def main():
from optparse import OptionParser
parser = OptionParser()
parser.add_option("-i", "--inputfile", dest="inputfile")
parser.add_option("-l", "--logfile", dest="logfile")
parser.add_option("-N", "--multiplicity", dest="N", type="int")
parser.add_option("", "--ST", dest="ST", type="int")
(options, args) = parser.parse_args()
from ROOT import gROOT, TFile, kFALSE, TGraph, TVectorD
gROOT.ProcessLine(".L roostats_cl95.C+")
from ROOT import roostats_cl95, roostats_cla, roostats_limit
import configurations as config
ilum = config.integrated_luminosity
slum = ilum * config.relative_luminosity_uncertainty
seff = config.nominal_signal_uncertainty
infile = TFile(options.inputfile, "READ")
from HistoStore import HistoStore
store = HistoStore()
hIntBkg = infile.Get("IntegralBackground_N%dup" % options.N)
hIntData = infile.Get("IntegralData_N%dup" % options.N)
ibin = hIntData.FindBin(options.ST)
nBkg = hIntBkg.GetBinContent(ibin)
sBkg = hIntBkg.GetBinError(ibin)
nData = int(hIntData.GetBinContent(ibin))
print "%10d %10.2f +/- %10.2f" % (nData, nBkg, sBkg)
print ilum, slum, 1.0, seff, nBkg, sBkg, nData
#CLs method
rl = roostats_limit(ilum, slum, 1.0, seff, nBkg, sBkg, nData, kFALSE, 1,
"cls", "my.png", 23576)
cl95 = rl.GetObservedLimit()
cla = rl.GetExpectedLimit()
exp_up = rl.GetOneSigmaHighRange()
exp_down = rl.GetOneSigmaLowRange()
exp_2up = rl.GetTwoSigmaHighRange()
exp_2down = rl.GetTwoSigmaLowRange()
#WAS Bayesian a la:
#cl95 = roostats_cl95(ilum, slum, 1.0, seff, nBkg, sBkg, nData, kFALSE, 1, "bayesian", "")
#cla = roostats_cla(ilum, slum, 1.0, seff, nBkg, sBkg, 1)
logfile = open(options.logfile, "w")
logfile.write(
"%-10d %-10.5f %-10.5f %-10.5f %-10.5f %-10.5f %-10.5f\n" %
(options.ST, cl95, cla, exp_up, exp_down, exp_2up, exp_2down))
logfile.close()
from ModelParser import XsecParser, ModelGroup
import csv
from ROOT import TGraphErrors, TVectorD
from HistoStore import HistoStore
store = HistoStore()
models = []
with open("xsec_tmp2.txt", "rb") as f:
reader = csv.reader(f, delimiter=" ", skipinitialspace=True)
for row in reader:
model = XsecParser(row)
models.append(model)
group = ModelGroup(models)
for generator, group_generator in group.items():
for n, group_n in group_generator.items():
for icolor, (MD, models) in enumerate(group_n.items()):
vsize = len(models)
vx = TVectorD(vsize)
vxsec = TVectorD(vsize)
vex = TVectorD(vsize)
vey = TVectorD(vsize)
for i, m in enumerate(models):
vx[i] = m.M
vex[i] = 0.0
vxsec[i] = m.xsec
vey[i] = m.xsecErr
def main():
from optparse import OptionParser
parser = OptionParser()
parser.add_option("-i", "--inputfile", dest="inputfile")
parser.add_option("-o", "--ouputfile", dest="outputfile")
parser.add_option("-b", "--batch", action="store_true",\
dest="isBatch", default=False)
parser.add_option("--normalization", nargs=2,\
type="float", dest="norm_range")
parser.add_option("--fit", nargs=2,\
type="float", dest="fit_range")
(options, args) = parser.parse_args()
isSaveOutput = options.outputfile is not None
if not (options.inputfile):
parser.error("Please specify inputfiles.")
