def Save(self):
# Save
#if not os.path.isdir(self.GetOutPath()):
# print 'Creating directory: ', self.GetOutPath()
gSystem.mkdir(self.GetOutPath(), True)
self.canvas.Print(self.GetOutName() + '.pdf', 'pdf')
self.canvas.Print(self.GetOutName() + '.png', 'png')
def parse_drawer_options(options):
# Create outdir if necessary
if not options.outdir.endswith("/"):
options.outdir += "/"
if gSystem.AccessPathName(options.outdir):
gSystem.mkdir(options.outdir)
# Make input file list
if not options.infile.endswith(".root") and not options.infile.endswith(
".txt"):
raise ValueError("infile must be .root file or .txt file")
if options.infile.endswith(".root"):
# Create a temporary file with one line
with tempfile.NamedTemporaryFile() as infile:
infile.write(options.infile)
infile.flush()
options.tfilecoll = TFileCollection("fc", "", infile.name)
else:
options.tfilecoll = TFileCollection("fc", "", options.infile)
# Batch mode
if options.batch:
gROOT.SetBatch(True)
# Trigger tower parameter space
if options.tower != 99:
ieta = options.tower / 8
iphi = options.tower % 8
options.etamin = -2.2 + (4.4 / 6) * ieta
options.etamax = -2.2 + (4.4 / 6) * (ieta + 1)
if iphi < 4:
options.phimin = (2 * pi / 8) * iphi
options.phimax = (2 * pi / 8) * (iphi + 1)
else:
options.phimin = -2 * pi - pi + (2 * pi / 8) * iphi
options.phimax = -2 * pi - pi + (2 * pi / 8) * (iphi + 1)
options.ptmin = 2.
options.ptmax = 2000.
options.vzmin = -15.
options.vzmax = +15.
def DrawPlot(fname,
hname,
outname,
outpath="./SFplots/",
inpath='InputFiles/',
doErrors=True):
if not outpath.endswith('/'): outpath += '/'
if not inpath.endswith('/'): inpath += '/'
if not fname.endswith('.root'): fname += '.root'
f = TFile.Open(inpath + fname)
h = f.Get(hname)
c = TCanvas("c", "c", 10, 10, 1600, 1200)
gStyle.SetOptStat(0)
gStyle.SetPalette(1)
gStyle.SetPaintTextFormat("1.2f")
if doErrors: h.Draw("colz, text, errors")
else: h.Draw("colz, text")
if not os.path.isdir(outpath): gSystem.mkdir(outpath, 1)
c.Print(outpath + outname + '.pdf', 'pdf')
c.Print(outpath + outname + '.png', 'png')
def parse_drawer_options(options):
# Create outdir if necessary
if not options.outdir.endswith("/"):
options.outdir += "/"
if gSystem.AccessPathName(options.outdir):
gSystem.mkdir(options.outdir)
# Make input file list
if not options.infile.endswith(".root") and not options.infile.endswith(".txt"):
raise ValueError("infile must be .root file or .txt file")
if options.infile.endswith(".root"):
# Create a temporary file with one line
with tempfile.NamedTemporaryFile() as infile:
infile.write(options.infile)
infile.flush()
options.tfilecoll = TFileCollection("fc", "", infile.name)
else:
options.tfilecoll = TFileCollection("fc", "", options.infile)
# Batch mode
if options.batch:
gROOT.SetBatch(True)
# Trigger tower parameter space
if options.tower != 99:
ieta = options.tower/8
iphi = options.tower%8
options.etamin = -2.2 + (4.4/6) * ieta
options.etamax = -2.2 + (4.4/6) * (ieta+1)
if iphi < 6:
options.phimin = -pi/2 + (2*pi/8) * iphi
options.phimax = -pi/2 + (2*pi/8) * (iphi+1)
else:
options.phimin = -2*pi -pi/2 + (2*pi/8) * iphi
options.phimax = -2*pi -pi/2 + (2*pi/8) * (iphi+1)
options.ptmin = 2.
options.ptmax = 2000.
