Merge pull request cms-sw#187 from gpetruc/CMGTools-from-CMSSW_7_2_3

TimeReport functionality, and heppy bugfixes.

CMGTools/TTHAnalysis/cfg/run_susyMultilepton_cfg.py

@@ -16,6 +16,9 @@
 #ttHLepSkim.ptCuts = []
 
 
+# switch off slow photon MC matching
+photonAna.do_mc_match = False
+
 # Event Analyzer for susy multi-lepton (at the moment, it's the TTH one)
 from CMGTools.TTHAnalysis.analyzers.ttHLepEventAnalyzer import ttHLepEventAnalyzer
 ttHEventAna = cfg.Analyzer(
@@ -57,9 +60,9 @@
 #-------- SAMPLES AND TRIGGERS -----------
 
 #-------- SEQUENCE
-from CMGTools.TTHAnalysis.samples.samples_13TeV_PHYS14 import TTH_PU40bx25, SMS_T1tttt_2J_mGl1500_mLSP100
+from CMGTools.TTHAnalysis.samples.samples_13TeV_PHYS14 import *
 
-selectedComponents = [ TTH_PU40bx25 ]
+selectedComponents = [ SMS_T1tttt_2J_mGl1200_mLSP800  ]
 
 sequence = cfg.Sequence(susyCoreSequence+[
     ttHEventAna,
@@ -68,11 +71,22 @@
 
 test = 1
 if test == 1:
-    comp = TTH_PU40bx25; comp.name = "TTH"
+    comp = TTH; comp.name = "TTH"
     #comp = SMS_T1tttt_2J_mGl1500_mLSP100
     comp.files = comp.files[:1]
     comp.splitFactor = 1
+    comp.fineSplitFactor = 4
     selectedComponents = [ comp ]
+elif test == 2:
+    for comp in selectedComponents:
+        comp.files = comp.files[:1]
+        comp.splitFactor = 1
+elif test == 3:
+    comp = TTJets
+    comp.files = comp.files[:1]
+    comp.splitFactor = 1
+    selectedComponents = [ comp ]
+
 
 # the following is declared in case this cfg is used in input to the heppy.py script
 from PhysicsTools.HeppyCore.framework.eventsfwlite import Events

CMGTools/TTHAnalysis/python/analyzers/susyCore_modules_cff.py

@@ -189,7 +189,8 @@
     photons='slimmedPhotons',
     ptMin = 20,
     etaMax = 2.5,
-    gammaID = "PhotonCutBasedIDLoose_CSA14"         
+    gammaID = "PhotonCutBasedIDLoose_CSA14",
+    do_mc_match = True,
 )
 
 

CMGTools/TTHAnalysis/python/plotter/bins/validation_plots.txt

@@ -13,8 +13,8 @@ lep1sip3d:  LepGood1_sip3d: 15,0,5;  Legend='TR'
 lep2sip3d:  LepGood2_sip3d: 15,0,5;  Legend='TR'
 lep1tightId:  LepGood1_tightId: [-0.5,0.5,1.5];  Legend='TL'
 lep2tightId:  LepGood2_tightId: [-0.5,0.5,1.5];  Legend='TL'
-lep1mvaTTH:  LepGood1_mvaTTH: 20,-1,1;  Legend='TL'
-lep2mvaTTH:  LepGood2_mvaTTH: 20,-1,1;  Legend='TL'
+lep1mvaSusy:  LepGood1_mvaSusy: 20,-1,1;  Legend='TL'
+lep2mvaSusy:  LepGood2_mvaSusy: 20,-1,1;  Legend='TL'
 lep1mcMatchId:  abs(LepGood1_mcMatchId): 28,-1.5,26.5;  Legend='TR'
 lep2mcMatchId:  abs(LepGood2_mcMatchId): 28,-1.5,26.5;  Legend='TR'
 lep1mcMatchAny:  LepGood1_mcMatchAny: [-1.5,-0.5,0.5,1.5,2.5,3.5];  Legend='TR'

PhysicsTools/Heppy/python/analyzers/core/AutoFillTreeProducer.py

@@ -50,7 +50,6 @@ def declareCoreVariables(self, tr, isMC):
         tr.var('lumi', int, storageType="i")
         tr.var('evt', int, storageType="i")
         tr.var('isData', int)
-        tr.var('xsec', float)
 
  #       self.triggerBitCheckers = []
  #       if hasattr(self.cfg_ana, 'triggerBits'):
@@ -62,6 +61,8 @@ def declareCoreVariables(self, tr, isMC):
 #                self.triggerBitCheckers.append( (T, TriggerBitChecker(trigVec)) )
 
         if isMC:
+            ## cross section
+            tr.var('xsec', float)
             ## PU weights
             tr.var("puWeight")
             ## number of true interactions
@@ -109,13 +110,14 @@ def fillCoreVariables(self, tr, event, isMC):
         tr.fill('lumi',event.input.eventAuxiliary().id().luminosityBlock())
         tr.fill('evt', event.input.eventAuxiliary().id().event())    
         tr.fill('isData', 0 if isMC else 1)
-        tr.fill('xsec',self.cfg_comp.xSection)
 
