| 1 |
+ |
|
| 2 |
+ |
__revision__ = "$Id$" |
| 3 |
+ |
__version__ = "$Revision$" |
| 4 |
+ |
|
| 5 |
|
import common |
| 2 |
– |
from crab_logger import Logger |
| 6 |
|
from crab_exceptions import * |
| 7 |
|
from crab_util import * |
| 8 |
+ |
|
| 9 |
+ |
from WMCore.DataStructs.File import File |
| 10 |
+ |
from WMCore.DataStructs.Fileset import Fileset |
| 11 |
+ |
from WMCore.DataStructs.Run import Run |
| 12 |
+ |
from WMCore.DataStructs.Subscription import Subscription |
| 13 |
+ |
from WMCore.DataStructs.Workflow import Workflow |
| 14 |
+ |
from WMCore.JobSplitting.SplitterFactory import SplitterFactory |
| 15 |
|
from WMCore.SiteScreening.BlackWhiteListParser import SEBlackWhiteListParser |
| 16 |
|
|
| 17 |
|
class JobSplitter: |
| 18 |
|
def __init__( self, cfg_params, args ): |
| 19 |
|
self.cfg_params = cfg_params |
| 20 |
|
self.args=args |
| 21 |
+ |
|
| 22 |
+ |
self.lumisPerJob = -1 |
| 23 |
+ |
self.totalNLumis = 0 |
| 24 |
+ |
self.theNumberOfJobs = 0 |
| 25 |
+ |
self.limitNJobs = False |
| 26 |
+ |
self.limitTotalLumis = False |
| 27 |
+ |
self.limitJobLumis = False |
| 28 |
+ |
|
| 29 |
|
#self.maxEvents |
| 30 |
|
# init BlackWhiteListParser |
| 31 |
< |
seWhiteList = cfg_params.get('EDG.se_white_list',[]) |
| 32 |
< |
seBlackList = cfg_params.get('EDG.se_black_list',[]) |
| 33 |
< |
self.blackWhiteListParser = SEBlackWhiteListParser(seWhiteList, seBlackList, common.logger) |
| 31 |
> |
self.seWhiteList = cfg_params.get('GRID.se_white_list',[]) |
| 32 |
> |
seBlackList = cfg_params.get('GRID.se_black_list',[]) |
| 33 |
> |
self.blackWhiteListParser = SEBlackWhiteListParser(self.seWhiteList, seBlackList, common.logger()) |
| 34 |
|
|
| 35 |
|
|
| 36 |
|
def checkUserSettings(self): |
| 62 |
|
self.selectTotalNumberEvents = 0 |
| 63 |
|
|
| 64 |
|
|
| 65 |
+ |
def checkLumiSettings(self): |
| 66 |
+ |
""" |
| 67 |
+ |
Check to make sure the user has specified enough information to |
| 68 |
+ |
perform splitting by Lumis to run the job |
| 69 |
+ |
""" |
| 70 |
+ |
settings = 0 |
| 71 |
+ |
if self.cfg_params.has_key('CMSSW.lumis_per_job'): |
| 72 |
+ |
self.lumisPerJob =int( self.cfg_params['CMSSW.lumis_per_job']) |
| 73 |
+ |
self.limitJobLumis = True |
| 74 |
+ |
settings += 1 |
| 75 |
+ |
|
| 76 |
+ |
if self.cfg_params.has_key('CMSSW.number_of_jobs'): |
| 77 |
+ |
self.theNumberOfJobs =int( self.cfg_params['CMSSW.number_of_jobs']) |
| 78 |
+ |
self.limitNJobs = True |
| 79 |
+ |
settings += 1 |
| 80 |
+ |
|
| 81 |
+ |
if self.cfg_params.has_key('CMSSW.total_number_of_lumis'): |
| 82 |
+ |
self.totalNLumis = int(self.cfg_params['CMSSW.total_number_of_lumis']) |
| 83 |
+ |
self.limitTotalLumis = (self.totalNLumis != -1) |
| 84 |
+ |
settings += 1 |
| 85 |
+ |
|
| 86 |
+ |
if settings != 2: |
| 87 |
+ |
msg = 'When running on analysis datasets you must specify two and only two of:\n' |
| 88 |
+ |
msg += ' number_of_jobs, lumis_per_job, total_number_of_lumis' |
| 89 |
+ |
raise CrabException(msg) |
| 90 |
+ |
if self.limitNJobs and self.limitJobLumis: |
| 91 |
+ |
self.limitTotalLumis = True |
| 92 |
+ |
self.totalNLumis = self.lumisPerJob * self.theNumberOfJobs |
| 93 |
+ |
|
| 94 |
+ |
|
| 95 |
+ |
def ComputeSubBlockSites( self, blockSites ): |
| 96 |
+ |
""" |
| 97 |
+ |
""" |
| 98 |
+ |
sub_blockSites = {} |
| 99 |
+ |
for k,v in blockSites.iteritems(): |
| 100 |
+ |
sites=self.blackWhiteListParser.checkWhiteList(v) |
| 101 |
+ |
if sites : sub_blockSites[k]=v |
| 102 |
+ |
if len(sub_blockSites) < 1: |
| 103 |
+ |
msg = 'WARNING: the sites %s is not hosting any part of data.'%self.seWhiteList |
| 104 |
+ |
raise CrabException(msg) |
| 105 |
+ |
return sub_blockSites |
| 106 |
+ |
|
| 107 |
|
######################################################################## |
| 108 |
|
def jobSplittingByEvent( self ): |
| 109 |
|
""" |
| 118 |
|
self.list_of_args - File(s) job will run on (a list of lists) |
| 119 |
|
""" |
| 120 |
|
|
| 121 |
< |
jobDestination=[] |
| 121 |
> |
jobDestination=[] |
| 122 |
|
self.checkUserSettings() |
| 123 |
|
if ( (self.selectTotalNumberEvents + self.selectEventsPerJob + self.selectNumberOfJobs) != 2 ): |
| 124 |
|
msg = 'Must define exactly two of total_number_of_events, events_per_job, or number_of_jobs.' |
| 125 |
|
raise CrabException(msg) |
| 126 |
< |
|
| 127 |
< |
blockSites = self.args['blockSites'] |
| 126 |
> |
|
| 127 |
> |
blockSites = self.args['blockSites'] |
| 128 |
|
pubdata = self.args['pubdata'] |
| 129 |
|
filesbyblock=pubdata.getFiles() |
| 130 |
|
|
| 138 |
