def data_mc_ratio(ratio_groups, var, cut, **kwargs):
ColorStyleGen.reset()
samples_mc = ratio_groups.keys()
ratio_groups["data"] = ["SingleMuAB", "SingleMuC", "SingleMuD"]
plot_name = kwargs.get("name", "plot")
samples_d = dict()
for name, group in ratio_groups.items():
samples_d[name] = [
metadata_iso.get_histogram(sample_name,
var,
cut_str=cut.cut_str,
weight=None)
for sample_name in group
]
if name != "data":
for hist in samples_d[name]:
hist.normalize_lumi(lumi_total)
merged = merge_hists(samples_d[name], name)
samples_d[name] = merged
samples_d[name].pretty_name = name
for name in samples_d.keys():
if name != "data":
samples_d[name].hist.Divide(samples_d["data"].hist)
samples_d["data"].hist.Divide(samples_d["data"].hist)
canv = plot_hists(samples_d.values(),
styles={"data": Styling.style["data"]},
**kwargs)
leg = legend([r for r in reversed(samples_d.values())],
styles=["p", "f"])
canv.SaveAs(plot_name + ".pdf")
return canv, leg
def mc_amount(cut, weight, lumi, ref=None):
histsD = dict()
for samp in samples_mc:
histsD[samp.name] = samp.drawHistogram("eta_lj",
str(cut),
weight=weight,
plot_range=[100, -5, 5])
for name, hist in histsD.items():
hist.normalize_lumi(lumi)
for name, hist in histsD.items():
histsD[name] = hist.hist
merge_cmd = copy.deepcopy(merge_cmds)
merge_cmd["QCD (MC)"] = ["QCDMu"]
merge_cmd["t#bar{t} incl."] = ["TTJets_MassiveBinDECAY"]
merge_cmd.pop("data")
merged_hists = merge_hists(histsD, merge_cmd)
interesting = [
"t#bar{t}", "t#bar{t} incl.", "W(#rightarrow l #nu) + jets",
"QCD (MC)", "t-channel"
]
for i in interesting:
r = 0.0
r0 = 0.0
if ref and i in ref.keys():
r0 = ref[i]
r = 100.0 * merged_hists[i].Integral() / float(r0)
print "%s | %d | %d | %.2f %%" % (i, merged_hists[i].Integral(),
r0, r)
def data_mc_ratio(ratio_groups, var, cut, **kwargs):
ColorStyleGen.reset()
samples_mc = ratio_groups.keys()
ratio_groups["data"] = ["SingleMuAB", "SingleMuC", "SingleMuD"]
plot_name = kwargs.get("name", "plot")
samples_d = dict()
for name, group in ratio_groups.items():
samples_d[name] = [
metadata_iso.get_histogram(sample_name, var, cut_str=cut.cut_str, weight=None) for
sample_name in group
]
if name != "data":
for hist in samples_d[name]:
hist.normalize_lumi(lumi_total)
merged = merge_hists(samples_d[name], name)
samples_d[name] = merged
samples_d[name].pretty_name = name
for name in samples_d.keys():
if name != "data":
samples_d[name].hist.Divide(samples_d["data"].hist)
samples_d["data"].hist.Divide(samples_d["data"].hist)
canv = plot_hists(
samples_d.values(),
styles={"data": Styling.style["data"]},
**kwargs
)
leg = legend([r for r in reversed(samples_d.values())], styles=["p", "f"])
canv.SaveAs(plot_name + ".pdf")
return canv, leg
def plot_hists_stacked(hist_groups, **kwargs):
import uuid
stacks = dict()
styles = kwargs.get("styles")
draw_styles = kwargs.get("draw_styles", {})
do_log_y = kwargs.get("do_log_y", False)
title = kwargs.get("title", "TITLE")
x_label = kwargs.get("x_label", "XLABEL")
name = kwargs.get("name", "NAME_%s" % uuid.uuid1())
min_bin = kwargs.get("min_bin", 10)
canv = ROOT.TCanvas(name, name)
