| 1 |
spiga |
1.1 |
import common
|
| 2 |
|
|
from crab_logger import Logger
|
| 3 |
|
|
from crab_exceptions import *
|
| 4 |
|
|
from crab_util import *
|
| 5 |
|
|
from WMCore.SiteScreening.BlackWhiteListParser import SEBlackWhiteListParser
|
| 6 |
|
|
|
| 7 |
|
|
class JobSplitter:
|
| 8 |
|
|
def __init__( self, cfg_params, args ):
|
| 9 |
|
|
self.cfg_params = cfg_params
|
| 10 |
spiga |
1.3 |
self.args=args
|
| 11 |
spiga |
1.1 |
#self.maxEvents
|
| 12 |
|
|
# init BlackWhiteListParser
|
| 13 |
|
|
seWhiteList = cfg_params.get('EDG.se_white_list',[])
|
| 14 |
|
|
seBlackList = cfg_params.get('EDG.se_black_list',[])
|
| 15 |
|
|
self.blackWhiteListParser = SEBlackWhiteListParser(seWhiteList, seBlackList, common.logger)
|
| 16 |
|
|
|
| 17 |
|
|
|
| 18 |
|
|
def checkUserSettings(self):
|
| 19 |
|
|
## Events per job
|
| 20 |
|
|
if self.cfg_params.has_key('CMSSW.events_per_job'):
|
| 21 |
|
|
self.eventsPerJob =int( self.cfg_params['CMSSW.events_per_job'])
|
| 22 |
|
|
self.selectEventsPerJob = 1
|
| 23 |
|
|
else:
|
| 24 |
|
|
self.eventsPerJob = -1
|
| 25 |
|
|
self.selectEventsPerJob = 0
|
| 26 |
|
|
|
| 27 |
|
|
## number of jobs
|
| 28 |
|
|
if self.cfg_params.has_key('CMSSW.number_of_jobs'):
|
| 29 |
|
|
self.theNumberOfJobs =int( self.cfg_params['CMSSW.number_of_jobs'])
|
| 30 |
|
|
self.selectNumberOfJobs = 1
|
| 31 |
|
|
else:
|
| 32 |
|
|
self.theNumberOfJobs = 0
|
| 33 |
|
|
self.selectNumberOfJobs = 0
|
| 34 |
|
|
|
| 35 |
|
|
if self.cfg_params.has_key('CMSSW.total_number_of_events'):
|
| 36 |
|
|
self.total_number_of_events = int(self.cfg_params['CMSSW.total_number_of_events'])
|
| 37 |
|
|
self.selectTotalNumberEvents = 1
|
| 38 |
|
|
if self.selectNumberOfJobs == 1:
|
| 39 |
|
|
if (self.total_number_of_events != -1) and int(self.total_number_of_events) < int(self.theNumberOfJobs):
|
| 40 |
|
|
msg = 'Must specify at least one event per job. total_number_of_events > number_of_jobs '
|
| 41 |
|
|
raise CrabException(msg)
|
| 42 |
|
|
else:
|
| 43 |
|
|
self.total_number_of_events = 0
|
| 44 |
|
|
self.selectTotalNumberEvents = 0
|
| 45 |
|
|
|
| 46 |
|
|
|
| 47 |
|
|
########################################################################
|
| 48 |
|
|
def jobSplittingByEvent( self ):
|
| 49 |
|
|
"""
|
| 50 |
|
|
Perform job splitting. Jobs run over an integer number of files
|
| 51 |
|
|
and no more than one block.
|
| 52 |
|
|
ARGUMENT: blockSites: dictionary with blocks as keys and list of host sites as values
|
| 53 |
|
|
REQUIRES: self.selectTotalNumberEvents, self.selectEventsPerJob, self.selectNumberofJobs,
|
| 54 |
|
|
self.total_number_of_events, self.eventsPerJob, self.theNumberOfJobs,
|
| 55 |
|
|
self.maxEvents, self.filesbyblock
|
| 56 |
spiga |
1.3 |
SETS: jobDestination - Site destination(s) for each job (a list of lists)
|
| 57 |
spiga |
1.1 |
self.total_number_of_jobs - Total # of jobs
|
| 58 |
|
|
self.list_of_args - File(s) job will run on (a list of lists)
|
| 59 |
|
|
"""
|
| 60 |
|
|
|
| 61 |
spiga |
1.3 |
jobDestination=[]
|
| 62 |
spiga |
1.1 |
self.checkUserSettings()
|
| 63 |
|
|
if ( (self.selectTotalNumberEvents + self.selectEventsPerJob + self.selectNumberOfJobs) != 2 ):
|
| 64 |
|
|
msg = 'Must define exactly two of total_number_of_events, events_per_job, or number_of_jobs.'