channel = "ZnunuHighPt"
region = "TT"
massH = 125
plotData = True
plotLog = False
plotSig = True
plotdir = "plots_etc/"
kRed = 632
kYellow = 400
kBlue = 600
TH1.SetDefaultSumw2(1)
gROOT.SetBatch(1)
if gSystem.AccessPathName(plotdir):
gSystem.mkdir(plotdir)
class EventsJ11(object):
pass
def Read(whatstep="Step4", cutallmc="", cutalldata=""):
if whatstep == "Step4":
indir = "/uscms_data/d3/lpchbb/jiafu/ZnnH_postHCP/Step4_20130404/stitch/"
#indir = "/uscms_data/d3/lpchbb/jiafu/ZnnH_postHCP/Step4_20130314/stitch/"
prefix = "Step4_"
suffix = ".root"
treename = "tree_%s_%s" % (channel, "ctrl")
ZH = TChain(treename)
def main():
# Default conf file, if running with a different one just provide it as an argument
configFile = 'plots.conf'
args = sys.argv[1:]
if len(args) >= 1:
configFile = args[0]
if os.path.exists(configFile):
print('running with config file', configFile)
else:
print('you are trying to use a config file (' + configFile +
') that does not exist!')
sys.exit(1)
# Parse the conf file
cfg = ConfigParser(dict_type=OrderedDict)
cfg.optionxform = str
cfg.read(configFile)
# Set up plotting
inputdir = cfg.get('setup', 'inputdir')
outputdir = cfg.get('setup', 'outputdir')
procs = cfg.get('setup', 'processes').replace(' ', '').split(',')
treename = cfg.get('setup', 'treename')
treename2d = cfg.get('setup', 'treename2d')
comparetype = cfg.get('setup', 'comparetype')
if not comparetype in ['processes', 'sels']:
print('comparetype must be either "processes" or "sels"!')
sys.exit(1)
plotnames = cfg.get('plotting', 'plotnames').replace(' ', '').split(',')
plotnames2d = cfg.get('plotting', 'plotnames2d').replace(' ',
'').split(',')
effplotnames = cfg.get('plotting', 'effplotnames').replace(' ',
'').split(',')
reversecutvars = cfg.get('plotting',
'reversecutvars').replace(' ', '').split(',')
logvars = cfg.get('plotting', 'logvars').replace(' ', '').split(',')
expr = {k: v for k, v in cfg['expressions'].items()}
sel = {k: v for k, v in cfg['sels'].items()}
sel2d = {k: v for k, v in cfg['sels2d'].items()}
proclabels = {k: v for k, v in cfg['proclabels'].items()}
plotlabels = {k: v for k, v in cfg['plotlabels'].items()}
binning = {k: ast.literal_eval(v) for k, v in cfg['binning'].items()}
colors = {k: st.colors[v] for k, v in cfg['colors'].items()}
# ROOT setup stuff
gROOT.SetBatch(True)
st.SetTDRStyle()
st.SetupColors()
gStyle.SetLabelSize(0.03, "XYZ")
gSystem.mkdir(outputdir, True)
print('Making plots!')
files = {
proc: TFile.Open(inputdir + '/' + proc + '_tree.root')
for proc in procs
}
# Loop over 1D variables to be plotted
for n in plotnames:
if n == '' or not n in binning:
continue
print('plotting', n)
(outerloopitems,
innerloopitems) = (procs, sel) if comparetype == 'sels' else (sel,
procs)
# Loop over outer loop items
for x in outerloopitems:
hists = []
ymax = -1.
infile = None
tree = None
selexp = ''
if comparetype == 'processes':
selexp = sel[x]
print('with selection', selexp)
# Loop over inner loop items
for y in innerloopitems:
if comparetype == 'sels':
infile = files[x]
selexp = sel[y]
print('with selection', selexp)
else:
infile = files[y]
tree = infile.FindObjectAny(treename)
hist = TH1D('_'.join(['h', n, x, y]), '', binning[n][0],
binning[n][1], binning[n][2])
hist.SetLineColor(colors[y])
# Check if variable name corresponds to an expression
if n in expr:
tree.Draw(expr[n] + '_'.join(['>>h', n, x, y]), selexp,
'histnorm')
else:
tree.Draw(n + '_'.join(['>>h', n, x, y]), selexp,
'histnorm')
# Histogram setup
st.addOverFlow(hist)
hist.GetXaxis().SetTitle(plotlabels[n])
hist.GetXaxis().SetTitleSize(0.04)
hist.GetYaxis().SetTitle("A.U.")
