| 1 |
nmohr |
1.1 |
#!/usr/bin/env python
|
| 2 |
|
|
import pickle
|
| 3 |
|
|
import ROOT
|
| 4 |
|
|
from BetterConfigParser import BetterConfigParser
|
| 5 |
|
|
import sys, os
|
| 6 |
|
|
from optparse import OptionParser
|
| 7 |
|
|
from copy import copy,deepcopy
|
| 8 |
|
|
from StackMaker import StackMaker
|
| 9 |
|
|
from math import sqrt
|
| 10 |
|
|
import math
|
| 11 |
|
|
from HiggsAnalysis.CombinedLimit.DatacardParser import *
|
| 12 |
|
|
from HiggsAnalysis.CombinedLimit.ShapeTools import *
|
| 13 |
|
|
|
| 14 |
|
|
ROOT.gROOT.SetBatch(True)
|
| 15 |
|
|
ROOT.gSystem.Load("libHiggsAnalysisCombinedLimit.so")
|
| 16 |
|
|
|
| 17 |
|
|
#CONFIGURE
|
| 18 |
|
|
argv = sys.argv
|
| 19 |
|
|
parser = OptionParser()
|
| 20 |
|
|
parser.add_option("-D", "--datacard", dest="dc", default="",
|
| 21 |
|
|
help="Datacard to be plotted")
|
| 22 |
|
|
parser.add_option("-B", "--bin", dest="bin", default="",
|
| 23 |
|
|
help="DC bin to plot")
|
| 24 |
|
|
parser.add_option("-M", "--mlfit", dest="mlfit", default="",
|
| 25 |
|
|
help="mlfit file for nuisances")
|
| 26 |
|
|
parser.add_option("-F", "--fitresult", dest="fit", default="s",
|
| 27 |
|
|
help="Fit result to be used, 's' (signal+background) or 'b' (background only), default is 's'")
|
| 28 |
|
|
parser.add_option("-C", "--config", dest="config", default=[], action="append",
|
| 29 |
|
|
help="configuration file")
|
| 30 |
|
|
(opts, args) = parser.parse_args(argv)
|
| 31 |
|
|
|
| 32 |
|
|
|
| 33 |
|
|
def readBestFit(theFile):
|
| 34 |
|
|
file = ROOT.TFile(theFile)
|
| 35 |
|
|
if file == None: raise RuntimeError, "Cannot open file %s" % args[0]
|
| 36 |
|
|
fit_s = file.Get("fit_s")
|
| 37 |
|
|
fit_b = file.Get("fit_b")
|
| 38 |
|
|
prefit = file.Get("nuisances_prefit")
|
| 39 |
|
|
if fit_s == None or fit_s.ClassName() != "RooFitResult": raise RuntimeError, "File %s does not contain the output of the signal fit 'fit_s'" % args[0]
|
| 40 |
|
|
if fit_b == None or fit_b.ClassName() != "RooFitResult": raise RuntimeError, "File %s does not contain the output of the background fit 'fit_b'" % args[0]
|
| 41 |
|
|
if prefit == None or prefit.ClassName() != "RooArgSet": raise RuntimeError, "File %s does not contain the prefit nuisances 'nuisances_prefit'" % args[0]
|
| 42 |
|
|
|
| 43 |
|
|
isFlagged = {}
|
| 44 |
|
|
table = {}
|
| 45 |
|
|
fpf_b = fit_b.floatParsFinal()
|
| 46 |
|
|
fpf_s = fit_s.floatParsFinal()
|
| 47 |
|
|
nuiVariation = {}
|
| 48 |
|
|
for i in range(fpf_s.getSize()):
|
| 49 |
|
|
nuis_s = fpf_s.at(i)
|
| 50 |
|
|
name = nuis_s.GetName();
