Merge pull request #1 from musella/hh_tag_80X_20180717_workspaces

cleanup double-H setup and turn on HH categorization
nadya-chernyavskaya · Sep 27, 2018 · 18c8a02 · 18c8a02
2 parents 4c164d9 + 9fe276a
commit 18c8a02
Show file tree

Hide file tree

Showing 17 changed files with 15 additions and 119 deletions.
diff --git a/.gitignore b/.gitignore
@@ -24,4 +24,7 @@ AutoDict_*
 *.txt
 *.log
 *.png
-*.pdf
+*.pdf
+
+# This is a symlink created by setup.sh
+DataFormats/src/classes_def.xml
diff --git a/MyAnalysis/test/MyAnalysis.py b/MyAnalysis/test/MyAnalysis.py
diff --git a/Systematics/test/workspaceStd.py b/Systematics/test/workspaceStd.py
@@ -109,7 +109,7 @@
                            'doPdfWeights'
                            )
 customize.options.register('dumpTrees',
-                           True,
+                           False,
                            VarParsing.VarParsing.multiplicity.singleton,
                            VarParsing.VarParsing.varType.bool,
                            'dumpTrees'
@@ -571,7 +571,7 @@
         producer =   cms.EDProducer('flashggbRegressionProducer80',
                                     JetTag=coll,
                                     rhoFixedGridCollection = cms.InputTag('fixedGridRhoFastjetAll'),
-                                    bRegressionWeightfile= cms.untracked.string(os.environ["CMSSW_BASE"]+"/src/flashgg/MetaData/data/DNN_models/model-18"),
+                                    bRegressionWeightfile= cms.untracked.string(os.environ["CMSSW_BASE"]+"/src/flashgg/Taggers/data/DNN_models/model-18"),
                                     y_mean = cms.untracked.double(1.0454729795455933),#check MetaData/data/DNN_models/config.json
                                     y_std = cms.untracked.double( 0.31628304719924927)
                                     )

diff --git a/MetaData/data/DNN_models/checkpoint → Taggers/data/DNN_models/checkpoint b/MetaData/data/DNN_models/checkpoint → Taggers/data/DNN_models/checkpoint
diff --git a/MetaData/data/DNN_models/config.json → Taggers/data/DNN_models/config.json b/MetaData/data/DNN_models/config.json → Taggers/data/DNN_models/config.json
diff --git a/...a/DNN_models/model-18.data-00000-of-00001 → ...a/DNN_models/model-18.data-00000-of-00001 b/...a/DNN_models/model-18.data-00000-of-00001 → ...a/DNN_models/model-18.data-00000-of-00001
diff --git a/MetaData/data/DNN_models/model-18.index → Taggers/data/DNN_models/model-18.index b/MetaData/data/DNN_models/model-18.index → Taggers/data/DNN_models/model-18.index
diff --git a/MetaData/data/DNN_models/model-18.meta → Taggers/data/DNN_models/model-18.meta b/MetaData/data/DNN_models/model-18.meta → Taggers/data/DNN_models/model-18.meta
diff --git a/MetaData/data/DNN_models/model-18.pb → Taggers/data/DNN_models/model-18.pb b/MetaData/data/DNN_models/model-18.pb → Taggers/data/DNN_models/model-18.pb
diff --git a/...Transformation_20180628_newcode_data.root → ...Transformation_20180628_newcode_data.root b/...Transformation_20180628_newcode_data.root → ...Transformation_20180628_newcode_data.root
diff --git a/...GluToHHTo2B2G_node_SM_13TeV-madgraph.root → ...GluToHHTo2B2G_node_SM_13TeV-madgraph.root b/...GluToHHTo2B2G_node_SM_13TeV-madgraph.root → ...GluToHHTo2B2G_node_SM_13TeV-madgraph.root
diff --git a/...raining_with_27_06_2018_fixed.weights.xml → ...raining_with_27_06_2018_fixed.weights.xml b/...raining_with_27_06_2018_fixed.weights.xml → ...raining_with_27_06_2018_fixed.weights.xml
diff --git a/...ng_with_27_06_2018_newcode_v2.weights.xml → ...ng_with_27_06_2018_newcode_v2.weights.xml b/...ng_with_27_06_2018_newcode_v2.weights.xml → ...ng_with_27_06_2018_newcode_v2.weights.xml
diff --git a/Taggers/python/flashggDoubleHTag_cfi.py b/Taggers/python/flashggDoubleHTag_cfi.py
@@ -34,14 +34,15 @@
                                    MJJBoundariesUpper = cms.vdouble(146.,153.,150.,155.,150.,147.,154.,151.,155.,155.,155.,148.), #for each category following the convention cat0=MX0 MVA0, cat1=MX1 MVA0, cat2=MX2 MVA0....
                                    MVAConfig = cms.PSet(variables=cms.VPSet(), # variables are added below
                                                         classifier=cms.string("BDT::bdt"), # classifier name
-                                                        weights=cms.FileInPath("flashgg/MetaData/data/HHTagger/training_with_27_06_2018_fixed.weights.xml"), # path to TMVA weights
+                                                        weights=cms.FileInPath("flashgg/Taggers/data/HHTagger/training_with_27_06_2018_fixed.weights.xml"), # path to TMVA weights
                                                         regression=cms.bool(False), # this is not a regression
                                                         multiclass=cms.bool(True), # this is multiclass 
                                                         multiclassSignalIdx=cms.int32(2), # this is multiclass index for Signal
                                                         ),
                                    doMVAFlattening=cms.bool(True),#do transformation of cumulative to make it flat
-                                   doCategorization=cms.bool(False),#do categorization based on MVA x MX or only fill first tree with all events
-                                   MVAFlatteningFileName=cms.untracked.FileInPath("flashgg/MetaData/data/HHTagger/cumulativeTransformation_20180628_newcode_data.root")#FIXME, this should be optional, is it?
+                                   doCategorization=cms.bool(True),#do categorization based on MVA x MX or only fill first tree with all events
+                                   MVAFlatteningFileName=cms.untracked.FileInPath("flashgg/Taggers/data/HHTagger/cumulativeTransformation_20180628_newcode_data.root")#FIXME, this should be optional, is it? 
+                                   ## 
                                   ) 
 
