[Feature]: Change or add extra doc for camera.get_quat()

[MandiZhao]

What feature or enhancement are you proposing?

Currently camera.get_quat()

Genesis/genesis/vis/camera.py

Line 741 in 84eeff5

def get_quat(self, env_idx=None):

directly returns values in self._multi_env_quat_tensor, which is a special quaternion representation specifically for Madrona, this is not the wxyz quaternion convention used in the rest of the codebase so might confuse first-time users.
(self._multi_env_quat_tensor gets updated by:

Genesis/genesis/vis/camera.py

Line 17 in 84eeff5

def _T_to_quat_for_madrona(T):

)

Motivation

As a [e.g. civil engineering researcher working on flood simulation]
I want to [e.g. be able to simulate city-scale fluid / terrain simulations]
Because [e.g. this enables the efficient design and validation of flood protection structures].

Potential Benefit

More clarity for users to access camera rotations (e.g. for randomizing camera poses)

What is the expected outcome of the implementation work?

One option is to just return wxyz quaternion, e.g.:

transform = self._multi_env_transform_tensor[env_idx]
return gu.R_to_quat(transform[..., :3, 3])

Additional information

No response

[duburcqa]

This issue will be fixed by #1563. No inconsistency anymore!