[FEA]: The new single-phase API for cuda.cccl.parallel should have more checks on the input arrays

### Is this a duplicate?

- [x] I confirmed there appear to be no [duplicate issues](https://github.com/NVIDIA/cccl/issues) for this request and that I agree to the [Code of Conduct](CODE_OF_CONDUCT.md)

### Area

cuda.cccl.parallel (Python)

### Is your feature request related to a problem? Please describe.

Currently the simpler, single-phase API lacks certain correctness checks on the inputs, like the array being contiguous and checking the stream in `__cuda_array_interface__`. There are potentially other things we need to check for as well.

Note that these checks should only be applied to the single-phase API, as the 3 phase API is meant to be a low-level and high-performance, and these checks introduce performance overhead.

### Describe the solution you'd like

We need to check the following for the input arrays passed to cuda.cccl.parallel:

* If they are contiguous
* If we accept `__cuda_array_interface__` arrays, we need to get the stream and make sure we enqueue the kernel on this stream (or potentially wait)

### Describe alternatives you've considered

_No response_

### Additional context

_No response_