[RTR] multinode penalties with OdeSolver.COLLOCATION (inclusion of cx_intermediates)

[EveCharbie]

This change is 

[pariterre]

Reviewed 8 of 8 files at r1, all commit messages.
Reviewable status: all files reviewed, 3 unresolved discussions (waiting on @EveCharbie)

---

bioptim/limits/penalty_helpers.py line 86 at r1 (raw file):

            idx = 0
            for phase, node, sub in zip(phases, nodes, subnodes):
                if (

Can you explain the actual case in a comment?

---

bioptim/limits/multinode_penalty.py line 286 at r1 (raw file):

                raise RuntimeError("This continuity function is only valid for 2 nodes")

            algebraic_states_end_interval = controllers[0].algebraic_states[key].cx_intermediates_list[-1]

Add a comment here to remind us why it makes sense not to use cx_end (as of cx_start)

---

bioptim/limits/penalty_option.py line 705 at r1 (raw file):

            if sn_idx.stop == 1:
                pass
            elif sn_idx.stop is None:

Typically None goes up to the actual last while -1 is the one before to last. This mean they could be reversed?

[EveCharbie]

@pariterre could you rereview this offline ? (I only changed the slicing with -1 and None to try to make it more intuitive)

[pariterre]

Reviewed 5 of 7 files at r2, 2 of 2 files at r3, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @EveCharbie)