def createSamples(mode, analysis_dir, total_weight, qcd_from_same_sign, w_qcd_mssm_method, r_qcd_os_ss):
hist_dict = {}
sample_dict = {}
samples_mc, samples_data, samples, all_samples, sampleDict = createSampleLists(analysis_dir=analysis_dir, mode=mode)
if mode == 'mssm_control' or not 'mssm' in mode:
all_samples = [s for s in all_samples if not 'ggH' in s.name and not 'bbH' in s.name]
sample_dict['all_samples'] = all_samples
if qcd_from_same_sign and not w_qcd_mssm_method:
samples_qcdfromss = [s for s in all_samples if s.name != 'QCD']
samples_ss = copy.deepcopy(samples_qcdfromss)
samples_ss = [s for s in samples_ss if not s.is_signal]
for sample in samples_ss:
if sample.name != 'data_obs':
# Subtract background from data
sample.scale = -1.
qcd = HistogramCfg(name='QCD', var=None, cfgs=samples_ss, cut=None, total_scale=r_qcd_os_ss, lumi=int_lumi, weight=total_weight)
samples_qcdfromss.append(qcd)
sample_dict['samples_qcdfromss'] = samples_qcdfromss
if w_qcd_mssm_method:
sample_dict['samples_mssm_method'] = createQCDWHistograms(samples, hist_dict, int_lumi, weight=total_weight, r_qcd_os_ss=r_qcd_os_ss)
return sample_dict, hist_dict
def qcd_estimation(B_cut,
C_cut,
D_cut,
all_samples,
int_lumi,
total_weight,
scale=1.,
verbose=True,
friend_func=None):
'''ABCD method.
A | B
-------
C | D
A is the signal region
B is where the shape is taken from
C/D gives the scale factor to be applied to the shape in B
Returns an updated list of samples that includes the QCD HistgramCfg.
'''
norm_var = dict_all_vars['_norm_']
QCD_C_region_cut = C_cut
QCD_D_region_cut = D_cut
QCD_B_region_cut = B_cut
samples_qcd_copy = copy.deepcopy(
[s for s in all_samples if s.name != 'QCD' and not s.is_signal])
samples_qcd_copy = [s for s in samples_qcd_copy if not s.is_signal]
for sample in samples_qcd_copy:
sample.scale = scale if sample.name == 'data_obs' else -scale
qcd_c_region = HistogramCfg(name='QCD_C_region',
var=norm_var,
cfgs=samples_qcd_copy,
cut=str(QCD_C_region_cut),
lumi=int_lumi,
weight=total_weight)
qcd_d_region = HistogramCfg(name='QCD_D_region',
var=norm_var,
cfgs=samples_qcd_copy,
cut=str(QCD_D_region_cut),
lumi=int_lumi,
weight=total_weight)
# samples_qcd = [qcd_c_region, qcd_d_region]
# cfg_qcd = HistogramCfg(name='QCD_aux', var=None, cfgs=samples_qcd, cut=None, lumi=int_lumi, weight=total_weight)
plot_qcd_c = createHistogram(qcd_c_region,
all_stack=True,
friend_func=friend_func)
plot_qcd_d = createHistogram(qcd_d_region,
all_stack=True,
friend_func=friend_func)
if verbose:
print 'Histogram C region'
print plot_qcd_c
print 'Histogram D region'
print plot_qcd_d
yield_c = plot_qcd_c.GetStack().totalHist.Yield()
yield_d = plot_qcd_d.GetStack().totalHist.Yield()
if yield_d == 0.:
print 'WARNING: no events left for the QCD estimation. Set to 0'
qcd_scale = 0.
else:
qcd_scale = yield_c / yield_d
if qcd_scale < 0.:
print 'WARNING: negative QCD scaling; set it to zero'
qcd_scale = 0.
verbose = True
if verbose:
print 'QCD estimation: '
print ' Yield C:', yield_c, ' yield D:', yield_d
print ' Ratio C/D', qcd_scale
qcd_b_region_hist = HistogramCfg(name='QCD',
var=None,
cfgs=samples_qcd_copy,
cut=str(QCD_B_region_cut),
lumi=int_lumi,
weight=total_weight,
total_scale=qcd_scale)
all_samples_qcd = copy.deepcopy(all_samples)
all_samples_qcd = [qcd_b_region_hist] + all_samples_qcd
return all_samples_qcd
analysis_dir = '/data/clange/ntuples/FixNeutrino/'
tree_prod_name = ''
samples_mc, samples_data, samples, all_samples, sampleDict = createSampleLists(analysis_dir, channel='WV', weight=weight_MC)
# Taken from Variables.py, can get subset with e.g. getVars(['mt', 'mvis'])
variables = generic_vars + lnujj_vars + lnujj_vbf_vars
# variables = [lnujj_vars[0]]
# variables = getVars(['l1_reliso05', 'l2_reliso05'])
# variables = [
# VariableCfg(name='mvis', binning={'nbinsx':35, 'xmin':0, 'xmax':350}, unit='GeV', xtitle='m_{vis}')
# ]
for cut_name in cuts:
cfg_example = HistogramCfg(name='example', var=None, cfgs=samples, cut='', lumi=int_lumi, weight=total_weight)
cfg_example.cut = cuts[cut_name].replace("||HLT_MET120", "")
cfg_example.vars = variables
channel = "l#nujj"
if cut_name.find("_mu") >= 0:
channel = "#mu#nujj"
elif cut_name.find("_e") >= 0:
channel = "e#nujj"
if cut_name.find("HP") >= 0:
channel += " HP"
elif cut_name.find("LP") >= 0:
channel += " LP"
plots = createHistograms(cfg_example, verbose=False)
'cut': '&& abs(l2_gen_pdgId) == 21 ',
'colour': 1,
'style': 1
},
]
for set_name, sample_defs in sample_def_sets.items():
for cut_name in cuts:
for variable in variables:
plot_inputs = []
for sample_def in sample_defs:
cut_extra = sample_def['cut']
cfg_tight = HistogramCfg(name='tight',
var=None,
cfgs=samples_mc_w,
cut=inc_cut,
lumi=int_lumi,
weight=total_weight +
' * ((l2_gen_match == 6))')
cfg_loose = HistogramCfg(name='loose',
var=None,
cfgs=samples_mc_w,
cut=inc_cut,
lumi=int_lumi,
weight=total_weight +
' * ((l2_gen_match == 6))')
tight_cut = inc_cut + cuts[cut_name] + cut_extra
loose_cut = tight_cut.replace(
'l2_byIsolationMVArun2v1DBoldDMwLT>2.5', '1.')
