Update get siblings

Configuration/python/configurationOptions.py

@@ -1811,7 +1811,7 @@
 
     'EGamma_2022A' : 6,
     'EGamma_2022B' : 112,
-    'EGamma_2022C' : 200,
+    'EGamma_2022C' : 3354,
     'EGamma_2022D' : 33,
     'EGamma_2022E' : 800,
     'EGamma_2022F' : 1000,

Configuration/python/histogramUtilities.py

@@ -102,7 +102,7 @@ def getHistIntegral(sample,condor_dir,channel,hist,xlo,xhi):
         # print "xhi is outside the range of the histogram, will include all the overflow instead"
         xhiBin = histogram.GetXaxis().FindBin(float(xhi))
     #intError = Double (0.0)
-    intError = 0.0
+    intError = c_double(0.0)
     integral = histogram.IntegralAndError(xloBin, xhiBin, intError)
 
     inputFile.Close()

DBTools/python/condorSubArgumentsSet.py

@@ -10,7 +10,7 @@
     CondorSubArgumentsSet = {
         1   :  {'Executable'             :  ''},
         2   :  {'Universe'               :  'vanilla'},
-        3   :  {'request_memory '        : '2048MB'},
+        3   :  {'request_memory '        : '4092MB'},
         4   :  {'Arguments'              :  ''},
         5   :  {'Output'                 :  'condor_$(Process).out'},
         6   :  {'Error'                  :  'condor_$(Process).err'},
@@ -23,7 +23,7 @@
         1   :  {'Executable'             :  ''},
         2   :  {'Universe'               :  'vanilla'},
         3   :  {'Getenv'                 :  'True'},
-        4   :  {'request_memory '        : '2048MB'},
+        4   :  {'request_memory '        : '4092MB'},
         5   :  {'Arguments'              :  ''},
         6   :  {'Output'                 :  'condor_$(Process).out'},
         7   :  {'Error'                  :  'condor_$(Process).err'},

DBTools/python/createEventLists.py

@@ -0,0 +1,57 @@
+#!/usr/bin/python3
+
+import json
+import ROOT as r
+import numpy as np
+import sys
+import os
+
+r.gInterpreter.Declare(
+    '''
+    #include "OSUT3Analysis/DBTools/interface/getEvents.h"
+
+    void dummyWrapper(){
+        getEvents();
+    }
+    '''
+)
+r.gSystem.Load('libOSUT3AnalysisDBTools.so')
+
+if __name__ == "__main__":
+
+
+    if not len(sys.argv) >= 3:
+        print("Error: Need to provide the input json file")
+        sys.exit(1)
+
+    outputDir = '/data/users/mcarrigan/condor/EventLists/'
+    if len(sys.argv) >= 4:
+        outputDir = sys.argv[3]
+
+    f_json = sys.argv[1]
+
+    job = int(sys.argv[2])
+
+    dataset = f_json.split('/')[-1].split('-')[0]
+
+    outputPath = outputDir + dataset + '/'
+
+    if not os.path.exists(outputPath):
+        os.mkdir(outputPath)
+
+    with open(f_json) as secondary_fin:
+        secondary_dict = json.load(secondary_fin)
+        filename = list(secondary_dict.keys())[job]
+
+    #for filename in secondary_dict.keys():
+    if not filename.startswith('root://'): filename = 'root://cmsxrootd.fnal.gov:/' + filename
+    print("Getting events for", filename)
+    events = r.getEventsInFile(filename)
+    events = sorted(events, key=lambda x: (x.runNum, x.lumiBlock, x.event))
+    tmpEvents = np.array([str(x.runNum)+':'+str(x.lumiBlock)+':'+str(x.event) for x in events])
+    print(tmpEvents, len(tmpEvents))
+    fileStr = filename.split('/')[-1]
+    outputFile = outputPath + fileStr
+    np.savez(outputFile, eventList=tmpEvents)
+    print("saved file {} with {} events".format(outputFile, len(tmpEvents)))
+    #break

DBTools/python/getSiblings.py

@@ -47,6 +47,7 @@ def __init__(self, inputJSON, secondaryJSON, label):
         self.nJobs = -1
         self.jobNumber = -1
         self.eventsPerJob = -1
+        self.redirector = None
 
         self.local=False
 
@@ -59,6 +60,7 @@ def getSiblings(self):
             self.getFilesFromList(args.jobNumber, args.totalJobs)
             self.findMatches()
         else:
+            print("Running default option...")
             self.getFilesFromList(args.jobNumber, args.totalJobs)
             self.findMatches()
 
