WMF extraction

[Andreas Beeker <ki...@apache.org>]

Hi,

POI will have a WMF module (org.apache.poi.hwmf.*) in the next beta.
Looking over the govdocs collection, those embedded wmfs might contain interesting
information for TIKA.

Although my main goal is to integrate the rendering for common sl,
it shouldn't be to laborious to provide something afterwards.

Should the output be part of the embedding document, e.g. ppt, or
does it make sense to crawl over various extensions and extract those metadata
separately?
(I haven't checked how the parsers are called, so this might be nonsense ...)

Andi

Re: WMF extraction

[Nick Burch <ap...@gagravarr.org>]

On Thu, 14 Jan 2016, Andreas Beeker wrote:
> POI will have a WMF module (org.apache.poi.hwmf.*) in the next beta. 
> Looking over the govdocs collection, those embedded wmfs might contain 
> interesting information for TIKA.
>
> Should the output be part of the embedding document, e.g. ppt, or does 
> it make sense to crawl over various extensions and extract those 
> metadata separately?

I'd suggest a two-step process. One is to update the current office 
parsers (especially HSLF) as needed to expose the embedded WMF files as 
embedded resources, much as they do for embedded jpegs, pngs etc

Next, add a WMF parser that uses HWMF to expose any useful metadata you 
can find

Tika will then call the WMF parser for embedded WMFs where requested

Nick

RE: WMF extraction

["Allison, Timothy B." <ta...@mitre.org>]

Great!  Thank you!

-----Original Message-----
From: Andreas Beeker [mailto:kiwiwings@apache.org] 
Sent: Thursday, January 14, 2016 5:41 PM
To: dev@tika.apache.org
Subject: WMF extraction

Hi,

POI will have a WMF module (org.apache.poi.hwmf.*) in the next beta.
Looking over the govdocs collection, those embedded wmfs might contain interesting information for TIKA.

Although my main goal is to integrate the rendering for common sl, it shouldn't be to laborious to provide something afterwards.

Should the output be part of the embedding document, e.g. ppt, or does it make sense to crawl over various extensions and extract those metadata separately?
(I haven't checked how the parsers are called, so this might be nonsense ...)

Andi