ViewVC Help
View File | Revision Log | Show Annotations | Root Listing
root/cvsroot/UserCode/VHbb/python/evaluateMVA.py
(Generate patch)

Comparing UserCode/VHbb/python/evaluateMVA.py (file contents):
Revision 1.16 by nmohr, Thu Oct 18 14:02:33 2012 UTC vs.
Revision 1.19 by bortigno, Fri Jan 11 13:47:24 2013 UTC

# Line 2 | Line 2
2   import sys
3   import os
4   import ROOT
5 from ROOT import TFile
5   from array import array
6   from math import sqrt
7   from copy import copy
# Line 11 | Line 10 | import warnings
10   warnings.filterwarnings( action='ignore', category=RuntimeWarning, message='creating converter.*' )
11   from optparse import OptionParser
12   from BetterConfigParser import BetterConfigParser
14 from samplesclass import sample
15 from mvainfos import mvainfo
13   import pickle
17 from progbar import progbar
18 from printcolor import printc
14  
15   #CONFIGURE
16 <
16 > ROOT.gROOT.SetBatch(True)
17 > print 'hello'
18   #load config
19   #os.mkdir(path+'/sys')
20   argv = sys.argv
# Line 27 | Line 23 | parser.add_option("-U", "--update", dest
23                        help="update infofile")
24   parser.add_option("-D", "--discr", dest="discr", default="",
25                        help="discriminators to be added")
26 < parser.add_option("-P", "--path", dest="path", default="",
27 <                      help="path to samples")
26 > #parser.add_option("-I", "--inpath", dest="inpath", default="",
27 > #                      help="path to samples")
28 > #parser.add_option("-O", "--outpath", dest="outpath", default="",
29 > #                      help="path where to store output samples")
30   parser.add_option("-S", "--samples", dest="names", default="",
31                        help="samples you want to run on")
32   parser.add_option("-C", "--config", dest="config", default=[], action="append",
33                        help="configuration file")
34   (opts, args) = parser.parse_args(argv)
35 +
36 + from samplesclass import sample
37 + from mvainfos import mvainfo
38 + from progbar import progbar
39 + from printcolor import printc
40 +
41   if opts.config =="":
42          opts.config = "config"
43   config = BetterConfigParser()
# Line 47 | Line 51 | MVASubdir=config.get('Directories','MVAd
51   samplesinfo=config.get('Directories','samplesinfo')
52  
53   #systematics
50 systematics=config.get('systematics','systematics')
51 systematics=systematics.split(' ')
54  
53 #TreeVar Array
54 #MVA_Vars={}
55 #for systematic in systematics:
56 #    MVA_Vars[systematic]=config.get('treeVars',systematic)
57 #    MVA_Vars[systematic]=MVA_Vars[systematic].split(' ')
55  
56 < ######################
57 < #Evaluate multi: Must Have same treeVars!!!
