| 8 |
|
from DBSAPI.dbsApiException import * |
| 9 |
|
import common |
| 10 |
|
from crab_util import * |
| 11 |
+ |
try: # Can remove when CMSSW 3.7 and earlier are dropped |
| 12 |
+ |
from FWCore.PythonUtilities.LumiList import LumiList |
| 13 |
+ |
except ImportError: |
| 14 |
+ |
from LumiList import LumiList |
| 15 |
+ |
|
| 16 |
|
import os |
| 17 |
|
|
| 18 |
|
|
| 93 |
|
# Attributes |
| 94 |
|
self.datasetPath = datasetPath |
| 95 |
|
# Analysis dataset is primary/processed/tier/definition |
| 96 |
< |
self.ads = len(self.datasetPath.split("/")) > 4 or len(self.datasetPath.split("/")) == 1 |
| 96 |
> |
self.ads = len(self.datasetPath.split("/")) > 4 |
| 97 |
|
self.cfg_params = cfg_params |
| 98 |
|
self.skipBlocks = skipAnBlocks |
| 99 |
|
|
| 106 |
|
self.maxLumis = 0 # DBS output: total number of lumis |
| 107 |
|
self.parent = {} # DBS output: parents of each file |
| 108 |
|
self.lumis = {} # DBS output: lumis in each file |
| 109 |
< |
|
| 109 |
> |
self.lumiMask = None |
| 110 |
> |
self.splitByLumi = False |
| 111 |
> |
self.splitDataByEvent = 0 |
| 112 |
|
|
| 113 |
|
def fetchDBSInfo(self): |
| 114 |
|
""" |
| 116 |
|
""" |
| 117 |
|
## get DBS URL |
| 118 |
|
global_url="http://cmsdbsprod.cern.ch/cms_dbs_prod_global/servlet/DBSServlet" |
| 119 |
< |
caf_url = "http://cmsdbsprod.cern.ch/cms_dbs_caf_analysis_01/servlet/DBSServlet" |
| 113 |
< |
dbs_url_map = {'glite': global_url, |
| 114 |
< |
'glite_slc5':global_url,\ |
| 115 |
< |
'glitecoll':global_url,\ |
| 116 |
< |
'condor': global_url,\ |
| 117 |
< |
'condor_g': global_url,\ |
| 118 |
< |
'glidein': global_url,\ |
| 119 |
< |
'lsf': global_url,\ |
| 120 |
< |
'caf': caf_url,\ |
| 121 |
< |
'sge': global_url,\ |
| 122 |
< |
'arc': global_url,\ |
| 123 |
< |
'pbs': global_url |
| 124 |
< |
} |
| 125 |
< |
|
| 126 |
< |
dbs_url_default = dbs_url_map[(common.scheduler.name()).lower()] |
| 127 |
< |
dbs_url= self.cfg_params.get('CMSSW.dbs_url', dbs_url_default) |
| 119 |
> |
dbs_url= self.cfg_params.get('CMSSW.dbs_url', global_url) |
| 120 |
|
common.logger.info("Accessing DBS at: "+dbs_url) |
| 121 |
|
|
| 122 |
|
## check if runs are selected |
| 123 |
|
runselection = [] |
| 124 |
|
if (self.cfg_params.has_key('CMSSW.runselection')): |
| 125 |
|
runselection = parseRange2(self.cfg_params['CMSSW.runselection']) |
| 126 |
< |
|
| 126 |
> |
if len(runselection)>1000000: |
| 127 |
> |
common.logger.info("ERROR: runselection range has more then 1M numbers") |
| 128 |
> |
common.logger.info("ERROR: Too large. runselection is ignored") |
| 129 |
> |
runselection=[] |
| 130 |
> |
|
| 131 |
> |
## check if various lumi parameters are set |
| 132 |
> |
self.lumiMask = self.cfg_params.get('CMSSW.lumi_mask',None) |
| 133 |
> |
self.lumiParams = self.cfg_params.get('CMSSW.total_number_of_lumis',None) or \ |
| 134 |
> |
self.cfg_params.get('CMSSW.lumis_per_job',None) |
| 135 |
> |
|
| 136 |
> |
lumiList = None |
| 137 |
> |
if self.lumiMask: |
| 138 |
> |
lumiList = LumiList(filename=self.lumiMask) |
| 139 |
> |
if runselection: |
| 140 |
> |
