32 bit integer limitation

[ZachAJohnson]

Every int specified here, presumably based on underlying Fortran type is _np.int32. The problem is, I, and probably many others, frequently work with tensors whose flattened tt.core attribute has size much in excess of the limited 2^31 signed integer representation. When using for example the .to_list(tensor_train) functionality, any tensor train whose core is larger than 2^31 throws a ValueError: cannot reshape array of size 0 into shape <actual_shape_of_tensor_core> because the tt.ps indices denoting the tensor train core boundary locations exceed the int32 limit.