1 |
|
#!/usr/bin/env python |
2 |
< |
from DBSInfo import * |
2 |
> |
|
3 |
> |
__revision__ = "$Id$" |
4 |
> |
__version__ = "$Revision$" |
5 |
> |
|
6 |
> |
import exceptions |
7 |
> |
import DBSAPI.dbsApi |
8 |
> |
from DBSAPI.dbsApiException import * |
9 |
> |
import common |
10 |
> |
from crab_util import * |
11 |
> |
import os |
12 |
> |
|
13 |
> |
|
14 |
> |
|
15 |
> |
class DBSError(exceptions.Exception): |
16 |
> |
def __init__(self, errorName, errorMessage): |
17 |
> |
args='\nERROR DBS %s : %s \n'%(errorName,errorMessage) |
18 |
> |
exceptions.Exception.__init__(self, args) |
19 |
> |
pass |
20 |
> |
|
21 |
> |
def getErrorMessage(self): |
22 |
> |
""" Return error message """ |
23 |
> |
return "%s" % (self.args) |
24 |
> |
|
25 |
> |
|
26 |
> |
|
27 |
> |
class DBSInvalidDataTierError(exceptions.Exception): |
28 |
> |
def __init__(self, errorName, errorMessage): |
29 |
> |
args='\nERROR DBS %s : %s \n'%(errorName,errorMessage) |
30 |
> |
exceptions.Exception.__init__(self, args) |
31 |
> |
pass |
32 |
> |
|
33 |
> |
def getErrorMessage(self): |
34 |
> |
""" Return error message """ |
35 |
> |
return "%s" % (self.args) |
36 |
> |
|
37 |
> |
|
38 |
> |
|
39 |
> |
class DBSInfoError: |
40 |
> |
def __init__(self, url): |
41 |
> |
print '\nERROR accessing DBS url : '+url+'\n' |
42 |
> |
pass |
43 |
> |
|
44 |
|
|
45 |
|
|
5 |
– |
# #################################### |
46 |
|
class DataDiscoveryError(exceptions.Exception): |
47 |
|
def __init__(self, errorMessage): |
8 |
– |
exceptions.Exception.__init__(self, self.args) |
48 |
|
self.args=errorMessage |
49 |
+ |
exceptions.Exception.__init__(self, self.args) |
50 |
|
pass |
51 |
|
|
52 |
|
def getErrorMessage(self): |
53 |
|
""" Return exception error """ |
54 |
|
return "%s" % (self.args) |
55 |
|
|
56 |
< |
# #################################### |
56 |
> |
|
57 |
> |
|
58 |
|
class NotExistingDatasetError(exceptions.Exception): |
59 |
|
def __init__(self, errorMessage): |
19 |
– |
exceptions.Exception.__init__(self, self.args) |
60 |
|
self.args=errorMessage |
61 |
+ |
exceptions.Exception.__init__(self, self.args) |
62 |
|
pass |
63 |
|
|
64 |
|
def getErrorMessage(self): |
65 |
|
""" Return exception error """ |
66 |
|
return "%s" % (self.args) |
67 |
|
|
68 |
< |
# #################################### |
68 |
> |
|
69 |
> |
|
70 |
|
class NoDataTierinProvenanceError(exceptions.Exception): |
71 |
|
def __init__(self, errorMessage): |
30 |
– |
exceptions.Exception.__init__(self, self.args) |
72 |
|
self.args=errorMessage |
73 |
+ |
exceptions.Exception.__init__(self, self.args) |
74 |
|
pass |
75 |
|
|
76 |
|
def getErrorMessage(self): |
77 |
|
""" Return exception error """ |
78 |
|
return "%s" % (self.args) |
79 |
|
|
80 |
< |
# #################################### |
81 |
< |
# class to find and extact info from published data |
80 |
> |
|
81 |
> |
|
82 |
|
class DataDiscovery: |
83 |
< |
def __init__(self, datasetPath, dataTiers, cfg_params): |
83 |
> |
""" |
84 |
> |
Class to find and extact info from published data |
85 |
> |
""" |
86 |
> |
def __init__(self, datasetPath, cfg_params, skipAnBlocks): |
87 |
|
|
88 |
< |
# Attributes |
88 |
> |
# Attributes |
89 |
|
self.datasetPath = datasetPath |
90 |
< |
self.dataTiers = dataTiers |
