pytensor.tensor.cartesian does not work on TensorVariables

[AuguB]

Describe the issue:

pytensor.tensor.cartesian does not work on TensorVariables

Reproduceable code example:

a = pm.Normal("a", shape=(7,))
b = pm.Normal("b", shape=(2,7))

pm.math.cartesian(a, b).shape.eval()
# Throws an error
# I need [7,2,7]

pytensor.tensor.mul(a,b).shape()
# [2,7] <- not what I need

Error message:

Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "/opt/anaconda3/envs/dev_refactor_2/lib/python3.12/site-packages/pymc/math.py", line 214, in cartesian
    arrays_integer = [np.arange(len(x)) for x in arrays_2d]
                                ^^^^^^
TypeError: len() of unsized object

PyMC version information:

5.19.1 conda

Context for the issue:

I don't know if there is another way to do this, but I have two TensorVariables with possibly matching dimension lengths, and I need the cartesian product of them. I can't seem to find a way to let the matching dimensions be combined.

[welcome]

]
🎉 Welcome to PyMC! 🎉 We're really excited to have your input into the project! 💖

If you haven't done so already, please make sure you check out our Contributing Guidelines and Code of Conduct.

[AuguB]

I came up with a workaround, but if there are more elegant soltions I am very happy to hear them:

with pm.Model() as model:
    a = pm.Normal("a", 0, 1, shape=(3, 4))
    b = pm.Normal("b", 0, 1, shape=(3))

    a_nones = len(a.shape.eval()) * [None]
    b_nones = len(b.shape.eval()) * [None]

    c = pt.tensor.mul(a[*b_nones], b[..., *a_nones])
    print(c.shape.eval())
-[3,3,4]

[ricardoV94]

pt.mul.outer(a, b), or pt.mul.outer(b, a), not sure about the order from the top of my mind