hist.GetYaxis().SetTitleSize(0.05)
hist.SetLineWidth(3)
hist.SetLineColor(colors[y])
if (hist.GetMaximum() > ymax):
ymax = hist.GetMaximum()
tempmax = 0.0
maxbin = 0
for ibin in range(1, hist.GetNbinsX() + 1):
binc = hist.GetBinContent(ibin)
if binc > tempmax:
tempmax = binc
maxbin = hist.GetBinLowEdge(ibin)
# Add this histogram to the list
hists.append(hist)
# Setup canvas
c = st.MakeCanvas("c", "", 600, 600)
leg = TLegend(0.65, 0.7, 0.95, 0.9)
st.SetLegendStyle(leg)
leg.SetTextSize(0.04)
# Check if plotting in log scale
logy = (n in logvars)
if (logy):
c.SetLogy()
labels = []
# Draw histograms and add each entry to the legend
for hist in hists:
if (logy):
hist.GetYaxis().SetRangeUser(1e-04, 15 * ymax)
else:
hist.GetYaxis().SetRangeUser(0, 1.2 * ymax)
hist.Draw('histsame')
hname = str(hist.GetName())
labelname = hname[hname.rindex('_') + 1:]
leg.AddEntry(hist, proclabels[labelname], 'L')
# Draw legend and save the plot
leg.Draw('same')
st.LDMX_lumi(c, 0, 'Simulation')
c.SaveAs(outputdir + '/' + n + '_' + x +
('_log.pdf' if logy else '.pdf'))
# Loop over 2D plots
for n in plotnames2d:
if n == '':
continue
xvar = n[n.rindex('_') + 1:]
yvar = n[0:n.index('_')]
if not xvar in binning or not yvar in binning:
continue
print('plotting', yvar, 'vs', xvar)
# Loop over processes
for proc in procs:
infile = files[proc]
tree = infile.FindObjectAny(treename2d)
c = st.MakeCanvas('c', '', 600, 600)
# Loop over cut strings
for seln in sel2d:
selexp = sel2d[seln]
print('with selection', selexp)
hist = TH2D('_'.join(['h', n, proc, seln]), '',
binning[xvar][0], binning[xvar][1],
binning[xvar][2], binning[yvar][0],
binning[yvar][1], binning[yvar][2])
c.cd()
logx, logy = False, False
if xvar in logvars:
logx = True
if yvar in logvars:
logy = True
if logx:
c.SetLogx()
if logy:
c.SetLogy()
c.SetLogz()
c.SetLeftMargin(0.13)
c.SetRightMargin(0.18)
print('Drawing', expr[n])
if n in expr:
#tree.Draw(expr[n]+'_'.join(['>>h',n,proc,seln]),selexp,'COLZnorm')
tree.Draw(expr[n] + '_'.join(['>>h', n, proc, seln]),
selexp, 'COLZ')
else:
#tree.Draw(n+'_'.join(['>>h',n,proc,seln]),selexp,'COLZnorm')
tree.Draw(n + '_'.join(['>>h', n, proc, seln]), selexp,
'COLZ')
# Histogram setup
hist.GetXaxis().SetTitle(plotlabels[xvar])
hist.GetXaxis().SetTitleSize(0.05)
hist.GetYaxis().SetTitle(plotlabels[yvar])
hist.GetYaxis().SetTitleSize(0.05)
# Save plot
c.SaveAs(outputdir + '/' + n + '_' + proc + '_' + seln +
('_log.pdf' if logx or logy else '.pdf'))
# Loop over efficiency variables to be plotted
for n in effplotnames:
if n == '' or not n in binning:
continue
print('plotting efficiency vs', n)
(outerloopitems,
innerloopitems) = (procs, sel) if comparetype == 'sels' else (sel,
procs)
# Loop over outer loop items
for x in outerloopitems:
hists = []
effs = []
infile = None
tree = None
selexp = ''
if comparetype == 'sels':
infile = files[x]
tree = infile.FindObjectAny(treename)
else:
selexp = sel[x]
print('with selection', selexp)
isel = -1
# Loop over inner loop items
for y in innerloopitems:
if comparetype == 'sels':
selexp = sel[y]