|
| 51 |
|
|
nuis_b = fpf_b.find(name)
|
| 52 |
|
|
nuis_p = prefit.find(name)
|
| 53 |
|
|
if nuis_p != None:
|
| 54 |
|
|
mean_p, sigma_p = (nuis_p.getVal(), nuis_p.getError())
|
| 55 |
|
|
for fit_name, nuis_x in [('b', nuis_b), ('s',nuis_s)]:
|
| 56 |
|
|
if nuis_p != None:
|
| 57 |
|
|
valShift = (nuis_x.getVal() - mean_p)/sigma_p
|
| 58 |
|
|
#sigShift = nuis_x.getError()/sigma_p
|
| 59 |
|
|
print fit_name, name
|
| 60 |
|
|
print valShift
|
| 61 |
|
|
nuiVariation['%s_%s'%(fit_name,name)] = valShift
|
| 62 |
|
|
#print valShift
|
| 63 |
|
|
return nuiVariation
|
| 64 |
|
|
|
| 65 |
|
|
|
| 66 |
|
|
def drawFromDC():
|
| 67 |
|
|
config = BetterConfigParser()
|
| 68 |
|
|
config.read(opts.config)
|
| 69 |
|
|
print config.sections()
|
| 70 |
|
|
region = 'BDT'
|
| 71 |
|
|
var = 'BDT'
|
| 72 |
|
|
ws_var = config.get('plotDef:%s'%var,'relPath')
|
| 73 |
|
|
blind = eval(config.get('Plot:%s'%region,'blind'))
|
| 74 |
|
|
Stack=StackMaker(config,var,region,True)
|
| 75 |
|
|
|
| 76 |
|
|
dataname = ''
|
| 77 |
|
|
if 'Zmm' in opts.bin: dataname = 'Zmm'
|
| 78 |
|
|
elif 'Zee' in opts.bin: dataname = 'Zee'
|
| 79 |
|
|
elif 'Wmn' in opts.bin: dataname = 'Wmn'
|
| 80 |
|
|
elif 'Wen' in opts.bin: dataname = 'Wen'
|
| 81 |
|
|
elif 'Znn' in opts.bin: dataname = 'Znn'
|
| 82 |
|
|
|
| 83 |
|
|
log = eval(config.get('Plot:%s'%region,'log'))
|
| 84 |
|
|
|
| 85 |
|
|
setup = config.get('Plot_general','setup').split(',')
|
| 86 |
|
|
Dict = eval(config.get('LimitGeneral','Dict'))
|
| 87 |
|
|
lumi = eval(config.get('Plot_general','lumi'))
|
| 88 |
|
|
|
| 89 |
|
|
options = copy(opts)
|
| 90 |
|
|
options.dataname = "data_obs"
|
| 91 |
|
|
options.mass = 0
|
| 92 |
|
|
options.format = "%8.3f +/- %6.3f"
|
| 93 |
nmohr |
1.2 |
options.channel = opts.bin
|
| 94 |
nmohr |
1.1 |
options.excludeSyst = []
|
| 95 |
|
|
options.norm = False
|
| 96 |
|
|
options.stat = False
|
| 97 |
|
|
options.bin = True # fake that is a binary output, so that we parse shape lines
|
| 98 |
|
|
options.out = "tmp.root"
|
| 99 |
|
|
options.fileName = args[0]
|
| 100 |
|
|
options.cexpr = False
|
| 101 |
|
|
options.fixpars = False
|
| 102 |
|
|
options.libs = []
|
| 103 |
|
|
options.verbose = 0
|
| 104 |
|
|
options.poisson = 0
|
| 105 |
|
|
options.nuisancesToExclude = []
|
| 106 |
|
|
options.noJMax = None
|
| 107 |
|
|
|
| 108 |
|
|
file = open(opts.dc, "r")
|
| 109 |
|
|
os.chdir(os.path.dirname(opts.dc))
|
| 110 |
|
|
DC = parseCard(file, options)
|
| 111 |
|
|
if not DC.hasShapes: DC.hasShapes = True
|
| 112 |
|
|
