = namespace;
}
public abstract Map extract(InputStream content) throws ExtractorException;
@@ -48,4 +54,9 @@
public String getUri() {
return uri;
}
-}
\ No newline at end of file
+
+ public String getNamespace() {
+ return namespace;
+ }
+
+}
1.5 +22 -13 jakarta-slide/src/share/org/apache/slide/extractor/ExtractorManager.java
Index: ExtractorManager.java
===================================================================
RCS file: /home/cvs/jakarta-slide/src/share/org/apache/slide/extractor/ExtractorManager.java,v
retrieving revision 1.4
retrieving revision 1.5
diff -u -r1.4 -r1.5
--- ExtractorManager.java 28 Jul 2004 09:36:01 -0000 1.4
+++ ExtractorManager.java 29 Sep 2004 15:28:06 -0000 1.5
@@ -34,7 +34,6 @@
/**
* The ExtractorManager class
- *
*/
public class ExtractorManager implements Configurable {
private final static ExtractorManager manager = new ExtractorManager();
@@ -51,11 +50,11 @@
extractors.add(extractor);
}
- public PropertyExtractor[] getPropertyExtractors(NodeRevisionDescriptors descriptors, NodeRevisionDescriptor descriptor) {
+ public PropertyExtractor[] getPropertyExtractors(String namespace, NodeRevisionDescriptors descriptors, NodeRevisionDescriptor descriptor) {
List matchingExtractors = new ArrayList();
for ( Iterator i = extractors.iterator(); i.hasNext(); ) {
Extractor extractor = (Extractor)i.next();
- if ( extractor instanceof PropertyExtractor && matches(extractor, descriptors, descriptor)) {
+ if ( extractor instanceof PropertyExtractor && matches(extractor, namespace, descriptors, descriptor)) {
matchingExtractors.add(extractor);
}
}
@@ -63,11 +62,11 @@
return (PropertyExtractor [])matchingExtractors.toArray(extractors);
};
- public ContentExtractor[] getContentExtractors(NodeRevisionDescriptors descriptors, NodeRevisionDescriptor descriptor) {
+ public ContentExtractor[] getContentExtractors(String namespace, NodeRevisionDescriptors descriptors, NodeRevisionDescriptor descriptor) {
List matchingExtractors = new ArrayList();
for ( Iterator i = extractors.iterator(); i.hasNext(); ) {
Extractor extractor = (Extractor)i.next();
- if ( extractor instanceof ContentExtractor && matches(extractor, descriptors, descriptor)) {
+ if ( extractor instanceof ContentExtractor && matches(extractor, namespace, descriptors, descriptor)) {
matchingExtractors.add(extractor);
}
}
@@ -75,11 +74,11 @@
return (ContentExtractor [])matchingExtractors.toArray(extractors);
};
- public Extractor[] getExtractors(NodeRevisionDescriptors descriptors, NodeRevisionDescriptor descriptor) {
+ public Extractor[] getExtractors(String namespace, NodeRevisionDescriptors descriptors, NodeRevisionDescriptor descriptor) {
List matchingExtractors = new ArrayList();
for ( Iterator i = extractors.iterator(); i.hasNext(); ) {
Extractor extractor = (Extractor)i.next();
- if ( matches(extractor, descriptors, descriptor)) {
+ if ( matches(extractor, namespace, descriptors, descriptor)) {
matchingExtractors.add(extractor);
}
}
@@ -87,7 +86,7 @@
return (Extractor [])matchingExtractors.toArray(extractors);
};
- public boolean matches(Extractor extractor, NodeRevisionDescriptors descriptors, NodeRevisionDescriptor descriptor) {
+ public boolean matches(Extractor extractor, String namespace, NodeRevisionDescriptors descriptors, NodeRevisionDescriptor descriptor) {
boolean matching = true;
if ( descriptor != null && extractor.getContentType() != null && !descriptor.getContentType().equals(extractor.getContentType()) ) {
matching = false;
@@ -95,6 +94,9 @@
if ( descriptors != null && extractor.getUri() != null && !descriptors.getUri().startsWith(extractor.getUri()) ) {
matching = false;
}
+ if ( descriptors != null && extractor.getNamespace() != null && !extractor.getNamespace().equals(namespace)) {
+ matching = false;
+ }
return matching;
}
@@ -105,11 +107,18 @@
String classname = extractorConfig.getAttribute("classname");
String uri = extractorConfig.getAttribute("uri", null);