print('with selection', selexp)
else:
infile = files[y]
tree = infile.FindObjectAny(treename)
isel += 1
hist = TH1D('_'.join(['h', n, x, y]), '', binning[n][0],
binning[n][1], binning[n][2])
hist.SetLineColor(colors[y])
hist.SetMarkerColor(colors[y])
# Check if variable name corresponds to an expression
if n in expr:
tree.Draw(expr[n] + '_'.join(['>>h', n, x, y]), selexp)
else:
tree.Draw(n + '_'.join(['>>h', n, x, y]), selexp)
# Histogram setup
st.addOverFlow(hist)
hist.GetXaxis().SetTitle(plotlabels[n])
hists.append(hist)
# Setup canvas
c = st.MakeCanvas("c", "", 600, 600)
leg = TLegend(0.65, 0.7, 0.95, 0.9)
st.SetLegendStyle(leg)
leg.SetTextSize(0.04)
logy = (n in logvars)
if (logy):
c.SetLogy()
if len(hists):
xmin = hists[0].GetXaxis().GetXmin()
xmax = hists[0].GetXaxis().GetXmax()
c.DrawFrame(xmin, 1e-6 if logy else 0, xmax, 1.1)
graphs = []
emptyhist = None
ihist = -1
for hist in hists:
ihist += 1
hname = str(hist.GetName())
labelname = hname[hname.rindex('_') + 1:]
xmin = hist.GetXaxis().GetXmin()
xmax = hist.GetXaxis().GetXmax()
effgr = pt.computeEffVsCutGraph(hist, n in reversecutvars)
effgr.SetLineWidth(3)
effgr.SetLineColor(colors[labelname])
effgr.SetMarkerColor(colors[labelname])
effgr.GetXaxis().SetTitle(plotlabels[n])
effgr.GetYaxis().SetTitle('Efficiency')
effgr.GetXaxis().SetTitleSize(0.04)
effgr.GetYaxis().SetTitleSize(0.05)
effgr.GetHistogram().GetXaxis().SetLimits(xmin, xmax)
effgr.GetHistogram().GetYaxis().SetRangeUser(
1e-6 if logy else 0.0, 1.1)
c.cd()
effgr.Draw('Csame')
if ihist == 0:
emptyhist = effgr.GetHistogram()
leg.AddEntry(effgr, proclabels[labelname], 'L')
graphs.append(effgr)
if emptyhist:
emptyhist.Draw('AXIS')
for graph in graphs:
graph.Draw('Csame')
c.cd()
leg.Draw('same')
st.LDMX_lumi(c, 0, 'Simulation')
c.SaveAs(outputdir + '/eff_vs_' + n + '_' + y + '.pdf')
for infile in files.values():
infile.Close()
chain.Add(f)
sections = {}
sections["makedataset"] = False
sections["makefits"] = True
sections["makenewfits"] = False
sections["makefits2"] = False
plotting = []
writing = []
imgdir = "figures_20140206/"
if not imgdir.endswith("/"): imgdir += "/"
if gSystem.AccessPathName(imgdir):
gSystem.mkdir(imgdir)
#_______________________________________________________________________________
# Classes/Functions
class Properties:
"""Stores stuff"""
def __init__(self, x, y):
self.x = x
self.y = y
def CMS_label():
old = (latex.GetTextFont(), latex.GetTextSize())
latex.SetTextFont(42); latex.SetTextSize(0.026)
latex.DrawLatex(0.665, 0.968, "Run2012D HLT_L1ETM40_v2")
def create_k_folds(input_root_file_name, input_tree_name = None, k_folds = None):
"""
Inspect a .root ntuple and resample a TTree into k-folds for cross-validation.
"""
input_root_file = TFile(input_root_file_name, 'READ')
# Check function arguments.
if (not input_tree_name) or (not input_root_file.GetListOfKeys().FindObject(input_tree_name)):
print "\nUnspecified or incorrect TTree name. Consider the file contents below.\n"
input_root_file.ls()
print "\n"
return
elif (k_folds == None):
print "\nUnspecified number of folds."
