-
Notifications
You must be signed in to change notification settings - Fork 5
/
tests_validation.py
188 lines (155 loc) · 6.91 KB
/
tests_validation.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
import os
from validation import *
from multi_agents import SimpleAgent, AlphaBetaAgent, SimpleMCTSAgent, \
StochasticSimpleMCTSAgent, PureMCTSAgent
from compare_agents import simple_func_names, simple_agent_names, \
ab_evaluation_func_names, ab_evaluation_agent_names
import numpy as np
import matplotlib.pyplot as plt
from time import perf_counter
def test_snapshot():
file = "bridge_data_1/Aulpll23.pbn"
# file = "bridge_data_1/Dkofro20.pbn"
files = Parser([file])
# out = files.games[0].snapshot(0, POSITIONS[0])
# out = files.games[0].snapshot(0, POSITIONS[2])
# out = files.games[0].snapshot(1, POSITIONS[2])
out = files.games[0].snapshot(1, POSITIONS[0])
# out = games[0].snapshot(6, POSITIONS[0])
return out
def test_all_snapshots_single_game():
file = "bridge_data_1/Aulpll23.pbn"
games = Parser([file])
# out = games.games[0].all_relevant_snapshots()
# out = games.games[2].all_relevant_snapshots()
# out = games.games[4].all_relevant_snapshots()
out = games.games[6].all_relevant_snapshots()
return out
def load_all_game():
games_dir = os.path.join(os.getcwd(), 'bridge_data_1')
files_name = os.listdir(games_dir)
pbn_files = []
for file in files_name:
if file[-3:] == 'pbn':
pbn_files.append(os.path.join('bridge_data_1', file))
files = Parser(np.random.permutation(pbn_files).tolist())
print("succeed loading")
return files
def test_valid_games():
games = load_all_game()
for i, g in enumerate(games.games):
trick_idx = max(0, len(g.tricks) - 1)
g.snapshot(trick_idx, POSITIONS[0])
print("all are valid")
def test_all_snapshots():
games = load_all_game()
for i, g in enumerate(games.games):
g.all_relevant_snapshots()
print("all snapshots for all games done")
return
def games_length():
games = load_all_game()
histogram = np.zeros(12)
for g in games.games:
histogram[len(g.tricks) - 1] += 1
print(histogram)
plt.bar(np.arange(12) + 1, histogram)
plt.title("Number of Tricks Histogram")
plt.xlabel("# tricks")
plt.ylabel("# games")
plt.show()
def test_validate_agent_action():
curr_agent = SimpleAgent('highest_first_action')
file = "bridge_data_1/Aulpll23.pbn"
files = Parser([file])
return validate_agent_action(dg=files.games[4],
trick_idx=1,
position=PositionEnum.N,
agent=curr_agent)
def test_validate_agent_by_the_whole_game():
curr_agent = SimpleAgent('highest_first_action')
file = "bridge_data_1/Aulpll23.pbn"
files = Parser([file])
return validate_agent_per_data_game(agent=curr_agent, dg=files.games[4])
def display_results(agents_scores: List[np.ndarray], names: List[str],
num_of_games: int, min_tricks_num: int=0):
labels = [i for i in range(13 - min_tricks_num, 1, -1)]
x = np.arange(len(labels)) # the label locations
num_agents = len(agents_scores)
width = 1 / (num_agents+2) # the width of the bars
fig, ax = plt.subplots()
rects_list = []
for i in range(num_agents):
rects_list.append(ax.bar(x + ((i - ((num_agents-1) / 2)) * width),
agents_scores[i], width, label=names[i]))
ax.set_ylabel('Match success rate (%)')
ax.set_xlabel('Hand size (cards)')
ax.set_title(f'Scores by trick and agent\n (data from {num_of_games} games)')
ax.set_title(f'Prediction success rate as func. of # cards in hand\naveraged over {num_of_games} games')
ax.set_xticks(x)
ax.set_xticklabels(labels)
ax.legend(loc='best')
ax.set_ylim([0, 119])
fig.tight_layout()
plt.show()
def run_validation(agents_list: List[IAgent], num_of_games: int,
min_tricks: int=0) -> List[np.ndarray]:
"""
:param agents_list:
:param num_of_games: if 0: run all games
:param min_tricks: trick index to start validation from (0 for start of game)
:return: list of arrays, one for every agent. the array is consist of 12
entries, each represents the score of agent respectivly to the number
of remaining cards in its hand.
"""
agents_scores = [np.zeros(12) for _ in agents_list]
games = load_all_game()
if num_of_games == 0:
num_of_games = len(games.games)
for agent_idx, agent in enumerate(agents_list):
start = perf_counter()
all_checks, all_succeeds = np.zeros(12), np.zeros(12)
for i in range(num_of_games):
curr_checks, curr_succeeds = validate_agent_per_data_game(
agent, games.games[i], min_tricks)
all_checks += curr_checks
all_succeeds += curr_succeeds
if (i+1) % 25 == 0:
print(f"Finished {i+1}/{num_of_games} games for agent {agent.__class__.__name__}")
agents_scores[agent_idx] = (all_succeeds / all_checks) * 100
print(f"agent {agent_idx}: {(perf_counter() - start)/num_of_games} seconds/game, {num_of_games} games")
print(agents_scores)
return agents_scores
if __name__ == '__main__':
### Tests
# test_snapshot()
# test_all_snapshots_single_game()
# test_validate_agent_action()
# test_load_all_game()
# test_valid_games()
# test_all_snapshots()
# games_length()
# test_validate_agent_by_the_whole_game()
num_of_games = 1 # how many data_games to use for validation
minimum_round = 8 # trick index to start validation from (0 for start of game)
### Run Simple agents
# agents_list = [SimpleAgent(kind) for kind in simple_func_names]
# agents_scores_list = run_validation(agents_list, num_of_games, min_tricks_num)
# display_results(agents_scores_list, simple_agent_names, num_of_games)
### Run AB agents
# agents_list = [AlphaBetaAgent(kind, depth=12) for kind in ab_evaluation_func_names]
# agents_scores_list = run_validation(agents_list, num_of_games, min_tricks_num)
# display_results([agent[min_tricks_num:] for agent in agents_scores_list],
# ab_evaluation_agent_names, num_of_games, min_tricks_num)
### Run MTCs (choose one agents_list every time not to be a comment)
num_of_simulations = 10
agents_list = [SimpleMCTSAgent(kind, num_simulations=num_of_simulations)
for kind in simple_func_names]
agent_names = [f"{agent_cls.__class__.__name__}({func_name})" for agent_cls, func_name in zip(agents_list, simple_func_names)]
# agents_list = [StochasticSimpleMCTSAgent(kind, num_simulations=num_of_simulations)
# for kind in simple_func_names]
# agents_list = [PureMCTSAgent(kind, num_simulations=num_of_simulations)
# for kind in simple_func_names]
agents_scores_list = run_validation(agents_list, num_of_games, minimum_round)
display_results([agent[minimum_round:] for agent in agents_scores_list],
agent_names, num_of_games, minimum_round)