kaggle-environments 0.2.1__py3-none-any.whl → 1.20.1__py3-none-any.whl

kaggle_environments/envs/football/test_football.py

@@ -0,0 +1,360 @@
+env = None
+
+# Temporarily disable these tests until we can fix the gfootball env
+"""
+def before_each(state=None, configuration=None, info=None):
+    if info is None:
+        info = {}
+    global env
+    steps = [] if state == None else [state]
+    env = make("football", steps=steps,
+               configuration=configuration, info=info, debug=True)
+
+
+def test_to_json():
+    before_each()
+    json = env.toJSON()
+    assert json["name"] == "football"
+    assert json["rewards"] == [0.0, 0.0]
+    assert json["statuses"] == ["ACTIVE", "ACTIVE"]
+    assert json["specification"]["reward"]["type"] == ["number", "null"]
+
+def clear_players_raw(state):
+  state = copy.deepcopy(state)
+  for entry in state:
+    entry.observation.players_raw = []
+  return state
+
+
+@helpers.human_readable_agent
+def human_readable_agent(obs):
+    if Action.Sprint not in obs['sticky_actions']:
+        return Action.Sprint
+    controlled_player_pos = obs['left_team'][obs['active']]
+    if obs['ball_owned_player'] == obs['active'] and obs['ball_owned_team'] == 0:
+        if controlled_player_pos[0] > 0.5:
+            return Action.Shot
+        return Action.Right
+    else:
+        return Action.Slide
+
+
+def test_single_agent():
+    before_each(configuration={"team_1": 1, "team_2": 0, "scenario_name": "11_vs_11_stochastic"})
+    x = env.reset()
+
+    assert clear_players_raw(env.reset()) == [
+        {
+            "action": [],
+            "status": "ACTIVE",
+            "reward": 0,
+            "info": {},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 1,
+                "players_raw": [],
+                "step": 0
+            }
+        },
+        {
+            "action": [],
+            "status": "ACTIVE",
+            "reward": 0,
+            "info": {},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 0,
+                "players_raw": []
+            }
+        }
+    ]
+
+    # Correct step from agent 0.
+    assert clear_players_raw(env.step([[0],[]])) == [
+        {
+            "action": [0],
+            "status": "ACTIVE",
+            "reward": 0,
+            "info": {},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 1,
+                "players_raw": [],
+                "step": 1
+            }
+        },
+        {
+            "action": [],
+            "status": "ACTIVE",
+            "reward": 0,
+            "info": {},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 0,
+                "players_raw": []
+            }
+        }
+    ]
+
+    # Incorrect step from agent 1 (it is forbidden to act in this scenario,
+    # as 'team_2' players is set to 0).
+    assert clear_players_raw(env.step([[0],[1]])) == [
+        {
+            "action": [0],
+            "status": "DONE",
+            "reward": 100,
+            'info': {'debug_info': 'Opponent forfeited. You win.'},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 1,
+                "players_raw": [],
+                "step": 2
+            }
+        },
+        {
+            "action": [1],
+            "status": "INVALID",
+            "reward": None,
+            'info': {'debug_info': 'Invalid number of actions provided: Expected 0, got 1.'},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 0,
+                "players_raw": []
+            }
+        }
+    ]
+
+    # Incorrect step from agent 1 (out of range).
+    before_each(configuration={"team_1": 1, "team_2": 0, "scenario_name": "11_vs_11_stochastic", "save_video": True})
+    x = env.reset()
+    assert clear_players_raw(env.step([[100],[]])) == [
+        {
+            "action": [100],
+            "status": "INVALID",
+            "reward": None,
+            'info': {'debug_info': 'Invalid action provided: [100].'},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 1,
+                "players_raw": [],
+                "step": 1
+            }
+        },
+        {
+            "action": [],
+            "status": "DONE",
+            "reward": 100,
+            'info': {'debug_info': 'Opponent forfeited. You win.'},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 0,
+                "players_raw": []
+            }
+        }
+    ]
+    # We can render even an "empty" episode...
+    env.render(mode="human", width=800, height=600)
+
+    # Incorrect step from agent 1 (not a list).
