ViewVC Help

View File | Revision Log | Show Annotations | Root Listing

root/cvsroot/COMP/CRAB/python/DataDiscovery.py

Comparing COMP/CRAB/python/DataDiscovery.py (file contents):
Revision 1.27 by spiga, Wed Feb 4 15:24:41 2009 UTC vs.
Revision 1.46 by spiga, Wed Jun 2 13:55:14 2010 UTC

#	Line 1 | Line 1
1		#!/usr/bin/env python
2	+
3	+	__revision__ = "$Id$"
4	+	__version__ = "$Revision$"
5	+
6		import exceptions
7		import DBSAPI.dbsApi
8	<	from DBSAPI.dbsApiException import *
8	>	from DBSAPI.dbsApiException import *
9		import common
10		from crab_util import *
11	<	import os
11	>	from LumiList import LumiList
12	>	import os
13	>
14
15
10	–	# #######################################
16		class DBSError(exceptions.Exception):
17		def __init__(self, errorName, errorMessage):
18		args='\nERROR DBS %s : %s \n'%(errorName,errorMessage)
19		exceptions.Exception.__init__(self, args)
20		pass
21	<
21	>
22		def getErrorMessage(self):
23		""" Return error message """
24		return "%s" % (self.args)
25
26	<	# #######################################
26	>
27	>
28		class DBSInvalidDataTierError(exceptions.Exception):
29		def __init__(self, errorName, errorMessage):
30		args='\nERROR DBS %s : %s \n'%(errorName,errorMessage)
31		exceptions.Exception.__init__(self, args)
32		pass
33	<
33	>
34		def getErrorMessage(self):
35		""" Return error message """
36		return "%s" % (self.args)
37
38	<	# #######################################
38	>
39	>
40		class DBSInfoError:
41		def __init__(self, url):
42		print '\nERROR accessing DBS url : '+url+'\n'
43		pass
44
45	<	# ####################################
45	>
46	>
47		class DataDiscoveryError(exceptions.Exception):
48		def __init__(self, errorMessage):
49		self.args=errorMessage
#	Line 46 | Line 54 | class DataDiscoveryError(exceptions.Exce
54		""" Return exception error """
55		return "%s" % (self.args)
56
57	<	# ####################################
57	>
58	>
59		class NotExistingDatasetError(exceptions.Exception):
60		def __init__(self, errorMessage):
61		self.args=errorMessage
#	Line 57 | Line 66 | class NotExistingDatasetError(exceptions
66		""" Return exception error """
67		return "%s" % (self.args)
68
69	<	# ####################################
69	>
70	>
71		class NoDataTierinProvenanceError(exceptions.Exception):
72		def __init__(self, errorMessage):
73		self.args=errorMessage
#	Line 68 | Line 78 | class NoDataTierinProvenanceError(except
78		""" Return exception error """
79		return "%s" % (self.args)
80
81	<	# ####################################
82	<	# class to find and extact info from published data
81	>
82	>
83		class DataDiscovery:
84	+	"""
85	+	Class to find and extact info from published data
86	+	"""
87		def __init__(self, datasetPath, cfg_params, skipAnBlocks):
88
89		#       Attributes
90		self.datasetPath = datasetPath
91	+	# Analysis dataset is primary/processed/tier/definition
92	+	self.ads = len(self.datasetPath.split("/")) > 4
93		self.cfg_params = cfg_params
94		self.skipBlocks = skipAnBlocks
95
96		self.eventsPerBlock = {}  # DBS output: map fileblocks-events for collection
97		self.eventsPerFile = {}   # DBS output: map files-events
98	<	self.blocksinfo = {}      # DBS output: map fileblocks-files
98	>	#         self.lumisPerBlock = {}   # DBS output: number of lumis in each block
99	>	#         self.lumisPerFile = {}    # DBS output: number of lumis in each file
100	>	self.blocksinfo = {}      # DBS output: map fileblocks-files
101		self.maxEvents = 0        # DBS output: max events