|
self.useParent = int(self.cfg_params.get('CMSSW.use_parent',0)) |
| 139 |
|
noBboundary = int(self.cfg_params.get('CMSSW.no_block_boundary',0)) |
| 140 |
|
|
| 141 |
+ |
if noBboundary == 1: |
| 142 |
+ |
if self.total_number_of_events== -1: |
| 143 |
+ |
msg = 'You are selecting no_block_boundary=1 which does not allow to set total_number_of_events=-1\n' |
| 144 |
+ |
msg +='\tYou shoud get the number of event from DBS web interface and use it for your configuration.' |
| 145 |
+ |
raise CrabException(msg) |
| 146 |
+ |
if len(self.seWhiteList) == 0 or len(self.seWhiteList.split(',')) != 1: |
| 147 |
+ |
msg = 'You are selecting no_block_boundary=1 which requires to choose one and only one site.\n' |
| 148 |
+ |
msg += "\tPlease set se_white_list with the site's storage element name." |
| 149 |
+ |
raise CrabException(msg) |
| 150 |
+ |
blockSites = self.ComputeSubBlockSites(blockSites) |
| 151 |
+ |
|
| 152 |
|
# ---- Handle the possible job splitting configurations ---- # |
| 153 |
|
if (self.selectTotalNumberEvents): |
| 154 |
|
totalEventsRequested = self.total_number_of_events |
| 163 |
|
# If user requested more events than are in the dataset |
| 164 |
|
elif (totalEventsRequested > self.maxEvents): |
| 165 |
|
eventsRemaining = self.maxEvents |
| 166 |
< |
common.logger.message("Requested "+str(self.total_number_of_events)+ " events, but only "+str(self.maxEvents)+" events are available.") |
| 166 |
> |
common.logger.info("Requested "+str(self.total_number_of_events)+ " events, but only "+str(self.maxEvents)+" events are available.") |
| 167 |
|
# If user requested less events than are in the dataset |
| 168 |
|
else: |
| 169 |
|
eventsRemaining = totalEventsRequested |
| 179 |
|
eventsPerJobRequested = int(eventsRemaining/self.theNumberOfJobs) |
| 180 |
|
|
| 181 |
|
if (self.selectNumberOfJobs): |
| 182 |
< |
common.logger.message("May not create the exact number_of_jobs requested.") |
| 182 |
> |
common.logger.info("May not create the exact number_of_jobs requested.") |
| 183 |
|
|
| 184 |
|
# old... to remove Daniele |
| 185 |
|
totalNumberOfJobs = 999999999 |
| 196 |
|
jobsOfBlock = {} |
| 197 |
|
|
| 198 |
|
parString = "" |
| 199 |
+ |
pString = "" |
| 200 |
|
filesEventCount = 0 |
| 201 |
+ |
msg='' |
| 202 |
|
|
| 203 |
|
# ---- Iterate over the blocks in the dataset until ---- # |
| 204 |
|
# ---- we've met the requested total # of events ---- # |
| 210 |
|
|
| 211 |
|
if self.eventsbyblock.has_key(block) : |
| 212 |
|
numEventsInBlock = self.eventsbyblock[block] |
| 213 |
< |
common.logger.debug(5,'Events in Block File '+str(numEventsInBlock)) |
| 213 |
> |
common.logger.debug('Events in Block File '+str(numEventsInBlock)) |
| 214 |
|
|
| 215 |
|
files = filesbyblock[block] |
| 216 |
|
numFilesInBlock = len(files) |
| 220 |
|
if noBboundary == 0: # DD |
| 221 |
|
# ---- New block => New job ---- # |
| 222 |
|
parString = "" |
| 223 |
+ |
pString="" |
| 224 |
|
# counter for number of events in files currently worked on |
| 225 |
|
filesEventCount = 0 |
| 226 |
|
# flag if next while loop should touch new file |
| 230 |
|
|
| 231 |
|
# ---- Iterate over the files in the block until we've met the requested ---- # |
| 232 |
|
# ---- total # of events or we've gone over all the files in this block ---- # |
| 233 |
< |
pString='' |
| 233 |
> |
msg='\n' |
| 234 |
|
while ( (eventsRemaining > 0) and (fileCount < numFilesInBlock) and (jobCount < totalNumberOfJobs) ): |
| 235 |
|
file = files[fileCount] |
| 236 |
|
if self.useParent==1: |
| 237 |
|
parent = self.parentFiles[file] |
| 238 |
< |
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)) |
| 238 |
> |
common.logger.log(10-1, "File "+str(file)+" has the following parents: "+str(parent)) |
| 239 |
|
if newFile : |
| 240 |
|
try: |
| 241 |
|
numEventsInFile = self.eventsbyfile[file] |
| 242 |
< |
common.logger.debug(6, "File "+str(file)+" has "+str(numEventsInFile)+" events") |
| 242 |
> |
common.logger.log(10-1, "File "+str(file)+" has "+str(numEventsInFile)+" events") |
| 243 |
|
# increase filesEventCount |
| 244 |
|
filesEventCount += numEventsInFile |
| 245 |
|
# Add file to current job |
| 246 |
< |
parString += '\\\"' + file + '\\\"\,' |
| 246 |
> |
parString += file + ',' |
| 247 |
> |
if self.useParent==1: |
| 248 |
> |
for f in parent : |
| 249 |
> |
pString += f + ',' |
| 250 |
|
newFile = 0 |
| 251 |
|
except KeyError: |
| 252 |
< |
common.logger.message("File "+str(file)+" has unknown number of events: skipping") |
| 252 |
> |