 cfgTools.addVariables(flashggDoubleHTag.MVAConfig.variables,

diff --git a/Taggers/python/flashggbRegressionProducer80_cfi.py b/Taggers/python/flashggbRegressionProducer80_cfi.py
@@ -11,7 +11,7 @@
 flashggbRegressionProducer80= cms.EDProducer('flashggbRegressionProducer80',
 #                                               JetTag=coll,
                                            JetTag=cms.InputTag("flashggUnpackedJets","0"),
-                                           bRegressionWeightfile= cms.untracked.string("flashgg/MetaData/data/DNN_models/model-18"), 
+                                           bRegressionWeightfile= cms.untracked.string("flashgg/Taggers/data/DNN_models/model-18"), 
                                            y_mean = cms.untracked.double(1.0454729795455933),
                                            y_std =cms.untracked.double( 0.31628304719924927)
                                            )

diff --git a/setup_8_0_X.sh b/setup_8_0_X.sh
@@ -156,6 +156,10 @@ cp $CMSSW_BASE/src/flashgg/Systematics/data/Winter_2016_reReco_v1_ele_smearings.
 cp $CMSSW_BASE/src/flashgg/Systematics/data/Moriond17_74x_pho_scales.dat $CMSSW_BASE/src/EgammaAnalysis/ElectronTools/data
 cp $CMSSW_BASE/src/flashgg/Systematics/data/Moriond17_74x_pho_smearings.dat $CMSSW_BASE/src/EgammaAnalysis/ElectronTools/data
 
+echo "downloading tensoflow interface"
+git clone -b tf_py_cpython https://gitlab.cern.ch/mrieger/CMSSW-DNN.git DNN
+./DNN/setup_legacy.sh
+
 echo "adding hook for indentation"
 ln -s $CMSSW_BASE/src/flashgg/Validation/scripts/flashgg_indent_check.sh $CMSSW_BASE/src/flashgg/.git/hooks/pre-commit
 

diff --git a/testHHbbgg/HHbbgg_MC_2016_80X.json b/testHHbbgg/HHbbgg_MC_2016_80X.json
-Original file line number
+Diff line change
@@ Expand Up / @@ -24,4 +24,7 @@ AutoDict_* @@
     *.txt
     *.log
     *.png
-    *.pdf
+    *.pdf
+    # This is a symlink created by setup.sh
+    DataFormats/src/classes_def.xml