def guess_available_selections_from_histofiles(inputDir, plot_group, verbose):
grp = plot_group.name
sys = plot_group.syst # default one, usually 'NOM' which is just fine
available_root_files = [
f for f in os.listdir(inputDir) if sys in f and grp in f
]
pre = commonPrefix(available_root_files)
suf = commonSuffix(available_root_files)
def abbreviate_suffix(s):
if s[1] in ['.', '_']:
return s[1:]
suf = abbreviate_suffix(suf)
if verbose:
print "guess_available_selections_from_histofiles: pre '{0}', suf '{1}'".format(
pre, suf)
available_selections = [
r.replace(pre, '').replace(suf, '') for r in available_root_files
]
selections = regions_to_plot(regions=available_selections)
if not selections:
raise UserWarning("guess_available_selections_from_histofiles:"
" failed to guess anything from '{0}'"
"\nTry using --region".format(inputDir))
return selections
def guess_out_name(input_names=[]):
pref = commonPrefix(hnames)
suff = commonSuffix(hnames)
out_name = (pref+label+suff) if (pref or suff) else (input_names[0]+'_'+label)
out_name = out_name.replace('__','_') if remove_extra_underscore else out_name
out_name = out_name[:-1] if remove_extra_underscore and out_name.endswith('_') else out_name
return out_name
def build_average_sys_yield(cls,
component_nominal_yields=[],
component_syst_yields_up=[],
component_syst_yields_do=[]):
"build the average sys yield of up/down variations"
cny = component_nominal_yields
csyu = component_syst_yields_up
csyd = component_syst_yields_do
cls.check_attribute_with_same_value('region', cny + csyu + csyd)
cls.check_attribute_with_same_value('variation', csyu)
cls.check_attribute_with_same_value('variation', csyd)
nom_yield = cls.build_merged_yield(cny, cny)
up_yield = cls.build_merged_yield(cny, csyu)
do_yield = cls.build_merged_yield(cny, csyd)
vup = csyu[0].variation
vdo = csyd[0].variation
variation = commonPrefix([vup, vdo]).strip('_') + commonSuffix(
[vup, vdo]).strip('_')
nom = nom_yield.get_yield()
up = up_yield.get_yield()
do = do_yield.get_yield()
avg_delta = 0.5 * (abs(nom - up) + abs(nom - do))
tot_yield = SignalYield(nom_yield, variation, nom + avg_delta)
return tot_yield
def guess_out_name(input_names=[]):
pref = commonPrefix(hnames)
suff = commonSuffix(hnames)
out_name = (pref+label+suff) if (pref or suff) else (input_names[0]+'_'+label)
out_name = out_name.replace('__','_') if remove_extra_underscore else out_name
out_name = out_name[:-1] if remove_extra_underscore and out_name.endswith('_') else out_name
return out_name
def fastSamplesFromFilenames(filenames=[], verbose=False):
"""Assume a list of filenames with some common prefix (basepath)
and suffix (tag): strip away the prefix&suffix, and return a list
of Dataset objects obtained by matching the name to the dataset
"""
prefix = utils.commonPrefix(filenames)
suffix = utils.commonSuffix(filenames)
if verbose : print "fastSamplesFromFilenames: prefix '%s' suffix '%s'"%(prefix, suffix)
if not prefix or not suffix : print "fastSamplesFromFilenames: warning prefix '%s' suffix '%s'"%(prefix, suffix)
filenames = [f.replace(prefix, '').replace(suffix, '') for f in filenames]
dsets = dict([(d.name, d) for d in datasets])
return [dsets[f] for f in filenames if f in dsets]
def fastSamplesFromFilenames(filenames=[], verbose=False):
"""Assume a list of filenames with some common prefix (basepath)
and suffix (tag): strip away the prefix&suffix, and return a list
