-
Notifications
You must be signed in to change notification settings - Fork 1
/
create_plots.py
159 lines (112 loc) · 4.67 KB
/
create_plots.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
import argparse
from pathlib import Path
import numpy as np
from metric_rl.utils import plot_data
import matplotlib.pyplot as plt
from contextlib import contextmanager
def create_figures(results_dir, env, data_dict, subfolder=None, use_median=False, display=False, legend_plot='R'):
for name in ['J', 'R', 'E']:
plot_legend = name == legend_plot
create_figure(results_dir, name, env, data_dict, subfolder=subfolder, use_median=use_median,
legend=plot_legend, display=display)
def create_figure(results_dir, name, env, data_dict, subfolder=None, use_median=False, legend=False, display=False):
fig, ax = plt.subplots()
legend_lab = []
for n, d in data_dict.items():
plot_data(ax, d[name], use_median)
legend_lab.append(n)
plt.grid(linestyle='dotted')
if legend:
plt.legend(legend_lab, fontsize='small', frameon=False)
plt.xlabel('Epoch')
plt.ylabel(name)
fig_name = name + '_' + env
fig_path = results_dir / 'plots'
if subfolder:
fig_path = fig_path / subfolder
fig_path.mkdir(exist_ok=True)
full_path = fig_path / (fig_name + '.png')
plt.savefig(full_path, bbox_inches='tight')
if display:
plt.suptitle(env)
plt.show()
plt.close(fig)
def load_data_metricrl(results_dir, env, n_seeds):
env_subfolder = 'env_id_' + env
env_dir = results_dir / env_subfolder
results_dict = dict()
for clusters_dir in sorted(env_dir.iterdir()):
n_clusters = clusters_dir.name.split('_')[-1]
data_dir = clusters_dir / 'MetricRL'
J, R, E = load_data(data_dir, n_seeds)
results_dict[f'MetricRL-{n_clusters}'] = dict(J=J, R=R, E=E)
return results_dict
@contextmanager
def ignore_missing_file():
try:
yield
except FileNotFoundError as e:
print(e)
return dict()
def load_data_metricrl_diff(results_dir, env, n_seeds):
env_subfolder = 'env_id_' + env
env_dir = results_dir / env_subfolder
results_dict = dict()
for alg_dir in sorted(env_dir.iterdir()):
alg_name = alg_dir.name.split('_')[-1]
for nb_centers_dir in sorted(alg_dir.iterdir()):
nb_centers = nb_centers_dir.name.split('_')[-1]
for init_cluster_noise_dir in sorted(nb_centers_dir.iterdir()):
init_cluster_noise = init_cluster_noise_dir.name.split('_')[-1]
data_dir = init_cluster_noise_dir / 'MetricRLDiff'
J, R, E = load_data(data_dir, n_seeds)
results_dict[f'MetricRLDiff-{alg_name}-{nb_centers}-{init_cluster_noise}'] = dict(J=J, R=R, E=E)
return results_dict
def load_data_deep_baselines(results_dir, env, n_seeds):
env_subfolder = 'env_id_' + env
env_dir = results_dir / env_subfolder
results_dict = dict()
for alg_dir in sorted(env_dir.iterdir()):
alg_name = alg_dir.name.split('_')[-1]
data_dir = alg_dir / alg_name
J, R, E = load_data(data_dir, n_seeds)
results_dict[alg_name] = dict(J=J, R=R, E=E)
return results_dict
def load_data(data_dir, n_seeds):
J_list = list()
R_list = list()
E_list = list()
for seed in range(n_seeds):
try:
J = np.load(data_dir / f'J-{seed}.npy')
R = np.load(data_dir / f'R-{seed}.npy')
E = np.load(data_dir / f'E-{seed}.npy')
J_list.append(J)
R_list.append(R)
E_list.append(E)
except FileNotFoundError as e:
print(e)
pass
return np.array(J_list), np.array(R_list), np.array(E_list)
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument('--envs', type=str, nargs='+')
parser.add_argument('--results-dir', type=str, default='logs')
parser.add_argument('--n-seeds', type=int, default=25)
parser.add_argument('--display', action='store_true')
args = parser.parse_args()
results_dir = Path(args.results_dir)
bullet_envs = ['HopperBulletEnv-v0', 'Walker2DBulletEnv-v0', 'HalfCheetahBulletEnv-v0', 'AntBulletEnv-v0']
envs = args.envs
if 'bullet' in envs:
envs.remove('bullet')
envs += bullet_envs
for env in envs:
data_dict = dict()
with ignore_missing_file():
data_dict.update(**load_data_metricrl(results_dir / 'metricrl', env, args.n_seeds))
with ignore_missing_file():
data_dict.update(**load_data_metricrl_diff(results_dir / 'metricrl_diff', env, args.n_seeds))
with ignore_missing_file():
data_dict.update(**load_data_deep_baselines(results_dir / 'deep_baselines', env, args.n_seeds))
create_figures(results_dir, env, data_dict, subfolder=None, display=args.display)