102	<	self.parent = {}       # DBS output: max events
102	>	self.maxLumis = 0         # DBS output: total number of lumis
103	>	self.parent = {}          # DBS output: parents of each file
104	>	self.lumis = {}           # DBS output: lumis in each file
105	>	self.lumiMask = None
106	>	self.splitByLumi = False
107
87	–	# ####################################
108		def fetchDBSInfo(self):
109		"""
110		Contact DBS
111		"""
112		## get DBS URL
113		global_url="http://cmsdbsprod.cern.ch/cms_dbs_prod_global/servlet/DBSServlet"
114	<	caf_url = "http://cmsdbsprod.cern.ch/cms_dbs_caf_analysis_01/servlet/DBSServlet"
115	<	dbs_url_map  =   {'glite':    global_url,
96	<	'glitecoll':global_url,\
97	<	'condor':   global_url,\
98	<	'condor_g': global_url,\
99	<	'glidein':  global_url,\
100	<	'lsf':      global_url,\
101	<	'caf':      caf_url,\
102	<	'sge':      global_url
103	<	}
104	<
105	<	dbs_url_default = dbs_url_map[(common.scheduler.name()).lower()]
106	<	dbs_url=  self.cfg_params.get('CMSSW.dbs_url', dbs_url_default)
107	<	common.logger.debug(3,"Accessing DBS at: "+dbs_url)
114	>	dbs_url=  self.cfg_params.get('CMSSW.dbs_url', global_url)
115	>	common.logger.info("Accessing DBS at: "+dbs_url)
116
117		## check if runs are selected
118		runselection = []
119		if (self.cfg_params.has_key('CMSSW.runselection')):
120		runselection = parseRange2(self.cfg_params['CMSSW.runselection'])
121
122	+	## check if various lumi parameters are set
123	+	self.lumiMask = self.cfg_params.get('CMSSW.lumi_mask',None)
124	+	self.lumiParams = self.cfg_params.get('CMSSW.total_number_of_lumis',None) or \
125	+	self.cfg_params.get('CMSSW.lumis_per_job',None)
126	+
127	+	lumiList = None
128	+	if self.lumiMask:
129	+	lumiList = LumiList(filename=self.lumiMask)
130	+	if runselection:
131	+	runList = LumiList(runs = runselection)
132
133		self.splitByRun = int(self.cfg_params.get('CMSSW.split_by_run', 0))
134	<
135	<	self.ads = int(self.cfg_params.get('CMSSW.ads', 0))
134	>	common.logger.log(10-1,"runselection is: %s"%runselection)
135	>
136	>	if not self.splitByRun:
137	>	self.splitByLumi = self.lumiMask or self.lumiParams or self.ads
138
119	–	common.logger.debug(6,"runselection is: %s"%runselection)
139		## service API
140		args = {}
141		args['url']     = dbs_url
#	Line 125 | Line 144 | class DataDiscovery:
144		## check if has been requested to use the parent info
145		useparent = int(self.cfg_params.get('CMSSW.use_parent',0))
146
147	<	## check if has been asked for a non default file to store/read analyzed fileBlocks
148	<	defaultName = common.work_space.shareDir()+'AnalyzedBlocks.txt'
147	>	## check if has been asked for a non default file to store/read analyzed fileBlocks
148	>	defaultName = common.work_space.shareDir()+'AnalyzedBlocks.txt'
149		fileBlocks_FileName = os.path.abspath(self.cfg_params.get('CMSSW.fileblocks_file',defaultName))
131	–
132	–	api = DBSAPI.dbsApi.DbsApi(args)
150
151	+	api = DBSAPI.dbsApi.DbsApi(args)
152		self.files = self.queryDbs(api,path=self.datasetPath,runselection=runselection,useParent=useparent)
153
154	+	# Check to see what the dataset is
155	+	pdsName = self.datasetPath.split("/")[1]
156	+	primDSs = api.listPrimaryDatasets(pdsName)
157	+	dataType = primDSs[0]['Type']
158	+	common.logger.debug("Datatype is %s" % dataType)
159	+	if dataType == 'data' and not (self.splitByRun or self.splitByLumi):
160	+	msg = 'Data must be split by lumi or by run. ' \
161	+	'Please see crab -help for the correct settings'
