ViewVC Help

View File | Revision Log | Show Annotations | Root Listing

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

Comparing COMP/CRAB/python/Publisher.py (file contents):
Revision 1.19 by afanfani, Mon Nov 17 13:02:07 2008 UTC vs.
Revision 1.49 by fanzago, Fri May 28 09:46:00 2010 UTC

#	Line 3 | Line 3 | import common
3		import time, glob
4		from Actor import *
5		from crab_util import *
6	–	from crab_logger import Logger
6		from crab_exceptions import *
7		from ProdCommon.FwkJobRep.ReportParser import readJobReport
8		from ProdCommon.FwkJobRep.ReportState import checkSuccess
#	Line 13 | Line 12 | from ProdCommon.DataMgmt.DBS.DBSErrors i
12		from ProdCommon.DataMgmt.DBS.DBSReader import DBSReader
13		from ProdCommon.DataMgmt.DBS.DBSWriter import DBSWriter,DBSWriterObjects
14		import sys
15	+	from DBSAPI.dbsApi import DbsApi
16	+	from DBSAPI.dbsMigrateApi import DbsMigrateApi
17	+	from DBSAPI.dbsApiException import *
18
19		class Publisher(Actor):
20		def __init__(self, cfg_params):
21		"""
22	<	Publisher class:
22	>	Publisher class:
23
24		- parses CRAB FrameworkJobReport on UI
25		- returns <file> section of xml in dictionary format for each xml file in crab_0_xxxx/res directory
26		- publishes output data on DBS and DLS
27		"""
28
29	<	try:
30	<	userprocessedData = cfg_params['USER.publish_data_name']
31	<	self.processedData = None
30	<	except KeyError:
29	>	self.cfg_params=cfg_params
30	>
31	>	if not cfg_params.has_key('USER.publish_data_name'):
32		raise CrabException('Cannot publish output data, because you did not specify USER.publish_data_name parameter in the crab.cfg file')
33	+	self.userprocessedData = cfg_params['USER.publish_data_name']
34	+	self.processedData = None
35
36	<	try:
37	<	if (int(cfg_params['USER.copy_data']) != 1): raise KeyError
38	<	except KeyError:
39	<	raise CrabException('You can not publish data because you did not selected *** copy_data = 1  *** in the crab.cfg file')
40	<	try:
41	<	self.pset = cfg_params['CMSSW.pset']
42	<	except KeyError:
36	>	if (not cfg_params.has_key('USER.copy_data') or int(cfg_params['USER.copy_data']) != 1 ) or \
37	>	(not cfg_params.has_key('USER.publish_data') or int(cfg_params['USER.publish_data']) != 1 ):
38	>	msg  = 'You can not publish data because you did not selected \n'
39	>	msg += '\t*** copy_data = 1 and publish_data = 1  *** in the crab.cfg file'
40	>	raise CrabException(msg)
41	>
42	>	if not cfg_params.has_key('CMSSW.pset'):
43		raise CrabException('Cannot publish output data, because you did not specify the psetname in [CMSSW] of your crab.cfg file')
44	<	try:
45	<	self.globalDBS=cfg_params['CMSSW.dbs_url']
46	<	except KeyError:
47	<	self.globalDBS="http://cmsdbsprod.cern.ch/cms_dbs_prod_global/servlet/DBSServlet"
48	<	try:
46	<	self.DBSURL=cfg_params['USER.dbs_url_for_publication']
47	<	common.logger.message('<dbs_url_for_publication> = '+self.DBSURL)
48	<	if (self.DBSURL == "http://cmsdbsprod.cern.ch/cms_dbs_prod_global/servlet/DBSServlet") or (self.DBSURL == "https://cmsdbsprod.cern.ch:8443/cms_dbs_prod_global_writer/servlet/DBSServlet"):
49	<	msg = "You can not publish your data in the globalDBS = " + self.DBSURL + "\n"
50	<	msg = msg + "Please write your local one in the [USER] section 'dbs_url_for_publication'"
51	<	raise CrabException(msg)
52	<	except KeyError:
44	>	self.pset = cfg_params['CMSSW.pset']
45	>
46	>	self.globalDBS=cfg_params.get('CMSSW.dbs_url',"http://cmsdbsprod.cern.ch/cms_dbs_prod_global/servlet/DBSServlet")
47	>
48	>	if not cfg_params.has_key('USER.dbs_url_for_publication'):
49		msg = "Warning. The [USER] section does not have 'dbs_url_for_publication'"
50		msg = msg + " entry, necessary to publish the data.\n"
51		msg = msg + "Use the command **crab -publish -USER.dbs_url_for_publication=dbs_url_for_publication*** \nwhere dbs_url_for_publication is your local dbs instance."