canv.SetWindowSize(1000, 1000)
canv.SetCanvasSize(1000, 1000)
for name, group in hist_groups.items():
stacks[name] = ROOT.THStack(name, name)
print name
for hist in group:
print hist.name
if not styles:
(color, style) = ColorStyleGen.next()
hist.hist.SetLineColor(color)
hist.hist.SetLineWidth(2)
hist.hist.SetFillColor(color)
hist.hist.SetFillStyle(style)
else:
styles[name](hist)
stacks[name].Add(hist.hist)
max_bin = get_max_bin(stacks.values())
for name, stack in stacks.items():
stack.Draw()
canv.Draw()
first = True
for name, stack in stacks.items():
if name in draw_styles.keys():
drawcmd = draw_styles[name]
else:
drawcmd = "BAR HIST goff"
if not first:
drawcmd += " SAME"
stack.Draw(drawcmd)
if first:
stack.SetMaximum(1.8 * max_bin)
if do_log_y:
stack.SetMinimum(min_bin)
stack.SetTitle(title)
stack.GetXaxis().SetTitle(x_label)
first = False
if do_log_y:
canv.SetLogy()
return canv, stacks
def plot_hists_stacked(hist_groups, **kwargs):
import uuid
stacks = dict()
styles = kwargs.get("styles")
draw_styles = kwargs.get("draw_styles", {})
do_log_y = kwargs.get("do_log_y", False)
title = kwargs.get("title", "TITLE")
x_label = kwargs.get("x_label", "XLABEL")
name = kwargs.get("name", "NAME_%s" % uuid.uuid1())
min_bin = kwargs.get("min_bin", 10)
canv = ROOT.TCanvas(name, name)
canv.SetWindowSize(1000,1000)
canv.SetCanvasSize(1000,1000)
for name, group in hist_groups.items():
stacks[name] = ROOT.THStack(name, name)
print name
for hist in group:
print hist.name
if not styles:
(color, style) = ColorStyleGen.next()
hist.hist.SetLineColor(color)
hist.hist.SetLineWidth(2)
hist.hist.SetFillColor(color)
hist.hist.SetFillStyle(style)
else:
styles[name](hist)
stacks[name].Add(hist.hist)
max_bin = get_max_bin(stacks.values())
for name, stack in stacks.items():
stack.Draw()
canv.Draw()
first = True
for name, stack in stacks.items():
if name in draw_styles.keys():
drawcmd = draw_styles[name]
else:
drawcmd = "BAR HIST goff"
if not first:
drawcmd += " SAME"
stack.Draw(drawcmd)
if first:
stack.SetMaximum(1.8*max_bin)
if do_log_y:
stack.SetMinimum(min_bin)
stack.SetTitle(title)
stack.GetXaxis().SetTitle(x_label)
first = False
if do_log_y:
canv.SetLogy()
return canv, stacks
def merge_hists_g(hists_d, merge_groups=g_merge_cmd):
out_d = dict()
for merge_name, items in merge_groups.items():
hist = hists_d[items[0]].hist.Clone()
for item in items[1:]:
hist.Add(hists_d[item].hist)
out_d[merge_name] = Histogram()
out_d[merge_name].setHist(hist, sample_name=hists_d[items[0]].sample_name, var=hists_d[items[0]].var, cut=hists_d[items[0]].cut)
integral, err = out_d[merge_name].calc_int_err()
out_d[merge_name].pretty_name = "%s : %.0f #pm %0.f" % (merge_name, integral, err)
return out_d
def merge_hists_g(hists_d, merge_groups=g_merge_cmd):
out_d = dict()
for merge_name, items in merge_groups.items():
hist = hists_d[items[0]].hist.Clone()
for item in items[1:]:
hist.Add(hists_d[item].hist)
out_d[merge_name] = Histogram()
out_d[merge_name].setHist(hist,
sample_name=hists_d[items[0]].sample_name,
var=hists_d[items[0]].var,
cut=hists_d[items[0]].cut)
integral, err = out_d[merge_name].calc_int_err()