|
| 65 |
|
|
raise CrabException(msg)
|
| 66 |
|
|
|
| 67 |
spiga |
1.4 |
blockSites = self.args['blockSites']
|
| 68 |
|
|
pubdata = self.args['pubdata']
|
| 69 |
spiga |
1.3 |
filesbyblock=pubdata.getFiles()
|
| 70 |
|
|
|
| 71 |
|
|
self.eventsbyblock=pubdata.getEventsPerBlock()
|
| 72 |
|
|
self.eventsbyfile=pubdata.getEventsPerFile()
|
| 73 |
|
|
self.parentFiles=pubdata.getParent()
|
| 74 |
spiga |
1.1 |
|
| 75 |
|
|
## get max number of events
|
| 76 |
spiga |
1.3 |
self.maxEvents=pubdata.getMaxEvents()
|
| 77 |
spiga |
1.1 |
|
| 78 |
|
|
self.useParent = int(self.cfg_params.get('CMSSW.use_parent',0))
|
| 79 |
|
|
noBboundary = int(self.cfg_params.get('CMSSW.no_block_boundary',0))
|
| 80 |
|
|
|
| 81 |
|
|
# ---- Handle the possible job splitting configurations ---- #
|
| 82 |
|
|
if (self.selectTotalNumberEvents):
|
| 83 |
|
|
totalEventsRequested = self.total_number_of_events
|
| 84 |
|
|
if (self.selectEventsPerJob):
|
| 85 |
|
|
eventsPerJobRequested = self.eventsPerJob
|
| 86 |
|
|
if (self.selectNumberOfJobs):
|
| 87 |
|
|
totalEventsRequested = self.theNumberOfJobs * self.eventsPerJob
|
| 88 |
|
|
|
| 89 |
|
|
# If user requested all the events in the dataset
|
| 90 |
|
|
if (totalEventsRequested == -1):
|
| 91 |
|
|
eventsRemaining=self.maxEvents
|
| 92 |
|
|
# If user requested more events than are in the dataset
|
| 93 |
|
|
elif (totalEventsRequested > self.maxEvents):
|
| 94 |
|
|
eventsRemaining = self.maxEvents
|
| 95 |
|
|
common.logger.message("Requested "+str(self.total_number_of_events)+ " events, but only "+str(self.maxEvents)+" events are available.")
|
| 96 |
|
|
# If user requested less events than are in the dataset
|
| 97 |
|
|
else:
|
| 98 |
|
|
eventsRemaining = totalEventsRequested
|
| 99 |
|
|
|
| 100 |
|
|
# If user requested more events per job than are in the dataset
|
| 101 |
|
|
if (self.selectEventsPerJob and eventsPerJobRequested > self.maxEvents):
|
| 102 |
|
|
eventsPerJobRequested = self.maxEvents
|
| 103 |
|
|
|
| 104 |
|
|
# For user info at end
|
| 105 |
|
|
totalEventCount = 0
|
| 106 |
|
|
|
| 107 |
|
|
if (self.selectTotalNumberEvents and self.selectNumberOfJobs):
|
| 108 |
|
|
eventsPerJobRequested = int(eventsRemaining/self.theNumberOfJobs)
|
| 109 |
|
|
|
| 110 |
|
|
if (self.selectNumberOfJobs):
|
| 111 |
|
|
common.logger.message("May not create the exact number_of_jobs requested.")
|
| 112 |
|
|
|
| 113 |
|
|
# old... to remove Daniele
|
| 114 |
|
|
totalNumberOfJobs = 999999999
|
| 115 |
|
|
|
| 116 |
spiga |
1.3 |
blocks = blockSites.keys()
|
| 117 |
spiga |
1.1 |
blockCount = 0
|
| 118 |
|
|
# Backup variable in case self.maxEvents counted events in a non-included block
|
| 119 |
|
|
numBlocksInDataset = len(blocks)
|
| 120 |
|
|
|
| 121 |
|
|
jobCount = 0
|
| 122 |
|
|
list_of_lists = []
|
| 123 |
|
|
|
| 124 |
|
|
# list tracking which jobs are in which jobs belong to which block
|
| 125 |
|
|
jobsOfBlock = {}
|
| 126 |
|
|
|
| 127 |
|
|
parString = ""
|
| 128 |
|
|
filesEventCount = 0
|
| 129 |
|
|
|
| 130 |
|
|
# ---- Iterate over the blocks in the dataset until ---- #
|
| 131 |
|
|
# ---- we've met the requested total # of events ---- #
|
| 132 |
|
|
while ( (eventsRemaining > 0) and (blockCount < numBlocksInDataset) and (jobCount < totalNumberOfJobs)):
