def plcLimit(obs_, poi_, model, ws, data, CL=0.95, verbose=False):
# obs : observable variable or RooArgSet of observables
# poi : parameter of interest or RooArgSet of parameters
# model : RooAbsPdf of model to consider including any constraints
# data : RooAbsData of the data
# CL : confidence level for interval
# returns a dictionary with the upper and lower limits for the first/only
# parameter in poi_ as well as the interval object and status flag
obs = RooArgSet(obs_)
obs.setName('observables')
poi = RooArgSet(poi_)
poi.setName('poi')
poi.setAttribAll('Constant', False)
nuis = model.getParameters(obs)
nuis.remove(poi)
nuis.remove(nuis.selectByAttrib('Constant', True))
nuis.setName('nuisance')
if verbose:
print 'observables'
obs.Print('v')
print 'parameters of interest'
poi.Print('v')
print 'nuisance parameters'
nuis.Print('v')
mc = RooStats.ModelConfig('mc')
mc.SetWorkspace(ws)
mc.SetPdf(model)
mc.SetObservables(obs)
mc.SetParametersOfInterest(poi)
mc.SetNuisanceParameters(nuis)
plc = RooStats.ProfileLikelihoodCalculator(data, mc)
plc.SetConfidenceLevel(CL)
interval = plc.GetInterval()
upperLimit = Double(999.)
lowerLimit = Double(0.)
Limits = {}
paramIter = poi.createIterator()
param = paramIter.Next()
while param:
ok = interval.FindLimits(param, lowerLimit, upperLimit)
Limits[param.GetName()] = {
'ok': ok,
'upper': float(upperLimit),
'lower': float(lowerLimit)
}
param = paramIter.Next()
if verbose:
print '%.0f%% CL limits' % (interval.ConfidenceLevel() * 100)
print Limits
Limits['interval'] = interval
return Limits
def GetPLRInterval(filename="workspace.root",
wsname='myWS',
interactive=False):
# this function loads a workspace and computes
# a Bayesian upper limit
pInFile = ROOT.TFile(filename, "read")
# load workspace
pWs = pInFile.Get(wsname)
if not pWs:
print "workspace ", wsname, " not found"
return -1
# printout workspace content
pWs.Print()
# load and print data from workspace
data = pWs.data("data")
data.Print()
# load and print S+B Model Config
pSbHypo = pWs.obj("SbHypo")
pSbHypo.Print()
# create RooStats Profiled Likelihood ratio calculator and set parameters
bCalc = RooStats.ProfileLikelihoodCalculator(data, pSbHypo)
bCalc.SetConfidenceLevel(0.95)
# estimate credible interval
# NOTE: unfortunate notation: the UpperLimit() name refers
# to the upper boundary of an interval,
# NOT to the upper limit on the parameter of interest
# (it just happens to be the same for the one-sided
# interval starting at 0)
pSInt = bCalc.GetInterval()
upper_bound = pSInt.UpperLimit(pWs.var('xsec'))
lower_bound = pSInt.LowerLimit(pWs.var('xsec'))
print "one-sided 95%.C.L. interval for xsec: ", "[", lower_bound, ", ", upper_bound, "]"
# make posterior PDF plot for POI
c1 = ROOT.TCanvas("plr", "likelihood furnction")
pPlot = RooStats.LikelihoodIntervalPlot(pSInt)
pPlot.SetRange(0., 0.006)
pPlot.SetMaximum(5.)
pPlot.SetNPoints(50)
pPlot.Draw()
ROOT.gPad.Update()
c1.SaveAs("plr_plot.pdf")
if interactive:
raw_input("\npress <enter> to continue")
def calculate_cls(self, workspace_name):
"""
returns a dictionary of CLs values
"""
from scharmfit import utils
utils.load_susyfit()
from ROOT import Util
from ROOT import RooStats
if not isfile(workspace_name):
raise OSError("can't find workspace {}".format(workspace_name))
workspace = Util.GetWorkspaceFromFile(workspace_name, 'combined')
Util.SetInterpolationCode(workspace,4)
# NOTE: We're completely silencing the fitter. Add an empty string
# to the accept_strings to get all output.
with OutputFilter(accept_strings={}):
limit = RooStats.get_Pvalue(
workspace,
True, # doUL
1, # n_toys
2, # asymtotic calculator
3, # test type (3 is atlas standard)
)
return dict(
cls=limit.GetCLs(),
cls_exp=limit.GetCLsexp(),
cls_up_1_sigma=limit.GetCLsu1S(),
cls_down_1_sigma=limit.GetCLsd1S(),
cls_up_2_sigma=limit.GetCLsu2S(),
cls_down_2_sigma=limit.GetCLsd2S())
def GetBayesianInterval(filename="workspace.root",
wsname='myWS',
interactive=False):
# this function loads a workspace and computes
# a Bayesian upper limit
pInFile = ROOT.TFile(filename, "read")
# load workspace
pWs = pInFile.Get(wsname)
if not pWs:
print "workspace ", wsname, " not found"
return -1
# printout workspace content
pWs.Print()
# load and print data from workspace
data = pWs.data("data")
data.Print()
# load and print S+B Model Config
pSbHypo = pWs.obj("SbHypo")
pSbHypo.Print()
# create RooStats Bayesian calculator and set parameters
bCalc = RooStats.BayesianCalculator(data, pSbHypo)
bCalc.SetName("myBC")
bCalc.SetConfidenceLevel(0.95)
bCalc.SetLeftSideTailFraction(0.0)
#bcalc->SetIntegrationType("ROOFIT")
# estimate credible interval
# NOTE: unfortunate notation: the UpperLimit() name refers
# to the upper boundary of an interval,
# NOT to the upper limit on the parameter of interest
# (it just happens to be the same for the one-sided
# interval starting at 0)
pSInt = bCalc.GetInterval()
upper_bound = pSInt.UpperLimit()
lower_bound = pSInt.LowerLimit()
print "one-sided 95%.C.L. bayesian credible interval for xsec: ", "[", lower_bound, ", ", upper_bound, "]"
# make posterior PDF plot for POI
c1 = ROOT.TCanvas("posterior", "posterior")
bCalc.SetScanOfPosterior(100)
pPlot = bCalc.GetPosteriorPlot()
pPlot.Draw()
ROOT.gPad.Update()
c1.SaveAs("bayesian_num_posterior.pdf")
if interactive:
raw_input("\npress <enter> to continue")
def get_significance(s, b1, b2, err_b1, err_b2, threshold):
'''
asimov significance
https://root.cern.ch/doc/v614/RooStatsUtils_8cxx_source.html
https://www.pp.rhul.ac.uk/~cowan/stat/medsig/medsigNote.pdf
eq[10], eq[20]
'''
err_b1_b2 = eff_error(b1, b2, err_b1, err_b2, threshold)
nB = max(b1 + b2, threshold)
significance = 0.
if s > 1.1e-10:
# value smaller than this cause NaN value when use vfunc
# please think and figure out a better way to treat negative weight
significance = RooStats.AsimovSignificance(s, nB, err_b1_b2)
return err_b1_b2, significance
def execute(self, ws, debug = 0):
print self.legend, 'Profile Likelihood calculation started...'
# time the action
t = TStopwatch()
t.Start()
# model config is guaranteed to be here
# the action won't configure without it
# and this code will never run unless valid model config is found
mconf = ws.obj(self._model_config_name)
_poi = mconf.GetParametersOfInterest()
plcInt = self._plc.GetInterval()
# stop watch and print how long it took to get the interval
t.Print()
# iterate over all parameters of interest and print out the intervals
# (could there be more than one?)
_iter = _poi.createIterator()
while True:
_poi_name = _iter().GetTitle()
lower_limit = plcInt.LowerLimit( _poi[_poi_name] )
upper_limit = plcInt.UpperLimit( _poi[_poi_name] )
print self.legend, 'Profile Likelihood interval for', _poi_name, 'is ['+ \
str(lower_limit) + ', ' + \
str(upper_limit) + ']'
if _iter.Next() == None:
break
if self._scan_plot:
# draw scan plot
print self.legend, 'making the likelihood scan plot'
_plot_name = _poi_name+'_plc_scan_exost.'+self._plot_format
c1 = TCanvas("c1", "c1", 600, 600)
plcPlot = RooStats.LikelihoodIntervalPlot(plcInt)
gROOT.SetStyle("Plain")
plcPlot.Draw()
c1.SaveAs(_plot_name)
return (lower_limit, upper_limit)
def model(self, ws, debug=0):
if self.sModel not in self.items:
if debug > 0:
print self.legend, 'no Model section defined in the config file'
print self.legend, 'no model will be created'
return
legend = '[' + self.sModel + ']:'
# FIXME: find a better way to load custom PDFs
#gROOT.ProcessLine(".L dijetQstarPdf.cxx+");
#gROOT.ProcessLine(".L Qstar_qg.cxx+");
#gROOT.ProcessLine(".L Qstar_qg_2.cxx+");
#gROOT.ProcessLine(".L Jacobian_mt.cxx+");
model_items = self.items[self.sModel]
if (debug > 0):
print 'Loading model to the workspace', ws.GetName()
# using RooStats::HLFactory to parse the model
# make a temp file with a card file
card_file_name = 'hlfCardFile.rs'
with open(card_file_name, 'w') as cardfile:
for item in model_items:
print >> cardfile, item[0], '=', item[1]
hlfVerbosity = debug > 0
#hlf = ROOT.RooStats.HLFactory('HLFactoryExost', card_file_name, True)
hlf = RooStats.HLFactory('HLFactoryExost', ws, hlfVerbosity)
self.modelAutoImportWarningExplanation(legend, debug)
current_messaging_level = RooMsgService.instance().globalKillBelow()
if debug <= 0:
RooMsgService.instance().setGlobalKillBelow(RooFit.ERROR)
hlf.ProcessCard(card_file_name)
print legend, 'importing nonstandard class code into the workspace...'
ws.importClassCode()
#ws.importClassCode('Qstar_qg', true);
print legend, '...code import done'
# return messaging to the original level
RooMsgService.instance().setGlobalKillBelow(current_messaging_level)
def limitBayesian(w=initializeWorkspace(), B=1, dB=0.35, S=1, dS=0.1):
w.var("B").setVal(B)
w.var("dB").setVal(dB)
w.var("S").setVal(S)
w.var("dS").setVal(dS)
w.var("N").setVal(B)
obs = ROOT.RooArgSet(w.var("N"))
data = ROOT.RooDataSet("data", "data", obs)
data.add(obs) # insert entry
rbackup = w.var("r").getMax()
nuisances = ROOT.RooArgSet()
if dB != 0:
nuisances.add(w.var("thetaB"))
w.var("thetaB").setConstant(False)
else:
w.var("thetaB").setConstant(True)
if dS != 0:
nuisances.add(w.var("thetaS"))
w.var("thetaS").setConstant(False)
else:
w.var("thetaS").setConstant(True)
poi = ROOT.RooArgSet(w.var("r"))
model = w.pdf({True: "model_s", False: "modelStat_s"}[dS > 0 or dB > 0])
prior = w.pdf("prior")
nuisSet = {True: nuisances, False: None}[dS > 0 or dB > 0]
bcalc = RooStats.BayesianCalculator(data, model, poi, prior, nuisSet)
bcalc.SetLeftSideTailFraction(0)
bcalc.SetConfidenceLevel(0.95)
interval = bcalc.GetInterval()
ret = interval.UpperLimit()
while ret > 0.5 * w.var("r").getMax():
#print "Got ",ret,"too close to the upper bound",w.var("r").getMax()
if (w.var("r").getMax() > 200): break
w.var("r").setMax(w.var("r").getMax() * 2)
interval = bcalc.GetInterval()
ret = interval.UpperLimit()
w.var("r").setMax(rbackup)
return ret
def calculate_cls(self, workspace_name):
"""
returns a dictionary of CLs values
"""
from scharmfit import utils
utils.load_susyfit()
from ROOT import Util
from ROOT import RooStats
if not isfile(workspace_name):
raise OSError("can't find workspace {}".format(workspace_name))
workspace = Util.GetWorkspaceFromFile(workspace_name, 'combined')
Util.SetInterpolationCode(workspace,4)
# NOTE: We're completely silencing the fitter. Add an empty string
# to the accept_strings to get all output.
with OutputFilter(accept_strings={}):
limit = RooStats.get_Pvalue(
workspace,
True, # doUL
1, # n_toys
2, # asymtotic calculator
3, # test type (3 is atlas standard)
)
ws_type = workspace_name.rsplit('_',1)[1].split('.')[0]
if ws_type == self.nominal:
return {
'obs':limit.GetCLs(),
'exp':limit.GetCLsexp(),
'exp_u1s':limit.GetCLsu1S(),
'exp_d1s':limit.GetCLsd1S(),
}
elif ws_type == self.up1s:
return {'obs_u1s':limit.GetCLs()}
elif ws_type == self.down1s:
return {'obs_d1s':limit.GetCLs()}
# should never get here
raise ValueError('can\'t classify {} as type of limit'.format(
workspace_name))
def GetLimits(tl,f):
from ROOT import RooStats,Util
#w=gDirectory.Get("w")
print "analysis name: ",tl.name
print "workspace name: ",tl.wsFileName
if not ("SU" in tl.name):
print "Do no hypothesis test for bkg only or discovery fit!\n"
return
print "Need to load workspace"
Util.ReadWorkspace(tl.wsFileName,"combined")
w=gDirectory.Get("w")
result = RooStats.MakeUpperLimitPlot(tl.name,w,2,3,1000,True,20,True)
if not result==0:
result.Print()
print result.UpperLimit()
return
b0SideSub.SetMarkerStyle(kFullCircle)
b0SideSub.SetMarkerColor(kBlack)
b0SideSub.SetMarkerSize(0.8)
b0SideSub.SetLineColor(kBlack)
b0SideSub.Draw("E0")
linezero = TLine(b0SideSub.GetBinCenter(1), 0.0,
b0SideSub.GetBinCenter(b0SideSub.GetNbinsX()), 0.0)
linezero.SetLineColor(kRed)
linezero.SetLineWidth(2)
linezero.SetLineStyle(kDotted)
linezero.Draw()
kcanvas.SaveAs(region + "/b0s_subtracted_" + region + ".png")
kcanvas.SaveAs(region + "/b0s_subtracted_" + region + ".root")
splot = RooStats.SPlot("sPlot", "sPlot", theData, kkTot,
RooArgList(nSigKK, nBkgKK))
dstree = theData.store().tree()
dstree.GetEntryNumber(88)
sPlot_B0_hist = TH1F("sPlot_B0_hist",
"#mu#muKK sPlot - " + region + "; m(#mu#muKK) [GeV]",
2000, 4.00, 6.0)
sPlot_B0_hist.Sumw2()
sPlot_B0_hist.SetLineColor(2)
sPlot_B0_hist.SetMarkerColor(2)
sPlot_B0_hist.SetMinimum(0.)
dstree.Project("sPlot_B0_hist", "xM", "nSig_sw")
sPlot_B0_hist.Draw("e0")
def getProfile(idx=None, update=False):
from ROOT import TFile, TCanvas
from ROOT import RooStats as RS
f = TFile("%s/data/fitWorkspace.root" % dsi.latSWDir)
fitWorkspace = f.Get("fitWorkspace")
fData = fitWorkspace.allData().front()
hitE = fitWorkspace.var("trapENFCal")
model = fitWorkspace.pdf("model")
fitResult = fitWorkspace.allGenericObjects().front()
fPars = fitResult.floatParsFinal()
nPars = fPars.getSize()
# === get fit results: {name : [nCts, err]} ===
fitVals = {}
for i in range(nPars):
fp = fitResult.floatParsFinal()
name = fp.at(i).GetName()
fitVal, fitErr = fp.at(i).getValV(), fp.at(i).getError()
if "amp" in name:
# fitVals[name.split('-')[1]] = [fitVal, fitErr]
fitVals[name] = [fitVal, fitErr, name.split('-')[1]]
# for f in fitVals:
# print(f, fitVals[f])
# === get "true" counts (reverse efficiency correction based on fit value) ===
hAx = getHistList("axion")
x, y, xpb = wl.npTH1D(hAx)
x, y = normPDF(x, y, eLo, eHi)
nCts, nErr, _ = fitVals["amp-axion"] # fit result
yc = nCts * getEffCorr(x, y, inv=True)
nCorr = np.sum(yc)
effCorr = nCorr / nCts
print("nCts %d nCorr %.2f effCorr %.2f" % (nCts, nCorr, effCorr))
# thesis plot, don't delete
# plt.step(x, y * nCts * (epb/xpb), c='r', lw=2, label="Eff-weighted: %.1f" % (nCts))
# plt.step(x, yc * (epb/xpb), c='b', lw=2, label="True counts: %.1f" % (nCorr))
# plt.xlabel("Energy (keV)", ha='right', x=1)
# plt.ylabel("Counts / keV", ha='right', y=1)
# plt.legend(loc=1)
# plt.tight_layout()
# plt.show()
tMode = "UPDATE" if update else "RECREATE"
tOut = TFile("%s/data/rs-plc.root" % dsi.latSWDir, "UPDATE")
start = time.clock()
name = "amp-axion"
fitVal = fitVals[name][0]
thisVar = fitWorkspace.var(name)
pCL = 0.9
plc = RS.ProfileLikelihoodCalculator(fData, model, ROOT.RooArgSet(thisVar))
plc.SetConfidenceLevel(0.90)
interval = plc.GetInterval()
lower = interval.LowerLimit(thisVar)
upper = interval.UpperLimit(thisVar)
plot = RS.LikelihoodIntervalPlot(interval)
plot.SetNPoints(50)
plot.Draw("tf1")
pName = "hP_%d" if idx is not None else "hP"
hProfile = plot.GetPlottedObject()
hProfile.SetName(pName)
hProfile.SetTitle("PL %.2f %s lo %.3f mid %.3f hi %.3f eC %.3f" %
(pCL, name, lower, fitVal, upper, effCorr))
hProfile.Write()
print(hProfile.GetTitle())
tOut.Close()
print("elapsed:", time.clock() - start)
tenIntegralSig = tenIntegralSig * nSig.getValV() / totIntegralSig
fiveIntegralBkg = fiveIntegralBkg * nBkg.getValV() / totIntegralBkg
tenIntegralBkg = tenIntegralBkg * nBkg.getValV() / totIntegralBkg
fiveIntegralTot = fiveIntegralTot * (nBkg.getValV()+nSig.getValV()) / totIntegralTot
tenIntegralTot = tenIntegralTot * (nBkg.getValV()+nSig.getValV()) / totIntegralTot
c.SaveAs('phimassSPlot_'+ outname + '.png')
c.SaveAs('phimassSPlot_' + outname + '.root')
c.Clear()
cD=TCanvas("cD","cD",750,600)
cD.cd()
splot = RooStats.SPlot ("sPlot","sPlot",splotData, tot, RooArgList(nSig,nBkg))
dstree = splotData.store().tree()
#S plot hist for signal
shistSig = TH1F('shistSig','shistSig', 50, 4.0, 6.0)
shistSig.Sumw2()
shistSig.SetLineColor(2)
shistSig.SetMarkerColor(2); shistSig.SetMinimum(0.)
dstree.Project('shistSig','dimuonditrk_m_rf_c','nSig_sw');
shistSig.Draw('e0');
cD.SaveAs('SigSPlotPhi_' + outname + '_' + outname + '.gif')
cD.SaveAs('SigSPlotPhi_' + outname + '.root')
def latexfitresults(filename,
poiname='mu_SIG',
lumiFB=1.0,
nTOYS=3000,
asimov=False,
wname='combined'):
"""
Calculate before/after-fit yields in all channels given
@param filename The filename containing afterFit workspace
@param poiname Name of ParameterOfInterest = POI (default='mu_SIG')
@param lumiFB Given lumi in fb-1 to translate upper limit on N_events into xsection limit (default='1.0')
@param nTOYS Number of toys to be run
@param asimov Boolean to run asimov or not (default=False)
@param wname RooWorkspace name in file (default='combined')
"""
"""
pick up workspace from file
"""
workspacename = wname
w = Util.GetWorkspaceFromFile(filename, workspacename)
if w == None:
print "ERROR : Cannot open workspace : ", workspacename
sys.exit(1)
"""
Set the POI in ModelConfig
"""
if len(poiname) == 0:
print " "
else:
modelConfig = w.obj("ModelConfig")
poi = w.var(poiname)
if poi == None:
print "ERROR : Cannot find POI with name: ", poiname, " in workspace from file ", filename
sys.exit(1)
modelConfig.SetParametersOfInterest(RooArgSet(poi))
modelConfig.GetNuisanceParameters().remove(poi)
"""
set some default values for nToys, calculator type and npoints to be scanned
"""
ntoys = 3000
calctype = 0 # toys = 0, asymptotic (asimov) = 2
npoints = 20
if nTOYS != 3000 and nTOYS > 0:
ntoys = nTOYS
if asimov:
calctype = 2
"""
set the range of POI to be scanned and perform HypoTest inversion
"""
murangelow = 0.0
murangehigh = 40.0
hti_result = RooStats.DoHypoTestInversion(w, ntoys, calctype, 3, True,
npoints, murangelow, murangehigh)
"""
save and print the HypoTest result
"""
outFileName = "./htiResult_poi_" + poiname + "_ntoys_" + str(
ntoys) + "_calctype_" + str(calctype) + "_npoints_" + str(
npoints) + ".root"
hti_result.SaveAs(outFileName)
hti_result.Print()
"""
get the upper limit on N_obs out of hypotest result, and transform to limit on visible xsection
"""
uL_nobsinSR = hti_result.UpperLimit()
uL_visXsec = uL_nobsinSR / lumiFB
"""
get the expected upper limit and one scan point up and down to calculate the error on upper limit
"""
uL_nexpinSR = hti_result.GetExpectedUpperLimit(0)
uL_nexpinSR_P = hti_result.GetExpectedUpperLimit(1)
uL_nexpinSR_M = hti_result.GetExpectedUpperLimit(-1)
if uL_nexpinSR > uL_nexpinSR_P or uL_nexpinSR < uL_nexpinSR_M:
print " \n something very strange, either the uL_nexpinSR > uL_nexpinSR_P or uL_nexpinSR < uL_nexpinSR_M"
print " uL_nexpinSR = ", uL_nexpinSR, " uL_nexpinSR_P = ", uL_nexpinSR_P, " uL_nexpinSR_M = ", uL_nexpinSR_M
uL_nexpinSRerrP = hti_result.GetExpectedUpperLimit(1) - uL_nexpinSR
uL_nexpinSRerrM = uL_nexpinSR - hti_result.GetExpectedUpperLimit(-1)
"""
find the CLB values at indexes above and below observed CLs p-value
"""
CLB_P = 0.
CLB_M = 0.
mu_P = 0.
mu_M = 0.
index_P = 0
indexFound = False
for iresult in range(hti_result.ArraySize()):
xval = hti_result.GetXValue(iresult)
yval = hti_result.GetYValue(iresult)
if xval > uL_nobsinSR and not indexFound:
index_P = iresult
CLB_P = hti_result.CLb(iresult)
mu_P = xval
if iresult > 0:
CLB_M = hti_result.CLb(iresult - 1)
mu_M = hti_result.GetXValue(iresult - 1)
indexFound = True
"""
interpolate (linear) the value of CLB to be exactly above upperlimit p-val
"""
try:
alpha_CLB = (CLB_P - CLB_M) / (mu_P - mu_M)
beta_CLB = CLB_P - alpha_CLB * mu_P
CLB = alpha_CLB * uL_nobsinSR + beta_CLB
except ZeroDivisionError:
print "WARNING ZeroDivisionError while calculating CLb. Setting CLb=0."
CLB = 0.0
print "\n\n\n\n ***--- now doing p-value (s=0) calculation ---*** \n\n\n\n"
"""
reset parameter values and errors for p(s=0) calculation by reopening workspace
"""
w2 = Util.GetWorkspaceFromFile(filename, workspacename)
if w2 == None:
print "ERROR : Cannot open workspace : ", workspacename
sys.exit(1)
"""
calculate p(s=0) from the workspace given
"""
pval = RooStats.get_Presult(w2, False, ntoys, calctype)
ulList = [
uL_visXsec, uL_nobsinSR, uL_nexpinSR, uL_nexpinSRerrP, uL_nexpinSRerrM,
CLB, pval
]
return ulList
if opts.nullFit:
fr_null.Write()
#fitter.ws.Print()
output.Close()
if pars.btagSelection:
c1.SaveAs("DibosonBoostedBtaglnuJ_%s_%ijets_Stacked.png" % (mode, opts.Nj))
c2.SaveAs("DibosonBoostedBtaglnuJ_%s_%ijets_Subtracted.png" %
(mode, opts.Nj))
cp1.SaveAs("DibosonBoostedBtaglnuJ_%s_%ijets_Pull.png" % (mode, opts.Nj))
else:
c1.SaveAs("DibosonBoostedlnuJ_%s_%ijets_Stacked.png" % (mode, opts.Nj))
c2.SaveAs("DibosonBoostedlnuJ_%s_%ijets_Subtracted.png" % (mode, opts.Nj))
cp1.SaveAs("DibosonBoostedlnuJ_%s_%ijets_Pull.png" % (mode, opts.Nj))
print 'Time elapsed: %.1f sec' % timer.RealTime()
print 'CPU time used: %.1f sec' % timer.CpuTime()
print '%i degrees of freedom' % ndf
print 'chi2: %.2f / %i = %.2f' % (chi2_1, ndf, (chi2 / ndf))
print 'chi2 probability: %.4g' % (TMath.Prob(chi2, ndf))
if opts.nullFit:
likelihoodRatio = 2. * fr_null.minNll() - 2. * fr.minNll()
print '2*nll_null - 2*nll: %.4f - %.4f = %.4f' % (
2. * fr_null.minNll(), 2. * fr.minNll(), likelihoodRatio)
pval = TMath.Prob(likelihoodRatio, 1)
print 'p-value: %.4g' % pval
print 'Gaussian significance: %.3g' % RooStats.PValueToSignificance(pval)
def latexfitresults(filename, poiname='mu_SIG', lumiFB=1.0, nTOYS=3000, asimov=False, wname='combined'):
workspacename=wname
w = Util.GetWorkspaceFromFile(filename,workspacename)
if w==None:
print "ERROR : Cannot open workspace : ", workspacename
sys.exit(1)
if len(poiname)==0:
print " "
else:
modelConfig = w.obj("ModelConfig")
poi = w.var(poiname)
if poi==None:
print "ERROR : Cannot find POI with name: ", poiname, " in workspace from file ", filename
sys.exit(1)
modelConfig.SetParametersOfInterest(RooArgSet(poi))
modelConfig.GetNuisanceParameters().remove(poi)
ntoys = 3000
calctype = 0 # toys = 0, asymptotic (asimov) = 2
npoints = 20
if nTOYS != 3000 and nTOYS>0:
ntoys = nTOYS
if asimov:
calctype = 2
# hti_result = RooStats.MakeUpperLimitPlot(poiname,w,calctype,3,ntoys,True,npoints)
## # RooStats::MakeUpperLimitPlot(const char* fileprefix,
## # RooWorkspace* w,
## # int calculatorType , # toys = 0, asymptotic (asimov) = 2
## # int testStatType ,
## # int ntoys,
## # bool useCLs ,
## # int npoints )
murangelow = 0.0
murangehigh = 40.0
hti_result = RooStats.DoHypoTestInversion(w,ntoys,calctype,3,True,npoints,murangelow,murangehigh)
outFileName = "./htiResult_poi_" + poiname + "_ntoys_" + str(ntoys) + "_calctype_" + str(calctype) + "_npoints_" + str(npoints) + ".root"
hti_result.SaveAs(outFileName)
hti_result.Print()
uL_nobsinSR = hti_result.UpperLimit()
uL_visXsec = uL_nobsinSR / lumiFB
# uL_visXsecErrorUp = uL_visXsec - uL_nobsinSR/(lumiFB * (1. + lumiRelUncert))
# uL_visXsecErrorDown = uL_nobsinSR/(lumiFB * (1. - lumiRelUncert)) - uL_visXsec
uL_nexpinSR = hti_result.GetExpectedUpperLimit(0)
uL_nexpinSR_P = hti_result.GetExpectedUpperLimit(1)
uL_nexpinSR_M = hti_result.GetExpectedUpperLimit(-1)
if uL_nexpinSR > uL_nexpinSR_P or uL_nexpinSR < uL_nexpinSR_M:
print " \n something very strange, either the uL_nexpinSR > uL_nexpinSR_P or uL_nexpinSR < uL_nexpinSR_M"
print " uL_nexpinSR = ", uL_nexpinSR , " uL_nexpinSR_P = ", uL_nexpinSR_P, " uL_nexpinSR_M = ", uL_nexpinSR_M
uL_nexpinSRerrP = hti_result.GetExpectedUpperLimit(1) - uL_nexpinSR
uL_nexpinSRerrM = uL_nexpinSR - hti_result.GetExpectedUpperLimit(-1)
# find the CLB values at indexes above and below observed CLs p-value
CLB_P = 0.
CLB_M = 0.
mu_P = 0.
mu_M = 0.
index_P = 0
indexFound = False
for iresult in range(hti_result.ArraySize()):
xval = hti_result.GetXValue(iresult)
yval = hti_result.GetYValue(iresult)
if xval>uL_nobsinSR and not indexFound:
index_P = iresult
CLB_P = hti_result.CLb(iresult)
mu_P = xval
if iresult>0:
CLB_M = hti_result.CLb(iresult-1)
mu_M = hti_result.GetXValue(iresult-1)
indexFound = True
# print " \n found the CLB values to interpolate"
# print " CLB_M =", CLB_M, " CLB_P =", CLB_P, " mu_P = ", mu_P, " mu_M = ", mu_M
# interpolate the value of CLB to be exactly above upperlimit p-val
try:
alpha_CLB = (CLB_P - CLB_M) / (mu_P - mu_M)
beta_CLB = CLB_P - alpha_CLB*mu_P
# CLB is taken as the point on the CLB curve for the same poi value, as the observed upperlimit
CLB = alpha_CLB * uL_nobsinSR + beta_CLB
except ZeroDivisionError:
print "WARNING ZeroDivisionError while calculating CLb. Setting CLb=0."
CLB=0.0
#print " CLB = " , CLB
print "\n\n\n\n ***--- now doing p-value calculation ---*** \n\n\n\n"
Util.resetAllValues(w)
Util.resetAllErrors(w)
Util.resetAllNominalValues(w)
pval = RooStats.get_Presult(w,False,1000,2)
# get_Presult( RooWorkspace* w,
# bool doUL, // = true, // true = exclusion, false = discovery
# int ntoys, //=1000,
# int calculatorType, // = 0,
# int testStatType, // = 3,
# const char * modelSBName, // = "ModelConfig",
# const char * modelBName, // = "",
# const char * dataName, // = "obsData",
# bool useCLs, // = true ,
# bool useNumberCounting, // = false,
# const char * nuisPriorName) // = 0
ulList = [uL_visXsec, uL_nobsinSR, uL_nexpinSR, uL_nexpinSRerrP, uL_nexpinSRerrM, CLB, pval ]
return ulList
observedZValue = [] tmpIntegratedLumi = 0.0 for ibin in xrange( h_den.GetNbinsX() ): if tmpIntegratedLumi/1000.>22.1: break tmpIntegratedLumi += h_den.GetBinContent(ibin) nobs = h_num["SRG3b"].Integral(0,ibin) nexp = 2.8*tmpIntegratedLumi/(22.1 * 1000.) nexpsig = 10*tmpIntegratedLumi/(22.1 * 1000.) integratedLumi.append(tmpIntegratedLumi/1000.) tmpPvalue = RooStats.NumberCountingUtils.BinomialObsP(nexpsig,nexp,0.27)/2. tmpZvalue = RooStats.PValueToSignificance( tmpPvalue ) if tmpPvalue else 0 expectedZValue.append( tmpZvalue ) tmpPvalue = RooStats.NumberCountingUtils.BinomialObsP(nexpsig,nexp*(1+0.27),0.27)/2. tmpZvalue = RooStats.PValueToSignificance( tmpPvalue ) if tmpPvalue else 0 expectedUpZValue.append( tmpZvalue ) tmpPvalue = RooStats.NumberCountingUtils.BinomialObsP(nexpsig,nexp*(1-0.27),0.27)/2. tmpZvalue = RooStats.PValueToSignificance( tmpPvalue ) if tmpPvalue else 0 expectedDownZValue.append( tmpZvalue ) tmpPvalue = RooStats.NumberCountingUtils.BinomialObsP(nobs,nexp,0.27)/2. tmpZvalue = RooStats.PValueToSignificance( tmpPvalue ) if tmpPvalue else 0 observedZValue.append( tmpZvalue )
c = TCanvas("canvas", "canvas", 1200, 800)
phiFrame = masskk.frame(Range(massmin, massmax))
xdataPrompt.plotOn(phiFrame)
tot.plotOn(phiFrame)
phiFrame.Draw()
c.SaveAs("phiMassSPlotPhi.png")
c.SaveAs("phiMassSPlotPhi.root")
c.Clear()
# In[ ]:
cD = TCanvas("cD", "cD", 750, 600)
cD.cd()
splot = RooStats.SPlot("sPlot", "sPlot", xdataPrompt, tot, alist2)
dstree = xdataPrompt.store().tree()
# In[19]:
shist = TH1F('shist', 'shist', 200, 0.97, 1.07)
# In[20]:
shist.Sumw2()
shist.SetLineColor(2)
shist.SetMarkerColor(2)
shist.SetMinimum(0.)
dstree.Project('shist', 'kkM', 'S1_sw + S2_sw')
# In[21]:
def latexfitresults(filename, poiname='mu_Sig', lumiFB=1.0,
nTOYS=3000, asimov=False, wname='combined'):
workspacename=wname
w = Util.GetWorkspaceFromFile(filename,workspacename)
if w==None:
print "ERROR : Cannot open workspace : ", workspacename
sys.exit(1)
if len(poiname)==0:
print " "
else:
modelConfig = w.obj("ModelConfig")
poi = w.var(poiname)
if poi==None:
print "ERROR : Cannot find POI with name: ", poiname, " in workspace from file ", filename
sys.exit(1)
modelConfig.SetParametersOfInterest(RooArgSet(poi))
modelConfig.GetNuisanceParameters().remove(poi)
ntoys = 3000
calctype = 0 # toys = 0, asymptotic (asimov) = 2
npoints = 20
if nTOYS != 3000 and nTOYS>0:
ntoys = nTOYS
if asimov:
calctype = 2
hti_result = RooStats.MakeUpperLimitPlot(
poiname,w,calctype,3,ntoys,True,npoints)
outFileName = "./htiResult_poi_" + poiname + "_ntoys_" + str(ntoys) + "_calctype_" + str(calctype) + "_npoints_" + str(npoints) + ".root"
hti_result.SaveAs(outFileName)
hti_result.Print()
uL_nobsinSR = hti_result.UpperLimit()
uL_visXsec = uL_nobsinSR / lumiFB
# uL_visXsecErrorUp = uL_visXsec - uL_nobsinSR/(lumiFB * (1. + lumiRelUncert))
# uL_visXsecErrorDown = uL_nobsinSR/(lumiFB * (1. - lumiRelUncert)) - uL_visXsec
uL_nexpinSR = hti_result.GetExpectedUpperLimit(0)
uL_nexpinSR_P = hti_result.GetExpectedUpperLimit(1)
uL_nexpinSR_M = hti_result.GetExpectedUpperLimit(-1)
if uL_nexpinSR > uL_nexpinSR_P or uL_nexpinSR < uL_nexpinSR_M:
print " \n something very strange, either the uL_nexpinSR > uL_nexpinSR_P or uL_nexpinSR < uL_nexpinSR_M"
print " uL_nexpinSR = ", uL_nexpinSR , " uL_nexpinSR_P = ", uL_nexpinSR_P, " uL_nexpinSR_M = ", uL_nexpinSR_M
uL_nexpinSRerrP = hti_result.GetExpectedUpperLimit(1) - uL_nexpinSR
uL_nexpinSRerrM = uL_nexpinSR - hti_result.GetExpectedUpperLimit(-1)
# find the CLB values at indexes above and below observed CLs p-value
CLB_P = 0.
CLB_M = 0.
mu_P = 0.
mu_M = 0.
index_P = 0
indexFound = False
for iresult in range(hti_result.ArraySize()):
xval = hti_result.GetXValue(iresult)
yval = hti_result.GetYValue(iresult)
if xval>uL_nobsinSR and not indexFound:
index_P = iresult
CLB_P = hti_result.CLb(iresult)
mu_P = xval
if iresult>0:
CLB_M = hti_result.CLb(iresult-1)
mu_M = hti_result.GetXValue(iresult-1)
indexFound = True
# print " \n found the CLB values to interpolate"
# print " CLB_M =", CLB_M, " CLB_P =", CLB_P, " mu_P = ", mu_P, " mu_M = ", mu_M
# interpolate the value of CLB to be exactly above upperlimit p-val
alpha_CLB = (CLB_P - CLB_M) / (mu_P - mu_M)
beta_CLB = CLB_P - alpha_CLB*mu_P
# CLB is taken as the point on the CLB curve for the same poi value,
# as the observed upperlimit
CLB = alpha_CLB * uL_nobsinSR + beta_CLB
#print " CLB = " , CLB
print "\n\n\n\n ***--- now doing p-value calculation ---*** \n\n\n\n"
pval = RooStats.get_Presult(w,False,1000,2)
print "p-value is: ", pval
ulList = [uL_visXsec, uL_nobsinSR, uL_nexpinSR, uL_nexpinSRerrP, uL_nexpinSRerrM, CLB, pval ]
return ulList
def MakeWorkspace(outfilename):
# use this area to implement your model for a counting experiment
print "building counting model..."
#create workspace
pWs = ROOT.RooWorkspace("myWS")
#observable
pWs.factory("n[0]") #this create a literal, real-valued container called n
#signal yield
# pWs.factory("nsig[0,0,100]") # name[value,min bound,max bound]
# pWs.factory("lumi[0]") #luminosity
pWs.factory("lumi_nom[20000, 10000, 30000]") #nominal lumi of 20 fb-1
pWs.factory("lumi_kappa[1.026]")
pWs.factory("lumi_beta[0,-5,5]")
pWs.factory("RooPowerFunction::lumi_alpha(lumi_kappa, lumi_beta)")
pWs.factory("prod::lumi(lumi_nom,lumi_alpha)") #the new lumi
pWs.factory("Gaussian::lumi_constr(lumi_beta,lumi_glob[0,-5,5],1)")
pWs.factory("xsec[0,0,0.1]") #cross-section (our new POI)
# pWs.factory("eff[0]") #efficiency
pWs.factory("eff_nom[0.1, 0.01, 0.20]") #nominal eff
pWs.factory("eff_kappa[1.1]")
pWs.factory("eff_beta[0,-5,5]")
pWs.factory("RooPowerFunction::eff_alpha(eff_kappa, eff_beta)")
pWs.factory("prod::eff(eff_nom,eff_alpha)") #the new eff
pWs.factory("Gaussian::eff_constr(eff_beta,eff_glob[0,-5,5],1)")
pWs.factory("prod::nsig(lumi,xsec,eff)") #new definition of nsig
#background yield
# pWs.factory("nbkg[10,0,100]")
pWs.factory("nbkg_nom[10.,5.,15.]")
pWs.factory("nbkg_kappa[1.1]")
pWs.factory("nbkg_beta[0,-5,5]")
pWs.factory("RooPowerFunction::nbkg_alpha(nbkg_kappa, nbkg_beta)")
pWs.factory("prod::nbkg(nbkg_nom,lumi_alpha,nbkg_alpha)")
pWs.factory("Gaussian::nbkg_constr(nbkg_beta, nbkg_glob[0,-5,5],1)")
#full yield
pWs.factory("sum::yield(nsig,nbkg)") # creates a function nsig+nbkg
# Bayesian prior
pWs.factory("Uniform::prior(xsec)")
# define the physics model a Poisson distribution
pWs.factory("Poisson::model_core(n,yield)")
# model and systematics
# pWs.factory("PROD::model(model_core,lumi_constr)") # all-cap PROD means PDF instead of only function
pWs.factory("PROD::model(model_core,lumi_constr,eff_constr,nbkg_constr)")
# create set of observables (needed for datasets and ModelConfig later)
pObs = pWs.var("n") # get the pointer to the observable
obs = ROOT.RooArgSet("observables")
obs.add(pObs)
# create the dataset
pObs.setVal(11) # this is your observed data: you counted eleven events
data = ROOT.RooDataSet("data", "data", obs)
data.add(obs)
# import dataset into workspace
getattr(pWs, 'import')(
data
) # we call it this way because "import" is a reserved word in python
# create set of global observables (need to be defined as constants!)
pWs.var("lumi_glob").setConstant(True)
pWs.var("eff_glob").setConstant(True)
pWs.var("nbkg_glob").setConstant(True)
globalObs = ROOT.RooArgSet("global_obs")
globalObs.add(pWs.var("lumi_glob"))
globalObs.add(pWs.var("eff_glob"))
globalObs.add(pWs.var("nbkg_glob"))
# create set of parameters of interest (POI)
poi = ROOT.RooArgSet("poi")
poi.add(pWs.var("xsec"))
# create set of nuisance parameters
nuis = ROOT.RooArgSet("nuis")
nuis.add(pWs.var("lumi_beta"))
nuis.add(pWs.var("eff_beta"))
nuis.add(pWs.var("nbkg_beta"))
# fix all other variables in model:
# everything except observables, POI, and nuisance parameters
# must be constant
pWs.var("lumi_nom").setConstant(True)
pWs.var("eff_nom").setConstant(True)
pWs.var("nbkg_nom").setConstant(True)
pWs.var("lumi_kappa").setConstant(True)
pWs.var("eff_kappa").setConstant(True)
pWs.var("nbkg_kappa").setConstant(True)
fixed = ROOT.RooArgSet("fixed")
fixed.add(pWs.var("lumi_nom"))
fixed.add(pWs.var("eff_nom"))
fixed.add(pWs.var("nbkg_nom"))
fixed.add(pWs.var("lumi_kappa"))
fixed.add(pWs.var("eff_kappa"))
fixed.add(pWs.var("nbkg_kappa"))
# create signal+background Model Config
sbHypo = RooStats.ModelConfig("SbHypo")
sbHypo.SetWorkspace(pWs)
sbHypo.SetPdf(pWs.pdf("model"))
sbHypo.SetObservables(obs)
sbHypo.SetGlobalObservables(globalObs)
sbHypo.SetParametersOfInterest(poi)
sbHypo.SetNuisanceParameters(nuis)
sbHypo.SetPriorPdf(pWs.pdf("prior")) # this is optional
# set parameter snapshot that corresponds to the best fit to data
pNll = sbHypo.GetPdf().createNLL(data)
pProfile = pNll.createProfile(
globalObs) # do not profile global observables
pProfile.getVal(
) # this will do fit and set POI and nuisance parameters to fitted values
pPoiAndNuisance = ROOT.RooArgSet("poiAndNuisance")
pPoiAndNuisance.add(sbHypo.GetNuisanceParameters())
pPoiAndNuisance.add(sbHypo.GetParametersOfInterest())
sbHypo.SetSnapshot(pPoiAndNuisance)
# import S+B ModelConfig into workspace
getattr(pWs, 'import')(sbHypo)
# create background-only Model Config from the S+B one
bHypo = RooStats.ModelConfig(sbHypo)
bHypo.SetName("BHypo")
bHypo.SetWorkspace(pWs)
# set parameter snapshot for bHypo, setting xsec=0
# it is useful to understand how this block of code works
# but you can also use it as a recipe to make a parameter snapshot
pNll = bHypo.GetPdf().createNLL(data)
poiAndGlobalObs = ROOT.RooArgSet("poiAndGlobalObs")
poiAndGlobalObs.add(poi)
poiAndGlobalObs.add(globalObs)
pProfile = pNll.createProfile(
poiAndGlobalObs) # do not profile POI and global observables
poi.first().setVal(0) # set xsec=0 here
pProfile.getVal(
) # this will do fit and set nuisance parameters to profiled values
pPoiAndNuisance = ROOT.RooArgSet("poiAndNuisance")
pPoiAndNuisance.add(nuis)
pPoiAndNuisance.add(poi)
bHypo.SetSnapshot(pPoiAndNuisance)
# import model config into the workspace
getattr(pWs, 'import')(bHypo)
#print the contents of the workspace
pWs.Print()
#save the workspace to a file
pWs.SaveAs(outfilename)
return 0
RooFit.LineStyle(kDashed))
bfr5.SetTitle("Jet slice of sim pdf")
bfr5.Draw()
#c.cd(6)
#bfr6 = w.var("bdt").frame()
#simPdf.plotOn(bfr6, RooFit.ProjWData(ROOT.RooArgSet(sample),combData)) ;
#simPdf.plotOn(bfr6, RooFit.ProjWData(ROOT.RooArgSet(sample),combData), RooFit.Slice(sample, "JKS"), RooFit.LineColor(2), RooFit.LineStyle(kDashed))
#simPdf.plotOn(bfr6, RooFit.ProjWData(ROOT.RooArgSet(sample),combData), RooFit.Slice(sample, "Jet"), RooFit.LineColor(4), RooFit.LineStyle(kDashed))
#bfr6.SetTitle("sim pdf")
#bfr6.Draw()
"""
sct = RooSimWSTool(w)
model_sim = sct.build("model_sim","model",SplitParam("m","c"))
"""
mc = RooStats.ModelConfig("mc", w)
mc.SetPdf(simPdf)
mc.SetParametersOfInterest("mu")
mc.SetObservables("bdt")
mc.SetNuisanceParameters("nbkg")
mc.SetNuisanceParameters("nbkg2")
c.cd(7)
bfr7 = w.var("bdt").frame()
cdata = ROOT.RooStats.AsymptoticCalculator.MakeAsimovData(
combData, mc, w.argSet("bdt,sample"), ROOT.RooArgSet())
dh = ROOT.RooDataHist("", "", w.argSet("bdt,sample"), cdata)
err_correction = [
dh.set(dh.get(i), dh.weight(dh.get(i)),
ROOT.TMath.Sqrt(dh.weight(dh.get(i))))
for i in range(0, dh.numEntries())
mean.setConstant(ROOT.kFALSE)
rfit = tot.fitTo(b0dataNonPrompt, Range(massmin, massmax), RooFit.NumCPU(8))
gamma.setConstant(ROOT.kFALSE)
rfit = tot.fitTo(b0dataNonPrompt, Range(phimean - 0.025, phimean + 0.025),
RooFit.NumCPU(8))
masskkFrame = masskk.frame(Range(phimean - 0.025, phimean + 0.025))
b0dataNonPrompt.plotOn(massFrame, RooLinkedList())
tot.plotOn(masskkFrame)
massFrame.Draw()
c.SaveAs("testmassFit.png")
cD = TCanvas("cD", "cD", 750, 600)
cD.cd()
splot = RooStats.SPlot("sPlot", "sPlot", b0dataNonPrompt, tot,
RooArgList(nSig, nBkg))
# In[18]:
dstree = b0dataNonPrompt.store().tree()
# In[19]:
shist = TH1F('shist', 'shist', 500, 1.00, 1.05)
# In[20]:
shist.Sumw2()
shist.SetLineColor(2)
shist.SetMarkerColor(2)
shist.SetMinimum(0.)
def variables(self, ws, debug):
var_items = self.items[self.sVariables]
legend = '[' + self.sVariables + ']:'
print legend, 'begin processing...'
print legend, 'define variables, constants, categories, sets and lists'
# dictionary of value lists for each item
# this is more detailed than the original items object
# because the multiple settings for a given item
# are already split
# NOTE: sets are special case and are not added here
item_values = {}
# using RooStats::HLFactory to parse the model
# make a temp file with a card file
card_file_name = 'hlfCardFile.rs'
with open(card_file_name, 'w') as cardfile:
vSets = {}
for item in var_items:
#
# special case for sets
# standard RooFactoryWSTool interface exists in newer versions
# need to switch at some point
# for now parsing and adding named sets by hand
#
if item[1].strip()[0:3] == 'set':
set_arg_list = item[1].strip()
set_arg_list = set_arg_list[3:len(set_arg_list)]
set_arg_list = set_arg_list.strip(' ;()').split(',')
set_args = []
for arg in set_arg_list:
set_args.append(arg.strip())
vSets[item[0].strip()] = set_args
if (debug > 2):
print legend, 'Set arg value is ->' + item[1] + '<-'
print legend, 'set args are', set_args
else:
# this is a way to handle multiple settings
# for each item: they are separated by semicolon
# in the value field of the config item
item_values[item[0]] = self.split_values(item)
#print >> cardfile, item[0], '=', item[1]
print >> cardfile, item[0], '=', item_values[item[0]][0]
# add variables using HLFactory cards
hlfVerbosity = debug > 2
hlf = RooStats.HLFactory('HLFactoryExost', ws, hlfVerbosity)
hlf.ProcessCard(card_file_name)
# process the secondary options (fix/float etc.)
for _key in item_values:
n_values = len(item_values[_key])
for _index in range(0, n_values):
self.process_value(ws, _key, item_values[_key][_index], debug)
# now add sets by hand
print legend, 'creating sets now...'
# fixme experimental
_sets = {}
for set_args_key in vSets.keys():
arg_list = ''
# fixme experimental
_sets[set_args_key] = RooArgSet(set_args_key)
for arg in vSets[set_args_key]:
if len(arg_list) > 0:
arg_list += ','
arg_list += arg
# fixme experimental
_sets[set_args_key].add(ws.var(arg))
#ws.defineSet(set_args_key, arg_list)
print legend, 'set', set_args_key, 'created'
print legend, 'done creating sets'
return {'sets': _sets}
def bayesian(self, ws, debug=0):
if self.sBayes not in self.items:
if debug > 0:
print self.legend, 'no Bayesian Calculator section in the config file'
print self.legend, 'cannot configure Bayesian calculator'
return {'status': 'fail'}
legend = '[' + self.sBayes + ']:'
_name = 'exostBayes'
# getting items as a dictionary
bayes_items = {}
bayes_items.update(self.items[self.sBayes])
print legend, 'Configuring Bayesian calculator... '
# check if a CL is specified
m_conf_name = self.check_value(self.sBayes, 'model_config')
if m_conf_name == -1:
print legend, 'Error: model config is not specified'
print legend, 'Error:', self.sBayes, 'cannot be configured'
return {'status': 'fail'}
# check if data is specified
data_name = self.check_value(self.sBayes, 'data')
data_valid = False
if data_name != -1:
data_valid = True
if ws.data(data_name) != None:
data = ws.data(data_name)
else:
data_valid = False
print legend, 'Error: dataset', data_name, 'is not defined'
if data_valid == False:
print legend, 'Error:', self.sBayes, 'cannot be configured'
return {'status': 'fail'}
# check if a CL is specified
conf_level = self.check_value(self.sBayes, 'confidence_level')
if conf_level == -1:
print legend, 'Warning: desired confidence level is not specified, setting to 0.90'
conf_level = 0.90
# check if a posterior plot is requested
post_plot = self.check_value(self.sBayes, 'posterior_plot')
b_post_plot = False
if post_plot == -1:
print legend, 'will generate a posterior plot'
b_post_plot = True
elif post_plot == 'True':
b_post_plot = True
else:
b_post_plot = False
# plot format
if post_plot:
plot_format = self.check_value(self.sBayes, 'plot_format')
if plot_format == -1:
plot_format = 'png'
# to suppress messgaes when pdf goes to zero
RooMsgService.instance().setGlobalKillBelow(RooFit.FATAL)
#mconf = ws.obj('exostModelConfig')
mconf = ws.obj(m_conf_name)
if mconf == None:
if debug > 0:
print self.legend, 'fail to get Model Config'
print self.legend, 'unable to configure Bayesian calculator'
print self.legend, 'calculator will not be imported into the workspace'
return {'status': 'fail'}
bCalc = RooStats.BayesianCalculator(data, mconf)
bCalc.SetConfidenceLevel(float(conf_level))
############# FIXME: debug test
#bInt = bCalc.GetInterval()
#print legend, 'DEBUG3*******************************'
#
#cl = bCalc.ConfidenceLevel()
#print legend, str(cl)+'% CL central interval: [', bInt.LowerLimit(), ' - ', bInt.UpperLimit(), ']'
#print legend, 'or', str(cl+(1.0-cl)/2), '% CL limits'
##############################
#size = 1.0 - float(conf_level)
#bCalc.SetTestSize(size)
# import the calculator into the Workspace
getattr(ws, 'import')(bCalc, _name)
print legend, 'Bayesian calculator', _name, 'is configured and added to workspace', ws.GetName(
)
print legend, 'done'
return {
'do_posterior_plot': b_post_plot,
'status': 'success',
'model_config_name': m_conf_name,
'plot_format': plot_format
}
def configure(self, ws, items, section, debug = 0):
print self.legend, 'configuring Profile Likelihood calculator...'
_items = items[section]
# check mandatory parameters
# dataset name in the workspace
_dataset_name = self.check_value(items, section, 'data')
if _dataset_name == None:
print self.legend, 'ERROR: dataset is not specified, cannot configure'
return
else:
if ws.data(_dataset_name) != None:
_data = ws.data(_dataset_name)
else:
print legend, 'Error: dataset', _dataset_name, 'is not defined, cannot configure'
return
# model config name in the workspace
self._model_config_name = self.check_value(items, section, 'model_config')
if self._model_config_name == None:
print self.legend, 'ERROR: Model Config name is not specified, cannot configure'
return
else:
if self.check_generic_object(ws, self._model_config_name, debug):
_mconf = ws.obj(self._model_config_name.strip())
else:
print self.legend, 'Error: model config object is not found, cannot configure'
return
# check optional parameters, set defaults if not specified
# desired confidence level
_conf_level = self.check_value(items, section, 'confidence_level')
if _conf_level == None:
print self.legend, 'no confidence level specified, setting to 0.95'
_conf_level = 0.95
# whether to make the posterior plot or not
self._scan_plot = self.check_value(items, section, 'make_scan_plot')
if self._scan_plot == None:
print self.legend, 'likelihood intervalscan plot is not requested'
self._scan_plot = False
elif self._scan_plot == 'True':
self._scan_plot = True
else:
print self.legend, 'invalid assignment make_scan_plot =', self._scan_plot
print self.legend, 'likelihood interval scan plot will not be made'
self._scan_plot = False
# plot format
if self._scan_plot:
self._plot_format = self.check_value(items, section, 'plot_format')
if self._plot_format == None:
print self.legend, 'no plot format specified, setting to PNG'
self._plot_format = png
# configuring now...
self._plc = RooStats.ProfileLikelihoodCalculator(_data,_mconf)
self._plc.SetConfidenceLevel( float(_conf_level) )
print legend, 'Profile Likelihood calculator is configured'
# import the calculator into the Workspace
#getattr(ws, 'import')(self._plc, 'exostMcmcCalc')
#print legend, 'Markov chain MC calculator', 'exostMcmcCalc', 'is added to workspace', ws.GetName()
#print legend, 'done'
ws.Print()
# flag that the action is configured successfully
self.configured = True
return {'do_scan_plot':self._scan_plot}
def latexfitresults(filename, poiname='mu_SIG', lumiFB=1.0, nTOYS=3000, asimov=False, wname='combined'):
"""
Calculate before/after-fit yields in all channels given
@param filename The filename containing afterFit workspace
@param poiname Name of ParameterOfInterest = POI (default='mu_SIG')
@param lumiFB Given lumi in fb-1 to translate upper limit on N_events into xsection limit (default='1.0')
@param nTOYS Number of toys to be run
@param asimov Boolean to run asimov or not (default=False)
@param wname RooWorkspace name in file (default='combined')
"""
"""
pick up workspace from file
"""
workspacename=wname
w = Util.GetWorkspaceFromFile(filename,workspacename)
if w==None:
print "ERROR : Cannot open workspace : ", workspacename
sys.exit(1)
"""
Set the POI in ModelConfig
"""
if len(poiname)==0:
print " "
else:
modelConfig = w.obj("ModelConfig")
poi = w.var(poiname)
if poi==None:
print "ERROR : Cannot find POI with name: ", poiname, " in workspace from file ", filename
sys.exit(1)
modelConfig.SetParametersOfInterest(RooArgSet(poi))
modelConfig.GetNuisanceParameters().remove(poi)
"""
set some default values for nToys, calculator type and npoints to be scanned
"""
ntoys = 3000
calctype = 0 # toys = 0, asymptotic (asimov) = 2
if nTOYS != 3000 and nTOYS>0:
ntoys = nTOYS
if asimov:
calctype = 2
"""
set the range of POI to be scanned and perform HypoTest inversion
"""
murangelow = 240
murangehigh = 350 #40.0 #set here -1. if you want to have automatic determined scan range, if using values != -1, please check the log file if the scan range was large enough
npoints = int((murangehigh-murangelow)*2.)
hti_result = RooStats.DoHypoTestInversion(w,ntoys,calctype,3,True,npoints,murangelow,murangehigh)
"""
save and print the HypoTest result
"""
outFileName = "./htiResult_poi_" + poiname + "_ntoys_" + str(ntoys) + "_calctype_" + str(calctype) + "_npoints_" + str(npoints) + ".root"
hti_result.SaveAs(outFileName)
hti_result.Print()
"""
get the upper limit on N_obs out of hypotest result, and transform to limit on visible xsection
"""
uL_nobsinSR = hti_result.UpperLimit()
uL_visXsec = uL_nobsinSR / lumiFB
"""
get the expected upper limit and one scan point up and down to calculate the error on upper limit
"""
uL_nexpinSR = hti_result.GetExpectedUpperLimit(0)
uL_nexpinSR_P = hti_result.GetExpectedUpperLimit(1)
uL_nexpinSR_M = hti_result.GetExpectedUpperLimit(-1)
if uL_nexpinSR > uL_nexpinSR_P or uL_nexpinSR < uL_nexpinSR_M:
print " \n something very strange, either the uL_nexpinSR > uL_nexpinSR_P or uL_nexpinSR < uL_nexpinSR_M"
print " uL_nexpinSR = ", uL_nexpinSR , " uL_nexpinSR_P = ", uL_nexpinSR_P, " uL_nexpinSR_M = ", uL_nexpinSR_M
uL_nexpinSRerrP = hti_result.GetExpectedUpperLimit(1) - uL_nexpinSR
uL_nexpinSRerrM = uL_nexpinSR - hti_result.GetExpectedUpperLimit(-1)
"""
find the CLB values at indexes above and below observed CLs p-value
"""
CLB_P = 0.
CLB_M = 0.
mu_P = 0.
mu_M = 0.
index_P = 0
indexFound = False
for iresult in range(hti_result.ArraySize()):
xval = hti_result.GetXValue(iresult)
yval = hti_result.GetYValue(iresult)
if xval>uL_nobsinSR and not indexFound:
index_P = iresult
CLB_P = hti_result.CLb(iresult)
mu_P = xval
if iresult>0:
CLB_M = hti_result.CLb(iresult-1)
mu_M = hti_result.GetXValue(iresult-1)
indexFound = True
"""
interpolate (linear) the value of CLB to be exactly above upperlimit p-val
"""
try:
alpha_CLB = (CLB_P - CLB_M) / (mu_P - mu_M)
beta_CLB = CLB_P - alpha_CLB*mu_P
CLB = alpha_CLB * uL_nobsinSR + beta_CLB
except ZeroDivisionError:
print "WARNING ZeroDivisionError while calculating CLb. Setting CLb=0."
CLB=0.0
print "\n\n\n\n ***--- now doing p-value (s=0) calculation ---*** \n\n\n\n"
"""
reset parameter values and errors for p(s=0) calculation by reopening workspace
"""
w2 = Util.GetWorkspaceFromFile(filename,workspacename)
if w2==None:
print "ERROR : Cannot open workspace : ", workspacename
sys.exit(1)
"""
calculate p(s=0) from the workspace given
"""
pval = RooStats.get_Presult(w2,False,ntoys,calctype)
ulList = [uL_visXsec, uL_nobsinSR, uL_nexpinSR, uL_nexpinSRerrP, uL_nexpinSRerrM, CLB, pval ]
return ulList
def plotProfiles():
f = TFile("./data/fitWorkspace.root")
fitWorkspace = f.Get("fitWorkspace")
fData = fitWorkspace.allData().front()
fEnergy = fitWorkspace.var("energy_keV")
model = fitWorkspace.pdf("model")
fitResult = fitWorkspace.allGenericObjects().front()
fitValsFinal = getRooArgDict(fitResult.floatParsFinal())
print "Fit Cov Qual:", fitResult.covQual()
nameList = ["amp-axion"]
# nameList = []
# for key in sorted(fitValsFinal):
# if "amp-" in key:
# nameList.append(key)
print "Generating profiles ..."
c = TCanvas("c", "c", 800, 600)
for name in nameList:
fitVal = fitValsFinal[name]
thisVar = fitWorkspace.var(name)
# Brian & Clint method
plc = RS.ProfileLikelihoodCalculator(fData, model,
ROOT.RooArgSet(thisVar))
plc.SetConfidenceLevel(0.90)
interval = plc.GetInterval()
lower = interval.LowerLimit(thisVar)
upper = interval.UpperLimit(thisVar)
print "%-10s = lo %-7.3f best %-7.3f hi %.3f" % (name, lower, fitVal,
upper)
plot = RS.LikelihoodIntervalPlot(interval)
# p1, pltHi = fitVal - 1.5*(fitVal - lower), fitVal + 1.5*(upper - fitVal)
# plot.SetRange(pltLo,pltHi)
plot.SetTitle("%s: lo %.3f fit %.3f hi %.3f" %
(name, lower, fitVal, upper))
plot.SetNPoints(50)
plot.Draw("tf1")
"""
# Lukas method
# note: lukas uses ModelConfig, to explicitly set observables, constraints, and nuisance parameters
# https://root-forum.cern.ch/t/access-toy-datasets-generated-by-roostat-frequentistcalculator/24465/8
mc = RS.ModelConfig('mc', fitWorkspace)
mc.SetPdf( model )
mc.SetParametersOfInterest( ROOT.RooArgSet(thisVar) )
mc.SetObservables( ROOT.RooArgSet(fEnergy) )
# lukas example
# mc.SetConstraintParameters( ROOT.RooArgSet(mean, sigma) )
# mc.SetNuisanceParameters( ROOT.RooArgSet(mean, sigma, n_bkg) )
# mc.SetGlobalObservables( ROOT.RooArgSet(mean_obs, sigma_obs) )
# need to make some RooArgSets from RooRealVars
# constraints, nuisances, globalobs = ROOT.RooArgSet(), ROOT.RooArgSet(), ROOT.RooArgSet()
# for parName in sorted(fitValsFinal):
# rrv = fitWorkspace.var(parName)
# if parName != name:
# constraints.add(rrv)
# # .. etc.
# pl = RS.ProfileLikelihoodCalculator(fData, mc)
# pl.SetConfidenceLevel(0.683) # lukas used 0.90
# interval = pl.GetInterval()
# plot = RS.LikelihoodIntervalPlot(interval)
# plot.SetNPoints(50)
# plot.Draw("")
"""
c.Print("./plots/profile_%s.pdf" % name)
################################################################################
poi = ROOT.RooArgSet("poi")
#poi.add( pWs.var("glob_nttbar") );
poi.add(pWs.var("nttbar_nom"))
#poi.add( pWs.var("beta_nttbar") );
################################################################################
# create set of nuisance parameters
################################################################################
nuis = ROOT.RooArgSet("nuis")
nuis.add(pWs.var("beta_nqcd"))
nuis.add(pWs.var("beta_nt"))
nuis.add(pWs.var("beta_ntbar"))
nuis.add(pWs.var("beta_nwjets"))
sbHypo = RooStats.ModelConfig("SbHypo")
sbHypo.SetWorkspace(pWs)
sbHypo.SetPdf(pWs.pdf("model"))
sbHypo.SetObservables(obs)
sbHypo.SetGlobalObservables(globalObs)
sbHypo.SetParametersOfInterest(poi)
sbHypo.SetNuisanceParameters(nuis)
pWs.var("beta_nttbar").setConstant(False)
pWs.var("beta_nqcd").setConstant(False)
pWs.var("beta_nwjets").setConstant(False)
pWs.var("beta_nt").setConstant(False)
pWs.var("beta_ntbar").setConstant(False)
pWs.var("nwjets_nom").setConstant(True)
pWs.var("nt_nom").setConstant(True)
xpoints1=array.array('d')
ypoints1=array.array('d')
xerrors1=array.array('d')
yerrors1=array.array('d')
for i in range(2,20):
x=i*0.25
ys=[]
for toy in range(jobs*toysPerJob):
ymax=0
for mass,significances in fWW:
#if mass<1200: continue
ymax=max(ymax,significances[toy])
ys+=[ymax]
y=RooStats.PValueToSignificance(float(len([s for s in ys if s>=x]))/len(ys))
if y>10 or y<-10: continue
xpoints1.append(x)
ypoints1.append(y)
xerrors1.append(0)
if len([s for s in ys if s>=x])>0:
yerror=2*(y-RooStats.PValueToSignificance(float(sqrt(len([s for s in ys if s>=x]))+len([s for s in ys if s>=x]))/len(ys)))
else:
yerror=y
yerrors1.append(yerror)
print x,y,yerror
canvas = TCanvas("","",0,0,200,200)
legend=TLegend(0.3,0.7,0.95,0.90,"")
def model_config(self, ws, debug=0, sets=None):
if self.sModelConfig not in self.items:
if debug > 0:
print self.legend, 'no Model Config section in the config file'
print self.legend, 'cannot create Model Config'
return
legend = '[' + self.sModelConfig + ']:'
# getting items as a dictionary
model_config_items = {}
model_config_items.update(self.items[self.sModelConfig])
print legend, 'Configuring the model... '
# get the desired name for the model config object
_model_config_name = self.check_value(self.sModelConfig, 'name')
if _model_config_name == -1:
print self.legend, 'using default name for Model Config: exostModelConfig'
_model_config_name = 'exostModelConfig'
else:
print self.legend, 'creating Model Config with name', _model_config_name
# create the ModelConfig object and associate it with the Workspace
model_config = RooStats.ModelConfig(_model_config_name,
_model_config_name)
model_config.SetWorkspace(ws)
# check if the model is defined
if 'model' in model_config_items.keys():
_model = model_config_items['model']
print legend, 'The full likelihood model:', _model
#model_config.SetPdf(ws.pdf(_model))
model_config.SetPdf(_model)
else:
print legend, 'Error: the model PDF is not specified'
print legend, 'Error: the model config is invalid'
print legend, 'Error: the model config is not added to the workspace'
return -1
# check if the observables are specified
observables_valid = False
if 'observables' in model_config_items.keys():
observables_valid = True
_observables = model_config_items['observables']
print legend, 'Observables:', _observables
# fixme experimental
#if ws.set(_observables) != None:
# model_config.SetObservables(ws.set(_observables))
if sets[_observables] != None:
model_config.SetObservables(sets[_observables])
else:
observables_valid = False
print legend, 'Warning: observables set', _observables, 'is not defined'
else:
observables_valid = False
print legend, 'Warning: no observables specified'
# observables is an optional field in ModelConfig, no need to invalidate
#if observables_valid == False:
# print legend, 'Error: the model config is invalid'
# print legend, 'Error: the model config is not added to the workspace'
# return -1
# check if the global global_observables are specified
global_observables_valid = False
if 'global_observables' in model_config_items.keys():
global_observables_valid = True
_global_observables = model_config_items['global_observables']
print legend, 'Global observables:', _global_observables
# fixme experimental
#if ws.set(_global_observables) != None:
# model_config.SetGlobalObservables(ws.set(_global_observables))
if sets[_global_observables] != None:
model_config.SetGlobalObservables(sets[_global_observables])
else:
global_observables_valid = False
print legend, 'Warning: global observables set', _global_observables, 'is not defined'
else:
global_observables_valid = False
print legend, 'Warning: no global observables specified'
#if global_observables_valid == False:
# print legend, 'Error: the model config is invalid'
# print legend, 'Error: the model config is not added to the workspace'
# return -1
# check if the parameters of interest are specified
poi_valid = False
if 'poi' in model_config_items.keys():
poi_valid = True
_poi = model_config_items['poi']
print legend, 'Parameters of interest:', _poi
# fixme experimental
#if ws.set(_poi) != None:
# model_config.SetParametersOfInterest(ws.set(_poi))
if sets[_poi] != None:
model_config.SetParametersOfInterest(sets[_poi])
else:
poi_valid = False
print legend, 'Error: POI set', _poi, 'is not defined'
else:
poi_valid = False
print legend, 'Error: no parameters of interest specified'
if poi_valid == False:
print legend, 'Error: the model config is invalid'
print legend, 'Error: the model config is not added to the workspace'
return -1
# check if the nuisance parameters are specified
nuis_valid = False
if 'nuisance_parameters' in model_config_items.keys():
nuis_valid = True
_nuis = model_config_items['nuisance_parameters']
print legend, 'Nuisance parameters:', _nuis
#fixme experimental
#if ws.set(_nuis) != None:
# model_config.SetNuisanceParameters(ws.set(_nuis))
if sets[_nuis] != None:
model_config.SetNuisanceParameters(sets[_nuis])
else:
nuis_valid = False
print legend, 'Error: Nuisance parameter set', _nuis, 'is not defined'
else:
nuis_valid = False
print legend, 'Error: no nuisance parameters specified'
if nuis_valid == False:
print legend, 'Error: the model config is invalid'
print legend, 'Error: the model config is not added to the workspace'
return -1
# check if a prior PDF is specified
prior_valid = False
if 'prior' in model_config_items.keys():
prior_valid = True
_prior = model_config_items['prior']
print legend, 'Prior:', _prior
if ws.pdf(_prior) != None:
model_config.SetPriorPdf(ws.pdf(_prior))
else:
prior_valid = False
print legend, 'Error: prior', _prior, 'is not defined'
else:
prior_valid = False
print legend, 'Error: no prior PDF specified'
if prior_valid == False:
print legend, 'Error: the model config is invalid'
print legend, 'Error: the model config is not added to the workspace'
return -1
# import the model config
getattr(ws, 'import')(model_config, _model_config_name)
print legend, 'Model config is added to workspace ', ws.GetName()
print legend, 'done'
def latexfitresults(filename, poiname='mu_SIG', lumiFB=1.0, nTOYS=3000, nPoints=20, muRange=40, asimov=False, wname='combined', outputPrefix=""):
"""
Calculate before/after-fit yields in all channels given
@param filename The filename containing afterFit workspace
@param poiname Name of ParameterOfInterest = POI (default='mu_SIG')
@param lumiFB Given lumi in fb-1 to translate upper limit on N_events into xsection limit (default='1.0')
@param nTOYS Number of toys to be run
@param asimov Boolean to run asimov or not (default=False)
@param nPoints Number of points of mu_SIG ranges to scan
@param muRange Maximim value of mu_SIG to probe
@param wname RooWorkspace name in file (default='combined')
@param outputPrefix Prefix of the output file name (default="")
"""
"""
pick up workspace from file
"""
workspacename=wname
w = Util.GetWorkspaceFromFile(filename,workspacename)
if w==None:
print "ERROR : Cannot open workspace : ", workspacename
sys.exit(1)
"""
Set the POI in ModelConfig
"""
if len(poiname)==0:
print " "
else:
modelConfig = w.obj("ModelConfig")
poi = w.var(poiname)
if poi==None:
print "ERROR : Cannot find POI with name: ", poiname, " in workspace from file ", filename
sys.exit(1)
modelConfig.SetParametersOfInterest(RooArgSet(poi))
modelConfig.GetNuisanceParameters().remove(poi)
"""
set some default values for nToys, calculator type and nPoints to be scanned
"""
ntoys = 3000
calctype = 0 # toys = 0, asymptotic (asimov) = 2
nPoints = nPoints
if nTOYS != 3000 and nTOYS>0:
ntoys = nTOYS
if asimov:
calctype = 2
"""
set the range of POI to be scanned and perform HypoTest inversion
"""
nCPUs = 8
murangelow = 0.0
murangehigh = muRange #set here -1. if you want to have automatic determined scan range, if using values != -1, please check the log file if the scan range was large enough
hti_result = RooStats.DoHypoTestInversion(w, ntoys, calctype, 3, True, nPoints, murangelow, murangehigh, False, False, "ModelConfig", "", "obsData", "")
#hti_result = RooStats.DoHypoTestInversion(w, ntoys, calctype, 3, True, nPoints, murangelow, murangehigh, False, False, "ModelConfig", "", "obsData", "", nCPUs)
nRemoved = hti_result.ExclusionCleanup()
if nRemoved > 0:
print "WARNING: removed %d points from hti_result" % nRemoved
#store plot
RooStats.AnalyzeHypoTestInverterResult( hti_result, calctype, 3, True, nPoints, "%s%s" % (outputPrefix, poiname), ".eps")
RooStats.AnalyzeHypoTestInverterResult( hti_result, calctype, 3, True, nPoints, "%s%s" % (outputPrefix, poiname), ".pdf")
RooStats.AnalyzeHypoTestInverterResult( hti_result, calctype, 3, True, nPoints, "%s%s" % (outputPrefix, poiname), ".png")
outFileName = "./%shtiResult_poi_%s_ntoys_%d_calctype_%s_nPoints_%d.root" % (outputPrefix, poiname, ntoys, calctype, nPoints)
hti_result.SaveAs(outFileName)
hti_result.Print()
"""
get the upper limit on N_obs out of hypotest result, and transform to limit on visible xsection
"""
uL_nobsinSR = hti_result.UpperLimit()
uL_visXsec = uL_nobsinSR / lumiFB
"""
get the expected upper limit and one scan point up and down to calculate the error on upper limit
"""
uL_nexpinSR = hti_result.GetExpectedUpperLimit(0)
uL_nexpinSR_P = hti_result.GetExpectedUpperLimit(1)
uL_nexpinSR_M = hti_result.GetExpectedUpperLimit(-1)
if uL_nexpinSR > uL_nexpinSR_P or uL_nexpinSR < uL_nexpinSR_M:
print " \n something very strange, either the uL_nexpinSR > uL_nexpinSR_P or uL_nexpinSR < uL_nexpinSR_M"
print " uL_nexpinSR = ", uL_nexpinSR , " uL_nexpinSR_P = ", uL_nexpinSR_P, " uL_nexpinSR_M = ", uL_nexpinSR_M
uL_nexpinSRerrP = hti_result.GetExpectedUpperLimit(1) - uL_nexpinSR
uL_nexpinSRerrM = uL_nexpinSR - hti_result.GetExpectedUpperLimit(-1)
"""
find the CLB values at indexes above and below observed CLs p-value
"""
CLB_P = 0.
CLB_M = 0.
mu_P = 0.
mu_M = 0.
index_P = 0
indexFound = False
for iresult in range(hti_result.ArraySize()):
xval = hti_result.GetXValue(iresult)
yval = hti_result.GetYValue(iresult)
if xval>uL_nobsinSR and not indexFound:
index_P = iresult
CLB_P = hti_result.CLb(iresult)
mu_P = xval
if iresult>0:
CLB_M = hti_result.CLb(iresult-1)
mu_M = hti_result.GetXValue(iresult-1)
indexFound = True
"""
interpolate (linear) the value of CLB to be exactly above upperlimit p-val
"""
try:
alpha_CLB = (CLB_P - CLB_M) / (mu_P - mu_M)
beta_CLB = CLB_P - alpha_CLB*mu_P
# CLB is taken as the point on the CLB curve for the same poi value, as the observed upperlimit
CLB = alpha_CLB * uL_nobsinSR + beta_CLB
except ZeroDivisionError:
print "WARNING ZeroDivisionError while calculating CLb. Setting CLb=0."
CLB=0.0
print "\n\n\n\n ***--- now doing p-value (s=0) calculation ---*** \n\n\n\n"
"""
reset parameter values and errors for p(s=0) calculation by reopening workspace
"""
w2 = Util.GetWorkspaceFromFile(filename,workspacename)
if w2==None:
print "ERROR : Cannot open workspace : ", workspacename
sys.exit(1)
"""
calculate p(s=0) from the workspace given
"""
pval = RooStats.get_Presult(w2,False,ntoys,calctype)
#print pval
#sigma = StatTools.GetSigma(pval)
#print sigma
UL = {}
UL["visXsec"] = uL_visXsec
UL["nObsInSR"] = uL_nobsinSR
UL["nExpInSR"] = uL_nexpinSR
UL["nExpInSRPlus1Sigma"] = uL_nexpinSRerrP
UL["nExpInSRMinus1Sigma"] = uL_nexpinSRerrM
UL["CLb"] = CLB
UL["p0"] = pval
UL["Z"] = StatTools.GetSigma(pval)
return UL
def latexfitresults(filename,
poiname='mu_SIG',
lumiFB=1.0,
nTOYS=3000,
nPoints=20,
muRange=40,
asimov=False,
wname='combined',
outputPrefix=""):
"""
Calculate before/after-fit yields in all channels given
@param filename The filename containing afterFit workspace
@param poiname Name of ParameterOfInterest = POI (default='mu_SIG')
@param lumiFB Given lumi in fb-1 to translate upper limit on N_events into xsection limit (default='1.0')
@param nTOYS Number of toys to be run
@param asimov Boolean to run asimov or not (default=False)
@param nPoints Number of points of mu_SIG ranges to scan
@param muRange Maximim value of mu_SIG to probe
@param wname RooWorkspace name in file (default='combined')
@param outputPrefix Prefix of the output file name (default="")
"""
"""
pick up workspace from file
"""
workspacename = wname
w = Util.GetWorkspaceFromFile(filename, workspacename)
if w == None:
print "ERROR : Cannot open workspace : ", workspacename
sys.exit(1)
"""
Set the POI in ModelConfig
"""
if len(poiname) == 0:
print " "
else:
modelConfig = w.obj("ModelConfig")
poi = w.var(poiname)
if poi == None:
print "ERROR : Cannot find POI with name: ", poiname, " in workspace from file ", filename
sys.exit(1)
modelConfig.SetParametersOfInterest(RooArgSet(poi))
modelConfig.GetNuisanceParameters().remove(poi)
"""
set some default values for nToys, calculator type and nPoints to be scanned
"""
ntoys = 3000
calctype = 0 # toys = 0, asymptotic (asimov) = 2
nPoints = nPoints
if nTOYS != 3000 and nTOYS > 0:
ntoys = nTOYS
if asimov:
calctype = 2
"""
set the range of POI to be scanned and perform HypoTest inversion
"""
nCPUs = 8
murangelow = 0.0
murangehigh = muRange #set here -1. if you want to have automatic determined scan range, if using values != -1, please check the log file if the scan range was large enough
hti_result = RooStats.DoHypoTestInversion(w, ntoys, calctype, 3, True,
nPoints, murangelow, murangehigh,
False, False, "ModelConfig", "",
"obsData", "")
#hti_result = RooStats.DoHypoTestInversion(w, ntoys, calctype, 3, True, nPoints, murangelow, murangehigh, False, False, "ModelConfig", "", "obsData", "", nCPUs)
nRemoved = hti_result.ExclusionCleanup()
if nRemoved > 0:
print "WARNING: removed %d points from hti_result" % nRemoved
#store plot
RooStats.AnalyzeHypoTestInverterResult(hti_result, calctype, 3, True,
nPoints,
"%s%s" % (outputPrefix, poiname),
".eps")
RooStats.AnalyzeHypoTestInverterResult(hti_result, calctype, 3, True,
nPoints,
"%s%s" % (outputPrefix, poiname),
".pdf")
RooStats.AnalyzeHypoTestInverterResult(hti_result, calctype, 3, True,
nPoints,
"%s%s" % (outputPrefix, poiname),
".png")
outFileName = "./%shtiResult_poi_%s_ntoys_%d_calctype_%s_nPoints_%d.root" % (
outputPrefix, poiname, ntoys, calctype, nPoints)
hti_result.SaveAs(outFileName)
hti_result.Print()
"""
get the upper limit on N_obs out of hypotest result, and transform to limit on visible xsection
"""
uL_nobsinSR = hti_result.UpperLimit()
uL_visXsec = uL_nobsinSR / lumiFB
"""
get the expected upper limit and one scan point up and down to calculate the error on upper limit
"""
uL_nexpinSR = hti_result.GetExpectedUpperLimit(0)
uL_nexpinSR_P = hti_result.GetExpectedUpperLimit(1)
uL_nexpinSR_M = hti_result.GetExpectedUpperLimit(-1)
if uL_nexpinSR > uL_nexpinSR_P or uL_nexpinSR < uL_nexpinSR_M:
print " \n something very strange, either the uL_nexpinSR > uL_nexpinSR_P or uL_nexpinSR < uL_nexpinSR_M"
print " uL_nexpinSR = ", uL_nexpinSR, " uL_nexpinSR_P = ", uL_nexpinSR_P, " uL_nexpinSR_M = ", uL_nexpinSR_M
uL_nexpinSRerrP = hti_result.GetExpectedUpperLimit(1) - uL_nexpinSR
uL_nexpinSRerrM = uL_nexpinSR - hti_result.GetExpectedUpperLimit(-1)
"""
find the CLB values at indexes above and below observed CLs p-value
"""
CLB_P = 0.
CLB_M = 0.
mu_P = 0.
mu_M = 0.
index_P = 0
indexFound = False
for iresult in range(hti_result.ArraySize()):
xval = hti_result.GetXValue(iresult)
yval = hti_result.GetYValue(iresult)
if xval > uL_nobsinSR and not indexFound:
index_P = iresult
CLB_P = hti_result.CLb(iresult)
mu_P = xval
if iresult > 0:
CLB_M = hti_result.CLb(iresult - 1)
mu_M = hti_result.GetXValue(iresult - 1)
indexFound = True
"""
interpolate (linear) the value of CLB to be exactly above upperlimit p-val
"""
try:
alpha_CLB = (CLB_P - CLB_M) / (mu_P - mu_M)
beta_CLB = CLB_P - alpha_CLB * mu_P
# CLB is taken as the point on the CLB curve for the same poi value, as the observed upperlimit
CLB = alpha_CLB * uL_nobsinSR + beta_CLB
except ZeroDivisionError:
print "WARNING ZeroDivisionError while calculating CLb. Setting CLb=0."
CLB = 0.0
print "\n\n\n\n ***--- now doing p-value (s=0) calculation ---*** \n\n\n\n"
"""
reset parameter values and errors for p(s=0) calculation by reopening workspace
"""
w2 = Util.GetWorkspaceFromFile(filename, workspacename)
if w2 == None:
print "ERROR : Cannot open workspace : ", workspacename
sys.exit(1)
"""
calculate p(s=0) from the workspace given
"""
pval = RooStats.get_Presult(w2, False, ntoys, calctype)
#print pval
#sigma = StatTools.GetSigma(pval)
#print sigma
UL = {}
UL["visXsec"] = uL_visXsec
UL["nObsInSR"] = uL_nobsinSR
UL["nExpInSR"] = uL_nexpinSR
UL["nExpInSRPlus1Sigma"] = uL_nexpinSRerrP
UL["nExpInSRMinus1Sigma"] = uL_nexpinSRerrM
UL["CLb"] = CLB
UL["p0"] = pval
UL["Z"] = StatTools.GetSigma(pval)
return UL
def frequentistCalculator( ):
"""Compute the significance (hypothesis test) using a frequentist calculator"""
warnings.filterwarnings( action='ignore', category=RuntimeWarning, message='.*class stack<RooAbsArg\*,deque<RooAbsArg\*> >' )
inFile = TFile('Rootfiles/workspace_QCD_RPVSt100tojj_FitP4Gaus_rooFit_1fb.root')
myWS = inFile.Get("myWS")
data = myWS.data("data_obs")
sbModel = myWS.obj("modelConfig")
sbModel.SetName("S+B Model")
#sbModel.GetParametersOfInterest().first().setVal(5.)
#sbModel.GetParametersOfInterest().first().setConstant()
sbModel.SetSnapshot( sbModel.GetParametersOfInterest() )
bModel = sbModel.Clone()
bModel.SetName("B Model")
bModel.GetParametersOfInterest().first().setVal(0.)
bModel.GetParametersOfInterest().first().setConstant()
bModel.SetSnapshot( bModel.GetParametersOfInterest() )
#c1 = TCanvas('c1', 'c1', 10, 10, 750, 500 )
#test = data.reduce( RooFit.Name('test'))
#testHisto = TH1D('test','test', 10, 80, 180)
#testHisto = test.fillHistogram( testHisto, RooArgList(x) )
#testHisto.Draw()
#x = RooRealVar( "x", "x", 50., 180. )
#testHisto = sbModel.GetPdf()
#xframe = x.frame()
#testHisto.plotOn( xframe )
#c1.SaveAs("test.pdf")
# create the AsymptoticCalculator from data,alt model, null model
################################### Here, roostats is fitting again and the parameters are wrong.
asym_calc = RooStats.AsymptoticCalculator(data, bModel, sbModel)
asym_calc.SetOneSidedDiscovery( True )
result = asym_calc.GetHypoTest()
result.Print()
tw = TStopwatch()
tw.Start()
# Frequentist calculator
fc = RooStats.FrequentistCalculator(data, bModel, sbModel)
fc.SetToys(10,10)
# create the test statistics
profll = RooStats.ProfileLikelihoodTestStat( sbModel.GetPdf() )
profll.SetOneSidedDiscovery( True )
# configure ToyMCSampler and set the test statistics
toymcs = RooStats.ToyMCSampler( fc.GetTestStatSampler() )
toymcs.SetTestStatistic( profll )
pc = RooStats.ProofConfig( myWS, 10, "", 0 ) #ROOT.kFALSE)
toymcs.SetProofConfig(pc) # enable proof
########################################## change to 0
if not sbModel.GetPdf().canBeExtended(): toymcs.SetNEventsPerToy(0)
# run the test
fqResult = RooStats.HypoTestResult( fc.GetHypoTest() )
fqResult.Print()
tw.Stop()
print tw.CpuTime(), tw.RealTime()
c1 = TCanvas('c1', 'c1', 10, 10, 750, 500 )
plot = RooStats.HypoTestPlot( fqResult )
plot.SetLogYaxis(true)
plot.Draw()
c1.SaveAs("Plots/Significance.pdf")
#Util.SetInterpolationCode(w,4)
print "Processing analysis " + sigSamples[0]
#if 'onestepCC' in sigSamples[0] and int(sigSamples[0].split("_")[3])>900:
# w.var("mu_SIG").setMax(50000.)
# print "Gluino mass above 900 - extending mu_SIG range \n"
## first asumptotic limit, to get a quick but reliable estimate for the upper limit
## dynamic evaluation of ranges
testStatType = 3
calcType = 2 #asympt
nToys = 1000
nPoints = 20 #mu sampling
hypo = RooStats.DoHypoTestInversion(w, 1, 2, testStatType, True, 20, 0,
-1)
# then reevaluate with proper settings
if (hypo != 0):
hypo.ExclusionCleanup()
i = 0
while (i < 5):
if min_CLs(hypo) > 0.05:
print "Starting rescan iteration: " + str(i)
hypo = RedoScan(w, testStatType, hypo)
if (hypo != 0):
hypo.ExclusionCleanup()
else:
break
[(4800, 5400)],
[(4200, 4800)],
[(3600, 4200)],
[(3000, 3600)],
[(2400, 3000)],
]
for signal, signalMass, massbinsset in [
("CIplusLL", "12000", massbinssets2),
("cs_ct14nlo_", "13000", massbinssets1),
("AntiCIplusLL", "12000", massbinssets2),
("DMAxial_Dijet_LO_Mphi_1_1p0_1p0_Mar5_gdmv_0_gdma_1p0_gv_0_ga_1", "6000",
""),
]:
for massbins in massbinsset:
limits = {}
name = "pvalue_LHCa" + signal + "_" + ("_".join([
s[0:4] for s in str(massbins).strip("[]").split("(")
])).strip("_") + "_exp_" + signalMass + "_2016"
print name
f1 = open(name + ".txt")
observed = 0
for l in f1.readlines():
if "CLb" in l:
pval = float(l.split("=")[1].split("+")[0].strip(" "))
print "pvalue", pval, "significance", RooStats.PValueToSignificance(
(1. - pval) / 2.)
if "Significance:" in l:
significance = float(l.strip().split(" ")[-1].strip(")"))
print "significance", significance
def GetBayesianInterval(filename="workspace.root",
wsname='myWS',
interactive=False):
pInFile = ROOT.TFile(filename, "read")
# load workspace
pWs = pInFile.Get("myWS")
if not pWs:
print "workspace ", wsname, " not found"
return -1
# printout workspace content
pWs.Print()
# load and print data from workspace
data = pWs.data("data")
data.Print()
# load and print S+B Model Config
pSbHypo = pWs.obj("SbHypo")
pSbHypo.Print()
# create RooStats Bayesian calculator and set parameters
# Metropolis-Hastings algorithm needs a proposal function
sp = RooStats.SequentialProposal(10.0)
mcmc = RooStats.MCMCCalculator(data, pSbHypo)
mcmc.SetConfidenceLevel(0.95)
mcmc.SetNumIters(100000) # Metropolis-Hastings algorithm iterations
mcmc.SetProposalFunction(sp)
mcmc.SetNumBurnInSteps(500) # first N steps to be ignored as burn-in
mcmc.SetLeftSideTailFraction(0.0)
mcmc.SetNumBins(
40) # for plotting only - does not affect limit calculation
# estimate credible interval
# NOTE: unfortunate notation: the UpperLimit() name refers
# to the upper boundary of an interval,
# NOT to the upper limit on the parameter of interest
# (it just happens to be the same for the one-sided
# interval starting at 0)
pMcmcInt = mcmc.GetInterval()
upper_bound = pMcmcInt.UpperLimit(pWs.var("xsec"))
lower_bound = pMcmcInt.LowerLimit(pWs.var("xsec"))
print "one-sided 95%.C.L. bayesian credible interval for xsec: ", "[", lower_bound, ", ", upper_bound, "]"
# make posterior PDF plot for POI
c1 = ROOT.TCanvas("posterior", "posterior")
plot = RooStats.MCMCIntervalPlot(pMcmcInt)
plot.Draw()
c1.SaveAs("bayesian_mcmc_posterior.pdf")
# make scatter plots to visualise the Markov chain
c2 = ROOT.TCanvas("xsec_vs_beta_lumi", "xsec vs lumi_beta")
plot.DrawChainScatter(pWs.var("xsec"), pWs.var("lumi_beta"))
c2.SaveAs("scatter_mcmc_xsec_vs_beta_lumi.pdf")
c3 = ROOT.TCanvas("xsec_vs_beta_efficiency", "xsec vs eff_beta")
plot.DrawChainScatter(pWs.var("xsec"), pWs.var("eff_beta"))
c3.SaveAs("scatter_mcmc_xsec_vs_beta_efficiency.pdf")
c4 = ROOT.TCanvas("xsec_vs_beta_nbkg", "xsec vs nbkg_beta")
plot.DrawChainScatter(pWs.var("xsec"), pWs.var("nbkg_beta"))
c4.SaveAs("scatter_mcmc_xsec_vs_beta_nbkg.pdf")
ROOT.gPad.Update()
if interactive:
raw_input("\npress <enter> to continue")