52		raise CrabException(msg)
53	+
54	+	self.DBSURL=cfg_params['USER.dbs_url_for_publication']
55	+	common.logger.info('<dbs_url_for_publication> = '+self.DBSURL)
56	+	if (self.DBSURL == "http://cmsdbsprod.cern.ch/cms_dbs_prod_global/servlet/DBSServlet") or (self.DBSURL == "https://cmsdbsprod.cern.ch:8443/cms_dbs_prod_global_writer/servlet/DBSServlet"):
57	+	msg = "You can not publish your data in the globalDBS = " + self.DBSURL + "\n"
58	+	msg = msg + "Please write your local one in the [USER] section 'dbs_url_for_publication'"
59	+	raise CrabException(msg)
60
61		self.content=file(self.pset).read()
62		self.resDir = common.work_space.resDir()
#	Line 72 | Line 75 | class Publisher(Actor):
75		dataset=string.strip(dataset)
76		self.dataset_to_import.append(dataset)
77		###
78	<
78	>
79	>	self.import_all_parents = cfg_params.get('USER.publish_with_import_all_parents',1)
80	>
81	>	if ( int(self.import_all_parents) == 0 ):
82	>	common.logger.info("WARNING: The option USER.publish_with_import_all_parents=0 has been deprecated. The import of parents is compulsory and done by default")
83	>	self.skipOcheck=cfg_params.get('CMSSW.publish_zero_event',1)
84	>	if ( int(self.skipOcheck) == 0 ):
85	>	common.logger.info("WARNING: The option CMSSW.publish_zero_event has been deprecated. The publication is done by default also for files with 0 events")
86		self.SEName=''
87		self.CMSSW_VERSION=''
88		self.exit_status=''
89		self.time = time.strftime('%y%m%d_%H%M%S',time.localtime(time.time()))
90	<	self.problemFiles=[]
90	>	self.problemFiles=[]
91		self.noEventsFiles=[]
92		self.noLFN=[]
93	<
93	>
94		def importParentDataset(self,globalDBS, datasetpath):
95		"""
96	<	"""
97	<	dbsWriter = DBSWriter(self.DBSURL,level='ERROR')
98	<
96	>	WARNING: it works only with DBS_2_0_9_patch_6
97	>	"""
98	>
99	>	args={'url':globalDBS}
100		try:
101	<	dbsWriter.importDatasetWithoutParentage(globalDBS, self.datasetpath, self.DBSURL)
102	<	except DBSWriterError, ex:
101	>	api_reader = DbsApi(args)
102	>	except DbsApiException, ex:
103	>	msg = "%s\n" % formatEx(ex)
104	>	raise CrabException(msg)
105	>
106	>	args={'url':self.DBSURL}
107	>	try:
108	>	api_writer = DbsApi(args)
109	>	except DbsApiException, ex:
110	>	msg = "%s\n" % formatEx(ex)
111	>	raise CrabException(msg)
112	>
113	>	try:
114	>	common.logger.info("--->>> Importing all parents level")
115	>	start = time.time()
116	>	common.logger.debug("start import parents time: " + str(start))
117	>	for block in api_reader.listBlocks(datasetpath):
118	>	api_writer.dbsMigrateBlock(globalDBS,self.DBSURL,block['Name'] )
119	>	stop = time.time()
120	>	common.logger.debug("stop import parents time: " + str(stop))
121	>	common.logger.info("--->>> duration of all parents import (sec): "+str(stop - start))
122	>	except DbsApiException, ex:
123		msg = "Error importing dataset to be processed into local DBS\n"
124		msg += "Source Dataset: %s\n" % datasetpath
125		msg += "Source DBS: %s\n" % globalDBS
126		msg += "Destination DBS: %s\n" % self.DBSURL
127	<	common.logger.message(msg)
127	>	common.logger.info(msg)
128	>	common.logger.info(str(ex))
129		return 1
130		return 0
131	<
131	>
132		def publishDataset(self,file):
133		"""
134		"""
#	Line 105 | Line 137 | class Publisher(Actor):
137		self.exit_status = '0'
138		except IndexError:
139		self.exit_status = '1'
