Allow backend.run method to take a list of circuits

[ColemanCollins]

What is the desired enhancement?

Some algorithm helpers (e.g. tomography) provide a list of circuits to be passed to the run method, rather than a single circuit, with all circuits run async in a single job and returned in a single result. It's currently possible to make this work this by just iterating over that list, but it would be nice for cross-compat and broader algorithm support to allow a list to be passed in directly.

[jplewa]

Some other spots in the code are also affected by the lack of support for circuit lists, which causes the problem that I reported here. Specifically, in this case, it's impossible to use Qiskit's amplitude estimation classes with IonQ via Azure, becase the qiskit_circ_to_ionq_circ helper method doesn't handle lists of circuits too well.

I created a pull request for this method to handle lists of size 1 (which will make it possible to use AmplitudeEstimation, IterativeAmplitudeEstimation, and FasterAmplitudeEstimation), but
MaximumLikelihoodAmplitudeEstimation still won't work, since in typically submits multiple circuits at once (so the circuit list is longer than 1). It might make sense to address the helper methods at the same time as backend.run?

EDIT: Upon further investigation, it looks like the issues reported in microsoft/azure-quantum-python#200 are caused by the fact that IonQBackend doesn't support circuit lists either (even those with a single element). If that is addressed, qiskit_circ_to_ionq_circ won't need to be updated to handle circuit lists, as no lists will presumably ever be passed to it by IonQBackend.