Here it is, ticket number 12155. http://trac.sagemath.org/sage_trac/ticket/12155
> Yes, it would appear that the generic complement() method is adding > edges to a bipartite graph object and the error is coming from the > bipartite graph routines complaining about an illegitimate edge. I > would imagine that temporarily you could convert your bipartite graph > to a generic graph before taking the complement. > > Long-term, maybe the bipartite graph object should implement its own > complement, which creates a generic graph and then calls the generic > complement on that? (Though I have not studied the situation too > carefully.) Can you make a Trac ticket and point back to this > discussion? It appears that BipartiteGraph (bipartite_graph.py) is using the complement method from Generic graph (generic_graph.py), which when constructing the complement G of self on line 11933 does G=copy(self) Maybe here is where the bipartite property gets transferred to G. If my counting is right, there are only 8 bipartite graphs whose complements are also bipartite (and each of them has at most 4 vertices). Would it be useful to preserve the bipartiteness property when taking complement in any of these instances? If so, maybe implement this in the BipartiteGraph complement method. I am also thinking that the complement method in the BipartiteGraph could just do something like return Graph(self.edges()).complement() Thanks, Fidel -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