loose_cut = tight_cut.replace(
def makePlots(variables,
cuts,
total_weight,
sample_dict,
hist_dict,
qcd_from_same_sign,
w_qcd_mssm_method,
mt_cut,
friend_func,
dc_postfix,
make_plots=True):
for cut in cuts:
if qcd_from_same_sign and not 'SS' in cut.name and not w_qcd_mssm_method:
cfg_main = HistogramCfg(name=cut.name,
var=None,
cfgs=sample_dict['samples_qcdfromss'],
cut=cut.cut,
lumi=int_lumi,
weight=total_weight)
elif w_qcd_mssm_method:
cfg_main = HistogramCfg(name=cut.name,
var=None,
cfgs=sample_dict['samples_mssm_method'],
cut=cut.cut,
lumi=int_lumi,
weight=total_weight)
hist_dict[
'wjets'].cut = cut.cut # since wjets is a sub-HistogramCfg
else:
cfg_main = HistogramCfg(name=cut.name,
var=None,
cfgs=sample_dict['all_samples'],
cut=cut.cut,
lumi=int_lumi,
weight=total_weight)
if qcd_from_same_sign and not 'SS' in cut.name:
hist_dict['qcd'].cut = cut.cut.replace('l1_charge != l2_charge',
'l1_charge == l2_charge')
if w_qcd_mssm_method:
hist_dict['stacknow_highmt_os'].name = 'HighMTOS' + cut.name
hist_dict['stacknow_highmt_ss'].name = 'HighMTSS' + cut.name
hist_dict['wjets_incl_os'].name = 'WInclOS' + cut.name
hist_dict['wjets_incl_ss'].name = 'WInclSS' + cut.name
hist_dict['wjets_highmt_os'].name = 'WHighMTOS' + cut.name
hist_dict['wjets_ss'].name = 'WJetsSS' + cut.name
hist_dict['qcd'].cut = cut.cut.replace('l1_charge != l2_charge',
'l1_charge == l2_charge')
hist_dict['wjets_ss'].cut = cut.cut.replace(
'l1_charge != l2_charge', 'l1_charge == l2_charge')
hist_dict['stacknow_highmt_os'].cut = cut.cut.replace(
mt_cut, 'mt>70') + '&& mt>70'
hist_dict['stacknow_highmt_ss'].cut = cut.cut.replace(
'l1_charge != l2_charge', 'l1_charge == l2_charge').replace(
mt_cut, 'mt>70') + '&& mt>70'
hist_dict['wjets_incl_os'].cut = cut.cut.replace(mt_cut, '1.')
hist_dict['wjets_incl_ss'].cut = cut.cut.replace(
'l1_charge != l2_charge',
'l1_charge == l2_charge').replace(mt_cut, '1.')
hist_dict['wjets_highmt_os'].cut = cut.cut.replace(
mt_cut, 'mt>70') + '&& mt>70'
plot_w_os = createHistogram(hist_dict['wjets_incl_os'],
verbose=False,
friend_func=friend_func)
plot_w_ss = createHistogram(hist_dict['wjets_incl_ss'],
verbose=False,
friend_func=friend_func)
r_w_os_ss = plot_w_os.GetStack().totalHist.Yield(
) / plot_w_ss.GetStack().totalHist.Yield()
print 'Inclusive W OS/SS ratio:', r_w_os_ss
plot_highmt_os = createHistogram(hist_dict['stacknow_highmt_os'],
all_stack=True,
verbose=False,
friend_func=friend_func)
plot_highmt_ss = createHistogram(hist_dict['stacknow_highmt_ss'],
all_stack=True,
verbose=False,
friend_func=friend_func)
createDefaultGroups(plot_highmt_os)
createDefaultGroups(plot_highmt_ss)
yield_highmt_os = plot_highmt_os.GetStack().totalHist.Yield()
yield_highmt_ss = plot_highmt_ss.GetStack().totalHist.Yield()
plot_w_highmt_os = createHistogram(hist_dict['wjets_highmt_os'],
verbose=False,
friend_func=friend_func)
yield_w_highmt_os = plot_w_highmt_os.GetStack().totalHist.Yield()
if r_w_os_ss < r_qcd_os_ss:
print 'WARNING, OS/SS ratio larger for QCD than for W+jets!', r_w_os_ss, r_qcd_os_ss
yield_estimation = r_w_os_ss * (yield_highmt_os -
r_qcd_os_ss * yield_highmt_ss) / (
r_w_os_ss - r_qcd_os_ss)
print 'High MT W+jets estimated yield', yield_estimation
print 'High MT W+jets MC yield', yield_w_highmt_os
w_sf = 0.
if yield_w_highmt_os:
w_sf = yield_estimation / yield_w_highmt_os
else:
print 'Warning: no MC events in high MT W+jets'
print 'W+jets scale factor:', w_sf, '\n'
if cut.name == 'vbf_highmva0':
print 'VBF very tight category, fixing W+jets SF to 1'
w_sf = 1.
hist_dict['wjets'].total_scale = w_sf
hist_dict['wjets_ss'].total_scale = -w_sf
cfg_main.vars = variables
if qcd_from_same_sign:
hist_dict[
'qcd'].vars = variables # Can put into function but we will not want it by default if we take normalisations from e.g. high MT
if w_qcd_mssm_method:
hist_dict[
'wjets'].vars = variables # Can put into function but we will not want it by default if we take normalisations from e.g. high MT
hist_dict['qcd'].vars = variables
hist_dict['wjets_ss'].vars = variables
for variable in variables:
if variable.name in [
'svfit_mass', 'svfit_transverse_mass', 'mvis'
] and 'mssm' in mode:
if cut.name in ['inclusive', 'nobtag']:
variable.binning = binning_mssm
elif cut.name in ['btag']:
variable.binning = binning_mssm_btag
plots = createHistograms(cfg_main,
verbose=False,
friend_func=friend_func)
for variable in variables:
# for plot in plots.itervalues():
plot = plots[variable.name]
createDefaultGroups(plot)
if not w_qcd_mssm_method:
plot.Group('W', ['W', 'W1Jets', 'W2Jets', 'W3Jets', 'W4Jets'])
# plot.Group('Electroweak', ['Diboson', 'W'])
# plot.Group('Single t', ['T_tWch', 'TBar_tWch', 'TToLeptons_sch', 'TToLeptons_tch'])
# plot.Group('ZLL', ['Ztt_ZL', 'Ztt_ZJ'], style=plot.Hist('Ztt_ZL').style)
if make_plots:
HistDrawer.draw(plot, plot_dir='plots/' + cut.name)
if variable.name in [
'mvis', 'svfit_transverse_mass', 'svfit_mass', 'mva',
'mva2div1', 'mva1', 'mva2', 'l2_nc_ratio'
]:
plot.WriteDataCard(
filename='datacard_{mode}_{var}.root'.format(
mode=mode, var=variable.name),
dir='mt_' + cut.name,
mode='UPDATE',
postfix=dc_postfix) #mt = mu-tau
def createQCDWHistograms(samples,
hist_dict,
int_lumi,
weight,
r_qcd_os_ss=1.17):
w_names = ['W', 'W1Jets', 'W2Jets', 'W3Jets', 'W4Jets']
samples_non_w = [
s for s in samples
if s.name != 'QCD' and s.name not in w_names and not s.is_signal
]
samples_non_w_ss = deepcopy(samples_non_w)
samples_signal = [s for s in samples if s.is_signal]
samples_w = deepcopy([s for s in samples if s.name in w_names])
samples_w_ss = deepcopy(samples_w)
# To calculate OS/SS ratio in inclusive W selection
samples_w_incl_os = deepcopy(samples_w)
samples_w_incl_ss = deepcopy(samples_w)
# To calculate W scale factor
samples_w_highmt_os = deepcopy(samples_w)
# Build a high MT region: OS - non-W/QCD OS - (SS - non-W/QCD SS)
samples_non_w_highmt_os = deepcopy(samples_non_w)
samples_non_w_highmt_ss = deepcopy(samples_non_w)
for sample in samples_non_w_highmt_os:
if sample.name != 'data_obs':
# Subtract background from data
sample.scale = -1.
