Performance: Faster network.copy(), perhaps using copy.deepcopy

[connortann]

At present, copying a network with pypsa.Network().copy() takes a surprisingly long time. For some workflows this actually seems to be a performance bottleneck.

I noticed that using the builtin copy.deepcopy seems to do the same job much faster:

import pypsa
import copy

n = pypsa.examples.ac_dc_meshed()
n2 = n.copy()  # takes about 158ms
n3 = copy.deepcopy(n)  # takes about 5 ms

Would this be a suitable alternative implementation? Are there any issues with using copy.deepcopy?

I can look into making a PR if that would be of interest.