String contentType = extractorConfig.getAttribute("content-type", null);
+ String namespace = extractorConfig.getAttribute("namespace", null);
try {
Class extractorClass = Class.forName(classname);
Extractor extractor = null;
- Constructor extractorConstructor = extractorClass.getConstructor(new Class[] { String.class, String.class } );
- extractor = (Extractor)extractorConstructor.newInstance(new String[] { uri, contentType });
+ try {
+ Constructor extractorConstructor = extractorClass.getConstructor(new Class[] { String.class, String.class, String.class } );
+ extractor = (Extractor) extractorConstructor.newInstance(new String[] { uri, contentType, namespace });
+ }
+ catch (NoSuchMethodException e) {
+ Constructor extractorConstructor = extractorClass.getConstructor(new Class[] { String.class, String.class } );
+ extractor = (Extractor) extractorConstructor.newInstance(new String[] { uri, contentType });
+ }
if ( extractor instanceof Configurable ) {
((Configurable)extractor).configure(extractorConfig.getConfiguration("configuration"));
}
1.4 +16 -5 jakarta-slide/src/share/org/apache/slide/extractor/AbstractContentExtractor.java
Index: AbstractContentExtractor.java
===================================================================
RCS file: /home/cvs/jakarta-slide/src/share/org/apache/slide/extractor/AbstractContentExtractor.java,v
retrieving revision 1.3
retrieving revision 1.4
diff -u -r1.3 -r1.4
--- AbstractContentExtractor.java 28 Jul 2004 09:36:01 -0000 1.3
+++ AbstractContentExtractor.java 29 Sep 2004 15:28:06 -0000 1.4
@@ -28,15 +28,21 @@
/**
* The AbstractPropertyExtractor class
- *
*/
public abstract class AbstractContentExtractor implements ContentExtractor {
+
private String contentType;
private String uri;
+ private String namespace;
public AbstractContentExtractor(String uri, String contentType) {
+ this(uri, contentType, null);
+ }
+
+ public AbstractContentExtractor(String uri, String contentType, String namespace) {
this.contentType = contentType;
this.uri = uri;
+ this.namespace = namespace;
}
public abstract Reader extract(InputStream content) throws ExtractorException;
@@ -48,4 +54,9 @@
public String getUri() {
return uri;
}
-}
\ No newline at end of file
+
+ public String getNamespace() {
+ return namespace;
+ }
+
+}
1.3 +83 -83 jakarta-slide/src/share/org/apache/slide/extractor/OfficeExtractor.java
Index: OfficeExtractor.java
===================================================================
RCS file: /home/cvs/jakarta-slide/src/share/org/apache/slide/extractor/OfficeExtractor.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -u -r1.2 -r1.3
--- OfficeExtractor.java 28 Jul 2004 09:36:01 -0000 1.2
+++ OfficeExtractor.java 29 Sep 2004 15:28:06 -0000 1.3
@@ -1,84 +1,84 @@
-package org.apache.slide.extractor;
-
-import java.io.InputStream;
-import java.util.*;
-
-import org.apache.poi.hpsf.*;
-import org.apache.poi.poifs.eventfilesystem.*;
-import org.apache.slide.util.conf.Configurable;
-import org.apache.slide.util.conf.Configuration;
-import org.apache.slide.util.conf.ConfigurationException;
-
-/**
- * The OfficeExtractor class
- *
- */
-public class OfficeExtractor extends AbstractPropertyExtractor implements Configurable {
- protected List instructions = new ArrayList();
- protected Map propertyMap = new HashMap();
-
- public OfficeExtractor(String uri, String contentType) {
- super(uri, contentType);
- }
-
- public Map extract(InputStream content) throws ExtractorException {
- OfficePropertiesListener listener = new OfficePropertiesListener();
- try {
- POIFSReader r = new POIFSReader();
- r.registerListener(listener);
- r.read(content);
- } catch (Exception e) {
- throw new ExtractorException("Exception while extracting properties in OfficeExtractor");
- }
- return listener.getProperties();
- }
-
- class OfficePropertiesListener implements POIFSReaderListener {
-
- private HashMap properties = new HashMap();
-
- public Map getProperties() {
- return properties;
- }
-