140	<	msg = "Error: Problem with "+file+" file"
141	<	common.logger.message(msg)
140	>	msg = "Error: Problem with "+file+" file"
141	>	common.logger.info(msg)
142		return self.exit_status
143
144		if (len(self.dataset_to_import) != 0):
145		for dataset in self.dataset_to_import:
146	<	common.logger.message("--->>> Importing parent dataset in the dbs: " +dataset)
146	>	common.logger.info("--->>> Importing parent dataset in the dbs: " +dataset)
147		status_import=self.importParentDataset(self.globalDBS, dataset)
148		if (status_import == 1):
149	<	common.logger.message('Problem with parent '+ dataset +' import from the global DBS '+self.globalDBS+ 'to the local one '+self.DBSURL)
149	>	common.logger.info('Problem with parent '+ dataset +' import from the global DBS '+self.globalDBS+ 'to the local one '+self.DBSURL)
150		self.exit_status='1'
151		return self.exit_status
152	<	else:
153	<	common.logger.message('Import ok of dataset '+dataset)
154	<
155	<	#// DBS to contact
156	<	dbswriter = DBSWriter(self.DBSURL)
125	<	try:
126	<	fileinfo= jobReport.files[0]
127	<	self.exit_status = '0'
128	<	except IndexError:
152	>	else:
153	>	common.logger.info('Import ok of dataset '+dataset)
154	>
155	>
156	>	if (len(jobReport.files) <= 0) :
157		self.exit_status = '1'
158	<	msg = "Error: No file to publish in xml file"+file+" file"
159	<	common.logger.message(msg)
158	>	msg = "Error: No EDM file to publish in xml file"+file+" file"
159	>	common.logger.info(msg)
160		return self.exit_status
161	+	else:
162	+	msg = "fjr contains some files to publish"
163	+	common.logger.debug(msg)
164
165	<	datasets=fileinfo.dataset
166	<	common.logger.debug(6,"FileInfo = " + str(fileinfo))
167	<	common.logger.debug(6,"DatasetInfo = " + str(datasets))
168	<	if len(datasets)<=0:
169	<	self.exit_status = '1'
170	<	msg = "Error: No info about dataset in the xml file "+file
171	<	common.logger.message(msg)
172	<	return self.exit_status
173	<	for dataset in datasets:
174	<	#### for production data
175	<	self.processedData = dataset['ProcessedDataset']
176	<	if (dataset['PrimaryDataset'] == 'null'):
177	<	dataset['PrimaryDataset'] = dataset['ProcessedDataset']
178	<	else: # add parentage from input dataset
179	<	dataset['ParentDataset']= self.datasetpath
180	<
181	<	dataset['PSetContent']=self.content
182	<	cfgMeta = {'name' : self.pset , 'Type' : 'user' , 'annotation': 'user cfg', 'version' : 'private version'} # add real name of user cfg
183	<	common.logger.message("PrimaryDataset = %s"%dataset['PrimaryDataset'])
184	<	common.logger.message("ProcessedDataset = %s"%dataset['ProcessedDataset'])
185	<	common.logger.message("<User Dataset Name> = /"+dataset['PrimaryDataset']+"/"+dataset['ProcessedDataset']+"/USER")
186	<
187	<	common.logger.debug(6,"--->>> Inserting primary: %s processed : %s"%(dataset['PrimaryDataset'],dataset['ProcessedDataset']))
188	<
189	<	primary = DBSWriterObjects.createPrimaryDataset( dataset, dbswriter.dbs)
190	<	common.logger.debug(6,"Primary:  %s "%primary)
191	<
192	<	algo = DBSWriterObjects.createAlgorithm(dataset, cfgMeta, dbswriter.dbs)
193	<	common.logger.debug(6,"Algo:  %s "%algo)
165	>	#### datasets creation in dbs
166	>	#// DBS to contact write and read of the same dbs
167	>	dbsReader = DBSReader(self.DBSURL,level='ERROR')
168	>	dbswriter = DBSWriter(self.DBSURL)