for sample in samples_non_w_highmt_ss:
if sample.name != 'data_obs':
sample.scale = -1.
for sample in samples_non_w_ss:
if sample.name != 'data_obs':
sample.scale = -1.
var_norm = VariableCfg(name='_norm_',
drawname='1.',
binning={
'nbinsx': 5,
'xmin': -1.5,
'xmax': 3.5
},
unit='',
xtitle='Normalisation')
hist_dict['stacknow_highmt_os'] = HistogramCfg(
name='HighMTOS',
var=var_norm,
cfgs=samples_non_w_highmt_os,
cut=None,
lumi=int_lumi,
weight=weight)
hist_dict['stacknow_highmt_ss'] = HistogramCfg(
name='HighMTSS',
var=var_norm,
cfgs=samples_non_w_highmt_ss,
cut=None,
lumi=int_lumi,
weight=weight)
hist_dict['wjets_incl_os'] = HistogramCfg(name='WInclOS',
var=var_norm,
cfgs=samples_w_incl_os,
cut=None,
lumi=int_lumi,
weight=weight)
hist_dict['wjets_incl_ss'] = HistogramCfg(name='WInclSS',
var=var_norm,
cfgs=samples_w_incl_ss,
cut=None,
lumi=int_lumi,
weight=weight)
hist_dict['wjets_highmt_os'] = HistogramCfg(name='WHighMTOS',
var=var_norm,
cfgs=samples_w_highmt_os,
cut=None,
lumi=int_lumi,
weight=weight)
hist_dict['wjets'] = HistogramCfg(name='W',
var=None,
cfgs=samples_w,
cut=None,
lumi=int_lumi,
weight=weight)
hist_dict['wjets_ss'] = HistogramCfg(name='WSS',
var=None,
cfgs=samples_w_ss,
cut=None,
lumi=int_lumi,
weight=weight)
hist_dict['qcd'] = HistogramCfg(name='QCD',
var=None,
cfgs=samples_non_w_ss +
[hist_dict['wjets_ss']],
cut=None,
total_scale=r_qcd_os_ss,
lumi=int_lumi,
weight=weight)
return samples_non_w + [hist_dict['wjets'], hist_dict['qcd']
] + samples_signal
if qcd_from_same_sign and not w_qcd_mssm_method:
samples_qcdfromss = [s for s in all_samples if s.name != 'QCD']
samples_ss = copy.deepcopy(samples_qcdfromss)
samples_ss = [s for s in samples_ss if not s.is_signal]
for sample in samples_ss:
if sample.name != 'data_obs':
# Subtract background from data
sample.scale = -1.
qcd = HistogramCfg(name='QCD',
var=None,
cfgs=samples_ss,
cut=inc_cut,
total_scale=r_qcd_os_ss,
lumi=int_lumi,
weight=total_weight)
samples_qcdfromss.append(qcd)
if w_qcd_mssm_method:
w_names = ['W', 'W1Jets', 'W2Jets', 'W3Jets', 'W4Jets']
samples_non_w = [
s for s in all_samples
if s.name != 'QCD' and s.name not in w_names and not s.is_signal
]
samples_non_w_ss = copy.deepcopy(samples_non_w)
samples_w = copy.deepcopy([s for s in all_samples if s.name in w_names])
# -> Can add cross sections for samples either explicitly, or from file, or from cfg
for sample in samples:
setSumWeights(sample)
if qcd_from_same_sign:
samples_ss = copy.deepcopy(samples)
samples = [s for s in samples if sample.name != 'QCD']
for sample in samples_ss:
if sample.name != 'Data':
# Subtract background from data
sample.scale = -1.
qcd = HistogramCfg(name='QCD', var=None, cfgs=samples_ss, cut=inc_cut, lumi=40.0)
samples.append(qcd)
# Taken from Variables.py, can get subset with e.g. getVars(['mt', 'mvis'])
variables = all_vars
cfg_example = HistogramCfg(name='example', var=None, cfgs=samples, cut=inc_cut, lumi=40.0)
for cut_name in cuts:
cfg_example.cut = cuts[cut_name]
if qcd_from_same_sign and 'OS' in cut_name:
qcd.cut = cuts[cut_name].replace('l1_charge != l2_charge', 'l1_charge == l2_charge')
for variable in variables:
'l1_pt', 'l2_pt', 'l1_gen_pdgId', 'l2_gen_pdgId', 'l1_reliso05_04',
'l1_reliso05', 'l2_byCombinedIsolationDeltaBetaCorrRaw3Hits'
])
variables += [v for v in all_vars if 'dxy' in v.name or 'dz' in v.name]
variables = getVars([
'_norm_', 'l2_pt', 'l2_eta', 'l2_mass', 'l2_decayMode', 'mvis', 'mt',
'delta_r_l1_l2', 'l2_gen_pdgId', 'l2_mt'
])
# variables = all_vars
for cut_name in cuts:
cfg_tight = HistogramCfg(name='tight',
var=None,
cfgs=samples,
cut=inc_cut,
lumi=int_lumi,
weight=total_weight +
' * ((l2_gen_match != 5) - (l2_gen_match == 5))')
cfg_loose = HistogramCfg(name='loose',
var=None,
cfgs=samples,
cut=inc_cut,
lumi=int_lumi,
weight=total_weight +
' * ((l2_gen_match != 5) - (l2_gen_match == 5))')
cfg_data_tight = HistogramCfg(
name='tight_data',
var=None,
cfgs=samples,
SampleCfg(name='WJetsToLNu', ana_dir=analysis_dir, tree_prod_name=tree_prod_name, xsec=WJetsToLNu.xSection, sumweights=WJetsToLNu.nGenEvents),
SampleCfg(name='TTJets', dir_name='TTLep_pow', ana_dir=analysis_dir, tree_prod_name=tree_prod_name, xsec=TT_pow.xSection, sumweights=TT_pow.nGenEvents),
SampleCfg(name='T_tWch', dir_name='T_tWch', ana_dir=analysis_dir, tree_prod_name=tree_prod_name, xsec=T_tWch.xSection, sumweights=T_tWch.nGenEvents),
SampleCfg(name='TBar_tWch', dir_name='TBar_tWch', ana_dir=analysis_dir, tree_prod_name=tree_prod_name, xsec=TBar_tWch.xSection, sumweights=TBar_tWch.nGenEvents),
