def fitNBkg(ibdt, fullbkg):
slope = RooRealVar("slope", "slope", 0.5, -10, 10)
bkg_exp = RooExponential("bkg_exp", "exponential", slope, theBMass)
cut = cut_base + '&& bdt_prob > %s' % (ibdt)
theBMass.setRange('sigRangeMC', B0Mass_ - 3 * dict_sigma[ibdt],
B0Mass_ + 3 * dict_sigma[ibdt])
databkg = fullbkg.reduce(RooArgSet(theBMass, mumuMass, mumuMassE), cut)
r = bkg_exp.fitTo(databkg, RooFit.Save(), ROOT.RooFit.Range('left,right'))
frame = theBMass.frame()
databkg.plotOn(frame, RooFit.Binning(70), RooFit.MarkerSize(.7))
bkg_exp.plotOn(frame, )
# bkg_exp.fixCoefRange('left,right')
nbkg = RooRealVar('nbkg', 'bkg n', 1000, 0, 550000)
ebkg = RooExtendPdf('ebkg', 'ebkg', bkg_exp, nbkg, 'sigRangeMC')
ebkg.fitTo(databkg, ROOT.RooFit.Range('left,right'))
ebkg.plotOn(frame, RooFit.LineStyle(ROOT.kDashed),
RooFit.LineColor(ROOT.kGreen + 1), RooFit.Range(4.9, 5.6))
frame.Draw()
dict_b_v1[ibdt] = [nbkg.getVal(), nbkg.getError()]
def fitNBkg(ibdt, fullbkg, isample):
slope = RooRealVar("slope", "slope", 0.5, -10, 10)
bkg_exp = RooExponential("bkg_exp", "exponential", slope, theBMass)
cut = cut_base + '&& bdt_prob > %s' % (ibdt)
theBMass.setRange('sigRangeMC', B0Mass_ - 3 * dict_sigma[ibdt],
B0Mass_ + 3 * dict_sigma[ibdt])
databkg = fullbkg.reduce(RooArgSet(theBMass, mumuMass, mumuMassE), cut)
r = bkg_exp.fitTo(databkg, RooFit.Save(), ROOT.RooFit.Range('left,right'),
RooFit.PrintLevel(-1))
frame = theBMass.frame()
databkg.plotOn(frame, RooFit.Binning(70), RooFit.MarkerSize(.7))
bkg_exp.plotOn(frame, )
canv = ROOT.TCanvas()
frame.Draw()
nbkg = RooRealVar('nbkg', 'bkg n', 1000, 0, 550000)
ebkg = RooExtendPdf(
'ebkg', 'ebkg', bkg_exp, nbkg,
'sigRangeMC') ## here imposing the range to calculate bkg yield
ebkg.fitTo(databkg, ROOT.RooFit.Range('left,right'), RooFit.PrintLevel(-1))
ebkg.plotOn(frame, RooFit.LineStyle(ROOT.kDashed),
RooFit.LineColor(ROOT.kGreen + 1), RooFit.Range(4.9, 5.6))
frame.Draw()
# canv.SaveAs('bkg_fit_bdt%f_sample%i.pdf'%(ibdt,isample))
dict_b_v1[ibdt] = [nbkg.getVal(), nbkg.getError()]
def main(infiles=None):
infile = infiles[0]
var = "leadmupt"
bounds = [25, 300]
c1 = ROOT.TCanvas("NLL", "NLL", 1000, 1000)
x = ROOT.RooRealVar(var, var, bounds[0], bounds[1])
aset = ROOT.RooArgSet(x, "aset")
frame = x.frame()
f = ROOT.TFile.Open(infile)
tree = f.Get(f.GetListOfKeys().At(0).GetName())
tree.Print()
nentries = tree.GetEntries()
y = []
dh2 = ROOT.TH1F()
data = ROOT.RooDataSet("Data", "Data", aset)
for n in range(nentries):
tree.GetEntry(n)
y.append(getattr(tree, var))
if y[n] <= bounds[1] and y[n] >= bounds[0]:
x.setVal(y[n])
data.add(aset)
dh2.Fill(y[n])
data.plotOn(frame)
dh = RooDataHist("dh", "dh", RooArgSet(x), data)
nbins = dh2.GetNbinsX()
nbinsy = dh2.GetNbinsX()
print("nbins: ", nbins)
print("nbinsy: ", nbinsy)
for i in range(nbins):
if dh2.GetBinContent(dh2.GetBin(i)) == 0:
print("bin: ", i)
#dh2.SetBinError(bin,0.01)
## CREATE GAUSSIAN MODEL
mx = RooRealVar("mx", "mx", 10, 0, 350)
sx = RooRealVar("sx", "sx", 3, 0, 10)
gx = RooGaussian("gx", "gx", x, mx, sx)
## CREATE EXPONENTIAL MODEL
lambda1 = RooRealVar("lambda1", "slope1", -100, 100)
expo1 = RooExponential("expo1", "exponential PDF 1", x, lambda1)
lambda2 = RooRealVar("lambda2", "slope2", -.03, -1000, 1000)
expo2 = RooExponential("expo2", "exponential PDF 2", x, lambda2)
l1 = RooRealVar("l1", "yield1", 100, 0, 10000)
l2 = RooRealVar("l2", "yield2", 100, 0, 10000)
#sum = RooAddPdf("sum","exp and gauss",RooArgList(expo1,gx),RooArgList(l1,l2))
sum = RooAddPdf("sum", "2 exps", RooArgList(expo1, expo2),
RooArgList(l1, l2))
## Construct binned likelihood
nll = RooNLLVar("nll", "nll", expo1, data, ROOT.RooFit.Extended(True))
## Start Minuit session on NLL
m = RooMinuit(nll)
m.migrad()
m.hesse()
r1 = m.save()
#sum.plotOn(frame,ROOT.RooFit.LineColor(1))
#sum.plotOn(frame,ROOT.RooFit.Components("expo1"),ROOT.RooFit.LineColor(2))
#sum.plotOn(frame,ROOT.RooFit.Components("expo2"),ROOT.RooFit.LineColor(3))
expo1.plotOn(frame)
## Construct Chi2
chi2 = RooChi2Var("chi2", "chi2", expo2, dh)
## Start Minuit session on Chi2
m2 = RooMinuit(chi2)
m2.migrad()
m2.hesse()
r2 = m2.save()
frame.Draw()
c2 = ROOT.TCanvas("Chi2", "Chi2", 1000, 1000)
frame2 = x.frame()
data.plotOn(frame2)
expo2.plotOn(frame2)
#sum.plotOn(frame2,ROOT.RooFit.LineColor(4))
#sum.plotOn(frame2,ROOT.RooFit.Components("expo1"),ROOT.RooFit.LineColor(5))
#sum.plotOn(frame2,ROOT.RooFit.Components("expo2"),ROOT.RooFit.LineColor(6))
## Print results
print("result of likelihood fit")
r1.Print("v")
print("result of chi2 fit")
r2.Print("v")
frame2.Draw()
c1.Draw()
c2.Draw()
rep = ''
while not rep in ['q', 'Q']:
rep = input('enter "q" to quit: ')
if 1 < len(rep):
rep = rep[0]
