Global phases in peephole optimization (disentangle passes)

[ritu-thombre99]

State-analysis figure from Relaxed Peephole Optimization paper
The phases are ignored since it considers both X|0> = Y|0> = |1>.

Is it possible to keep track of phases by considering eigenstates of PauliY i.e. |+i> and |-i> instead of treating them the same as |+> and |->

Originally posted by @ritu-thombre99 in #1823 (comment)