Tidy up tests now that information set sorting is obsolete.

Author: tturocy

tests/games.py

@@ -5,7 +5,6 @@
 from abc import ABC, abstractmethod
 
 import numpy as np
-import pytest
 
 import pygambit as gbt
 
@@ -353,7 +352,6 @@ def create_3_player_with_internal_outcomes_efg(nonterm_outcomes: bool = False) -
         o = g.add_outcome([0, 0, 0])
         g.set_outcome(g.root.children[0].children[0].children[1].children[0], o)
         g.set_outcome(g.root.children[0].children[0].children[1].children[1], o)
-    # g.sort_infosets()
     return g
 
 
@@ -651,10 +649,6 @@ def create_kuhn_poker_efg(nonterm_outcomes: bool = False) -> gbt.Game:
         g = _create_kuhn_poker_efg_nonterm_outcomes()
     else:
         g = _create_kuhn_poker_efg_only_term_outcomes()
-
-    # Ensure infosets are in the same order as if game was written to efg and read back in
-    with pytest.warns(FutureWarning):
-        g.sort_infosets()
     return g
 
 
@@ -909,7 +903,6 @@ def create_STOC_simplified() -> gbt.Game:
         g.root.children[1].children[1],
         outcome=g.add_outcome(payoffs=[20, -20], label="d"),
     )
-    # g.sort_infosets()
     return g
 
 
@@ -949,7 +942,6 @@ def create_STOC_simplified2() -> gbt.Game:
         g.root.children[1].children[0],
         outcome=g.add_outcome(payoffs=[10, -10], label="c"),
     )
-    # g.sort_infosets()
     return g
 
 
@@ -1031,7 +1023,6 @@ def create_seq_form_STOC_paper_zero_sum_2_player_efg() -> gbt.Game:
     g.root.children[0].children[1].infoset.label = "01"
     g.root.children[2].children[0].infoset.label = "20"
     g.root.children[0].children[1].children[0].infoset.label = "010"
-    g.sort_infosets()
     return g
 
 

tests/test_nash.py

@@ -608,7 +608,6 @@ def test_lcp_behavior_rational(game: gbt.Game, mixed_behav_prof_data: list):
     using max_regret and agent_max_regret (internal consistency); and
     comparison to a previously computed equilibrium using this function (regression test).
     """
-    game.sort_infosets()
     result = gbt.nash.lcp_solve(game, use_strategic=False, rational=True)
     assert len(result.equilibria) == 1
     eq = result.equilibria[0]
@@ -786,15 +785,12 @@ def test_lp_behavior_rational(game: gbt.Game, mixed_behav_prof_data: list):
     using max_regret and agent_max_regret (internal consistency); and
     comparison to a previously computed equilibrium using this function (regression test).
     """
-    game.sort_infosets()
     result = gbt.nash.lp_solve(game, use_strategic=False, rational=True)
     assert len(result.equilibria) == 1
     eq = result.equilibria[0]
     assert eq.max_regret() == 0
     assert eq.agent_max_regret() == 0
     expected = game.mixed_behavior_profile(rational=True, data=mixed_behav_prof_data)
-    print(expected)
-    print(eq)
     assert eq == expected
 
 
@@ -975,11 +971,8 @@ def test_regrets_tmp2():
     ],
 )
 def test_repeat_max_regret(game: gbt.Game, mixed_behav_prof_data: list):
-    game.sort_infosets()
     profile1 = game.mixed_behavior_profile(rational=True, data=mixed_behav_prof_data)
     mr1 = profile1.max_regret()
     profile2 = game.mixed_behavior_profile(rational=True, data=mixed_behav_prof_data)
     mr2 = profile2.max_regret()
-    print()
-    print(mr1, mr2)
     assert mr1 == mr2