|
| 133 |
|
|
block = blocks[blockCount]
|
| 134 |
|
|
blockCount += 1
|
| 135 |
|
|
if block not in jobsOfBlock.keys() :
|
| 136 |
|
|
jobsOfBlock[block] = []
|
| 137 |
|
|
|
| 138 |
|
|
if self.eventsbyblock.has_key(block) :
|
| 139 |
|
|
numEventsInBlock = self.eventsbyblock[block]
|
| 140 |
|
|
common.logger.debug(5,'Events in Block File '+str(numEventsInBlock))
|
| 141 |
|
|
|
| 142 |
spiga |
1.4 |
files = filesbyblock[block]
|
| 143 |
spiga |
1.1 |
numFilesInBlock = len(files)
|
| 144 |
|
|
if (numFilesInBlock <= 0):
|
| 145 |
|
|
continue
|
| 146 |
|
|
fileCount = 0
|
| 147 |
|
|
if noBboundary == 0: # DD
|
| 148 |
|
|
# ---- New block => New job ---- #
|
| 149 |
|
|
parString = ""
|
| 150 |
|
|
# counter for number of events in files currently worked on
|
| 151 |
|
|
filesEventCount = 0
|
| 152 |
|
|
# flag if next while loop should touch new file
|
| 153 |
|
|
newFile = 1
|
| 154 |
|
|
# job event counter
|
| 155 |
|
|
jobSkipEventCount = 0
|
| 156 |
|
|
|
| 157 |
|
|
# ---- Iterate over the files in the block until we've met the requested ---- #
|
| 158 |
|
|
# ---- total # of events or we've gone over all the files in this block ---- #
|
| 159 |
|
|
pString=''
|
| 160 |
|
|
while ( (eventsRemaining > 0) and (fileCount < numFilesInBlock) and (jobCount < totalNumberOfJobs) ):
|
| 161 |
|
|
file = files[fileCount]
|
| 162 |
|
|
if self.useParent==1:
|
| 163 |
|
|
parent = self.parentFiles[file]
|
| 164 |
|
|
for f in parent :
|
| 165 |
|
|
pString += '\\\"' + f + '\\\"\,'
|
| 166 |
|
|
common.logger.debug(6, "File "+str(file)+" has the following parents: "+str(parent))
|
| 167 |
|
|
common.logger.write("File "+str(file)+" has the following parents: "+str(parent))
|
| 168 |
|
|
if newFile :
|
| 169 |
|
|
try:
|
| 170 |
|
|
numEventsInFile = self.eventsbyfile[file]
|
| 171 |
|
|
common.logger.debug(6, "File "+str(file)+" has "+str(numEventsInFile)+" events")
|
| 172 |
|
|
# increase filesEventCount
|
| 173 |
|
|
filesEventCount += numEventsInFile
|
| 174 |
|
|
# Add file to current job
|
| 175 |
|
|
parString += '\\\"' + file + '\\\"\,'
|
| 176 |
|
|
newFile = 0
|
| 177 |
|
|
except KeyError:
|
| 178 |
|
|
common.logger.message("File "+str(file)+" has unknown number of events: skipping")
|
| 179 |
|
|
|
| 180 |
|
|
eventsPerJobRequested = min(eventsPerJobRequested, eventsRemaining)
|
| 181 |
|
|
# if less events in file remain than eventsPerJobRequested
|
| 182 |
|
|
if ( filesEventCount - jobSkipEventCount < eventsPerJobRequested):
|
| 183 |
|
|
if noBboundary == 1: ## DD
|
| 184 |
|
|
newFile = 1
|
| 185 |
|
|
fileCount += 1
|
| 186 |
|
|
else:
|
| 187 |
|
|
# if last file in block
|
| 188 |
|
|
if ( fileCount == numFilesInBlock-1 ) :
|
| 189 |
|
|
# end job using last file, use remaining events in block
|
| 190 |
|
|
# close job and touch new file
|
| 191 |
|
|
fullString = parString[:-2]
|
| 192 |
|
|
if self.useParent==1:
|
| 193 |
|
|
fullParentString = pString[:-2]
|
| 194 |
|
|
list_of_lists.append([fullString,fullParentString,str(-1),str(jobSkipEventCount)])
|
| 195 |
|
|
else:
|
| 196 |
|
|
list_of_lists.append([fullString,str(-1),str(jobSkipEventCount)])
|
| 197 |
|
|
common.logger.debug(3,"Job "+str(jobCount+1)+" can run over "+str(filesEventCount - jobSkipEventCount)+" events (last file in block).")
