-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathpreprocess.py
executable file
·154 lines (127 loc) · 5.87 KB
/
preprocess.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
import numpy as np
from scipy.spatial.distance import directed_hausdorff
from polygon_generation import generate_polygon
from random_polygon_basis_encoder import RandomBasisEncoder, RandomSVDEncoder, RandomRotScaleEncoder
from tqdm import tqdm
def construct_vertices(number_vertices, dataset_size=100):
# irregs, spiks = [], []
vs = []
for _ in tqdm(range(dataset_size)):
irreg, spik = np.random.rand(), np.random.rand()
vertices = generate_polygon(center=(0, 0),
avg_radius=1,
irregularity=irreg,
spikiness=spik,
num_vertices=number_vertices)
v = np.array(vertices)
# irregs.append(irreg)
# spiks.append(spik)
vs.append(v)
# vertices = np.array(vs)
# return vertices
return vs
def construct_encodings(encoder, vertices):
dataset_size = len(vertices)
es = []
for i in tqdm(range(dataset_size)):
v = vertices[i]
e = encoder.encode(v)
es.append(e)
encodings = np.array(es)
return encodings
def construct_hausdorffs(vertices):
dataset_size = len(vertices)
hausdorffs = np.zeros((dataset_size, dataset_size))
for i, v1 in tqdm(enumerate(vertices), total=dataset_size):
for j in range(i, dataset_size):
v2 = vertices[j]
raw_hausdorff = directed_hausdorff(v1, v2)[0]
hausdorffs[i, j] = raw_hausdorff
hausdorffs[j, i] = raw_hausdorff
return hausdorffs
# idxs = []
# hausdorffs = []
# for i, v1 in tqdm(enumerate(vs), total=dataset_size):
# random_idx = np.random.choice(dataset_size, num_cross, replace=False)
# for j in random_idx:
# v2 = vs[j]
#
# idxs.append([i, j])
#
# raw_hausdorff = directed_hausdorff(v1, v2)[0]
# hausdorffs.append(raw_hausdorff)
#
# return vertices, encodings, idxs, hausdorffs
data_num_vertices= 5
basis_num_vertices= 5
train_size = 10000
test_size = 1000
# print("Data Preparation:")
# print("Constructing Training Vertices...\n")
# train_vertices = construct_vertices(data_num_vertices, train_size)
# print("Constructing Training Hausdorffs...\n")
# train_targets = construct_hausdorffs(train_vertices)
#
# print("Constructing Test Vertices...\n")
# test_vertices = construct_vertices(data_num_vertices, test_size)
# print("Constructing Test Hausdorffs...\n")
# test_targets = construct_hausdorffs(test_vertices)
# print("Data Preparation:")
# print("Constructing Training Vertices...\n")
# penta_vertices = construct_vertices(5, train_size // 2)
# hexagon_vertices = construct_vertices(6, train_size // 2)
# train_vertices = penta_vertices + hexagon_vertices
# print("Constructing Training Hausdorffs...\n")
# train_targets = construct_hausdorffs(train_vertices)
#
# print("Constructing Test Vertices...\n")
# penta_vertices = construct_vertices(5, test_size // 2)
# hexagon_vertices = construct_vertices(6, test_size // 2)
# test_vertices = penta_vertices + hexagon_vertices
#
# print("Constructing Test Hausdorffs...\n")
# test_targets = construct_hausdorffs(test_vertices)
# np.savez("data/vertices-targets.npz", train_vertices=train_targets, test_vertices=test_targets, train_targets=train_targets, test_targets=test_targets)
data = np.load("dataset/de_5-be_5-et_basis-bs_32-trs_10000-tes_1000.npz")
train_vertices, train_targets, test_vertices, test_targets = data["train_vertices"], data["train_targets"], data["test_vertices"], data["test_targets"]
# all_anchors = data["basis"]
all_anchors = [generate_polygon(center=(0, 0),
avg_radius=1,
irregularity=np.random.rand(),
spikiness=np.random.rand(),
num_vertices=basis_num_vertices) for _ in range(32)]
# five_anchors = [generate_polygon(center=(0, 0),
# avg_radius=1,
# irregularity=np.random.rand(),
# spikiness=np.random.rand(),
# num_vertices=5) for _ in range(16)]
#
# six_anchors = [generate_polygon(center=(0, 0),
# avg_radius=1,
# irregularity=np.random.rand(),
# spikiness=np.random.rand(),
# num_vertices=6) for _ in range(16)]
# for encoder_type in ["basis", "svd"]:
for basis_size in [32]:
anchors = all_anchors[:basis_size]
# anchors = five_anchors[:basis_size//2] + six_anchors[:basis_size//2]
for encoder_type in ["basis", "svd", "rotscale"]:
print("Encoder type: {}, Basis size: {}".format(encoder_type, basis_size))
if encoder_type == "basis":
encoder = RandomBasisEncoder(basis_size, basis_num_vertices, anchors)
elif encoder_type == "svd":
encoder = RandomSVDEncoder(basis_size, basis_num_vertices, anchors)
elif encoder_type == "rotscale":
encoder = RandomRotScaleEncoder(basis_size, basis_num_vertices, anchors)
else:
assert False, "Unknown encoder type: {}".format(encoder_type)
print("Constructing Training Encodings...\n")
train_encodings = construct_encodings(encoder, train_vertices)
print("Constructing Test Encodings...\n")
test_encodings = construct_encodings(encoder, test_vertices)
np.savez("dataset/de_{}-be_{}-et_{}-bs_{}-trs_{}-tes_{}".format(data_num_vertices, basis_num_vertices, encoder_type, basis_size, train_size, test_size),
# basis=encoder.anchors,
# train_vertices=train_vertices,
train_encodings=train_encodings, train_targets=train_targets,
# test_vertices=test_vertices,
test_encodings=test_encodings, test_targets=test_targets)