Search for Office 2007 in OSGi environment

[Marc Speck <ma...@gmail.com>]

We are testing an update from Jackrabbit 2.1.1 to 2.2.4 in a OSGi
environment similar to the sling.jcr.server bundle. I found the following
issues with search:

- JCR-2838 introduces a dependency of Tika to
org.apache.jackrabbit.core.query.pdf.PDFParser. tika-bundle-0.8 cannot
import that class.
- tika-config.xml configures now org.apache.tika.parser.DefaultParser but
defaultParser.getParser().size() == 0. This seems to be a classloading issue
because new DefaultParser(DefaultParser.class.getClassLoader()) works fine
(see patch below)
- Falling back to SearchIndex.textFilterClasses does not work for Office
2007: JackrabbitParser defines OfficeParser for Office 2007 even though
OfficeParser fails to read the format according to log warnings. I
guess OOXMLParser would work (see patch below)
- tika-bundle-0.9 does not embed xmlbeans-qname that xmlbeans seems to
require.

What works fine so far is applying the patch below and using a patched
tika-bundle-0.9. Not sure whether you want to fix this in Jackrabbit or Tika
or whether I'm just messing up things... but maybe it helps someone :)
Marc



   Index: src/main/java/org/apache/jackrabbit/core/query/lucene
/JackrabbitParser.java

===================================================================

--- src/main/java/org/apache/jackrabbit/core/query/lucene
/JackrabbitParser.java (revision 1071702)

+++ src/main/java/org/apache/jackrabbit/core/query/lucene
/JackrabbitParser.java (working copy)

@@ -33,6 +33,7 @@

 import org.apache.tika.parser.html.HtmlParser;

 import org.apache.tika.parser.image.ImageParser;

 import org.apache.tika.parser.microsoft.OfficeParser;

+import org.apache.tika.parser.microsoft.ooxml.OOXMLParser;

 import org.apache.tika.parser.odf.OpenDocumentParser;

 import org.apache.tika.parser.pdf.PDFParser;

 import org.apache.tika.parser.rtf.RTFParser;

@@ -79,7 +80,13 @@

         try {

             if (stream != null) {

                 try {

-                    parser = new AutoDetectParser(new TikaConfig(stream));

+                ClassLoader classLoader =
Thread.currentThread().getContextClassLoader();

+
Thread.currentThread().setContextClassLoader(TikaConfig.class.getClassLoader());

+             try {

+             parser = new AutoDetectParser(new TikaConfig(stream));

+             } finally {

+             Thread.currentThread().setContextClassLoader(classLoader);

+             }

                 } finally {

                     stream.close();

                 }

@@ -134,9 +141,10 @@

                 parsers.put(MediaType.application("vnd.ms-powerpoint"),
parser);

                 parsers.put(MediaType.application("mspowerpoint"), parser);

                 parsers.put(MediaType.application("vnd.ms-excel"), parser);

-
parsers.put(MediaType.application("vnd.openxmlformats-officedocument.wordprocessingml.document"),
parser);

-
parsers.put(MediaType.application("vnd.openxmlformats-officedocument.presentationml.presentation"),
parser);

-
parsers.put(MediaType.application("vnd.openxmlformats-officedocument.spreadsheetml.sheet"),
parser);

+                OOXMLParser ooxmlParser = new OOXMLParser();

+
parsers.put(MediaType.application("vnd.openxmlformats-officedocument.wordprocessingml.document"),
ooxmlParser);

+
parsers.put(MediaType.application("vnd.openxmlformats-officedocument.presentationml.presentation"),
ooxmlParser);

+
parsers.put(MediaType.application("vnd.openxmlformats-officedocument.spreadsheetml.sheet"),
ooxmlParser);

             } else if
(name.equals("org.apache.jackrabbit.extractor.OpenOfficeTextExtractor")) {

                 Parser parser = new OpenDocumentParser();


parsers.put(MediaType.application("vnd.oasis.opendocument.database"),
parser);

Index: src/main/resources/org/apache/jackrabbit/core/query/lucene/tika
-config.xml

===================================================================

--- src/main/resources/org/apache/jackrabbit/core/query/lucene/tika
-config.xml (revision 1071702)

+++ src/main/resources/org/apache/jackrabbit/core/query/lucene/tika
-config.xml (working copy)

@@ -23,11 +23,6 @@



     <parser class="org.apache.tika.parser.DefaultParser"/>



-    <parser class="org.apache.jackrabbit.core.query.pdf.PDFParser">

-      <!-- JCR-2838: Override the faulty PDF parser in Tika 0.8 -->

-      <mime>application/pdf</mime>

-    </parser>

-

     <parser class="org.apache.tika.parser.EmptyParser">

       <!-- Disable package extraction as it's too resource-intensive -->

       <mime>application/x-archive</mime>