runList = LumiList(runs = runselection) |
| 141 |
|
|
| 142 |
|
self.splitByRun = int(self.cfg_params.get('CMSSW.split_by_run', 0)) |
| 143 |
< |
|
| 143 |
> |
self.splitDataByEvent = int(self.cfg_params.get('CMSSW.split_by_event', 0)) |
| 144 |
|
common.logger.log(10-1,"runselection is: %s"%runselection) |
| 145 |
+ |
|
| 146 |
+ |
if not self.splitByRun: |
| 147 |
+ |
self.splitByLumi = self.lumiMask or self.lumiParams or self.ads |
| 148 |
+ |
|
| 149 |
+ |
if self.splitByRun and not runselection: |
| 150 |
+ |
msg = "Error: split_by_run must be combined with a runselection" |
| 151 |
+ |
raise CrabException(msg) |
| 152 |
+ |
|
| 153 |
|
## service API |
| 154 |
|
args = {} |
| 155 |
|
args['url'] = dbs_url |
| 156 |
|
args['level'] = 'CRITICAL' |
| 143 |
– |
args['adshome'] = '$HOME/DBSADS' |
| 157 |
|
|
| 158 |
|
## check if has been requested to use the parent info |
| 159 |
|
useparent = int(self.cfg_params.get('CMSSW.use_parent',0)) |
| 165 |
|
api = DBSAPI.dbsApi.DbsApi(args) |
| 166 |
|
self.files = self.queryDbs(api,path=self.datasetPath,runselection=runselection,useParent=useparent) |
| 167 |
|
|
| 168 |
+ |
# Check to see what the dataset is |
| 169 |
+ |
pdsName = self.datasetPath.split("/")[1] |
| 170 |
+ |
primDSs = api.listPrimaryDatasets(pdsName) |
| 171 |
+ |
dataType = primDSs[0]['Type'] |
| 172 |
+ |
common.logger.debug("Datatype is %s" % dataType) |
| 173 |
+ |
if dataType == 'data' and not \ |
| 174 |
+ |
(self.splitByRun or self.splitByLumi or self.splitDataByEvent): |
| 175 |
+ |
msg = 'Data must be split by lumi or by run. ' \ |
| 176 |
+ |
'Please see crab -help for the correct settings' |
| 177 |
+ |
raise CrabException(msg) |
| 178 |
+ |
|
| 179 |
+ |
|
| 180 |
+ |
|
| 181 |
|
anFileBlocks = [] |
| 182 |
|
if self.skipBlocks: anFileBlocks = readTXTfile(self, fileBlocks_FileName) |
| 183 |
|
|
| 184 |
|
# parse files and fill arrays |
| 185 |
|
for file in self.files : |
| 186 |
|
parList = [] |
| 187 |
< |
lumiList = [] # List of tuples |
| 187 |
> |
fileLumis = [] # List of tuples |
| 188 |
|
# skip already analyzed blocks |
| 189 |
|
fileblock = file['Block']['Name'] |
| 190 |
|
if fileblock not in anFileBlocks : |
| 192 |
|
# asked retry the list of parent for the given child |
| 193 |
|
if useparent==1: |
| 194 |
|
parList = [x['LogicalFileName'] for x in file['ParentList']] |
| 195 |
< |
if self.ads: |
| 196 |
< |
lumiList = [ (x['RunNumber'], x['LumiSectionNumber']) |
| 195 |
> |
if self.splitByLumi: |
| 196 |
> |
fileLumis = [ (x['RunNumber'], x['LumiSectionNumber']) |
| 197 |
|
for x in file['LumiList'] ] |
| 198 |
|
self.parent[filename] = parList |
| 199 |
< |
self.lumis[filename] = lumiList |
| 199 |
> |
# For LumiMask, intersection of two lists. |
| 200 |
> |
if self.lumiMask and runselection: |
| 201 |
> |
self.lumis[filename] = runList.filterLumis(lumiList.filterLumis(fileLumis)) |
| 202 |
> |
elif runselection: |
| 203 |
> |
self.lumis[filename] = runList.filterLumis(fileLumis) |
| 204 |
> |
elif self.lumiMask: |
| 205 |
> |
self.lumis[filename] = lumiList.filterLumis(fileLumis) |
