Index: tika-parsers/src/main/java/org/apache/tika/detect/ContainerAwareDetector.java
===================================================================
--- tika-parsers/src/main/java/org/apache/tika/detect/ContainerAwareDetector.java (revision 987225)
+++ tika-parsers/src/main/java/org/apache/tika/detect/ContainerAwareDetector.java (working copy)
@@ -78,13 +78,45 @@
// Is this an ole2 file?
long ole2Signature = LittleEndian.getLong(first8, 0);
+ MediaType poiDetectorMediaType = null;
if(ole2Signature == HeaderBlockConstants._signature) {
- return poifsDetector.detect(input, metadata);
+ poiDetectorMediaType = poifsDetector.detect(input, metadata);
+ input = TikaInputStream.get(input.getFile());
}
- // Not a supported container, ask our fall back
- // detector to figure it out
- return fallbackDetector.detect(input, metadata);
+ if (poiDetectorMediaType == null ||
+ poiDetectorMediaType.equals(MediaType.application("x-tika-msoffice"))) {
+ // Not a supported container, ask our fall back
+ // detector to figure it out, it's also possible that the fallback container
+ // can return a more specific result than the container-aware
+ MediaType fallbackDetectorMediaType = fallbackDetector.detect(input, metadata);
+
+ if (poiDetectorMediaType == null && fallbackDetectorMediaType == null) {
+ return null;
+ } else if (poiDetectorMediaType == null && fallbackDetectorMediaType != null) {
+ return fallbackDetectorMediaType;
+ } else if (poiDetectorMediaType != null && fallbackDetectorMediaType == null) {
+ return poiDetectorMediaType;
+ } else if (fallbackDetector instanceof MimeTypes) {
+ // use mime types to see which one is more specific
+ MimeTypes mimeTypes = (MimeTypes)fallbackDetector;
+ if (mimeTypes.getMediaTypeRegistry().isSpecializationOf(poiDetectorMediaType, fallbackDetectorMediaType)) {
+ // return the more specific one
+ return poiDetectorMediaType;
+ } else if (mimeTypes.getMediaTypeRegistry().isSpecializationOf(fallbackDetectorMediaType, poiDetectorMediaType)) {
+ // return the more specific one
+ return fallbackDetectorMediaType;
+ } else {
+ // they aren't related, return the one from fallback
+ return fallbackDetectorMediaType;
+ }
+ } else {
+ // we can't decide which is more specific, return the one from fallback
+ return fallbackDetectorMediaType;
+ }
+ } else {
+ return poiDetectorMediaType; // specific enough
+ }
}
}
Index: tika-parsers/src/test/java/org/apache/tika/detect/TestContainerAwareDetector.java
===================================================================
--- tika-parsers/src/test/java/org/apache/tika/detect/TestContainerAwareDetector.java (revision 987225)
+++ tika-parsers/src/test/java/org/apache/tika/detect/TestContainerAwareDetector.java (working copy)
@@ -16,6 +16,7 @@
*/
package org.apache.tika.detect;
+import java.io.ByteArrayInputStream;
import java.io.InputStream;
import junit.framework.TestCase;
@@ -26,6 +27,8 @@
import org.apache.tika.io.TikaInputStream;
import org.apache.tika.metadata.Metadata;
import org.apache.tika.mime.MediaType;
+import org.apache.tika.mime.MimeTypes;
+import org.apache.tika.mime.MimeTypesFactory;
/**
* Junit test class for {@link ContainerAwareDetector}
@@ -154,4 +157,46 @@
d.detect(tis, new Metadata())
);
}
+
+ public void testCorelPresentationsX3() throws Exception {
+ MimeTypes mimeTypes = MimeTypesFactory.create("tika-mimetypes.xml");
+ ContainerAwareDetector detector = new ContainerAwareDetector(mimeTypes);
+ InputStream input = getTestDoc("corel-presentations-x3.shw");
+ Metadata metadata = new Metadata();
+ metadata.add(Metadata.RESOURCE_NAME_KEY, "corel-presentations-x3.shw");
+ MediaType mt = detector.detect(input, metadata);
+ assertEquals(MediaType.application("x-corelpresentations"),mt);
+ }
+
+ public void testQuattroPro7() throws Exception {
+ MimeTypes mimeTypes = MimeTypesFactory.create("tika-mimetypes.xml");
+ ContainerAwareDetector detector = new ContainerAwareDetector(mimeTypes);
+ InputStream input = getTestDoc("corel-quattro-pro-7.wb3");
+ Metadata metadata = new Metadata();
+ metadata.add(Metadata.RESOURCE_NAME_KEY, "corel-quattro-pro-7.wb3");
+ MediaType mt = detector.detect(input, metadata);
+ assertEquals(MediaType.application("x-quattropro"),mt);
+ }
+
+ public void testQuattroProX3() throws Exception {
+ MimeTypes mimeTypes = MimeTypesFactory.create("tika-mimetypes.xml");
+ ContainerAwareDetector detector = new ContainerAwareDetector(mimeTypes);
+ InputStream input = getTestDoc("corel-quattro-pro-x3.qpw");
+ Metadata metadata = new Metadata();
+ metadata.add(Metadata.RESOURCE_NAME_KEY, "corel-quattro-pro-x3.qpw");
+ MediaType mt = detector.detect(input, metadata);
+ assertEquals(MediaType.application("x-quattropro"),mt);
+ }
+
+ public void testWorksWordProcessor70() throws Exception {
+ MimeTypes mimeTypes = MimeTypesFactory.create("tika-mimetypes.xml");
+ ContainerAwareDetector detector = new ContainerAwareDetector(mimeTypes);
+ InputStream input = getTestDoc("microsoft-works-word-processor-7.0.wps");
+ Metadata metadata = new Metadata();
+ metadata.add(Metadata.RESOURCE_NAME_KEY, "microsoft-works-word-processor-7.0.wps");
+ MediaType mt = detector.detect(input, metadata);
+ assertEquals(MediaType.application("vnd.ms-works"),mt);
+ }
+
+
}
Index: tika-core/src/main/resources/org/apache/tika/mime/tika-mimetypes.xml
===================================================================
--- tika-core/src/main/resources/org/apache/tika/mime/tika-mimetypes.xml (revision 987225)
+++ tika-core/src/main/resources/org/apache/tika/mime/tika-mimetypes.xml (working copy)
@@ -1310,10 +1310,30 @@
+
+
+
+
+
+
+
+
+
+
+
@@ -2241,6 +2261,16 @@
+
+
+
+
+
+
+
+
+
+