import configurations as config
if options.fit_range:
fit_range = options.fit_range
norm_range = (fit_range[1] - 200., fit_range[1])
else:
fit_range = config.fit_range
norm_range = config.norm_range
# Override normalization range from input
if options.norm_range:
norm_range = options.norm_range
from Styles import formatST, formatTemplate, formatUncertainty
from ROOT import TFile, TF1, TH1D, TMath, TCanvas, TLegend,\
TGraphAsymmErrors, TVectorD, gStyle
gStyle.SetPadTopMargin(0.05)
gStyle.SetPadRightMargin(0.05)
#gStyle.SetOptFit(1111111)
#input file name
infile = TFile(options.inputfile, "READ")
from HistoStore import HistoStore
store = HistoStore()
canvas = HistoStore()
print "Fit range: %d - %d GeV" % fit_range
print "Normalization range: %d - %d GeV" % norm_range
# Fit
for N in config.exclusive_multiplicities:
hST = infile.Get("plots%dJets/ST" % N)
if not options.isBatch:
c = TCanvas("TemplateN%d" % N, "TemplateN%d" % N, 500, 500)
canvas.book(c)
formatST(hST)
hST.Draw("e")
hST.GetXaxis().SetRangeUser(fit_range[0], config.maxST)
hST.GetYaxis().SetRangeUser(1e-2, 2e4)
hST.GetYaxis().SetTitleOffset(1.2)
c.SetLogy(1)
for i, formula in enumerate(config.templates):
if N == 2:
f = TF1("templateN%d_%d" % (N, i), formula, 0, 10000)
elif N == 3:
f = store.get("templateN2_%d" % i).Clone("templateN%d_%d" %
(N, i))
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
hST.Fit(f, "QN0", "", fit_range[0], fit_range[1])
if i == 0:
hST.Fit(f, "Q0", "", fit_range[0], fit_range[1])
formatTemplate(f, N, i)
store.book(f)
if not options.isBatch:
f.Draw("same")
hTemplate = hST.Clone("histoTemplateN%d_%d" % (N, i))
hTemplate.Reset()
hTemplate.Eval(f)
formatTemplate(hTemplate, N, i)
store.book(hTemplate)
if i == 0:
hRef = hTemplate.Clone("ReferenceTemplateN%d_0" % N)
store.book(hRef)
# Print Chi-squre/Ndof
print "N = %d, Chi^2/Ndof = %0.2f/%d" %\
(N, f.GetChisquare(), f.GetNDF())
if not options.isBatch:
c.Update()
# Calculate scale/error
from OptimizationTools import OptimizeScale
for histoN, templateN in [[2, 3]]:
hST = store.get("ReferenceTemplateN%d_0" % histoN)
hTemplate = store.get("ReferenceTemplateN%d_0" % templateN)
hlnL, scale, err = OptimizeScale(hST, hTemplate, norm_range)
hlnL.SetName("LogLikelihood_%dto%d" % (templateN, histoN))
store.book(hlnL)
for i in range(len(config.templates)):
hTemplate = store.get("histoTemplateN%d_%d" % (templateN, i))
hTemplate_ = hTemplate.Clone("histoTemplateN%d_%d__RescaledToN%d" %
(templateN, i, histoN))
hTemplate_.Scale(scale)
store.book(hTemplate_)
# Shape Uncertainty
hBkgTemplate = store.get("histoTemplateN2_0")
hBkgTemplate.Rebin(config.rebin)
nbins = hBkgTemplate.GetNbinsX()
vST = TVectorD(nbins)
vBkg = TVectorD(nbins)
vexl = TVectorD(nbins)
vexh = TVectorD(nbins)
shape_el = TVectorD(nbins)
shape_eh = TVectorD(nbins)
rel_shape_el = TVectorD(nbins)
rel_shape_eh = TVectorD(nbins)
for i in range(nbins):
vST[i] = hBkgTemplate.GetBinCenter(i + 1)
if (vST[i] < config.com):
vBkg[i] = hBkgTemplate.GetBinContent(i + 1)
else:
vBkg[i] = 0.0
vexl[i] = 0.0
vexh[i] = 0.0
shape_el[i] = 0.0
shape_eh[i] = 0.0
rel_shape_el[i] = 0.0
rel_shape_eh[i] = 0.0
for i in range(len(config.templates)):
for label in [
"histoTemplateN2_%d", "histoTemplateN3_%d__RescaledToN2"
]:
if label % i == "histoTemplateN2_0":
continue
h = store.get(label % i)
h.Rebin(config.rebin)
for ibin in range(nbins):
diff = h.GetBinContent(ibin + 1) - vBkg[ibin]
if diff > 0 and diff > shape_eh[ibin]:
shape_eh[ibin] = diff
elif diff < 0 and abs(diff) > shape_el[ibin]:
