Parallel bond querying

[hmedina]

In KappaComplex._edge_match(), parallel edges pose a problem with the function used: nx.edges(net, node) returns a list of (a,b) node identifier pairs, and so parallel bonds appear as duplicates in this list. The problem is when querying for bond data with net.get_edge_data(a,b), as this will return only the first of the parallel bonds.