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 1.12 by nmohr, Thu Sep 20 13:08:07 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()
44 + #config.read('./config7TeV_ZZ')
45   config.read(opts.config)
46   anaTag = config.get("Analysis","tag")
47  
48   #get locations:
49   Wdir=config.get('Directories','Wdir')
50 < MVAdir=config.get('Directories','MVAdir')
50 > MVASubdir=config.get('Directories','MVAdir')
51 > samplesinfo=config.get('Directories','samplesinfo')
52  
53   #systematics
48 systematics=config.get('systematics','systematics')
49 systematics=systematics.split(' ')
54  
51 #TreeVar Array
52 #MVA_Vars={}
53 #for systematic in systematics:
54 #    MVA_Vars[systematic]=config.get('treeVars',systematic)
55 #    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  
59 < Apath=opts.path
61 < infofile = open(Apath+'/samples.info','r')
59 > infofile = open(samplesinfo,'r')
60   info = pickle.load(infofile)
61   infofile.close()
62   arglist=opts.discr #RTight_blavla,bsbsb
# Line 69 | 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
78 #MVAnames=[]
79 #for MVA in MVAlist:
80 #    print MVA
81 #    MVAnames.append(config.get(MVA,'MVAname'))
82 #print Wdir+'/weights/'+factoryname+'_'+MVAname+'.info'
83 #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      
90 treeVarSet=MVAinfos[0].varset
91 #variables
92 #TreeVar Array
93 MVA_Vars={}
94 for systematic in systematics:
95    MVA_Vars[systematic]=config.get(treeVarSet,systematic)
96    MVA_Vars[systematic]=MVA_Vars[systematic].split(' ')
97 #Spectators:
98 #spectators=config.get(treeVarSet,'spectators')
99 #spectators=spectators.split(' ')
100 #progbar quatsch
88   longe=40
89   #Workdir
90   workdir=ROOT.gDirectory.GetPath()
104 #os.mkdir(Apath+'/MVAout')
91  
106 #Book TMVA readers: MVAlist=["MMCC_bla","CC5050_bla"]
107 readers=[]
108 for MVA in MVAlist:
109    readers.append(ROOT.TMVA.Reader("!Color:!Silent"))
110
111 #define variables and specatators
112 MVA_var_buffer = []
113 MVA_var_buffer4 = []
114 for i in range(len( MVA_Vars['Nominal'])):
115    MVA_var_buffer.append(array( 'f', [ 0 ] ))
116    for reader in readers:
117        reader.AddVariable( MVA_Vars['Nominal'][i],MVA_var_buffer[i])
118 #MVA_spectator_buffer = []
119 #for i in range(len(spectators)):
120 #    MVA_spectator_buffer.append(array( 'f', [ 0 ] ))
121 #    for reader in readers:
122 #        reader.AddSpectator(spectators[i],MVA_spectator_buffer[i])
123 #Load raeder
124 for i in range(0,len(readers)):
125    readers[i].BookMVA(MVAinfos[i].MVAname,MVAdir+'/data/'+MVAinfos[i].getweightfile())
126 #--> Now the MVA is booked
92  
93   #Apply samples
94 < infofile = open(Apath+'/samples.info','r')
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(job.path+'/'+MVAdir+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 152 | 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]=[]
164 <                    #create TTreeFormulas
165 <                    for j in range(len( MVA_Vars['Nominal'])):
166 <                        MVA_formulas[systematic].append(ROOT.TTreeFormula("MVA_formula%s_%s"%(j,systematic),MVA_Vars[systematic][j],tree))
167 <                        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
168 <                outfile.cd()
169 <                #Setup Branches
170 <                MVAbranches=[]
171 <                MVAbranches4=[]
172 <                for i in range(0,len(readers)):
173 <                    MVAbranches.append(array('f',[0]*9))
174 <                    MVAbranches4.append(array('f',[0]*9))
175 <                    newtree.Branch(MVAinfos[i].MVAname,MVAbranches[i],'nominal:JER_up:JER_down:JES_up:JES_down:beff_up:beff_down:bmis_up:bmis_down/F')
176 <                    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/F')
177 <                print '\n--> ' + job.name +':'
178 <                #progbar setup
179 <                if nEntries >= longe:
180 <                    step=int(nEntries/longe)
181 <                    long=longe
182 <                else:
183 <                    long=nEntries
184 <                    step = 1
185 <                bar=progbar(long)
186 <                #Fill event by event:
187 <                for entry in range(0,nEntries):
188 <                    if entry % step == 0:
189 <                        bar.move()
190 <                    #load entry
191 <                    tree.GetEntry(entry)
192 <                    for systematic in systematics:
193 <                        for j in range(len( MVA_Vars['Nominal'])):
194 <                            MVA_var_buffer[j][0] = MVA_formulas[systematic][j].EvalInstance()
195 <                            
196 <                        for j in range(0,len(readers)):
197 <                            MVAbranches[j][systematics.index(systematic)] = readers[j].EvaluateMVA(MVAinfos[j].MVAname)
198 <                            
199 <                        for j in range(len( MVA_Vars['Nominal'])):
200 <                            MVA_var_buffer[j][0] = MVA_formulas4[systematic][j].EvalInstance()
201 <                            
202 <                        for j in range(0,len(readers)):
203 <                            MVAbranches4[j][systematics.index(systematic)] = readers[j].EvaluateMVA(MVAinfos[j].MVAname)
204 <                    #Fill:
205 <                    newtree.Fill()
206 <                newtree.AutoSave()
207 <                outfile.Close()
208 <                
209 <            #DATA:
210 <            if job.type == 'DATA':
211 <                #MVA Formulas
212 <                MVA_formulas_Nominal = []
213 <                #create TTreeFormulas
214 <                for j in range(len( MVA_Vars['Nominal'])):
215 <                    MVA_formulas_Nominal.append(ROOT.TTreeFormula("MVA_formula%s_Nominal"%j, MVA_Vars['Nominal'][j],tree))
216 <                outfile.cd()
217 <                MVAbranches=[]
218 <                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')
221                    newtree.Branch(MVAinfos[i].MVAname+'_4',MVAbranches[i],'nominal/F')
222                #progbar          
223                print '\n--> ' + job.name +':'
224                if nEntries >= longe:
225                    step=int(nEntries/longe)
226                    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:
# Line 252 | Line 210 | if doinfo:
210          for MVAinfo in MVAinfos:
211              job.addcomment('Added MVA %s'%MVAinfo.MVAname)
212          job.addpath(MVAdir)
213 <    infofile = open(Apath+MVAdir+'/samples.info','w')
213 >    infofile = open(samplesinfo,'w')
214      pickle.dump(Ainfo,infofile)
215      infofile.close()
216  

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines