Generalize row/column index subsetting to any `Sequence[int]`

[LTLA]

On 0.3.6:

import biocframe
x=biocframe.BiocFrame({"A": [1,2,3]})
import numpy
x[numpy.array([0,2]),:] # fails
x[:,numpy.array([0,2])] # fails

There's no reason to restrict this to lists of integers if you can accept arbitrary sequences of integers.

(Also Sequence[bool]; replace the row indices with numpy.array([True, False, True]) and it still fails.