Radians vs degrees in quaternion logic

Hi Brax,

When converting between quaternions and Euler angles, the function [euler_to_quat](https://github.com/google/brax/blob/673a41f780fe0f137507d1c35286577c48bde4d1/brax/math.py#L69) takes the Euler angles in degrees, but the inverse function, [quat_to_euler](https://github.com/google/brax/blob/673a41f780fe0f137507d1c35286577c48bde4d1/brax/math.py#L81), returns the Euler angles in radians.

It is documented, and it is a relatively small fix to make in your own code to wrap one of these functions to work either exclusively in degrees or radians; but the code is not really self-consistent (and cost me a day's head-scratching, although I suppose I should have read the docstring...!).

Changing this interface would be a _wildly_ breaking change, so I don't know if there is a quick-fix -- maybe explicitly adding wrapped pairs functions that work in exclusively radians _or_ degrees? Maybe this is something for a later release?

Thanks,
Andy