-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathinteraction_filter.py
159 lines (131 loc) · 5.98 KB
/
interaction_filter.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
# INT2: Interactive Trajectory Prediction at Intersections
# Published at ICCV 2023
# Written by Zhijie Yan
# All Rights Reserved
import pandas as pd
import numpy as np
from math import sqrt
import os
import pickle
from p_tqdm import p_map
import warnings
import argparse
warnings.filterwarnings("ignore")
from utils.interaction_utils import *
def parse_config():
parser = argparse.ArgumentParser(description='INT2 Dataset Interaction Filter Visualization.')
parser.add_argument('--scenario_path', '--s', type=str, default='int2_dataset_example/scenario/8/010213355106-010213364106.pickle',
help='The scenario path to be visualized')
parser.add_argument('--output_dir', type=str, default='int2_dataset_example/interaction_scenario/complete_scenario', help='')
args = parser.parse_args()
return args
def interaction_define(scenario_path, output_dir):
with open(scenario_path, 'rb+') as f:
scenario_info = pickle.load(f)
AGENT_INFO = scenario_info['AGENT_INFO']
object_id = AGENT_INFO['object_id']
object_type = AGENT_INFO['object_type']
object_sub_type = AGENT_INFO['object_sub_type']
state = AGENT_INFO['state']
interaction_info = {}
position_x = state['position_x']
position_y = state['position_y']
position_z = state['position_z']
theta = state['theta']
velocity_x = state['velocity_x']
velocity_y = state['velocity_y']
length = state['length']
width = state['width']
height = state['height']
valid = state['valid']
inter_info_dict = {}
inter_pair_info_dict = {}
inter_pair_index = 0
agent_num = valid.shape[0]
ir_indices_list = []
for i in range(0, agent_num - 1, 1):
if object_type[i] != 2:
continue
valid_i = valid[i].nonzero()[0]
if len(valid_i) < scenario_min_len:
continue
for j in range(i + 1, agent_num, 1):
valid_j = valid[j].nonzero()[0]
if len(valid_j) < scenario_min_len:
continue
coexistence_time = np.array([x for x in valid_i if x in valid_j])
if len(coexistence_time) < scenario_min_len:
continue
agent_i_x = position_x[i][coexistence_time]
agent_i_y = position_y[i][coexistence_time]
agent_i_w = width[i][coexistence_time]
agent_i_l = length[i][coexistence_time]
agent_i_t = theta[i][coexistence_time]
agent_i_vx = velocity_x[i][coexistence_time]
agent_i_vy = velocity_y[i][coexistence_time]
agent_i_info = np.stack([agent_i_x, agent_i_y, agent_i_w, agent_i_l, agent_i_t, agent_i_vx, agent_i_vy], axis=0)
agent_j_x = position_x[j][coexistence_time]
agent_j_y = position_y[j][coexistence_time]
agent_j_w = width[j][coexistence_time]
agent_j_l = length[j][coexistence_time]
agent_j_t = theta[j][coexistence_time]
agent_j_vx = velocity_x[j][coexistence_time]
agent_j_vy = velocity_y[j][coexistence_time]
agent_j_info = np.stack([agent_j_x, agent_j_y, agent_j_w, agent_j_l, agent_j_t, agent_j_vx, agent_j_vy], axis=0)
inter_is_ok, relation_type, interaction_time_valid = is_interaction_valid(i, j, agent_i_info, agent_j_info)
if inter_is_ok:
interaction_time_truth = np.array(coexistence_time)[interaction_time_valid]
if relation_type == 0:
influencer_id = i
reactor_id = j
else:
influencer_id = j
reactor_id = i
ir_indices_list.append([influencer_id, reactor_id])
inter_pair_info_dict[inter_pair_index] = {
'influencer_id': influencer_id,
'reactor_id': reactor_id,
'influencer_type': object_type[influencer_id],
'reactor_type': object_type[reactor_id],
'coexistence_time': coexistence_time,
'interaction_time': interaction_time_truth
}
inter_pair_index += 1
interested_agents = set()
for i in range(len(ir_indices_list)):
interested_agents.add(ir_indices_list[i][0])
interested_agents.add(ir_indices_list[i][1])
inter_info_dict['interaction_pair_info'] = inter_pair_info_dict
inter_info_dict['interested_agents'] = list(interested_agents)
output_path = os.path.join(output_dir, scenario_path.split('/')[-1])
scenario_info['INTERACTION_INFO'] = inter_info_dict
with open(output_path, 'wb+') as f:
f.write(pickle.dumps(scenario_info))
# print(output_path)
def single_process(scenario_path, output_floder):
error_scenario_path = 'error_scenario.txt'
try:
hdmap_id = scenario_path.split('/')[-2]
output_dir = os.path.join(output_floder, hdmap_id)
os.makedirs(output_dir, exist_ok=True)
interaction_define(scenario_path, output_dir)
except:
with open(error_scenario_path, 'a') as f:
f.write(scenario_path + '\n')
def multi_thread_process():
scenario_floder = 'int2_dataset/scenario'
output_floder = 'int2_dataset/interaction_scenario/complete_scenario'
scenario_dir_names = sorted(os.listdir(scenario_floder), key=lambda x: int(x))
for idx, scenario_id in enumerate(scenario_dir_names):
print(f'now are processed in {scenario_id}th')
scenario_files = [os.path.join(scenario_floder, scenario_id, f) for f
in os.listdir(os.path.join(scenario_floder, scenario_id))]
p_map(single_process, scenario_files, [output_floder] * len(scenario_files), num_cpus=0.2)
def main():
# multi_thread_process()
args = parse_config()
assert args.scenario_path != None
assert args.output_dir != None
single_process(args.scenario_path, args.output_dir)
if __name__ == "__main__":
main()