SampleCfg(name='TToLeptons_tch', dir_name='TToLeptons_tch', ana_dir=analysis_dir, tree_prod_name=tree_prod_name, xsec=TToLeptons_tch.xSection, sumweights=TBar_tWch.nGenEvents),
SampleCfg(name='TToLeptons_sch', dir_name='TToLeptons_sch', ana_dir=analysis_dir, tree_prod_name=tree_prod_name, xsec=TToLeptons_sch.xSection, sumweights=TToLeptons_sch.nGenEvents),
SampleCfg(name='ZZ', dir_name='ZZp8', ana_dir=analysis_dir, tree_prod_name=tree_prod_name, xsec=ZZp8.xSection, sumweights=ZZp8.nGenEvents),
SampleCfg(name='WZ', dir_name='WZp8', ana_dir=analysis_dir, tree_prod_name=tree_prod_name, xsec=WZp8.xSection, sumweights=WZp8.nGenEvents),
SampleCfg(name='WW', dir_name='WWTo2L2Nu', ana_dir=analysis_dir, tree_prod_name=tree_prod_name, xsec=WWTo2L2Nu.xSection, sumweights=WWTo2L2Nu.nGenEvents),
SampleCfg(name='QCD', dir_name='QCD_Mu15', ana_dir=analysis_dir, tree_prod_name=tree_prod_name, xsec=QCD_Mu15.xSection),
SampleCfg(name='Data', dir_name='SingleMuon_Run2015B', ana_dir=data_dir, tree_prod_name=tree_prod_name, is_data=True),
]
# -> Can add cross sections for samples either explicitly, or from file, or from cfg
for sample in samples:
setSumWeights(sample)
# Taken from Variables.py, can get subset with e.g. getVars(['mt', 'mvis'])
variables = all_vars
cfg_example = HistogramCfg(name='example', var=None, cfgs=samples, cut=example_cut, lumi=40.0)
for variable in variables:
cfg_example.var = variable
plot = createHistogram(cfg_example)
plot.Group('Diboson', ['ZZ', 'WZ', 'WW'])
plot.Group('Single t', ['T_tWch', 'TBar_tWch', 'TToLeptons_sch', 'TToLeptons_tch'])
HistDrawer.draw(plot)
# HistDrawer.drawRatio(plot)
# cuts.append(myCut('baseline_btd', 'abs(cand_refit_charge) == 1 & cand_refit_tau_mass > 1.6 & cand_refit_tau_mass < 2.0 '))
# cuts.append(myCut('tight' , 'tau_sv_prob > 0.10 & cand_refit_tau_dBetaIsoCone0p8strength0p2_rel < 0.2 & cand_refit_dRtauMuonMax < 0.2 & cand_refit_mttau > 40 & abs(cand_refit_charge)==1'))
# cuts.append(myCut('final' , 'tau_sv_prob > 0.10 & cand_refit_tau_dBetaIsoCone0p8strength0p2_rel < 0.2 & cand_refit_dRtauMuonMax < 0.2 & cand_refit_mttau > 40 & mu1_muonid_tight & mu2_muonid_tight & mu3_muonid_tight & abs(cand_refit_charge)==1 & cand_refit_tau_pt>10 & mu1_refit_reliso05<0.2 & mu2_refit_reliso05<0.2 & mu3_refit_reliso05<0.2'))
# cuts.append(myCut('bdt0p5' , 'bdt_score > 0.5'))
# cuts.append(myCut('bdt0p7' , 'bdt_score > 0.7'))
# cuts.append(myCut('bdt0p8' , 'bdt_score > 0.8'))
# cuts.append(myCut('bdt0p9' , 'bdt_score > 0.9'))
cuts.append(myCut('mauro_bdt', '1.'))
variables = generic_vars
sample_names = set()
for cut in cuts:
cfg_total = HistogramCfg(name=cut.name,
vars=variables,
cfgs=all_samples,
cut=str(cut.cut),
lumi=int_lumi,
weight=total_weight)
plots = createHistograms(
cfg_total,
verbose=True,
)
for variable in variables:
if variable.name == 'cand_refit_tau_mass' or variable.name == 'cand_tau_mass':
blindxmin = 1.777 - 0.040
blindxmax = 1.777 + 0.040
else:
blindxmin = None
blindxmax = None
xtitle='PUPPI ave cone 0.3'),
]
VarSet = namedtuple('VariableSet', ['name', 'vars', 'cut_s', 'cut_b'])
var_sets = [
VarSet('tau_iso', vars_tau, '&& l2_gen_match == 5',
'&& l2_gen_match == 6'),
# VarSet('muon_iso', vars_mu, '&& (l2_gen_match == 2 || l2_gen_match == 4)', '&& l2_gen_match == 6')
]
# samples = [sampleDict['TTJets']]#, sampleDict['QCD']]
cfg_signal = HistogramCfg(name='signal',
var=None,
cfgs=samples,
cut=inc_cut,
lumi=int_lumi,
weight='1.')
cfg_bg = HistogramCfg(name='bg',
var=None,
cfgs=samples,
cut=inc_cut,
lumi=int_lumi,
weight='1.')
for var_set in var_sets:
print 'Variable set', var_set.name
cfg_signal.cut += var_set.cut_s
cfg_bg.cut += var_set.cut_b
if qcd_from_same_sign and not w_qcd_mssm_method:
samples_qcdfromss = [s for s in all_samples if s.name != 'QCD']
samples_ss = copy.deepcopy(samples_qcdfromss)
samples_ss = [s for s in samples_ss if not s.is_signal]
for sample in samples_ss:
if sample.name != 'data_obs':
# Subtract background from data
sample.scale = -1.
qcd = HistogramCfg(name='QCD',
var=None,
cfgs=samples_ss,
cut=inc_cut,
total_scale=r_qcd_os_ss,
lumi=int_lumi,
weight=total_weight)
samples_qcdfromss.append(qcd)
if w_qcd_mssm_method:
w_names = ['W', 'W1Jets', 'W2Jets', 'W3Jets', 'W4Jets']
samples_non_w = [
s for s in all_samples
if s.name != 'QCD' and s.name not in w_names and not s.is_signal
]
samples_non_w_ss = copy.deepcopy(samples_non_w)
samples_signal = [s for s in all_samples if s.is_signal]
xsec=sample.get(XSection, 1),
sumweights=sample.get(SumWeights, 1))
if sample[Name] == 'data_obs': config_qcd.scale = 1.
else: config_qcd.scale = -1.
