[FEA]: DeviceReduceTransformReduceByKey

### 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

CUB

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

I'm trying to use DeviceReduce to efficiently compute the bounding box for each set of inputs which have a matching key. The input value is a lot smaller than the structure I'm reducing over.

### Describe the solution you'd like

I would like to see the `ByKey` equivalent of `DeviceReduce::TransformReduce`.
I'm able to do this by using a TransformInputIterator, but that consumes more memory.

### Describe alternatives you've considered

`DeviceReduce::ReduceByKey` can be persuaded to work with a `cub::TransformInputIterator`, but has the same tradeoffs that made it worth building `DeviceReduce::TransformReduce`.

### Additional context

_No response_