Select an item from a list of object of any type when using tensorflow 2.x

Given a list of instances of class A, [A() for _ in range(5)], I want to randomly select one of them (see the following code for an example)

class A:
    def __init__(self, a):
        self.a = a
    def __call__(self):
        return self.a
def f():
    a_list = [A(i) for i in range(5)]
    a = a_list[random.randint(0, 5)]()
    return a

f()

Is there is a way to decorate f with  `@tf.function` without changing what f does and without calling all items in `a_list`?

Note that directly decorating f with `@tf.function` without any other changing to the above code is infeasible as it will always return the same result. Also, I know that this can be achieved by calling all elements in `a_list` first and then index them using `tf.gather_nd`. But this will incur a large amount of overhead if calling an object of type A involves a deep neural network.