Sparse TensorVariables are not using the Sparse Elemwise methods

Description

I think during the merging of SparseVariable and TensorVariables Type classes still in Aesara https://github.com/aesara-devs/aesara/pull/766, there was a mistake where the Structured methods were not borrowed, instead we always convert to dense with a warning

import numpy as np
import pytensor.tensor as pt
import pytensor.sparse as ps


X_sp = ps.csc_from_dense(pt.arange(5))
X_sp.sin()  # Method sin is not implemented for sparse variables. The variable will be converted to dense.

Either way, we have a sparse Sin implemented, so we should use them:

out = ps.sin(X_sp)
out.eval()  # <1x5 sparse matrix of type '<class 'numpy.float64'>'

https://github.com/pymc-devs/pytensor/blob/0b632bdb64b9a5f8e0fe8e98491ce2a6f2bca27f/pytensor/sparse/basic.py#L3028-L3051

Reproduce the sparse TensorVariable example from the issue, then inspect pytensor/sparse/basic.py at lines 3028-3051 and the Structured methods mentioned in the report. Compare X_sp.sin() with ps.sin(X_sp); done means the method uses the sparse implementation, returns a sparse result, and no longer warns about dense conversion.