return
else:
input_tree = input_root_file.Get(input_tree_name)
input_tree_n_entries = input_tree.GetEntriesFast()
# Specify output directory, creating it if it doesn't exist.
output_directory = 'k_folds/'
if (gSystem.AccessPathName(output_directory)):
gSystem.mkdir(output_directory)
# Determine the number of entries per fold, assuming the total number of entries is exactly divisible into k folds.
# Create a list holding the entry indices specifying the endpoints for each fold.
n_entries_per_fold = int(input_tree_n_entries / k_folds)
range_list = range(0, n_entries_per_fold * (k_folds + 1), n_entries_per_fold)
# If a remainder exists, we modify accordingly.
n_entries_remainder = int(input_tree_n_entries % k_folds)
if (n_entries_remainder != 0):
for extra in range(0, len(range_list)):
if (extra < n_entries_remainder):
range_list[extra] += extra
else:
range_list[extra] += n_entries_remainder
# Loop over each fold.
for k in range(0, k_folds):
# Create output .root file for the current fold.
output_root_file_name = output_directory + 'CVFold_%s_of_%s.root' % (k+1, k_folds)
output_root_file = TFile(output_root_file_name, 'RECREATE')
# Copy an empty tree to hold the training sample.
output_train_tree = input_tree.CloneTree(0)
output_train_tree.SetName('%s_CV_Train' % input_tree.GetName())
# Copy an empty tree to hold the "testing" sample.
output_test_tree = input_tree.CloneTree(0)
output_test_tree.SetName('%s_CV_Test' % input_tree.GetName())
# Loop over the input tree entries, storing it for testing if it falls within the range
# and training otherwise. I find this faster than conditionals within the loop.
for entry in range(0, range_list[k]):
input_tree.GetEntry(entry, 1) # The second argument 1 means get all branches.
output_train_tree.Fill()
for entry in range(range_list[k], range_list[k+1]):
input_tree.GetEntry(entry, 1)
output_test_tree.Fill()
for entry in range(range_list[k+1], input_tree_n_entries):
input_tree.GetEntry(entry, 1)
output_train_tree.Fill()
# Write the output trees and save the output file.
output_train_tree.Write()
output_test_tree.Write()
output_root_file.Close()
# Close the input file.
input_root_file.Close()
return "Successfully split entries for %s-fold cross-validation." % k_folds
def RunPAF(samples,
selection,
xsec,
nSumOfWeights,
year,
outname,
nSlots=1,
outpath='',
options='',
isamcatnlo=False,
isData=False,
nEvents=0,
FirstEvent=0,
workingdir=''):
from ROOT import PAFProject, PAFIExecutionEnvironment, PAFSequentialEnvironment, PAFPROOFLiteEnvironment, PAFPoDEnvironment
from ROOT import vector, TString, gSystem
# PAF mode selection (based on number of slots)
pafmode = PAFSequentialEnvironment()
if nSlots <= 1: pafmode = PAFSequentialEnvironment()
else: pafmode = PAFPROOFLiteEnvironment(nSlots)
myProject = PAFProject(pafmode)
myProject.AddDataFiles(samples)
myProject.SetDefaultTreeName("Events")
# Deal with first and last event
if nEvents > 0: myProject.SetNEvents(nEvents)
if FirstEvent > 0: myProject.SetFirstEvent(FirstEvent)
if workingdir == "": workingdir = os.getcwd()
# Set output file
if outpath == '': outpath = workingdir + "/" + selection + "_temp"
gSystem.mkdir(outpath, 1)
myProject.SetOutputFile(outpath + "/" + outname + ".root")
if verbose: print '## Output to: %s' % (outpath + "/" + outname + ".root")
# Parameters for the analysis
myProject.SetInputParam("sampleName", outname)
myProject.SetInputParam("IsData", isData)
myProject.SetInputParam("weight", xsec / nSumOfWeights)
myProject.SetInputParam("IsMCatNLO", isamcatnlo)
myProject.SetInputParam("selection", selection)
myProject.SetInputParam("WorkingDir", workingdir)
myProject.SetInputParam("xsec", xsec)
myProject.SetInputParam("_options", options)
myProject.SetInputParam("year", str(year))
# Name of analysis class
myProject.AddSelectorPackage("LeptonSelector")
myProject.AddSelectorPackage("JetSelector")
myProject.AddSelectorPackage("EventBuilder")
# Analysis selector
if selection != "":
myProject.AddSelectorPackage(selection)
#if selection == 'TopAnalysis':
# myProject.AddSelectorPackage("LepEffTop")
# Additional packages
myProject.AddPackage("Lepton")
myProject.AddPackage("Jet")
myProject.AddPackage("mt2")
myProject.AddPackage("Functions")
myProject.AddPackage("LeptonSF")
myProject.AddPackage("BTagSFUtil")
myProject.Run()