56 > INpath = config.get('Directories','MVAin')
57 > OUTpath = config.get('Directories','MVAout')
58  
62 Apath=opts.path
59   infofile = open(samplesinfo,'r')
60   info = pickle.load(infofile)
61   infofile.close()
# Line 71 | Line 67 | namelist=namelistIN.split(',')
67   doinfo=bool(int(opts.update))
68  
69   MVAlist=arglist.split(',')
70 < MVAdir=config.get('Directories','vhbbpath')
70 >
71  
72   #CONFIG
73   #factory
74   factoryname=config.get('factory','factoryname')
75 +
76 + #load the namespace
77 + VHbbNameSpace=config.get('VHbbNameSpace','library')
78 + ROOT.gSystem.Load(VHbbNameSpace)
79 +
80   #MVA
80 #MVAnames=[]
81 #for MVA in MVAlist:
82 #    print MVA
83 #    MVAnames.append(config.get(MVA,'MVAname'))
84 #print Wdir+'/weights/'+factoryname+'_'+MVAname+'.info'
85 #MVAinfofiles=[]
81   MVAinfos=[]
82 + MVAdir=config.get('Directories','vhbbpath')
83   for MVAname in MVAlist:
84      MVAinfofile = open(MVAdir+'/data/'+factoryname+'_'+MVAname+'.info','r')
85      MVAinfos.append(pickle.load(MVAinfofile))
86      MVAinfofile.close()
87      
92 treeVarSet=MVAinfos[0].varset
93 #variables
94 #TreeVar Array
95 MVA_Vars={}
96 for systematic in systematics:
97    MVA_Vars[systematic]=config.get(treeVarSet,systematic)
98    MVA_Vars[systematic]=MVA_Vars[systematic].split(' ')
99 #Spectators:
100 #spectators=config.get(treeVarSet,'spectators')
101 #spectators=spectators.split(' ')
102 #progbar quatsch
88   longe=40
89   #Workdir
90   workdir=ROOT.gDirectory.GetPath()
106 #os.mkdir(Apath+'/MVAout')
91  
108 #Book TMVA readers: MVAlist=["MMCC_bla","CC5050_bla"]
109 readers=[]
110 for MVA in MVAlist:
111    readers.append(ROOT.TMVA.Reader("!Color:!Silent"))
112
113 #define variables and specatators
114 MVA_var_buffer = []
115 MVA_var_buffer4 = []
116 for i in range(len( MVA_Vars['Nominal'])):
117    MVA_var_buffer.append(array( 'f', [ 0 ] ))
118    for reader in readers:
119        reader.AddVariable( MVA_Vars['Nominal'][i],MVA_var_buffer[i])
120 #MVA_spectator_buffer = []
121 #for i in range(len(spectators)):
122 #    MVA_spectator_buffer.append(array( 'f', [ 0 ] ))
123 #    for reader in readers:
124 #        reader.AddSpectator(spectators[i],MVA_spectator_buffer[i])
125 #Load raeder
126 for i in range(0,len(readers)):
127    readers[i].BookMVA(MVAinfos[i].MVAname,MVAdir+'/data/'+MVAinfos[i].getweightfile())
128 #--> Now the MVA is booked
92  
93   #Apply samples
94   infofile = open(samplesinfo,'r')
95   Ainfo = pickle.load(infofile)
96   infofile.close()
97  
98 +
99 + class MvaEvaluater:
100 +    def __init__(self, config, MVAinfo):
101 +        self.varset = MVAinfo.varset
102 +        #Define reader
103 +        self.reader = ROOT.TMVA.Reader("!Color:!Silent")
104 +        MVAdir=config.get('Directories','vhbbpath')
105 +        self.systematics=config.get('systematics','systematics').split(' ')
106 +        self.MVA_Vars={}
107 +        self.MVAname = MVAinfo.MVAname
108 +        for systematic in self.systematics:
109 +            self.MVA_Vars[systematic]=config.get(self.varset,systematic)
110 +            self.MVA_Vars[systematic]=self.MVA_Vars[systematic].split(' ')
111 +        #define variables and specatators
112 +        self.MVA_var_buffer = []
113 +        for i in range(len( self.MVA_Vars['Nominal'])):
114 +            self.MVA_var_buffer.append(array( 'f', [ 0 ] ))
115 +            self.reader.AddVariable( self.MVA_Vars['Nominal'][i],self.MVA_var_buffer[i])
116 +        self.reader.BookMVA(MVAinfo.MVAname,MVAdir+'/data/'+MVAinfo.getweightfile())
117 +        #--> Now the MVA is booked
118 +
119 +    def setBranches(self,tree,job):
120 +        #Set formulas for all vars
121 +        self.MVA_formulas={}
122 +        for systematic in self.systematics:
123 +            if job.type == 'DATA' and not systematic == 'Nominal': continue
124 +            self.MVA_formulas[systematic]=[]
125 +            for j in range(len( self.MVA_Vars['Nominal'])):
126 +                self.MVA_formulas[systematic].append(ROOT.TTreeFormula("MVA_formula%s_%s"%(j,systematic),self.MVA_Vars[systematic][j],tree))
127 +