@@ -72,12 +74,14 @@ def getDASInfo(dataset, jsonName=None):
 
         test = ast.literal_eval(miniaod)
         files = [x['file'][0]['name'] for x in test]
+        events = [len(x['events'][0]['number']) if 'events' in list(x.keys()) else 0 for x in test]
         lumis = [[x['lumi'][0]['number']] if isinstance(x['lumi'][0]['number'], int) else x['lumi'][0]['number'] for x in test]
         runs = [[x['run'][0]['run_number']] if isinstance(x['run'][0]['run_number'], int) else x['run'][0]['run_number'] for x in test]
 
         miniDict = {}
 
-        for f, l, r in zip(files, lumis, runs):
+        for f, l, r, e in zip(files, lumis, runs, events):
+            if e == 0: continue
             miniDict[f] = {'lumis': l, 'runs': r}
 
         if jsonName:
@@ -125,20 +129,32 @@ def findMatches(self, jsonName='default.json'):
             with open(self.secondaryJSON) as secondary_fin:
                 secondary_dict = json.load(secondary_fin)
 
+                secondary_dataset = self.secondaryJSON.split('/')[-1].split('-')[0]
+                eventsDir = '/data/users/mcarrigan/condor/EventLists/' + secondary_dataset
+
                 for inputFile in self.inputFiles:
-                    if inputFile not in primary_dict.keys():
-                        continue
                     p_file = inputFile
+                    if 'root://' in inputFile: 
+                        self.redirector = 'root://'+inputFile.split('://')[1]
+                        p_file = '/'+inputFile.split('://')[-1]
+                        print("saving redirector as", self.redirector)
+                    #print("looking for file", p_file)
+                    if p_file not in primary_dict.keys():
+                        continue
+                    #print("found file ", p_file)
                     primary_info = primary_dict[p_file]
-                    if p_file not in self.inputFiles: continue
+                    if p_file not in self.inputFiles and inputFile not in self.inputFiles: continue
                     sibs = []
+                    #print("looking for siblings")
                     for s_file, secondary_info in secondary_dict.items():
                         if len(np.intersect1d(primary_info['runs'], secondary_info['runs'])) == 0: continue
                         if len(np.intersect1d(primary_info['lumis'], secondary_info['lumis'])) != 0:
                             sibs.append(s_file)
-
+
+                    #print("There are {} siblings".format(len(sibs))) 
                     siblings[p_file] = sibs
-                    self.getEventList(p_file, sibs)
+
+                    self.getEventList(inputFile, sibs, eventsDir)
 
                     if self.eventsPerJob != -1 and len(self.sharedEvents) > self.eventsPerJob:
                         break                        
@@ -152,24 +168,33 @@ def findMatches(self, jsonName='default.json'):
 
         self.siblingDict = siblings
 
-    def getEventList(self, primaryFile, siblings):
+    def getEventList(self, primaryFile, siblings, eventsDir):
 
         primaryEvents = np.array([])
         if not primaryFile in self.inputFiles: 
             print("File is missing from input file list")
             return
         if not primaryFile.startswith("root://") and not self.local: 
             primaryFile = 'root://cms-xrd-global.cern.ch:/' + primaryFile
+        print("getting primary file events")
         events = r.getEventsInFile(primaryFile)
         tmpEvents = np.array([str(x.runNum)+':'+str(x.lumiBlock)+':'+str(x.event) for x in events])
         primaryEvents = np.concatenate((primaryEvents, tmpEvents))
 
         secondaryEvents = np.array([])
-        for filename in siblings:
-            if not filename.startswith("root://") and not self.local: 
-                filename = 'root://cms-xrd-global.cern.ch:/' + filename
-            events = r.getEventsInFile(filename)
-            tmpEvents = np.array([str(x.runNum)+':'+str(x.lumiBlock)+':'+str(x.event) for x in events])
+        for ifile, filename in enumerate(siblings):
+            fileStr = filename.split('/')[-1]
+            filePath = '/'.join([eventsDir,fileStr+'.npz'])
+            print("Looking for event list for file", filePath, filename)
+            if os.path.exists(filePath):
+                print("found event list!")
+                tmpEvents = np.load(filePath)['eventList']
+            else:
+                print("getting secondary file events", filename)
+                if not filename.startswith("root://") and not self.local: 
+                    filename = 'root://cms-xrd-global.cern.ch:/' + filename
+                events = r.getEventsInFile(filename)
+                tmpEvents = np.array([str(x.runNum)+':'+str(x.lumiBlock)+':'+str(x.event) for x in events])
             secondaryEvents = np.concatenate((secondaryEvents, tmpEvents))
 