169	>	#####
170	>
171	>	self.published_datasets = []
172	>	for fileinfo in jobReport.files:
173	>	datasets_info=fileinfo.dataset
174	>	if len(datasets_info)<=0:
175	>	self.exit_status = '1'
176	>	msg = "Error: No info about dataset in the xml file "+file
177	>	common.logger.info(msg)
178	>	return self.exit_status
179	>	else:
180	>	for dataset in datasets_info:
181	>	#### for production data
182	>	self.processedData = dataset['ProcessedDataset']
183	>	if (dataset['PrimaryDataset'] == 'null'):
184	>	dataset['PrimaryDataset'] = self.userprocessedData
185	>	elif self.datasetpath.upper() != 'NONE':
186	>	dataset['ParentDataset']= self.datasetpath
187	>
188	>	dataset['PSetContent']=self.content
189	>	cfgMeta = {'name' : self.pset , 'Type' : 'user' , 'annotation': 'user cfg', 'version' : 'private version'} # add real name of user cfg
190	>	common.logger.info("PrimaryDataset = %s"%dataset['PrimaryDataset'])
191	>	common.logger.info("ProcessedDataset = %s"%dataset['ProcessedDataset'])
192	>	common.logger.info("<User Dataset Name> = /"+dataset['PrimaryDataset']+"/"+dataset['ProcessedDataset']+"/USER")
193	>
194	>	self.dataset_to_check="/"+dataset['PrimaryDataset']+"/"+dataset['ProcessedDataset']+"/USER"
195
196	<	processed = DBSWriterObjects.createProcessedDataset(primary, algo, dataset, dbswriter.dbs)
197	<	common.logger.debug(6,"Processed:  %s "%processed)
198	<
199	<	common.logger.debug(6,"Inserted primary %s processed %s"%(primary,processed))
200	<
201	<	common.logger.debug(6,"exit_status = %s "%self.exit_status)
202	<	return self.exit_status
196	>
197	>	self.published_datasets.append(self.dataset_to_check)
198	>
199	>	common.logger.log(10-1,"--->>> Inserting primary: %s processed : %s"%(dataset['PrimaryDataset'],dataset['ProcessedDataset']))
200	>
201	>	#### check if dataset already exists in the DBS
202	>	result = dbsReader.matchProcessedDatasets(dataset['PrimaryDataset'], 'USER', dataset['ProcessedDataset'])
203	>	if (len(result) != 0):
204	>	result = dbsReader.listDatasetFiles(self.dataset_to_check)
205	>
206	>	primary = DBSWriterObjects.createPrimaryDataset( dataset, dbswriter.dbs)
207	>	common.logger.log(10-1,"Primary:  %s "%primary)
208	>	print "primary = ", primary
209	>
210	>	algo = DBSWriterObjects.createAlgorithm(dataset, cfgMeta, dbswriter.dbs)
211	>	common.logger.log(10-1,"Algo:  %s "%algo)
212	>
213	>	processed = DBSWriterObjects.createProcessedDataset(primary, algo, dataset, dbswriter.dbs)
214	>	common.logger.log(10-1,"Processed:  %s "%processed)
215	>	print "processed = ", processed
216	>
217	>	common.logger.log(10-1,"Inserted primary %s processed %s"%(primary,processed))
218	>	#######################################################################################
219	>
220	>	common.logger.log(10-1,"exit_status = %s "%self.exit_status)
221	>	return self.exit_status
222
223		def publishAJobReport(self,file,procdataset):
224		"""
225		input:  xml file, processedDataset
226		"""
227	+	common.logger.debug("FJR = %s"%file)
228		try:
229		jobReport = readJobReport(file)[0]
230		self.exit_status = '0'
#	Line 180 | Line 232 | class Publisher(Actor):
232		self.exit_status = '1'
233		msg = "Error: Problem with "+file+" file"
234		raise CrabException(msg)
183	–	### overwrite ProcessedDataset with user defined value
184	–	### overwrite lumisections with no value
