-
Notifications
You must be signed in to change notification settings - Fork 1
/
classes.py
193 lines (144 loc) · 6.67 KB
/
classes.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
189
190
191
192
193
import numpy as np
class dataType:
'''
Put data into proper format.
'''
def __init__(self, data_arrays):
self.init_features = ['trk_pt','trk_phi','trk_eta','trk_z0','trk_chi2pdof',
'trk_bendchi2','trk_nstub','trk_stubs_layer','trk_stubs_ps']
self.trks = trks(data_arrays, self.init_features, False)
#self.tps = tps(data_arrays)
def summarize(self):
print('L1 tracks info:')
print('\t',len(self.trks.y),'total tracks')
print('\t',len(self.trks.y[self.trks.y==0]),'fake tracks (',len(self.trks.y[self.trks.y==0])
/len(self.trks.y),')')
print('\t',len(self.trks.y[self.trks.y==1]),'real tracks (',len(self.trks.y[self.trks.y==1])
/len(self.trks.y),')')
#print('tp matched tracks info:')
#print('\t',self.tps.num,'total tps')
#print('\t',len(self.tps.matchtrks.y),'total matched tracks (',len(self.tps.matchtrks.y)/self.tps.num,')')
#print('\t',len(self.tps.matchtrks.y[self.tps.matchtrks.y==0]),'fake matched tracks (',
# len(self.tps.matchtrks.y[self.tps.matchtrks.y==0])/len(self.tps.matchtrks.y),')')
#print('\t',len(self.tps.matchtrks.y[self.tps.matchtrks.y==1]),'real matched tracks (',
# len(self.tps.matchtrks.y[self.tps.matchtrks.y==1])/len(self.tps.matchtrks.y),')')
return
class trks:
def __init__(self, data_arrays, feats, match_bool):
# data for ML
self.tp_match = match_bool
self.X_feats = None
self.X = self.setX(data_arrays, feats)
self.y = self.setY(data_arrays)
self.pdgid = data_arrays['trk_matchtp_pdgid'].flatten()
self.removeBad()
def setX(self, arrays, feats):
self.X_feats = []
tmp_X = [None]*len(feats)
# loop through features and extract them from array
nstub_idx = -1
for ii in range(len(feats)):
feat = feats[ii]
if 'nstub' in feat:
nstub_idx = ii
if 'stubs_layer' in feat:
tmp_X[ii] = self.setStubsLayer(arrays[feat].flatten(), arrays[feats[nstub_idx]].flatten())
elif 'stubs_ps' in feat:
tmp_X[ii] = self.isPs(arrays[feat].flatten(), arrays[feats[nstub_idx]].flatten())
elif 'stubs_barrel' in feat:
tmp_X[ii] = self.isBarrel(arrays[feat].flatten(), arrays[feats[nstub_idx]].flatten())
else:
tmp_X[ii] = arrays[feat].flatten()
if self.tp_match:
self.X_feats.append(feat[5:])
else:
self.X_feats.append(feat)
# put features in proper format
X = tmp_X[0]
for ii in range(1,len(tmp_X)):
X = np.column_stack((X,tmp_X[ii]))
return X
def setY(self, arrays):
if self.tp_match:
y = np.ones(len(self.X[:,0]))
else:
y = arrays['trk_fake'].flatten()
# both hard (1) and soft interactions (2) labeled as 1
y[y==2] = 1
return y
def setStubsLayer(self, layers, nstubs):
'''
Create individual counters for # of stubs in each layer (0-6).
'''
self.X_feats.extend(('trk_stubs_layer1','trk_stubs_layer2','trk_stubs_layer3',
'trk_stubs_layer4','trk_stubs_layer5','trk_stubs_layer6','trk_nlayer_miss'))
n_trks = len(nstubs)
stubs_layers = np.zeros((n_trks,7), dtype=int)
kk = 0
for ii in range(n_trks):
nstub = nstubs[ii]
for jj in range(kk,nstub+kk):
stubs_layers[ii,layers[jj]-1] = stubs_layers[ii,layers[jj]-1]+1
kk = kk+nstub
seq_idx = np.where(stubs_layers[ii,0:-1]>0)[0]
stubs_layers[ii,-1] = np.count_nonzero(stubs_layers[ii,seq_idx[0]:seq_idx[-1]+1]==0)
return stubs_layers
def isBarrel(self, labels, nstubs):
'''
Find majority label for barrel (1)/endcap (0) in all stubs.
For a tie, label assigned to endcap.
'''
self.X_feats.append('trk_stubs_barrel')
n_trks = len(nstubs)
stubs_barrel = np.zeros((n_trks,), dtype=int)
kk = 0
for ii in range(n_trks):
nstub = nstubs[ii]
counts = np.bincount(labels[kk:kk+nstub].astype(int))
stubs_barrel[ii] = np.argmax(counts)
kk = kk+nstub
return stubs_barrel
def isPs(self, labels, nstubs):
'''
Find majority label for ps module (1)/2s module (0) in all stubs.
For a tie, label assigned to 2s module.
'''
self.X_feats.append('trk_stubs_ps')
n_trks = len(nstubs)
stubs_ps = np.zeros((n_trks,), dtype=int)
kk = 0
for ii in range(n_trks):
nstub = nstubs[ii]
counts = np.bincount(labels[kk:kk+nstub].astype(int))
stubs_ps[ii] = np.argmax(counts)
kk = kk+nstub
return stubs_ps
def removeBad(self):
'''
Take out instances of tracks with nan as a feature and where |z0| is greater than 20 cm.
'''
bad_idx = np.argwhere(np.isnan(self.X))[0]
while np.isnan(self.X).any():
bad_i = np.argwhere(np.isnan(self.X))[0][0]
self.X = np.delete(self.X,bad_i,0)
self.y = np.delete(self.y,bad_i,0)
self.pdgid = np.delete(self.pdgid,bad_i,0)
#if self.z0idx != None:
# bad_i = np.where(abs(self.X[:,self.z0idx])>20)[0]
# if len(bad_i)>0:
# self.X = np.delete(self.X,bad_i,0)
# self.y = np.delete(self.y,bad_i,0)
return
class tps:
def __init__(self, data_arrays):
self.init_features = ['matchtrk_pt','matchtrk_phi','matchtrk_eta','matchtrk_z0','matchtrk_chi2pdof',
'matchtrk_bendchi2','matchtrk_nstub','matchtrk_stubs_layer','matchtrk_stubs_ps']
self.matchtrks = trks(self.takeOutFake(data_arrays), self.init_features, True)
self.num = len(data_arrays['tp_pt'].flatten())
def takeOutFake(self, arrays):
good_idx = (arrays['matchtrk_fake']==1)
fixed_feats = ['matchtrk_pt','matchtrk_phi','matchtrk_eta','matchtrk_z0',
'matchtrk_chi2pdof','matchtrk_bendchi2','matchtrk_nstub']
for feat in fixed_feats:
arrays[feat] = arrays[feat][good_idx]
return arrays