Dynamic circuit & meas level 1 support

[jyu00]

What is the expected feature or enhancement?

Qiskit Runtime primitives will soon support dynamic circuits and measurement level 1 in SamplerV2. qiskit-ibm-runtime will need to make corresponding changes for that.

Acceptance criteria

• Document that dynamic circuits are supported in SamplerV2 without DD
• Raise an error if dynamic circuits are passed with DD
• Separate execution options for SamplerV2 that includes measurement level and return (under meas_type)

[ihincks]

The two options in the third bullet will actually be lumped into one option called meas_type as follows:

class SamplerExecutionOptions(ExecutionOptions):
    r"""Extension of :class:~.ExecutionOptions` for the sampler primitive."""

    meas_type: Literal["classified", "kerneled", "avg_kerneled"] = "classified"
    r"""How to process and return measurement results.
    
    This option sets the return type of all classical registers in all :class:`~.PubResult`\s.
    If a sampler pub with shape ``pub_shape`` has a circuit that contains a classical register 
    with size ``creg_size``, then the returned data associated with this register will have one of
    the following formats depending on the value of this option.

    * ``"classified"``: A :class:`~.BitArray` of shape ``pub_shape`` over ``num_shots`` with a 
      number of bits equal to ``creg_size``.

    * ``"kerneled"``: A complex NumPy array of shape ``(*pub_shape, num_shots, creg_size)``, where
      each entry represents an IQ data point (resulting from kerneling the measurement trace) in 
      arbitrary units.

    * ``"avg_kerneled"``: A complex NumPy array of shape ``(*pub_shape, creg_size)``, where
      each entry represents an IQ data point (resulting from kerneling the measurement trace and 
      averaging over shots) in arbitrary units. This option is equivalent to selecting
      ``"kerneled"`` and then averaging over the shots axis, but requires less data bandwidth.

    See https://pubs.aip.org/aip/rsi/article/88/10/104703/836456 for a description of kerneling.
    """

This can be further word-smithed. We could also add a note about how "kerneled" is essentially (meas_level=1, meas_return=avg), "avg_kerneled" is (meas_level=1, meas=single), and "classified" is meas_level 2.

[jyu00]

Closing since both PRs are merged.