-
Notifications
You must be signed in to change notification settings - Fork 6
/
TrafoProbNN.py
258 lines (219 loc) · 8.85 KB
/
TrafoProbNN.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
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
#!/usr/bin/python
import ROOT
import array
import itertools
import numpy as np
import math
from argparse import ArgumentParser
import os.path
def get_any_tree(tfilepath):
'''
If given a root file with only one tree, this function will return the name the tree.
'''
from root_numpy import list_trees
trees = list_trees(tfilepath)
if len(trees) == 1:
tree_name = trees[0]
else:
raise ValueError('More than one tree found in {}'.format(tfilepath))
return tree_name
def back_transform(var):
'''
Performs the back transformation of ProbNN variables, using numpy magic
for huge speedup when passing arrays to this function.
This function does not return anything, but changes the passed array in
place.
For the per-event back transformation when looping over a TTree use
UntrafoProbNN
'''
error_mask = var < -500
return -2 * error_mask.astype(int) + (~error_mask).astype(int) * (
np.exp(var) / (1 + np.exp(var)))
def trafoProbNN(variable_name, event):
'''
calculates the transformation of a given variable
'''
variable = event.__getattr__(variable_name)
#Check for problematic ProbNN values
if variable <= 0.0 or variable >= 1.0: #-1000 or -2 is the default value when ProbNN could not be read when creating the tuples
problem_branches[variable_name] = problem_branches.get(
variable_name, 0) + 1
return -1000. #also return -1000
return np.log(variable / (1 - variable))
def UntrafoProbNN(variable_name, event):
'''
calculates the inverse transformation of a given variable
'''
variable = event.__getattr__(variable_name)
#Check for problematic transformed ProbNN values
if variable < -500:
return -2.
else:
return np.exp(variable) / (1 + np.exp(variable))
if __name__ == '__main__':
#Read options
#Create optionparser
parser = ArgumentParser(
usage=
"Tool to transform ProbNN-variables according to log(X_ProbNNY/(1-X_ProbNNY))"
)
parser.add_argument(
"-i",
"--input",
dest="Input",
action="store",
required=True,
help="Input ROOT-file")
parser.add_argument(
"-t",
"--tree",
dest="Tree",
action="store",
required=False,
help=
"Input TTree (optional if there is only one tree in the input-file)")
parser.add_argument(
"-o",
"--output",
dest="Output",
action="store",
required=True,
help="Output ROOT-file")
#parser.add_argument( "--progress", dest="progress", action="store_true", default=False, required=False, help="Show detailed progressbar")
parser.add_argument(
"Variables",
metavar='V',
type=str,
nargs='*',
help="Explicit variable name or variable names to be transformed")
parser.add_argument(
"-m",
"--match",
dest="Patterns",
action="append",
required=False,
help="Transform all variables matching this pattern")
parser.add_argument(
"-r",
"--reverse",
dest="Reverse",
action="store_true",
help="Perform inverse transformation exp(X_ProbNNY)/(1+exp(X_ProbNNY))"
)
#Parse arguments from command line
options = parser.parse_args()
#Open tree and clone it
inputfile = ROOT.TFile(options.Input, "READ")
if not inputfile.IsOpen():
raise SystemExit("Could not open inputfile!")
if options.Tree == None:
inputtreename = get_any_tree(options.Input)
else:
inputtreename = options.Tree
from root_numpy import list_trees
trees = list_trees(options.Input)
# If data was downloaded with grab_data method, you end up with many trees in your file
if len(trees) > 1:
inputtree=ROOT.TChain("tree")
for i in range(1,len(trees)+1):
filename = options.Input + "/tree;{}".format(i)
status = inputtree.Add(filename,-1)
if status == 0:
break
else:
inputtree = inputfile.Get(inputtreename)
print 'Entries', inputtree.GetEntries()
outputfile = options.Output
#Clone tree
outputfile = ROOT.TFile(options.Output, "RECREATE")
outputtree = inputtree.CloneTree(0)
#Get explicit variable names
variable_names = options.Variables
#Create new branches for explicit variable names
variable_trafo_branches = []
for variable_name in variable_names:
variable_trafo_branches.append(array.array("d", [0.0]))
if options.Reverse:
#If var-name has "Trafo" in it, change it to "Untrafo", otherwise append Untrafo
if "Trafo" in variable_name:
outputbranchname = variable_name.replace("Trafo", "Untrafo")
else:
outputbranchname = variable_name + "_Untrafo"
outputtree.Branch(outputbranchname, variable_trafo_branches[-1],
outputbranchname + "/D")
else:
outputtree.Branch(variable_name + "_Trafo",
variable_trafo_branches[-1],
variable_name + "_Trafo/D")
#Get variable names matching the patterns (if applicable)
if options.Patterns:
for pattern in options.Patterns:
for branch in inputtree.GetListOfBranches():
if pattern in branch.GetName():
variable_name = branch.GetName()
variable_names.append(variable_name)
variable_trafo_branches.append(array.array("d", [0.0]))
if options.Reverse:
#If var-name has "Trafo" in it, change it to "Untrafo", otherwise append Untrafo
if "Trafo" in variable_name:
outputbranchname = variable_name.replace(
"Trafo", "Untrafo")
else:
outputbranchname = variable_name + "_Untrafo"
outputtree.Branch(outputbranchname,
variable_trafo_branches[-1],
outputbranchname + "/D")
else:
outputtree.Branch(variable_name + "_Trafo",
variable_trafo_branches[-1],
variable_name + "_Trafo/D")
if len(variable_names) is 0:
raise SystemExit("No variables for transformation given/found for {}".
format(options.Input))
# Create dictionary that will contain the branch names, where invalid ProbNN values were encountered (i.e. outside of 0 to 1)
# and attached to the branch names the number of events there an error occured
problem_branches = {}
#Progressbar
entries = inputtree.GetEntries()
print("Processing {0} entries in {2} /{1}".format(entries, inputtreename,
options.Input))
print("\nThe following variables will be transformed:")
print(", ".join(variable_names))
print("\n")
#widgets = [os.path.basename(options.Output), progressbar.Percentage(), ' ', progressbar.Bar(), ' ', progressbar.ETA()]
#pbar = progressbar.ProgressBar(widgets=widgets, maxval=entries).start()
#pbar.update(0)
#Iterate through tree
for (i, event) in itertools.izip(xrange(entries), inputtree):
inputtree.GetEntry(i)
#iterate through variables
for (variable_name, variable_trafo_branch) in itertools.izip(
variable_names, variable_trafo_branches):
if options.Reverse:
variable_trafo_branch[0] = UntrafoProbNN(variable_name, event)
else:
variable_trafo_branch[0] = trafoProbNN(variable_name, event)
#Fill tree
outputtree.Fill()
#Progressbar
# if options.progress: #detailed progress
# pbar.update(i+1)
# else:
# if (i+1) % (entries/10) == 0:
# pbar.update(i+1)
if problem_branches:
print(
"\n\nWARNING: There were events with ProbNN-values outside the allowed region of [0,1] for {}:".
format(options.Input))
print("{:<20} {:>15} {:>10}".format('Branch', 'Probl. Events',
'Percent'))
for branch, events in problem_branches.iteritems():
print("{:<20} {:>15} {:>10}%".format(
branch, events, float(events) / entries * 100.))
print(
"\t=> Setting these to -1000 (Default value for ProbNN-variables if none was found when creating the ntuple)"
)
print("\nFinished processing {0} entries in {2} /{1}\n\t=> Writing to {3}".
format(entries, inputtreename, options.Input, options.Output))
outputtree.Write()
outputfile.Close()