out_d[merge_name].pretty_name = "%s : %.0f #pm %0.f" % (merge_name,
integral, err)
return out_d
def mc_amount(cut, weight, lumi, ref=None):
histsD = dict()
for samp in samples_mc:
histsD[samp.name] = samp.drawHistogram("eta_lj", str(cut), weight=weight, plot_range=[100,-5,5])
for name, hist in histsD.items():
hist.normalize_lumi(lumi)
for name, hist in histsD.items():
histsD[name] = hist.hist
merge_cmd = copy.deepcopy(merge_cmds)
merge_cmd["QCD (MC)"] = ["QCDMu"]
merge_cmd["t#bar{t} incl."] = ["TTJets_MassiveBinDECAY"]
merge_cmd.pop("data")
merged_hists = merge_hists(histsD, merge_cmd)
interesting = ["t#bar{t}", "t#bar{t} incl.", "W(#rightarrow l #nu) + jets", "QCD (MC)", "t-channel"]
for i in interesting:
r = 0.0
r0 = 0.0
if ref and i in ref.keys():
r0 = ref[i]
r = 100.0*merged_hists[i].Integral() / float(r0)
print "%s | %d | %d | %.2f %%" % (i, merged_hists[i].Integral(), r0, r)
n_true.Scale(1.0/sum(xs_s))
n_tagged.Scale(1.0/sum(xs_s))
#print "int_true=%.2f, int_tagged=%.2f" % (n_true.Integral(), n_tagged.Integral())
div = n_tagged.Clone("eff_%s" % flavour)
div.Divide(n_true)
#print "div=%.2f" % div.Integral()
return div
if __name__=="__main__":
from glob import glob
from common.utils import get_sample_dict
import os
import sys
samples_d = dict()
samples_d["ttbar"] = ["TTJets_FullLept", "TTJets_SemiLept"]
samples_d["wjets"] = ["W1Jets_exclusive", "W2Jets_exclusive", "W3Jets_exclusive", "W4Jets_exclusive"]
samples_d["t-channel"] = ["T_t_ToLeptons", "Tbar_t_ToLeptons"]
sample_name = sys.argv[1]
out_dir = "b_eff_plots/"
import os
if not os.path.exists(out_dir):
os.mkdir(out_dir)
axis = sys.argv[2]
if axis == "x":
xlab = "p_{T,q}"
elif axis == "y":
xlab = "|#eta_{q}|"
else:
interesting = ["t#bar{t}", "t#bar{t} incl.", "W(#rightarrow l #nu) + jets", "QCD (MC)", "t-channel"]
for i in interesting:
r = 0.0
r0 = 0.0
if ref and i in ref.keys():
r0 = ref[i]
r = 100.0*merged_hists[i].Integral() / float(r0)
print "%s | %d | %d | %.2f %%" % (i, merged_hists[i].Integral(), r0, r)
weight = "pu_weight*muon_IDWeight*muon_IsoWeight"
doCutFlow = True
if doCutFlow:
print "1 iso mu, 0 veto mu/ele"
ref = dict()
ref["t#bar{t}"] = 261429
ref["t#bar{t} incl."] = ref["t#bar{t}"]
ref["W(#rightarrow l #nu) + jets"] = 47758738
ref["QCD (MC)"] = 1873957
ref["t-channel"] = 54769
previous = Cuts.hlt_isomu * Cuts.one_muon * Cuts.lepton_veto
mc_amount(previous, weight, 12210, ref=ref)
print
print "2 jets"
ref = dict()
ref["t#bar{t}"] = 78201
ref["t#bar{t} incl."] = ref["t#bar{t}"]
ref["W(#rightarrow l #nu) + jets"] = 921051
ref["QCD (MC)"] = 67463
def stack_plot(var, cut, weight=None, **kwargs):
qcd_weight = kwargs.get("qcd_weight", None)
skip_samples = kwargs.get("skip_samples", [])
doDataDrivenQCD = kwargs.get("doDataDrivenQCD", True)
hists_mc = [
metadata_iso.get_histogram(sample_name,
var,
cut_str=cut.cut_str,
weight=weight)
for sample_name in mc_sample_names
]
hists_qcd_iso = [
metadata_iso.get_histogram(sample_name,
var,
cut_str=cut.cut_str,
weight=qcd_weight)
for sample_name in qcd_sample_names
]
if doDataDrivenQCD:
hists_data_antiiso = [