162	+	raise  CrabException(msg)
163	+
164	+
165	+
166		anFileBlocks = []
167	<	if self.skipBlocks: anFileBlocks = readTXTfile(self, fileBlocks_FileName)
167	>	if self.skipBlocks: anFileBlocks = readTXTfile(self, fileBlocks_FileName)
168
169		# parse files and fill arrays
170		for file in self.files :
171	<	parList = []
171	>	parList  = []
172	>	fileLumis = [] # List of tuples
173		# skip already analyzed blocks
174		fileblock = file['Block']['Name']
175		if fileblock not in anFileBlocks :
176		filename = file['LogicalFileName']
177	<	# asked retry the list of parent for the given child
178	<	if useparent==1: parList = [x['LogicalFileName'] for x in file['ParentList']]
179	<	self.parent[filename] = parList
177	>	# asked retry the list of parent for the given child
178	>	if useparent==1:
179	>	parList = [x['LogicalFileName'] for x in file['ParentList']]
180	>	if self.splitByLumi:
181	>	fileLumis = [ (x['RunNumber'], x['LumiSectionNumber'])
182	>	for x in file['LumiList'] ]
183	>	self.parent[filename] = parList
184	>	# For LumiMask, intersection of two lists.
185	>	if self.lumiMask and runselection:
186	>	self.lumis[filename] = runList.filterLumis(lumiList.filterLumis(fileLumis))
187	>	elif runselection:
188	>	self.lumis[filename] = runList.filterLumis(fileLumis)
189	>	elif self.lumiMask:
190	>	self.lumis[filename] = lumiList.filterLumis(fileLumis)
191	>	else:
192	>	self.lumis[filename] = fileLumis
193		if filename.find('.dat') < 0 :
194		events    = file['NumberOfEvents']
195	<	# number of events per block
195	>	# Count number of events and lumis per block
196		if fileblock in self.eventsPerBlock.keys() :
197		self.eventsPerBlock[fileblock] += events
198		else :
199		self.eventsPerBlock[fileblock] = events
200	<	# number of events per file
200	>	# Number of events per file
201		self.eventsPerFile[filename] = events
202	<
203	<	# number of events per block
202	>
203	>	# List of files per block
204		if fileblock in self.blocksinfo.keys() :
205		self.blocksinfo[fileblock].append(filename)
206		else :
207		self.blocksinfo[fileblock] = [filename]
208	<
208	>
209		# total number of events
210		self.maxEvents += events
211	+	self.maxLumis  += len(self.lumis[filename])
212	+
213		if  self.skipBlocks and len(self.eventsPerBlock.keys()) == 0:
214		msg = "No new fileblocks available for dataset: "+str(self.datasetPath)
215	<	raise  CrabException(msg)
215	>	raise  CrabException(msg)
216	>
217
171	–	saveFblocks=''
172	–	for block in self.eventsPerBlock.keys() :
173	–	saveFblocks += str(block)+'\n'
174	–	common.logger.debug(6,"DBSInfo: total nevts %i in block %s "%(self.eventsPerBlock[block],block))
175	–	writeTXTfile(self, fileBlocks_FileName , saveFblocks)
176	–
218		if len(self.eventsPerBlock) <= 0:
219		raise NotExistingDatasetError(("\nNo data for %s in DBS\nPlease check"
220		+ " dataset path variables in crab.cfg")
221		% self.datasetPath)
222
223
183	–	###########################
184	–
224		def queryDbs(self,api,path=None,runselection=None,useParent=None):
225	<
226	<	allowedRetriveValue = [#'retrive_parent',
227	<	'retrive_block',
228	<	#'retrive_lumi',
229	<	'retrive_run'
230	<	]
225	>
226	>	allowedRetriveValue = ['retrive_block', 'retrive_run']
227	>	if self.ads or self.lumiMask or self.lumiParams:
228	>	allowedRetriveValue.append('retrive_lumi')
229	>	if useParent == 1: allowedRetriveValue.append('retrive_parent')
230	>	common.logger.debug("Set of input parameters used for DBS query: %s" % allowedRetriveValue)
231		try:
