def parse(self):
if self.parsed:
return
self.options.parseArguments()
if self.options.processIdMap != "":
self.readProcessIdMap(self.options.processIdMap)
if self.useAAA:
self.filePrepend = "root://xrootd-cms.infn.it/"
elif self.useEOS:
self.filePrepend = "root://eoscms//eos/cms"
dataset = None
if self.dataset != "":
print "Reading dataset (%s) %s" % (self.campaign, self.dataset)
if self.dryRun and self.getMaxJobs:
dataset = SamplesManager(
"$CMSSW_BASE/src/%s/MetaData/data/%s/datasets.json" %
(self.metaDataSrc, self.campaign),
self.crossSections,
).getDatasetMetaData(self.maxEvents,
self.dataset,
jobId=-1,
nJobs=self.nJobs)
else:
dataset = SamplesManager(
"$CMSSW_BASE/src/%s/MetaData/data/%s/datasets.json" %
(self.metaDataSrc, self.campaign),
self.crossSections,
).getDatasetMetaData(self.maxEvents,
self.dataset,
jobId=self.jobId,
nJobs=self.nJobs)
if not dataset:
print "Could not find dataset %s in campaing %s/%s" % (
self.dataset, self.metaDataSrc, self.campaing)
sys.exit(-1)
self.dataset = dataset
# auto-detect data from xsec = 0
if self.dataset:
name, xsec, totEvents, files, maxEvents = self.dataset
if type(xsec) != dict or type(xsec.get("xs", None)) != float:
print "Warning: you are running on a dataset for which you specified no cross section: \n %s " % name
else:
if self.processType == "" and xsec["xs"] == 0.:
self.processType = "data"
self.processId = self.getProcessId(name)
outputFile = self.outputFile
if self.jobId != -1:
outputFile = "%s_%d.root" % (outputFile.replace(".root",
""), self.jobId)
self.setType("outputFile", VarParsing.VarParsing.varType.string)
self.outputFile = outputFile
self.parsed = True
def parse(self):
if self.parsed:
return
self.options.parseArguments()
if self.options.processIdMap != "":
self.readProcessIdMap(self.options.processIdMap)
if self.useAAA:
self.filePrepend = "root://xrootd-cms.infn.it/"
elif self.useEOS:
self.filePrepend = "root://eoscms//eos/cms"
dataset = None
if self.dataset != "":
print "JOBCONFIG", self.dataset
print "Reading dataset (%s) %s" % (self.campaign, self.dataset)
#MATTEO
if self.dryRun and self.getMaxJobs:
dataset = SamplesManager(
"$CMSSW_BASE/src/%s/MetaData/data/%s/datasets.json" %
(self.metaDataSrc, self.campaign),
self.crossSections,
).getDatasetMetaData(self.maxEvents,
self.dataset,
jobId=-1,
nJobs=self.nJobs)
else:
dataset = SamplesManager(
"$CMSSW_BASE/src/%s/MetaData/data/%s/datasets.json" %
(self.metaDataSrc, self.campaign),
self.crossSections,
).getDatasetMetaData(self.maxEvents,
self.dataset,
jobId=self.jobId,
nJobs=self.nJobs)
self.dataset = dataset
outputFile = self.outputFile
if self.jobId != -1:
outputFile = "%s_%d.root" % (outputFile.replace(".root",
""), self.jobId)
self.setType("outputFile", VarParsing.VarParsing.varType.string)
self.outputFile = outputFile
def checkCrossSections(self):
# checks if all cross sections (apart from data)
# were specified
import re
from flashgg.MetaData.samples_utils import SamplesManager
from flashgg.MetaData.JobConfig import JobConfig
# ignore any escaping etc.
kwargs = {}
for part in re.split('\s+', self.options.cmdLine):
key, value = part.split('=', 1)
kwargs[key] = value
if not kwargs.has_key('campaign'):
print >> sys.stderr, "WARNING: campaign not set"
# automatically get metadata locations right for both high-mass and SM Higgs
try:
from diphotons.MetaData.JobConfig import customize
except ImportError:
from flashgg.MetaData.JobConfig import customize
if not kwargs.has_key("metaDataSrc"):
kwargs["metaDataSrc"] = customize.metaDataSrc
if not kwargs.has_key("crossSections"):
kwargs["crossSections"] = customize.crossSections
# TODO: the following line is duplicated from
# JobConfig(..) but we can't make JobConfig.options.parseArguments()
# to parse anything other than sys.argv without modifying sys.argv temporarily...
campaign = kwargs.get('campaign', "")
jobConfig = JobConfig(**kwargs)
sm = SamplesManager(
"$CMSSW_BASE/src/%s/MetaData/data/%s/datasets.json" %
(jobConfig.metaDataSrc, campaign),
jobConfig.crossSections,
)
hasProblem = False
# loop over all types of processes (data, signal, background)
for dsList in self.options.processes.values():
for idsName in dsList:
#if args were provided for this dataset, then it is a list...
if isinstance(idsName, list):
dsName = idsName[0]
#... or just a string
else:
dsName = idsName
# check if this datasets was selected
if not self.isSelectedDataset(dsName):
# skip this dataset
continue
#----------
try:
name, xsec, totEvents, files, maxEvents = sm.getDatasetMetaData(
jobConfig.maxEvents,
dsName,
jobId=-1,
nJobs=jobConfig.nJobs)
if xsec == None:
print >> sys.stderr, "cross section for", dsName, "not found in", jobConfig.crossSections
hasProblem = True
except Exception, ex:
if type(ex) == Exception and (
ex.message.startswith(
'No dataset matched the request:')
or ex.message.startswith(
'More then one dataset matched the request:')):
print >> sys.stderr, ex.message
else:
raise
