No divide

__analysis__.py

@@ -139,7 +139,7 @@ def sampleSpecs(self, tag = "") :
         def sampleSpecDict(looper) :
             looper.setupSteps(minimal = True)
             sampleSpec = next( s for s in confSamples if s.weightedName == looper.name ) 
-            return {"name":looper.name, "outputFileName":looper.steps[0].outputFileName,
+            return {"name":looper.name, "outputFileName":looper.steps[0].outputFileName, "nInDivide": sampleSpec.nInDivide,
                     "color":sampleSpec.color, "markerStyle":sampleSpec.markerStyle, "nCheck":self.sampleDict[sampleSpec.name].nCheck }
         return [ sampleSpecDict(looper) for looper in self.listsOfLoopers[tag] ]
 ############

__organizer__.py

@@ -27,8 +27,12 @@ def __init__(self,samples,dirs = None,keys = [], prefixesNoScale=[]) :
                     [ hist.Scale(  1.0 / sample['nJobs'] ) for hist,sample in zip(self[key],samples) if hist ]
                 elif any([key.startswith(prefix) for prefix in prefixesNoScale]):
                     continue
-                else: [ hist.Scale(sample["xs"]/sample['nEventsIn']) for hist,sample in zip(self[key],samples)
-                        if hist and sample['nEventsIn'] and "xs" in sample ]
+                else:
+                    for hist, sample in zip(self[key], samples):
+                        if hist and "xs" in sample:
+                            hist.Scale(sample["xs"])
+                            if sample['nEventsIn'] and sample["nInDivide"]:
+                                hist.Scale(1.0 / sample['nEventsIn'])
 
         def __str__(self) : return "%s: %s"%self.nameTitle
 

__plotter__.py

@@ -20,13 +20,6 @@ def combineBinContentAndError(histo, binToContainCombo, binToBeKilled) :
     histo.SetBinError(binToContainCombo, math.sqrt(xflowError**2+currentError**2))
 
 
-def mcLumi(nEvents=None, w=None, xs=None):
-    """see docs/mcLumi.txt"""
-    if not w*xs:
-        return 0.0
-    return nEvents**2/(w*xs)
-
-
 def sampleName(sample):
     if len(sample.get("sources", [])) == 1:
         return sample["sources"][0]["name"]
@@ -38,24 +31,30 @@ def sampleInfo(samples, trim=""):
     out = []
     for sample in samples:
         if "xs" in sample:
-            xs = sample["xs"]
-            lumi = mcLumi(nEvents=sample["nEventsIn"],
-                          w=sample["weightIn"],
-                          xs=sample["xs"])
+            xsIn = sample["xs"]
+            xsEff = sample["xs"] * sample["weightIn"]
+            if sample["nInDivide"] and sample["nEventsIn"]:
+                xsEff /= sample["nEventsIn"]
+
+            # see docs/mcLumi.txt
+            lumi = sample["nEventsIn"] / xsEff if xsEff else 0.0
         elif "lumi" in sample:
-            xs = None
+            xsIn = None
+            xsEff = None
             lumi = sample["lumi"]
         else:
             assert False, sample
 
         name = sampleName(sample)
         if trim:
             name = name.replace(trim, "")
+
         out.append((name,
-                    "%d" % sample['nEventsIn'],
+                    "%g" % sample['nEventsIn'],  # %g handles 'float epsilon less than nearest int' better than %d
                     "%3.2e" % sample['weightIn'],
                     "%3.2e" % (lumi/1.0e3),
-                    "%3.2e" % (xs*1.0e3) if xs is not None else "",
+                    "%3.2e" % (xsIn*1.0e3) if xsIn is not None else "",
+                    "%3.2e" % (xsEff*1.0e3) if xsEff is not None else "",
                     ))
     return out
 
@@ -508,7 +507,7 @@ def printSampleList(self, x, rows=[], header="", sigma=None):
         text.SetTextSize(0.38*text.GetTextSize())
         defSize = text.GetTextSize()
 
