Spin intro changes

[hannorein]

I've changed the variable names to better correspond to the param names. Please check that this makes sense before merging.

This is NOT urgent, but there is one thing I was also stumbling upon while going over things. The total energy is calculated this was:

sim.calculate_energy() + rebx.spin_potential(sf)

But the total angular momentum is calculated this way:

rebx.calculate_total_angular_momentum()

Is there a reason for so many inconsistencies? Specifically:

• The rebound python function calculate_energy has calculate in it (note: the rebound C function does not). The reboundx function calculate_total_angular_momentum has it too. But the reboundx function spin_potential does not. 🤯
• spin_potential needs the force object sf as an argument, but calculate_total_angular_momentum does not.
• Why is there a reboundx function that calculates the total angular momentum, but there is none for the total energy.

On the REBOUND side, I suggest adding a new function sim.energy() and adding a deprecation warning to sim.calculate_energy(). I can make that change.

Not sure what the best approach is to the REBOUNDx side. I suggest:

• Remove rebx.calculate_total_angular_momentum and just have rebx.spin_angular_momentum(), then let the user add the normal angular momentum and the spin angular momentum like in the energy case.
• Remove the sf argument in rebx.spin_potential(sf). I think this is actually a 🐞 and should not be there in the first place (the c version doesn't have that argument and ctypes cannot check the argument list of a c function so this does not result in a warning/error).

[dtamayo]

Thank you so much for catching the naming mismatches left in Spinsintro.ipynb. And I definitely agree with everything else. I sent a pull request with your suggestion for REBOUND (also calculate_angular_momentum -> angular_momentum), and I've taken all of your suggestions. I did combine spin_potential and spin_kinetic energy into one and I think more consistently named it rebx.tides_spin_energy(). Angular momentum is a bit annoying in that sim.angular_momentum() + rebx.spin_angular_momentum() doesn't work since each of them returns a list, but we can cast them to np.array, or maybe switch them to reb_vec3d if someone has a good answer on stackoverflow

[hannorein]

Great! Thanks!

I agree, it's a bit annoying that the addition of lists doesn't work. One workaround would be to have the Vec3D class not be a subclass of ctypes.Structure but our own. We could write our own setters/getters for x, y, z, and keep a reference to the actual ctypes.Structure in our class but hidden from the user. Clearly a hack to circumvent some weird behaviour in numpy. Not sure if it's worth it. Let's wait a little longer to see if there is a simpler solution.