def test_can_update_mata_game(RPS_task):
psro = PSRONashResponse(task=RPS_task, benchmarking_episodes=2)
psro.menagerie = [rockAgent, paperAgent, scissorsAgent]
psro.meta_game = np.array([[0.5, 0], [1, 0.5]])
expected_meta_game = np.array([[0.5, 0, 1], [1, 0.5, 0], [0, 1, 0.5]])
actual_meta_game = psro.update_meta_game()
np.testing.assert_array_equal(expected_meta_game, actual_meta_game)
def test_can_fill_missing_game_entries_upon_adding_new_policy(RPS_task):
psro = PSRONashResponse(task=RPS_task, benchmarking_episodes=2)
psro.menagerie = [rockAgent, paperAgent, scissorsAgent]
meta_game = np.array([[0.5, 0, np.nan], [1, 0.5, np.nan],
[np.nan, np.nan, np.nan]])
expected_updated_metagame = np.array([[0.5, 0, 1], [1, 0.5, 0],
[0, 1, 0.5]])
actual_updated_metagame = psro.fill_meta_game_missing_entries(
policies=psro.menagerie,
updated_meta_game=meta_game,
benchmarking_episodes=psro.benchmarking_episodes,
task=RPS_task)
np.testing.assert_array_equal(expected_updated_metagame,
actual_updated_metagame)