2 qubit unitary_to_observation has extra factor of 1/2

[Farquhar13]

Is this correct @choihrM3?

In the single qubit env:

    def unitary_to_observation(self, U):
        return (
            np.array(
                [(abs(x), (cmath.phase(x) / np.pi + 1) / 2) for x in U.flatten()],
                dtype=np.float64,
            )
            .squeeze()
            .reshape(-1)
        )

In the two qubit env:

def unitary_to_observation(self, U):
        return (
            np.array(
                [(abs(x), (cmath.phase(x) / 2 / np.pi + 1) / 2) for x in U.flatten()], 
                dtype=np.float64,
                )
            .squeeze()
            .reshape(-1)  # cmath phase gives -2pi to 2pi (?)
        )

[Farquhar13]

If the extra /2 is not needed than we can have the two qubit env inherit this method.

[choihrM3]

extra "/2" is not needed, and I believe "self.observation_space" should be also fixed?

[Farquhar13]

@choihrM3 I don't think this should affect self.observation_space. If the extra \2 factor is fixed on your current 2-qubit work, then I think we can close this.