| 206 |
> |
else: |
| 207 |
> |
self.lumis[filename] = fileLumis |
| 208 |
|
if filename.find('.dat') < 0 : |
| 209 |
|
events = file['NumberOfEvents'] |
| 210 |
|
# Count number of events and lumis per block |
| 223 |
|
|
| 224 |
|
# total number of events |
| 225 |
|
self.maxEvents += events |
| 226 |
< |
self.maxLumis += len(lumiList) |
| 226 |
> |
self.maxLumis += len(self.lumis[filename]) |
| 227 |
|
|
| 228 |
|
if self.skipBlocks and len(self.eventsPerBlock.keys()) == 0: |
| 229 |
|
msg = "No new fileblocks available for dataset: "+str(self.datasetPath) |
| 230 |
|
raise CrabException(msg) |
| 231 |
|
|
| 198 |
– |
saveFblocks='' |
| 199 |
– |
for block in self.eventsPerBlock.keys() : |
| 200 |
– |
saveFblocks += str(block)+'\n' |
| 201 |
– |
common.logger.log(10-1,"DBSInfo: total nevts %i in block %s "%(self.eventsPerBlock[block],block)) |
| 202 |
– |
writeTXTfile(self, fileBlocks_FileName , saveFblocks) |
| 232 |
|
|
| 233 |
|
if len(self.eventsPerBlock) <= 0: |
| 234 |
|
raise NotExistingDatasetError(("\nNo data for %s in DBS\nPlease check" |
| 238 |
|
|
| 239 |
|
def queryDbs(self,api,path=None,runselection=None,useParent=None): |
| 240 |
|
|
| 241 |
< |
allowedRetriveValue = ['retrive_block', 'retrive_run'] |
| 242 |
< |
if self.ads: allowedRetriveValue.append('retrive_lumi') |
| 243 |
< |
if useParent == 1: allowedRetriveValue.append('retrive_parent') |
| 241 |
> |
|
| 242 |
> |
allowedRetriveValue = [] |
| 243 |
> |
if self.splitByLumi or self.splitByRun or useParent == 1: |
| 244 |
> |
allowedRetriveValue.extend(['retrive_block', 'retrive_run']) |
| 245 |
> |
if self.splitByLumi: |
| 246 |
> |
allowedRetriveValue.append('retrive_lumi') |
| 247 |
> |
if useParent == 1: |
| 248 |
> |
allowedRetriveValue.append('retrive_parent') |
| 249 |
|
common.logger.debug("Set of input parameters used for DBS query: %s" % allowedRetriveValue) |
| 250 |
|
try: |
| 251 |
< |
if len(runselection) <=0 : |
| 252 |
< |
if useParent==1 or self.splitByRun==1 or self.ads: |
| 219 |
< |
if self.ads: |
| 220 |
< |
files = api.listFiles(analysisDataset=path, retriveList=allowedRetriveValue) |
| 221 |
< |
else : |
| 222 |
< |
files = api.listFiles(path=path, retriveList=allowedRetriveValue) |
| 223 |
< |
else: |
| 224 |
< |
files = api.listDatasetFiles(self.datasetPath) |
| 225 |
< |
else : |
| 226 |
< |
files=[] |
| 251 |
> |
if self.splitByRun: |
| 252 |
> |
files = [] |
| 253 |
|
for arun in runselection: |
| 254 |
|
try: |
| 255 |
|
if self.ads: |
| 262 |
|
common.logger.info(msg) |
| 263 |
|
pass |
| 264 |
|
|
| 265 |
+ |
else: |
| 266 |
+ |
if allowedRetriveValue: |
| 267 |
+ |
if self.ads: |
| 268 |
+ |
files = api.listFiles(analysisDataset=path, retriveList=allowedRetriveValue) |
| 269 |
+ |
else : |
| 270 |
+ |
files = api.listFiles(path=path, retriveList=allowedRetriveValue) |
| 271 |
+ |
else: |
| 272 |
+ |
files = api.listDatasetFiles(self.datasetPath) |
| 273 |
+ |
|
| 274 |
|
except DbsBadRequest, msg: |
| 275 |
|
raise DataDiscoveryError(msg) |
| 276 |
|
except DBSError, msg: |
