Author: nick
Date: Tue Oct 19 14:56:54 2010
New Revision: 1024255
URL: http://svn.apache.org/viewvc?rev=1024255&view=rev
Log:
Add iWork support to the Container Aware Detector (TIKA-533)
It's a bit icky for now, but it works and it's quick...
Added:
tika/trunk/tika-parsers/src/test/resources/test-documents/test-zip-of-zip.zip
(with props)
Modified:
tika/trunk/tika-parsers/src/main/java/org/apache/tika/detect/ZipContainerDetector.java
tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/iwork/IWorkPackageParser.java
tika/trunk/tika-parsers/src/test/java/org/apache/tika/detect/TestContainerAwareDetector.java
Modified:
tika/trunk/tika-parsers/src/main/java/org/apache/tika/detect/ZipContainerDetector.java
URL:
http://svn.apache.org/viewvc/tika/trunk/tika-parsers/src/main/java/org/apache/tika/detect/ZipContainerDetector.java?rev=1024255&r1=1024254&r2=1024255&view=diff
==============================================================================
---
tika/trunk/tika-parsers/src/main/java/org/apache/tika/detect/ZipContainerDetector.java
(original)
+++
tika/trunk/tika-parsers/src/main/java/org/apache/tika/detect/ZipContainerDetector.java
Tue Oct 19 14:56:54 2010
@@ -31,6 +31,7 @@ import org.apache.tika.io.IOUtils;
import org.apache.tika.io.TikaInputStream;
import org.apache.tika.metadata.Metadata;
import org.apache.tika.mime.MediaType;
+import org.apache.tika.parser.iwork.IWorkPackageParser;
/**
@@ -87,7 +88,12 @@ public class ZipContainerDetector implem
throw new IOException("Office Open XML File detected, but
corrupted - " + e.getMessage());
}
} else if(entry.getName().equals("buildVersionHistory.plist")) {
- // TODO - iWork
+ // This is an iWork document
+
+ // Reset and ask
+ zip.close();
+ zip = new ZipFile(input.getFile());
+ return IWorkPackageParser.identifyType(zip);
} else if(entry.getName().equals("META-INF/")) {
// Java Jar
return MediaType.application("java-archive");
Modified:
tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/iwork/IWorkPackageParser.java
URL:
http://svn.apache.org/viewvc/tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/iwork/IWorkPackageParser.java?rev=1024255&r1=1024254&r2=1024255&view=diff
==============================================================================
---
tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/iwork/IWorkPackageParser.java
(original)
+++
tika/trunk/tika-parsers/src/main/java/org/apache/tika/parser/iwork/IWorkPackageParser.java
Tue Oct 19 14:56:54 2010
@@ -16,6 +16,15 @@
*/
package org.apache.tika.parser.iwork;
+import java.io.IOException;
+import java.io.InputStream;
+import java.util.Arrays;
+import java.util.Collections;
+import java.util.HashSet;
+import java.util.Set;
+import java.util.zip.ZipEntry;
+import java.util.zip.ZipFile;
+
import org.apache.commons.compress.archivers.ArchiveEntry;
import org.apache.commons.compress.archivers.ArchiveInputStream;
import org.apache.commons.compress.archivers.zip.ZipArchiveInputStream;
@@ -29,13 +38,6 @@ import org.apache.tika.parser.Parser;
import org.xml.sax.ContentHandler;
import org.xml.sax.SAXException;
-import java.io.IOException;
-import java.io.InputStream;
-import java.util.Arrays;
-import java.util.Collections;
-import java.util.HashSet;
-import java.util.Set;
-
/**
* A parser for the IWork container files. This includes *.key, *.pages and
*.numbers files.
* This parser delegates the relevant files to {...@link IWorkParser} that
parsers the content.
@@ -78,4 +80,39 @@ public class IWorkPackageParser implemen
throws IOException, SAXException, TikaException {
parse(stream, handler, metadata, new ParseContext());
}
+
+ /**
+ * Locates the appropriate index file entry, and reads from that
+ * the root element of the document. That is used to the identify
+ * the correct type of the keynote container.
+ */
+ public static MediaType identifyType(ZipFile zip)
+ throws IOException {
+ for (ZipEntry entry : Collections.list(zip.entries())) {
+ if (relevantFileNames.contains(entry.getName())) {
+ // Bingo, found the right entry
+
+ // Grab the first few hundred bytes of the file
+ // This is quite sick, but it is much quicker and less
+ // memory intensive than the "proper" way to do it!
+ byte[] data = new byte[400];
+ InputStream stream = zip.getInputStream(entry);
+ stream.read(data);
+
+ String docStart = new String(data, "UTF-8");
+ if(docStart.contains("ls:document")) {
+ return MediaType.application("vnd.apple.numbers");
+ }
+ if(docStart.contains("sl:document")) {
+ return MediaType.application("vnd.apple.pages");
+ }
+ if(docStart.contains("key:presentation")) {
+ return MediaType.application("vnd.apple.keynote");
+ }
+ }
+ }
+
+ // Not sure, fallback to the container type
+ return MediaType.application("vnd.apple.iwork");
+ }
}
Modified:
tika/trunk/tika-parsers/src/test/java/org/apache/tika/detect/TestContainerAwareDetector.java
URL:
http://svn.apache.org/viewvc/tika/trunk/tika-parsers/src/test/java/org/apache/tika/detect/TestContainerAwareDetector.java?rev=1024255&r1=1024254&r2=1024255&view=diff
==============================================================================
---
tika/trunk/tika-parsers/src/test/java/org/apache/tika/detect/TestContainerAwareDetector.java
(original)
+++
tika/trunk/tika-parsers/src/test/java/org/apache/tika/detect/TestContainerAwareDetector.java
Tue Oct 19 14:56:54 2010
@@ -172,7 +172,25 @@ public class TestContainerAwareDetector
}
public void testDetectIWork() throws Exception {
- // TODO
+ TikaInputStream tis;
+
+ tis = TikaInputStream.get(getTestDoc("testKeynote.key"));
+ assertEquals(
+ MediaType.application("vnd.apple.keynote"),
+ d.detect(tis, new Metadata())
+ );
+
+ tis = TikaInputStream.get(getTestDoc("testNumbers.numbers"));
+ assertEquals(
+ MediaType.application("vnd.apple.numbers"),
+ d.detect(tis, new Metadata())
+ );
+
+ tis = TikaInputStream.get(getTestDoc("testPages.pages"));
+ assertEquals(
+ MediaType.application("vnd.apple.pages"),
+ d.detect(tis, new Metadata())
+ );
}
public void testDetectZip() throws Exception {
@@ -184,6 +202,12 @@ public class TestContainerAwareDetector
d.detect(tis, new Metadata())
);
+ tis = TikaInputStream.get(getTestDoc("test-zip-of-zip.zip"));
+ assertEquals(
+ MediaType.application("zip"),
+ d.detect(tis, new Metadata())
+ );
+
tis = TikaInputStream.get(getTestDoc("testJAR.jar"));
assertEquals(
MediaType.application("java-archive"),
Added:
tika/trunk/tika-parsers/src/test/resources/test-documents/test-zip-of-zip.zip
URL:
http://svn.apache.org/viewvc/tika/trunk/tika-parsers/src/test/resources/test-documents/test-zip-of-zip.zip?rev=1024255&view=auto
==============================================================================
Binary file - no diff available.
Propchange:
tika/trunk/tika-parsers/src/test/resources/test-documents/test-zip-of-zip.zip
------------------------------------------------------------------------------
svn:mime-type = application/octet-stream