of Dataset objects obtained by matching the name to the dataset
"""
prefix = utils.commonPrefix(filenames)
suffix = utils.commonSuffix(filenames)
if verbose:
print "fastSamplesFromFilenames: prefix '%s' suffix '%s'" % (prefix,
suffix)
if not prefix or not suffix:
print "fastSamplesFromFilenames: warning prefix '%s' suffix '%s'" % (
prefix, suffix)
filenames = [f.replace(prefix, '').replace(suffix, '') for f in filenames]
dsets = dict([(d.name, d) for d in datasets])
return [dsets[f] for f in filenames if f in dsets]
def guess_available_selections_from_histofiles(inputDir, plot_group, verbose):
grp = plot_group.name
sys = plot_group.syst # default one, usually 'NOM' which is just fine
available_root_files = [f for f in os.listdir(inputDir) if sys in f and grp in f]
pre = commonPrefix(available_root_files)
suf = commonSuffix(available_root_files)
def abbreviate_suffix(s):
if s[1] in ['.','_']:
return s[1:]
suf = abbreviate_suffix(suf)
if verbose:
print "guess_available_selections_from_histofiles: pre '{0}', suf '{1}'".format(pre, suf)
available_selections = [r.replace(pre, '').replace(suf, '') for r in available_root_files]
selections = regions_to_plot(regions=available_selections)
if not selections:
raise UserWarning("guess_available_selections_from_histofiles:"
" failed to guess anything from '{0}'"
"\nTry using --region".format(inputDir))
return selections
def buildRatioHistos(histosNum={}, histosDen={}) :
"assume that the histos are organized in two dict[vt][sample] provided by buildHistos"
histosPerVtypePerSample = {}
def sameLists(l1=[], l2=[]) : return len(l1)==len(l2) and sorted(l1)==sorted(l2)
assert sameLists(histosNum.keys(), histosDen.keys()),"num and den w/ different vtype keys"
result = {}
for vt in histosDen.keys() :
resultPerSample = {}
hvn, hvd = histosNum[vt], histosDen[vt]
assert sameLists(hvn.keys(), hvd.keys()),"num and den w/ different sample keys"
for s in hvd.keys() :
hnum, hden = hvn[s], hvd[s]
hNames = [h.GetName() for h in hnum, hden]
pre, suf = commonPrefix(hNames), commonSuffix(hNames)
hName = pre+'_ratio_'+suf if len(pre) and len(suf) else '_over_'.join(hNames)
h = buildRatioHistogram(hnum, hden, hName)
resetErrors(h)
resultPerSample[s] = h
result[vt] = resultPerSample
return result
def buildRatioHistos(histosNum={}, histosDen={}) :
"assume that the histos are organized in two dict[vt][sample] provided by buildHistos"
histosPerVtypePerSample = {}
def sameLists(l1=[], l2=[]) : return len(l1)==len(l2) and sorted(l1)==sorted(l2)
assert sameLists(histosNum.keys(), histosDen.keys()),"num and den w/ different vtype keys"
result = {}
for vt in histosDen.keys() :
resultPerSample = {}
hvn, hvd = histosNum[vt], histosDen[vt]
assert sameLists(hvn.keys(), hvd.keys()),"num and den w/ different sample keys"
for s in hvd.keys() :
hnum, hden = hvn[s], hvd[s]
hNames = [h.GetName() for h in hnum, hden]
pre, suf = commonPrefix(hNames), commonSuffix(hNames)
hName = pre+'_ratio_'+suf if len(pre) and len(suf) else '_over_'.join(hNames)
h = buildRatioHistogram(hnum, hden, hName)
resetErrors(h)
resultPerSample[s] = h
result[vt] = resultPerSample
return result
def merged_sample_name(cls, names=[]):
return commonPrefix(names) + commonSuffix(names).replace('__', '_')
def guessBaseHistoname(histonames=[]) :
basename = commonSuffix(histonames)
basename = commonPrefix(histonames) if not len(basename) else basename
basename = histonames[0] if not len(basename) else basename
return basename.strip(' _')