Precise locating rendered items

[Jacek Gajek]

Hello!

I am developing an app in which user should be capable to place notes
on a staff. I don't want to implement a whole rendering engine, I
decided to use lilypond. Now.

My idea is to allow user to place notes by clicking on a prerendered
image by lily. All I need is to determine where did the user clicked.
Is it possible to get such characteristics from lily? Of course I
could scan an image and find staff lines etc., but that would be
implementation-dependant and ugly, so I would prefer to gather
objects' locations directly from lily. Is it possible?

Thanks -- Jacek