def __init__(self):
self.filenamesByReqid = collections.defaultdict(list)
self.masspointByReqid = collections.defaultdict(list)
self.reqidProvider = ModeAWhDbMergedFake2Lreqid()
self.parDb = ModeAWhDbPar()
self.reqDb = ModeAWhDbReqid()
self.overwrite = False
self.verbose = False
self.regexHist = '.*'
class SamplesMerger :
def __init__(self) :
self.filenamesByReqid = collections.defaultdict(list)
self.masspointByReqid = collections.defaultdict(list)
self.reqidProvider = ModeAWhDbMergedFake2Lreqid()
self.parDb = ModeAWhDbPar()
self.reqDb = ModeAWhDbReqid()
self.overwrite = False
self.verbose = False
self.regexHist = '.*'
def idTagFromFilename(self, fname) :
'parse something like wA_noslep_WH_2Lep_9_May16_n0139.AnaHists'
match = re.search('wA_noslep_WH_2Lep_(?P<id>\d+?)_(?P<tag>.*?).AnaHist', fname)# nongreedy
return match.group('id'), match.group('tag')
def addFile(self, filename) :
sample = guessSampleFromFilename(filename)
mc1, mn1 = self.parDb.mc1Mn1ByReqid(self.reqDb.reqidBySample(sample))
fakeReqid = self.reqidProvider.reqidByMc1Mn1(mc1, mn1)
self.filenamesByReqid[fakeReqid].append(filename)
self.masspointByReqid[fakeReqid].append((mc1, mn1))
def mergeAndWrite(self, outdir) :
"""write merged histos, unless !overwrite, in which case just
compute merged coords and fnames"""
def mergeFiles(targetFile, filenames, regex) :
infiles = [r.TFile.Open(f) for f in filenames]
histonames = [h for h in getAllHistoNames(infiles[0]) if re.search(regex, h)]
scale = 1.0/len(infiles)
outFile = r.TFile.Open(targetFile, 'recreate')
outFile.cd()
for hn in histonames :
h = infiles[0].Get(hn).Clone()
h.Scale(scale)
for inf in infiles[1:] : h.Add(inf.Get(hn), scale)
h.Write()
outFile.Close()
self.mergedPointsByReqid = {}
self.outFnameByReqid = {}
reqids = self.filenamesByReqid.keys()
for rid in reqids :
fnames = self.filenamesByReqid[rid]
points = self.masspointByReqid[rid]
npoints = len(points)
assert npoints,"cannot merge 0 samples"
def avgPoint(points) :
np = float(len(points))
return (sum([x for x,y in points])/np,
sum([y for x,y in points])/np)
point = avgPoint(points)
ids, tags = zip(*[self.idTagFromFilename(f) for f in fnames]) # vertical slice
assert len(set(tags))==1,"cannot merge samples with different tags %s"%str(fnames)
outfname = outdir+'/'+os.path.basename(fnames[0]).replace(str(ids[0]), str(rid))
self.outFnameByReqid[rid], self.mergedPointsByReqid[rid] = outfname, point
if self.verbose : print str(point)+' : '+outfname
skipMerge = os.path.exists(outfname) and not self.overwrite
if skipMerge : continue
else : mergeFiles(outfname, fnames, self.regexHist)
def printMergedFiles(self) :
for rid in self.outFnameByReqid.keys() :
print self.mergedPointsByReqid[rid], ' : ', self.outFnameByReqid[rid]
def __init__(self) :
self.filenamesByReqid = collections.defaultdict(list)
self.masspointByReqid = collections.defaultdict(list)
self.reqidProvider = ModeAWhDbMergedFake2Lreqid()
self.parDb = ModeAWhDbPar()
self.reqDb = ModeAWhDbReqid()
self.overwrite = False
self.verbose = False
self.regexHist = '.*'
import collections, optparse, re, sys
import ROOT as r
