Bug 61296

Summary: Bring over missing constants from Tika
Product: POI Reporter: Nick Burch <apache>
Component: POI OverallAssignee: POI Developers List <dev>
Status: NEW ---    
Severity: enhancement    
Priority: P2    
Version: 3.17-dev   
Target Milestone: ---   
Hardware: All   
OS: All   
Attachments: a quick comparison of Tika and POI constants
a quick comparison of Tika and POI constants

Description Nick Burch 2017-07-13 20:31:45 UTC
In Apache Tika, under tika-parsers/src/main/java/org/apache/tika/parser/microsoft/, there's now a surprisingly large number of POI and OOXML constants in the parser codebase

We should review these, add our own constants where we don't already have them (eg relationships or types we don't have defined), then swap the Tika classes to using our constants after a release
Comment 1 Javen O'Neal 2017-07-14 03:15:49 UTC
Created attachment 35138 [details]
a quick comparison of Tika and POI constants


git clone https://github.com/apache/tika.git apache-tika
pushd apache-tika
cd tika-parsers/src/main/java/org/apache/tika/parser/microsoft/
grep -r -P "(static final|final static|http://schemas|vnd|urn)" .

Most notably,
* ./ooxml/AbstractOOXMLExtractor.java has 8 relationship schema URLS and 1 ooxml mime type
* ./ooxml/OOXMLWordAndPowerPointTextHandler.java has 6 schema urls and 2 urns
* ./POIFSContainerDetector.java has several mime types
And a few others
See attachment for a list of current constants that could be copied over.
Comment 2 Javen O'Neal 2017-07-14 03:16:27 UTC
Comment 3 Javen O'Neal 2017-07-14 03:40:12 UTC
Comment 4 Javen O'Neal 2017-07-14 03:43:44 UTC
Created attachment 35139 [details]
a quick comparison of Tika and POI constants
Comment 5 Tim Allison 2017-07-14 11:10:03 UTC
Yup.  Sorry.  I've been meaning to do this.  Thank you, Nick and Javen!

Speaking of which...is there any interest in moving over the SAX-based docx/pptx code from Tika into POI?
Comment 6 Javen O'Neal 2017-07-14 14:54:48 UTC
Yes, absolutely!