Tseitin over multigraphs

[marcvinyals]

Should we allow multigraphs as input for Tseitin formulas? Currently we do not, but the only (technical) reason is that we do not know how to label multiedges. I do not think this issue has a high priority, so I would suggest to leave the behaviour undefined (i.e. skip tests) but leave the issue open as a reminder and address it later.