128 +    def evaluate(self,MVAbranches,job):
129 +        #Evaluate all vars and fill the branches
130 +        for systematic in self.systematics:
131 +            for j in range(len( self.MVA_Vars['Nominal'])):
132 +                if job.type == 'DATA' and not systematic == 'Nominal': continue
133 +                self.MVA_var_buffer[j][0] = self.MVA_formulas[systematic][j].EvalInstance()                
134 +            MVAbranches[self.systematics.index(systematic)] = self.reader.EvaluateMVA(self.MVAname)
135 +
136 +
137 + theMVAs = []
138 + for mva in MVAinfos:
139 +    theMVAs.append(MvaEvaluater(config,mva))
140 +
141 +
142   #eval
143   for job in Ainfo:
144      if eval(job.active):
145          if job.name in namelist:
146              #get trees:
147 <            input = TFile.Open(Apath+'/'+job.getpath(),'read')
148 <            outfile = TFile.Open(Apath+'/'+MVASubdir+job.prefix+job.identifier+'.root','recreate')
147 >            print INpath+'/'+job.prefix+job.identifier+'.root'
148 >            input = ROOT.TFile.Open(INpath+'/'+job.prefix+job.identifier+'.root','read')
149 >            print OUTpath+'/'+job.prefix+job.identifier+'.root'
150 >            outfile = ROOT.TFile.Open(OUTpath+'/'+job.prefix+job.identifier+'.root','recreate')
151              input.cd()
152              obj = ROOT.TObject
153              for key in ROOT.gDirectory.GetListOfKeys():
# Line 154 | Line 163 | for job in Ainfo:
163              nEntries = tree.GetEntries()
164              outfile.cd()
165              newtree = tree.CloneTree(0)
166 +            
167  
168 <            #MCs:
169 <            if job.type != 'DATA':
170 <                MVA_formulas={}
171 <                MVA_formulas4={}
172 <                for systematic in systematics:
173 <                    #print '\t\t - ' + systematic
174 <                    MVA_formulas[systematic]=[]
175 <                    MVA_formulas4[systematic]=[]
166 <                    #create TTreeFormulas
167 <                    for j in range(len( MVA_Vars['Nominal'])):
168 <                        MVA_formulas[systematic].append(ROOT.TTreeFormula("MVA_formula%s_%s"%(j,systematic),MVA_Vars[systematic][j],tree))
169 <                        MVA_formulas4[systematic].append(ROOT.TTreeFormula("MVA_formula4%s_%s"%(j,systematic),MVA_Vars['Nominal'][j]+'+('+MVA_Vars[systematic][j]+'-'+MVA_Vars['Nominal'][j]+')*4',tree))#HERE change
170 <                outfile.cd()
171 <                #Setup Branches
172 <                MVAbranches=[]
173 <                MVAbranches4=[]
174 <                for i in range(0,len(readers)):
175 <                    MVAbranches.append(array('f',[0]*11))
176 <                    MVAbranches4.append(array('f',[0]*11))
177 <                    newtree.Branch(MVAinfos[i].MVAname,MVAbranches[i],'nominal:JER_up:JER_down:JES_up:JES_down:beff_up:beff_down:bmis_up:bmis_down:beff1_up:beff1_down/F')
178 <                    newtree.Branch(MVAinfos[i].MVAname+'_4',MVAbranches4[i],'nominal:JER_up:JER_down:JES_up:JES_down:beff_up:beff_down:bmis_up:bmis_down:beff1_up:beff1_down/F')
179 <                print '\n--> ' + job.name +':'
180 <                #progbar setup
181 <                if nEntries >= longe:
182 <                    step=int(nEntries/longe)
183 <                    long=longe
184 <                else:
185 <                    long=nEntries
186 <                    step = 1
187 <                bar=progbar(long)
188 <                #Fill event by event:
189 <                for entry in range(0,nEntries):
190 <                    if entry % step == 0:
191 <                        bar.move()
192 <                    #load entry
193 <                    tree.GetEntry(entry)
194 <                    for systematic in systematics:
195 <                        for j in range(len( MVA_Vars['Nominal'])):
196 <                            MVA_var_buffer[j][0] = MVA_formulas[systematic][j].EvalInstance()
197 <                            
198 <                        for j in range(0,len(readers)):