90 |
> |
# Analysis dataset is primary/processed/tier/definition |
91 |
> |
self.ads = len(self.datasetPath.split("/")) > 4 |
92 |
|
self.cfg_params = cfg_params |
93 |
+ |
self.skipBlocks = skipAnBlocks |
94 |
|
|
95 |
|
self.eventsPerBlock = {} # DBS output: map fileblocks-events for collection |
96 |
|
self.eventsPerFile = {} # DBS output: map files-events |
97 |
< |
self.blocksinfo = {} # DBS output: map fileblocks-files |
98 |
< |
#DBS output: max events computed by method getMaxEvents |
97 |
> |
# self.lumisPerBlock = {} # DBS output: number of lumis in each block |
98 |
> |
# self.lumisPerFile = {} # DBS output: number of lumis in each file |
99 |
> |
self.blocksinfo = {} # DBS output: map fileblocks-files |
100 |
> |
self.maxEvents = 0 # DBS output: max events |
101 |
> |
self.maxLumis = 0 # DBS output: total number of lumis |
102 |
> |
self.parent = {} # DBS output: parents of each file |
103 |
> |
self.lumis = {} # DBS output: lumis in each file |
104 |
> |
|
105 |
|
|
53 |
– |
# #################################### |
106 |
|
def fetchDBSInfo(self): |
107 |
|
""" |
108 |
|
Contact DBS |
109 |
|
""" |
58 |
– |
|
110 |
|
## get DBS URL |
111 |
< |
try: |
112 |
< |
dbs_url=self.cfg_params['CMSSW.dbs_url'] |
113 |
< |
except KeyError: |
114 |
< |
dbs_url="http://cmsdoc.cern.ch/cms/test/aprom/DBS/CGIServer/prodquery" |
111 |
> |
global_url="http://cmsdbsprod.cern.ch/cms_dbs_prod_global/servlet/DBSServlet" |
112 |
> |
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) |
128 |
> |
common.logger.info("Accessing DBS at: "+dbs_url) |
129 |
> |
|
130 |
> |
## check if runs are selected |
131 |
> |
runselection = [] |
132 |
> |
if (self.cfg_params.has_key('CMSSW.runselection')): |
133 |
> |
runselection = parseRange2(self.cfg_params['CMSSW.runselection']) |
134 |
> |
|
135 |
> |
|
136 |
> |
self.splitByRun = int(self.cfg_params.get('CMSSW.split_by_run', 0)) |
137 |
> |
|
138 |
> |
common.logger.log(10-1,"runselection is: %s"%runselection) |
139 |
> |
## service API |
140 |
> |
args = {} |
141 |
> |
args['url'] = dbs_url |
142 |
> |
args['level'] = 'CRITICAL' |
143 |
> |
|
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' |
149 |
> |
fileBlocks_FileName = os.path.abspath(self.cfg_params.get('CMSSW.fileblocks_file',defaultName)) |
150 |
> |
|
151 |
> |
api = DBSAPI.dbsApi.DbsApi(args) |
152 |
> |
self.files = self.queryDbs(api,path=self.datasetPath,runselection=runselection,useParent=useparent) |
153 |
> |
|
154 |
> |
anFileBlocks = [] |
155 |
> |
if self.skipBlocks: anFileBlocks = readTXTfile(self, fileBlocks_FileName) |
156 |
> |
|
157 |
> |
# parse files and fill arrays |
158 |
> |
for file in self.files : |
159 |
> |
parList = [] |
160 |
> |
lumiList = [] # List of tuples |
161 |
> |
# skip already analyzed blocks |
162 |
> |
fileblock = file['Block']['Name'] |
163 |
> |
if fileblock not in anFileBlocks : |
164 |
> |
filename = file['LogicalFileName'] |
165 |
> |
# asked retry the list of parent for the given child |
166 |
> |
if useparent==1: |
167 |
> |
parList = [x['LogicalFileName'] for x in file['ParentList']] |
168 |
> |
if self.ads: |
169 |
> |
lumiList = [ (x['RunNumber'], x['LumiSectionNumber']) |
170 |
> |
for x in file['LumiList'] ] |
171 |
> |
self.parent[filename] = parList |
