chore: handle new values of loops argument of degree() and max_degree()

[maelle]

No description provided.

[schochastics]

2.2.0?

[szhorvat]

LGTM. Note that I did not take a careful look at the code, I just checked if it seemed to do what I expected.

I have one comment: I am not sure whether TRUE/FALSE should be deprecated. Perhaps it's not bad to keep allowing these, with the current interpretation (TRUE = twice, FALSE = none). I'll leave the decision to you.

Some functions do have a different kind of "loops" parameter which only takes TRUE/FALSE. Perhaps it's nicer to stay a bit more consistent with those functions. To be clear, for most of these functions twice/once/none wouldn't even make sense, which is why they stay with TRUE/FALSE. An example is some graph generators. Twice/once/none is mostly used with functions that convert a graph to a different representation (adjacency matrix, adjacency list) or compute a quantity that is affected by this interpretation (degree is probably the only one).

I don't personally have a strong opinion whether TRUE/FALSE should be deprecated or kept supported.

[maelle]

Thanks both!

@szhorvat to me it's "cleaner" when an argument only takes one kind of values although I know there are plenty of well justified exceptions to that.