[Nauty] Canonical representations for annotated graphs

[Falk Hueffner]

Hi,

I have an algorithm that includes a search tree over annotated
graphs. To recognize graphs already encountered, I want to convert
them to a canonical form so I can look them up in a hash table.

In particular, I have

* graphs with 1 or 2 extra bits per vertex,
* graphs with 1 or 2 extra bits per edge.

1 extra bit per vertex can easily be modeled with a self loop. Is
there any possibility for the other options? The graphs are rather
small, < 32 nodes.

        Falk