indexes() might return an empty float64 array

This is the current code in indexes

    # check if the signal is totally flat
    if len(zeros) == len(y) - 1:
        return np.array([])

It should be fixed to be

    # check if the signal is totally flat
    if len(zeros) == len(y) - 1:
        return np.array([], dtype=np.int64)

Default for an numpy array is float64, so til will raise an error for anyone trying to use this as an index, which is what this is for

Comments (2)