metadata_antiiso.get_histogram(sample_name,
var,
cut_str=cut.cut_str,
weight=None)
for sample_name in data_sample_names
]
hists_d = dict()
for hist in hists_mc:
hists_d[hist.sample_name] = hist
if not doDataDrivenQCD:
for hist in hists_qcd_iso:
hists_d[hist.sample_name] = hist
# for hist in hists_mc:
# if hist.sample_name in mc_sample_titles.keys():
# hist.pretty_name = mc_sample_titles[hist.sample_name]
# else:
# hist.pretty_name = hist.sample_name
hists_data = [
metadata_iso.get_histogram(sample_name,
var,
cut_str=cut.cut_str,
weight=None)
for sample_name in data_sample_names
]
for hist in hists_data:
hists_d[hist.sample_name] = hist
for hist in hists_data:
hist.pretty_name = "SingleMu"
stack_group = dict()
stack_group["mc"] = hists_mc
#stack_group["data"] = hists_data
stack_group["data"] = [hists_data[0]]
for hist in stack_group["mc"]:
hist.normalize_lumi(lumi_total)
for hist in hists_qcd_iso:
hist.normalize_lumi(lumi_total)
merge_cmd = dict()
merge_cmd["data"] = ["SingleMuAB", "SingleMuC", "SingleMuD"]
merge_cmd["diboson"] = ["WW", "WZ", "ZZ"]
merge_cmd["DY-jets"] = ["DYJets"]
merge_cmd["s-channel"] = ["T_s", "Tbar_s"]
merge_cmd["tW-channel"] = ["T_tW", "Tbar_tW"]
merge_cmd["t#bar{t}"] = ["TTJets_FullLept", "TTJets_SemiLept"]
merge_cmd["W+jets"] = [
"W1Jets_exclusive", "W2Jets_exclusive", "W3Jets_exclusive",
"W4Jets_exclusive"
]
#merge_cmd["W+jets"] = ["WJets_inclusive"]
merge_cmd["t-channel"] = ["T_t", "Tbar_t"]
if not doDataDrivenQCD:
merge_cmd["QCD (MC)"] = ["QCDMu"]
for sample in skip_samples:
if sample in merge_cmd.keys():
merge_cmd.pop(sample)
merged_hists = merge_hists_g(hists_d, merge_cmd)
if doDataDrivenQCD:
hist_data_antiiso = merge_hists(hists_data_antiiso, "data_antiiso")
hist_qcd_iso = merge_hists(hists_qcd_iso, "qcd_iso")
hist_data_antiiso.normalize(hist_qcd_iso.hist.Integral())
print hist_data_antiiso.hist.Integral()
hist_data_antiiso.sample_name = "QCDMu"
hist_data_antiiso.pretty_name = "QCD (dd.)"
#for hist in hists_data[1:]:
# stack_group["data"][0].hist.Add(hist.hist)
stack_d = dict()
if doDataDrivenQCD:
stack_d["mc"] = [hist_data_antiiso]
else:
stack_d["mc"] = []
stack_d["mc"] += [
merged_hists[name] for name in merged_hists.keys()
if name != "data"
]
stack_d["data"] = [merged_hists["data"]]
canv, stacks = plot_hists_stacked(stack_d,
styles=Styling.style,
draw_styles={"data": "E1"},
**kwargs)
#canvas_margin(canv, side="R", margin=0.3)
# leg_hists = ([hists_data[0]] +
# [hists_d["T_t"]] +
# [hists_d["TTJets_FullLept"]] +
# [hists_d["W1Jets_exclusive"]] +
# [hists_d["T_tW"]] +
# [hists_d["T_s"]] +
# [hists_d["WW"]] +
# [hists_d["QCDMu"]])
stack_d["mc"].reverse()
leg = legend(stack_d["data"] + stack_d["mc"], styles=["p", "f"])
text = lumi_textbox(lumi=lumi_total)
canv.SaveAs(canv.GetName() + ".pdf")
def lumi_textbox(lumi=10.435, pos="top-center"):
if pos == "top-center":
coords = [0.25, 0.73, 0.71, 0.88]
text = ROOT.TPaveText(coords[0], coords[1], coords[2], coords[3], "NDC")
text.AddText("CMS preliminary #sqrt{s} = 8 TeV, #int L dt = %.1f fb^{-1}" %
(float(lumi) / 1000.0))
text.SetShadowColor(ROOT.kWhite)
text.SetLineColor(ROOT.kWhite)
