DAOD_PHYSLITE PRW Update/Fix, 21.2 branch (2020.03.19.)
Removed the generation of pileup event weights for DAOD_PHYSLITE
.
This will require analysers to use the PRW tool to generate their "correct" pileup weights when running over the DAOD_PHYSLITE
files. But it will simplify the DAOD_PHYSLITE
production a whole lot, as none of the tools running in the DAOD_PHYSLITE
production actually need the event weights. They just need the random run numbers to be generated.
With this change in, we go from the PRW tool printing this during initialisation:
...
11:12:48 Info in <CP::TPileupReweighting::UsePeriodConfig>: Using MC16 Period configuration
11:12:48 Info in <CP::TPileupReweighting::UsePeriodConfig>: Using MC16 Period configuration
11:12:48 Info in <CP::TPileupReweighting::UsePeriodConfig>: Using MC16 Period configuration
11:12:48 PathResolver ERROR Locating dev file dev/PileupReweighting/share/DSID410xxx/pileup_mc16d_dsid410250_FS.root. Do not let this propagate to a release
11:12:48 Info in <CP::TPileupReweighting::AddLumiCalcFile>: Adding LumiMetaData (scale factor=0.970874)...
11:12:48 Info in <CP::TPileupReweighting::AddLumiCalcFile>: Adding LumiMetaData (scale factor=1.010101)...
11:12:48 Info in <CP::TPileupReweighting::AddLumiCalcFile>: Adding LumiMetaData (scale factor=0.934579)...
...
To rather printing:
...
11:33:01 ToolSvc.PHYSLITE_PRWTool WARNING No config files provided, but 1 lumicalc file provided. Assuming a period config of MC16
11:33:01 Info in <CP::TPileupReweighting::UsePeriodConfig>: Using MC16 Period configuration
11:33:01 Info in <CP::TPileupReweighting::UsePeriodConfig>: Using MC16 Period configuration
11:33:01 Info in <CP::TPileupReweighting::UsePeriodConfig>: Using MC16 Period configuration
11:33:01 Info in <CP::TPileupReweighting::AddLumiCalcFile>: Adding LumiMetaData (scale factor=0.970874)...
11:33:01 Info in <CP::TPileupReweighting::AddLumiCalcFile>: Adding LumiMetaData (scale factor=1.010101)...
11:33:01 Info in <CP::TPileupReweighting::AddLumiCalcFile>: Adding LumiMetaData (scale factor=0.934579)...
...
(I was using an AOD from mc16_13TeV.410250.Sherpa_221_NNPDF30NNLO_ttbar_SingleLeptonP_MEPS_NLO.merge.AOD.e5450_e5984_s3126_r10201_r10210
for these tests.
When comparing the DAOD_PHYSLITE
files from the two runs, I got this:
Singularity> acmd.py diff-root DAOD_PHYSLITE.test1.pool.root DAOD_PHYSLITE.test2.pool.root
Py:Acmd INFO running sub-command [diff-root]...
Py:diff-root INFO comparing tree [CollectionTree] in files:
Py:diff-root INFO old: [DAOD_PHYSLITE.test1.pool.root]
Py:diff-root INFO new: [DAOD_PHYSLITE.test2.pool.root]
Py:diff-root INFO ignore leaves: ('Token',)
Py:diff-root INFO enforce leaves: ('BCID',)
Py:diff-root INFO hacks: ('m_athenabarcode', 'm_token')
Py:diff-root INFO entries: -1
Py:diff-root INFO mode: detailed
Py:diff-root INFO error mode: bailout
Py:diff-root INFO order trees: False
Py:diff-root INFO comparing over [100] entries...
Py:diff-root WARNING the following variables exist only in the old file !
Py:diff-root WARNING - [EventInfoAuxDyn.PileupWeight]
Py:diff-root INFO comparing [1434] leaves over entries...
/cvmfs/atlas-nightlies.cern.ch/repo/sw/21.2_AthDerivation_x86_64-slc6-gcc62-opt/2020-03-19T0022/AthDerivation/21.2.92.0/InstallArea/x86_64-slc6-gcc62-opt/python/PyUtils/RootUtils.py:327: RuntimeWarning: creating executor for unknown type "unsigned char*"
vals = _pythonize(val, py_name, True)
/cvmfs/atlas-nightlies.cern.ch/repo/sw/21.2_AthDerivation_x86_64-slc6-gcc62-opt/2020-03-19T0022/AthDerivation/21.2.92.0/InstallArea/x86_64-slc6-gcc62-opt/python/PyUtils/RootUtils.py:327: RuntimeWarning: creating executor for unknown type "ULong64_t*"
vals = _pythonize(val, py_name, True)
Py:diff-root INFO Found [634538] identical leaves
Py:diff-root INFO Found [0] different leaves
Py:diff-root INFO all good.
Singularity>
Which, I believe, is what we want. The EventInfoAuxDyn.PileupWeight
variable is gone, but every other variable remained exactly the same.