235		### skip publication for 0 events files
236		filestopublish=[]
237		for file in jobReport.files:
#	Line 191 | Line 241 | class Publisher(Actor):
241		elif (file['LFN'] == ''):
242		self.noLFN.append(file['PFN'])
243		else:
244	<	if int(file['TotalEvents']) != 0 :
195	<	#file.lumisections = {}
196	<	# lumi info are now in run hash
197	<	file.runs = {}
198	<	for ds in file.dataset:
199	<	### FEDE FOR NEW LFN ###
200	<	#ds['ProcessedDataset']=procdataset
201	<	########################
202	<	### Fede for production
203	<	if (ds['PrimaryDataset'] == 'null'):
204	<	ds['PrimaryDataset']=procdataset
205	<	filestopublish.append(file)
206	<	else:
244	>	if int(file['TotalEvents']) == 0:
245		self.noEventsFiles.append(file['LFN'])
246	+	for ds in file.dataset:
247	+	### Fede for production
248	+	if (ds['PrimaryDataset'] == 'null'):
249	+	ds['PrimaryDataset']=self.userprocessedData
250	+	filestopublish.append(file)
251	+
252		jobReport.files = filestopublish
253	+	for file in filestopublish:
254	+	common.logger.debug("--->>> LFN of file to publish =  " + str(file['LFN']))
255		### if all files of FJR have number of events = 0
256		if (len(filestopublish) == 0):
257	<	return None
257	>	return None
258
259		#// DBS to contact
260		dbswriter = DBSWriter(self.DBSURL)
261		# insert files
262		Blocks=None
263		try:
264	<	Blocks=dbswriter.insertFiles(jobReport)
265	<	common.logger.message("Inserting file in blocks = %s"%Blocks)
264	>	### FEDE added insertDetectorData = True to propagate in DBS info about run and lumi
265	>	Blocks=dbswriter.insertFiles(jobReport, insertDetectorData = True)
266	>	#Blocks=dbswriter.insertFiles(jobReport)
267	>	common.logger.debug("--->>> Inserting file in blocks = %s"%Blocks)
268		except DBSWriterError, ex:
269	<	common.logger.error("Insert file error: %s"%ex)
269	>	common.logger.debug("--->>> Insert file error: %s"%ex)
270		return Blocks
271
272		def run(self):
#	Line 227 | Line 275 | class Publisher(Actor):
275		"""
276
277		file_list = glob.glob(self.resDir+"crab_fjr*.xml")
278	+
279		## Select only those fjr that are succesfull
280	+	if (len(file_list)==0):
281	+	common.logger.info("--->>> "+self.resDir+" empty: no file to publish on DBS")
282	+	self.exit_status = '1'
283	+	return self.exit_status
284	+
285		good_list=[]
286		for fjr in file_list:
287		reports = readJobReport(fjr)
288	<	if reports[0].status == "Success":
289	<	good_list.append(fjr)
288	>	if len(reports)>0:
289	>	if reports[0].status == "Success":
290	>	good_list.append(fjr)
291		file_list=good_list
292		##
293	<	common.logger.debug(6, "file_list = "+str(file_list))
294	<	common.logger.debug(6, "len(file_list) = "+str(len(file_list)))
295	<
293	>	common.logger.log(10-1, "fjr with FrameworkJobReport Status='Success', file_list = "+str(file_list))
294	>	common.logger.log(10-1, "len(file_list) = "+str(len(file_list)))
295	>
296		if (len(file_list)>0):
297		BlocksList=[]
298	<	common.logger.message("--->>> Start dataset publication")
298	>	common.logger.info("--->>> Start dataset publication")
299		self.exit_status=self.publishDataset(file_list[0])
300		if (self.exit_status == '1'):
301	<	return self.exit_status
302	<	common.logger.message("--->>> End dataset publication")
301	>	return self.exit_status
302	>	common.logger.info("--->>> End dataset publication")