text.SetFillColor(ROOT.kWhite)
text.Draw()
return text
g_merge_cmd = dict()
g_merge_cmd["data"] = ["SingleMuAB", "SingleMuC", "SingleMuD"]
g_merge_cmd["diboson"] = ["WW", "WZ", "ZZ"]
g_merge_cmd["DY-jets"] = ["DYJets"]
g_merge_cmd["s-channel"] = ["T_s", "Tbar_s"]
g_merge_cmd["tW-channel"] = ["T_tW", "Tbar_tW"]
g_merge_cmd["t#bar{t}"] = ["TTJets_FullLept", "TTJets_SemiLept"]
g_merge_cmd["W+jets"] = [
"W1Jets_exclusive", "W2Jets_exclusive", "W3Jets_exclusive",
"W4Jets_exclusive"
]
#g_merge_cmd["W+jets"] = ["WJets_inclusive"]
g_merge_cmd["t-channel"] = ["T_t", "Tbar_t"]
def merge_hists_g(hists_d, merge_groups=g_merge_cmd):
def stack_plot(var, cut, weight=None, **kwargs):
qcd_weight = kwargs.get("qcd_weight", None)
skip_samples = kwargs.get("skip_samples", [])
doDataDrivenQCD = kwargs.get("doDataDrivenQCD", True)
hists_mc = [metadata_iso.get_histogram(sample_name, var, cut_str=cut.cut_str, weight=weight) for sample_name in mc_sample_names]
hists_qcd_iso = [metadata_iso.get_histogram(sample_name, var, cut_str=cut.cut_str, weight=qcd_weight) for sample_name in qcd_sample_names]
if doDataDrivenQCD:
hists_data_antiiso = [metadata_antiiso.get_histogram(sample_name, var, cut_str=cut.cut_str, weight=None) for sample_name in data_sample_names]
hists_d = dict()
for hist in hists_mc:
hists_d[hist.sample_name] = hist
if not doDataDrivenQCD:
for hist in hists_qcd_iso:
hists_d[hist.sample_name] = hist
# for hist in hists_mc:
# if hist.sample_name in mc_sample_titles.keys():
# hist.pretty_name = mc_sample_titles[hist.sample_name]
# else:
# hist.pretty_name = hist.sample_name
hists_data = [metadata_iso.get_histogram(sample_name, var, cut_str=cut.cut_str, weight=None) for sample_name in data_sample_names]
for hist in hists_data:
hists_d[hist.sample_name] = hist
for hist in hists_data:
hist.pretty_name = "SingleMu"
stack_group = dict()
stack_group["mc"] = hists_mc
#stack_group["data"] = hists_data
stack_group["data"] = [hists_data[0]]
for hist in stack_group["mc"]:
hist.normalize_lumi(lumi_total)
for hist in hists_qcd_iso:
hist.normalize_lumi(lumi_total)
merge_cmd = dict()
merge_cmd["data"] = ["SingleMuAB", "SingleMuC", "SingleMuD"]
merge_cmd["diboson"] = ["WW", "WZ", "ZZ"]
merge_cmd["DY-jets"] = ["DYJets"]
merge_cmd["s-channel"] = ["T_s", "Tbar_s"]
merge_cmd["tW-channel"] = ["T_tW", "Tbar_tW"]
merge_cmd["t#bar{t}"] = ["TTJets_FullLept", "TTJets_SemiLept"]
merge_cmd["W+jets"] = ["W1Jets_exclusive", "W2Jets_exclusive", "W3Jets_exclusive", "W4Jets_exclusive"]
#merge_cmd["W+jets"] = ["WJets_inclusive"]
merge_cmd["t-channel"] = ["T_t", "Tbar_t"]
if not doDataDrivenQCD:
merge_cmd["QCD (MC)"] = ["QCDMu"]
for sample in skip_samples:
if sample in merge_cmd.keys():
merge_cmd.pop(sample)
merged_hists = merge_hists_g(hists_d, merge_cmd)
if doDataDrivenQCD:
hist_data_antiiso = merge_hists(hists_data_antiiso, "data_antiiso")
hist_qcd_iso = merge_hists(hists_qcd_iso, "qcd_iso")
hist_data_antiiso.normalize(hist_qcd_iso.hist.Integral())
print hist_data_antiiso.hist.Integral()
hist_data_antiiso.sample_name = "QCDMu"
hist_data_antiiso.pretty_name = "QCD (dd.)"
