`params_broadcast_shapes` logic is wrong

[ricardoV94]

This helper uses max to broadcast dimensions, but this is wrong because a dim of length 1 should broadcasted to length 0 if that's the length of the other dim

pytensor/pytensor/tensor/random/utils.py

Lines 25 to 71 in 9cf2d18

	def params_broadcast_shapes(param_shapes, ndims_params, use_pytensor=True):
	"""Broadcast parameters that have different dimensions.
	Parameters
	==========
	param_shapes : list of ndarray or Variable
	The shapes of each parameters to broadcast.
	ndims_params : list of int
	The expected number of dimensions for each element in `params`.
	use_pytensor : bool
	If ``True``, use PyTensor `Op`; otherwise, use NumPy.
	Returns
	=======
	bcast_shapes : list of ndarray
	The broadcasted values of `params`.
	"""
	max_fn = maximum if use_pytensor else max
	rev_extra_dims = []
	for ndim_param, param_shape in zip(ndims_params, param_shapes):
	# We need this in order to use `len`
	param_shape = tuple(param_shape)
	extras = tuple(param_shape[: (len(param_shape) - ndim_param)])
	def max_bcast(x, y):
	if getattr(x, "value", x) == 1:
	return y
	if getattr(y, "value", y) == 1:
	return x
	return max_fn(x, y)
	rev_extra_dims = [
	max_bcast(a, b)
	for a, b in zip_longest(reversed(extras), rev_extra_dims, fillvalue=1)
	]
	extra_dims = tuple(reversed(rev_extra_dims))
	bcast_shapes = [
	(extra_dims + tuple(param_shape)[-ndim_param:])
	if ndim_param > 0
	else extra_dims
	for ndim_param, param_shape in zip(ndims_params, param_shapes)
	]
	return bcast_shapes

[ricardoV94]

This is not relevant if we want to forbid runtime broadcasting on RVs as well