CRAB/python/DataDiscovery.py

#!/usr/bin/env python
import exceptions
import DBSAPI.dbsApi
from DBSAPI.dbsApiException import * 
import common
from crab_util import *
import os 


# #######################################
class DBSError(exceptions.Exception):
    def __init__(self, errorName, errorMessage):
        args='\nERROR DBS %s : %s \n'%(errorName,errorMessage)
        exceptions.Exception.__init__(self, args)
        pass
    
    def getErrorMessage(self):
        """ Return error message """
        return "%s" % (self.args)

# #######################################
class DBSInvalidDataTierError(exceptions.Exception):
    def __init__(self, errorName, errorMessage):
        args='\nERROR DBS %s : %s \n'%(errorName,errorMessage)
        exceptions.Exception.__init__(self, args)
        pass
    
    def getErrorMessage(self):
        """ Return error message """
        return "%s" % (self.args)

# #######################################
class DBSInfoError:
    def __init__(self, url):
        print '\nERROR accessing DBS url : '+url+'\n'
        pass

# ####################################
class DataDiscoveryError(exceptions.Exception):
    def __init__(self, errorMessage):
        self.args=errorMessage
        exceptions.Exception.__init__(self, self.args)
        pass

    def getErrorMessage(self):
        """ Return exception error """
        return "%s" % (self.args)

# ####################################
class NotExistingDatasetError(exceptions.Exception):
    def __init__(self, errorMessage):
        self.args=errorMessage
        exceptions.Exception.__init__(self, self.args)
        pass

    def getErrorMessage(self):
        """ Return exception error """
        return "%s" % (self.args)

# ####################################
class NoDataTierinProvenanceError(exceptions.Exception):
    def __init__(self, errorMessage):
        self.args=errorMessage
        exceptions.Exception.__init__(self, self.args)
        pass

    def getErrorMessage(self):
        """ Return exception error """
        return "%s" % (self.args)

# ####################################
# class to find and extact info from published data
class DataDiscovery:
    def __init__(self, datasetPath, cfg_params, skipAnBlocks):

        #       Attributes
        self.datasetPath = datasetPath
        self.cfg_params = cfg_params
        self.skipBlocks = skipAnBlocks

        self.eventsPerBlock = {}  # DBS output: map fileblocks-events for collection
        self.eventsPerFile = {}   # DBS output: map files-events
        self.blocksinfo = {}      # DBS output: map fileblocks-files 
        self.maxEvents = 0        # DBS output: max events
        self.parent = {}       # DBS output: max events

# ####################################
    def fetchDBSInfo(self):
        """
        Contact DBS
        """

        ## get DBS URL
        dbs_url="http://cmsdbsprod.cern.ch/cms_dbs_prod_global/servlet/DBSServlet"
        if (self.cfg_params.has_key('CMSSW.dbs_url')):
            dbs_url=self.cfg_params['CMSSW.dbs_url']

        common.logger.debug(3,"Accessing DBS at: "+dbs_url)

        ## check if runs are selected
        runselection = []
        if (self.cfg_params.has_key('CMSSW.runselection')):
            runselection = parseRange2(self.cfg_params['CMSSW.runselection'])

        common.logger.debug(6,"runselection is: %s"%runselection)
        ## service API
        args = {}
        args['url']     = dbs_url
        args['level']   = 'CRITICAL'

        ## check if has been requested to use the parent info
        useParent = self.cfg_params.get('CMSSW.use_parent',False)

        ## check if has been asked for a non default file to store/read analyzed fileBlocks   
        defaultName = common.work_space.shareDir()+'AnalyzedBlocks.txt'  
        fileBlocks_FileName = os.path.abspath(self.cfg_params.get('CMSSW.fileblocks_file',defaultName))
 
        api = DBSAPI.dbsApi.DbsApi(args)
        try:
            if len(runselection) <= 0 :
                if useParent:
                    files = api.listFiles(path=self.datasetPath,retriveList= [
                                                                              'retrive_parent', 
                                                                              'retrive_block',
                                                                              'retrive_lumi',
                                                                              'retrive_run'
                                                                              ])
                    common.logger.debug(5,"Set of input parameters used for DBS query : \n"+str(allowedRetriveValue)) 
                    common.logger.write("Set of input parameters used for DBS query : \n"+str(allowedRetriveValue)) 
                else:
                    files = api.listDatasetFiles(self.datasetPath)
            else :
                files=[]
                for arun in runselection:
                    try:
                        filesinrun = api.listFiles(path=self.datasetPath,retriveList=allowedRetriveValue,runNumber=arun)
                        files.extend(filesinrun)
                    except:
                        msg="WARNING: problem extracting info from DBS for run %s "%arun
                        common.logger.message(msg)
                        pass

        except DbsBadRequest, msg:
            raise DataDiscoveryError(msg)
        except DBSError, msg:
            raise DataDiscoveryError(msg)

        anFileBlocks = []
        if self.skipBlocks: anFileBlocks = readTXTfile(self, fileBlocks_FileName) 

