Does bct.efficiency_wei also work for directed networks?

[JohannesWiesner]

I would like to compute the global efficiency for a set of directed matrices. Looking at the appendix of this this paper, it seems that bct.efficiency_wei is agnostic to direction (aka. can be used for both directed and undirected networks?). If so, I guess the documentation should be changed to undirected/directed weighted connection matrix?:

bctpy/bct/algorithms/distance.py

Lines 560 to 562 in afac15f

	W : NxN np.ndarray
	undirected weighted connection matrix
	(all weights in W must be between 0 and 1)

[aestrivex]

I never thought of doing it for directed networks. It appears to me that the algorithm could accommodate directed matrices. However, that algorithm is a bit more complicated than the one currently written, it would have to compute distances between pairs of vertices directionally. A bit of work will be needed to make it do that.

Pull requests are always welcome.