         this_sharedEvents = np.intersect1d(primaryEvents, secondaryEvents)
@@ -180,6 +205,7 @@ def getEventList(self, primaryFile, siblings):
 
 
     def getFilesFromList(self, jobNumber, nJobs):
+        print("getting files from list")
         sys.path.append(os.getcwd())
 
         self.jobNumber = jobNumber
@@ -191,6 +217,7 @@ def getFilesFromList(self, jobNumber, nJobs):
         #If no job number or number of jobs is passed use the full file list
         if jobNumber == -1 or nJobs == -1:
             self.inputFiles = datasetInfo.listOfFiles
+            print("Using full file list")
             return
 
         filesPerJob = int (math.floor (len (datasetInfo.listOfFiles) / nJobs))
@@ -204,7 +231,7 @@ def getFilesFromList(self, jobNumber, nJobs):
         if runList[0].startswith('file:'):
             runList = [x.split('file:')[1] for x in runList]
 
-        #print("This is the run list:\n",runList)
+        print("This is the run list:\n",runList)
 
         self.inputFiles = runList
 

DBTools/python/osusub_cfg.py

@@ -198,7 +198,7 @@ def skimListExists(dataset):
 
 def getSiblingList(sibList, runList, siblingDataset):
 
-  print("Trying to get sibling list")
+  print("Trying to get sibling list from", sibList)
   siblings = []
 
   fin = open(sibList, 'r')

DBTools/scripts/eventListWrapper.sh

@@ -0,0 +1,13 @@
+#!/usr/bin/bash
+
+(>&2 echo "Starting job on " `date`) # Date/time of start of job
+(>&2 echo "Running on: `uname -a`") # Condor job is running on this node
+(>&2 echo "System software: `cat /etc/redhat-release`") # Operating System on that node
+(>&2 echo "Arguments passed to this script are: $@") #print all arguments
+
+export CMSSW_VERSION_LOCAL=$1
+source /cvmfs/cms.cern.ch/cmsset_default.sh
+export SCRAM_ARCH=slc7_amd64_gcc820
+#eval `scramv1 runtime -sh`
+
+python3 createEventLists.py $2 $3

DBTools/scripts/osusub.py

@@ -907,6 +907,8 @@ def MakeSpecificConfig(Dataset, Directory, SkimDirectory, Label, SkimChannelName
         ConfigFile.write("    print(\"No valid grid proxy. Not adding sibling files.\")\n")
         if arguments.localSkim != None:
             ConfigFile.write("siblings = ['file:{0}{1}'.format(" + arguments.localSkim + ", sib.split('/')[-1]) for sib in siblings] \n")
+        elif arguments.Redirector != "":
+            ConfigFile.write("siblings = ['root://{0}:/{1}'.format(\'" + RedirectorDic[arguments.Redirector] + "\', sib) for sib in siblings] \n")
         ConfigFile.write("pset.process.source.secondaryFileNames.extend(siblings)\n\n")
 
     # If the dataset has a Run3 skim sibling defined and not run over skim, add the corresponding files to the secondary file names
@@ -923,6 +925,8 @@ def MakeSpecificConfig(Dataset, Directory, SkimDirectory, Label, SkimChannelName
         ConfigFile.write("    print( \"No valid grid proxy. Not adding sibling files.\")\n" )
         if arguments.localSkim != None:
             ConfigFile.write("siblings = ['file:{0}{1}'.format(\'" + arguments.localSkim + "\', sib.split('/')[-1]) for sib in siblings] \n")
+        elif arguments.Redirector != "":
+            ConfigFile.write("siblings = ['root://{0}:/{1}'.format(\'" + RedirectorDic[arguments.Redirector] + "\', sib) for sib in siblings] \n")
         ConfigFile.write("pset.process.source.secondaryFileNames.extend(siblings)\n\n")
 
     #if ...: make this an if statement for running over no cuts