Run python CI tests across all backends

Required prerequisites

• Search the issue tracker to check if your feature has already been mentioned or rejected in other issues.

Describe the feature

We currently only test our python API on the simulator that is set within each individual test (most use the default). We should extend this to run across the same targets that our C++ suite runs on. This will require sorting out issues wherein certain tests have hard-coded count values that we assert against. E.g

kernel = cudaq.make_kernel()
... kernel ops ...

result = cudaq.sample(kernel)
assert result["01010101"] == 505
assert result["10101010"] == 495

The actual count values will have minor variations across backends, so a new strategy will be needed to either better generalize these asserts, or to have different checks per backend.

Start by comparing the Python test setup with the targets used by the C++ suite, then identify tests with hard-coded count assertions such as the example in the issue. Done means Python CI runs across the same backends and its result checks accommodate backend-specific count variation.