 #       triggerResults = self.handles['TriggerResults'].product()
 #       for T,TC in self.triggerBitCheckers:
 #           tr.fill("HLT_"+T, TC.check(event.object(), triggerResults))
 
         if isMC:
+            ## xsection, if available
+            tr.fill('xsec', getattr(self.cfg_comp,'xSection',1.0))
             ## PU weights, check if a PU analyzer actually filled it
             if hasattr(event,"nPU"):
                     tr.fill("nTrueInt", event.nPU)

PhysicsTools/Heppy/python/physicsobjects/GenParticle.py

@@ -14,4 +14,5 @@ def printGenParticle(self):
         return theStr
 setattr(ROOT.reco.GenParticle,"__str__",printGenParticle)
 
-from ROOT.reco import GenParticle
+#from ROOT.reco import GenParticle   # sometimes doesn't work
+GenParticle = ROOT.reco.GenParticle  # this instead does

PhysicsTools/HeppyCore/python/framework/heppy.py

@@ -48,7 +48,8 @@ def runLoop( comp, outDir, config, options):
     loop = Looper( fullName,
                    config,
                    options.nevents, 0,
-                   nPrint = options.nprint)
+                   nPrint = options.nprint,
+                   timeReport = options.timeReport)
     print loop
     if options.iEvent is None:
         loop.loop()
@@ -202,6 +203,12 @@ def main( options, args ):
                       action='store_true',
                       help="stay in the command line prompt instead of exiting",
                       default=False)
+    parser.add_option("-t", "--timereport", 
+                      dest="timeReport",
+                      action='store_true',
+                      help="Make a report of the time used by each analyzer",
+                      default=False)
+
 
 
     (options,args) = parser.parse_args()

PhysicsTools/HeppyCore/python/framework/looper.py

@@ -10,7 +10,7 @@
 from platform import platform 
 from math import ceil
 from event import Event
-import time
+import timeit
 
 class Setup(object):
     '''The Looper creates a Setup object to hold information relevant during 
@@ -52,7 +52,7 @@ class Looper(object):
     def __init__( self, name,
                   config, 
                   nEvents=None,
-                  firstEvent=0, nPrint=0 ):
+                  firstEvent=0, nPrint=0, timeReport=False ):
         """Handles the processing of an event sample.
         An Analyzer is built for each Config.Analyzer present
         in sequence. The Looper can then be used to process an event,
@@ -79,6 +79,7 @@ def __init__( self, name,
         self.nEvents = nEvents
         self.firstEvent = firstEvent
         self.nPrint = int(nPrint)
+        self.timeReport = [ {'time':0.0,'events':0} for a in self.analyzers ] if timeReport else False
         tree_name = None
         if( hasattr(self.cfg_comp, 'tree_name') ):
             tree_name = self.cfg_comp.tree_name
@@ -158,10 +159,10 @@ def loop(self):
                     # print 'event', iEv
                     if not hasattr(self,'start_time'):
                         print 'event', iEv
-                        self.start_time = time.time()
+                        self.start_time = timeit.default_timer()
                         self.start_time_event = iEv
                     else:
-                        print 'event %d (%.1f ev/s)' % (iEv, (iEv-self.start_time_event)/float(time.time() - self.start_time))
+                        print 'event %d (%.1f ev/s)' % (iEv, (iEv-self.start_time_event)/float(timeit.default_timer() - self.start_time))
 
                 self.process( iEv )
                 if iEv<self.nPrint:
@@ -186,10 +187,16 @@ def process(self, iEv ):
         """
         self.event = Event(iEv, self.events[iEv], self.setup)
         self.iEvent = iEv
-        for analyzer in self.analyzers:
+        for i,analyzer in enumerate(self.analyzers):
             if not analyzer.beginLoopCalled:
                 analyzer.beginLoop()
-            if analyzer.process( self.event ) == False:
+            start = timeit.default_timer()
+            ret = analyzer.process( self.event )
+            if self.timeReport:
+                self.timeReport[i]['events'] += 1
+                if self.timeReport[i]['events'] > 0:
+                    self.timeReport[i]['time'] += timeit.default_timer() - start
+            if ret == False:
                 return (False, analyzer.name)
         return (True, analyzer.name)
 
@@ -201,6 +208,15 @@ def write(self):
         for analyzer in self.analyzers:
             analyzer.write(self.setup)
         self.setup.close() 
+
+        if self.timeReport:
+            allev = max([x['events'] for x in self.timeReport])
+            print "\n      ---- TimeReport (all times in ms; first evt is skipped) ---- "
+            print "%9s   %9s    %9s   %9s   %s" % ("processed","all evts","time/proc", " time/all", "analyer")
+            print "%9s   %9s    %9s   %9s   %s" % ("---------","--------","---------", "---------", "-------------")
+            for ana,rep in zip(self.analyzers,self.timeReport):
+                print "%9d   %9d   %10.2f  %10.2f   %s" % ( rep['events'], allev, 1000*rep['time']/(rep['events']-1) if rep['events']>1 else 0, 1000*rep['time']/(allev-1) if allev > 1 else 0, ana.name)
+            print ""
         pass