-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathdata_processing.py
More file actions
170 lines (132 loc) · 9.98 KB
/
data_processing.py
File metadata and controls
170 lines (132 loc) · 9.98 KB
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
from dataclasses import dataclass
from typing import Optional
import torch
import pandas as pd
@dataclass
class PatchParam:
name: str
p_type: str # 'm', 'b', 'c', 'x'
p_min: int
p_max: int
n_classes: Optional[int] = None # for categorical
def get_params():
p_type = ["m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "c", "c", "m", "m", "m", "m", "b", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "c", "c", "m", "m", "m", "m", "b", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "c", "c", "m", "m", "m", "m", "b", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "c", "c", "m", "m", "m", "m", "b", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "c", "c", "m", "m", "m", "m", "b", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "c", "c", "m", "m", "m", "m", "b", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "m", "x", "m", "b", "m", "m", "m", "m", "b", "c", "m", "m"]
p_name = ["operator6rate1", "operator6rate2", "operator6rate3", "operator6rate4", "operator6level1", "operator6level2", "operator6level3", "operator6level4", "operator6keyboardlevelscalingbreakpoint", "operator6keyboardlevelscalingleftdepth", "operator6keyboardlevelscalingrightdepth", "operator6keyboardlevelscalingleftcurve", "operator6keyboardlevelscalingrightcurve", "operator6keyboardratescaling", "operator6amplitudemodulationsensitivity", "operator6keyvelocitysensitivity", "operator6operatoroutputlevel", "operator6oscillatormode", "operator6frequencycoarse", "operator6frequencyfine", "operator6frequencydetune", "operator5rate1", "operator5rate2", "operator5rate3", "operator5rate4", "operator5level1", "operator5level2", "operator5level3", "operator5level4", "operator5keyboardlevelscalingbreakpoint", "operator5keyboardlevelscalingleftdepth", "operator5keyboardlevelscalingrightdepth", "operator5keyboardlevelscalingleftcurve", "operator5keyboardlevelscalingrightcurve", "operator5keyboardratescaling", "operator5amplitudemodulationsensitivity", "operator5keyvelocitysensitivity", "operator5operatoroutputlevel", "operator5oscillatormode", "operator5frequencycoarse", "operator5frequencyfine", "operator5frequencydetune", "operator4rate1", "operator4rate2", "operator4rate3", "operator4rate4", "operator4level1", "operator4level2", "operator4level3", "operator4level4", "operator4keyboardlevelscalingbreakpoint", "operator4keyboardlevelscalingleftdepth", "operator4keyboardlevelscalingrightdepth", "operator4keyboardlevelscalingleftcurve", "operator4keyboardlevelscalingrightcurve", "operator4keyboardratescaling", "operator4amplitudemodulationsensitivity", "operator4keyvelocitysensitivity", "operator4operatoroutputlevel", "operator4oscillatormode", "operator4frequencycoarse", "operator4frequencyfine", "operator4frequencydetune", "operator3rate1", "operator3rate2", "operator3rate3", "operator3rate4", "operator3level1", "operator3level2", "operator3level3", "operator3level4", "operator3keyboardlevelscalingbreakpoint", "operator3keyboardlevelscalingleftdepth", "operator3keyboardlevelscalingrightdepth", "operator3keyboardlevelscalingleftcurve", "operator3keyboardlevelscalingrightcurve", "operator3keyboardratescaling", "operator3amplitudemodulationsensitivity", "operator3keyvelocitysensitivity", "operator3operatoroutputlevel", "operator3oscillatormode", "operator3frequencycoarse", "operator3frequencyfine", "operator3frequencydetune", "operator2rate1", "operator2rate2", "operator2rate3", "operator2rate4", "operator2level1", "operator2level2", "operator2level3", "operator2level4", "operator2keyboardlevelscalingbreakpoint", "operator2keyboardlevelscalingleftdepth", "operator2keyboardlevelscalingrightdepth", "operator2keyboardlevelscalingleftcurve", "operator2keyboardlevelscalingrightcurve", "operator2keyboardratescaling", "operator2amplitudemodulationsensitivity", "operator2keyvelocitysensitivity", "operator2operatoroutputlevel", "operator2oscillatormode", "operator2frequencycoarse", "operator2frequencyfine", "operator2frequencydetune", "operator1rate1", "operator1rate2", "operator1rate3", "operator1rate4", "operator1level1", "operator1level2", "operator1level3", "operator1level4", "operator