Organize two_qubit_decompose code

[ShellyGarion]

Summary

Following the discussions on #15833 and #15835, this PR continues the code organization of crates/synthesis/src/two_qubit_decompose

Details and comments

[qiskit-bot]

One or more of the following people are relevant to this code:

• @Qiskit/terra-core
• @levbishop

[Cryoris]

In general I agree to move re-used logic into a common.rs module, but if there's logic that's specific to a single file then I think it's clearer to keep it there. That shows where the logic is required, and importantly does not require making functions or consts pub. Having pub functions like this makes it harder to refactor and understand where it's actually used.

[ShellyGarion]

In general I agree to move re-used logic into a common.rs module, but if there's logic that's specific to a single file then I think it's clearer to keep it there. That shows where the logic is required, and importantly does not require making functions or consts pub. Having pub functions like this makes it harder to refactor and understand where it's actually used.

My motivation here is that the main files for the decomposers (basis decomposer, controlled-u decomposer and weyl decomposer) should mainly include code that is specifically relevant for these decomposers.
This is the reason that I moved functions that seem to be more general (u4_to_su4, real_trace_transform) to common.rs, as well as some static matrices. I thought that putting these functions and matrices in common would be useful in case one adds more synthesis methods in the future (to prevent code duplications).

[Cryoris]

I thought that putting these functions and matrices in common would be useful in case one adds more synthesis methods in the future (to prevent code duplications).

I agree we should move re-used code there, but it's not re-used yet -- do we have any other methods in the pipeline we want to add that would use this? There's obviously no hard rule on how to do this, but we also don't need to change things unnecessarily since we do have to review these changes in the end 🙂

[ShellyGarion]

OK, so I moved back the consts and functions from common.rs to their original files, and added comments indicating these are helper functions / constant matrices.

[Cryoris]

LGTM, thanks for the updates!