|
| 198 |
spiga |
1.3 |
jobDestination.append(blockSites[block])
|
| 199 |
|
|
common.logger.debug(5,"Job "+str(jobCount+1)+" Destination: "+str(jobDestination[jobCount]))
|
| 200 |
spiga |
1.1 |
# fill jobs of block dictionary
|
| 201 |
|
|
jobsOfBlock[block].append(jobCount+1)
|
| 202 |
|
|
# reset counter
|
| 203 |
|
|
jobCount = jobCount + 1
|
| 204 |
|
|
totalEventCount = totalEventCount + filesEventCount - jobSkipEventCount
|
| 205 |
|
|
eventsRemaining = eventsRemaining - filesEventCount + jobSkipEventCount
|
| 206 |
|
|
jobSkipEventCount = 0
|
| 207 |
|
|
# reset file
|
| 208 |
|
|
pString = ""
|
| 209 |
|
|
parString = ""
|
| 210 |
|
|
filesEventCount = 0
|
| 211 |
|
|
newFile = 1
|
| 212 |
|
|
fileCount += 1
|
| 213 |
|
|
else :
|
| 214 |
|
|
# go to next file
|
| 215 |
|
|
newFile = 1
|
| 216 |
|
|
fileCount += 1
|
| 217 |
|
|
# if events in file equal to eventsPerJobRequested
|
| 218 |
|
|
elif ( filesEventCount - jobSkipEventCount == eventsPerJobRequested ) :
|
| 219 |
|
|
# close job and touch new file
|
| 220 |
|
|
fullString = parString[:-2]
|
| 221 |
|
|
if self.useParent==1:
|
| 222 |
|
|
fullParentString = pString[:-2]
|
| 223 |
|
|
list_of_lists.append([fullString,fullParentString,str(eventsPerJobRequested),str(jobSkipEventCount)])
|
| 224 |
|
|
else:
|
| 225 |
|
|
list_of_lists.append([fullString,str(eventsPerJobRequested),str(jobSkipEventCount)])
|
| 226 |
|
|
common.logger.debug(3,"Job "+str(jobCount+1)+" can run over "+str(eventsPerJobRequested)+" events.")
|
| 227 |
spiga |
1.3 |
jobDestination.append(blockSites[block])
|
| 228 |
|
|
common.logger.debug(5,"Job "+str(jobCount+1)+" Destination: "+str(jobDestination[jobCount]))
|
| 229 |
spiga |
1.1 |
jobsOfBlock[block].append(jobCount+1)
|
| 230 |
|
|
# reset counter
|
| 231 |
|
|
jobCount = jobCount + 1
|
| 232 |
|
|
totalEventCount = totalEventCount + eventsPerJobRequested
|
| 233 |
|
|
eventsRemaining = eventsRemaining - eventsPerJobRequested
|
| 234 |
|
|
jobSkipEventCount = 0
|
| 235 |
|
|
# reset file
|
| 236 |
|
|
pString = ""
|
| 237 |
|
|
parString = ""
|
| 238 |
|
|
filesEventCount = 0
|
| 239 |
|
|
newFile = 1
|
| 240 |
|
|
fileCount += 1
|
| 241 |
|
|
|
| 242 |
|
|
# if more events in file remain than eventsPerJobRequested
|
| 243 |
|
|
else :
|
| 244 |
|
|
# close job but don't touch new file
|
| 245 |
|
|
fullString = parString[:-2]
|
| 246 |
|
|
if self.useParent==1:
|
| 247 |
|
|
fullParentString = pString[:-2]
|
| 248 |
|
|
list_of_lists.append([fullString,fullParentString,str(eventsPerJobRequested),str(jobSkipEventCount)])
|
| 249 |
|
|
else:
|
| 250 |
|
|
list_of_lists.append([fullString,str(eventsPerJobRequested),str(jobSkipEventCount)])
|
| 251 |
|
|
common.logger.debug(3,"Job "+str(jobCount+1)+" can run over "+str(eventsPerJobRequested)+" events.")
|
| 252 |
spiga |
1.3 |
jobDestination.append(blockSites[block])
|
| 253 |
|
|
common.logger.debug(5,"Job "+str(jobCount+1)+" Destination: "+str(jobDestination[jobCount]))
|
| 254 |
spiga |
1.1 |
jobsOfBlock[block].append(jobCount+1)
|
| 255 |
|
|
# increase counter
|
| 256 |
|
|
jobCount = jobCount + 1
|
| 257 |
|
|
totalEventCount = totalEventCount + eventsPerJobRequested
|
| 258 |
|
|
eventsRemaining = eventsRemaining - eventsPerJobRequested
|
| 259 |
|
|
# calculate skip events for last file
|
| 260 |
|
|
# use filesEventCount (contains several files), jobSkipEventCount and eventsPerJobRequest
|
| 261 |
|
|
jobSkipEventCount = eventsPerJobRequested - (filesEventCount - jobSkipEventCount - self.eventsbyfile[file])
|
| 262 |
|
|
# remove all but the last file
|
| 263 |
|
|
filesEventCount = self.eventsbyfile[file]
|
| 264 |
|
|
if self.useParent==1:
|
| 265 |
|
|
for f in parent : pString += '\\\"' + f + '\\\"\,'
|
| 266 |
|
|
parString = '\\\"' + file + '\\\"\,'
|
| 267 |
|
|
pass # END if
|
| 268 |
|
|
pass # END while (iterate over files in the block)
|
| 269 |
|
|
pass # END while (iterate over blocks in the dataset)
|
| 270 |
|
|
self.ncjobs = self.total_number_of_jobs = jobCount
|
| 271 |
|
|
if (eventsRemaining > 0 and jobCount < totalNumberOfJobs ):
|
| 272 |
|
|
common.logger.message("Could not run on all requested events because some blocks not hosted at allowed sites.")