# shift also background subtraction in QCD
if config_qcd.name in shift:
config_qcd.scale *= shift[config.name]
if 'QCD' in shift:
config_qcd.scale *= shift['QCD']
qcd.append(config_qcd)
## Discard non fake MC
if not sample[HistoDir] in non_fakes:
samples_tmp.append(config)
# Add QCD component
config_qcd_total = HistogramCfg(name='QCD',
var=variable,
cfgs=qcd,
lumi=int_lumi)
samples_tmp.append(config_qcd_total)
config = HistogramCfg(name='config',
var=variable,
cfgs=samples_tmp,
lumi=int_lumi)
histos = createHistograms(config)
for histo in histos.histos.values():
removeNegativeValues2D(histo)
if 'Stat' in shiftname:
## Fluctuate up/down each bin
for sign in [-1, 1]:
for bx in xrange(1, histos.histos['QCD'].GetNbinsX() + 1):
for by in xrange(1,
def makePlots(variables,
cuts,
total_weight,
all_samples,
samples,
friend_func,
mode='control',
dc_postfix='',
make_plots=True,
optimisation=False):
sample_names = set()
ams_dict = {}
from CMGTools.H2TauTau.proto.plotter.cut import Cut
# def_iso_cut = inc_sig_tau1_iso & inc_sig_tau2_iso
iso_cuts = {
# 'vvtight':(Cut('l1_byIsolationMVArun2v1DBoldDMwLT>5.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>5.5'), Cut('l1_byIsolationMVArun2v1DBoldDMwLT>3.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>3.5')),
'vtight_relax2nd': (Cut('l1_byIsolationMVArun2v1DBoldDMwLT>4.5')
& Cut('l2_byIsolationMVArun2v1DBoldDMwLT>4.5'),
Cut('l1_byIsolationMVArun2v1DBoldDMwLT>4.5')
& Cut('l2_byIsolationMVArun2v1DBoldDMwLT>2.5')),
'loose_not_vtight':
(Cut(
'l1_byIsolationMVArun2v1DBoldDMwLT>1.5 && l1_byIsolationMVArun2v1DBoldDMwLT<4.5'
) & Cut(
'l2_byIsolationMVArun2v1DBoldDMwLT>1.5&&l2_byIsolationMVArun2v1DBoldDMwLT<4.5'
),
Cut('l1_byIsolationMVArun2v1DBoldDMwLT<1.5 && l1_byIsolationMVArun2v1DBoldDMwLT>0.5'
) &
Cut('l2_byIsolationMVArun2v1DBoldDMwLT<1.5 && l2_byIsolationMVArun2v1DBoldDMwLT>0.5'
)),
'one_loose_other_vtight':
(Cut(
'(l1_byIsolationMVArun2v1DBoldDMwLT>4.5 && (l2_byIsolationMVArun2v1DBoldDMwLT>1.5&&l2_byIsolationMVArun2v1DBoldDMwLT<4.5)) || (l2_byIsolationMVArun2v1DBoldDMwLT>4.5 && (l1_byIsolationMVArun2v1DBoldDMwLT>1.5&&l1_byIsolationMVArun2v1DBoldDMwLT<4.5)) '
),
Cut('l1_byIsolationMVArun2v1DBoldDMwLT<1.5 && l1_byIsolationMVArun2v1DBoldDMwLT>0.5'
) &
Cut('l2_byIsolationMVArun2v1DBoldDMwLT<1.5 && l2_byIsolationMVArun2v1DBoldDMwLT>0.5'
)),
# 'vtight':(Cut('l1_byIsolationMVArun2v1DBoldDMwLT>4.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>4.5'), Cut('l1_byIsolationMVArun2v1DBoldDMwLT>2.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>2.5')),
# 'tight':(Cut('l1_byIsolationMVArun2v1DBoldDMwLT>3.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>3.5'), Cut('l1_byIsolationMVArun2v1DBoldDMwLT>3.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>1.5')),
# 'medium':(Cut('l1_byIsolationMVArun2v1DBoldDMwLT>2.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>2.5'), Cut('l1_byIsolationMVArun2v1DBoldDMwLT>0.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>0.5')),
# 'loose':(Cut('l1_byIsolationMVArun2v1DBoldDMwLT>1.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>1.5'), Cut('1')),
# 'vloose':(Cut('l1_byIsolationMVArun2v1DBoldDMwLT>0.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>0.5'), Cut('1')),
}
# iso_cuts = {
# 'l1_vvtight':(Cut('l1_byIsolationMVArun2v1DBoldDMwLT>5.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>4.5'), Cut('l1_byIsolationMVArun2v1DBoldDMwLT>5.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>3.5')),
# 'l1_vtight':(Cut('l1_byIsolationMVArun2v1DBoldDMwLT>4.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>4.5'), Cut('l1_byIsolationMVArun2v1DBoldDMwLT>4.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>2.5')),
# 'l1_tight':(Cut('l1_byIsolationMVArun2v1DBoldDMwLT>3.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>4.5'), Cut('l1_byIsolationMVArun2v1DBoldDMwLT>3.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>1.5')),
# 'l1_medium':(Cut('l1_byIsolationMVArun2v1DBoldDMwLT>2.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>4.5'), Cut('l1_byIsolationMVArun2v1DBoldDMwLT>2.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>0.5')),
# 'l1_loose':(Cut('l1_byIsolationMVArun2v1DBoldDMwLT>1.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>4.5'), Cut('l1_byIsolationMVArun2v1DBoldDMwLT>1.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>0.5')),
# 'l1_vloose':(Cut('l1_byIsolationMVArun2v1DBoldDMwLT>0.5') & Cut('l2_byIsolationMVArun2v1DBoldDMwLT>4.5'),Cut('l1_byIsolationMVArun2v1DBoldDMwLT>0.5') & Cut('1')),
# }
for cut in cuts:
for iso_cut_name, (iso_cut, max_iso_cut) in iso_cuts.items():
# iso and charge cuts, need to have them explicitly for the QCD estimation
# max_iso_cut = Cut('l1_byIsolationMVArun2v1DBoldDMwLT > 2.5 && l2_byIsolationMVArun2v1DBoldDMwLT > 2.5')
iso_sideband_cut = (~iso_cut) & max_iso_cut
charge_cut = Cut('l1_charge != l2_charge')
isSS = 'SS' in cut.name
all_samples_qcd = qcd_estimation(
cut.cut & iso_sideband_cut &
