Currently the proof of concept does the minimum by returning classes defined in Qiskit (terra and nature). But, implementing an interface so that qiskit-alt works as a drop-in replacement needs to be done. This only makes sense at the moment for the functionality that is available on the Julia side: jordan-wigner transform.
There may be other useful smaller scale algorithms. This would involve benchmarking the overhead in translating between Julia and Python types compared to the increased efficiency operations on the Julia side.
Currently the proof of concept does the minimum by returning classes defined in Qiskit (terra and nature). But, implementing an interface so that qiskit-alt works as a drop-in replacement needs to be done. This only makes sense at the moment for the functionality that is available on the Julia side: jordan-wigner transform.
There may be other useful smaller scale algorithms. This would involve benchmarking the overhead in translating between Julia and Python types compared to the increased efficiency operations on the Julia side.