        # parse files and fill arrays
        for file in files :
            parList = []
            # skip already analyzed blocks
            fileblock = file['Block']['Name']
            if fileblock not in anFileBlocks :
                filename = file['LogicalFileName']
                # asked retry the list of parent for the given child 
                if useParent: parList = [x['LogicalFileName'] for x in file['ParentList']] 
                self.parent[filename] = parList 
                if filename.find('.dat') < 0 :
                    events    = file['NumberOfEvents']
                    # number of events per block
                    if fileblock in self.eventsPerBlock.keys() :
                        self.eventsPerBlock[fileblock] += events
                    else :
                        self.eventsPerBlock[fileblock] = events
                    # number of events per file
                    self.eventsPerFile[filename] = events
            
                    # number of events per block
                    if fileblock in self.blocksinfo.keys() :
                        self.blocksinfo[fileblock].append(filename)
                    else :
                        self.blocksinfo[fileblock] = [filename]
            
                    # total number of events
                    self.maxEvents += events
        if  self.skipBlocks and len(self.eventsPerBlock.keys()) == 0:
            msg = "No new fileblocks available for dataset: "+str(self.datasetPath)
            raise  CrabException(msg)    

        saveFblocks='' 
        for block in self.eventsPerBlock.keys() :
            saveFblocks += str(block)+'\n' 
            common.logger.debug(6,"DBSInfo: total nevts %i in block %s "%(self.eventsPerBlock[block],block))
        writeTXTfile(self, fileBlocks_FileName , saveFblocks) 
                      
        if len(self.eventsPerBlock) <= 0:
            raise NotExistingDatasetError(("\nNo data for %s in DBS\nPlease check"
                                            + " dataset path variables in crab.cfg")
                                            % self.datasetPath)


# #################################################
    def getMaxEvents(self):
        """
        max events 
        """
        return self.maxEvents

# #################################################
    def getEventsPerBlock(self):
        """
        list the event collections structure by fileblock 
        """
        return self.eventsPerBlock

# #################################################
    def getEventsPerFile(self):
        """
        list the event collections structure by file 
        """
        return self.eventsPerFile

# #################################################
    def getFiles(self):
        """
        return files grouped by fileblock 
        """
        return self.blocksinfo        

# #################################################
    def getParent(self):
        """
        return parent grouped by file 
        """
        return self.parent        