- public void processPOIFSReaderEvent(POIFSReaderEvent event) {
- PropertySet ps = null;
- try {
- ps = PropertySetFactory.create(event.getStream());
- } catch (NoPropertySetStreamException ex) {
- return;
- } catch (Exception ex) {
- throw new RuntimeException("Property set stream \"" + event.getPath() + event.getName() + "\": " + ex);
- }
- String eventName = event.getName().trim();
- final long sectionCount = ps.getSectionCount();
- List sections = ps.getSections();
- int nr = 0;
- for (Iterator i = sections.iterator(); i.hasNext();) {
- Section sec = (Section) i.next();
- int propertyCount = sec.getPropertyCount();
- Property[] props = sec.getProperties();
- for (int i2 = 0; i2 < props.length; i2++) {
- Property p = props[i2];
- int id = p.getID();
- long type = p.getType();
- Object value = p.getValue();
- String key = eventName + "-" + nr + "-" + id;
- if ( propertyMap.containsKey(key) ) {
- properties.put(propertyMap.get(key), value);
- }
- }
- }
- }
- }
-
- public void configure(Configuration configuration) throws ConfigurationException {
- Enumeration instructions = configuration.getConfigurations("instruction");
- while (instructions.hasMoreElements()) {
- Configuration extract = (Configuration)instructions.nextElement();
- String property = extract.getAttribute("property");
- String id = extract.getAttribute("id");
- propertyMap.put(id, property);
- }
- }
+package org.apache.slide.extractor;
+
+import java.io.InputStream;
+import java.util.*;
+
+import org.apache.poi.hpsf.*;
+import org.apache.poi.poifs.eventfilesystem.*;
+import org.apache.slide.util.conf.Configurable;
+import org.apache.slide.util.conf.Configuration;
+import org.apache.slide.util.conf.ConfigurationException;
+
+/**
+ * The OfficeExtractor class
+ *
+ */
+public class OfficeExtractor extends AbstractPropertyExtractor implements Configurable {
+ protected List instructions = new ArrayList();
+ protected Map propertyMap = new HashMap();
+
+ public OfficeExtractor(String uri, String contentType, String namespace) {
+ super(uri, contentType, namespace);
+ }
+
+ public Map extract(InputStream content) throws ExtractorException {
+ OfficePropertiesListener listener = new OfficePropertiesListener();
+ try {
+ POIFSReader r = new POIFSReader();
+ r.registerListener(listener);
+ r.read(content);
+ } catch (Exception e) {
+ throw new ExtractorException("Exception while extracting properties in OfficeExtractor");
+ }
+ return listener.getProperties();
+ }
+
+ class OfficePropertiesListener implements POIFSReaderListener {
+
+ private HashMap properties = new HashMap();
+
+ public Map getProperties() {
+ return properties;
+ }
+
+ public void processPOIFSReaderEvent(POIFSReaderEvent event) {
+ PropertySet ps = null;
+ try {
+ ps = PropertySetFactory.create(event.getStream());
+ } catch (NoPropertySetStreamException ex) {
+ return;
+ } catch (Exception ex) {
+ throw new RuntimeException("Property set stream \"" + event.getPath() + event.getName() + "\": " + ex);
+ }
+ String eventName = event.getName().trim();
+ final long sectionCount = ps.getSectionCount();
+ List sections = ps.getSections();
+ int nr = 0;
+ for (Iterator i = sections.iterator(); i.hasNext();) {
+ Section sec = (Section) i.next();
+ int propertyCount = sec.getPropertyCount();
+ Property[] props = sec.getProperties();
+ for (int i2 = 0; i2 < props.length; i2++) {
+ Property p = props[i2];
+ int id = p.getID();
+ long type = p.getType();
+ Object value = p.getValue();
+ String key = eventName + "-" + nr + "-" + id;
+ if ( propertyMap.containsKey(key) ) {
+ properties.put(propertyMap.get(key), value);
+ }
+ }
+ }
+ }
+ }
+
+ public void configure(Configuration configuration) throws ConfigurationException {
+ Enumeration instructions = configuration.getConfigurations("instruction");
+ while (instructions.hasMoreElements()) {
+ Configuration extract = (Configuration)instructions.nextElement();
+ String property = extract.getAttribute("property");
+ String id = extract.getAttribute("id");