232	<	if len(runselection) <=0 :
233	<	if useParent==1 or self.splitByRun==1 :
234	<	if self.ads==1 :
235	<	files = api.listFiles(analysisDataset=path, retriveList=allowedRetriveValue)
232	>	if len(runselection) <=0 or self.splitByLumi:
233	>	if useParent==1 or self.splitByRun==1 or self.splitByLumi:
234	>	if self.ads:
235	>	files = api.listFiles(analysisDataset=path, retriveList=allowedRetriveValue)
236		else :
237	<	files = api.listFiles(path=path, retriveList=allowedRetriveValue)
199	<	common.logger.debug(5,"Set of input parameters used for DBS query : \n"+str(allowedRetriveValue))
200	<	common.logger.write("Set of input parameters used for DBS query : \n"+str(allowedRetriveValue))
237	>	files = api.listFiles(path=path, retriveList=allowedRetriveValue)
238		else:
202	–	print 'MALE2'
239		files = api.listDatasetFiles(self.datasetPath)
240		else :
241		files=[]
242		for arun in runselection:
243		try:
244	<	if self.ads==1 : filesinrun = api.listFiles(analysisDataset=path,retriveList=allowedRetriveValue,runNumber=arun)
245	<	else: filesinrun = api.listFiles(path=path,retriveList=allowedRetriveValue,runNumber=arun)
244	>	if self.ads:
245	>	filesinrun = api.listFiles(analysisDataset=path,retriveList=allowedRetriveValue,runNumber=arun)
246	>	else:
247	>	filesinrun = api.listFiles(path=path,retriveList=allowedRetriveValue,runNumber=arun)
248		files.extend(filesinrun)
249		except:
250		msg="WARNING: problem extracting info from DBS for run %s "%arun
251	<	common.logger.message(msg)
251	>	common.logger.info(msg)
252		pass
253
254		except DbsBadRequest, msg:
#	Line 220 | Line 258 | class DataDiscovery:
258
259		return files
260
261	<	# #################################################
261	>
262		def getMaxEvents(self):
263		"""
264	<	max events
264	>	max events
265		"""
266		return self.maxEvents
267
268	<	# #################################################
268	>
269	>	def getMaxLumis(self):
270	>	"""
271	>	Return the number of lumis in the dataset
272	>	"""
273	>	return self.maxLumis
274	>
275	>
276		def getEventsPerBlock(self):
277		"""
278	<	list the event collections structure by fileblock
278	>	list the event collections structure by fileblock
279		"""
280		return self.eventsPerBlock
281
282	<	# #################################################
282	>
283		def getEventsPerFile(self):
284		"""
285	<	list the event collections structure by file
285	>	list the event collections structure by file
286		"""
287		return self.eventsPerFile
288
289	<	# #################################################
289	>
290		def getFiles(self):
291		"""
292	<	return files grouped by fileblock
292	>	return files grouped by fileblock
293		"""
294	<	return self.blocksinfo
294	>	return self.blocksinfo
295	>
296
251	–	# #################################################
297		def getParent(self):
298		"""
299	<	return parent grouped by file
299	>	return parent grouped by file
300		"""
301	<	return self.parent
301	>	return self.parent
302
303	<	# #################################################
304	<	def getListFiles(self):
303	>
304	>	def getLumis(self):
305		"""
306	<	return parent grouped by file
306	>	return lumi sections grouped by file
307		"""
308	<	return self.files
308	>	return self.lumis
309	>
310
311	<	########################################################################
311	>	def getListFiles(self):
312	>	"""
313	>	return parent grouped by file
314	>	"""
315	>	return self.files

Diff Legend

–	Removed lines
+	Added lines
<	Changed lines
>	Changed lines