|
| 273 |
|
|
common.logger.message(str(jobCount)+" job(s) can run on "+str(totalEventCount)+" events.\n")
|
| 274 |
|
|
|
| 275 |
|
|
# skip check on block with no sites DD
|
| 276 |
spiga |
1.5 |
if noBboundary == 0 : self.checkBlockNoSite(blocks,jobsOfBlock,blockSites)
|
| 277 |
spiga |
1.1 |
|
| 278 |
|
|
# prepare dict output
|
| 279 |
|
|
dictOut = {}
|
| 280 |
|
|
dictOut['args'] = list_of_lists
|
| 281 |
spiga |
1.3 |
dictOut['jobDestination'] = jobDestination
|
| 282 |
spiga |
1.1 |
dictOut['njobs']=self.total_number_of_jobs
|
| 283 |
|
|
|
| 284 |
|
|
return dictOut
|
| 285 |
|
|
|
| 286 |
|
|
# keep trace of block with no sites to print a warning at the end
|
| 287 |
|
|
|
| 288 |
spiga |
1.5 |
def checkBlockNoSite(self,blocks,jobsOfBlock,blockSites):
|
| 289 |
spiga |
1.1 |
# screen output
|
| 290 |
|
|
screenOutput = "List of jobs and available destination sites:\n\n"
|
| 291 |
|
|
noSiteBlock = []
|
| 292 |
|
|
bloskNoSite = []
|
| 293 |
|
|
|
| 294 |
|
|
blockCounter = 0
|
| 295 |
|
|
for block in blocks:
|
| 296 |
|
|
if block in jobsOfBlock.keys() :
|
| 297 |
|
|
blockCounter += 1
|
| 298 |
|
|
screenOutput += "Block %5i: jobs %20s: sites: %s\n" % (blockCounter,spanRanges(jobsOfBlock[block]),
|
| 299 |
spiga |
1.3 |
','.join(self.blackWhiteListParser.checkWhiteList(self.blackWhiteListParser.checkBlackList(blockSites[block],[block]),[block])))
|
| 300 |
|
|
if len(self.blackWhiteListParser.checkWhiteList(self.blackWhiteListParser.checkBlackList(blockSites[block],[block]),[block])) == 0:
|
| 301 |
spiga |
1.1 |
noSiteBlock.append( spanRanges(jobsOfBlock[block]) )
|
| 302 |
|
|
bloskNoSite.append( blockCounter )
|
| 303 |
|
|
|
| 304 |
|
|
common.logger.message(screenOutput)
|
| 305 |
|
|
if len(noSiteBlock) > 0 and len(bloskNoSite) > 0:
|
| 306 |
|
|
msg = 'WARNING: No sites are hosting any part of data for block:\n '
|
| 307 |
|
|
virgola = ""
|
| 308 |
|
|
if len(bloskNoSite) > 1:
|
| 309 |
|
|
virgola = ","
|
| 310 |
|
|
for block in bloskNoSite:
|
| 311 |
|
|
msg += ' ' + str(block) + virgola
|
| 312 |
|
|
msg += '\n Related jobs:\n '
|
| 313 |
|
|
virgola = ""
|
| 314 |
|
|
if len(noSiteBlock) > 1:
|
| 315 |
|
|
virgola = ","
|
| 316 |
|
|
for range_jobs in noSiteBlock:
|
| 317 |
|
|
msg += str(range_jobs) + virgola
|
| 318 |
|
|
msg += '\n will not be submitted and this block of data can not be analyzed!\n'
|
| 319 |
|
|
if self.cfg_params.has_key('EDG.se_white_list'):
|
| 320 |
|
|
msg += 'WARNING: SE White List: '+self.cfg_params['EDG.se_white_list']+'\n'
|
| 321 |
|
|
msg += '(Hint: By whitelisting you force the job to run at this particular site(s).\n'
|
| 322 |
|
|
msg += 'Please check if the dataset is available at this site!)\n'
|
| 323 |
|
|
if self.cfg_params.has_key('EDG.ce_white_list'):
|
| 324 |
|
|
msg += 'WARNING: CE White List: '+self.cfg_params['EDG.ce_white_list']+'\n'
|
| 325 |
|
|
msg += '(Hint: By whitelisting you force the job to run at this particular site(s).\n'
|
| 326 |
|
|
