Hello Nathann, > Well, for the slooooooooooooow relabeling there is a ready answer : #14000. > It's a patch I wrote for Christian during the Sage days. You can apply it, > and (even better) review it !
It is reviewed! > With this it should be almost instantaneous. Or at least the time will be > spent smartly, i.e. on the relabeling and not on checking the input of > relabel() :-P Is it also possible to create the relabelled digraph from the beginning, rather than creating it first and then doing the relabelling? > 2) The way you look for common subgraphs wastes *a lot* of computations. Not > using the information that you already have (that the graphs up to a given > level are isomorphic) is almost criminal ! :-P > You're saved by the efficiency of the graph isomorphism test in this case, > but it really isn't the best way to do it... You would save much, much more > time by thinking of a dedicated way to test > isomorphism in your graphs, or by writing another version of your code that > uses this information. Hence if you want a better performance, I think the > most important is to improve how this algorithm > works. And perhaps its implementation, later. Yes, in my case it is definitely better to store the intermediate computations since they already give most of the isomorphism! Best, Anne -- You received this message because you are subscribed to the Google Groups "sage-combinat-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-combinat-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-combinat-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-combinat-devel?hl=en. For more options, visit https://groups.google.com/groups/opt_out.