172 |
> |
self.lumis[filename] = lumiList |
173 |
> |
if filename.find('.dat') < 0 : |
174 |
> |
events = file['NumberOfEvents'] |
175 |
> |
# Count number of events and lumis per block |
176 |
> |
if fileblock in self.eventsPerBlock.keys() : |
177 |
> |
self.eventsPerBlock[fileblock] += events |
178 |
> |
else : |
179 |
> |
self.eventsPerBlock[fileblock] = events |
180 |
> |
# Number of events per file |
181 |
> |
self.eventsPerFile[filename] = events |
182 |
> |
|
183 |
> |
# List of files per block |
184 |
> |
if fileblock in self.blocksinfo.keys() : |
185 |
> |
self.blocksinfo[fileblock].append(filename) |
186 |
> |
else : |
187 |
> |
self.blocksinfo[fileblock] = [filename] |
188 |
> |
|
189 |
> |
# total number of events |
190 |
> |
self.maxEvents += events |
191 |
> |
self.maxLumis += len(lumiList) |
192 |
> |
|
193 |
> |
if self.skipBlocks and len(self.eventsPerBlock.keys()) == 0: |
194 |
> |
msg = "No new fileblocks available for dataset: "+str(self.datasetPath) |
195 |
> |
raise CrabException(msg) |
196 |
> |
|
197 |
> |
saveFblocks='' |
198 |
> |
for block in self.eventsPerBlock.keys() : |
199 |
> |
saveFblocks += str(block)+'\n' |
200 |
> |
common.logger.log(10-1,"DBSInfo: total nevts %i in block %s "%(self.eventsPerBlock[block],block)) |
201 |
> |
writeTXTfile(self, fileBlocks_FileName , saveFblocks) |
202 |
|
|
203 |
< |
## get info about the requested dataset |
204 |
< |
try: |
205 |
< |
dbs_instance=self.cfg_params['CMSSW.dbs_instance'] |
206 |
< |
except KeyError: |
207 |
< |
dbs_instance="MCGlobal/Writer" |
208 |
< |
|
209 |
< |
dbs = DBSInfo(dbs_url, dbs_instance) |
203 |
> |
if len(self.eventsPerBlock) <= 0: |
204 |
> |
raise NotExistingDatasetError(("\nNo data for %s in DBS\nPlease check" |
205 |
> |
+ " dataset path variables in crab.cfg") |
206 |
> |
% self.datasetPath) |
207 |
> |
|
208 |
> |
|
209 |
> |
def queryDbs(self,api,path=None,runselection=None,useParent=None): |
210 |
> |
|
211 |
> |
allowedRetriveValue = ['retrive_block', 'retrive_run'] |
212 |
> |
if self.ads: allowedRetriveValue.append('retrive_lumi') |
213 |
> |
if useParent == 1: allowedRetriveValue.append('retrive_parent') |
214 |
> |
common.logger.debug("Set of input parameters used for DBS query: %s" % allowedRetriveValue) |
215 |
|
try: |
216 |
< |
self.datasets = dbs.getMatchingDatasets(self.datasetPath) |
217 |
< |
except dbsCgiApi.DbsCgiExecutionError, msg: |
216 |
> |
if len(runselection) <=0 : |
217 |
> |
if useParent==1 or self.splitByRun==1 : |
218 |
> |
if self.ads: |
219 |
> |
files = api.listFiles(analysisDataset=path, retriveList=allowedRetriveValue) |
220 |
> |
else : |
221 |
> |
files = api.listFiles(path=path, retriveList=allowedRetriveValue) |
222 |
> |
else: |
223 |
> |
files = api.listDatasetFiles(self.datasetPath) |
224 |
> |
else : |
225 |
> |
files=[] |
226 |
> |
for arun in runselection: |
227 |
> |
try: |
228 |
> |
if self.ads: |
229 |
> |
filesinrun = api.listFiles(analysisDataset=path,retriveList=allowedRetriveValue,runNumber=arun) |
230 |
> |
else: |
231 |
> |
filesinrun = api.listFiles(path=path,retriveList=allowedRetriveValue,runNumber=arun) |
232 |
> |
files.extend(filesinrun) |
233 |
> |
except: |
234 |
> |
msg="WARNING: problem extracting info from DBS for run %s "%arun |