+ propertyMap.put(id, property);
+ }
+ }
}
1.5 +7 -3 jakarta-slide/src/share/org/apache/slide/extractor/Extractor.java
Index: Extractor.java
===================================================================
RCS file: /home/cvs/jakarta-slide/src/share/org/apache/slide/extractor/Extractor.java,v
retrieving revision 1.4
retrieving revision 1.5
diff -u -r1.4 -r1.5
--- Extractor.java 28 Jul 2004 09:36:01 -0000 1.4
+++ Extractor.java 29 Sep 2004 15:28:06 -0000 1.5
@@ -28,7 +28,11 @@
*
*/
public interface Extractor {
+
public String getContentType();
public String getUri();
+
+ public String getNamespace();
+
}
1.3 +6 -6 jakarta-slide/src/share/org/apache/slide/extractor/MSPowerPointExtractor.java
Index: MSPowerPointExtractor.java
===================================================================
RCS file: /home/cvs/jakarta-slide/src/share/org/apache/slide/extractor/MSPowerPointExtractor.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -u -r1.2 -r1.3
--- MSPowerPointExtractor.java 14 Jul 2004 08:01:31 -0000 1.2
+++ MSPowerPointExtractor.java 29 Sep 2004 15:28:06 -0000 1.3
@@ -40,8 +40,8 @@
{
private ByteArrayOutputStream writer = new ByteArrayOutputStream();
- public MSPowerPointExtractor(String uri, String contentType) {
- super(uri, contentType);
+ public MSPowerPointExtractor(String uri, String contentType, String namespace) {
+ super(uri, contentType, namespace);
}
public Reader extract(InputStream content) throws ExtractorException {
@@ -91,7 +91,7 @@
{
FileInputStream in = new FileInputStream(args[0]);
- MSPowerPointExtractor ex = new MSPowerPointExtractor(null, null);
+ MSPowerPointExtractor ex = new MSPowerPointExtractor(null, null, null);
Reader reader = ex.extract(in);
1.2 +6 -6 jakarta-slide/src/share/org/apache/slide/extractor/PDFExtractor.java
Index: PDFExtractor.java
===================================================================
RCS file: /home/cvs/jakarta-slide/src/share/org/apache/slide/extractor/PDFExtractor.java,v
retrieving revision 1.1
retrieving revision 1.2
diff -u -r1.1 -r1.2
--- PDFExtractor.java 29 Jun 2004 08:10:57 -0000 1.1
+++ PDFExtractor.java 29 Sep 2004 15:28:06 -0000 1.2
@@ -37,9 +37,9 @@
public class PDFExtractor extends AbstractContentExtractor
{
- public PDFExtractor(String uri, String contentType)
+ public PDFExtractor(String uri, String contentType, String namespace)
{
- super(uri, contentType);
+ super(uri, contentType, namespace);
}
public Reader extract(InputStream content) throws ExtractorException
@@ -72,7 +72,7 @@
{
FileInputStream in = new FileInputStream(args[0]);
- PDFExtractor ex = new PDFExtractor(null, null);
+ PDFExtractor ex = new PDFExtractor(null, null, null);
Reader reader = ex.extract(in);
1.8 +6 -5 jakarta-slide/src/share/org/apache/slide/extractor/SimpleXmlExtractor.java
Index: SimpleXmlExtractor.java
===================================================================
RCS file: /home/cvs/jakarta-slide/src/share/org/apache/slide/extractor/SimpleXmlExtractor.java,v
retrieving revision 1.7
retrieving revision 1.8
diff -u -r1.7 -r1.8
--- SimpleXmlExtractor.java 28 Jul 2004 09:36:01 -0000 1.7
+++ SimpleXmlExtractor.java 29 Sep 2004 15:28:06 -0000 1.8
@@ -42,10 +42,11 @@
*
*/
public class SimpleXmlExtractor extends AbstractPropertyExtractor implements Configurable {
+
protected List instructions = new ArrayList();
- public SimpleXmlExtractor(String uri, String contentType) {
- super(uri, contentType);
+ public SimpleXmlExtractor(String uri, String contentType, String namespace) {
+ super(uri, contentType, namespace);
}
public Map extract(InputStream content) throws ExtractorException {
1.6 +6 -5 jakarta-slide/src/share/org/apache/slide/extractor/PropertyExtractorTrigger.java
Index: PropertyExtractorTrigger.java
===================================================================
RCS file: /home/cvs/jakarta-slide/src/share/org/apache/slide/extractor/PropertyExtractorTrigger.java,v
retrieving revision 1.5
retrieving revision 1.6
