Bug in analyticactionangle function

[sophialilleengen]

The analyticactionangle function does not take a single PhaseSpacePosition point:

pot = gp.IsochronePotential(m=1.24e+11, b=4.02, units = galactic)
ics = gd.PhaseSpacePosition(pos=[7.5,0,0.] * u.kpc, vel=[0,-200,0] * u.km/u.s)
pot.action_angle(ics)

TypeError Traceback (most recent call last)
/var/folders/gx/w54ng9cs5k3bfxcphmqpmc740001vz/T/ipykernel_72625/1131334737.py in
----> 1 pot.action_angle(ics)

gala/potential/potential/builtin/cybuiltin.pyx in gala.potential.potential.builtin.cybuiltin.IsochronePotential.action_angle()

/opt/anaconda3/lib/python3.9/site-packages/gala/dynamics/analyticactionangle.py in isochrone_to_aa(w, potential)
105 # Compute theta_z
106 psi = np.arctan2(np.cos(theta), -np.sin(theta)rvtheta/L)
--> 107 psi[np.abs(vtheta) <= 1e-10] = np.pi/2. # blows up for small vtheta
108
109 omega_th = 0.5 * (1 + L/np.sqrt(LL + 4GM*b))

TypeError: numpy.float64 object does not support item assignment

while this works:
pot = gp.IsochronePotential(m=1.24e+11, b=4.02, units = galactic)
ics = gd.PhaseSpacePosition(pos=[7.5,0,0.] * u.kpc, vel=[0,-200,0] * u.km/u.s)
orbits = gp.Hamiltonian(pot).integrate_orbit(ics, dt=4, n_steps=750)
pot.action_angle(orbits)