MB = ShapeBuilder(DC, options)
|
| 113 |
|
|
theShapes = {}
|
| 114 |
|
|
theSyst = {}
|
| 115 |
|
|
if opts.mlfit:
|
| 116 |
|
|
nuiVar = readBestFit(opts.mlfit)
|
| 117 |
|
|
for b in DC.bins:
|
| 118 |
|
|
if options.channel != None and (options.channel != b): continue
|
| 119 |
|
|
exps = {}
|
| 120 |
|
|
expNui = {}
|
| 121 |
|
|
shapeNui = {}
|
| 122 |
|
|
for (p,e) in DC.exp[b].items(): # so that we get only self.DC.processes contributing to this bin
|
| 123 |
|
|
exps[p] = [ e, [] ]
|
| 124 |
|
|
expNui[p] = [ e, [] ]
|
| 125 |
|
|
for (lsyst,nofloat,pdf,pdfargs,errline) in DC.systs:
|
| 126 |
|
|
if pdf in ('param', 'flatParam'): continue
|
| 127 |
|
|
# begin skip systematics
|
| 128 |
|
|
skipme = False
|
| 129 |
|
|
for xs in options.excludeSyst:
|
| 130 |
|
|
if re.search(xs, lsyst):
|
| 131 |
|
|
skipme = True
|
| 132 |
|
|
if skipme: continue
|
| 133 |
|
|
# end skip systematics
|
| 134 |
|
|
counter = 0
|
| 135 |
|
|
for p in DC.exp[b].keys(): # so that we get only self.DC.processes contributing to this bin
|
| 136 |
|
|
if errline[b][p] == 0: continue
|
| 137 |
|
|
if pdf == 'gmN':
|
| 138 |
|
|
exps[p][1].append(1/sqrt(pdfargs[0]+1));
|
| 139 |
|
|
elif pdf == 'gmM':
|
| 140 |
|
|
exps[p][1].append(errline[b][p]);
|
| 141 |
|
|
elif type(errline[b][p]) == list:
|
| 142 |
|
|
kmax = max(errline[b][p][0], errline[b][p][1], 1.0/errline[b][p][0], 1.0/errline[b][p][1]);
|
| 143 |
|
|
exps[p][1].append(kmax-1.);
|
| 144 |
|
|
elif pdf == 'lnN':
|
| 145 |
|
|
exps[p][1].append(max(errline[b][p], 1.0/errline[b][p])-1.);
|
| 146 |
|
|
if not nuiVar.has_key('%s_%s'%(opts.fit,lsyst)):
|
| 147 |
|
|
nui = 0.
|
| 148 |
|
|
else:
|
| 149 |
|
|
nui= nuiVar['%s_%s'%(opts.fit,lsyst)]
|
| 150 |
|
|
expNui[p][1].append(abs(1-errline[b][p])*nui);
|
| 151 |
|
|
elif ("shape" in pdf) and not 'CMS_vhbb_stats_' in lsyst:
|
| 152 |
|
|
s0 = MB.getShape(b,p)
|
| 153 |
|
|
sUp = MB.getShape(b,p,lsyst+"Up")
|
| 154 |
|
|
sDown = MB.getShape(b,p,lsyst+"Down")
|
| 155 |
|
|
if (s0.InheritsFrom("RooDataHist")):
|
| 156 |
|
|
s0 = ROOT.RooAbsData.createHistogram(s0,ws_var)
|
| 157 |
|
|
s0.SetName(p)
|
| 158 |
|
|
sUp = ROOT.RooAbsData.createHistogram(sUp,ws_var)
|
| 159 |
|
|
sUp.SetName(p+lsyst+'Up')
|
| 160 |
|
|
sDown = ROOT.RooAbsData.createHistogram(sDown,ws_var)
|
| 161 |
|
|
sDown.SetName(p+lsyst+'Down')
|
| 162 |
|
|
theShapes[p] = s0.Clone()
|
| 163 |
|
|
theShapes[p+lsyst+'Up'] = sUp.Clone()
|
| 164 |
|
|
theShapes[p+lsyst+'Down'] = sDown.Clone()
|
| 165 |
|
|
if not nuiVar.has_key('%s_%s'%(opts.fit,lsyst)):
|
| 166 |
|
|
nui = 0.