diff -u -r1.5 -r1.6
--- PropertyExtractorTrigger.java 28 Jul 2004 09:36:01 -0000 1.5
+++ PropertyExtractorTrigger.java 29 Sep 2004 15:28:06 -0000 1.6
@@ -45,11 +45,12 @@
public void store(ContentEvent event) throws VetoException {
try {
+ String namespaceName = event.getNamespace() == null ? null : event.getNamespace().getName();
NodeRevisionDescriptors descriptors = event.getRevisionDescriptors();
NodeRevisionDescriptor descriptor = event.getRevisionDescriptor();
NodeRevisionContent content = event.getRevisionContent();
if ( content != null && descriptor != null ) {
- PropertyExtractor[] extractor = ExtractorManager.getInstance().getPropertyExtractors(descriptors, descriptor);
+ PropertyExtractor[] extractor = ExtractorManager.getInstance().getPropertyExtractors(namespaceName, descriptors, descriptor);
for ( int i = 0; i < extractor.length; i++ ) {
Map extractedProperties = extractor[i].extract(new ByteArrayInputStream(content.getContentBytes()));
for ( Iterator j = extractedProperties.entrySet().iterator(); j.hasNext(); ) {
1.2 +6 -6 jakarta-slide/src/share/org/apache/slide/extractor/MSExcelExtractor.java
Index: MSExcelExtractor.java
===================================================================
RCS file: /home/cvs/jakarta-slide/src/share/org/apache/slide/extractor/MSExcelExtractor.java,v
retrieving revision 1.1
retrieving revision 1.2
diff -u -r1.1 -r1.2
--- MSExcelExtractor.java 29 Jun 2004 08:10:57 -0000 1.1
+++ MSExcelExtractor.java 29 Sep 2004 15:28:06 -0000 1.2
@@ -40,8 +40,8 @@
public class MSExcelExtractor extends AbstractContentExtractor
{
- public MSExcelExtractor(String uri, String contentType) {
- super(uri, contentType);
+ public MSExcelExtractor(String uri, String contentType, String namespace) {
+ super(uri, contentType, namespace);
}
public Reader extract(InputStream content) throws ExtractorException
@@ -92,7 +92,7 @@
{
FileInputStream in = new FileInputStream(args[0]);
- MSExcelExtractor ex = new MSExcelExtractor(null, null);
+ MSExcelExtractor ex = new MSExcelExtractor(null, null, null);
Reader reader = ex.extract(in);
1.68 +41 -4 jakarta-slide/src/share/org/apache/slide/common/Namespace.java
Index: Namespace.java
===================================================================
RCS file: /home/cvs/jakarta-slide/src/share/org/apache/slide/common/Namespace.java,v
retrieving revision 1.67
retrieving revision 1.68
diff -u -r1.67 -r1.68
--- Namespace.java 23 Sep 2004 10:46:56 -0000 1.67
+++ Namespace.java 29 Sep 2004 15:28:06 -0000 1.68
@@ -23,15 +23,20 @@
package org.apache.slide.common;
+import java.lang.reflect.Constructor;
import java.lang.reflect.Method;
import java.util.Enumeration;
import java.util.Hashtable;
import java.util.Vector;
+
import javax.transaction.Status;
import javax.transaction.SystemException;
import javax.transaction.TransactionManager;
+
import org.apache.slide.authenticate.CredentialsToken;
import org.apache.slide.content.ContentInterceptor;
+import org.apache.slide.extractor.Extractor;
+import org.apache.slide.extractor.ExtractorManager;
import org.apache.slide.store.ContentStore;
import org.apache.slide.store.DefaultIndexer;
import org.apache.slide.store.IndexStore;
@@ -45,6 +50,7 @@
import org.apache.slide.structure.ObjectAlreadyExistsException;
import org.apache.slide.structure.SubjectNode;
import org.apache.slide.transaction.SlideTransactionManager;
+import org.apache.slide.util.conf.Configurable;
import org.apache.slide.util.conf.Configuration;
import org.apache.slide.util.conf.ConfigurationException;
import org.apache.slide.util.logger.Logger;
@@ -883,6 +889,37 @@
}
+
+ void loadExtractors(Configuration namespaceExtractorsDefinition)
+ throws SlideException {
+
+ getLogger().log("Loading namespace " + getName() + " extractors",LOG_CHANNEL,Logger.INFO);
+
+ Enumeration extractorConfigs = namespaceExtractorsDefinition.getConfigurations("extractor");
+ while (extractorConfigs.hasMoreElements()) {