########################################################################
Revision:	1.22
Committed:	Mon Jun 9 15:53:05 2008 UTC (16 years, 10 months ago) by spiga
Content type:	text/x-python
Branch:	MAIN
Changes since 1.21:	+54 -39 lines
Log Message:	improved support for two tiers input. Added support to skip already analyzed Bloks
#	User	Rev	Content
1	gutsche	1.6	#!/usr/bin/env python
2	slacapra	1.18	import exceptions
3			import DBSAPI.dbsApi
4			from DBSAPI.dbsApiException import *
5			import common
6			from crab_util import *
7	spiga	1.22	import os
8	afanfani	1.1
9	afanfani	1.3
10	slacapra	1.18	# #######################################
11			class DBSError(exceptions.Exception):
12			def __init__(self, errorName, errorMessage):
13			args='\nERROR DBS %s : %s \n'%(errorName,errorMessage)
14			exceptions.Exception.__init__(self, args)
15			pass
16
17			def getErrorMessage(self):
18			""" Return error message """
19			return "%s" % (self.args)
20
21			# #######################################
22			class DBSInvalidDataTierError(exceptions.Exception):
23			def __init__(self, errorName, errorMessage):
24			args='\nERROR DBS %s : %s \n'%(errorName,errorMessage)
25			exceptions.Exception.__init__(self, args)
26			pass
27
28			def getErrorMessage(self):
29			""" Return error message """
30			return "%s" % (self.args)
31
32			# #######################################
33			class DBSInfoError:
34			def __init__(self, url):
35			print '\nERROR accessing DBS url : '+url+'\n'
36			pass
37
38	afanfani	1.1	# ####################################
39	afanfani	1.3	class DataDiscoveryError(exceptions.Exception):
40	slacapra	1.7	def __init__(self, errorMessage):
41	gutsche	1.15	self.args=errorMessage
42	slacapra	1.14	exceptions.Exception.__init__(self, self.args)
43	slacapra	1.7	pass
44
45			def getErrorMessage(self):
46			""" Return exception error """
47			return "%s" % (self.args)
48	afanfani	1.3
49	afanfani	1.1	# ####################################
50	afanfani	1.3	class NotExistingDatasetError(exceptions.Exception):
51	slacapra	1.7	def __init__(self, errorMessage):
52	gutsche	1.15	self.args=errorMessage
53	slacapra	1.14	exceptions.Exception.__init__(self, self.args)
54	slacapra	1.7	pass
55
56			def getErrorMessage(self):
57			""" Return exception error """
58			return "%s" % (self.args)
59	afanfani	1.1
60			# ####################################
61	afanfani	1.3	class NoDataTierinProvenanceError(exceptions.Exception):
62	slacapra	1.7	def __init__(self, errorMessage):
63	gutsche	1.15	self.args=errorMessage
64	slacapra	1.14	exceptions.Exception.__init__(self, self.args)
65	slacapra	1.7	pass
66
67			def getErrorMessage(self):
68			""" Return exception error """
69			return "%s" % (self.args)
70	afanfani	1.1
71			# ####################################
72			# class to find and extact info from published data
73			class DataDiscovery:
74	spiga	1.22	def __init__(self, datasetPath, cfg_params, skipAnBlocks):
75	afanfani	1.1
76	slacapra	1.18	# Attributes
77	slacapra	1.11	self.datasetPath = datasetPath
78	afanfani	1.1	self.cfg_params = cfg_params
79	spiga	1.22	self.skipBlocks = skipAnBlocks
80	afanfani	1.1
81	slacapra	1.11	self.eventsPerBlock = {} # DBS output: map fileblocks-events for collection
82			self.eventsPerFile = {} # DBS output: map files-events
83	slacapra	1.18	self.blocksinfo = {} # DBS output: map fileblocks-files
84			self.maxEvents = 0 # DBS output: max events
85	spiga	1.21	self.parent = {} # DBS output: max events
86	afanfani	1.1
87			# ####################################
88			def fetchDBSInfo(self):
89			"""
90			Contact DBS
91			"""
92
93	slacapra	1.11	## get DBS URL
94	slacapra	1.19	dbs_url="http://cmsdbsprod.cern.ch/cms_dbs_prod_global/servlet/DBSServlet"
95			if (self.cfg_params.has_key('CMSSW.dbs_url')):
96	slacapra	1.11	dbs_url=self.cfg_params['CMSSW.dbs_url']
97	afanfani	1.3
98	slacapra	1.18	common.logger.debug(3,"Accessing DBS at: "+dbs_url)
99
100			## check if runs are selected
101	slacapra	1.19	runselection = []
102			if (self.cfg_params.has_key('CMSSW.runselection')):
103	slacapra	1.18	runselection = parseRange2(self.cfg_params['CMSSW.runselection'])
104
105	afanfani	1.20	common.logger.debug(6,"runselection is: %s"%runselection)
106	slacapra	1.18	## service API
107			args = {}
108			args['url'] = dbs_url
109			args['level'] = 'CRITICAL'
110
111	spiga	1.21	## check if has been requested to use the parent info
112	spiga	1.22	useParent = self.cfg_params.get('CMSSW.use_parent',False)
113	spiga	1.21
114	spiga	1.22	## check if has been asked for a non default file to store/read analyzed fileBlocks