(charge_cut if not isSS else ~charge_cut), # shape sideband
cut.cut & iso_cut & (~charge_cut), # norm sideband 1
cut.cut & iso_sideband_cut & (~charge_cut), # norm sideband 2
all_samples if mode in ['mssm'] else samples,
int_lumi,
total_weight,
verbose=verbose,
friend_func=friend_func)
# now include charge and isolation too
the_cut = MyCut(
cut.name + iso_cut_name,
cut.cut & iso_cut & (charge_cut if not isSS else ~charge_cut))
# for variable in variables:
cfg_total = HistogramCfg(name=the_cut.name,
vars=variables,
cfgs=all_samples_qcd,
cut=str(the_cut.cut),
lumi=int_lumi,
weight=total_weight)
# all_samples_qcd[-1].vars = variables
if mode == 'mva_train':
createTrees(cfg_total,
'/data1/steggema/tt/MVATrees',
verbose=True)
continue
plots = createHistograms(cfg_total,
verbose=True,
friend_func=friend_func)
for variable in variables:
plot = plots[variable.name]
plot.Group('Single t', [
'T_tWch', 'TBar_tWch', 'TToLeptons_tch_powheg',
'TBarToLeptons_tch_powheg'
]) # 'TToLeptons_sch',
plot.Group('VV', [
'VVTo2L2Nu', 'ZZTo2L2Q', 'WWTo1L1Nu2Q', 'WZTo1L3Nu',
'ZZTo4L', 'WZTo2L2Q', 'WZTo1L1Nu2Q', 'Single t'
]) # 'WZTo3L',
plot.Group(
'ZTT',
['ZTT', 'ZTT1Jets', 'ZTT2Jets', 'ZTT3Jets', 'ZTT4Jets'])
plot.Group('ZJ',
['ZJ', 'ZJ1Jets', 'ZJ2Jets', 'ZJ3Jets', 'ZJ4Jets'])
plot.Group('ZL',
['ZL', 'ZL1Jets', 'ZL2Jets', 'ZL3Jets', 'ZL4Jets'])
plot.Group(
'W',
['WJetsToLNu', 'W1Jets', 'W2Jets', 'W3Jets', 'W4Jets'])
plot.Group('Electroweak', ['W', 'VV', 'Single t', 'ZJ'])
if optimisation:
plot.DrawStack('HIST')
print plot
for signal_hist in plot.SignalHists():
sample_names.add(signal_hist.name)
ams = ams_hists_rebin(signal_hist.weighted,
plot.BGHist().weighted)
if variable.name == 'mt_total_mssm' and signal_hist.name == 'ggH1800':
print ams_hists_rebin(signal_hist.weighted,
plot.BGHist().weighted,
debug=True)
# import pdb; pdb.set_trace()
ams_dict[variable.name + '__' + the_cut.name + '__' +
signal_hist.name + '_'] = ams
if not make_plots:
continue
blindxmin = 0.7 if 'mva' in variable.name else None
blindxmax = 1.00001 if 'mva' in variable.name else None
if variable.name == 'mt2':
blindxmin = 60.
blindxmax = variable.binning['xmax']
if variable.name == 'mt_sum':
blindxmin = 250.
blindxmax = variable.binning['xmax']
if variable.name == 'mt_total':
blindxmin = 200.
blindxmax = variable.binning['xmax']
plot_dir = 'plot_' + the_cut.name
HistDrawer.draw(plot,
channel='#tau_{h}#tau_{h}',
plot_dir=plot_dir,
blindxmin=blindxmin,
blindxmax=blindxmax)
# HistDrawer.drawRatio(plot, channel='#tau_{h}#tau_{h}')
plot.UnGroup('Electroweak') #, ['W', 'VV', 'Single t', 'ZJ'])
plot.Group('VV', ['VV', 'Single t'])
if variable.name in [
'mt_total', 'svfit_mass', 'mt_total_mssm',
'mt_total_mssm_fine'
]:
plot.WriteDataCard(
filename=plot_dir +
'/htt_tt.inputs-sm-13TeV_{var}{postfix}.root'.format(
var=variable.name, postfix=dc_postfix),
dir='tt_' + cut.name,
mode='UPDATE')
# Save AMS dict
import pickle
pickle.dump(ams_dict, open('opt.pkl', 'wb'))
if optimisation:
print '\nOptimisation results:'
all_vals = ams_dict.items()
for sample_name in sample_names:
vals = [v for v in all_vals if sample_name + '_' in v[0]]
vals.sort(key=itemgetter(1))
for key, item in vals:
print item, key
print '\nBy variable'
for variable in variables:
name = variable.name
print '\nResults for variable', name
for key, item in vals:
if key.startswith(name + '__'):
print item, key
'nbinsx': 25,
'xmin': -2.5,
'xmax': 2.5
},
unit='',
xtitle='jet #eta'),
]
# variables = [
# VariableCfg(name='mvis', binning={'nbinsx':35, 'xmin':0, 'xmax':350}, unit='GeV', xtitle='m_{vis}')
# ]
for cut_name in cuts:
cfg_tight = HistogramCfg(
name='tight',
var=None,
cfgs=samples,
cut=inc_cut,
lumi=int_lumi,
weight=total_weight +
'*( (abs(tau1_gen_pdgId)!=15) - (abs(tau1_gen_pdgId)==15))')
cfg_loose = HistogramCfg(
name='loose',
var=None,
cfgs=samples,
cut=inc_cut,
lumi=int_lumi,
weight=total_weight +
'*( (abs(tau1_gen_pdgId)!=15) - (abs(tau1_gen_pdgId)==15))')
cfg_tight_data = HistogramCfg(
name='tight_data',
var=None,
cfgs=samples,
'l1_pt', 'l2_pt', 'l1_gen_pdgId', 'l2_gen_pdgId', 'l1_reliso05_04',
'l1_reliso05', 'l2_byCombinedIsolationDeltaBetaCorrRaw3Hits'
])
variables += [v for v in all_vars if 'dxy' in v.name or 'dz' in v.name]
variables = getVars([
'_norm_', 'l2_pt', 'l2_eta', 'l2_mass', 'l2_decayMode', 'mvis', 'mt',
'delta_r_l1_l2', 'l2_gen_pdgId', 'l2_mt'
])
# variables = all_vars
for cut_name in cuts:
cfg_tight = HistogramCfg(name='tight',
var=None,
cfgs=samples,
cut=inc_cut,
lumi=int_lumi,
weight=total_weight +
' * ((l2_gen_match != 5) - (l2_gen_match == 5))')
cfg_loose = HistogramCfg(name='loose',
var=None,
cfgs=samples,
cut=inc_cut,
lumi=int_lumi,
weight=total_weight +
' * ((l2_gen_match != 5) - (l2_gen_match == 5))')
cfg_data_tight = HistogramCfg(
name='tight_data',
var=None,
cfgs=samples,
def makePlots(variables,
cuts,
total_weight,
sample_dict,
hist_dict,
qcd_from_same_sign,
w_qcd_mssm_method,
mt_cut,
friend_func,
dc_postfix,
make_plots=True,
create_trees=False):
