Fix e_string_to_ground_state for type A2 even dual

[tscrim]

The ground state for type A_2n⁽²⁾⁺ should be based on \Lambda_n rather than \Lambda_0 due to the equivalence with type A_2n⁽²⁾. Since this is based on \Lambda_0, we have:

sage: K = crystals.KirillovReshetikhin(CartanType(['A',6,2]).dual(), 1,1)
sage: T = tensor([K,K,K])
sage: hw = [x for x in T if x.is_highest_weight([1,2,3])]
sage: for elt in hw:
....:     elt.energy_function()
....:     elt.energy_function('definition')
....:     print ""
0
0

1
2

2
2

1
1

2
3

2
3

1
1

[tscrim]

New commits:

​f2edd24

Fix energy function for A2 even dual.

[aschilling]

Looks good! For A_{2n}^{{(2)} dual there was indeed a bug in the path to the ground state beforehand.}