Integrate QRAO

[t-imamichi]

What should we add?

Integrate qrao-prototype into Qiskit optimization.

• discuss whether to_ising returns identity matrix as offset instead of returning offset float value
  • Currently: operator, offset = qp.to_ising()
  • Newly: operator = qp.to_ising()
  • In order not to break API, introducing an option to control output may be useful, e.g., to_ising(offset: bool) to return operator and offset if offset=True, otherwise return only operator.
• magic rounding needs to be migrated to sampler
• refactor solve so that QRAO solver can access the problem to be encoded
• discuss whether allow pytest for unit tests because qrao-prototype uses pytest for unit tests.

[garrison]

Some comments to follow up on the above points:

• There is an open PR to deprecate offset: Deprecate encoding.offset while incorporating it into encoding.qubit_op prototype-qrao#14
• There is an open issue for migrating magic rounding to the sampler: Incorporate qiskit primitives into magic rounding prototype-qrao#21. I have not started to work on it.
• There is a draft PR for generating the encoding on the fly: Allow the encoding to be generated on the fly in solve() prototype-qrao#13, with the corresponding issue at Implement "implicit" solve(problem) interface, able to generate encoding on the fly prototype-qrao#9.

[t-imamichi]

This might be related too #325