ams_dict = {}
sample_names = set()
for cut in cuts:
if qcd_from_same_sign and not 'SS' in cut.name and not w_qcd_mssm_method:
cfg_main = HistogramCfg(name=cut.name,
var=None,
cfgs=sample_dict['samples_qcdfromss'],
cut=cut.cut,
lumi=int_lumi,
weight=total_weight)
elif w_qcd_mssm_method:
cfg_main = HistogramCfg(name=cut.name,
var=None,
cfgs=sample_dict['samples_mssm_method'],
cut=cut.cut,
lumi=int_lumi,
weight=total_weight)
hist_dict[
'wjets'].cut = cut.cut # since wjets is a sub-HistogramCfg
else:
cfg_main = HistogramCfg(name=cut.name,
var=None,
cfgs=sample_dict['all_samples'],
cut=cut.cut,
lumi=int_lumi,
weight=total_weight)
if qcd_from_same_sign and not 'SS' in cut.name:
hist_dict['qcd'].cut = cut.cut.replace('l1_charge != l2_charge',
'l1_charge == l2_charge')
if w_qcd_mssm_method:
estimateQCDWMSSM(hist_dict,
cut,
mt_cut,
friend_func=friend_func,
r_qcd_os_ss=r_qcd_os_ss)
cfg_main.vars = variables
if qcd_from_same_sign:
hist_dict[
'qcd'].vars = variables # Can put into function but we will not want it by default if we take normalisations from e.g. high MT
if w_qcd_mssm_method:
hist_dict[
'wjets'].vars = variables # Can put into function but we will not want it by default if we take normalisations from e.g. high MT
hist_dict['qcd'].vars = variables
hist_dict['wjets_ss'].vars = variables
for variable in variables:
if variable.name in [
'svfit_mass', 'svfit_transverse_mass', 'mvis'
] and 'mssm' in mode:
if cut.name in ['inclusive', 'nobtag']:
variable.binning = binning_mssm
elif cut.name in ['btag']:
variable.binning = binning_mssm_btag
if create_trees:
createTrees(cfg_main, '/data1/steggema/mt/MVATrees', verbose=True)
continue
plots = createHistograms(cfg_main,
verbose=False,
friend_func=friend_func)
for variable in variables:
# for plot in plots.itervalues():
plot = plots[variable.name]
createDefaultGroups(plot)
if not w_qcd_mssm_method:
plot.Group('W', ['W', 'W1Jets', 'W2Jets', 'W3Jets', 'W4Jets'])
plot.Group('Electroweak', ['VV', 'W'])
# plot.Group('Single t', ['T_tWch', 'TBar_tWch', 'TToLeptons_sch', 'TToLeptons_tch'])
plot.Group('ZTT', ['ZTT', 'ZJ'], style=plot.Hist('ZTT').style)
if make_plots:
HistDrawer.draw(plot, plot_dir='plots/' + cut.name)
if variable.name in [
'mvis', 'svfit_transverse_mass', 'svfit_mass', 'mva',
'mva2div1', 'mva1', 'mva2', 'l2_nc_ratio'
]:
plot.WriteDataCard(
filename='datacard_{mode}_{var}.root'.format(
mode=mode, var=variable.name),
dir='mt_' + cut.name,
mode='UPDATE',
postfix=dc_postfix) #mt = mu-tau
for signal_hist in plot.SignalHists():
sample_names.add(signal_hist.name)
ams_dict[variable.name + '__' + cut.name + '__' +
signal_hist.name + '_'] = ams_hists(
signal_hist.weighted,
plot.BGHist().weighted)
print '\nOptimisation results:'
all_vals = ams_dict.items()
for sample_name in sample_names:
vals = [v for v in all_vals if sample_name + '_' in v[0]]
vals.sort(key=itemgetter(1))
for key, item in vals:
print item, key
print '\nBy variable'
for variable in variables:
name = variable.name
print '\nResults for variable', name
for key, item in vals:
if key.startswith(name + '__'):
print item, key
samples_mc, samples_data, samples, all_samples, sampleDict = createSampleLists(analysis_dir, channel='em', ztt_cut='(l1_gen_match>2 && l2_gen_match>3)', zl_cut='l2_gen_match==99',zj_cut='l2_gen_match==99', data2016G=data2016G)
if qcd_from_same_sign:
samples_qcdfromss = [s for s in samples if s.name != 'QCD']
samples_ss = copy.deepcopy(samples_qcdfromss)
scale = 2.0
for sample in samples_ss:
sample.scale = scale
if sample.name != 'data_obs':
# Subtract background from data
sample.scale = -scale
qcd = HistogramCfg(name='QCD', var=None, cfgs=samples_ss, cut=inc_cut, lumi=lumi)
samples_qcdfromss.append(qcd)
# Taken from Variables.py, can get subset with e.g. getVars(['mt', 'mvis'])
variables = emu_vars
# variables = getVars(['n_vertices'])
# variables = getVars(['l1_reliso05', 'l2_reliso05'])
# variables = [
# VariableCfg(name='mvis', binning={'nbinsx':35, 'xmin':0, 'xmax':350}, unit='GeV', xtitle='m_{vis}')
# ]
for cut_name in cuts:
if qcd_from_same_sign and not 'SS' in cut_name :
cfg_example = HistogramCfg(name='example', var=None, cfgs=samples_qcdfromss, cut=inc_cut, lumi=lumi, weight=total_weight)
else:
def createSamples(mode, analysis_dir, total_weight, qcd_from_same_sign,
w_qcd_mssm_method, r_qcd_os_ss):
hist_dict = {}
sample_dict = {}
samples_mc, samples_data, samples, all_samples, sampleDict = createSampleLists(
analysis_dir=analysis_dir)
if mode == 'mssm_control' or not 'mssm' in mode:
all_samples = [
s for s in all_samples
if not 'ggH' in s.name and not 'bbH' in s.name
]
sample_dict['all_samples'] = all_samples
if qcd_from_same_sign and not w_qcd_mssm_method:
samples_qcdfromss = [s for s in all_samples if s.name != 'QCD']
samples_ss = copy.deepcopy(samples_qcdfromss)
samples_ss = [s for s in samples_ss if not s.is_signal]
for sample in samples_ss:
if sample.name != 'data_obs':
# Subtract background from data
sample.scale = -1.