+    before_each(configuration={"team_1": 1, "team_2": 0, "scenario_name": "11_vs_11_stochastic"})
+    x = env.reset()
+    assert clear_players_raw(env.step([100, []])) == [
+        {
+            "action": [],
+            "status": "INVALID",
+            "reward": None,
+            'info': {'debug_info': 'Invalid number of actions provided: Expected 1, got 0.'},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 1,
+                "players_raw": [],
+                "step": 1
+            }
+        },
+        {
+            "action": [],
+            "status": "DONE",
+            "reward": 100,
+            'info': {'debug_info': 'Opponent forfeited. You win.'},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 0,
+                "players_raw": []
+            }
+        }
+    ]
+
+
+def test_multi_agent():
+    before_each(configuration={"team_1": 2, "team_2": 1, "scenario_name": "11_vs_11_stochastic"})
+    x = env.reset()
+
+    assert clear_players_raw(env.reset()) == [
+        {
+            "action": [],
+            "status": "ACTIVE",
+            "reward": 0,
+            "info": {},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 2,
+                "players_raw": [],
+                "step": 0
+            }
+        },
+        {
+            "action": [],
+            "status": "ACTIVE",
+            "reward": 0,
+            "info": {},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 1,
+                "players_raw": []
+            }
+        }
+    ]
+
+    # Correct step from both agents.
+    assert clear_players_raw(env.step([[0, 2],[4]])) == [
+        {
+            "action": [0, 2],
+            "status": "ACTIVE",
+            "reward": 0,
+            "info": {},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 2,
+                "players_raw": [],
+                "step": 1
+            }
+        },
+        {
+            "action": [4],
+            "status": "ACTIVE",
+            "reward": 0,
+            "info": {},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 1,
+                "players_raw": []
+            }
+        }
+    ]
+
+    # Incorrect step from agent 1  - too many actions passed.
+    assert clear_players_raw(env.step([[0, 1, 2], [1]])) == [
+        {
+            "action": [0, 1, 2],
+            "status": "INVALID",
+            "reward": None,
+            'info': {'debug_info': 'Invalid number of actions provided: Expected 2, got 3.'},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 2,
+                "players_raw": [],
+                "step": 2
+            }
+        },
+        {
+            "action": [1],
+            "status": "DONE",
+            "reward": 100,
+            'info': {'debug_info': 'Opponent forfeited. You win.'},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 1,
+                "players_raw": []
+            }
+        }
+    ]
+
+def test_deadline():
+    before_each(configuration={"team_1": 1, "team_2": 0, "scenario_name": "11_vs_11_stochastic"})
+    x = env.reset()
+
+    assert clear_players_raw(env.reset()) == [
+        {
+            "action": [],
+            "status": "ACTIVE",
+            "reward": 0,
+            "info": {},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 1,
+                "players_raw": [],
+                "step": 0
+            }
+        },
+        {
+            "action": [],
+            "status": "ACTIVE",
+            "reward": 0,
+            "info": {},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 0,
+                "players_raw": []
+            }
+        }
+    ]
+
+    # Correct step from agent 0.
+    assert clear_players_raw(env.step([DeadlineExceeded(),[]])) == [
+        {
+            "action": None,
+            "status": "TIMEOUT",
+            "reward": None,
+            "info": {},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 1,
+                "players_raw": [],
+                "step": 1
+            }
+        },
+        {
+            "action": [],
+            "status": "DONE",
+            "reward": 100,
+            "info": {'debug_info': 'Opponent forfeited. You win.'},
+            "observation": {
+                "remainingOverageTime": 60,
+                "controlled_players": 0,
+                "players_raw": []
+            }
+        }
+    ]
+
+
+def test_render():
+    video_file = "/tmp/video.webm"
+    before_each(configuration={"team_1": 1, "team_2": 1, "scenario_name": "tests.penalty", "save_video": True},
+        info={"LiveVideoPath": video_file})
+    env.step([[0],[0]])
+    env.render(mode="human", width=800, height=600)
+    output = env.step([[0],[0]])
+    env.render(mode="human", width=800, height=600)
+    while output[0]['status'] == 'ACTIVE':
+        output = env.step([[0],[0]])
+    assert output[0]['reward'] == 0
+    assert output[1]['reward'] == 0
+    assert output[0]['status'] == "DONE"
+    assert output[1]['status'] == "DONE"
+    assert os.path.isfile(video_file)
+
+
+def test_human_readable_agent():
+    before_each(configuration={"team_1": 1, "team_2": 1, "scenario_name": "tests.penalty"})
+    action = [0]
+    for _ in range(10):
+      obs = env.step([action, [0]])
+      action  = human_readable_agent(obs[0]['observation'])
+
+
+def test_score():
+    before_each(configuration={"team_1": 1, "team_2": 1, "scenario_name": "tests.goal_test", "save_video": True, "episode_number": 0})
+    res = env.run(["run_right", "run_right"])
+    assert res[-1][0]['reward'] == -1
+    assert res[-1][1]['reward'] == 1
+"""