[BUG] Various Operator.pow methods with a batched exponent fail

[albi3ro]

Expected behavior

I expect to get a decomposition with a batch dimension.

Actual behavior

Source code and traceback below

Additional information

We have a moderate amount of support for Pow operators with a batched exponent. For example:

(qml.X(0) ** np.array([0,1])).matrix()

Works as expected.

But many definitions of Operator.pow do not work when pow has a batch dimension.

Source code

qml.X(0).pow(np.array([0,1]))

Tracebacks

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
Cell In[6], line 1
----> 1 qml.X(0).pow(np.array([0,1]))

File ~/Prog/pennylane/pennylane/ops/qubit/non_parametric_ops.py:337, in PauliX.pow(self, z)
    335 def pow(self, z):
    336     z_mod2 = z % 2
--> 337     if abs(z_mod2 - 0.5) < 1e-6:
    338         return [SX(wires=self.wires)]
    339     return super().pow(z_mod2)

ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

System information

PL master.

Existing GitHub issues

• I have searched existing GitHub issues to make sure the issue does not already exist.