common.logger.info("File "+str(file)+" has unknown number of events: skipping") |
| 253 |
|
|
| 254 |
|
eventsPerJobRequested = min(eventsPerJobRequested, eventsRemaining) |
| 255 |
|
# if less events in file remain than eventsPerJobRequested |
| 262 |
|
if ( fileCount == numFilesInBlock-1 ) : |
| 263 |
|
# end job using last file, use remaining events in block |
| 264 |
|
# close job and touch new file |
| 265 |
< |
fullString = parString[:-2] |
| 265 |
> |
fullString = parString[:-1] |
| 266 |
|
if self.useParent==1: |
| 267 |
< |
fullParentString = pString[:-2] |
| 267 |
> |
fullParentString = pString[:-1] |
| 268 |
|
list_of_lists.append([fullString,fullParentString,str(-1),str(jobSkipEventCount)]) |
| 269 |
|
else: |
| 270 |
|
list_of_lists.append([fullString,str(-1),str(jobSkipEventCount)]) |
| 271 |
< |
common.logger.debug(3,"Job "+str(jobCount+1)+" can run over "+str(filesEventCount - jobSkipEventCount)+" events (last file in block).") |
| 271 |
> |
msg += "Job %s can run over %s events (last file in block).\n"%(str(jobCount+1), str(filesEventCount - jobSkipEventCount)) |
| 272 |
|
jobDestination.append(blockSites[block]) |
| 273 |
< |
common.logger.debug(5,"Job "+str(jobCount+1)+" Destination: "+str(jobDestination[jobCount])) |
| 273 |
> |
msg += "Job %s Destination: %s\n"%(str(jobCount+1),str(SE2CMS(jobDestination[jobCount]))) |
| 274 |
|
# fill jobs of block dictionary |
| 275 |
|
jobsOfBlock[block].append(jobCount+1) |
| 276 |
|
# reset counter |
| 291 |
|
# if events in file equal to eventsPerJobRequested |
| 292 |
|
elif ( filesEventCount - jobSkipEventCount == eventsPerJobRequested ) : |
| 293 |
|
# close job and touch new file |
| 294 |
< |
fullString = parString[:-2] |
| 294 |
> |
fullString = parString[:-1] |
| 295 |
|
if self.useParent==1: |
| 296 |
< |
fullParentString = pString[:-2] |
| 296 |
> |
fullParentString = pString[:-1] |
| 297 |
|
list_of_lists.append([fullString,fullParentString,str(eventsPerJobRequested),str(jobSkipEventCount)]) |
| 298 |
|
else: |
| 299 |
|
list_of_lists.append([fullString,str(eventsPerJobRequested),str(jobSkipEventCount)]) |
| 300 |
< |
common.logger.debug(3,"Job "+str(jobCount+1)+" can run over "+str(eventsPerJobRequested)+" events.") |
| 300 |
> |
msg += "Job %s can run over %s events.\n"%(str(jobCount+1),str(eventsPerJobRequested)) |
| 301 |
|
jobDestination.append(blockSites[block]) |
| 302 |
< |
common.logger.debug(5,"Job "+str(jobCount+1)+" Destination: "+str(jobDestination[jobCount])) |
| 302 |
> |
msg+= "Job %s Destination: %s\n"%(str(jobCount+1),str(SE2CMS(jobDestination[jobCount]))) |
| 303 |
|
jobsOfBlock[block].append(jobCount+1) |
| 304 |
|
# reset counter |
| 305 |
|
jobCount = jobCount + 1 |
| 316 |
|
# if more events in file remain than eventsPerJobRequested |
| 317 |
|
else : |
| 318 |
|
# close job but don't touch new file |
| 319 |
< |
fullString = parString[:-2] |
| 319 |
> |
fullString = parString[:-1] |
| 320 |
|
if self.useParent==1: |
| 321 |
< |
fullParentString = pString[:-2] |
| 321 |
> |
fullParentString = pString[:-1] |
| 322 |
|
list_of_lists.append([fullString,fullParentString,str(eventsPerJobRequested),str(jobSkipEventCount)]) |
| 323 |
|
else: |
| 324 |
|
list_of_lists.append([fullString,str(eventsPerJobRequested),str(jobSkipEventCount)]) |
| 325 |
< |
common.logger.debug(3,"Job "+str(jobCount+1)+" can run over "+str(eventsPerJobRequested)+" events.") |
| 325 |
> |
msg += "Job %s can run over %s events.\n"%(str(jobCount+1),str(eventsPerJobRequested)) |
| 326 |
|
jobDestination.append(blockSites[block]) |
| 327 |
< |
common.logger.debug(5,"Job "+str(jobCount+1)+" Destination: "+str(jobDestination[jobCount])) |
| 327 |
> |
msg+= "Job %s Destination: %s\n"%(str(jobCount+1),str(SE2CMS(jobDestination[jobCount]))) |
| 328 |
|
jobsOfBlock[block].append(jobCount+1) |
| 329 |
|
# increase counter |
| 330 |
|
jobCount = jobCount + 1 |
| 335 |
|
jobSkipEventCount = eventsPerJobRequested - (filesEventCount - jobSkipEventCount - self.eventsbyfile[file]) |
| 336 |
|
# remove all but the last file |
| 337 |
|
filesEventCount = self.eventsbyfile[file] |
| 338 |
+ |
pString_tmp='' |
| 339 |
|
if self.useParent==1: |
| 340 |
< |
for f in parent : pString += '\\\"' + f + '\\\"\,' |
| 341 |
< |
parString = '\\\"' + file + '\\\"\,' |
| 340 |
> |
for f in parent : pString_tmp += f + ',' |
| 341 |
> |
pString = pString_tmp |
| 342 |
> |
parString = file + ',' |
| 343 |
|
pass # END if |
| 344 |
|
pass # END while (iterate over files in the block) |
| 345 |
|
pass # END while (iterate over blocks in the dataset) |
| 346 |
+ |
common.logger.debug(msg) |
| 347 |
|
self.ncjobs = self.total_number_of_jobs = jobCount |
| 348 |
|
if (eventsRemaining > 0 and jobCount < totalNumberOfJobs ): |
| 349 |
< |
common.logger.message("Could not run on all requested events because some blocks not hosted at allowed sites.") |
| 350 |
< |
common.logger.message(str(jobCount)+" job(s) can run on "+str(totalEventCount)+" events.\n") |
| 351 |
< |
|
| 349 |
> |
common.logger.info("Could not run on all requested events because some blocks not hosted at allowed sites.") |
| 350 |
> |
common.logger.info(str(jobCount)+" job(s) can run on "+str(totalEventCount)+" events.\n") |
| 351 |
> |
|
| 352 |
|
# skip check on block with no sites DD |
| 353 |
|
if noBboundary == 0 : self.checkBlockNoSite(blocks,jobsOfBlock,blockSites) |
| 354 |
|
|
| 355 |
|
# prepare dict output |
| 356 |
|
dictOut = {} |
| 357 |
+ |
dictOut['params']= ['InputFiles','MaxEvents','SkipEvents'] |
| 358 |
+ |
if self.useParent: dictOut['params']= ['InputFiles','ParentFiles','MaxEvents','SkipEvents'] |
| 359 |
|
dictOut['args'] = list_of_lists |
| 360 |
|
dictOut['jobDestination'] = jobDestination |
| 361 |
|
dictOut['njobs']=self.total_number_of_jobs |
| 364 |
|
|
| 365 |
|
# keep trace of block with no sites to print a warning at the end |
| 366 |
|
|
| 367 |
< |
def checkBlockNoSite(self,blocks,jobsOfBlock,blockSites): |
| 367 |
> |
def checkBlockNoSite(self,blocks,jobsOfBlock,blockSites): |
| 368 |
|
# screen output |
| 369 |
|
screenOutput = "List of jobs and available destination sites:\n\n" |
| 370 |
|
noSiteBlock = [] |
| 371 |
|
bloskNoSite = [] |
| 372 |
+ |
allBlock = [] |
| 373 |
|
|
| 374 |
|
blockCounter = 0 |
| 375 |
|
for block in blocks: |
| 376 |
|
if block in jobsOfBlock.keys() : |
| 377 |
|
blockCounter += 1 |
| 378 |
+ |
allBlock.append( blockCounter ) |
| 379 |
+ |
sites=self.blackWhiteListParser.checkWhiteList(self.blackWhiteListParser.checkBlackList(blockSites[block],[block]),[block]) |
| 380 |
|
screenOutput += "Block %5i: jobs %20s: sites: %s\n" % (blockCounter,spanRanges(jobsOfBlock[block]), |
| 381 |
< |
','.join(self.blackWhiteListParser.checkWhiteList(self.blackWhiteListParser.checkBlackList(blockSites[block],[block]),[block]))) |
| 382 |
< |
if len(self.blackWhiteListParser.checkWhiteList(self.blackWhiteListParser.checkBlackList(blockSites[block],[block]),[block])) == 0: |
| 381 |
> |
', '.join(SE2CMS(sites))) |
| 382 |
> |
if len(sites) == 0: |
| 383 |
|
noSiteBlock.append( spanRanges(jobsOfBlock[block]) ) |
| 384 |
|
bloskNoSite.append( blockCounter ) |
| 385 |
|
|
| 386 |
< |
common.logger.message(screenOutput) |
| 386 |
> |
common.logger.info(screenOutput) |
| 387 |
|
if len(noSiteBlock) > 0 and len(bloskNoSite) > 0: |
| 388 |
|
msg = 'WARNING: No sites are hosting any part of data for block:\n ' |
| 389 |
|
virgola = "" |
| 391 |
|
virgola = "," |
| 392 |
|
for block in bloskNoSite: |
| 393 |
|
msg += ' ' + str(block) + virgola |
| 394 |
< |
msg += '\n Related jobs:\n ' |
| 394 |
> |
msg += '\n\t\tRelated jobs:\n ' |
| 395 |
|
virgola = "" |
| 396 |
|
if len(noSiteBlock) > 1: |
| 397 |
|
virgola = "," |
| 398 |
|
for range_jobs in noSiteBlock: |
| 399 |
|
msg += str(range_jobs) + virgola |
| 400 |
< |
msg += '\n will not be submitted and this block of data can not be analyzed!\n' |
| 401 |
< |
if self.cfg_params.has_key('EDG.se_white_list'): |
| 402 |
< |
msg += 'WARNING: SE White List: '+self.cfg_params['EDG.se_white_list']+'\n' |
| 403 |
< |
msg += '(Hint: By whitelisting you force the job to run at this particular site(s).\n' |
| 404 |
< |
msg += 'Please check if the dataset is available at this site!)\n' |
| 405 |
< |
if self.cfg_params.has_key('EDG.ce_white_list'): |
| 406 |
< |
msg += 'WARNING: CE White List: '+self.cfg_params['EDG.ce_white_list']+'\n' |
| 407 |
< |
msg += '(Hint: By whitelisting you force the job to run at this particular site(s).\n' |
| 408 |
< |
msg += 'Please check if the dataset is available at this site!)\n' |
| 400 |
> |
msg += '\n\t\twill not be submitted and this block of data can not be analyzed!\n' |
| 401 |
> |
if self.cfg_params.has_key('GRID.se_white_list'): |
| 402 |
> |
msg += '\tWARNING: SE White List: '+self.cfg_params['GRID.se_white_list']+'\n' |
| 403 |
> |
msg += '\t(Hint: By whitelisting you force the job to run at this particular site(s).\n' |
| 404 |
> |
msg += '\tPlease check if the dataset is available at this site!)' |
| 405 |
> |
if self.cfg_params.has_key('GRID.ce_white_list'): |
| 406 |
