1 |
import os
|
2 |
import sys
|
3 |
import getopt
|
4 |
import popen2
|
5 |
import time
|
6 |
|
7 |
|
8 |
# Fixme: do the same for createProdWorkflow
|
9 |
def createPreProdWorkflow (PyCfg = None,
|
10 |
PSetHash = None,
|
11 |
prodName = None,
|
12 |
cmsswVersion = None,
|
13 |
category = "PreProd",
|
14 |
timestamp = int(time.time()),
|
15 |
FakeHash = False,
|
16 |
pileupDS = None,
|
17 |
pileupFilesPerJob = 1,
|
18 |
dbsAddress = None,
|
19 |
dbsUrl = None,
|
20 |
dlsAddress = None,
|
21 |
dlsType = None,
|
22 |
pileupSkipLocation = False):
|
23 |
spec = WorkflowSpec()
|
24 |
spec.setWorkflowName(prodName)
|
25 |
spec.setRequestCategory(category)
|
26 |
spec.setRequestTimestamp(timestamp)
|
27 |
|
28 |
# //
|
29 |
# // This value was created by running the EdmConfigHash tool
|
30 |
# // on the original cfg file.
|
31 |
|
32 |
|
33 |
cmsRun = spec.payload
|
34 |
cmsRun.name = "cmsRun1" # every node in the workflow needs a unique name
|
35 |
cmsRun.type = "CMSSW" # Nodes that are CMSSW based should set the name
|
36 |
cmsRun.application["Project"] = "CMSSW" # project
|
37 |
cmsRun.application["Version"] = cmsswVersion # version
|
38 |
cmsRun.application["Architecture"] = "slc3_ia32_gcc323" # arch (not needed)
|
39 |
cmsRun.application["Executable"] = "cmsRun" # binary name
|
40 |
cmsRun.configuration = PyCfg # Python PSet file
|
41 |
|
42 |
|
43 |
# //
|
44 |
# // Pileup sample?
|
45 |
# //
|
46 |
if pileupDS != None:
|
47 |
puPrimary = pileupDS.split("/")[1]
|
48 |
puTier = pileupDS.split("/")[2]
|
49 |
puProc = pileupDS.split("/")[3]
|
50 |
puDataset = cmsRun.addPileupDataset(puPrimary, puTier, puProc)
|
51 |
puDataset['FilesPerJob'] = pileupFilesPerJob
|
52 |
if customDBSDLS:
|
53 |
puDataset['DBSAddress'] = dbsAddress
|
54 |
puDataset['DBSURL'] = dbsUrl
|
55 |
puDataset['DLSType'] = dlsType
|
56 |
puDataset['DLSAddress'] = dlsAddress
|
57 |
if pileupSkipLocation:
|
58 |
puDataset['SkipLocation'] = pileupSkipLocation
|
59 |
|
60 |
# //
|
61 |
# // Pull all the output modules from the configuration file,
|
62 |
#// treat the output module name as DataTier and AppFamily,
|
63 |
# //since there isnt a standard way to derive these things yet.
|
64 |
# //
|
65 |
#// For each module a dataset declaration is created in the spec
|
66 |
cfgInt = CfgInterface(cmsRun.configuration, True)
|
67 |
datasetList = []
|
68 |
for outModName, val in cfgInt.outputModules.items():
|
69 |
# //
|
70 |
# // Check for Multi Tiers.
|
71 |
#// If Output module contains - characters, we split based on it
|
72 |
# //And create a different tier for each basic tier
|
73 |
# //
|
74 |
#//
|
75 |
|
76 |
tierList = splitMultiTier(outModName)
|
77 |
for dataTier in tierList:
|
78 |
processedDS = "%s-%s-%s-unmerged" % (
|
79 |
cmsRun.application['Version'], outModName, timestamp)
|
80 |
outDS = cmsRun.addOutputDataset(prodName,
|
81 |
processedDS,
|
82 |
outModName)
|
83 |
|
84 |
outDS['DataTier'] = dataTier
|
85 |
outDS["ApplicationName"] = cmsRun.application["Executable"]
|
86 |
outDS["ApplicationProject"] = cmsRun.application["Project"]
|
87 |
outDS["ApplicationVersion"] = cmsRun.application["Version"]
|
88 |
outDS["ApplicationFamily"] = outModName
|
89 |
if FakeHash:
|
90 |
guid = MCPayloadsUUID.uuidgen()
|
91 |
if guid == None:
|
92 |
guid = MCPayloadsUUID.uuid()
|
93 |
hashValue = "hash=%s;guid=%s" % (PSetHash, guid)
|
94 |
outDS['PSetHash'] = hashValue
|
95 |
else:
|
96 |
outDS['PSetHash'] = PSetHash
|
97 |
datasetList.append(outDS.name())
|
98 |
|
99 |
stageOut = cmsRun.newNode("stageOut1")
|
100 |
stageOut.type = "StageOut"
|
101 |
stageOut.application["Project"] = ""
|
102 |
stageOut.application["Version"] = ""
|
103 |
stageOut.application["Architecture"] = ""
|
104 |
stageOut.application["Executable"] = "RuntimeStageOut.py" # binary name
|
105 |
stageOut.configuration = ""
|
106 |
|
107 |
mergedLFNBase(spec)
|
108 |
unmergedLFNBase(spec)
|
109 |
|
110 |
return spec
|