[
https://issues.apache.org/jira/browse/TIKA-2191?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Tim Allison updated TIKA-2191:
------------------------------
Attachment: element_counts_ooxml-docx.xlsx
I counted the elements in the main story .xml file (mostly document.xml) in
~150k doc[xm] files in our regression corpus. I optimized the if/else
branching in startElement and endElement to test for the most common elements
earlier.
There are a few other interesting things in these stats...including rare "dev"
name spaces like
{{http://schemas.openxmlformats.org/wordprocessingml/2006/2/main}}
> Apply current .docx unit tests to experimental SAX parser and fix or document
> as necessary
> ------------------------------------------------------------------------------------------
>
> Key: TIKA-2191
> URL: https://issues.apache.org/jira/browse/TIKA-2191
> Project: Tika
> Issue Type: Improvement
> Reporter: Tim Allison
> Priority: Minor
> Attachments: element_counts_ooxml-docx.xlsx
>
>
> There are many areas for clean up to ensure that the new SAX .docx parser
> yields similar results to the legacy DOM .docx parser. Let's use this issue
> to track work on improvements.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