r.PyConfig.IgnoreCommandLineOptions = True
r.gROOT.SetBatch(1)
from PickleUtils import readFromPickle
from SampleUtils import ModeAWhDbPar, ModeAWhDbReqid, ModeAWhDbMergedFake2Lreqid
parser = optparse.OptionParser()
parser.add_option("-v", "--verbose", action="store_true", dest="verbose", default=False,
help="print more details about what is going on")
(options, args) = parser.parse_args()
verbose = options.verbose
reqDb = ModeAWhDbReqid()
parDb = ModeAWhDbPar()
allMc1 = [float(m) for m in parDb.allMc1()]
allMn1 = [float(m) for m in parDb.allMn1()]
def roundup(val) : return round(float(val)+0.5)
def rounddo(val) : return round(float(val)+0.0)
mc1Range = {'min': rounddo(min(allMc1)), 'max' : roundup(max(allMc1))}
mn1Range = {'min': rounddo(min(allMn1)), 'max' : roundup(max(allMn1))}
histo2l = r.TH2F('mc1mn1_2lep_reqids',
'ReqIds for the WH 2lep grid ;mc_{1};mn_{1}',
50, float(mc1Range['min']), float(mc1Range['max']),
50, float(mn1Range['min']), float(mn1Range['max']))
histo2lnth = r.TH2F('mc1mn1_2lep_notauhad_reqids',
'ReqIds for the WH notauhad 2lep grid ;mc_{1};mn_{1}',
sigScale = options.sigScale
verbose = options.verbose
countsSigSampleSel = readFromPickle(sigInputFname)
countsBkgSampleSel = readFromPickle(bkgInputFname)
interestingSelections = ["sr%d" % i for i in range(6, 9 + 1)]
countBkgTot = collections.defaultdict(float)
for sample, countsSel in countsBkgSampleSel.iteritems():
if sample in ['data', 'totbkg']: continue
for sel, counts in countsSel.iteritems():
if sel not in interestingSelections: continue
print 'adding ' + sample + ' to ' + sel + ' (' + str(counts) + ')'
countBkgTot[sel] += counts
print countBkgTot
reqDb = ModeAWhDbReqid()
parDb = ModeAWhDbPar()
def selIsRelevant(sel):
return sel.startswith('sr')
def selIsFinal(sel):
return sel in interestingSelections
mc1Range = {'min': min(parDb.allMc1()), 'max': max(parDb.allMc1())}
mn1Range = {'min': min(parDb.allMn1()), 'max': max(parDb.allMn1())}
histos = dict()
sigScale = options.sigScale
verbose = options.verbose
countsSigSampleSel = readFromPickle(sigInputFname)
countsBkgSampleSel = readFromPickle(bkgInputFname)
interestingSelections = ["sr%d"%i for i in range(6,9+1)]
countBkgTot = collections.defaultdict(float)
for sample, countsSel in countsBkgSampleSel.iteritems() :
if sample in ['data', 'totbkg'] : continue
for sel, counts in countsSel.iteritems() :
if sel not in interestingSelections : continue
print 'adding '+sample+' to '+sel+' ('+str(counts)+')'
countBkgTot[sel] += counts
print countBkgTot
reqDb = ModeAWhDbReqid()
parDb = ModeAWhDbPar()
def selIsRelevant(sel) : return sel.startswith('sr')
def selIsFinal(sel) : return sel in interestingSelections
mc1Range = {'min': min(parDb.allMc1()), 'max' : max(parDb.allMc1())}
mn1Range = {'min': min(parDb.allMn1()), 'max' : max(parDb.allMn1())}
histos = dict()
backgroundRelErr = 0.2
for sample, countsSel in countsSigSampleSel.iteritems() :
mc1, mn1 = parDb.mc1Mn1ByReqid(reqDb.reqidBySample(sample))
print "%s (%.1f, %.1f) " % (sample, mc1, mn1)
for sel, counts in sorted(countsSel.iteritems()) :
if not selIsFinal(sel) : continue
r.gROOT.SetBatch(1)