#for hist in hists_data[1:]:
# stack_group["data"][0].hist.Add(hist.hist)
stack_d = dict()
if doDataDrivenQCD:
stack_d["mc"] = [hist_data_antiiso]
else:
stack_d["mc"] = []
stack_d["mc"] += [merged_hists[name] for name in merged_hists.keys() if name!="data"]
stack_d["data"] = [merged_hists["data"]]
canv, stacks = plot_hists_stacked(
stack_d,
styles=Styling.style, draw_styles={"data": "E1"}, **kwargs
)
#canvas_margin(canv, side="R", margin=0.3)
# leg_hists = ([hists_data[0]] +
# [hists_d["T_t"]] +
# [hists_d["TTJets_FullLept"]] +
# [hists_d["W1Jets_exclusive"]] +
# [hists_d["T_tW"]] +
# [hists_d["T_s"]] +
# [hists_d["WW"]] +
# [hists_d["QCDMu"]])
stack_d["mc"].reverse()
leg = legend(stack_d["data"] + stack_d["mc"], styles=["p", "f"])
text = lumi_textbox(lumi=lumi_total)
canv.SaveAs(canv.GetName() + ".pdf")
if do_log_y:
canv.SetLogy()
return canv, stacks
def lumi_textbox(lumi=10.435, pos="top-center"):
if pos=="top-center":
coords = [0.25, 0.73, 0.71, 0.88]
text = ROOT.TPaveText(coords[0], coords[1], coords[2], coords[3], "NDC")
text.AddText("CMS preliminary #sqrt{s} = 8 TeV, #int L dt = %.1f fb^{-1}" % (float(lumi)/1000.0))
text.SetShadowColor(ROOT.kWhite)
text.SetLineColor(ROOT.kWhite)
text.SetFillColor(ROOT.kWhite)
text.Draw()
return text
g_merge_cmd = dict()
g_merge_cmd["data"] = ["SingleMuAB", "SingleMuC", "SingleMuD"]
g_merge_cmd["diboson"] = ["WW", "WZ", "ZZ"]
g_merge_cmd["DY-jets"] = ["DYJets"]
g_merge_cmd["s-channel"] = ["T_s", "Tbar_s"]
g_merge_cmd["tW-channel"] = ["T_tW", "Tbar_tW"]
g_merge_cmd["t#bar{t}"] = ["TTJets_FullLept", "TTJets_SemiLept"]
g_merge_cmd["W+jets"] = ["W1Jets_exclusive", "W2Jets_exclusive", "W3Jets_exclusive", "W4Jets_exclusive"]
#g_merge_cmd["W+jets"] = ["WJets_inclusive"]
g_merge_cmd["t-channel"] = ["T_t", "Tbar_t"]
def merge_hists_g(hists_d, merge_groups=g_merge_cmd):
out_d = dict()
for merge_name, items in merge_groups.items():
hist = hists_d[items[0]].hist.Clone()
]
for i in interesting:
r = 0.0
r0 = 0.0
if ref and i in ref.keys():
r0 = ref[i]
r = 100.0 * merged_hists[i].Integral() / float(r0)
print "%s | %d | %d | %.2f %%" % (i, merged_hists[i].Integral(),
r0, r)
weight = "pu_weight*muon_IDWeight*muon_IsoWeight"
doCutFlow = True
if doCutFlow:
print "1 iso mu, 0 veto mu/ele"
ref = dict()
ref["t#bar{t}"] = 261429
ref["t#bar{t} incl."] = ref["t#bar{t}"]
ref["W(#rightarrow l #nu) + jets"] = 47758738
ref["QCD (MC)"] = 1873957
ref["t-channel"] = 54769
previous = Cuts.hlt_isomu * Cuts.one_muon * Cuts.lepton_veto
mc_amount(previous, weight, 12210, ref=ref)
print
print "2 jets"
ref = dict()
ref["t#bar{t}"] = 78201
ref["t#bar{t} incl."] = ref["t#bar{t}"]
ref["W(#rightarrow l #nu) + jets"] = 921051
ref["QCD (MC)"] = 67463
