explore refactoring array handling in a separate module

[sgrava]

Explore refactoring array handling in a separate module.

• Ideally similar to pulser.math
• Type aliases could simplify type hints

from typing import TypeAlias

Array: TypeAlias = npt.NDArray[np.float64] | torch.Tensor
ArrayLike: TypeAlias = npt.ArrayLike | Array

Benefits:

• reusing numpy/torch array handling in other parts of the library, mostly waveforms