199 <                            MVAbranches[j][systematics.index(systematic)] = readers[j].EvaluateMVA(MVAinfos[j].MVAname)
200 <                            
201 <                        for j in range(len( MVA_Vars['Nominal'])):
202 <                            MVA_var_buffer[j][0] = MVA_formulas4[systematic][j].EvalInstance()
203 <                            
204 <                        for j in range(0,len(readers)):
205 <                            MVAbranches4[j][systematics.index(systematic)] = readers[j].EvaluateMVA(MVAinfos[j].MVAname)
206 <                    #Fill:
207 <                    newtree.Fill()
208 <                newtree.AutoSave()
209 <                outfile.Close()
210 <                
211 <            #DATA:
212 <            if job.type == 'DATA':
213 <                #MVA Formulas
214 <                MVA_formulas_Nominal = []
215 <                #create TTreeFormulas
216 <                for j in range(len( MVA_Vars['Nominal'])):
217 <                    MVA_formulas_Nominal.append(ROOT.TTreeFormula("MVA_formula%s_Nominal"%j, MVA_Vars['Nominal'][j],tree))
218 <                outfile.cd()
219 <                MVAbranches=[]
220 <                for i in range(0,len(readers)):
168 >            #Set branch adress for all vars
169 >            for i in range(0,len(theMVAs)):
170 >                theMVAs[i].setBranches(tree,job)
171 >            outfile.cd()
172 >            #Setup Branches
173 >            MVAbranches=[]
174 >            for i in range(0,len(theMVAs)):
175 >                if job.type == 'Data':
176                      MVAbranches.append(array('f',[0]))
177                      newtree.Branch(MVAinfos[i].MVAname,MVAbranches[i],'nominal/F')
223                    newtree.Branch(MVAinfos[i].MVAname+'_4',MVAbranches[i],'nominal/F')
224                #progbar          
225                print '\n--> ' + job.name +':'
226                if nEntries >= longe:
227                    step=int(nEntries/longe)
228                    long=longe
178                  else:
179 <                    long=nEntries
180 <                    step = 1
181 <                bar=progbar(long)
182 <                #Fill event by event:
183 <                for entry in range(0,nEntries):
184 <                    if entry % step == 0:
185 <                        bar.move()
186 <                    #load entry
187 <                    tree.GetEntry(entry)
188 <                    #nominal:
189 <                    for j in range(len( MVA_Vars['Nominal'])):
190 <                            MVA_var_buffer[j][0] = MVA_formulas_Nominal[j].EvalInstance()
179 >                    MVAbranches.append(array('f',[0]*11))
180 >                    newtree.Branch(theMVAs[i].MVAname,MVAbranches[i],'nominal:JER_up:JER_down:JES_up:JES_down:beff_up:beff_down:bmis_up:bmis_down:beff1_up:beff1_down/F')
181 >                MVA_formulas_Nominal = []
182 >            print '\n--> ' + job.name +':'
183 >            #progbar setup
184 >            if nEntries >= longe:
185 >                step=int(nEntries/longe)
186 >                long=longe
187 >            else:
188 >                long=nEntries
189 >                step = 1
190 >            bar=progbar(long)
191 >            #Fill event by event:
192 >            for entry in range(0,nEntries):
193 >                if entry % step == 0:
194 >                    bar.move()
195 >                #load entry
196 >                tree.GetEntry(entry)
197                              
198 <                    for j in range(0,len(readers)):
199 <                        MVAbranches[j][0]= readers[j].EvaluateMVA(MVAinfos[j].MVAname)
200 <                    newtree.Fill()
201 <                newtree.AutoSave()
202 <                outfile.Close()
203 <
198 >                for i in range(0,len(theMVAs)):
199 >                    theMVAs[i].evaluate(MVAbranches[i],job)
200 >                #Fill:
201 >                newtree.Fill()
202 >            newtree.AutoSave()
203 >            outfile.Close()
204 >                
205   print '\n'
206  
207   #Update Info:

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines