createDataCard.py

#!/usr/bin/env python

from __future__ import print_function, division

import math, argparse, pdb, os, sys
import ROOT
import tmGeneralUtils, tmCombineDataCardInterface
import MCTemplateReader

inputArgumentsParser = argparse.ArgumentParser(description='Create data card from MC and data systematics and nEvents data.')
inputArgumentsParser.add_argument('--outputPrefix', required=True, help='Prefix to output files.', type=str)
inputArgumentsParser.add_argument('--outputDirectory', default="analysis/dataCards/", help='Output directory.', type=str)
inputArgumentsParser.add_argument('--eventProgenitorMassBin', required=True, help='Event progenitor mass bin.', type=int)
inputArgumentsParser.add_argument('--neutralinoMassBin', required=True, help='Neutralino mass bin.', type=int)
inputArgumentsParser.add_argument('--crossSectionsFile', required=True, help='Path to dat file that contains cross-sections as a function of eventProgenitor mass, from which to get the fractional uncertainties.',type=str)
inputArgumentsParser.add_argument('--crossSectionsScale', required=True, help='Cross sections shift scale. Accepted values: +/- 1 or 0. Signal is scaled by the factor cross_section*(1+this_scale*cross_section_uncertainty)', type=int)
inputArgumentsParser.add_argument('--MCTemplatePath', default="plot_susyMasses_template.root", help='Path to root file that contains a TH2F with bins containing points with generated masses set to 1 and all other bins set to 0.', type=str)
inputArgumentsParser.add_argument('--inputFile_STRegionBoundaries', default="STRegionBoundaries.dat", help='Path to file with ST region boundaries. First bin is the normalization bin, and the last bin is the last boundary to infinity.', type=str)
inputArgumentsParser.add_argument('--inputFile_MCEventHistograms_signal', required=True, help='Input MC event histograms, signal.', type=str)
inputArgumentsParser.add_argument('--inputFile_MCEventHistograms_signal_loose', required=True, help='Input MC event histograms, loose signal.', type=str)
# inputArgumentsParser.add_argument('--inputFile_MCEventHistograms_control', required=True, help='Input MC event histograms, control.', type=str)
inputArgumentsParser.add_argument('--inputFile_MCUncertainties_signal', required=True, help='Input MC uncertainties, signal.', type=str)
inputArgumentsParser.add_argument('--inputFile_MCUncertainties_signal_loose', required=True, help='Input MC uncertainties, loose signal.', type=str)
# inputArgumentsParser.add_argument('--inputFile_MCUncertainties_control', required=True, help='Input MC uncertainties, control.', type=str)
inputArgumentsParser.add_argument('--inputFile_dataSystematics_signal', required=True, help='Input file containing fractional uncertainties from signal data.', type=str)
inputArgumentsParser.add_argument('--inputFile_dataSystematics_signal_loose', required=True, help='Input file containing fractional uncertainties from loose signal data.', type=str)
# inputArgumentsParser.add_argument('--inputFile_dataSystematics_control', required=True, help='Input file containing fractional uncertainties from control data.', type=str)
# inputArgumentsParser.add_argument('--inputFile_dataSystematics_MCShapeAdjustment_signal', required=True, help='Input file containing MC shape adjustments for the signal selection.', type=str)
# inputArgumentsParser.add_argument('--inputFile_dataSystematics_MCShapeAdjustment_signal_loose', required=True, help='Input file containing MC shape adjustments for the loose signal selection.', type=str)
# inputArgumentsParser.add_argument('--inputFile_dataSystematics_MCShapeAdjustment_control', required=True, help='Input file containing MC shape adjustments for the control selection.', type=str)
# inputArgumentsParser.add_argument('--inputFile_dataSystematics_dataMCRatioAdjustment_QCD_signal', required=True, help='Input file containing data/QCD MC ratio adjustments for the signal selection.', type=str)
# inputArgumentsParser.add_argument('--inputFile_dataSystematics_dataMCRatioAdjustment_QCD_signal_loose', required=True, help='Input file containing data/QCD MC ratio adjustments for the loose signal selection.', type=str)
# inputArgumentsParser.add_argument('--inputFile_dataSystematics_dataMCRatioAdjustment_diphoton_signal', required=True, help='Input file containing data/diphoton MC ratio adjustments for the signal selection.', type=str)
# inputArgumentsParser.add_argument('--inputFile_dataSystematics_dataMCRatioAdjustment_diphoton_signal_loose', required=True, help='Input file containing data/diphoton MC ratio adjustments for the loose signal selection.', type=str)
inputArgumentsParser.add_argument('--inputFile_dataSystematics_expectedEventCounters_signal', required=True, help='Input file containing expected number of events from signal data.', type=str)
inputArgumentsParser.add_argument('--inputFile_dataSystematics_expectedEventCounters_signal_loose', required=True, help='Input file containing expected number of events from loose signal data.', type=str)
# inputArgumentsParser.add_argument('--inputFile_dataSystematics_expectedEventCounters_control', required=True, help='Input file containing expected number of events from control data.', type=str)
inputArgumentsParser.add_argument('--inputFile_dataSystematics_observedEventCounters_signal', required=True, help='Input file containing observed number of events from signal data.', type=str)
inputArgumentsParser.add_argument('--inputFile_dataSystematics_observedEventCounters_signal_loose', required=True, help='Input file containing observed number of events from loose signal data.', type=str)
# inputArgumentsParser.add_argument('--inputFile_dataSystematics_observedEventCounters_control', required=True, help='Input file containing observed number of events from control data.', type=str)
inputArgumentsParser.add_argument('--mismodelingUncertainty', required=True, help='Uncertainty to account for mismodeling of the adjustments in MC, uncorrelated across all bins.', type=float)
inputArgumentsParser.add_argument('--totalLumi', required=True, help='Total luminosity in inv pb (required for MC stats gmN uncertainty).', type=float)
inputArgumentsParser.add_argument('--luminosityUncertainty', required=True, help='Uncertainty on the luminosity.', type=float)
inputArgumentsParser.add_argument('--addSignalToBackground', required=False, action='store_true', help='If this argument is passed, a signal with a strength 1 is added to the background. USE WITH CAUTION.')
inputArgumentsParser.add_argument('--runUnblinded', action='store_true', help="If this flag is set, then the signal region data is unblinded. Specifically, the entry for the observed number of events is filled from the data, rather than from the expectation values.")
inputArgumentsParser.add_argument('--usePoissonForAsimov', action='store_true', help="By default, if the observations are blinded, then the number of observed events in each bin is set to the QCD background expectation. If this flag is set, the observations are instead set to a random numbers in a Poisson distribution with the QCD expectation as the mean of the Poisson.")
inputArgumentsParser.add_argument('--treatMETUncertaintiesAsUncorrelated', action='store_true', help="By default, MET-associated uncertainties are treated as fully correlated. This flag sets all MET-associated uncertainties to fully uncorrelated.")
inputArgumentsParser.add_argument('--regionsToUse', required=True, help="Comma-separated list of regions to run on.", type=str)
inputArgumentsParser.add_argument('--allowLargeSystematics', action='store_true', help="By default, very large systematic errors (more than 1000%) are replaced with a more sane value, and a warning is printed. (They are probably because of one-off problems like unphysical event weights.) Setting this flag disables all such replacements.")
inputArguments = inputArgumentsParser.parse_args()

SYSTEMATIC_SIGNIFICANCE_THRESHOLD = 0.0005
LOGNORMAL_REGULARIZE_THRESHOLD = 0.01 # To "regularize" the lognormal error; the error in the column is supposed to be (1 plus deltaX/X), and if this is 0, the combine tool complains

if not((inputArguments.crossSectionsScale == 0) or (abs(inputArguments.crossSectionsScale) == 1)): sys.exit("ERROR: argument crossSectionsScale can only take values 0 or +/- 1. Currently, crossSectionsScale={cSS}".format(cSS=inputArguments.crossSectionsScale))

def get_dict_fromFile(inputPath, localSignalLabels, inputPrefix):
    fileContents_nEvents = tmGeneralUtils.getConfigurationFromFile(inputFilePath=inputPath)
    outputDict = {}
    for signalBinLabel in localSignalLabels:
        outputDict[signalBinLabel] = fileContents_nEvents["{iP}_{l}".format(iP=inputPrefix, l=signalBinLabel)]
    return outputDict

def get_dict_expectedNEvents_stealth(stealthNEventsHistograms, eventProgenitorMassBin, neutralinoMassBin, localSignalLabels, scaleFactor):
    outputDict = {}
    for signalBinLabel in localSignalLabels:
        outputDict[signalBinLabel] = scaleFactor*((stealthNEventsHistograms[signalBinLabel]).GetBinContent(eventProgenitorMassBin, neutralinoMassBin))
    return outputDict

def get_symmetric_data_systematics_from_file(localSignalLabels, dataSystematicLabels, sourceFile):
    sourceSystematics = tmGeneralUtils.getConfigurationFromFile(sourceFile)
    outputDict = {}
    for dataSystematicLabel in dataSystematicLabels:
        outputDict[dataSystematicLabel] = {}
        for signalBinLabel in localSignalLabels:
            outputDict[dataSystematicLabel][signalBinLabel] = 1.0 + sourceSystematics["fractionalUncertainty_{dSL}_{sBL}".format(dSL=dataSystematicLabel, sBL=signalBinLabel)]
    return outputDict

def get_asymmetric_data_systematics_from_file(localSignalLabels, dataSystematicLabels, sourceFile):
    sourceSystematics = tmGeneralUtils.getConfigurationFromFile(sourceFile)
    outputDict = {}
    for dataSystematicLabel in dataSystematicLabels:
        outputDict[dataSystematicLabel] = {}
        for signalBinLabel in localSignalLabels:
            outputDict[dataSystematicLabel][signalBinLabel] = {}
            for upDownLabel in ["Down", "Up"]:
                outputDict[dataSystematicLabel][signalBinLabel][upDownLabel] = 1.0 + sourceSystematics["fractionalUncertainty{uDL}_{dSL}_{sBL}".format(uDL=upDownLabel, dSL=dataSystematicLabel, sBL=signalBinLabel)]
    return outputDict

def get_gmN_MC_systematic_from_histogram(localSignalLabels, dict_localToGlobalBinLabels, inputHistograms_unweighted, expectedNEventsDict, eventProgenitorMassBin, neutralinoMassBin, defaultMCWeight, signalType):
    unweighted_nevts = {}
    alpha_values = {}
    for signalBinLabel in localSignalLabels:
        global_label = dict_localToGlobalBinLabels[signalType][signalBinLabel]
        unweighted_nevts[signalBinLabel] = int(0.5 + (inputHistograms_unweighted[signalBinLabel]).GetBinContent(eventProgenitorMassBin, neutralinoMassBin))
        if (unweighted_nevts[signalBinLabel] > 0):
            weighted_nevts = expectedNEventsDict[global_label]
            alpha_values[signalBinLabel] = weighted_nevts/(1.0*unweighted_nevts[signalBinLabel])
        else:
            alpha_values[signalBinLabel] = defaultMCWeight
    return (unweighted_nevts, alpha_values)

def build_data_systematic_with_check(list_signalTypes, dict_localToGlobalBinLabels, dict_localSignalLabelsToUse, dict_sources_dataSystematics, print_debug=False):
    if (print_debug):
        print("build_data_systematic_with_check called with the following arguments:")
        print("list_signalTypes: {l}".format(l=list_signalTypes))
        print("dict_localToGlobalBinLabels: {d}".format(d=str(dict_localToGlobalBinLabels)))
        print("dict_localSignalLabelsToUse: {d}".format(d=str(dict_localSignalLabelsToUse)))
        print("dict_sources_dataSystematics: {d}".format(d=str(dict_sources_dataSystematics)))
    outputDict = {}
    isSignificant = False
    for signalType in list_signalTypes:
        if (not(signalType in dict_localSignalLabelsToUse)): continue
        if (len(dict_localSignalLabelsToUse[signalType]) == 0): continue
        sourceDict_dataSystematics = dict_sources_dataSystematics[signalType]
        for localSignalBinLabel in dict_localSignalLabelsToUse[signalType]:
            globalSignalBinLabel = dict_localToGlobalBinLabels[signalType][localSignalBinLabel]
            outputDict[globalSignalBinLabel] = {}
            if isinstance(sourceDict_dataSystematics[localSignalBinLabel], dict):
                for bkgLabel in ["qcd"]:
                    outputDict[globalSignalBinLabel][bkgLabel] = {}
                    for upDownLabel in ["Down", "Up"]:
                        try:
                            if (inputArguments.allowLargeSystematics):
                                outputDict[globalSignalBinLabel][bkgLabel][upDownLabel] = sourceDict_dataSystematics[localSignalBinLabel][upDownLabel]
                            else:
                                outputDict[globalSignalBinLabel][bkgLabel][upDownLabel] = min(100.0, max(0.01, sourceDict_dataSystematics[localSignalBinLabel][upDownLabel]))
                            if (abs(outputDict[globalSignalBinLabel][bkgLabel][upDownLabel] - 1.0) > SYSTEMATIC_SIGNIFICANCE_THRESHOLD): isSignificant = True
                            if (abs(outputDict[globalSignalBinLabel][bkgLabel][upDownLabel]) < LOGNORMAL_REGULARIZE_THRESHOLD): outputDict[globalSignalBinLabel][bkgLabel][upDownLabel] = LOGNORMAL_REGULARIZE_THRESHOLD
                        except KeyError:
                            sys.exit("ERROR: sourceDict_dataSystematics[localSignalBinLabel] is a dict but does not have elements named \"Up\" or \"Down\". dict contents: {c}".format(c=sourceDict_dataSystematics[localSignalBinLabel]))
            else:
                if (inputArguments.allowLargeSystematics):
                    for bkgLabel in ["qcd"]:
                        outputDict[globalSignalBinLabel][bkgLabel] = sourceDict_dataSystematics[localSignalBinLabel]
                else:
                    for bkgLabel in ["qcd"]:
                        outputDict[globalSignalBinLabel][bkgLabel] = min(100.0, max(0.01, sourceDict_dataSystematics[localSignalBinLabel]))
                for bkgLabel in ["qcd"]:
                    if (abs(outputDict[globalSignalBinLabel][bkgLabel] - 1.0) > SYSTEMATIC_SIGNIFICANCE_THRESHOLD): isSignificant = True
                    if (abs(outputDict[globalSignalBinLabel][bkgLabel]) < LOGNORMAL_REGULARIZE_THRESHOLD): outputDict[globalSignalBinLabel][bkgLabel] = LOGNORMAL_REGULARIZE_THRESHOLD
    return (isSignificant, outputDict)

def build_MC_constant_systematic(list_signalTypes, dict_localToGlobalBinLabels, constantFractionalUncertainty):
    outputDict = {}
    for signalType in list_signalTypes:
        for localSignalBinLabel in (dict_localToGlobalBinLabels[signalType].keys()):
            globalSignalBinLabel = dict_localToGlobalBinLabels[signalType][localSignalBinLabel]
            outputDict[globalSignalBinLabel] = {}
            outputDict[globalSignalBinLabel]["stealth"] = 1.0 + constantFractionalUncertainty
    return outputDict

def get_MC_systematic_from_histogram(localSignalLabels, inputHistograms, eventProgenitorMassBin, neutralinoMassBin):
    outputDict = {}
    for signalBinLabel in localSignalLabels:
        outputDict[signalBinLabel] = 1.0 + (inputHistograms[signalBinLabel]).GetBinContent(eventProgenitorMassBin, neutralinoMassBin)
    return outputDict

# def get_signal_contamination_systematic_from_histogram(nSTSignalBins, inputHistograms, eventProgenitorMassBin, neutralinoMassBin):
#     # what's saved in the histogram = r = S/B
#     # we want 1 + (S/(S+B)) = 1 + (1+1/r)^(-1)
#     fractionalErrors_2Jets = {}
#     for STRegionIndex in range(1, 2 + nSTSignalBins):
#         localLabel = "STRegion{r}_2Jets".format(r=STRegionIndex)
#         contamination = (inputHistograms[localLabel]).GetBinContent(eventProgenitorMassBin, neutralinoMassBin)
#         fractionalErrors_2Jets[STRegionIndex] = 0.
#         try:
#             fractionalErrors_2Jets[STRegionIndex] = math.pow((1.0 + (1.0/contamination)), -1)
#         except ZeroDivisionError:
#             fractionalErrors_2Jets[STRegionIndex] = 0.
#     outputDict = {}
#     for nJetsBin in range(4, 7):
#         localLabel = "STRegion1_{n}Jets".format(n=nJetsBin)
#         contamination = (inputHistograms[localLabel]).GetBinContent(eventProgenitorMassBin, neutralinoMassBin)
#         fractionalError = 0.
#         try:
#             fractionalError = math.pow((1.0 + (1.0/contamination)), -1)
#         except ZeroDivisionError:
#             fractionalError = 0.
#         for STRegionIndex in range(2, 2 + nSTSignalBins):
#             localLabel = "STRegion{r}_{n}Jets".format(r=STRegionIndex, n=nJetsBin)
#             outputDict[localLabel] = 1.0 + math.sqrt(math.pow(fractionalError, 2) + math.pow(fractionalErrors_2Jets[1], 2) + math.pow(fractionalErrors_2Jets[STRegionIndex], 2))
#     return outputDict

def get_signal_contamination_dict(nSTSignalBins, inputHistograms, eventProgenitorMassBin, neutralinoMassBin):
    output_signal_contamination_dict = {}
    for nJetsBin in range(2, 7):
        output_signal_contamination_dict[nJetsBin] = {}
        STRegionsToFetch = None
        if (nJetsBin == 2): STRegionsToFetch = range(1, 8)
        elif (nJetsBin == 3): STRegionsToFetch = []
        else: STRegionsToFetch = [1]
        for STRegionIndex in STRegionsToFetch:
            localSignalBinLabel = "STRegion{r}_{n}Jets".format(r=STRegionIndex, n=nJetsBin)
            output_signal_contamination_dict[nJetsBin][STRegionIndex] = (inputHistograms[localSignalBinLabel]).GetBinContent(eventProgenitorMassBin, neutralinoMassBin)
    return output_signal_contamination_dict

def get_asymmetric_MC_systematic_from_histogram(localSignalLabels, inputHistograms, eventProgenitorMassBin, neutralinoMassBin):
    outputDict = {}
    for signalBinLabel in localSignalLabels:
        outputDict[signalBinLabel] = {}
        for UpDownShift in ["Down", "Up"]:
            outputDict[signalBinLabel][UpDownShift] = 1.0 + (inputHistograms[signalBinLabel][UpDownShift]).GetBinContent(eventProgenitorMassBin, neutralinoMassBin)
    return outputDict

def build_MC_systematic_with_check(list_signalTypes, dict_localToGlobalBinLabels, dict_localSignalLabelsToUse, dict_sources_dataSystematics):
    outputDict = {}
    isSignificant = False
    for signalType in list_signalTypes:
        if (not(signalType in dict_localSignalLabelsToUse)): continue
        if (len(dict_localSignalLabelsToUse[signalType]) == 0): continue
        sourceDict_MCSystematics = dict_sources_dataSystematics[signalType]
        for localSignalBinLabel in dict_localSignalLabelsToUse[signalType]:
            globalSignalBinLabel = dict_localToGlobalBinLabels[signalType][localSignalBinLabel]
            outputDict[globalSignalBinLabel] = {}
            if isinstance(sourceDict_MCSystematics[localSignalBinLabel], dict):
                outputDict[globalSignalBinLabel]["stealth"] = {}
                for upDownLabel in ["Down", "Up"]:
                    try:
                        if (inputArguments.allowLargeSystematics):
                            outputDict[globalSignalBinLabel]["stealth"][upDownLabel] = sourceDict_MCSystematics[localSignalBinLabel][upDownLabel]
                        else:
                            outputDict[globalSignalBinLabel]["stealth"][upDownLabel] = min(100.0, max(0.01, sourceDict_MCSystematics[localSignalBinLabel][upDownLabel]))
                        if (abs(outputDict[globalSignalBinLabel]["stealth"][upDownLabel] - 1.0) > SYSTEMATIC_SIGNIFICANCE_THRESHOLD): isSignificant = True
                        if (abs(outputDict[globalSignalBinLabel]["stealth"][upDownLabel]) < LOGNORMAL_REGULARIZE_THRESHOLD): outputDict[globalSignalBinLabel]["stealth"][upDownLabel] = LOGNORMAL_REGULARIZE_THRESHOLD
                    except KeyError:
                        sys.exit("ERROR: sourceDict_MCSystematics[localSignalBinLabel] is a dict but does not have elements named \"Up\" or \"Down\". dict contents: {c}".format(c=sourceDict_MCSystematics[localSignalBinLabel]))
            else:
                if (inputArguments.allowLargeSystematics):
                    outputDict[globalSignalBinLabel]["stealth"] = sourceDict_MCSystematics[localSignalBinLabel]
                else:
                    outputDict[globalSignalBinLabel]["stealth"] = min(100.0, max(0.01, sourceDict_MCSystematics[localSignalBinLabel]))
                if (abs(outputDict[globalSignalBinLabel]["stealth"] - 1.0) > SYSTEMATIC_SIGNIFICANCE_THRESHOLD): isSignificant = True
                if (abs(outputDict[globalSignalBinLabel]["stealth"]) < LOGNORMAL_REGULARIZE_THRESHOLD): outputDict[globalSignalBinLabel]["stealth"] = LOGNORMAL_REGULARIZE_THRESHOLD
    return (isSignificant, outputDict)

def build_MC_gmN_stats_systematic_with_check(list_signalTypes, dict_localToGlobalBinLabels, dict_localSignalLabelsToUse, dict_alpha_values):
    outputDict = {}
    for signalType in list_signalTypes:
        if (not(signalType in dict_localSignalLabelsToUse)): continue
        if (len(dict_localSignalLabelsToUse[signalType]) == 0): continue
        sourceDict_alphaValues = dict_alpha_values[signalType]
        for localSignalBinLabel in dict_localSignalLabelsToUse[signalType]:
            globalSignalBinLabel = dict_localToGlobalBinLabels[signalType][localSignalBinLabel]
            outputDict[globalSignalBinLabel] = {}
            outputDict[globalSignalBinLabel]["stealth"] = sourceDict_alphaValues[localSignalBinLabel]
    return (True, outputDict)

def createDataCard(outputPath,
                   signalBinLabels,
                   observedNEvents,
                   expectedNEvents_qcd,
                   systematics_data,
                   systematics_data_labels,
                   systematics_data_types,
                   expectedNEvents_stealth,
                   systematics_MC,
                   systematics_MC_labels,
                   systematics_MC_types,
                   rateParamLabels,
                   rateParamProperties
):
    expectedNEvents = {}
    for signalBinLabel in signalBinLabels:
        expectedNEvents[signalBinLabel] = {}
        expectedNEvents[signalBinLabel]["qcd"] = expectedNEvents_qcd[signalBinLabel]
        expectedNEvents[signalBinLabel]["stealth"] = expectedNEvents_stealth[signalBinLabel]

    systematics = {}
    systematicsLabels = []
    systematicsTypes = {}
    for dataSystematicLabel in systematics_data_labels:
        systematics[dataSystematicLabel] = systematics_data[dataSystematicLabel]
        systematicsLabels.append(dataSystematicLabel)
        systematicsTypes[dataSystematicLabel] = systematics_data_types[dataSystematicLabel]
    for MCSystematicLabel in systematics_MC_labels:
        systematics[MCSystematicLabel] = systematics_MC[MCSystematicLabel]
        systematicsLabels.append(MCSystematicLabel)
        systematicsTypes[MCSystematicLabel] = systematics_MC_types[MCSystematicLabel]

    combineInterface = tmCombineDataCardInterface.tmCombineDataCardInterface(list_signalBinLabels=signalBinLabels, list_backgroundProcessLabels=["qcd"], list_signalProcessLabels=["stealth"], list_systematicsLabels=systematicsLabels, list_rateParamLabels=rateParamLabels, dict_rateParamProperties=rateParamProperties, dict_observedNEvents=observedNEvents, dict_expectedNEvents=expectedNEvents, dict_systematicsTypes=systematicsTypes, dict_systematics=systematics)
    combineInterface.writeToFile(outputFilePath=outputPath)

def get_MC_adjustment_file_path(signalType):
    return ("data/adjustments_all_MC_Bkg_{sT}.dat".format(sT=signalType))

def get_bkg_residual_adjustment_file_path(bkg, shift, signalType):
    return ("data/ratio_adjustment_all_MC_{b}_shift_{s}_{sT}.dat".format(b=bkg, s=shift, sT=signalType))

crossSectionsInputFileObject = open(inputArguments.crossSectionsFile, 'r')
crossSectionsDictionary = {}
crossSectionsFractionalUncertaintyDictionary = {}
for line in crossSectionsInputFileObject:
    crossSectionsData = line.split()
    eventProgenitorMassInt = int(0.5 + float(crossSectionsData[0]))
    crossSection = float(crossSectionsData[1])
    crossSectionFractionalUncertainty = 0.01*float(crossSectionsData[2])
    crossSectionsDictionary[eventProgenitorMassInt] = crossSection
    crossSectionsFractionalUncertaintyDictionary[eventProgenitorMassInt] = crossSectionFractionalUncertainty
crossSectionsInputFileObject.close()

STRegionBoundariesFileObject = open(inputArguments.inputFile_STRegionBoundaries, 'r')
nSTBoundaries = 0
STBoundaries = []
for STBoundaryString in STRegionBoundariesFileObject:
    if (STBoundaryString.strip()):
        nSTBoundaries += 1
        STBoundary = float(STBoundaryString.strip())
        STBoundaries.append(STBoundary)
STBoundaries.append(3500.0) # needed because we need a meaningful central value for last bin
nSTSignalBins = nSTBoundaries - 2 + 1 # First two lines are for the normalization bin, last boundary is at 3500
print("Using {n} signal bins for ST.".format(n = nSTSignalBins))
STRegionBoundariesFileObject.close()

signalTypesToUse = []
for signalType in ((inputArguments.regionsToUse).strip()).split(","):
    if not(signalType in ["signal", "signal_loose", "control"]): sys.exit("ERROR: Unrecognized region to use: {r}".format(r=signalType))
    signalTypesToUse.append(signalType)

abbreviated_signalTypes = {
    "signal": "s",
    "signal_loose": "l",
    # "control": "c"
}

inputDataFilePaths = {
    "signal": {
        "observations": inputArguments.inputFile_dataSystematics_observedEventCounters_signal,
        "expectations": inputArguments.inputFile_dataSystematics_expectedEventCounters_signal
    },
    "signal_loose": {
        "observations": inputArguments.inputFile_dataSystematics_observedEventCounters_signal_loose,
        "expectations": inputArguments.inputFile_dataSystematics_expectedEventCounters_signal_loose
    },
    # "control": {
    #     "observations": inputArguments.inputFile_dataSystematics_observedEventCounters_control,
    #     "expectations": inputArguments.inputFile_dataSystematics_expectedEventCounters_control
    # }
}

inputDataSystematicsFilePaths = {
    "signal": inputArguments.inputFile_dataSystematics_signal,
    "signal_loose": inputArguments.inputFile_dataSystematics_signal_loose,
    # "control": inputArguments.inputFile_dataSystematics_control
}

# inputMCShapeAdjustmentFilePaths = {
#     "signal": inputArguments.inputFile_dataSystematics_MCShapeAdjustment_signal,
#     "signal_loose": inputArguments.inputFile_dataSystematics_MCShapeAdjustment_signal_loose,
#     "control": inputArguments.inputFile_dataSystematics_MCShapeAdjustment_control
# }

# inputDataMCRatioAdjustmentFilePaths = {
#     "QCD": {
#         "signal": inputArguments.inputFile_dataSystematics_dataMCRatioAdjustment_QCD_signal,
#         "signal_loose": inputArguments.inputFile_dataSystematics_dataMCRatioAdjustment_QCD_signal_loose
#     },
#     "diphoton": {
#         "signal": inputArguments.inputFile_dataSystematics_dataMCRatioAdjustment_diphoton_signal,
#         "signal_loose": inputArguments.inputFile_dataSystematics_dataMCRatioAdjustment_diphoton_signal_loose
#     }
# }

inputMCFilePaths = {
    "signal": {
        "eventHistograms": inputArguments.inputFile_MCEventHistograms_signal,
        "uncertainties": inputArguments.inputFile_MCUncertainties_signal
    },
    "signal_loose": {
        "eventHistograms": inputArguments.inputFile_MCEventHistograms_signal_loose,
        "uncertainties": inputArguments.inputFile_MCUncertainties_signal_loose
    },
    # "control": {
    #     "eventHistograms": inputArguments.inputFile_MCEventHistograms_control,
    #     "uncertainties": inputArguments.inputFile_MCUncertainties_control
    # }
}

localSignalBinLabels = []
globalSignalBinLabels = []
dict_localToGlobalBinLabels = {}
for signalType in signalTypesToUse:
    dict_localToGlobalBinLabels[signalType] = {}

for STRegionIndex in range(2, 2 + nSTSignalBins):
    for nJetsBin in range(4, 7):
        localSignalBinLabel = "STRegion{r}_{n}Jets".format(r=STRegionIndex, n=nJetsBin)
        localSignalBinLabels.append(localSignalBinLabel)
        for signalType in signalTypesToUse:
            dict_localToGlobalBinLabels[signalType][localSignalBinLabel] = (abbreviated_signalTypes[signalType] + "ST{r}J{n}".format(r=STRegionIndex, n=nJetsBin))
            globalSignalBinLabels.append(dict_localToGlobalBinLabels[signalType][localSignalBinLabel])

# Fetch MC histograms from input files
MCHistograms_weightedNEvents = {}
MCHistograms_unweightedNEvents = {}
MCHistograms_MCStatUncertainties = {}
MCHistograms_JECUncertainties = {}
MCHistograms_UnclusteredMETUncertainties = {}
MCHistograms_JERMETUncertainties = {}
MCHistograms_prefiringWeightsUncertainties = {}
MCHistograms_HLTUncertainties = {}
MCHistograms_photonScaleFactorUncertainties = {}
MCHistograms_signalContamination = {}
for signalType in signalTypesToUse:
    MCHistograms_weightedNEvents[signalType] = {}
    MCHistograms_unweightedNEvents[signalType] = {}
    MCHistograms_MCStatUncertainties[signalType] = {}
    MCHistograms_JECUncertainties[signalType] = {}
    MCHistograms_UnclusteredMETUncertainties[signalType] = {}
    MCHistograms_JERMETUncertainties[signalType] = {}
    MCHistograms_prefiringWeightsUncertainties[signalType] = {}
    MCHistograms_HLTUncertainties[signalType] = {}
    MCHistograms_photonScaleFactorUncertainties[signalType] = {}
    MCHistograms_signalContamination[signalType] = {}

MCEventHistograms = {}
MCUncertainties = {}
for signalType in signalTypesToUse:
    MCEventHistograms[signalType] = ROOT.TFile.Open(inputMCFilePaths[signalType]["eventHistograms"], "READ")
    MCUncertainties[signalType] = ROOT.TFile.Open(inputMCFilePaths[signalType]["uncertainties"], "READ")
    for signalBinLabel in localSignalBinLabels:
        MCHistograms_weightedNEvents[signalType][signalBinLabel] = ROOT.TH2F()
        MCEventHistograms[signalType].GetObject("h_lumiBasedYearWeightedNEvents_{sBL}".format(sBL=signalBinLabel), MCHistograms_weightedNEvents[signalType][signalBinLabel])
        if (not(MCHistograms_weightedNEvents[signalType][signalBinLabel])):
            sys.exit("ERROR: Histogram MCHistograms_weightedNEvents[signalType][signalBinLabel] appears to be a nullptr at STRegionIndex={r}, nJets={n}".format(r=STRegionIndex, n=nJetsBin))
        MCHistograms_unweightedNEvents[signalType][signalBinLabel] = ROOT.TH2I()
        MCEventHistograms[signalType].GetObject("h_totalNEvents_{sBL}".format(sBL=signalBinLabel), MCHistograms_unweightedNEvents[signalType][signalBinLabel])
        if (not(MCHistograms_unweightedNEvents[signalType][signalBinLabel])):
            sys.exit("ERROR: Histogram MCHistograms_unweightedNEvents[signalType][signalBinLabel] appears to be a nullptr at STRegionIndex={r}, nJets={n}".format(r=STRegionIndex, n=nJetsBin))
        MCHistograms_MCStatUncertainties[signalType][signalBinLabel] = {}
        MCHistograms_JECUncertainties[signalType][signalBinLabel] = {}
        MCHistograms_UnclusteredMETUncertainties[signalType][signalBinLabel] = {}
        MCHistograms_JERMETUncertainties[signalType][signalBinLabel] = {}
        MCHistograms_prefiringWeightsUncertainties[signalType][signalBinLabel] = {}
        MCHistograms_HLTUncertainties[signalType][signalBinLabel] = {}
        MCHistograms_photonScaleFactorUncertainties[signalType][signalBinLabel] = {}
        for UpDownShift in ["Down", "Up"]:
            MCHistograms_MCStatUncertainties[signalType][signalBinLabel][UpDownShift] = ROOT.TH2F()
            MCUncertainties[signalType].GetObject("h_MCStatisticsFractionalError{UDS}_{sBL}".format(UDS=UpDownShift, sBL=signalBinLabel), MCHistograms_MCStatUncertainties[signalType][signalBinLabel][UpDownShift])
            if (not(MCHistograms_MCStatUncertainties[signalType][signalBinLabel][UpDownShift])):
                sys.exit("ERROR: Histogram MCHistograms_MCStatUncertainties[signalType][signalBinLabel][UpDownShift] appears to be a nullptr at STRegionIndex={r}, nJets={n}".format(r=STRegionIndex, n=nJetsBin))
            MCHistograms_JECUncertainties[signalType][signalBinLabel][UpDownShift] = ROOT.TH2F()
            MCUncertainties[signalType].GetObject("h_JECUncertainty{UDS}_{sBL}".format(UDS=UpDownShift, sBL=signalBinLabel), MCHistograms_JECUncertainties[signalType][signalBinLabel][UpDownShift])
            if (not(MCHistograms_JECUncertainties[signalType][signalBinLabel][UpDownShift])):
                sys.exit("ERROR: Histogram MCHistograms_JECUncertainties[signalType][signalBinLabel][UpDownShift] appears to be a nullptr at STRegionIndex={r}, nJets={n}".format(r=STRegionIndex, n=nJetsBin))
            MCHistograms_UnclusteredMETUncertainties[signalType][signalBinLabel][UpDownShift] = ROOT.TH2F()
            MCUncertainties[signalType].GetObject("h_UnclusteredMETUncertainty{UDS}_{sBL}".format(UDS=UpDownShift, sBL=signalBinLabel), MCHistograms_UnclusteredMETUncertainties[signalType][signalBinLabel][UpDownShift])
            if (not(MCHistograms_UnclusteredMETUncertainties[signalType][signalBinLabel][UpDownShift])):
                sys.exit("ERROR: Histogram MCHistograms_UnclusteredMETUncertainties[signalType][signalBinLabel][UpDownShift] appears to be a nullptr at STRegionIndex={r}, nJets={n}".format(r=STRegionIndex, n=nJetsBin))
            MCHistograms_JERMETUncertainties[signalType][signalBinLabel][UpDownShift] = ROOT.TH2F()
            MCUncertainties[signalType].GetObject("h_JERMETUncertainty{UDS}_{sBL}".format(UDS=UpDownShift, sBL=signalBinLabel), MCHistograms_JERMETUncertainties[signalType][signalBinLabel][UpDownShift])
            if (not(MCHistograms_JERMETUncertainties[signalType][signalBinLabel][UpDownShift])):
                sys.exit("ERROR: Histogram MCHistograms_JERMETUncertainties[signalType][signalBinLabel][UpDownShift] appears to be a nullptr at STRegionIndex={r}, nJets={n}".format(r=STRegionIndex, n=nJetsBin))
            MCHistograms_prefiringWeightsUncertainties[signalType][signalBinLabel][UpDownShift] = ROOT.TH2F()
            MCUncertainties[signalType].GetObject("h_prefiringWeightsUncertainty{UDS}_{sBL}".format(UDS=UpDownShift, sBL=signalBinLabel), MCHistograms_prefiringWeightsUncertainties[signalType][signalBinLabel][UpDownShift])
            if (not(MCHistograms_prefiringWeightsUncertainties[signalType][signalBinLabel][UpDownShift])):
                sys.exit("ERROR: Histogram MCHistograms_prefiringWeightsUncertainties[signalType][signalBinLabel][UpDownShift] appears to be a nullptr at STRegionIndex={r}, nJets={n}".format(r=STRegionIndex, n=nJetsBin))
            MCHistograms_HLTUncertainties[signalType][signalBinLabel][UpDownShift] = ROOT.TH2F()
            MCUncertainties[signalType].GetObject("h_HLTUncertainty{UDS}_{sBL}".format(UDS=UpDownShift, sBL=signalBinLabel), MCHistograms_HLTUncertainties[signalType][signalBinLabel][UpDownShift])
            if (not(MCHistograms_HLTUncertainties[signalType][signalBinLabel][UpDownShift])):
                sys.exit("ERROR: Histogram MCHistograms_HLTUncertainties[signalType][signalBinLabel][UpDownShift] appears to be a nullptr at STRegionIndex={r}, nJets={n}".format(r=STRegionIndex, n=nJetsBin))
            MCHistograms_photonScaleFactorUncertainties[signalType][signalBinLabel][UpDownShift] = ROOT.TH2F()
            MCUncertainties[signalType].GetObject("h_photonMCScaleFactorUncertainty{UDS}_{sBL}".format(UDS=UpDownShift, sBL=signalBinLabel), MCHistograms_photonScaleFactorUncertainties[signalType][signalBinLabel][UpDownShift])
            if (not(MCHistograms_photonScaleFactorUncertainties[signalType][signalBinLabel][UpDownShift])):
                sys.exit("ERROR: Histogram MCHistograms_photonScaleFactorUncertainties[signalType][signalBinLabel][UpDownShift] appears to be a nullptr at STRegionIndex={r}, nJets={n}".format(r=STRegionIndex, n=nJetsBin))

    for STRegionIndex in range(1, 2 + nSTSignalBins):
        localSignalBinLabel = "STRegion{r}_2Jets".format(r=STRegionIndex)
        MCHistograms_signalContamination[signalType][localSignalBinLabel] = ROOT.TH2F()
        MCUncertainties[signalType].GetObject("h_signalContamination_{sBL}".format(sBL=localSignalBinLabel), MCHistograms_signalContamination[signalType][localSignalBinLabel])
        if not(MCUncertainties[signalType]):
            sys.exit("ERROR: Histogram MCHistograms_signalContamination[signalType][localSignalBinLabel] appears to be a nullptr at STRegionIndex={r}, nJets=2".format(r=STRegionIndex))
    for nJetsBin in range(4, 7):
        localSignalBinLabel = "STRegion1_{n}Jets".format(n=nJetsBin)
        MCHistograms_signalContamination[signalType][localSignalBinLabel] = ROOT.TH2F()
        MCUncertainties[signalType].GetObject("h_signalContamination_{sBL}".format(sBL=localSignalBinLabel), MCHistograms_signalContamination[signalType][localSignalBinLabel])
        if not(MCUncertainties[signalType]):
            sys.exit("ERROR: Histogram MCHistograms_signalContamination[signalType][localSignalBinLabel] appears to be a nullptr at STRegionIndex={r}, nJets=2".format(r=STRegionIndex))

templateReader = MCTemplateReader.MCTemplateReader(inputArguments.MCTemplatePath)
eventProgenitorBinIndex = inputArguments.eventProgenitorMassBin
eventProgenitorMass = (templateReader.eventProgenitorMasses)[eventProgenitorBinIndex]
neutralinoBinIndex = inputArguments.neutralinoMassBin
neutralinoMass = (templateReader.neutralinoMasses)[neutralinoBinIndex]

# This extra scale factor is the correction to the nominal signal (calculated assuming the nominal cross-section).
# To move the cross-section down(up), we use cross_section_corrected = cross_section_nominal*{1 -(+) fractional_uncertainty}
crossSectionCorrectionScale = 1.0 + (inputArguments.crossSectionsScale)*crossSectionsFractionalUncertaintyDictionary[int(0.5+eventProgenitorMass)]

# default MC weight (without considering HLT etc., useful for those bins for which we have 0 stats)
default_MC_weight = (crossSectionsDictionary[int(0.5+eventProgenitorMass)]*(inputArguments.totalLumi))/(templateReader.getTotalNEvents(eventProgenitorBinIndex, neutralinoBinIndex))

# MC systematics
systematics_MC = {}
systematics_MC_labels = []
systematics_MC_types = {}
# lumi systematic is fully correlated across all bins
systematics_MC_labels.append("lumi")
systematics_MC_types["lumi"] = "lnN"
systematics_MC["lumi"] = build_MC_constant_systematic(list_signalTypes=signalTypesToUse, dict_localToGlobalBinLabels=dict_localToGlobalBinLabels, constantFractionalUncertainty=inputArguments.luminosityUncertainty)

MCSystematicsSource_MCStatistics = {}
MCSystematicsSource_JEC = {}
MCSystematicsSource_Unclstrd = {}
MCSystematicsSource_JER = {}
MCSystematicsSource_pref = {}
MCSystematicsSource_HLT = {}
MCSystematicsSource_phoSF = {}
signal_contamination_dict = {}
# dataSystematicsSource_signalContamination = {}
for signalType in signalTypesToUse:
    MCSystematicsSource_MCStatistics[signalType] = get_asymmetric_MC_systematic_from_histogram(localSignalLabels=localSignalBinLabels, inputHistograms=MCHistograms_MCStatUncertainties[signalType], eventProgenitorMassBin=eventProgenitorBinIndex, neutralinoMassBin=neutralinoBinIndex)
    MCSystematicsSource_JEC[signalType] = get_asymmetric_MC_systematic_from_histogram(localSignalLabels=localSignalBinLabels, inputHistograms=MCHistograms_JECUncertainties[signalType], eventProgenitorMassBin=eventProgenitorBinIndex, neutralinoMassBin=neutralinoBinIndex)
    MCSystematicsSource_Unclstrd[signalType] = get_asymmetric_MC_systematic_from_histogram(localSignalLabels=localSignalBinLabels, inputHistograms=MCHistograms_UnclusteredMETUncertainties[signalType], eventProgenitorMassBin=eventProgenitorBinIndex, neutralinoMassBin=neutralinoBinIndex)
    MCSystematicsSource_JER[signalType] = get_asymmetric_MC_systematic_from_histogram(localSignalLabels=localSignalBinLabels, inputHistograms=MCHistograms_JERMETUncertainties[signalType], eventProgenitorMassBin=eventProgenitorBinIndex, neutralinoMassBin=neutralinoBinIndex)
    MCSystematicsSource_pref[signalType] = get_asymmetric_MC_systematic_from_histogram(localSignalLabels=localSignalBinLabels, inputHistograms=MCHistograms_prefiringWeightsUncertainties[signalType], eventProgenitorMassBin=eventProgenitorBinIndex, neutralinoMassBin=neutralinoBinIndex)
    MCSystematicsSource_HLT[signalType] = get_asymmetric_MC_systematic_from_histogram(localSignalLabels=localSignalBinLabels, inputHistograms=MCHistograms_HLTUncertainties[signalType], eventProgenitorMassBin=eventProgenitorBinIndex, neutralinoMassBin=neutralinoBinIndex)
    MCSystematicsSource_phoSF[signalType] = get_asymmetric_MC_systematic_from_histogram(localSignalLabels=localSignalBinLabels, inputHistograms=MCHistograms_photonScaleFactorUncertainties[signalType], eventProgenitorMassBin=eventProgenitorBinIndex, neutralinoMassBin=neutralinoBinIndex)
    signal_contamination_dict[signalType] = get_signal_contamination_dict(nSTSignalBins=nSTSignalBins, inputHistograms=MCHistograms_signalContamination[signalType], eventProgenitorMassBin=eventProgenitorBinIndex, neutralinoMassBin=neutralinoBinIndex)
    # dataSystematicsSource_signalContamination[signalType] = get_signal_contamination_systematic_from_histogram(nSTSignalBins=nSTSignalBins, inputHistograms=MCHistograms_signalContamination[signalType], eventProgenitorMassBin=eventProgenitorBinIndex, neutralinoMassBin=neutralinoBinIndex)

for signalType in signalTypesToUse:
    # MCStatistics and HLT systematics in each bin are uncorrelated
    for signalBinLabel in localSignalBinLabels:
        localLabelsToUse = {signalType: [signalBinLabel]}
        tmp = build_MC_systematic_with_check(list_signalTypes=signalTypesToUse, dict_localToGlobalBinLabels=dict_localToGlobalBinLabels, dict_localSignalLabelsToUse=localLabelsToUse, dict_sources_dataSystematics=MCSystematicsSource_MCStatistics)
        if (tmp[0]):
            systematicLabel = "MCStatistics_{sBL}_{sT}".format(sBL=signalBinLabel, sT=signalType)
            systematics_MC_labels.append(systematicLabel)
            systematics_MC_types[systematicLabel] = "lnN"
            systematics_MC[systematicLabel] = tmp[1]
        tmp = build_MC_systematic_with_check(list_signalTypes=signalTypesToUse, dict_localToGlobalBinLabels=dict_localToGlobalBinLabels, dict_localSignalLabelsToUse=localLabelsToUse, dict_sources_dataSystematics=MCSystematicsSource_HLT)
        if (tmp[0]):
            systematicLabel = "HLT_{sBL}_{sT}".format(sBL=signalBinLabel, sT=signalType)
            systematics_MC_labels.append(systematicLabel)
            systematics_MC_types[systematicLabel] = "lnN"
            systematics_MC[systematicLabel] = tmp[1]
        if (inputArguments.treatMETUncertaintiesAsUncorrelated):
            tmp = build_MC_systematic_with_check(list_signalTypes=signalTypesToUse, dict_localToGlobalBinLabels=dict_localToGlobalBinLabels, dict_localSignalLabelsToUse=localLabelsToUse, dict_sources_dataSystematics=MCSystematicsSource_Unclstrd)
            if (tmp[0]):
                systematicsLabel = "Unclstrd_{sBL}_{sT}".format(sBL=signalBinLabel, sT=signalType)
                systematics_MC_labels.append(systematicsLabel)
                systematics_MC_types[systematicsLabel] = "lnN"
                systematics_MC[systematicsLabel] = tmp[1]
            tmp = build_MC_systematic_with_check(list_signalTypes=signalTypesToUse, dict_localToGlobalBinLabels=dict_localToGlobalBinLabels, dict_localSignalLabelsToUse=localLabelsToUse, dict_sources_dataSystematics=MCSystematicsSource_JER)
            if (tmp[0]):
                systematicsLabel = "JER_{sBL}_{sT}".format(sBL=signalBinLabel, sT=signalType)
                systematics_MC_labels.append(systematicsLabel)
                systematics_MC_types[systematicsLabel] = "lnN"
                systematics_MC[systematicsLabel] = tmp[1]

# All other MC uncertainties are correlated across all bins
localLabelsToUse = {signalType: localSignalBinLabels}
tmp = build_MC_systematic_with_check(list_signalTypes=signalTypesToUse, dict_localToGlobalBinLabels=dict_localToGlobalBinLabels, dict_localSignalLabelsToUse=localLabelsToUse, dict_sources_dataSystematics=MCSystematicsSource_JEC)
if (tmp[0]):
    systematics_MC_labels.append("JEC")
    systematics_MC_types["JEC"] = "lnN"
    systematics_MC["JEC"] = tmp[1]
if not(inputArguments.treatMETUncertaintiesAsUncorrelated):
    tmp = build_MC_systematic_with_check(list_signalTypes=signalTypesToUse, dict_localToGlobalBinLabels=dict_localToGlobalBinLabels, dict_localSignalLabelsToUse=localLabelsToUse, dict_sources_dataSystematics=MCSystematicsSource_Unclstrd)
    if (tmp[0]):
        systematics_MC_labels.append("Unclstrd")
        systematics_MC_types["Unclstrd"] = "lnN"
        systematics_MC["Unclstrd"] = tmp[1]
    tmp = build_MC_systematic_with_check(list_signalTypes=signalTypesToUse, dict_localToGlobalBinLabels=dict_localToGlobalBinLabels, dict_localSignalLabelsToUse=localLabelsToUse, dict_sources_dataSystematics=MCSystematicsSource_JER)
    if (tmp[0]):
        systematics_MC_labels.append("JER")
        systematics_MC_types["JER"] = "lnN"
        systematics_MC["JER"] = tmp[1]
tmp = build_MC_systematic_with_check(list_signalTypes=signalTypesToUse, dict_localToGlobalBinLabels=dict_localToGlobalBinLabels, dict_localSignalLabelsToUse=localLabelsToUse, dict_sources_dataSystematics=MCSystematicsSource_pref)
if (tmp[0]):
    systematics_MC_labels.append("pref")
    systematics_MC_types["pref"] = "lnN"
    systematics_MC["pref"] = tmp[1]
tmp = build_MC_systematic_with_check(list_signalTypes=signalTypesToUse, dict_localToGlobalBinLabels=dict_localToGlobalBinLabels, dict_localSignalLabelsToUse=localLabelsToUse, dict_sources_dataSystematics=MCSystematicsSource_phoSF)
if (tmp[0]):
    systematics_MC_labels.append("phoSF")
    systematics_MC_types["phoSF"] = "lnN"
    systematics_MC["phoSF"] = tmp[1]

expectedNEvents_stealth = {}
for signalType in signalTypesToUse:
    expectedNEventsLocal_stealth = get_dict_expectedNEvents_stealth(stealthNEventsHistograms=MCHistograms_weightedNEvents[signalType], eventProgenitorMassBin=eventProgenitorBinIndex, neutralinoMassBin=neutralinoBinIndex, localSignalLabels=localSignalBinLabels, scaleFactor=crossSectionCorrectionScale)
    for localLabel in localSignalBinLabels:
        globalLabel = dict_localToGlobalBinLabels[signalType][localLabel]
        expectedNEvents_stealth[globalLabel] = expectedNEventsLocal_stealth[localLabel]

