def main():
parser = prepare_parser()
options, args = parser.parse_args()
if options.wmtest:
print "Setting to injection in cmswebtest : ", options.wmtesturl
wma.testbed(options.wmtesturl)
for cfg_name in args:
wma.upload_to_couch(cfg_name, options.label, options.user, options.group)
def main():
parser = prepare_parser()
options, args = parser.parse_args()
if options.wmtest:
print("Setting to injection in cmswebtest : ", options.wmtesturl)
wma.testbed(options.wmtesturl)
for cfg_name in args:
wma.upload_to_couch(cfg_name, options.label, options.user, options.group)
def uploadConf(self, filePath, label, where):
labelInCouch = self.label + '_' + label
cacheName = filePath.split('/')[-1]
if self.testMode:
self.count += 1
print '\tFake upload to couch with label', labelInCouch
return self.count
else:
try:
from modules.wma import upload_to_couch
except:
print '\n\tUnable to find wmcontrol modules. Please include it in your python path\n'
print '\n\t QUIT\n'
sys.exit(-16)
if cacheName in self.couchCache:
print "Not re-uploading", filePath, "to", where, "for", label
cacheId = self.couchCache[cacheName]
else:
print "Loading", filePath, "to", where, "for", label
cacheId = upload_to_couch(filePath,
labelInCouch,
self.user,
self.group,
test_mode=False,
url=where)
self.couchCache[cacheName] = cacheId
return cacheId
def uploadConf(self,filePath,label,where):
labelInCouch=self.label+'_'+label
cacheName=filePath.split('/')[-1]
if self.testMode:
self.count+=1
print '\tFake upload of',filePath,'to couch with label',labelInCouch
return self.count
else:
try:
from modules.wma import upload_to_couch
except:
print '\n\tUnable to find wmcontrol modules. Please include it in your python path\n'
print '\n\t QUIT\n'
sys.exit(-16)
if cacheName in self.couchCache:
print "Not re-uploading",filePath,"to",where,"for",label
cacheId=self.couchCache[cacheName]
else:
print "Loading",filePath,"to",where,"for",label
cacheId=upload_to_couch(filePath,
labelInCouch,
self.user,
self.group,
test_mode=False,
url=where
)
self.couchCache[cacheName]=cacheId
return cacheId
def upload_to_couch_oneArg(arguments):
from modules.wma import upload_to_couch
(filePath,labelInCouch,user,group,where) = arguments
cacheId=upload_to_couch(filePath,
labelInCouch,
user,
group,
test_mode=False,
url=where)
return cacheId
def upload_to_couch_oneArg(arguments):
from modules.wma import upload_to_couch
(filePath, labelInCouch, user, group, where) = arguments
cacheId = upload_to_couch(filePath,
labelInCouch,
user,
group,
test_mode=False,
url=where)
return cacheId
def build_params_dict(section,cfg):
global couch_pass
'''
Build the parameters dictionary for the request. Assumes the presence of an input .conf file or commandline.
For the moment the defaults of the parameters are stored here.
Put a dictionary on top?
'''
#wm testing
wmtest = cfg.get_param('wmtest', False, section)
url_dict = cfg.get_param('url_dict',"",section)
# fetch some important parameters
#this trick is to make the uniformation smoother and be able to read old cfgfiles
doc_id = step1_docID = ''
doc_id = step1_docID = cfg.get_param('docID','',section)
dummy = cfg.get_param('step1_docID','',section)
if dummy!='':
doc_id = step1_docID = dummy
step2_docid = cfg.get_param('step2_docID','',section)
#print step2_docid
step3_docid = cfg.get_param('step3_docID','',section)
#print step3_docid
# elaborate the file containing the name docid pairs
cfg_db_file = cfg.get_param('cfg_db_file','',section)
#print cfg_db_file
cfg_docid_dict = make_cfg_docid_dict(cfg_db_file)
release = cfg.get_param('release','',section)
globaltag = cfg.get_param('globaltag','',section)
pileup_dataset = cfg.get_param('pu_dataset','',section)
primary_dataset = cfg.get_param('primary_dataset','',section)
filter_eff = cfg.get_param('filter_eff','',section)
if not filter_eff:
filter_eff = 1.0
number_events = int(cfg.get_param('number_events',0,section))
#number_events = cfg.get_param('number_events',0,section)
version = cfg.get_param('version','',section)
##new values for renewed Request Agent
time_event = float(cfg.get_param('time_event',20,section))
size_memory = int(float(cfg.get_param('size_memory',2300,section)))
