@_get_vector_length.register(TensorSharedVariable)
def _get_vector_length_TensorSharedVariable(var_inst, var):
    return len(var.get_value(borrow=True))

def get_vector_length(v: TensorLike) -> int:
    """Return the run-time length of a symbolic vector, when possible.

    Parameters
    ----------
    v
        A rank-1 `TensorType` variable.

    Raises
    ------
    TypeError
        `v` hasn't the proper type.
    ValueError
        No special case applies, the length is not known.
        In general this is not possible, but for a number of special cases
        the length can be determined at compile / graph-construction time.
        This function implements these special cases.

    """
    v = as_tensor_variable(v)
    
    if v.type.ndim != 1:
        raise TypeError(f"Argument must be a vector; got {v.type}")

    static_shape: int | None = v.type.shape[0]
    if static_shape is not None:
        return static_shape

    return _get_vector_length(getattr(v.owner, "op", v), v)

import pytensor
import numpy as np
from pytensor import tensor as pt

val = np.zeros(3, dtype='float32')
x = pytensor.shared(val, name='x', shape = val.shape, strict=True)
print(pt.get_vector_length(x)) # returns 3

pt.get_vector_length(pytensor.shared(val))# Raises ValueError : "ValueError: Length of <Vector(float32, shape=(?,))> cannot be determined"