Vectorization of RandomVariables failing

[ricardoV94]

Description

from pytensor.graph.replace import vectorize_graph
import pytensor.tensor as pt

mu = pt.tensor("mu", shape=(5, 3))
sigma = pt.tensor("sigma", shape=())
x = pt.random.normal(mu, sigma)

new_x = vectorize_graph(
    x,
    replace={
        mu: pt.tensor("mu", shape=(10, 5, 3)),
        sigma: pt.tensor("sigma", shape=(10,)),
    }
)

ValueError: Could not broadcast dimensions. Incompatible shapes were [(ScalarConstant(ScalarType(int64), data=10), ScalarConstant(ScalarType(int64), data=5), ScalarConstant(ScalarType(int64), data=3)), (ScalarConstant(ScalarType(int64), data=1), ScalarConstant(ScalarType(int64), data=1), ScalarConstant(ScalarType(int64), data=10))].

With size we get another error

x = pt.random.normal(mu, sigma, size=mu.shape)
new_x = vectorize_graph(
    x,
    replace={
        mu: pt.tensor("mu", shape=(10, 5, 3)),
        sigma: pt.tensor("sigma", shape=(10,)),
    }
)

ValueError: Tensor of type Matrix(int64, shape=(10, 2)) could not be cast to have 1 dimensions