-        rows = [("name", "nEventsIn", "weightIn", "lumi(/fb)", "xs(fb)")] + rows
+        rows = [("name", "nEventsIn", "weightIn", "lumi(/fb)", "xsIn(fb)", "xsEff(fb)")] + rows
         realRows = filter(lambda x: len(x[1]), rows)
         if len(realRows) == 1:
             return

samples/__init__.py

@@ -13,14 +13,21 @@ def printNumberEvents(sampleHolder, treeName='tree'):
 def test(sampleHolder) :
     return [(name,len(eval(ss.filesCommand))) for name,ss in sorted(sampleHolder.items())]
 
-def specify(names = [], overrideLumi = None, xsPostWeights = None, effectiveLumi = None, nFilesMax = None, nEventsMax = None, weights = [], color = 1, markerStyle = 1 , weightedName = None) :
+def specify(names=[], overrideLumi=None, xsPostWeights=None, effectiveLumi=None, nFilesMax=None, nEventsMax=None, weights=[], nInDivide=True, color=1, markerStyle=1 , weightedName=None):
     assert not (overrideLumi and type(names)==list)
     if type(names) != list : names = [names]
     if type(weights) != list : weights = [weights]
-    samplespec = collections.namedtuple("samplespec", "name weightedName overrideLumi xsPostWeights effectiveLumi nFilesMax nEventsMax weights color markerStyle")
+
+    if not nInDivide:  # typically, all events are required for correct normalization
+        for s in ["nFilesMax", "nEventsMax", "effectiveLumi"]:
+            x = eval(s)
+            if x is not None:
+                print "WARNING: nInDivide=False, but %s=%s (%s)" % (s, x, str(names))
+
+    samplespec = collections.namedtuple("samplespec", "name weightedName overrideLumi xsPostWeights effectiveLumi nFilesMax nEventsMax weights nInDivide color markerStyle")
     weightNames = [w if type(w)==str else w.name for w in weights]
-    return [samplespec(name,'.'.join([name]+weightNames),overrideLumi,xsPostWeights,effectiveLumi,nFilesMax,nEventsMax,weights,color,markerStyle) for name in names]
-    
+    return [samplespec(name, '.'.join([name]+weightNames), overrideLumi, xsPostWeights, effectiveLumi, nFilesMax, nEventsMax, weights, nInDivide, color, markerStyle) for name in names]
+
 class SampleHolder(dict) :
     sample = collections.namedtuple("sample", "filesCommand xs lumi ptHatMin nCheck")
     def __init__(self) : self.inclusiveGroups = []
@@ -35,6 +42,7 @@ def update(self, other) :
         for group in other.inclusiveGroups : self.addInclusiveGroup(group)
 
     def add(self, name, filesCommand = None, xs = None, lumi = None, ptHatMin = None, nCheck = None) :
+        assert "/" not in name,  "name '%s' contains at least one /" % name
         assert name not in self,  "%s already specified" % name
         assert lumi or xs,                      "Underspecified sample: %s"%name
         assert not (lumi and (xs or ptHatMin)), "Overspecified sample: %s"%name

sites/uw_cmsTemplate.condor

@@ -6,6 +6,8 @@ transfer_output_files = OUTFLAG
 Should_Transfer_Files = YES
 WhenToTransferOutput = ON_EXIT
 
+Requirements = HAS_CMS_HDFS
+
 Output = JOBFLAG_$(Cluster)_$(Process).stdout
 Error = JOBFLAG_$(Cluster)_$(Process).stderr
 Log = JOBFLAG_$(Cluster)_$(Process).log

steps/other.py

@@ -46,14 +46,13 @@ class skimmer(analysisStep):
     created.
     """
 
-    def __init__(self, mainChain=True, otherChains=True, extraVars=[], haddOutput=False):
+    def __init__(self, mainChain=True, otherChains=True, extraVars=[], haddOutput=False, suffix="skim"):
         assert mainChain or extraVars
         self.outputTree = None
         self.moreName = "(see below)"
-        for var in ["mainChain", "otherChains", "extraVars"]:
+        for var in ["mainChain", "otherChains", "extraVars", "haddOutput", "suffix"]:
             setattr(self, var, eval(var))
         self.addresses = None
-        self.haddOutput = haddOutput
 
     def requiresNoSetBranchAddress(self):
         return True  # check
@@ -141,7 +140,7 @@ def endFunc(self, chains):
         self.outputFile.Close()
 
     def outputSuffix(self):
-        return "_skim.root"
+        return "_%s.root" % self.suffix
 
     def modifiedFileName(self, s):
         l = s.split("/")