As reported in Issue #2, many VisibleDeprecationWarnings were raised
when running the tests of ltfatpy with a recent version of numpy.

This is now solved by insuring that values used for indexes in numpy
arrays are integers, either by using integer division (//) where needed,
or by explicitely converting to integer type (changing np.func(X/2) to
int(np.func(X/2))), or by using int as the array dtype.