> |
msg += '\tWARNING: CE White List: '+self.cfg_params['GRID.ce_white_list']+'\n' |
| 407 |
> |
msg += '\t(Hint: By whitelisting you force the job to run at this particular site(s).\n' |
| 408 |
> |
msg += '\tPlease check if the dataset is available at this site!)\n' |
| 409 |
> |
|
| 410 |
> |
common.logger.info(msg) |
| 411 |
|
|
| 412 |
< |
common.logger.message(msg) |
| 412 |
> |
if bloskNoSite == allBlock: |
| 413 |
> |
raise CrabException('No jobs created') |
| 414 |
|
|
| 415 |
|
return |
| 416 |
|
|
| 417 |
|
|
| 418 |
|
######################################################################## |
| 419 |
< |
def jobSplittingByRun(self): |
| 419 |
> |
def jobSplittingByRun(self): |
| 420 |
|
""" |
| 421 |
|
""" |
| 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 |
| 422 |
|
|
| 423 |
|
self.checkUserSettings() |
| 424 |
< |
blockSites = self.args['blockSites'] |
| 424 |
> |
blockSites = self.args['blockSites'] |
| 425 |
|
pubdata = self.args['pubdata'] |
| 426 |
|
|
| 427 |
|
if self.selectNumberOfJobs == 0 : |
| 428 |
|
self.theNumberOfJobs = 9999999 |
| 429 |
|
blocks = {} |
| 430 |
< |
runList = [] |
| 430 |
> |
runList = [] |
| 431 |
|
thefiles = Fileset(name='FilesToSplit') |
| 432 |
|
fileList = pubdata.getListFiles() |
| 433 |
|
for f in fileList: |
| 357 |
– |
# print f |
| 434 |
|
block = f['Block']['Name'] |
| 435 |
< |
# if not blocks.has_key(block): |
| 360 |
< |
# blocks[block] = reader.listFileBlockLocation(block) |
| 361 |
< |
try: |
| 435 |
> |
try: |
| 436 |
|
f['Block']['StorageElementList'].extend(blockSites[block]) |
| 437 |
|
except: |
| 438 |
|
continue |
| 440 |
|
[ wmbsFile['locations'].add(x) for x in blockSites[block] ] |
| 441 |
|
wmbsFile['block'] = block |
| 442 |
|
runNum = f['RunsList'][0]['RunNumber'] |
| 443 |
< |
runList.append(runNum) |
| 443 |
> |
runList.append(runNum) |
| 444 |
|
myRun = Run(runNumber=runNum) |
| 445 |
|
wmbsFile.addRun( myRun ) |
| 446 |
|
thefiles.addFile( |
| 447 |
|
wmbsFile |
| 448 |
|
) |
| 449 |
< |
|
| 449 |
> |
|
| 450 |
|
work = Workflow() |
| 451 |
|
subs = Subscription( |
| 452 |
|
fileset = thefiles, |
| 455 |
|
type = "Processing") |
| 456 |
|
splitter = SplitterFactory() |
| 457 |
|
jobfactory = splitter(subs) |
| 458 |
< |
|
| 459 |
< |
#loop over all runs |
| 386 |
< |
set = Set(runList) |
| 458 |
> |
|
| 459 |
> |
#loop over all runs |
| 460 |
|
list_of_lists = [] |
| 461 |
|
jobDestination = [] |
| 389 |
– |
|
| 462 |
|
count = 0 |
| 463 |
< |
for i in list(set): |
| 463 |
> |
for jobGroup in jobfactory(): |
| 464 |
|
if count < self.theNumberOfJobs: |
| 465 |
< |
res = self.getJobInfo(jobfactory()) |
| 466 |
< |
parString = '' |
| 465 |
> |
res = self.getJobInfo(jobGroup) |
| 466 |
> |
parString = '' |
| 467 |
|
for file in res['lfns']: |
| 468 |
< |
parString += '\\\"' + file + '\\\"\,' |
| 469 |
< |
fullString = parString[:-2] |
| 470 |
< |
list_of_lists.append([fullString,str(-1),str(0)]) |
| 468 |
> |
parString += file + ',' |
| 469 |
> |
fullString = parString[:-1] |
| 470 |
> |
list_of_lists.append([fullString,str(-1),str(0)]) |
| 471 |
|
#need to check single file location |
| 472 |
< |
jobDestination.append(res['locations']) |
| 472 |
> |
jobDestination.append(res['locations']) |
| 473 |
|
count +=1 |
| 402 |
– |
#print jobDestination |
| 474 |
|
# prepare dict output |
| 475 |
|
dictOut = {} |
| 476 |
+ |
dictOut['params']= ['InputFiles','MaxEvents','SkipEvents'] |
| 477 |
|
dictOut['args'] = list_of_lists |
| 478 |
|
dictOut['jobDestination'] = jobDestination |
| 479 |
|
dictOut['njobs']=count |
| 482 |
|
|
| 483 |
|
def getJobInfo( self,jobGroup ): |
| 484 |
|
res = {} |
| 485 |
< |
lfns = [] |
| 486 |
< |
locations = [] |
| 485 |
> |
lfns = [] |
| 486 |
> |
locations = [] |
| 487 |
|
tmp_check=0 |
| 488 |
|
for job in jobGroup.jobs: |
| 489 |
|
for file in job.getFiles(): |
| 490 |
< |
lfns.append(file['lfn']) |
| 490 |
> |
lfns.append(file['lfn']) |
| 491 |
|
for loc in file['locations']: |
| 492 |
|
if tmp_check < 1 : |
| 493 |
|
locations.append(loc) |
| 494 |
< |
tmp_check = tmp_check + 1 |
| 495 |
< |
### qui va messo il check per la locations |
| 496 |
< |
res['lfns'] = lfns |
| 497 |
< |
res['locations'] = locations |
| 498 |
< |
return res |
| 499 |
< |
|
| 494 |
> |
tmp_check = tmp_check + 1 |
| 495 |
> |
### qui va messo il check per la locations |
| 496 |
> |
res['lfns'] = lfns |
| 497 |
> |
res['locations'] = locations |
| 498 |
> |
return res |
| 499 |
> |
|
| 500 |
|
######################################################################## |