235 |
> |
common.logger.info(msg) |
236 |
> |
pass |
237 |
> |
|
238 |
> |
except DbsBadRequest, msg: |
239 |
|
raise DataDiscoveryError(msg) |
240 |
|
except DBSError, msg: |
241 |
|
raise DataDiscoveryError(msg) |
242 |
|
|
243 |
< |
if len(self.datasets) == 0: |
80 |
< |
raise DataDiscoveryError("DatasetPath=%s unknown to DBS" %self.datasetPath) |
81 |
< |
if len(self.datasets) > 1: |
82 |
< |
raise DataDiscoveryError("DatasetPath=%s is ambiguous" %self.datasetPath) |
243 |
> |
return files |
244 |
|
|
84 |
– |
try: |
85 |
– |
self.dbsdataset = self.datasets[0].get('datasetPathName') |
245 |
|
|
246 |
< |
self.eventsPerBlock = dbs.getEventsPerBlock(self.dbsdataset) |
247 |
< |
self.blocksinfo = dbs.getDatasetFileBlocks(self.dbsdataset) |
248 |
< |
self.eventsPerFile = dbs.getEventsPerFile(self.dbsdataset) |
249 |
< |
except DBSError, ex: |
250 |
< |
raise DataDiscoveryError(ex.getErrorMessage()) |
92 |
< |
|
93 |
< |
if len(self.eventsPerBlock) <= 0: |
94 |
< |
raise NotExistingDatasetError (("\nNo data for %s in DBS\nPlease check" |
95 |
< |
+ " dataset path variables in crab.cfg") |
96 |
< |
% self.dbsdataset) |
246 |
> |
def getMaxEvents(self): |
247 |
> |
""" |
248 |
> |
max events |
249 |
> |
""" |
250 |
> |
return self.maxEvents |
251 |
|
|
252 |
|
|
253 |
< |
# ################################################# |
100 |
< |
def getMaxEvents(self): |
253 |
> |
def getMaxLumis(self): |
254 |
|
""" |
255 |
< |
max events |
255 |
> |
Return the number of lumis in the dataset |
256 |
|
""" |
257 |
< |
## loop over the event collections |
105 |
< |
nevts=0 |
106 |
< |
for evc_evts in self.eventsPerBlock.values(): |
107 |
< |
nevts=nevts+evc_evts |
257 |
> |
return self.maxLumis |
258 |
|
|
109 |
– |
return nevts |
259 |
|
|
111 |
– |
# ################################################# |
260 |
|
def getEventsPerBlock(self): |
261 |
|
""" |
262 |
< |
list the event collections structure by fileblock |
262 |
> |
list the event collections structure by fileblock |
263 |
|
""" |
264 |
|
return self.eventsPerBlock |
265 |
|
|
266 |
< |
# ################################################# |
266 |
> |
|
267 |
|
def getEventsPerFile(self): |
268 |
|
""" |
269 |
< |
list the event collections structure by file |
269 |
> |
list the event collections structure by file |
270 |
|
""" |
271 |
|
return self.eventsPerFile |
272 |
|
|
273 |
< |
# ################################################# |
273 |
> |
|
274 |
|
def getFiles(self): |
275 |
|
""" |
276 |
< |
return files grouped by fileblock |
276 |
> |
return files grouped by fileblock |
277 |
> |
""" |
278 |
> |
return self.blocksinfo |
279 |
> |
|
280 |
> |
|
281 |
> |
def getParent(self): |
282 |
> |
""" |
283 |
> |
return parent grouped by file |
284 |
|
""" |
285 |
< |
return self.blocksinfo |
285 |
> |
return self.parent |
286 |
|
|
287 |
< |
######################################################################## |
287 |
> |
|
288 |
> |
def getLumis(self): |
289 |
> |
""" |
290 |
> |
return lumi sections grouped by file |
291 |
> |
""" |
292 |
> |
return self.lumis |
293 |
> |
|
294 |
> |
|
295 |
> |
def getListFiles(self): |
296 |
> |
""" |
297 |
> |
return parent grouped by file |
298 |
> |
""" |
299 |
> |
return self.files |