+ Configuration extractorConfig = (Configuration) extractorConfigs.nextElement();
+ String classname = extractorConfig.getAttribute("classname");
+ String uri = extractorConfig.getAttribute("uri", null);
+ String contentType = extractorConfig.getAttribute("content-type", null);
+ String namespace = getName();
+ try {
+ Class extractorClass = Class.forName(classname);
+ Extractor extractor = null;
+ Constructor extractorConstructor = extractorClass.getConstructor(new Class[] { String.class, String.class, String.class } );
+ extractor = (Extractor)extractorConstructor.newInstance(new String[] { uri, contentType, namespace });
+ if ( extractor instanceof Configurable ) {
+ ((Configurable)extractor).configure(extractorConfig.getConfiguration("configuration"));
+ }
+ ExtractorManager.getInstance().addExtractor(extractor);
+ } catch (ClassCastException e) {
+ throw new ConfigurationException("Extractor '"+classname+"' is not of type Extractor", namespaceExtractorsDefinition);
+ } catch (ConfigurationException e) {
+ throw e;
+ } catch (Exception e) {
+ throw new ConfigurationException("Extractor '"+classname+"' could not be loaded", namespaceExtractorsDefinition);
+ }
+ }
+ }
// -------------------------------------------------------- Private Methods
1.49 +13 -8 jakarta-slide/src/share/org/apache/slide/common/Domain.java
Index: Domain.java
===================================================================
RCS file: /home/cvs/jakarta-slide/src/share/org/apache/slide/common/Domain.java,v
retrieving revision 1.48
retrieving revision 1.49
diff -u -r1.48 -r1.49
--- Domain.java 28 Jul 2004 09:38:20 -0000 1.48
+++ Domain.java 29 Sep 2004 15:28:06 -0000 1.49
@@ -848,10 +848,15 @@
}
namespace.loadConfiguration(namespaceConfigurationDefinition);
-
-
-
-
+
+ try {
+ Configuration namespaceExtractorsDefinition =
+ configuration.getConfiguration("extractors");
+ namespace.loadExtractors(namespaceExtractorsDefinition);
+ } catch (ConfigurationException e) {
+ // ignore
+ }
+
// preparation to add services, please ignore now
try {
Configuration services = configuration.getConfiguration("services");
1.7 +5 -5 jakarta-slide/src/stores/org/apache/slide/index/TextContentIndexer.java
Index: TextContentIndexer.java
===================================================================
RCS file: /home/cvs/jakarta-slide/src/stores/org/apache/slide/index/TextContentIndexer.java,v
retrieving revision 1.6
retrieving revision 1.7
diff -u -r1.6 -r1.7
--- TextContentIndexer.java 13 Sep 2004 16:52:23 -0000 1.6
+++ TextContentIndexer.java 29 Sep 2004 15:28:06 -0000 1.7
@@ -142,7 +142,7 @@
doc.add(Field.Text(CONTENT_TEXT, readContent(revisionDescriptor, revisionContent)));
if ( revisionContent != null && revisionDescriptor != null ) {
- ContentExtractor[] extractor = ExtractorManager.getInstance().getContentExtractors(null, revisionDescriptor);
+ ContentExtractor[] extractor = ExtractorManager.getInstance().getContentExtractors(uri.getNamespace().getName(), null, revisionDescriptor);
for ( int i = 0; i < extractor.length; i++ ) {
Reader reader = extractor[i].extract(new ByteArrayInputStream(revisionContent.getContentBytes()));
doc.add(Field.Text(CONTENT_TEXT, reader));
@@ -217,7 +217,7 @@
doc.add(Field.Text(CONTENT_TEXT, readContent(revisionDescriptor, revisionContent)));
if ( revisionContent != null && revisionDescriptor != null ) {
- ContentExtractor[] extractor = ExtractorManager.getInstance().getContentExtractors(null, revisionDescriptor);
+ ContentExtractor[] extractor = ExtractorManager.getInstance().getContentExtractors(uri.getNamespace().getName(), null, revisionDescriptor);
for ( int i = 0; i < extractor.length; i++ ) {
Reader reader = extractor[i].extract(new ByteArrayInputStream(revisionContent.getContentBytes()));
doc.add(Field.Text(CONTENT_TEXT, reader));
---------------------------------------------------------------------
To unsubscribe, e-mail: slide-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: slide-dev-help@jakarta.apache.org