Impractical to use Partition with CDType (always needs a QAny)

[Axel-pappalardo]

Partition when "splitting" (partition = True) requires an input of dtype: QAny, thus not working when giving it a classical register (such as CUInt).
When "merging" (partition = False), it outputs a QAny(bitsize) even when given CBit (which should at least raise an error).

It is still possible to use Partition in these cases using Cast, but it messes up qubit count and feels like an unnecessary work around.

Possible solutions would be to add a classical_bits: bool = False argument to Partition which would use a dtype: CAny instead of QAny.
Another possibility is a new Partition method for Classical bits.

[mpharrigan]

If you'd like to take a stab at this, Partition is now a thin wrapper around _PartitionBase and you could make a classical partition without too much boilerplate

[Axel-pappalardo]

Thank you, I will take a stab at this once the pull request #1812 is resolved.