Remove restrictions on operations; Numpy already has them built in

[298A-E9E3]

This will allow for all operations to function on scalars and vectors

[ikz87]

I'd prefer the error handling to be at the level of the library itself. NumPy would show ValueError: operands could not be broadcast together with shapes (2,) (3,) which might not be good devex. I guess the double checking on vector sizes could be avoided, but at least a try/catch block should be added to change the error message coming from NumPy to something like ValueError("Vectors must have the same size")

The checks using the unions are a bit less verbose but would crash on python versions 3.6 to 3.9, I'd rather have the compatibility.

The addition of toList() is actually desirable.

Apply those changes and I'd gladly merge