def split(self, n, nSub=None):
''' Split sample into n sub-samples
'''
if n == 1: return self
if not n >= 1:
raise ValueError("Can not split into: '%r'" % n)
chunks = helpers.partition(self.files, min(n, len(self.files)))
splitSamps = [
FWLiteSample(name=self.name + "_%i" % n_sample,
files=chunks[n_sample],
color=self.color,
texName=self.texName)
for n_sample in xrange(len(chunks))
]
if nSub == None:
return splitSamps
else:
if nSub < len(chunks):
return splitSamps[nSub]
else:
return None
def split( self, n, nSub=None, clear = True, shuffle = False):
''' Split sample into n sub-samples
'''
if n==1: return self
if not n>=1:
raise ValueError( "Can not split into: '%r'" % n )
files = self.files
if shuffle: random.shuffle( files )
chunks = helpers.partition( files, min(n , len(files) ) )
if clear: self.clear() # Kill yourself.
if nSub == None:
return [ HEPMCSample(
name = self.name+"_%i" % n_sample,
files = chunks[n_sample],
normalization = self.normalization,
isData = self.isData,
color = self.color,
texName = self.texName ) for n_sample in xrange(len(chunks)) ]
else:
return HEPMCSample(
name = self.name,
files = chunks[nSub],
normalization = self.normalization,
isData = self.isData,
color = self.color,
texName = self.texName )
def split(self, n, nSub=None, clear=True, shuffle=False):
''' Split sample into n sub-samples
'''
if n == 1: return self
if not n >= 1:
raise ValueError("Can not split into: '%r'" % n)
files = self.files
if shuffle: random.shuffle(files)
chunks = helpers.partition(files, min(n, len(files)))
if clear: self.clear() # Kill yourself.
if nSub == None:
return [
HEPMCSample(name=self.name + "_%i" % n_sample,
files=chunks[n_sample],
normalization=self.normalization,
isData=self.isData,
color=self.color,
texName=self.texName)
for n_sample in xrange(len(chunks))
]
else:
return HEPMCSample(name=self.name,
files=chunks[nSub],
normalization=self.normalization,
isData=self.isData,
color=self.color,
texName=self.texName)
def split(self, n, nSub=None, clear=True, shuffle=False):
''' Split sample into n sub-samples
'''
if n == 1: return self
if not n >= 1:
raise ValueError("Cannot split into: '%r'" % n)
files = self.files
if shuffle: random.shuffle(files)
chunks = helpers.partition(files, min(n, len(files)))
if clear: self.clear() # Kill yourself.
splitSamps = [
Sample(name=self.name + "_%i" % n_sample,
treeName=self.treeName,
files=chunks[n_sample],
xSection=self.xSection,
normalization=self.normalization,
selectionString=self.selectionString,
weightString=self.weightString,
isData=self.isData,
color=self.color,
texName=self.texName) for n_sample in xrange(len(chunks))
]
if hasattr(self, 'json'):
for s in splitSamps:
s.json = self.json
if nSub == None:
return splitSamps
else:
if nSub < len(chunks):
return splitSamps[nSub]
else:
return None
def split(self, n, nSub = None, clear = True, shuffle = False):
''' Split sample into n sub-samples
'''
if n==1: return self
if not n>=1:
raise ValueError( "Cannot split into: '%r'" % n )
files = self.files
if shuffle: random.shuffle( files )
chunks = helpers.partition( files, min(n , len(files) ) )
if clear: self.clear() # Kill yourself.
splitSamps = [Sample(
name = self.name + "_%i" % n_sample,
treeName = self.treeName,
files = chunks[n_sample],
xSection = self.xSection,
normalization = self.normalization,
selectionString = self.selectionString,
weightString = self.weightString,
isData = self.isData,
color = self.color,
texName = self.texName) for n_sample in xrange(len(chunks))]
if hasattr(self, 'json'):
for s in splitSamps:
s.json = self.json
if nSub == None:
return splitSamps
else:
if nSub<len(chunks):
return splitSamps[nSub]
else:
return None
def split( self, n, nSub = None):
''' Split sample into n sub-samples
'''
if n==1: return self
if not n>=1:
raise ValueError( "Can not split into: '%r'" % n )
chunks = helpers.partition( self.files, min(n , len(self.files) ) )
splitSamps = [ FWLiteSample(
name = self.name+"_%i" % n_sample,
files = chunks[n_sample],
color = self.color,
texName = self.texName ) for n_sample in xrange(len(chunks)) ]
if nSub == None:
return splitSamps
else:
if nSub<len(chunks):
return splitSamps[nSub]
else:
return None
for c in channels:
logger.debug("Queuing jobs for channel %s", c)
jobs.append((region, c, setup, False))
for var in variations:
jobs.append(
(region, c, setup.sysClone(sys={"reweight": [var]}), False))
logger.debug("Queuing jobs for channel %s", c)
jobs.append((noRegions[0], "all", setupIncl, True))
for var in variations:
jobs.append((noRegions[0], "all",
setupIncl.sysClone(sys={"reweight": [var]}), True))
logger.info("Created %s jobs", len(jobs))
jobs = partition(jobs, args.nJobs)[args.job]
logger.info("Running over %s jobs", len(jobs))
if not args.combine:
if args.cores == 1:
results = map(wrapper, jobs)
else:
from multiprocessing import Pool
pool = Pool(processes=args.cores)
results = pool.map(wrapper, jobs)
pool.close()
pool.join()
logger.info("All done.")