Qiskit / qiskit / 11683622570

88%

Build Type

push

github

Committed by web-flow

Commit Message

Add meas_level, meas_return, and noise_model options to BackendSamplerV2 (#13357)

* Add meas_level, meas_return, and noise_model options to BackendSamaplerV2

This change adds support to the `BackendSamplerV2` class so that it will
pass through the `meas_level`, `meas_return`, and `noise_model` options
passed to it through to the underlying `BackendV2`'s `run()` method. For
the sake of compatibility with backends that might not expect those
options, it does not pass default values for them (like the previously
defined options for the class) if the default `None` value is not
overridden.

Additionally, to support `meas_level=1`, the results processing code
checks the `meas_level` option and handles `meas_level=1` data
appropriately, rather than always assuming the returned data is level 2.

* Fix type signature for BackendSamplerV2 internal result handling

* Switch from individual new options to a run_options option

run_options is a dict passed on to backend.run as it is for SamplerV2 in
qiskit-aer.

* Add test of using level 1 data with BackendSamplerV2

* Do not clear circuit metadata for BackendSamplerV2

All of the backend primitives use the same helper function for calling
`backend.run` and this function has been clearing metadata because the
estimator primitives can add large objects to the metadata that payload
large for running the circuits on a remote service. In some cases, it is
helpful to have the circuit metadata make it to the `backend.run` call.
Since the concern about large metadata entries should not affect the
sampler case, an option is added here to skip clearing the metadata and
`BackendSamplerV2` is updated to use this option.

* Add release notes

* black

Run Details

24 of 25 new or added lines in 2 files covered. (96.0%)

26 existing lines in 3 files now uncovered.

76772 of 86491 relevant lines covered (88.76%)

355280.85 hits per line