|
| 167 |
|
|
else:
|
| 168 |
|
|
nui= nuiVar['%s_%s'%(opts.fit,lsyst)]
|
| 169 |
|
|
shapeNui[p] = nui
|
| 170 |
|
|
if counter == 0:
|
| 171 |
|
|
theSyst[lsyst] = s0.Clone()
|
| 172 |
|
|
theSyst[lsyst+'Up'] = sUp.Clone()
|
| 173 |
|
|
theSyst[lsyst+'Down'] = sDown.Clone()
|
| 174 |
|
|
else:
|
| 175 |
|
|
theSyst[lsyst].Add(s0)
|
| 176 |
|
|
theSyst[lsyst+'Up'].Add(sUp.Clone())
|
| 177 |
|
|
theSyst[lsyst+'Down'].Add(sDown.Clone())
|
| 178 |
|
|
counter += 1
|
| 179 |
|
|
procs = DC.exp[b].keys(); procs.sort()
|
| 180 |
|
|
fmt = ("%%-%ds " % max([len(p) for p in procs]))+" "+options.format;
|
| 181 |
|
|
#Compute norm uncertainty and best fit
|
| 182 |
|
|
theNormUncert = {}
|
| 183 |
|
|
theBestFit = {}
|
| 184 |
|
|
for p in procs:
|
| 185 |
|
|
relunc = sqrt(sum([x*x for x in exps[p][1]]))
|
| 186 |
|
|
print fmt % (p, exps[p][0], exps[p][0]*relunc)
|
| 187 |
|
|
theNormUncert[p] = relunc
|
| 188 |
|
|
absBestFit = sum([x for x in expNui[p][1]])
|
| 189 |
|
|
theBestFit[p] = 1.+absBestFit
|
| 190 |
|
|
|
| 191 |
|
|
histos = []
|
| 192 |
|
|
typs = []
|
| 193 |
|
|
|
| 194 |
|
|
setup2=copy(setup)
|
| 195 |
|
|
|
| 196 |
|
|
shapesUp = [[] for _ in range(0,len(setup2))]
|
| 197 |
|
|
shapesDown = [[] for _ in range(0,len(setup2))]
|
| 198 |
|
|
|
| 199 |
|
|
for p in procs:
|
| 200 |
nmohr |
1.2 |
b = opts.bin
|
| 201 |
nmohr |
1.1 |
for s in setup:
|
| 202 |
|
|
if not Dict[s] == p: continue
|
| 203 |
|
|
if 'ZH' == s:
|
| 204 |
|
|
Overlay=copy(theShapes[Dict[s]])
|
| 205 |
|
|
else:
|
| 206 |
|
|
histos.append(theShapes[Dict[s]])
|
| 207 |
|
|
typs.append(s)
|
| 208 |
|
|
print s
|
| 209 |
|
|
for (lsyst,nofloat,pdf,pdfargs,errline) in DC.systs:
|
| 210 |
|
|
if errline[b][p] == 0: continue
|
| 211 |
|
|
if ("shape" in pdf) and not 'CMS_vhbb_stats_' in lsyst:
|
| 212 |
|
|
print 'syst %s'%lsyst
|
| 213 |
|
|
shapesUp[setup2.index(s)].append(theShapes[Dict[s]+lsyst+'Up'])
|
| 214 |
|
|
shapesDown[setup2.index(s)].append(theShapes[Dict[s]+lsyst+'Down'])
|
| 215 |
|
|
|
| 216 |
|
|
#-------------
|
| 217 |
|
|
#Compute absolute uncertainty from shapes
|
| 218 |
|
|
counter = 0
|
| 219 |
|
|
for (lsyst,nofloat,pdf,pdfargs,errline) in DC.systs:
|
| 220 |
|
|
if ("shape" in pdf) and not 'CMS_vhbb_stats_' in lsyst:
|
| 221 |
|
|
theSystUp = theSyst[lsyst+'Up'].Clone()
|
| 222 |
|
|
theSystUp.Add(theSyst[lsyst].Clone(),-1.)
|
| 223 |
|
|
theSystUp.Multiply(theSystUp)
|
| 224 |
|
|
theSystDown = theSyst[lsyst+'Down'].Clone()
|
| 225 |
|
|
theSystDown.Add(theSyst[lsyst].Clone(),-1.)