msg += 'Please check if the dataset is available at this site!)\n'
|
| 327 |
|
|
|
| 328 |
|
|
common.logger.message(msg)
|
| 329 |
|
|
|
| 330 |
|
|
return
|
| 331 |
|
|
|
| 332 |
|
|
|
| 333 |
|
|
########################################################################
|
| 334 |
|
|
def jobSplittingByRun(self):
|
| 335 |
|
|
"""
|
| 336 |
|
|
"""
|
| 337 |
|
|
from sets import Set
|
| 338 |
|
|
from WMCore.JobSplitting.RunBased import RunBased
|
| 339 |
|
|
from WMCore.DataStructs.Workflow import Workflow
|
| 340 |
|
|
from WMCore.DataStructs.File import File
|
| 341 |
|
|
from WMCore.DataStructs.Fileset import Fileset
|
| 342 |
|
|
from WMCore.DataStructs.Subscription import Subscription
|
| 343 |
|
|
from WMCore.JobSplitting.SplitterFactory import SplitterFactory
|
| 344 |
|
|
from WMCore.DataStructs.Run import Run
|
| 345 |
|
|
|
| 346 |
|
|
self.checkUserSettings()
|
| 347 |
spiga |
1.4 |
blockSites = self.args['blockSites']
|
| 348 |
|
|
pubdata = self.args['pubdata']
|
| 349 |
spiga |
1.1 |
|
| 350 |
|
|
if self.selectNumberOfJobs == 0 :
|
| 351 |
|
|
self.theNumberOfJobs = 9999999
|
| 352 |
|
|
blocks = {}
|
| 353 |
|
|
runList = []
|
| 354 |
|
|
thefiles = Fileset(name='FilesToSplit')
|
| 355 |
spiga |
1.3 |
fileList = pubdata.getListFiles()
|
| 356 |
spiga |
1.1 |
for f in fileList:
|
| 357 |
|
|
block = f['Block']['Name']
|
| 358 |
|
|
try:
|
| 359 |
spiga |
1.3 |
f['Block']['StorageElementList'].extend(blockSites[block])
|
| 360 |
spiga |
1.1 |
except:
|
| 361 |
|
|
continue
|
| 362 |
|
|
wmbsFile = File(f['LogicalFileName'])
|
| 363 |
spiga |
1.3 |
[ wmbsFile['locations'].add(x) for x in blockSites[block] ]
|
| 364 |
spiga |
1.1 |
wmbsFile['block'] = block
|
| 365 |
|
|
runNum = f['RunsList'][0]['RunNumber']
|
| 366 |
|
|
runList.append(runNum)
|
| 367 |
|
|
myRun = Run(runNumber=runNum)
|
| 368 |
|
|
wmbsFile.addRun( myRun )
|
| 369 |
|
|
thefiles.addFile(
|
| 370 |
|
|
wmbsFile
|
| 371 |
|
|
)
|
| 372 |
|
|
|
| 373 |
|
|
work = Workflow()
|
| 374 |
|
|
subs = Subscription(
|
| 375 |
|
|
fileset = thefiles,
|
| 376 |
|
|
workflow = work,
|
| 377 |
|
|
split_algo = 'RunBased',
|
| 378 |
|
|
type = "Processing")
|
| 379 |
|
|
splitter = SplitterFactory()
|
| 380 |
|
|
jobfactory = splitter(subs)
|
| 381 |
|
|
|
| 382 |
|
|
#loop over all runs
|
| 383 |
|
|
set = Set(runList)
|
| 384 |
|
|
list_of_lists = []
|
| 385 |
|
|
jobDestination = []
|
| 386 |
|
|
|
| 387 |
|
|
count = 0
|
| 388 |
|
|
for i in list(set):
|
| 389 |
|
|
if count < self.theNumberOfJobs:
|
| 390 |
|
|
res = self.getJobInfo(jobfactory())
|
| 391 |
|
|
parString = ''
|
| 392 |
|
|
for file in res['lfns']:
|
| 393 |
|
|
parString += '\\\"' + file + '\\\"\,'
|
| 394 |
|
|
fullString = parString[:-2]
|
| 395 |
|
|
list_of_lists.append([fullString,str(-1),str(0)])
|
| 396 |
spiga |
1.2 |
#need to check single file location
|
| 397 |
spiga |
1.1 |
jobDestination.append(res['locations'])
|
| 398 |
|
|
count +=1
|
| 399 |
|
|
# prepare dict output
|
| 400 |
|
|
dictOut = {}
|
| 401 |
