I was adding a lot of edges on one of my graph, using
vertex() to get vertices and my code was kind of slow.
I did some benchmarking and found that a lot of time was spent inside the
vertex() function (used by my code and
add_edge()), especialy by calling
Since I wasen't using any filter, I just simplfied
def vertex(self, i, use_index=True): return libcore.get_vertex(weakref.ref(self), int(i))
This is an ugly hack, but I got x4 speedup.
Would be possible to add an option to disable any
*_vertex_filter call ? Eg. if the user didn't set a filter or with an option.
I can provide a merge request if needed :)