|
| 226 |
|
|
theSystDown.Multiply(theSystDown)
|
| 227 |
|
|
if counter == 0:
|
| 228 |
|
|
theAbsSystUp = theSystUp.Clone()
|
| 229 |
|
|
theAbsSystDown = theSystDown.Clone()
|
| 230 |
|
|
else:
|
| 231 |
|
|
theAbsSystUp.Add(theSystUp.Clone())
|
| 232 |
|
|
theAbsSystDown.Add(theSystDown.Clone())
|
| 233 |
|
|
counter +=1
|
| 234 |
|
|
|
| 235 |
|
|
#-------------
|
| 236 |
|
|
#Best fit for shapes
|
| 237 |
|
|
for p in procs:
|
| 238 |
|
|
counter = 0
|
| 239 |
|
|
nom = theShapes[p].Clone()
|
| 240 |
|
|
for (lsyst,nofloat,pdf,pdfargs,errline) in DC.systs:
|
| 241 |
|
|
if errline[b][p] == 0: continue
|
| 242 |
|
|
if ("shape" in pdf) and not 'CMS_vhbb_stats_' in lsyst and not 'CMS_vhbb_model_VV' in lsyst:
|
| 243 |
|
|
if shapeNui > 0.:
|
| 244 |
|
|
theVari = 'Up'
|
| 245 |
|
|
else:
|
| 246 |
|
|
theVari = 'Down'
|
| 247 |
|
|
bestNuiVar = theShapes[p+lsyst+theVari].Clone()
|
| 248 |
|
|
bestNuiVar.Add(nom,-1.)
|
| 249 |
|
|
bestNuiVar.Scale(abs(shapeNui[p]))
|
| 250 |
|
|
if counter == 0:
|
| 251 |
|
|
bestNui = bestNuiVar.Clone()
|
| 252 |
|
|
else:
|
| 253 |
|
|
bestNui.Add(bestNuiVar)
|
| 254 |
|
|
counter +=1
|
| 255 |
|
|
nom.Add(bestNui)
|
| 256 |
nmohr |
1.2 |
nom.Scale(theShapes[p].Integral()/nom.Integral())
|
| 257 |
nmohr |
1.1 |
nBins = nom.GetNbinsX()
|
| 258 |
|
|
for bin in range(1,nBins+1):
|
| 259 |
|
|
nom.SetBinError(bin,theShapes[p].GetBinError(bin))
|
| 260 |
|
|
theShapes['%s_%s'%(opts.fit,p)] = nom.Clone()
|
| 261 |
|
|
histos = []
|
| 262 |
|
|
typs = []
|
| 263 |
|
|
for s in setup:
|
| 264 |
|
|
if 'ZH' == s:
|
| 265 |
|
|
Overlay=copy(theShapes[Dict[s]])
|
| 266 |
|
|
else:
|
| 267 |
|
|
histos.append(theShapes['%s_%s'%(opts.fit,Dict[s])])
|
| 268 |
|
|
typs.append(s)
|
| 269 |
|
|
|
| 270 |
|
|
counter = 0
|
| 271 |
|
|
errUp=[]
|
| 272 |
|
|
total=[]
|
| 273 |
|
|
errDown=[]
|
| 274 |
|
|
nBins = histos[0].GetNbinsX()
|
| 275 |
|
|
print 'total bins %s'%nBins
|
| 276 |
|
|
Error = ROOT.TGraphAsymmErrors(histos[0])
|
| 277 |
|
|
theTotalMC = histos[0].Clone()
|
| 278 |
|
|
for h in range(1,len(histos)):
|
| 279 |
|
|
theTotalMC.Add(histos[h])
|
| 280 |
|
|
|
| 281 |
|
|
total = [[]]*nBins
|
| 282 |
|
|
errUp = [[]]*nBins
|
| 283 |
|
|
errDown = [[]]*nBins
|
| 284 |
|
|
for bin in range(1,nBins+1):
|
| 285 |
|
|
binError = theTotalMC.GetBinError(bin)
|
| 286 |
|
|
if math.isnan(binError):
|
| 287 |
|
|
binError = 0.