| 501 |
< |
def jobSplittingNoInput(self): |
| 501 |
> |
def prepareSplittingNoInput(self): |
| 502 |
|
""" |
| 431 |
– |
Perform job splitting based on number of event per job |
| 503 |
|
""" |
| 433 |
– |
common.logger.debug(5,'Splitting per events') |
| 434 |
– |
self.checkUserSettings() |
| 435 |
– |
jobDestination=[] |
| 436 |
– |
if ( (self.selectTotalNumberEvents + self.selectEventsPerJob + self.selectNumberOfJobs) != 2 ): |
| 437 |
– |
msg = 'Must define exactly two of total_number_of_events, events_per_job, or number_of_jobs.' |
| 438 |
– |
raise CrabException(msg) |
| 439 |
– |
|
| 440 |
– |
managedGenerators =self.args['managedGenerators'] |
| 441 |
– |
generator = self.args['generator'] |
| 442 |
– |
firstRun = self.cfg_params.get('CMSSW.first_run',None) |
| 443 |
– |
|
| 504 |
|
if (self.selectEventsPerJob): |
| 505 |
< |
common.logger.message('Required '+str(self.eventsPerJob)+' events per job ') |
| 505 |
> |
common.logger.info('Required '+str(self.eventsPerJob)+' events per job ') |
| 506 |
|
if (self.selectNumberOfJobs): |
| 507 |
< |
common.logger.message('Required '+str(self.theNumberOfJobs)+' jobs in total ') |
| 507 |
> |
common.logger.info('Required '+str(self.theNumberOfJobs)+' jobs in total ') |
| 508 |
|
if (self.selectTotalNumberEvents): |
| 509 |
< |
common.logger.message('Required '+str(self.total_number_of_events)+' events in total ') |
| 509 |
> |
common.logger.info('Required '+str(self.total_number_of_events)+' events in total ') |
| 510 |
|
|
| 511 |
|
if (self.total_number_of_events < 0): |
| 512 |
|
msg='Cannot split jobs per Events with "-1" as total number of events' |
| 523 |
|
self.total_number_of_jobs = self.theNumberOfJobs |
| 524 |
|
self.eventsPerJob = int(self.total_number_of_events/self.total_number_of_jobs) |
| 525 |
|
|
| 526 |
< |
common.logger.debug(5,'N jobs '+str(self.total_number_of_jobs)) |
| 526 |
> |
|
| 527 |
> |
def jobSplittingNoInput(self): |
| 528 |
> |
""" |
| 529 |
> |
Perform job splitting based on number of event per job |
| 530 |
> |
""" |
| 531 |
> |
common.logger.debug('Splitting per events') |
| 532 |
> |
self.checkUserSettings() |
| 533 |
> |
jobDestination=[] |
| 534 |
> |
if ( (self.selectTotalNumberEvents + self.selectEventsPerJob + self.selectNumberOfJobs) != 2 ): |
| 535 |
> |
msg = 'Must define exactly two of total_number_of_events, events_per_job, or number_of_jobs.' |
| 536 |
> |
raise CrabException(msg) |
| 537 |
> |
|
| 538 |
> |
managedGenerators =self.args['managedGenerators'] |
| 539 |
> |
generator = self.args['generator'] |
| 540 |
> |
firstLumi = self.cfg_params.get('CMSSW.first_lumi', 1) |
| 541 |
> |
|
| 542 |
> |
self.prepareSplittingNoInput() |
| 543 |
> |
|
| 544 |
> |
common.logger.debug('N jobs '+str(self.total_number_of_jobs)) |
| 545 |
|
|
| 546 |
|
# is there any remainder? |
| 547 |
|
check = int(self.total_number_of_events) - (int(self.total_number_of_jobs)*self.eventsPerJob) |
| 548 |
|
|
| 549 |
< |
common.logger.debug(5,'Check '+str(check)) |
| 549 |
> |
common.logger.debug('Check '+str(check)) |
| 550 |
|
|
| 551 |
< |
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') |
| 551 |
> |
common.logger.info(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') |
| 552 |
|
if check > 0: |
| 553 |
< |
common.logger.message('Warning: asked '+str(self.total_number_of_events)+' but can do only '+str(int(self.total_number_of_jobs)*self.eventsPerJob)) |
| 553 |
> |
common.logger.info('Warning: asked '+str(self.total_number_of_events)+' but can do only '+str(int(self.total_number_of_jobs)*self.eventsPerJob)) |
| 554 |
|
|
| 555 |
|
# argument is seed number.$i |
| 556 |
|
self.list_of_args = [] |
| 557 |
|
for i in range(self.total_number_of_jobs): |
| 558 |
|
## Since there is no input, any site is good |
| 559 |
< |
jobDestination.append([""]) #must be empty to write correctly the xml |
| 559 |
> |
jobDestination.append([""]) # must be empty to correctly write the XML |
| 560 |
|
args=[] |
| 561 |
< |
if (firstRun): |
| 562 |
< |
## pythia first run |
| 485 |
< |
args.append(str(firstRun)+str(i)) |
| 561 |
> |
if (firstLumi): # Pythia first lumi |
| 562 |
> |
args.append(str(int(firstLumi)+i)) |
| 563 |
|
if (generator in managedGenerators): |
| 564 |
< |
if (generator == 'comphep' and i == 0): |
| 564 |
> |
args.append(generator) |
| 565 |
> |
if (generator == 'comphep' and i == 0): |
| 566 |
|
# COMPHEP is brain-dead and wants event #'s like 1,100,200,300 |
| 567 |
|
args.append('1') |
| 568 |
< |
else: |
