def __call__(self, pnode):
if pnode.type != "CMSSW":
return
#Can't take datasetInfo from getOutputDatasetsWithPSet as that changes
#the AppFamily to the processing configuration, hence file insertions
#fail due to a missing algo. WARNING: relies on identical dataset order
glblTags = [
x['Conditions']
for x in getOutputDatasetsWithPSet(pnode, sorted=True)
]
for dataset, globalTag in zip(getOutputDatasets(pnode, sorted=True),
glblTags):
dataset['Conditions'] = globalTag
primary = DBSWriterObjects.createPrimaryDataset(
dataset, self.apiRef)
mergeAlgo = DBSWriterObjects.createMergeAlgorithm(
dataset, self.apiRef)
DBSWriterObjects.createProcessedDataset(primary, mergeAlgo,
dataset, self.apiRef)
inputDataset = dataset.get('ParentDataset', None)
if inputDataset == None:
continue
processedDataset = dataset["ProcessedDataset"]
self.apiRef.insertMergedDataset(inputDataset, processedDataset,
mergeAlgo)
# algorithm used when process jobs produce merged files directly
# doesnt contain pset content - taken from processing (same hash)
mergeDirectAlgo = DBSWriterObjects.createAlgorithm(
dataset, None, self.apiRef)
self.apiRef.insertAlgoInPD(makeDSName2(dataset), mergeDirectAlgo)
logging.debug("ProcessedDataset: %s" % processedDataset)
logging.debug("inputDataset: %s" % inputDataset)
logging.debug("mergeAlgo: %s" % mergeAlgo)
return
def __call__(self, pnode):
if pnode.type != "CMSSW":
return
#Can't take datasetInfo from getOutputDatasetsWithPSet as that changes
#the AppFamily to the processing configuration, hence file insertions
#fail due to a missing algo. WARNING: relies on identical dataset order
glblTags = [x['Conditions'] for x in getOutputDatasetsWithPSet(pnode,
sorted = True)]
for dataset, globalTag in zip(getOutputDatasets(pnode, sorted = True),
glblTags):
dataset['Conditions'] = globalTag
primary = DBSWriterObjects.createPrimaryDataset(
dataset, self.apiRef)
mergeAlgo = DBSWriterObjects.createMergeAlgorithm(dataset,
self.apiRef)
DBSWriterObjects.createProcessedDataset(
primary, mergeAlgo, dataset, self.apiRef)
inputDataset = dataset.get('ParentDataset', None)
if inputDataset == None:
continue
processedDataset = dataset["ProcessedDataset"]
self.apiRef.insertMergedDataset(
inputDataset, processedDataset, mergeAlgo)
# algorithm used when process jobs produce merged files directly
# doesnt contain pset content - taken from processing (same hash)
mergeDirectAlgo = DBSWriterObjects.createAlgorithm(
dataset, None, self.apiRef)
self.apiRef.insertAlgoInPD(makeDSName2(dataset), mergeDirectAlgo)
logging.debug("ProcessedDataset: %s"%processedDataset)
logging.debug("inputDataset: %s"%inputDataset)
logging.debug("mergeAlgo: %s"%mergeAlgo)
return
class _CreateDatasetOperator:
"""
_CreateDatasetOperator_
Operator for creating datasets from a workflow node
"""
def __init__(self, apiRef, workflow):
self.apiRef = apiRef
self.workflow = workflow
def __call__(self, pnode):
if pnode.type != "CMSSW":
return
datasets = getOutputDatasetsWithPSet(pnode, sorted=True)
cfgMeta = None
try:
cfgInt = pnode.cfgInterface
cfgMeta = cfgInt.configMetadata
cfgMeta['Type'] = self.workflow.parameters["RequestCategory"]
except Exception, ex:
msg = "Unable to Extract cfg data from workflow"
msg += str(ex)
logging.error(msg)
return
for dataset in datasets:
primary = DBSWriterObjects.createPrimaryDataset(
dataset, self.apiRef)
algo = DBSWriterObjects.createAlgorithm(dataset, cfgMeta,
self.apiRef)
processed = DBSWriterObjects.createProcessedDataset(
primary, algo, dataset, self.apiRef)
return
jobReportFile['TotalEvents'] = nEvents
jobReportFile['SEName'] = "srm.cern.ch"
##jobReport.write('FrameworkJobReport.xml')
localDbsUrl = "https://cmst0dbs.cern.ch:8443/cms_dbs_prod_tier0_writer/servlet/DBSServlet"
dbswriter = DBSWriter(localDbsUrl,level='ERROR')
primary = DBSWriterObjects.createPrimaryDataset(datasetStrmr, dbswriter.dbs)
datasetStrmr['ApplicationName'] = appName
datasetStrmr['ApplicationVersion'] = appVersion
datasetStrmr['ApplicationFamily'] = 'DAQ'
datasetStrmr['PSetHash'] = 'NA'
datasetStrmr['PSetContent'] = 'NA'
algo = DBSWriterObjects.createAlgorithm(datasetStrmr, None, dbswriter.dbs)
processed = DBSWriterObjects.createProcessedDataset(primary, algo, datasetStrmr, dbswriter.dbs)
try:
blocks = dbswriter.insertFiles(jobReport, insertDetectorData = True)
except DBSWriterError, ex:
print "%s"%ex
# limit block size to 500
for blockName in blocks:
dbswriter.manageFileBlock(blockName, maxFiles=500)