|
| 288 |
|
|
total[bin-1]=theTotalMC.GetBinContent(bin)
|
| 289 |
|
|
#Stat uncertainty of the MC outline
|
| 290 |
|
|
errUp[bin-1] = [binError]
|
| 291 |
|
|
errDown[bin-1] = [binError]
|
| 292 |
|
|
#Relative norm uncertainty of the individual MC
|
| 293 |
|
|
for h in range(0,len(histos)):
|
| 294 |
|
|
errUp[bin-1].append(histos[h].GetBinContent(bin)*theNormUncert[histos[h].GetName()])
|
| 295 |
|
|
errDown[bin-1].append(histos[h].GetBinContent(bin)*theNormUncert[histos[h].GetName()])
|
| 296 |
|
|
#Shape uncertainty of the MC
|
| 297 |
|
|
for bin in range(1,nBins+1):
|
| 298 |
|
|
#print sqrt(theSystUp.GetBinContent(bin))
|
| 299 |
|
|
errUp[bin-1].append(sqrt(theAbsSystUp.GetBinContent(bin)))
|
| 300 |
|
|
errDown[bin-1].append(sqrt(theAbsSystDown.GetBinContent(bin)))
|
| 301 |
|
|
|
| 302 |
|
|
|
| 303 |
|
|
#Add all in quadrature
|
| 304 |
|
|
totErrUp=[sqrt(sum([x**2 for x in bin])) for bin in errUp]
|
| 305 |
|
|
totErrDown=[sqrt(sum([x**2 for x in bin])) for bin in errDown]
|
| 306 |
|
|
|
| 307 |
|
|
#Make TGraph with errors
|
| 308 |
|
|
for bin in range(1,nBins+1):
|
| 309 |
|
|
if not total[bin-1] == 0:
|
| 310 |
|
|
point=histos[0].GetXaxis().GetBinCenter(bin)
|
| 311 |
|
|
Error.SetPoint(bin-1,point,1)
|
| 312 |
|
|
Error.SetPointEYlow(bin-1,totErrDown[bin-1]/total[bin-1])
|
| 313 |
|
|
print 'down %s'%(totErrDown[bin-1]/total[bin-1])
|
| 314 |
|
|
Error.SetPointEYhigh(bin-1,totErrUp[bin-1]/total[bin-1])
|
| 315 |
|
|
print 'up %s'%(totErrUp[bin-1]/total[bin-1])
|
| 316 |
|
|
|
| 317 |
|
|
#-----------------------
|
| 318 |
|
|
#Read data
|
| 319 |
nmohr |
1.2 |
data0 = MB.getShape(opts.bin,'data_obs')
|
| 320 |
nmohr |
1.1 |
if (data0.InheritsFrom("RooDataHist")):
|
| 321 |
|
|
data0 = ROOT.RooAbsData.createHistogram(data0,ws_var)
|
| 322 |
|
|
data0.SetName('data_obs')
|
| 323 |
|
|
datas=[data0]
|
| 324 |
|
|
datatyps = [None]
|
| 325 |
|
|
datanames=[dataname]
|
| 326 |
|
|
|
| 327 |
|
|
|
| 328 |
|
|
if blind:
|
| 329 |
|
|
for bin in range(10,datas[0].GetNbinsX()+1):
|
| 330 |
|
|
datas[0].SetBinContent(bin,0)
|
| 331 |
|
|
|
| 332 |
|
|
histos.append(copy(Overlay))
|
| 333 |
|
|
typs.append('ZH')
|
| 334 |
|
|
|
| 335 |
|
|
Stack.histos = histos
|
| 336 |
|
|
Stack.typs = typs
|
| 337 |
|
|
Stack.datas = datas
|
| 338 |
|
|
Stack.datatyps = datatyps
|
| 339 |
|
|
Stack.datanames= datanames
|
| 340 |
|
|
Stack.overlay = Overlay
|
| 341 |
|
|
Stack.AddErrors=Error
|
| 342 |
|
|
Stack.lumi = lumi
|
| 343 |
|
|
Stack.doPlot()
|
| 344 |
|
|
|
| 345 |
|
|
print 'i am done!\n'
|
| 346 |
|
|
#-------------------------------------------------
|
| 347 |
|
|
|
| 348 |
|
|
|
| 349 |
|
|
if __name__ == "__main__":
|
| 350 |
|
|
drawFromDC()
|
| 351 |
|
|
sys.exit(0)
|
