Add specialized Sampler primitive

[airwoodix]

Summary

This patch adds a specialized version of the Qiskit Sampler primitive that enables the use of Sampler-based algorithms that have parametric circuits.

Depends on #57. The branch will be rebased after #57 is merged.

Details and comments

• usage of the AQTSampler is mandatory with AQT resources because the custom transpilation passes need to work on parameter-bound circuits.
• AQTSampler implementation works around that limitation by delaying the application of the custom transpilation passes to after the parameters are bound to the circuit. This way, the partially transpiled circuits can still be cached by the underlying BackendSampler implementation.
• the offline_simulator_no_noise fixture is reworked to use a new MockSimulator class that encapsulates the specialized-mock post-processing rather than relying on unittest.mock. It is now arguable whether offline_simulator_no_noise should still be a fixture. We could clean this up at a later point if relevant.

[github-actions]

☂️ Python Coverage

current status: ✅

Overall Coverage

Lines	Covered	Coverage	Threshold	Status
1347	1328	99%	97%	🟢

New Files

File	Coverage	Status
qiskit_aqt_provider/primitives/init.py	100%	🟢
qiskit_aqt_provider/primitives/sampler.py	100%	🟢
TOTAL	100%	🟢

Modified Files

File	Coverage	Status
qiskit_aqt_provider/aqt_resource.py	100%	🟢
qiskit_aqt_provider/test/fixtures.py	100%	🟢
qiskit_aqt_provider/transpiler_plugin.py	100%	🟢
test/test_primitives.py	100%	🟢
test/test_resource.py	100%	🟢
test/test_transpilation.py	100%	🟢
TOTAL	100%	🟢

updated for commit: 9ae84b0 by action🐍