[#SAK-13752] update search to use POI 3.5 to index OOXML docs (docx, etc.)

Sakai

update search to use POI 3.5 to index OOXML docs (docx, etc.)

Created: 13-Jun-2008 08:43 Updated: 28-Sep-2009 16:06

Component/s:

Affects Version/s:

2.5.2

Fix Version/s:

2.6.0, 2.7.0

Time Tracking:

Not Specified

Issue Links:

Relate

This issue is related to:

~~KNL-98~~ Add MIME types for MS OOXML formats (docx, etc.)

2.6.x Status:	Closed
2.5.x Status:	None
2.4.x Status:	None

Description

« Hide

POI v3.1 (or is it 3.5?) is due for release during summer 08, it will support the new office open XML formats.

Indexing such documents in 2.6 would be rather spiffing.

All

Comments

Work Log

Change History

Subversion Commits

git Commits

Sort Order:

[ Permlink | « Hide ]

Stephen Marquard added a comment - 10-Jan-2009 06:59 - edited

Seems we could bump poi and pdfbox to latest versions. Not sure if poi-ooxml will require code changes to support OOXML digesting.

Index: search-impl/impl/pom.xml
===================================================================
--- search-impl/impl/pom.xml (revision 56007)
+++ search-impl/impl/pom.xml (working copy)
@@ -155,21 +155,27 @@
       <type>jar</type>
     </dependency>
     <dependency>
- <groupId>poi</groupId>
+ <groupId>org.apache.poi</groupId>
       <artifactId>poi</artifactId>
- <version>3.0-alpha3-20070301</version>
+ <version>3.5-beta4</version>
       <type>jar</type>
     </dependency>
     <dependency>
- <groupId>poi</groupId>
+ <groupId>org.apache.poi</groupId>
+ <artifactId>poi-ooxml</artifactId>
+ <version>3.5-beta4</version>
+ <type>jar</type>
+ </dependency>
+ <dependency>
+ <groupId>org.apache.poi</groupId>
       <artifactId>poi-scratchpad</artifactId>
- <version>3.0-alpha3-20070301</version>
+ <version>3.5-beta4</version>
       <type>jar</type>
     </dependency>
     <dependency>
       <groupId>pdfbox</groupId>
       <artifactId>pdfbox</artifactId>
- <version>0.7.1</version>
+ <version>0.7.3</version>
       <type>jar</type>
     </dependency>
     <dependency>

[ Permlink | « Hide ]

Stephen Marquard added a comment - 10-Jan-2009 07:50

Looks like to get .docx etc. digesting to work may require a new digester to use the poi code for this (as opposed to textmining for word docs). Another issue is that testing with FF3 on Mac, uploading a .docx into Resources results in a mime type of application/binary rather than application/msword, so the search digester should probably check the file extension in addition to the MIME type.

[ Permlink | « Hide ]

Stephen Marquard added a comment - 10-Jan-2009 12:26

Another library for OOXML:
http://www.openxml4j.org/

[ Permlink | « Hide ]

Stephen Marquard added a comment - 10-Jan-2009 22:50

Some info on poi for docx:

http://www.nabble.com/java.lang.NegativeArraySizeException-tt18729532.html#a18754926

[ Permlink | « Hide ]

Stephen Marquard added a comment - 12-Jan-2009 13:00

Initial support in r56076

[ Permlink | « Hide ]

David Horwitz added a comment - 14-Jan-2009 03:54

tested on local build

[ Permlink | « Hide ]

Anthony Whyte added a comment - 28-Sep-2009 16:06

in 2.6.0.