qcd = HistogramCfg(name='QCD',
var=None,
cfgs=samples_ss,
cut=None,
total_scale=r_qcd_os_ss,
lumi=int_lumi,
weight=total_weight)
samples_qcdfromss.append(qcd)
sample_dict['samples_qcdfromss'] = samples_qcdfromss
if w_qcd_mssm_method:
w_names = ['W', 'W1Jets', 'W2Jets', 'W3Jets', 'W4Jets']
samples_non_w = [
s for s in all_samples
if s.name != 'QCD' and s.name not in w_names and not s.is_signal
]
samples_non_w_ss = copy.deepcopy(samples_non_w)
samples_signal = [s for s in all_samples if s.is_signal]
samples_w = copy.deepcopy(
[s for s in all_samples if s.name in w_names])
samples_w_ss = copy.deepcopy(samples_w)
# To calculate OS/SS ratio in inclusive W selection
samples_w_incl_os = copy.deepcopy(samples_w)
samples_w_incl_ss = copy.deepcopy(samples_w)
# To calculate W scale factor
samples_w_highmt_os = copy.deepcopy(samples_w)
# Build a high MT region: OS - non-W/QCD OS - (SS - non-W/QCD SS)
samples_non_w_highmt_os = copy.deepcopy(samples_non_w)
samples_non_w_highmt_ss = copy.deepcopy(samples_non_w)
for sample in samples_non_w_highmt_os:
if sample.name != 'data_obs':
# Subtract background from data
sample.scale = -1.
for sample in samples_non_w_highmt_ss:
if sample.name != 'data_obs':
sample.scale = -1.
for sample in samples_non_w_ss:
if sample.name != 'data_obs':
sample.scale = -1.
var_norm = VariableCfg(name='_norm_',
drawname='1.',
binning={
'nbinsx': 5,
'xmin': -1.5,
'xmax': 3.5
},
unit='',
xtitle='Normalisation')
hist_dict['stacknow_highmt_os'] = HistogramCfg(
name='HighMTOS',
var=var_norm,
cfgs=samples_non_w_highmt_os,
cut=None,
lumi=int_lumi,
weight=total_weight)
hist_dict['stacknow_highmt_ss'] = HistogramCfg(
name='HighMTSS',
var=var_norm,
cfgs=samples_non_w_highmt_ss,
cut=None,
lumi=int_lumi,
weight=total_weight)
hist_dict['wjets_incl_os'] = HistogramCfg(name='WInclOS',
var=var_norm,
cfgs=samples_w_incl_os,
cut=None,
lumi=int_lumi,
weight=total_weight)
hist_dict['wjets_incl_ss'] = HistogramCfg(name='WInclSS',
var=var_norm,
cfgs=samples_w_incl_ss,
cut=None,
lumi=int_lumi,
weight=total_weight)
hist_dict['wjets_highmt_os'] = HistogramCfg(name='WHighMTOS',
var=var_norm,
cfgs=samples_w_highmt_os,
cut=None,
lumi=int_lumi,
weight=total_weight)
hist_dict['wjets'] = HistogramCfg(name='W',
var=None,
cfgs=samples_w,
cut=None,
lumi=int_lumi,
weight=total_weight)
hist_dict['wjets_ss'] = HistogramCfg(name='WSS',
var=None,
cfgs=samples_w_ss,
cut=None,
lumi=int_lumi,
weight=total_weight)
hist_dict['qcd'] = HistogramCfg(name='QCD',
var=None,
cfgs=samples_non_w_ss +
[hist_dict['wjets_ss']],
cut=None,
total_scale=r_qcd_os_ss,
lumi=int_lumi,
weight=total_weight)
sample_dict['samples_mssm_method'] = samples_non_w + [
hist_dict['wjets'], hist_dict['qcd']
] + samples_signal
return sample_dict, hist_dict
samples_mc, samples_data, samples, all_samples, sampleDict = createSampleLists(
analysis_dir, channel='WV', weight=weight_MC, reweightVJets=False)
# Taken from Variables.py, can get subset with e.g. getVars(['mt', 'mvis'])
variables = generic_vars + lnujj_vars + lnujj_vbf_vars
# variables = [lnujj_vars[0]]
# variables = getVars(['l1_reliso05', 'l2_reliso05'])
# variables = [
# VariableCfg(name='mvis', binning={'nbinsx':35, 'xmin':0, 'xmax':350}, unit='GeV', xtitle='m_{vis}')
# ]
for cut_name in cuts:
cfg_example = HistogramCfg(name='example',
var=None,
cfgs=samples,
cut='',
lumi=int_lumi,
weight=total_weight)
cfg_example.cut = cuts[cut_name]
cfg_example.vars = variables
channel = "l#nujj"
if cut_name.find("_mu") >= 0:
channel = "#mu#nujj"
elif cut_name.find("_e") >= 0:
channel = "e#nujj"
if cut_name.find("HP") >= 0:
channel += " HP"
elif cut_name.find("LP") >= 0:
channel += " LP"
'&& tau_decayModeFinding>0.5&& tau_byCombinedIsolationDeltaBetaCorr3Hits>1.5 && tau_againstMuon3>0.5 && tau_pt>20.',
# 'dbetamedium_antimuloose_antietight_20':'&& tau_decayModeFinding>0.5&& tau_byCombinedIsolationDeltaBetaCorr3Hits>1.5 && tau_againstMuon3>0.5 && tau_pt>20. && tau_againstElectronMVA6>2.5',
# 'dbetaloose_antimuloose':'&& tau_decayModeFinding>0.5&& tau_byCombinedIsolationDeltaBetaCorr3Hits>0.5 && tau_againstMuon3>0.5',
# 'mvaolddmvloose_antimuloose':'&& tau_decayModeFinding>0.5&& tau_byIsolationMVA3oldDMwLT>0.5 && tau_againstMuon3>0.5',
# 'mvaolddmloose_antimuloose':'&& tau_decayModeFinding>0.5&& tau_byIsolationMVA3oldDMwLT>1.5 && tau_againstMuon3>0.5',
# 'mvaolddmmedium_antimuloose':'&& tau_decayModeFinding>0.5&& tau_byIsolationMVA3oldDMwLT>2.5 && tau_againstMuon3>0.5',
# 'mvaolddmtight_antimuloose':'&& tau_decayModeFinding>0.5&& tau_byIsolationMVA3oldDMwLT>3.5 && tau_againstMuon3>0.5',
# 'mvaolddmvtight_antimuloose':'&& tau_decayModeFinding>0.5&& tau_byIsolationMVA3oldDMwLT>4.5 && tau_againstMuon3>0.5',
}
for cut_name in cuts:
for signal_cut_name, signal_cut in signal_cuts.items():
cfg_tight = HistogramCfg(
name='tight' + cut_name + signal_cut_name,
var=None,
cfgs=samples_mc,
cut=inc_cut,
lumi=int_lumi,
weight=total_weight +
' * ((tau_gen_match != 5) - (tau_gen_match == 5))')
cfg_loose = HistogramCfg(
name='loose' + cut_name + signal_cut_name,
var=None,
cfgs=samples_mc,
cut=inc_cut,
lumi=int_lumi,
weight=total_weight +
' * ((tau_gen_match != 5) - (tau_gen_match == 5))')
cfg_data_tight = HistogramCfg(
name='tight_data' + cut_name + signal_cut_name,
var=None,
