Enable random functions to return arrays

[enekomartinmartinez]

Currently random functions work returning a float

dim1: A,B,C~~|
var[dim1] = RANDOM UNIFORM(1,3) ~~|

returns a length 3 vector with different random numbers, however this is translated in PySD as:

@(["dim1"], _subscript_dict)
def var():
    return np.random.uniform(1, 3)

which will produce a unique random value and repeat it to generate a vector.

This should be done as:

@(["dim1"], _subscript_dict)
def var():
    return xr.DataArray(np.random.uniform(1, 3, size=(3,)), {"dim1": ["A", "B", "C"]}, ["dim1"])

I am working on it for a future release.