# Data systematics
systematics_data = {}
systematics_data_labels = []
systematics_data_types = {}
expectedNEvents_qcd = {}
for signalType in signalTypesToUse:
    expectedNEventsLocal_qcd = get_dict_fromFile(inputPath=inputDataFilePaths[signalType]["expectations"], localSignalLabels=localSignalBinLabels, inputPrefix="expectedNEvents")
    adjustments_from_MC = tmGeneralUtils.getConfigurationFromFile(inputFilePath=get_MC_adjustment_file_path(signalType))
    residual_adjustments_systematic_dict = {}
    for bkg in ["Diph", "GJet", "QCD"]:
        residual_adjustments_systematic_dict[bkg] = {}
        for shift in ["up", "down"]:
            residual_adjustments_systematic_dict[bkg][shift] = tmGeneralUtils.getConfigurationFromFile(inputFilePath=get_bkg_residual_adjustment_file_path(bkg, shift, signalType))
    # adjustments_ratio_dataToMC_QCD = tmGeneralUtils.getConfigurationFromFile(inputFilePath=inputDataMCRatioAdjustmentFilePaths["QCD"][signalType])
    # adjustments_ratio_dataToMC_diphoton = tmGeneralUtils.getConfigurationFromFile(inputFilePath=inputDataMCRatioAdjustmentFilePaths["diphoton"][signalType])
    for nJetsBin in range(4, 7):
        localLabelsToUse = {signalType: []}
        scalingMCShapeSystematic = {mode: {signalType: {}} for mode in ["mode0", "mode1"]}
        for STRegionIndex in range(2, 2 + nSTSignalBins):
            localLabel = "STRegion{r}_{n}Jets".format(r=STRegionIndex, n=nJetsBin)
            globalLabel = dict_localToGlobalBinLabels[signalType][localLabel]
            scalingBkgCompositionResidualSystematic = {signalType: {}}
            # dict_localSignalLabelsToUse_signalContamination = {}
            # dict_localSignalLabelsToUse_signalContamination[signalType] = [localLabel]
            # tmp = build_data_systematic_with_check(list_signalTypes=[signalType], dict_localToGlobalBinLabels=dict_localToGlobalBinLabels, dict_localSignalLabelsToUse=dict_localSignalLabelsToUse_signalContamination, dict_sources_dataSystematics=dataSystematicsSource_signalContamination)
            # if (tmp[0]):
            #     systematicsLabel = "sigContam_{n}Jets_STRegion{i}_{sT}".format(n=nJetsBin, i=STRegionIndex, sT=signalType)
            #     systematics_data_labels.append(systematicsLabel)
            #     systematics_data_types[systematicsLabel] = "lnN"
            #     systematics_data[systematicsLabel] = tmp[1]
            adjustment_MCShape = adjustments_from_MC["nominalAdjustment_{l}".format(l=localLabel)]
            adjustment_deviationsFrom1 = []
            residual_adjustment_string = "ratio_adjustment_{l}".format(l=localLabel)
            for bkg in ["Diph", "GJet", "QCD"]:
                for shift in ["up", "down"]:
                    adjustment_ratio = residual_adjustments_systematic_dict[bkg][shift][residual_adjustment_string]
                    adjustment_deviationsFrom1.append(abs(adjustment_ratio - 1.0))
            # adjustment_ratio_dataToMC_QCD = adjustments_ratio_dataToMC_QCD["ratio_adjustment_{l}".format(l=localLabel)]
            # adjustment_ratio_dataToMC_QCD_deviationFrom1 = abs(adjustment_ratio_dataToMC_QCD - 1.0)
            # adjustment_ratio_dataToMC_diphoton = adjustments_ratio_dataToMC_diphoton["ratio_adjustment_{l}".format(l=localLabel)]
            # adjustment_ratio_dataToMC_diphoton_deviationFrom1 = abs(adjustment_ratio_dataToMC_diphoton - 1.0)
            # residual_systematic_bkgComposition = max(adjustment_ratio_dataToMC_QCD_deviationFrom1, adjustment_ratio_dataToMC_diphoton_deviationFrom1)
            residual_systematic_bkgComposition = max(adjustment_deviationsFrom1)
            # scalingBkgCompositionResidualSystematic[signalType][localLabel] = 1.0 + 2.0*residual_systematic_bkgComposition # "twice the adjustment" to be conservative
            scalingBkgCompositionResidualSystematic[signalType][localLabel] = 1.0 + residual_systematic_bkgComposition
            expectedNEvents_qcd[globalLabel] = max(0., adjustment_MCShape*(expectedNEventsLocal_qcd[localLabel]))
            for mode in ["mode0", "mode1"]:
                adjustment_MCShape_plus1Sigma = adjustments_from_MC["fractionalUncertaintyUp_{m}_{l}".format(m=mode, l=localLabel)]
                adjustment_MCShape_minus1Sigma = adjustments_from_MC["fractionalUncertaintyDown_{m}_{l}".format(m=mode, l=localLabel)]
                scalingMCShapeSystematic[mode][signalType][localLabel] = {}
                scalingMCShapeSystematic[mode][signalType][localLabel]["Up"] = 1.0 + adjustment_MCShape_plus1Sigma
                scalingMCShapeSystematic[mode][signalType][localLabel]["Down"] = 1.0 + adjustment_MCShape_minus1Sigma
            localLabelsToUse[signalType].append(localLabel)
            dict_localSignalLabelsToUse_tmp = {}
            dict_localSignalLabelsToUse_tmp[signalType] = [localLabel]
            tmp = build_data_systematic_with_check(list_signalTypes=[signalType], dict_localToGlobalBinLabels=dict_localToGlobalBinLabels, dict_localSignalLabelsToUse=dict_localSignalLabelsToUse_tmp, dict_sources_dataSystematics=scalingBkgCompositionResidualSystematic)
            if (tmp[0]):
                systematicsLabel = "bkg_comp_{n}Jets_STRegion{r}_{sT}".format(n=nJetsBin, r=STRegionIndex, sT=signalType)
                systematics_data_labels.append(systematicsLabel)
                systematics_data_types[systematicsLabel] = "lnN"
                systematics_data[systematicsLabel] = tmp[1]
            STMismodelingSystematic = {signalType: {}}
            STMismodelingSystematic[signalType][localLabel] = 1.0 + inputArguments.mismodelingUncertainty
            tmp = build_data_systematic_with_check(list_signalTypes=[signalType], dict_localToGlobalBinLabels=dict_localToGlobalBinLabels, dict_localSignalLabelsToUse=dict_localSignalLabelsToUse_tmp, dict_sources_dataSystematics=STMismodelingSystematic)
            if (tmp[0]):
                systematicsLabel = "mismodeling_{n}Jets_STRegion{r}_{sT}".format(n=nJetsBin, r=STRegionIndex, sT=signalType)
                systematics_data_labels.append(systematicsLabel)
                systematics_data_types[systematicsLabel] = "lnN"
                systematics_data[systematicsLabel] = tmp[1]
        # tmp = build_data_systematic_with_check(list_signalTypes=[signalType], dict_localToGlobalBinLabels=dict_localToGlobalBinLabels, dict_localSignalLabelsToUse=localLabelsToUse, dict_sources_dataSystematics=scalingBkgCompositionResidualSystematic)
        # if (tmp[0]):
        #     systematicsLabel = "scaling_DataMCRatio_{n}Jets_{sT}".format(n=nJetsBin, sT=signalType)
        #     systematics_data_labels.append(systematicsLabel)
        #     systematics_data_types[systematicsLabel] = "lnN"
        #     systematics_data[systematicsLabel] = tmp[1]
        for mode in ["mode0", "mode1"]:
            tmp = build_data_systematic_with_check(list_signalTypes=[signalType], dict_localToGlobalBinLabels=dict_localToGlobalBinLabels, dict_localSignalLabelsToUse=localLabelsToUse, dict_sources_dataSystematics=scalingMCShapeSystematic[mode])
            if (tmp[0]):
                systematicsLabel = "scaling_MCShape_{m}_{n}Jets_{sT}".format(m=mode, n=nJetsBin, sT=signalType)
                systematics_data_labels.append(systematicsLabel)
                systematics_data_types[systematicsLabel] = "lnN"
                systematics_data[systematicsLabel] = tmp[1]

# Correct signal expectation for potential signal contamination
monitoredQuantities = []
for signalType in signalTypesToUse:
    for nJetsBin in range(4, 7):
        for STRegionIndex in range(2, 2 + nSTSignalBins):
            localLabel = "STRegion{r}_{n}Jets".format(r=STRegionIndex, n=nJetsBin)
            globalLabel = dict_localToGlobalBinLabels[signalType][localLabel]
            stealth_expectation_uncorrected = expectedNEvents_stealth[globalLabel]
            signal_contamination_correction = (expectedNEvents_qcd[globalLabel])*(signal_contamination_dict[signalType][2][1] - signal_contamination_dict[signalType][2][STRegionIndex] - signal_contamination_dict[signalType][nJetsBin][1])
            expectedNEvents_stealth[globalLabel] = max(0.0, stealth_expectation_uncorrected + signal_contamination_correction) # signal expectation should be positive
            # print("At signalType={sT}, nJetsBin={n}, STRegionIndex={i}, uncorrected signal expectation: {u}, corrected signal expectation: {c}".format(sT=signalType, n=nJetsBin, i=STRegionIndex, u=stealth_expectation_uncorrected, c=expectedNEvents_stealth[globalLabel]))
            if (stealth_expectation_uncorrected > 0.0):
                monitoredQuantities.append(tuple(["float", "fractionalSignalCorrection_{s}_STRegion{r}_{n}Jets".format(s=signalType, r=STRegionIndex, n=nJetsBin), abs(signal_contamination_correction)/(stealth_expectation_uncorrected + expectedNEvents_qcd[globalLabel])]))
            else:
                monitoredQuantities.append(tuple(["float", "fractionalSignalCorrection_{s}_STRegion{r}_{n}Jets".format(s=signalType, r=STRegionIndex, n=nJetsBin), -1.0]))
            monitoredQuantities.append(tuple(["float", "signalCorrectionOverBackground_{s}_STRegion{r}_{n}Jets".format(s=signalType, r=STRegionIndex, n=nJetsBin), math.fabs(signal_contamination_correction)/(expectedNEvents_qcd[globalLabel])]))
            monitoredQuantities.append(tuple(["float", "signalCorrectionSignificance_{s}_STRegion{r}_{n}Jets".format(s=signalType, r=STRegionIndex, n=nJetsBin), math.fabs(signal_contamination_correction)/math.sqrt(expectedNEvents_qcd[globalLabel])]))
            if (math.fabs(signal_contamination_correction) > 0.0):
                monitoredQuantities.append(tuple(["float", "signalCorrectionNormTermsOverFull_{s}_STRegion{r}_{n}Jets".format(s=signalType, r=STRegionIndex, n=nJetsBin), math.fabs(signal_contamination_dict[signalType][2][1] - signal_contamination_dict[signalType][nJetsBin][1])/math.fabs(signal_contamination_correction)]))
            else:
                monitoredQuantities.append(tuple(["float", "signalCorrectionNormTermsOverFull_{s}_STRegion{r}_{n}Jets".format(s=signalType, r=STRegionIndex, n=nJetsBin), -1.0]))
tmGeneralUtils.writeConfigurationParametersToFile(configurationParametersList=monitoredQuantities, outputFilePath="{oD}/{oP}_signal_contamination_monitor_eventProgenitorMassBin{gBI}_neutralinoMassBin{nBI}.txt".format(oD=inputArguments.outputDirectory, oP=inputArguments.outputPrefix, gBI=eventProgenitorBinIndex, nBI=neutralinoBinIndex))

observedNEvents = {}
randomNumberGenerator = ROOT.TRandom3(1234)
for signalType in signalTypesToUse:
    observedNEventsLocal = {}
    if (inputArguments.runUnblinded): observedNEventsLocal = get_dict_fromFile(inputPath=inputDataFilePaths[signalType]["observations"], localSignalLabels=localSignalBinLabels, inputPrefix="observedNEvents")
    else:
        for localLabel in localSignalBinLabels:
            globalLabel = dict_localToGlobalBinLabels[signalType][localLabel]
            observedNEventsLocal[localLabel] = expectedNEvents_qcd[globalLabel]
    for localLabel in localSignalBinLabels:
        globalLabel = dict_localToGlobalBinLabels[signalType][localLabel]
        if (not(inputArguments.runUnblinded) and (inputArguments.usePoissonForAsimov)):
            observedNEvents[globalLabel] = randomNumberGenerator.Poisson(observedNEventsLocal[localLabel])
        else:
            observedNEvents[globalLabel] = observedNEventsLocal[localLabel]
        if (not(inputArguments.runUnblinded) and (inputArguments.addSignalToBackground)):
            if (inputArguments.usePoissonForAsimov):
                observedNEvents[globalLabel] = randomNumberGenerator.Poisson(observedNEventsLocal[localLabel] + expectedNEvents_stealth[globalLabel])
            else:
                observedNEvents[globalLabel] = observedNEventsLocal[localLabel] + expectedNEvents_stealth[globalLabel]