from PickleUtils import readFromPickle
from SampleUtils import ModeAWhDbPar, ModeAWhDbReqid, ModeAWhDbMergedFake2Lreqid
parser = optparse.OptionParser()
parser.add_option("-v",
"--verbose",
action="store_true",
dest="verbose",
default=False,
help="print more details about what is going on")
(options, args) = parser.parse_args()
verbose = options.verbose
reqDb = ModeAWhDbReqid()
parDb = ModeAWhDbPar()
allMc1 = [float(m) for m in parDb.allMc1()]
allMn1 = [float(m) for m in parDb.allMn1()]
def roundup(val):
return round(float(val) + 0.5)
def rounddo(val):
return round(float(val) + 0.0)
mc1Range = {'min': rounddo(min(allMc1)), 'max': roundup(max(allMc1))}
class SamplesMerger:
def __init__(self):
self.filenamesByReqid = collections.defaultdict(list)
self.masspointByReqid = collections.defaultdict(list)
self.reqidProvider = ModeAWhDbMergedFake2Lreqid()
self.parDb = ModeAWhDbPar()
self.reqDb = ModeAWhDbReqid()
self.overwrite = False
self.verbose = False
self.regexHist = '.*'
def idTagFromFilename(self, fname):
'parse something like wA_noslep_WH_2Lep_9_May16_n0139.AnaHists'
match = re.search(
'wA_noslep_WH_2Lep_(?P<id>\d+?)_(?P<tag>.*?).AnaHist',
fname) # nongreedy
return match.group('id'), match.group('tag')
def addFile(self, filename):
sample = guessSampleFromFilename(filename)
mc1, mn1 = self.parDb.mc1Mn1ByReqid(self.reqDb.reqidBySample(sample))
fakeReqid = self.reqidProvider.reqidByMc1Mn1(mc1, mn1)
self.filenamesByReqid[fakeReqid].append(filename)
self.masspointByReqid[fakeReqid].append((mc1, mn1))
def mergeAndWrite(self, outdir):
"""write merged histos, unless !overwrite, in which case just
compute merged coords and fnames"""
def mergeFiles(targetFile, filenames, regex):
infiles = [r.TFile.Open(f) for f in filenames]
histonames = [
h for h in getAllHistoNames(infiles[0]) if re.search(regex, h)
]
scale = 1.0 / len(infiles)
outFile = r.TFile.Open(targetFile, 'recreate')
outFile.cd()
for hn in histonames:
h = infiles[0].Get(hn).Clone()
h.Scale(scale)
for inf in infiles[1:]:
h.Add(inf.Get(hn), scale)
h.Write()
outFile.Close()
self.mergedPointsByReqid = {}
self.outFnameByReqid = {}
reqids = self.filenamesByReqid.keys()
for rid in reqids:
fnames = self.filenamesByReqid[rid]
points = self.masspointByReqid[rid]
npoints = len(points)
assert npoints, "cannot merge 0 samples"
def avgPoint(points):
np = float(len(points))
return (sum([x for x, y in points]) / np,
sum([y for x, y in points]) / np)
point = avgPoint(points)
ids, tags = zip(*[self.idTagFromFilename(f)
for f in fnames]) # vertical slice
assert len(
set(tags)
) == 1, "cannot merge samples with different tags %s" % str(fnames)
outfname = outdir + '/' + os.path.basename(fnames[0]).replace(
str(ids[0]), str(rid))
self.outFnameByReqid[rid], self.mergedPointsByReqid[
rid] = outfname, point
if self.verbose: print str(point) + ' : ' + outfname
skipMerge = os.path.exists(outfname) and not self.overwrite
if skipMerge: continue
else: mergeFiles(outfname, fnames, self.regexHist)
def printMergedFiles(self):
for rid in self.outFnameByReqid.keys():
print self.mergedPointsByReqid[rid], ' : ', self.outFnameByReqid[
rid]