size_event = int(float(cfg.get_param('size_event',2000,section)))
if size_event <0:
size_event = 2000
# parameters with fallback
scramarch = cfg.get_param('scramarch',default_parameters['scramarch'],section)
#group = cfg.get_param('group',default_parameters['group'],section)
#requestor = cfg.get_param('requestor',default_parameters['requestor'],section)
identifier = cfg.get_param('identifier','',section)
dbsurl = cfg.get_param('dbsurl',default_parameters['dbsurl'],section)
includeparents=cfg.get_param('includeparents',default_parameters['includeparents'],section)
req_name=cfg.get_param('req_name','',section)
process_string = cfg.get_param('process_string','',section)
processing_string = cfg.get_param('processing_string','',section)
batch = cfg.get_param('batch','',section)
open_running_timeout = int(float(cfg.get_param('open_running_timeout','43200',section))) # 12h is legacy
# for the user and group
user,group = get_user_group(cfg,section)
# for the skims
skim_cfg = cfg.get_param('skim_cfg','',section)
skim_docid = cfg.get_param('skim_docID','',section)
skim_name = cfg.get_param('skim_name','',section)
skim_input = cfg.get_param('skim_input','RECOoutput',section)
if not skim_docid and skim_cfg:
if cfg_docid_dict.has_key(skim_cfg):
skim_docid=cfg_docid_dict[skim_cfg]
else:
skim_docid=wma.upload_to_couch(skim_cfg, section, user, group,test_mode)
# priority
priority = cfg.get_param('priority',default_parameters['priority'],section)
#blocks
blocks = cfg.get_param('blocks', [], section)
# Now the service ones
# Service
step1_cfg = cfg_path = ''
step1_cfg = cfg_path = cfg.get_param('cfg_path','',section)
dummy = cfg.get_param('step1_cfg','',section)
if dummy != '':
step1_cfg = cfg_path = dummy
harvest_cfg = cfg.get_param('harvest_cfg','',section)
harvest_docID = cfg.get_param('harvest_docID','',section)
step1_output = cfg.get_param('step1_output','',section)
keep_step1 = cfg.get_param('keep_step1',False,section)
step2_cfg = cfg.get_param('step2_cfg','',section)
step2_docID = cfg.get_param('step2_docID','',section)
step2_output = cfg.get_param('step2_output','',section)
keep_step2 = cfg.get_param('keep_step2',False,section)
step3_cfg = cfg.get_param('step3_cfg','',section)
step3_docID = cfg.get_param('step3_docID','',section)
step3_output = cfg.get_param('step3_output','',section)
transient_output = cfg.get_param('transient_output',[],section)
request_type = cfg.get_param('request_type',default_parameters['request_type'],section)
request_id = cfg.get_param('request_id','',section)
events_per_job = cfg.get_param('events_per_job','',section)
events_per_lumi = int(float(cfg.get_param('events_per_lumi',100,section))) # 100 is legacy
force_lumis = cfg.get_param('force_lumis', False, section)
brute_force = cfg.get_param('brute_force', False, section)
margin = cfg.get_param('margin', 0.05, section)
# Upload to couch if needed or check in the cfg dict if there
docIDs=[step1_docID,step2_docID,step3_docID]
cfgs=[step1_cfg,step2_cfg,step3_cfg]
for step in xrange(3):
step_cfg_name= cfgs[step]
step_docid = docIDs[step]
#print step_cfg_name, step_docid
if step_cfg_name!='' and step_docid=='' :
#print step_cfg_name, step_docid
# try to see if it is in the cfg name dict
if cfg_docid_dict.has_key(step_cfg_name):
print "Using the one in the cfg-docid dictionary."
docIDs[step] = cfg_docid_dict[step_cfg_name]
else:
print "No DocId found for section %s. Uploading the cfg to the couch." %section
docIDs[step]= wma.upload_to_couch(step_cfg_name, section, user, group,test_mode)
step1_docID,step2_docID,step3_docID=docIDs
if harvest_docID=='' and harvest_cfg!='':
harvest_docID= wma.upload_to_couch(harvest_cfg , section, user, group,test_mode)
# check if the request is valid
if step1_docID=='' and url_dict=="":
print "Invalid request, no docID configuration specified."
sys.stderr.write("[wmcontrol exception] Invalid request, no docID configuration specified.")
sys.exit(-1)
# Extract Campaign from PREP-ID if necessary
campaign = cfg.get_param('campaign','',section)
if campaign =="" and request_id=="":
print "Campaign and request-id are not set. Provide at least the Campaign."