| 568 |
> |
else: |
| 569 |
|
args.append(str(i*self.eventsPerJob)) |
| 570 |
|
args.append(str(self.eventsPerJob)) |
| 571 |
|
self.list_of_args.append(args) |
| 572 |
|
# prepare dict output |
| 573 |
+ |
|
| 574 |
|
dictOut = {} |
| 575 |
+ |
dictOut['params'] = ['MaxEvents'] |
| 576 |
+ |
if (firstLumi): |
| 577 |
+ |
dictOut['params'] = ['FirstLumi','MaxEvents'] |
| 578 |
+ |
if (generator in managedGenerators): |
| 579 |
+ |
dictOut['params'] = ['FirstLumi', 'Generator', 'FirstEvent', 'MaxEvents'] |
| 580 |
+ |
else: |
| 581 |
+ |
if (generator in managedGenerators) : |
| 582 |
+ |
dictOut['params'] = ['Generator', 'FirstEvent', 'MaxEvents'] |
| 583 |
|
dictOut['args'] = self.list_of_args |
| 584 |
|
dictOut['jobDestination'] = jobDestination |
| 585 |
|
dictOut['njobs']=self.total_number_of_jobs |
| 596 |
|
msg = 'must specify number_of_jobs.' |
| 597 |
|
raise crabexception(msg) |
| 598 |
|
jobDestination = [] |
| 599 |
< |
common.logger.debug(5,'Splitting per job') |
| 600 |
< |
common.logger.message('Required '+str(self.theNumberOfJobs)+' jobs in total ') |
| 599 |
> |
common.logger.debug('Splitting per job') |
| 600 |
> |
common.logger.info('Required '+str(self.theNumberOfJobs)+' jobs in total ') |
| 601 |
|
|
| 602 |
< |
self.total_number_of_jobs = self.theNumberOfJobs |
| 602 |
> |
# self.total_number_of_jobs = self.theNumberOfJobs |
| 603 |
|
|
| 604 |
< |
common.logger.debug(5,'N jobs '+str(self.total_number_of_jobs)) |
| 604 |
> |
self.prepareSplittingNoInput() |
| 605 |
|
|
| 606 |
< |
common.logger.message(str(self.total_number_of_jobs)+' jobs can be created') |
| 606 |
> |
common.logger.debug('N jobs '+str(self.total_number_of_jobs)) |
| 607 |
> |
|
| 608 |
> |
common.logger.info(str(self.total_number_of_jobs)+' jobs can be created') |
| 609 |
|
|
| 610 |
|
# argument is seed number.$i |
| 611 |
|
self.list_of_args = [] |
| 612 |
|
for i in range(self.total_number_of_jobs): |
| 613 |
+ |
args=[] |
| 614 |
|
jobDestination.append([""]) |
| 615 |
< |
self.list_of_args.append([str(i)]) |
| 615 |
> |
if self.eventsPerJob != 0 : |
| 616 |
> |
args.append(str(self.eventsPerJob)) |
| 617 |
> |
self.list_of_args.append(args) |
| 618 |
|
|
| 619 |
|
# prepare dict output |
| 620 |
|
dictOut = {} |
| 621 |
< |
dictOut['args'] = self.list_of_args |
| 621 |
> |
dictOut['params'] = ['MaxEvents'] |
| 622 |
> |
dictOut['args'] = self.list_of_args |
| 623 |
|
dictOut['jobDestination'] = jobDestination |
| 624 |
|
dictOut['njobs']=self.total_number_of_jobs |
| 625 |
|
return dictOut |
| 533 |
– |
|
| 626 |
|
|
| 627 |
< |
def jobSplittingByLumi(self): |
| 627 |
> |
|
| 628 |
> |
def jobSplittingByLumi(self): |
| 629 |
|
""" |
| 630 |
+ |
Split task into jobs by Lumi section paying attention to which |
| 631 |
+ |
lumis should be run (according to the analysis dataset). |
| 632 |
+ |
This uses WMBS job splitting which does not split files over jobs |
| 633 |
+ |
so the job will have AT LEAST as many lumis as requested, perhaps |
| 634 |
+ |
more |
| 635 |
|
""" |
| 636 |
< |
return |
| 636 |
> |
|
| 637 |
> |
common.logger.debug('Splitting by Lumi') |
| 638 |
> |
self.checkLumiSettings() |
| 639 |
> |
|
| 640 |
> |
blockSites = self.args['blockSites'] |
| 641 |
> |
pubdata = self.args['pubdata'] |
| 642 |
> |
|
| 643 |
> |
lumisPerFile = pubdata.getLumis() |
| 644 |
> |
|
| 645 |
> |
# Make the list of WMBS files for job splitter |
| 646 |
> |
fileList = pubdata.getListFiles() |
| 647 |
> |
thefiles = Fileset(name='FilesToSplit') |
| 648 |
> |
for jobFile in fileList: |
| 649 |
> |
block = jobFile['Block']['Name'] |
| 650 |
> |
try: |
| 651 |
> |
jobFile['Block']['StorageElementList'].extend(blockSites[block]) |
| 652 |
> |
except: |
| 653 |
> |
continue |
| 654 |
> |
wmbsFile = File(jobFile['LogicalFileName']) |
| 655 |
> |
[ wmbsFile['locations'].add(x) for x in blockSites[block] ] |
| 656 |
> |
wmbsFile['block'] = block |
| 657 |
> |
for lumi in lumisPerFile[jobFile['LogicalFileName']]: |
| 658 |
> |
wmbsFile.addRun(Run(lumi[0], lumi[1])) |
| 659 |
> |
thefiles.addFile(wmbsFile) |
| 660 |
> |
|
| 661 |
> |
# Create the factory and workflow |
| 662 |
> |
work = Workflow() |
| 663 |
> |
subs = Subscription(fileset = thefiles, workflow = work, |
| 664 |
> |
split_algo = 'LumiBased', type = "Processing") |
| 665 |
> |
splitter = SplitterFactory() |
| 666 |
> |
jobFactory = splitter(subs) |
| 667 |
> |
|
| 668 |
> |
list_of_lists = [] |
| 669 |
> |
jobDestination = [] |
| 670 |
> |
jobCount = 0 |
| 671 |
> |
lumisCreated = 0 |
| 672 |