|
|
dictOut['args'] = list_of_lists
|
| 402 |
|
|
dictOut['jobDestination'] = jobDestination
|
| 403 |
|
|
dictOut['njobs']=count
|
| 404 |
|
|
|
| 405 |
|
|
return dictOut
|
| 406 |
|
|
|
| 407 |
|
|
def getJobInfo( self,jobGroup ):
|
| 408 |
|
|
res = {}
|
| 409 |
|
|
lfns = []
|
| 410 |
|
|
locations = []
|
| 411 |
|
|
tmp_check=0
|
| 412 |
|
|
for job in jobGroup.jobs:
|
| 413 |
|
|
for file in job.getFiles():
|
| 414 |
|
|
lfns.append(file['lfn'])
|
| 415 |
|
|
for loc in file['locations']:
|
| 416 |
|
|
if tmp_check < 1 :
|
| 417 |
|
|
locations.append(loc)
|
| 418 |
spiga |
1.8 |
tmp_check = tmp_check + 1
|
| 419 |
spiga |
1.1 |
### qui va messo il check per la locations
|
| 420 |
|
|
res['lfns'] = lfns
|
| 421 |
|
|
res['locations'] = locations
|
| 422 |
|
|
return res
|
| 423 |
|
|
|
| 424 |
|
|
########################################################################
|
| 425 |
|
|
def jobSplittingNoInput(self):
|
| 426 |
|
|
"""
|
| 427 |
|
|
Perform job splitting based on number of event per job
|
| 428 |
|
|
"""
|
| 429 |
|
|
common.logger.debug(5,'Splitting per events')
|
| 430 |
spiga |
1.3 |
self.checkUserSettings()
|
| 431 |
|
|
jobDestination=[]
|
| 432 |
spiga |
1.6 |
if ( (self.selectTotalNumberEvents + self.selectEventsPerJob + self.selectNumberOfJobs) != 2 ):
|
| 433 |
|
|
msg = 'Must define exactly two of total_number_of_events, events_per_job, or number_of_jobs.'
|
| 434 |
spiga |
1.3 |
raise CrabException(msg)
|
| 435 |
|
|
|
| 436 |
|
|
managedGenerators =self.args['managedGenerators']
|
| 437 |
|
|
generator = self.args['generator']
|
| 438 |
|
|
firstRun = self.cfg_params.get('CMSSW.first_run',None)
|
| 439 |
spiga |
1.1 |
|
| 440 |
|
|
if (self.selectEventsPerJob):
|
| 441 |
|
|
common.logger.message('Required '+str(self.eventsPerJob)+' events per job ')
|
| 442 |
|
|
if (self.selectNumberOfJobs):
|
| 443 |
|
|
common.logger.message('Required '+str(self.theNumberOfJobs)+' jobs in total ')
|
| 444 |
|
|
if (self.selectTotalNumberEvents):
|
| 445 |
|
|
common.logger.message('Required '+str(self.total_number_of_events)+' events in total ')
|
| 446 |
|
|
|
| 447 |
|
|
if (self.total_number_of_events < 0):
|
| 448 |
|
|
msg='Cannot split jobs per Events with "-1" as total number of events'
|
| 449 |
|
|
raise CrabException(msg)
|
| 450 |
|
|
|
| 451 |
|
|
if (self.selectEventsPerJob):
|
| 452 |
|
|
if (self.selectTotalNumberEvents):
|
| 453 |
|
|
self.total_number_of_jobs = int(self.total_number_of_events/self.eventsPerJob)
|
| 454 |
|
|
elif(self.selectNumberOfJobs) :
|
| 455 |
|
|
self.total_number_of_jobs =self.theNumberOfJobs
|
| 456 |
|
|
self.total_number_of_events =int(self.theNumberOfJobs*self.eventsPerJob)
|
| 457 |
|
|
|
| 458 |
|
|
elif (self.selectNumberOfJobs) :
|
| 459 |
|
|
self.total_number_of_jobs = self.theNumberOfJobs
|
| 460 |
|
|
self.eventsPerJob = int(self.total_number_of_events/self.total_number_of_jobs)
|
| 461 |
|
|
|
| 462 |
|
|
common.logger.debug(5,'N jobs '+str(self.total_number_of_jobs))
