statistical analysis

[Richard Fournier]

Dear lilypond developpers,

I am a physicist envolved in research studies dealing with self-organisation
mecanismes and I find myself following a project concerning the statistical
analysis of musical improvisation.

We have the need to transform one monodic voice (or several voices), entered
under the lilypond format, into any simple code for successive hights and
durations, in order to perform dynamical statistical analysis. For the scores
that we enter ourselves, we can impose strong coding constraints so that it
becomes easy to parse the lilypond file (with a simple perl code for instance).
But if we want to perform comparative studies using available lilypond input
files, we will have to deal with all the freedom of lilypond coding and a simple
perl code will never be sufficient.

We therefore wonder wether the lilypond parser could be extracted
and do the job ?
[extract only the successive hights and durations]

Thank you very much in advance for any advice.
Yours sincerely,
Richard Fournier