> |
|
| 673 |
> |
if not self.limitJobLumis: |
| 674 |
> |
self.lumisPerJob = pubdata.getMaxLumis() // self.theNumberOfJobs + 1 |
| 675 |
> |
common.logger.info('Each job will process about %s lumis.' % |
| 676 |
> |
self.lumisPerJob) |
| 677 |
> |
|
| 678 |
> |
for jobGroup in jobFactory(lumis_per_job = self.lumisPerJob): |
| 679 |
> |
for job in jobGroup.jobs: |
| 680 |
> |
if (self.limitNJobs and jobCount >= self.theNumberOfJobs): |
| 681 |
> |
common.logger.info('Limit on number of jobs reached.') |
| 682 |
> |
break |
| 683 |
> |
if (self.limitTotalLumis and lumisCreated >= self.totalNLumis): |
| 684 |
> |
common.logger.info('Limit on number of lumis reached.') |
| 685 |
> |
break |
| 686 |
> |
lumis = [] |
| 687 |
> |
lfns = [] |
| 688 |
> |
locations = [] |
| 689 |
> |
firstFile = True |
| 690 |
> |
# Collect information from all the files |
| 691 |
> |
for jobFile in job.getFiles(): |
| 692 |
> |
if firstFile: # Get locations from first file in the job |
| 693 |
> |
for loc in jobFile['locations']: |
| 694 |
> |
locations.append(loc) |
| 695 |
> |
firstFile = False |
| 696 |
> |
# Accumulate Lumis from all files |
| 697 |
> |
for lumiList in jobFile['runs']: |
| 698 |
> |
theRun = lumiList.run |
| 699 |
> |
for theLumi in list(lumiList): |
| 700 |
> |
lumis.append( (theRun, theLumi) ) |
| 701 |
> |
|
| 702 |
> |
lfns.append(jobFile['lfn']) |
| 703 |
> |
fileString = ','.join(lfns) |
| 704 |
> |
lumiString = compressLumiString(lumis) |
| 705 |
> |
list_of_lists.append([fileString, str(-1), str(0), lumiString]) |
| 706 |
> |
|
| 707 |
> |
jobDestination.append(locations) |
| 708 |
> |
jobCount += 1 |
| 709 |
> |
lumisCreated += len(lumis) |
| 710 |
> |
common.logger.debug('Job %s will run on %s files and %s lumis ' |
| 711 |
> |
% (jobCount, len(lfns), len(lumis) )) |
| 712 |
> |
|
| 713 |
> |
common.logger.info('%s jobs created to run on %s lumis' % |
| 714 |
> |
(jobCount, lumisCreated)) |
| 715 |
> |
|
| 716 |
> |
# Prepare dict output matching back to non-WMBS job creation |
| 717 |
> |
dictOut = {} |
| 718 |
> |
dictOut['params'] = ['InputFiles', 'MaxEvents', 'SkipEvents', 'Lumis'] |
| 719 |
> |
dictOut['args'] = list_of_lists |
| 720 |
> |
dictOut['jobDestination'] = jobDestination |
| 721 |
> |
dictOut['njobs'] = jobCount |
| 722 |
> |
|
| 723 |
> |
return dictOut |
| 724 |
> |
|
| 725 |
> |
|
| 726 |
|
def Algos(self): |
| 727 |
|
""" |
| 728 |
|
Define key splittingType matrix |
| 729 |
|
""" |
| 730 |
< |
SplitAlogs = { |
| 731 |
< |
'EventBased' : self.jobSplittingByEvent, |
| 730 |
> |
SplitAlogs = { |
| 731 |
> |
'EventBased' : self.jobSplittingByEvent, |
| 732 |
|
'RunBased' : self.jobSplittingByRun, |
| 733 |
< |
'LumiBased' : self.jobSplittingByLumi, |
| 734 |
< |
'NoInput' : self.jobSplittingNoInput, |
| 733 |
> |
'LumiBased' : self.jobSplittingByLumi, |
| 734 |
> |
'NoInput' : self.jobSplittingNoInput, |
| 735 |
|
'ForScript' : self.jobSplittingForScript |
| 736 |
< |
} |
| 736 |
> |
} |
| 737 |
|
return SplitAlogs |
| 738 |
|
|
| 739 |
+ |
|
| 740 |
+ |
|
| 741 |
+ |
def compressLumiString(lumis): |
| 742 |
+ |
""" |
| 743 |
+ |
Turn a list of 2-tuples of run/lumi numbers into a list of the format |
| 744 |
+ |
R1:L1,R2:L2-R3:L3 which is acceptable to CMSSW LumiBlockRange variable |
| 745 |
+ |
""" |
| 746 |
+ |
|
| 747 |
+ |
lumis.sort() |
| 748 |
+ |
parts = [] |
| 749 |
+ |
startRange = None |
| 750 |
+ |
endRange = None |
| 751 |
+ |
|
| 752 |
+ |
for lumiBlock in lumis: |
| 753 |
+ |
if not startRange: # This is the first one |
| 754 |
+ |
startRange = lumiBlock |
| 755 |
+ |
endRange = lumiBlock |
| 756 |
+ |
elif lumiBlock == endRange: # Same Lumi (different files?) |
| 757 |
+ |
pass |
| 758 |
+ |
elif lumiBlock[0] == endRange[0] and lumiBlock[1] == endRange[1] + 1: # This is a continuation |
| 759 |
+ |
endRange = lumiBlock |
| 760 |
+ |
else: # This is the start of a new range |
| 761 |
+ |
part = ':'.join(map(str, startRange)) |
| 762 |
+ |
if startRange != endRange: |
| 763 |
+ |
part += '-' + ':'.join(map(str, endRange)) |
| 764 |
+ |
parts.append(part) |
| 765 |
+ |
startRange = lumiBlock |
| 766 |
+ |
endRange = lumiBlock |
| 767 |
+ |
|
| 768 |
+ |
# Put out what's left |
| 769 |
+ |
if startRange: |
| 770 |
+ |
part = ':'.join(map(str, startRange)) |
| 771 |
+ |
if startRange != endRange: |
| 772 |
+ |
part += '-' + ':'.join(map(str, endRange)) |
| 773 |
+ |
parts.append(part) |
| 774 |
+ |
|
| 775 |
+ |
output = ','.join(parts) |
| 776 |
+ |
return output |
| 777 |
+ |
|
| 778 |
+ |
|
