[TIKA-2458] Unify number of pages metadata key?

• Type: Improvement
• Status: Open
• Priority: Minor
• Resolution: Unresolved
• Component/s: core
• Assignee: Unassigned
• Reporter: Tim Allison
• Created: 01/Sep/17 14:07
• Updated: 01/Sep/17 14:24

Description

On TIKA-2451, we're adding a metadata value for the number of images in a tiff. This raises the broader (admittedly minor) question of how we want to handle "number of pages".

I'm opening this issue for discussion and feedback.

Unfortunately Dublin Core doesn't have a number of pages element as far as a I can tell.

Do we want to have a single key in TikaCoreProperties that is "number of pages" that would be used for:

1. number of pages in a PDF
2. number of pages that a .docx alleges it has
3. the number of slides in a PPT
4. the number of sheets in an XLS
5. the number of tiffs in a multi-image tiff

Others?

Or, do we want to have different keys MSOffice.PageCount, PagedText.N_PAGES, TIFF.NUM_TIFFS

Or, thanks to the beauty of composite keys, do we want to have both a unified key and the above individual keys?

*I would propose using PagedText's N_PAGES as the unifying key, but the definition of that seems to be strictly within XMP-land and it should be a sum of the pages in the container document and all embedded documents according to our javadocs.