303
304
305	<	common.logger.message("--->>> Start files publication")
305	>	common.logger.info("--->>> Start files publication")
306		for file in file_list:
252	–	common.logger.debug(1, "file = "+file)
307		Blocks=self.publishAJobReport(file,self.processedData)
308		if Blocks:
309		for x in Blocks: # do not allow multiple entries of the same block
310		if x not in BlocksList:
311		BlocksList.append(x)
312	<
312	>
313		# close the blocks
314	<	common.logger.debug(6, "BlocksList = %s"%BlocksList)
261	<	# dbswriter = DBSWriter(self.DBSURL,level='ERROR')
314	>	common.logger.log(10-1, "BlocksList = %s"%BlocksList)
315		dbswriter = DBSWriter(self.DBSURL)
316	<
316	>
317		for BlockName in BlocksList:
318	<	try:
318	>	try:
319		closeBlock=dbswriter.manageFileBlock(BlockName,maxFiles= 1)
320	<	common.logger.debug(6, "closeBlock %s"%closeBlock)
268	<	#dbswriter.dbs.closeBlock(BlockName)
320	>	common.logger.log(10-1, "closeBlock %s"%closeBlock)
321		except DBSWriterError, ex:
322	<	common.logger.message("Close block error %s"%ex)
322	>	common.logger.info("Close block error %s"%ex)
323
324		if (len(self.noEventsFiles)>0):
325	<	common.logger.message("--->>> WARNING: "+str(len(self.noEventsFiles))+" files not published because they contain 0 events are:")
325	>	common.logger.info("--->>> WARNING: "+str(len(self.noEventsFiles))+" published files contain 0 events are:")
326		for lfn in self.noEventsFiles:
327	<	common.logger.message("------ LFN: %s"%lfn)
327	>	common.logger.info("------ LFN: %s"%lfn)
328		if (len(self.noLFN)>0):
329	<	common.logger.message("--->>> WARNING: there are "+str(len(self.noLFN))+" files not published because they have empty LFN")
329	>	common.logger.info("--->>> WARNING: there are "+str(len(self.noLFN))+" files not published because they have empty LFN")
330		for pfn in self.noLFN:
331	<	common.logger.message("------ pfn: %s"%pfn)
331	>	common.logger.info("------ pfn: %s"%pfn)
332		if (len(self.problemFiles)>0):
333	<	common.logger.message("--->>> WARNING: "+str(len(self.problemFiles))+" files not published because they had problem with copy to SE")
333	>	common.logger.info("--->>> WARNING: "+str(len(self.problemFiles))+" files not published because they had problem with copy to SE")
334		for lfn in self.problemFiles:
335	<	common.logger.message("------ LFN: %s"%lfn)
336	<	common.logger.message("--->>> End files publication")
337	<	common.logger.message("--->>> To check data publication please use: InspectDBS2.py --DBSURL=<dbs_url_for_publication> --datasetPath=<User Dataset Name>")
335	>	common.logger.info("------ LFN: %s"%lfn)
336	>	common.logger.info("--->>> End files publication")
337	>
338	>	#### FEDE for MULTI ####
339	>	for dataset_to_check in self.published_datasets:
340	>	self.cfg_params['USER.dataset_to_check']=dataset_to_check
341	>	from InspectDBS import InspectDBS
342	>	check=InspectDBS(self.cfg_params)
343	>	check.checkPublication()
344	>	#########################
345	>
346		return self.exit_status
347
348		else:
349	<	common.logger.message("--->>> "+self.resDir+" empty: no file to publish on DBS")
349	>	common.logger.info("--->>> No valid files to publish on DBS. Your jobs do not report exit codes = 0")
350		self.exit_status = '1'
351		return self.exit_status
352

Diff Legend

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