shape_el[ibin] = abs(diff)
# Relative Shape Uncertaincy
for i in range(nbins):
if vBkg[i] > 0:
#rel_shape_el[i] = rel_shape_el[i] / vBkg[i]
#hape_eh[i] = rel_shape_eh[i] / vBkg[i]
max_err = max(shape_el[i], shape_eh[i])
shape_el[i] = max_err
shape_eh[i] = max_err
rel_shape_el[i] = max_err / vBkg[i]
rel_shape_eh[i] = max_err / vBkg[i]
else:
rel_shape_el[i] = 0.0
rel_shape_eh[i] = 0.0
#print vST[i], vBkg[i], rel_shape_el[i], rel_shape_eh[i]
gShapeUncertainty = TGraphAsymmErrors(vST, vBkg, vexl, vexh, shape_el,
shape_eh)
gShapeUncertainty.SetName("Shape_Uncertainty")
formatUncertainty(gShapeUncertainty)
store.book(gShapeUncertainty)
gRelShapeUncertainty = TGraphAsymmErrors(vST, vexl, vexl, vexh,
rel_shape_el, rel_shape_eh)
gRelShapeUncertainty.SetName("Relative_Shape_Uncertainty")
formatUncertainty(gRelShapeUncertainty)
store.book(gRelShapeUncertainty)
# Generate Backgrouds
for N in config.label_for_data:
hST = infile.Get("plotsN%s/ST" % N)
rel_scale_err2 = 0.0
scale_factor = 1.0
for Nref in config.label_for_ref:
if N == Nref:
continue
template = store.get("ReferenceTemplateN%s_0" % Nref)
hlnL, scale, err = OptimizeScale(hST, template, norm_range)
hlnL.SetName("LogLikelihood_%sto%s" % (Nref, N))
store.book(hlnL)
if Nref == "2":
scale_factor = scale
rel_scale_err2 += err / scale * err / scale
print "%s/%s %.3f +/- %.3f" % (N, Nref, scale, err)
vy = TVectorD(nbins)
veyh = TVectorD(nbins)
veyl = TVectorD(nbins)
for i in range(nbins):
vy[i] = vBkg[i] * scale_factor
veyh[i] = vy[i] * TMath.Sqrt(rel_scale_err2 +
rel_shape_eh[i] * rel_shape_eh[i])
veyl[i] = vy[i] * TMath.Sqrt(rel_scale_err2 +
rel_shape_el[i] * rel_shape_el[i])
print "Scaling uncertainty (%s): %.2f" %\
(N, TMath.sqrt(rel_scale_err2) * 100.0)
gBkg = TGraphAsymmErrors(vST, vy, vexl, vexh, veyl, veyh)
gBkg.SetName("BackgroundGraph_N%s" % N)
formatUncertainty(gBkg)
store.book(gBkg)
hST.Rebin(config.rebin)
hST.SetName("Data_N%s" % N)
formatST(hST)
store.book(hST)
hBkg = hST.Clone("Background_N%s" % N)
hBkg.Reset()
store.book(hBkg)
for i in range(nbins):
ibin = hBkg.FindBin(vST[i])
hBkg.SetBinContent(ibin, vy[i])
hBkg.SetBinError(ibin, max(veyh[i], vexl[i]))
from OptimizationTools import Integral
hIntBkg = hBkg.Clone("IntegralBackground_N%s" % N)
Integral(hIntBkg)
store.book(hIntBkg)
hIntData = hST.Clone("IntegralData_N%s" % N)
Integral(hIntData)
store.book(hIntData)
# Plot Shape Uncertainty
if not options.isBatch:
legend_shape = TLegend(0.5544355, 0.5741525, 0.9395968, 0.9152542)
legend_shape.SetTextFont(42)
legend_shape.SetFillColor(0)
legend_shape.SetLineColor(0)
c = TCanvas("ShapeUncertaintyN2", "ShapeUncertaintyN2", 500, 500)
canvas.book(c)
gShapeUncertainty.Draw("AC3")
gShapeUncertainty.GetXaxis().SetRangeUser(fit_range[0], config.maxST)
gShapeUncertainty.GetYaxis().SetRangeUser(5e-2, 1.2e6)
legend_shape.AddEntry(store.get("Data_N2"), "Data (N = 2)", "p")
legend_shape.AddEntry(gShapeUncertainty, "Shape Uncertainty", "f")
for i in range(len(config.templates)):
for label in [
"histoTemplateN2_%d", "histoTemplateN3_%d__RescaledToN2"
]:
h = store.get(label % i)
h.GetXaxis().SetRangeUser(fit_range[0], config.maxST)
h.Draw("histcsame")
if label == "histoTemplateN2_%d":
N = 2
else:
N = 3
legend_shape.AddEntry(h,
"Parametrization %d (N = %d)" % (i, N),
"l")
store.get("Data_N2").Draw("esame")
c.SetLogy(1)
legend_shape.Draw("plain")
c.Update()
if isSaveOutput:
store.saveAs(options.outputfile)
if not options.isBatch:
raw_input("Press Enter to continue...")