115			defaultName = common.work_space.shareDir()+'AnalyzedBlocks.txt'
116			fileBlocks_FileName = os.path.abspath(self.cfg_params.get('CMSSW.fileblocks_file',defaultName))
117
118	slacapra	1.18	api = DBSAPI.dbsApi.DbsApi(args)
119	afanfani	1.5	try:
120	slacapra	1.18	if len(runselection) <= 0 :
121	spiga	1.22	if useParent:
122			files = api.listFiles(path=self.datasetPath,retriveList= [
123			'retrive_parent',
124			'retrive_block',
125			'retrive_lumi',
126			'retrive_run'
127			])
128			common.logger.debug(5,"Set of input parameters used for DBS query : \n"+str(allowedRetriveValue))
129			common.logger.write("Set of input parameters used for DBS query : \n"+str(allowedRetriveValue))
130			else:
131			files = api.listDatasetFiles(self.datasetPath)
132	slacapra	1.18	else :
133	afanfani	1.20	files=[]
134			for arun in runselection:
135			try:
136	spiga	1.21	filesinrun = api.listFiles(path=self.datasetPath,retriveList=allowedRetriveValue,runNumber=arun)
137	afanfani	1.20	files.extend(filesinrun)
138			except:
139			msg="WARNING: problem extracting info from DBS for run %s "%arun
140			common.logger.message(msg)
141			pass
142
143	slacapra	1.18	except DbsBadRequest, msg:
144	slacapra	1.11	raise DataDiscoveryError(msg)
145	slacapra	1.13	except DBSError, msg:
146			raise DataDiscoveryError(msg)
147	slacapra	1.11
148	spiga	1.22	anFileBlocks = []
149			if self.skipBlocks: anFileBlocks = readTXTfile(self, fileBlocks_FileName)
150
151	slacapra	1.18	# parse files and fill arrays
152			for file in files :
153	spiga	1.21	parList = []
154	spiga	1.22	# skip already analyzed blocks
155			fileblock = file['Block']['Name']
156			if fileblock not in anFileBlocks :
157			filename = file['LogicalFileName']
158			# asked retry the list of parent for the given child
159			if useParent: parList = [x['LogicalFileName'] for x in file['ParentList']]
160			self.parent[filename] = parList
161			if filename.find('.dat') < 0 :
162			events = file['NumberOfEvents']
163			# number of events per block
164			if fileblock in self.eventsPerBlock.keys() :
165			self.eventsPerBlock[fileblock] += events
166			else :
167			self.eventsPerBlock[fileblock] = events
168			# number of events per file
169			self.eventsPerFile[filename] = events
170
171			# number of events per block
172			if fileblock in self.blocksinfo.keys() :
173			self.blocksinfo[fileblock].append(filename)
174			else :
175			self.blocksinfo[fileblock] = [filename]
176
177			# total number of events
178			self.maxEvents += events
179			if self.skipBlocks and len(self.eventsPerBlock.keys()) == 0:
180			msg = "No new fileblocks available for dataset: "+str(self.datasetPath)
181			raise CrabException(msg)
182	slacapra	1.11
183	spiga	1.22	saveFblocks=''
184	slacapra	1.18	for block in self.eventsPerBlock.keys() :
185	spiga	1.22	saveFblocks += str(block)+'\n'
186	slacapra	1.18	common.logger.debug(6,"DBSInfo: total nevts %i in block %s "%(self.eventsPerBlock[block],block))
187	spiga	1.22	writeTXTfile(self, fileBlocks_FileName , saveFblocks)
188
189	slacapra	1.11	if len(self.eventsPerBlock) <= 0:
190	slacapra	1.18	raise NotExistingDatasetError(("\nNo data for %s in DBS\nPlease check"
191	slacapra	1.11	+ " dataset path variables in crab.cfg")
192	slacapra	1.18	% self.datasetPath)
193	afanfani	1.1
194
195			# #################################################
196			def getMaxEvents(self):
197			"""
198	slacapra	1.11	max events
199	afanfani	1.1	"""
200	slacapra	1.18	return self.maxEvents
201	afanfani	1.1
202			# #################################################
203	slacapra	1.11	def getEventsPerBlock(self):
204	afanfani	1.1	"""
205	slacapra	1.11	list the event collections structure by fileblock
206	afanfani	1.1	"""
207	slacapra	1.11	return self.eventsPerBlock
208	afanfani	1.1
209			# #################################################
210	slacapra	1.11	def getEventsPerFile(self):
211	afanfani	1.1	"""
212	slacapra	1.11	list the event collections structure by file
213	afanfani	1.1	"""
214	slacapra	1.11	return self.eventsPerFile
215	afanfani	1.1
216			# #################################################
217	slacapra	1.11	def getFiles(self):
218	afanfani	1.1	"""
219	slacapra	1.11	return files grouped by fileblock
220	afanfani	1.1	"""
221	slacapra	1.11	return self.blocksinfo
222	afanfani	1.1
223	spiga	1.21	# #################################################
224			def getParent(self):
225			"""
226			return parent grouped by file
227			"""
228			return self.parent
229
230	afanfani	1.1	########################################################################