Update IonQBackend to handle circuit lists of size one

[jplewa]

closes #200

Qiskit's BackendV1 supports lists of circuits in its run method. This isn't implemented in azure-quantum and qiskit-ionq, which means that passing a list results in confusing exceptions. A while ago, qiskit-ionq added some changes that make it possible to handle circuit lists of length 1 and also added more explicit error messages (see qiskit-community/qiskit-ionq#71). They also have an issue for adding multi-circuit experiment support in the future (see qiskit-community/qiskit-ionq#70).

I believe it would make sense to do the same in azure-quantum, because currently some of Qiskit's features are not usable. I reported this issue in #200.

To make Qiskit's features such as MaximumLikelihoodAmplitudeEstimation usable, azure-quantum would actually need to learn how to handle multi-circuit experiments, but with this PR we can at least make AmplitudeEstimation, IterativeAmplitudeEstimation, and FasterAmplitudeEstimation usable, since they pass single-element circuit lists to the run method.

EDIT: With this PR, every amplitude estimation method from Qiskit should start working.

[ghost]

All CLA requirements met.

[azure-pipelines]

Azure Pipelines successfully started running 2 pipeline(s).

[guenp]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 2 pipeline(s).

[jplewa]

To make Qiskit's features such as MaximumLikelihoodAmplitudeEstimation usable, azure-quantum would actually need to learn how to handle multi-circuit experiments, but with this PR we can at least make AmplitudeEstimation, IterativeAmplitudeEstimation, and FasterAmplitudeEstimation usable, since they pass single-element circuit lists to the run method.

I re-verified the claim I'd made about MaximumLikelihoodAmplitudeEstimation. This actually isn't true! Once this PR is merged, MLAE should work due to the code changes Terra introducted in Qiskit/qiskit#6391. There's a clever bit of code that splits circuit lists into one-element batches if the backend requires that (which azure-quantum does).

[guenp]

Thanks @jplewa for fixing this issue and adding tests as well! It looks good to me but before we merge to main we also want to support this on the Honeywell backend, i.e. in HoneywellBackend.run.

[guenp]

/azp run

[azure-pipelines]

Azure Pipelines successfully started running 2 pipeline(s).