1keyboardlevelscalingbreakpoint", "operator1keyboardlevelscalingleftdepth", "operator1keyboardlevelscalingrightdepth", "operator1keyboardlevelscalingleftcurve", "operator1keyboardlevelscalingrightcurve", "operator1keyboardratescaling", "operator1amplitudemodulationsensitivity", "operator1keyvelocitysensitivity", "operator1operatoroutputlevel", "operator1oscillatormode", "operator1frequencycoarse", "operator1frequencyfine", "operator1frequencydetune", "pitchegrate1", "pitchegrate2", "pitchegrate3", "pitchegrate4", "pitcheglevel1", "pitcheglevel2", "pitcheglevel3", "pitcheglevel4", "algorithm", "feedback", "oscillatorkeysync", "lfospeed", "lfodelay", "lfopitchmodulationdepth", "lfoamplitudemodulationdepth", "lfokeysync", "lfowave", "lfopitchmodulationsensitivity", "transpose"]
p_min = [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
p_max = [99, 99, 99, 99, 99, 99, 99, 99, 99, 99, 99, 3, 3, 7, 3, 7, 99, 1, 31, 99, 14, 99, 99, 99, 99, 99, 99, 99, 99, 99, 99, 99, 3, 3, 7, 3, 7, 99, 1, 31, 99, 14, 99, 99, 99, 99, 99, 99, 99, 99, 99, 99, 99, 3, 3, 7, 3, 7, 99, 1, 31, 99, 14, 99, 99, 99, 99, 99, 99, 99, 99, 99, 99, 99, 3, 3, 7, 3, 7, 99, 1, 31, 99, 14, 99, 99, 99, 99, 99, 99, 99, 99, 99, 99, 99, 3, 3, 7, 3, 7, 99, 1, 31, 99, 14, 99, 99, 99, 99, 99, 99, 99, 99, 99, 99, 99, 3, 3, 7, 3, 7, 99, 1, 31, 99, 14, 99, 99, 99, 99, 99, 99, 99, 99, 31, 7, 1, 99, 99, 99, 99, 1, 5, 7, 48]
params = [
PatchParam(n, t, mn, mx)
for n, t, mn, mx in zip(p_name, p_type, p_min, p_max)
]
c_lengths = [4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 32, 6]
c_idx = 0
for p in params:
if p.p_type in ('c', 'x'):
p.n_classes = c_lengths[c_idx]
c_idx += 1
return params
def get_algorithms():
algorithms = []
with open("E:\\Coding\\vae-main\\dx7\\dx7.algorithms", "r") as f:
for line in f:
size = 7
algorithm = [[(2**24-1) for _ in range(size)] for _ in range(size)]
edges = line.split(",")
for edge in edges:
start = int(edge[0]) - 1
end = int(edge[1]) - 1
algorithm[start][end] = 1
algorithms.append(algorithm)
all_graphs = []
for dist in algorithms:
n = len(dist)
for k in range(n):
for i in range(n):
for j in range(n):
dist[i][j] = min(dist[i][j], dist[i][k] + dist[k][j])
all_graphs.append(dist)
alibi_distances = -1 * (torch.tensor(all_graphs)) + 1
algorithms = torch.tensor(algorithms)
algorithms = algorithms.where(algorithms == 1, 0)
return algorithms, alibi_distances
class PatchProcessor:
def __init__(self, params):
self.params = params
def normalize(self, df):
df = df.copy()
for p in self.params:
if p.p_type == 'm':
df[p.name] = (
df[p.name].clip(lower=p.p_min) / p.p_max
)
return df
def one_hot_dataframe(self, df):
columns = []
new_types = []
for p in self.params:
col = df[p.name]
if p.p_type == 'm':
columns.append(col)
new_types.append('m')
elif p.p_type == 'b':
d = pd.get_dummies(col, prefix=p.name)
columns.append(d)
new_types.extend(['b'] * d.shape[1])
elif p.p_type in ('c', 'x'):
d = pd.get_dummies(col, prefix=p.name)
expected = [f"{p.name}_{i}" for i in range(p.n_classes)]
for e in expected:
if e not in d.columns:
d[e] = 0
d = d[expected]
columns.append(d)
new_types.extend([p.p_type] * p.n_classes)
df_out = pd.concat(columns, axis=1).astype('float32')
return df_out, new_types
def make_masks(self, expanded_types):
t = torch.tensor([ord(c) for c in expanded_types], dtype=torch.int32)
return (
t == ord('m'),
t == ord('b'),
t == ord('c'),
t == ord('x')
)
import torch
import torch.nn.functional as F
class PatchProcessorTensor:
def __init__(self, params):
self.params = params
def normalize(self, x):
x = x.clone().float()
for i, p in enumerate(self.params):
if p.p_type == 'm':
x[:, i] = x[:, i].clamp(min=p.p_min) / p.p_max
return x
def one_hot_tensor(self, x):
columns = []
new_types = []
for i, p in enumerate(self.params):
col = x[:, i]
if p.p_type == 'm':
columns.append(col.unsqueeze(1).float())
new_types.append('m')
elif p.p_type == 'b':
col = col.long()
d = F.one_hot(col, num_classes=2).float()
columns.append(d)
new_types.extend(['b'] * 2)
elif p.p_type in ('c', 'x'):
col = col.long()
d = F.one_hot(col, num_classes=p.n_classes).float()
columns.append(d)
new_types.extend([p.p_type] * p.n_classes)
x_out = torch.cat(columns, dim=1)
return x_out, new_types
def make_masks(self, expanded_types):
t = torch.tensor([ord(c) for c in expanded_types], dtype=torch.int32)
return (
t == ord('m'),
t == ord('b'),
t == ord('c'),
t == ord('x')
)