# rate params for ST scaling
rateParamLabels = []
rateParamProperties = {}
# Disabling
# for signalType in signalTypesToUse:
#     for nJetsBin in range(4, 7):
#         rateParamLabels.append("const_{s}_{n}Jets".format(s=abbreviated_signalTypes[signalType], n=nJetsBin))
#         rateParamLabels.append("slope_{s}_{n}Jets".format(s=abbreviated_signalTypes[signalType], n=nJetsBin))
#         list_globalLabels_rateParams = []
#         for STRegionIndex in range(2, 2 + nSTSignalBins):
#             localSignalBinLabel = "STRegion{r}_{n}Jets".format(r=STRegionIndex, n=nJetsBin)
#             globalSignalBinLabel = dict_localToGlobalBinLabels[signalType][localSignalBinLabel]
#             list_globalLabels_rateParams.append(globalSignalBinLabel)
#         rateParamProperties["const_{s}_{n}Jets".format(s=abbreviated_signalTypes[signalType], n=nJetsBin)] = (list_globalLabels_rateParams, ["qcdconst"], 1.0, 0.0, 5.0)
#         # min allowed value of slope is the value that makes the ratio drop by 1.0 over the full ST range
#         # 0.5*(STBoundaries[-1] + STBoundaries[-2]) is the midpoint of the last ST bin
#         # 0.5*(STBoundaries[0] + STBoundaries[1]) is the midpoint of the first ST bin
#         # Disabling for now, easier to interpret if you just use 0 as the minimum slope
#         # rateParamProperties["slope_{s}_{n}Jets".format(s=abbreviated_signalTypes[signalType], n=nJetsBin)] = (list_globalLabels_rateParams, ["qcdlin"], 0.0, -(1000.0/(0.5*(STBoundaries[-1] + STBoundaries[-2]) - 0.5*(STBoundaries[0] + STBoundaries[1]))), 5.0)
#         rateParamProperties["slope_{s}_{n}Jets".format(s=abbreviated_signalTypes[signalType], n=nJetsBin)] = (list_globalLabels_rateParams, ["qcdlin"], 0.0, 0.0, 5.0)

# Set all other data systematics
for signalType in signalTypesToUse:
    sources_symmetricDataSystematicsLocal = get_symmetric_data_systematics_from_file(localSignalLabels=localSignalBinLabels, dataSystematicLabels=["shape", "rho"], sourceFile=inputDataSystematicsFilePaths[signalType])
    sources_asymmetricDataSystematicsLocal = get_asymmetric_data_systematics_from_file(localSignalLabels=localSignalBinLabels, dataSystematicLabels=["normEvents"], sourceFile=inputDataSystematicsFilePaths[signalType])
    sources_dataSystematics = {}
    for dataSystematic in ["shape", "rho"]:
        sources_dataSystematics[dataSystematic] = {signalType: sources_symmetricDataSystematicsLocal[dataSystematic]}
    for dataSystematic in ["normEvents"]:
        sources_dataSystematics[dataSystematic] = {signalType: sources_asymmetricDataSystematicsLocal[dataSystematic]}
    for nJetsBin in range(4, 7):
        localLabelsToUse = {signalType: []}
        for STRegionIndex in range(2, 2+nSTSignalBins):
            signalBinLabel = "STRegion{r}_{n}Jets".format(r=STRegionIndex, n=nJetsBin)
            localLabelsToUse[signalType].append(signalBinLabel)
        sources_normSystematics = {signalType: sources_asymmetricDataSystematicsLocal["normEvents"]}
        tmp = build_data_systematic_with_check(list_signalTypes=[signalType], dict_localToGlobalBinLabels=dict_localToGlobalBinLabels, dict_localSignalLabelsToUse=localLabelsToUse, dict_sources_dataSystematics=sources_dataSystematics["normEvents"])
        if (tmp[0]):
            systematicsLabel = "normEvents_{n}Jets_{sT}".format(n=nJetsBin, sT=signalType)
            systematics_data_labels.append(systematicsLabel)
            systematics_data_types[systematicsLabel] = "lnN"
            systematics_data[systematicsLabel] = tmp[1]
    # shape and rho systematics are correlated across all nJets bins
    for dataSystematic in ["shape", "rho"]:
        for STRegionIndex in range(2, 2+nSTSignalBins):
            localLabelsToUse = {signalType: []}
            for nJetsBin in range(4, 7):
                signalBinLabel = "STRegion{r}_{n}Jets".format(r=STRegionIndex, n=nJetsBin)
                localLabelsToUse[signalType].append(signalBinLabel)
            tmp = build_data_systematic_with_check(list_signalTypes=[signalType], dict_localToGlobalBinLabels=dict_localToGlobalBinLabels, dict_localSignalLabelsToUse=localLabelsToUse, dict_sources_dataSystematics=sources_dataSystematics[dataSystematic])
            if (tmp[0]):
                systematicsLabel = "{dS}_STRegion{r}_{sT}".format(dS=dataSystematic, r=STRegionIndex, sT=signalType)
                systematics_data_labels.append(systematicsLabel)
                systematics_data_types[systematicsLabel] = "lnN"
                systematics_data[systematicsLabel] = tmp[1]

# Finally: if there are low statistics in some bins, replace lnN MC stats systematics with gmN systematics
# This has to be done at the end because it depends on the corrected values of the stealth expectations, for example

MCSystematicsSource_stats_gmN_N = {}
MCSystematicsSource_stats_gmN_alpha = {}
for signalType in signalTypesToUse:
    unweighted_nevts, alpha_values = get_gmN_MC_systematic_from_histogram(localSignalLabels=localSignalBinLabels, dict_localToGlobalBinLabels=dict_localToGlobalBinLabels, inputHistograms_unweighted=MCHistograms_unweightedNEvents[signalType], expectedNEventsDict=expectedNEvents_stealth, eventProgenitorMassBin=eventProgenitorBinIndex, neutralinoMassBin=neutralinoBinIndex, defaultMCWeight=default_MC_weight, signalType=signalType)
    MCSystematicsSource_stats_gmN_N[signalType] = unweighted_nevts
    MCSystematicsSource_stats_gmN_alpha[signalType] = alpha_values
    for signalBinLabel in localSignalBinLabels:
        localLabelsToUse = {signalType: [signalBinLabel]}
        gmN_N = MCSystematicsSource_stats_gmN_N[signalType][signalBinLabel]
        if (gmN_N < 25):
            tmp = build_MC_gmN_stats_systematic_with_check(list_signalTypes=signalTypesToUse, dict_localToGlobalBinLabels=dict_localToGlobalBinLabels, dict_localSignalLabelsToUse=localLabelsToUse, dict_alpha_values=MCSystematicsSource_stats_gmN_alpha)
            if (tmp[0]):
                systematicLabel = "MCStatistics_{sBL}_{sT}".format(sBL=signalBinLabel, sT=signalType)
                # this label has already been created, no need to append to systematics_MC_labels
                systematics_MC_types[systematicLabel] = "gmN {n}".format(n=gmN_N)
                systematics_MC[systematicLabel] = tmp[1]
        if (gmN_N == 0):
            systematics_to_remove_labels = ["HLT_{sBL}_{sT}".format(sBL=signalBinLabel, sT=signalType)]
            if (inputArguments.treatMETUncertaintiesAsUncorrelated):
                systematics_to_remove_labels.extend(["Unclstrd_{sBL}_{sT}".format(sBL=signalBinLabel, sT=signalType), "JER_{sBL}_{sT}".format(sBL=signalBinLabel, sT=signalType)])
            for systematics_to_remove_label in systematics_to_remove_labels:
                systematics_MC_types.pop(systematics_to_remove_label, None)
                systematics_MC.pop(systematics_to_remove_label, None)
                if systematics_to_remove_label in systematics_MC_labels:
                    systematics_MC_labels.remove(systematics_to_remove_label)

output_path="{oD}/{oP}_dataCard_eventProgenitorMassBin{gBI}_neutralinoMassBin{nBI}.txt".format(oD=inputArguments.outputDirectory, oP=inputArguments.outputPrefix, gBI=eventProgenitorBinIndex, nBI=neutralinoBinIndex)
if (inputArguments.addSignalToBackground):
    output_path = "{oD}/WITH_ADDED_SIGNAL_{oP}_dataCard_eventProgenitorMassBin{gBI}_neutralinoMassBin{nBI}.txt".format(oD=inputArguments.outputDirectory, oP=inputArguments.outputPrefix, gBI=eventProgenitorBinIndex, nBI=neutralinoBinIndex)
print("Saving data card at path: {p}".format(p=output_path))
createDataCard(outputPath=output_path,
               signalBinLabels=globalSignalBinLabels,
               observedNEvents=observedNEvents,
               expectedNEvents_qcd=expectedNEvents_qcd,
               systematics_data=systematics_data,
               systematics_data_labels=systematics_data_labels,
               systematics_data_types=systematics_data_types,
               expectedNEvents_stealth=expectedNEvents_stealth,
               systematics_MC=systematics_MC,
               systematics_MC_labels=systematics_MC_labels,
               systematics_MC_types=systematics_MC_types,
               rateParamLabels=rateParamLabels,
               rateParamProperties=rateParamProperties
)

for signalType in signalTypesToUse:
    MCEventHistograms[signalType].Close()
    MCUncertainties[signalType].Close()