[TIKA-1823] Support detecting DWF format

• Type: Improvement
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: 1.13
• Component/s: detector, mime
• Labels: detection dwf mime
• Assignee: Unassigned
• Reporter: Luca Moretti
• Created: 05/Jan/16 23:32
• Updated: 26/Jan/16 16:35
• Resolved: 26/Jan/16 16:35

Description

Tika currently detects dwf files as application/octect-stream.
To make Tika mime magic detector correctly recognize dwf files it should be added this code fragment in tika-mimetypes.xml registry:

<mime-type type="model/vnd.dwf">
	<acronym>dwf</acronym>
	<_comment>Design Web Format</_comment>
	<magic priority="50">
		<match type="string" offset="0" value="(DWF V">
			<match type="string" offset="8" value=".">
				<match type="string" offset="11" value=")" />
			</match>
		</match>
	</magic>
	<glob pattern="*.dwf" />
</mime-type>

In current version (DWF 6.0), dwf file is a ZIP-compressed container for vector-based CAD drawings. It is basically a ZIP archive with the (DWF V06.00) signature added before the regular ZIP magic number. For this reason, the match value to detect dwf files should be: (DWF V06.00)PK.
In the previous versions, the dwf data transport isn't a ZIP file format, so the magic number is only the (DWF V00.55) signature in the file header.
To make Tika detect dwf files with this version too I propose the match value in the code above.

Thanks,

Luca

P.S.: The DWF format specification is included in the DWF Toolkit. The DWF Toolkit is available for free at http://www.autodesk.com/dwftoolkit