sketching out some thoughts on representing coordinates in Python

astropy_coordinates_api.py

position = np.array([15, 18.5, -0.7]) * u.kpc
c = Cartesian(position, frame=None)
c.transform_system(Spherical)
c.transform_frame(Heliocentric) # raises error - no input frame specified
 
# the system/representation classes -- e.g. Cartesian, Spherical, 
#   Cylindrical, Bispherical, etc. -- accept any Quantity object
velocity = np.array([10.4, 57.134, -81.124]) * u.km/u.s
v = Cartesian(velocity, frame=None)
v.transform_system(Cylindrical)
 
v = Cartesian(velocity, frame=GSR)
c.transform_frame(Heliocentric)
c.transform_system(Spherical)
 
# the more high level API would have the system/representation and
#   frame set automatically, e.g. for ICRS
c = ICRS('5:10:20.52 +23:23:23.5', units=(u.hourangle, u.degree))
print(c.frame) # <ICRSFrame  ...>
print(c.system) # <Spherical ...>

# the kind of thing I have been dealing a lot with is dealing with 
#   simulated observations of simulation data, so somethings like
#   this would be very nice:

# read in table of particle data, always in ASCII b.c. we're masochists
tbl = ascii.read("nbody_output.txt")

# this works because Table Columns will know about units, right?
pos = Cartesian(tbl["x","y","z"], frame=GSR)
vel = Cartesian(tbl["vx","vy","vz"], frame=GSR)

# now I need to "observe" these particles from heliocentric frame,
#   presumably as Spherical coordinates

# this goes GSR to LSR, then LSR to Heliocentric. One gotcha is how do we
#   let the user specify what circular velocity and peculiar velocity to use? 
pos.transform_frame(Heliocentric)
heliocentric_pos = pos.spherical # same as pos.transform_system(Spherical)?

vel.transform_frame(Heliocentric)
heliocentric_vel = vel.spherical

# ...