elif campaign =="" and request_id!="":
campaign = re.match(".*-(.*)-.*",request_id).group(1)
elif campaign !="" and request_id!="":
print "Campaign and request-id are set. Using %s as campaign." %campaign
time_per_campaign=wma.time_per_events(campaign)
if time_per_campaign:
time_event=time_per_campaign
service_params={"section": section,
"version": version,
"request_type": request_type,
"step1_cfg": step1_cfg,
"step1_output": step1_output,
"keep_step1":keep_step1,
#
"step2_cfg": step2_cfg,
"step2_output": step2_output,
"keep_step2":keep_step2,
#
"step3_cfg": step3_cfg,
"step3_output": step3_output,
#
'cfg_docid_dict' : cfg_docid_dict,
'req_name': req_name,
"batch": batch,
"process_string": process_string,
'force_lumis': force_lumis,
'brute_force': brute_force,
'margin': margin
}
# According to the rerquest type, cook a request!
params={"CMSSWVersion": release,
"ScramArch": scramarch,
"RequestPriority": priority,
"RunWhitelist": ['Will Be replaced'],
"InputDataset": 'Will Be replaced',
"RunBlacklist": [],
"BlockWhitelist": blocks,
"BlockBlacklist": [],
"DbsUrl": dbsurl,
"RequestType": request_type,
"GlobalTag": globaltag,
"inputMode": "couchDB",
"RequestString": "Will Be dynamically created",
"Group": group,
"Requestor": user,
"Campaign": campaign,
"Memory": size_memory,
"SizePerEvent": size_event,
"TimePerEvent": time_event,
"OpenRunningTimeout" : open_running_timeout,
#"ConfigCacheUrl": wma.COUCH_DB_ADDRESS,
#"EnableHarvesting" : False
"ProcessingString": processing_string
}
if url_dict != "":
#print "This is the url",url_dict,"to get the dict from"
params = json.loads(os.popen('curl -s --insecure %s'%(url_dict)).read())
#print params
service_params["request_type"] = params["RequestType"]
service_params["version"] = params["ProcessingVersion"]
service_params["process_string"] = "T"
service_params["pid"] = params["RequestString"]
params["DbsUrl"] = "https://"+wma.WMAGENT_URL+wma.DBS3_URL
params["CouchURL"] = wma.COUCH_DB_ADDRESS
params["ConfigCacheURL"] = wma.COUCH_DB_ADDRESS
elif request_type == "ReReco":
if number_events:
if blocks:
## cannot perform automatic block selection
print "\n\n\n WARNING number_events is not functionnal because you specified blocks in input\n\n\n"
else:
print "\n\n\n WARNING automated block selection performed \n\n\n"
params.update({"RequestNumEvents" : number_events})
params.update({"ConfigCacheID": step1_docID,
"Scenario": "pp",
"IncludeParents" : includeparents,
"TransientOutputModules":transient_output})
if skim_docid != '':
print "This is a skim"
params.update({"SkimName1" : skim_name,
"SkimInput1" : skim_input,
"Skim1ConfigCacheID" : skim_docid,
"nskims" : 1})
elif request_type == 'MonteCarlo':
params.update({"RequestString": identifier,
"FirstEvent": 1,
"FirstLumi": 1,
"TimePerEvent": time_event,
"FilterEfficiency": filter_eff,
"LheInputFiles" : cfg.get_param('lhe_input',False,section),
"RequestNumEvents": number_events,
"ConfigCacheID": step1_docID,
"PrimaryDataset": primary_dataset,
"PrepID": request_id,
}
)
events_per_lumi = int(float( events_per_lumi ) / float(filter_eff))
params.update({
"EventsPerLumi" : events_per_lumi,
})
if wmtest:
params.pop("EventsPerLumi")
if params["LheInputFiles"]=='True' or params["LheInputFiles"]==True:
#max out to 500K for "lhe step zero"
print "Setting events per job here !!!!",type(params["LheInputFiles"]),params["LheInputFiles"]
events_per_job=500000
if wmtest:
events_per_job=15000
if events_per_job and int(events_per_job):
params.update({
"EventsPerJob" : int(events_per_job)
})
params.pop('BlockBlacklist')
params.pop('BlockWhitelist')
params.pop('InputDataset')
params.pop('RunBlacklist')
elif request_type == 'MonteCarloFromGEN':
params.update({"TimePerEvent": time_event,
"FilterEfficiency": filter_eff,
"ConfigCacheID": step1_docID,
"PrepID": request_id,
"TotalTime": 28800 })
if primary_dataset:
params.update({"PrimaryDataset": primary_dataset})
