How to extract rendered glyphs (graphemes?) with Pango?

[Alex Kerr <alex phonething com>]

Hi,

Can anyone suggest a method for extracting the rendered glyph images in 
image form (i.e. after combining and shaping) for any script please? I 
need them in visual order, not logical order.
i.e. what APIs to use, or even source code (/pseudocode). Context is any 
script, but especially complex scripts, e.g. Indic, Arabic etc.

I need to be able to uniquely identify each rendered glyph (/grapheme) 
so I know whether it's already been rendered before in a document. 
Obviously each glyph could consist of several unicode code points, so 
not sure that's the best way to track them if there's another a unique 
identifier for the rendered grapheme.

Also, I don't suppose it's possible to extract rendered diacritics is 
it, and positioning information for them relative to the base character 
they're attached to? Would be amazing if so :)

Many thanks,
Alex