Half von mises distribution wrapped on half circle

Modeling
1

I am currently analysing stress data which is bi-directional, i.e. 30 deg = 30 + 180 deg. So the stress direction data obtained is between 0 and 180 deg, and therefore I need a von mises prior that is wrapped between 0 and 180. However, the von mises distribution in Stan appears to work within the range of -pi and pi.

Is it legitimate if I declare my parameter mu to be within (0, pi) like:

parameter{
real<lower=0, upper=pi()> mu;
}

model{
mu ~ von_mises(mu0, kappa0);
angle ~ von_mises(mu, kappa);
}

I also tried using truncated function as von_mises(mu0, kappa0) T[0, pi()], but it seems Stan does not have truncated von_mises distribution.

Does anyone know how to specify such a half von mises distribution in Stan? Thanks a lot!

2

If you’re saying that data lie on a half-circle, and wrap to the other side, can’t you just multiply the angle by 2 and use a standard full circle?