|
| 463 |
|
|
|
| 464 |
|
|
# is there any remainder?
|
| 465 |
|
|
check = int(self.total_number_of_events) - (int(self.total_number_of_jobs)*self.eventsPerJob)
|
| 466 |
|
|
|
| 467 |
|
|
common.logger.debug(5,'Check '+str(check))
|
| 468 |
|
|
|
| 469 |
|
|
common.logger.message(str(self.total_number_of_jobs)+' jobs can be created, each for '+str(self.eventsPerJob)+' for a total of '+str(self.total_number_of_jobs*self.eventsPerJob)+' events')
|
| 470 |
|
|
if check > 0:
|
| 471 |
|
|
common.logger.message('Warning: asked '+str(self.total_number_of_events)+' but can do only '+str(int(self.total_number_of_jobs)*self.eventsPerJob))
|
| 472 |
|
|
|
| 473 |
|
|
# argument is seed number.$i
|
| 474 |
|
|
self.list_of_args = []
|
| 475 |
|
|
for i in range(self.total_number_of_jobs):
|
| 476 |
|
|
## Since there is no input, any site is good
|
| 477 |
spiga |
1.3 |
jobDestination.append([""]) #must be empty to write correctly the xml
|
| 478 |
spiga |
1.1 |
args=[]
|
| 479 |
spiga |
1.3 |
if (firstRun):
|
| 480 |
spiga |
1.1 |
## pythia first run
|
| 481 |
spiga |
1.3 |
args.append(str(firstRun)+str(i))
|
| 482 |
|
|
if (generator in managedGenerators):
|
| 483 |
|
|
if (generator == 'comphep' and i == 0):
|
| 484 |
spiga |
1.1 |
# COMPHEP is brain-dead and wants event #'s like 1,100,200,300
|
| 485 |
|
|
args.append('1')
|
| 486 |
|
|
else:
|
| 487 |
|
|
args.append(str(i*self.eventsPerJob))
|
| 488 |
spiga |
1.7 |
args.append(str(self.eventsPerJob))
|
| 489 |
spiga |
1.1 |
self.list_of_args.append(args)
|
| 490 |
|
|
# prepare dict output
|
| 491 |
|
|
dictOut = {}
|
| 492 |
|
|
dictOut['args'] = self.list_of_args
|
| 493 |
spiga |
1.3 |
dictOut['jobDestination'] = jobDestination
|
| 494 |
spiga |
1.1 |
dictOut['njobs']=self.total_number_of_jobs
|
| 495 |
|
|
|
| 496 |
|
|
return dictOut
|
| 497 |
|
|
|
| 498 |
|
|
|
| 499 |
|
|
def jobSplittingForScript(self):
|
| 500 |
|
|
"""
|
| 501 |
|
|
Perform job splitting based on number of job
|
| 502 |
|
|
"""
|
| 503 |
|
|
self.checkUserSettings()
|
| 504 |
spiga |
1.3 |
if (self.selectNumberOfJobs == 0):
|
| 505 |
spiga |
1.1 |
msg = 'must specify number_of_jobs.'
|
| 506 |
|
|
raise crabexception(msg)
|
| 507 |
spiga |
1.3 |
jobDestination = []
|
| 508 |
spiga |
1.1 |
common.logger.debug(5,'Splitting per job')
|
| 509 |
|
|
common.logger.message('Required '+str(self.theNumberOfJobs)+' jobs in total ')
|
| 510 |
|
|
|
| 511 |
|
|
self.total_number_of_jobs = self.theNumberOfJobs
|
| 512 |
|
|
|
| 513 |
|
|
common.logger.debug(5,'N jobs '+str(self.total_number_of_jobs))
|
| 514 |
|
|
|
| 515 |
|
|
common.logger.message(str(self.total_number_of_jobs)+' jobs can be created')
|
| 516 |
|
|
|
| 517 |
|
|
# argument is seed number.$i
|
| 518 |
|
|
self.list_of_args = []
|
| 519 |
|
|
for i in range(self.total_number_of_jobs):
|
| 520 |
spiga |
1.3 |
jobDestination.append([""])
|
| 521 |
spiga |
1.1 |
self.list_of_args.append([str(i)])
|
| 522 |
|
|
|
| 523 |
|
|
# prepare dict output
|
| 524 |
|
|
dictOut = {}
|
| 525 |
|
|
dictOut['args'] = self.list_of_args
|
| 526 |
spiga |
1.3 |
dictOut['jobDestination'] = jobDestination
|
| 527 |
spiga |
1.1 |
dictOut['njobs']=self.total_number_of_jobs
|
| 528 |
|
|
return dictOut
|
| 529 |
|
|
|
| 530 |
|
|
|
| 531 |
|
|
def jobSplittingByLumi(self):
|
| 532 |
|
|
"""
|
| 533 |
|
|
"""
|
| 534 |
|
|
return
|
| 535 |
|
|
def Algos(self):
|
| 536 |
|
|
"""
|
| 537 |
|
|
Define key splittingType matrix
|
| 538 |
|
|
"""
|
| 539 |
|
|
SplitAlogs = {
|
| 540 |
|
|
'EventBased' : self.jobSplittingByEvent,
|
| 541 |
|
|
'RunBased' : self.jobSplittingByRun,
|
| 542 |
|
|
'LumiBased' : self.jobSplittingByLumi,
|
| 543 |
|
|
'NoInput' : self.jobSplittingNoInput,
|
| 544 |
|
|
'ForScript' : self.jobSplittingForScript
|
| 545 |
|
|
}
|
| 546 |
|
|
return SplitAlogs
|
| 547 |
|
|
|