if int(number_events):
params.update({"RequestNumEvents": number_events})
elif request_type == 'LHEStepZero':
params.update({"RequestString": identifier,
"TimePerEvent": time_event,
"FirstEvent": 1,
"FirstLumi": 1,
"LheInputFiles" : cfg.get_param('lhe_input',False,section),
"Memory": 2300,
"SizePerEvent": size_event,
"ConfigCacheID": step1_docID,
"RequestNumEvents": number_events,
"PrimaryDataset": primary_dataset,
"PrepID": request_id,
"TotalTime": 28800 ,
"EventsPerLumi":300,
"ProdJobSplitAlgo" : "EventBased",
"ProdJobSplitArgs" : {"events_per_job": int(events_per_job),"events_per_lumi": int(events_per_lumi)}
})
params.pop('BlockBlacklist')
params.pop('BlockWhitelist')
params.pop('InputDataset')
params.pop('RunBlacklist')
params.pop('RunWhitelist')
elif request_type == 'ReDigi':
if number_events:
if blocks:
## cannot perform automatic block selection
print "\n\n\n WARNING number_events is not functionnal because you specified blocks in input\n\n\n"
else:
print "\n\n\n WARNING automated block selection performed \n\n\n"
params.update({"RequestNumEvents" : number_events})
params.update({"RequestString": identifier,
"StepOneConfigCacheID": step1_docID,
"KeepStepOneOutput": keep_step1,
"StepOneOutputModuleName": step1_output,
#"DataPileup": "",
"MCPileup": pileup_dataset,
#"Scenario": "pp",
"PrepID": request_id})
if primary_dataset:
params.update({"PrimaryDataset": primary_dataset})
if step2_cfg != '' or step2_docID !='':
params.update({"StepTwoConfigCacheID": step2_docID,
"KeepStepTwoOutput": keep_step2,
"StepTwoOutputModuleName": step2_output})
if step3_cfg !='' or step3_docID!='':
params['StepThreeConfigCacheID'] = step3_docID
else:
if not keep_step2:
print 'Request not keeping its step 2 output'
raise Exception("The request has a second step, no third step and not keeping it's second output")
else:
if not keep_step1:
print 'Request not keeping anything'
raise Exception('The request has one step and not keeping anything')
elif request_type == 'TaskChain':
params.pop('RunBlacklist')
params.pop('BlockWhitelist')
params.pop('BlockBlacklist')
task1_dict={'SplittingAlgorithm': 'LumiBased',
'SplittingArguments': {'lumis_per_job': 8},
'TaskName':'Task1'
}
task1_dict['GlobalTag'] = cfg.get_param('step1_globaltag',globaltag,section)
task1_dict['ConfigCacheID'] = step1_docID
task1_dict['KeepOutput'] = keep_step1
params['Task1']=task1_dict
params['TaskChain']=1
if step2_cfg or step2_docID:
task2_dict={'SplittingAlgorithm': 'LumiBased',
'SplittingArguments': {'lumis_per_job': 8},
'TaskName':'Task2'
}
task2_dict['GlobalTag'] = cfg.get_param('step2_globaltag',globaltag,section)
task2_dict['ConfigCacheID'] = step2_docID
task2_dict['InputFromOutputModule'] = step2_output
task2_dict['InputTask'] = cfg.get_param('step2_input','Task1',section)
#task2_dict['KeepOutput'] = keep_step2
params['Task2']=task2_dict
params['TaskChain']=2
if step3_cfg or step3_docID:
task3_dict={'SplittingAlgorithm': 'LumiBased',
'SplittingArguments': {'lumis_per_job': 8},
'TaskName':'Task3'
}
task3_dict['GlobalTag'] = cfg.get_param('step3_globaltag',globaltag,section)
task3_dict['ConfigCacheID'] = step3_docID
task3_dict['InputFromOutputModule'] = step3_output
task3_dict['InputTask'] = cfg.get_param('step3_input','Task2',section)
#task3_dict['KeepOutput'] = keep_step3
params['Task3']=task3_dict
params['TaskChain']=3
#from pprint import pformat
#print "\n current dictionnary \n",pformat(params),'\n\n'
###raise Exception('Unknown request type, aborting')
else:
print "Request type chose: "+str(request_type)
raise Exception('Unknown request type, aborting')
if harvest_docID:
##setup automatic harvesting
params.update({"EnableDQMHarvest" : 1,
"DQMUploadUrl" : "https://cmsweb.cern.ch/dqm/offline",
"DQMConfigCacheID" : harvest_docID})
## pop any empty parameters
for (param,value) in params.items():
if value in ["",[]]:
params.pop(param)
return params,service_params
