def initialize(self):
"""Load the necessary settings from the config file."""
self.p_objects = self.p_settings.get(self.p_cfg_name,'objects')
self.p_scalefactor = config.str2bool(self.p_settings.get(self.p_cfg_name,'eventweights'))
self.p_lepton = self.p_settings.get(self.p_cfg_name,'lepton')
self.p_extra_save = self.p_settings.get(self.p_cfg_name,'extra_saveAs')
self.p_plot_type = self.p_settings.get(self.p_cfg_name,'plot_framework') # python/root
self.p_plot1d = self.p_settings.get(self.p_cfg_name,'1dplot') # for saving data
self.p_plot2d = self.p_settings.get(self.p_cfg_name,'2dplot') # for saving data
self.p_cuts = self.p_settings.get(self.p_cfg_name,'cutsfile')
self.p_cuts = config.processCuts(self.p_cuts)
self.p_variables = self.p_settings.get(self.p_cfg_name,'variables').split(',')
self.p_eff_x_vars = self.p_settings.get(self.p_cfg_name,'eff_x').split(',')
self.p_treename = self.p_settings.get(self.p_cfg_name,'treename')
self.eff_conditions = self.p_settings.get(self.p_cfg_name,'eff_y').split(',')
self.p_nevents = int(self.p_settings.get(self.p_cfg_name,'NEvents'))
p_files = self.p_settings.get(self.p_cfg_name,'files')
self.p_files = open(p_files,'r').readlines()
self.p_read_data = config.str2bool(self.p_settings.get(self.p_cfg_name,'read_data'))
p_custom_vars = self.p_settings.get(self.p_cfg_name,'custom_variables')
if p_custom_vars == 'None':
self.p_custom_vars = ''
else:
self.p_custom_vars = p_custom_vars
## -- Easy variables to initialize -- ##
self.p_btag_wkpt = info.btagging_WP(self.p_settings.get(self.p_cfg_name,'btag_wkpt'))
self.outputfilenames = ['data/'+fname.split('/')[-1].split('.')[0]+'_'+self.p_lepton+'_'+self.p_extra_save for fname in self.p_files]
return self.p_settings
def main(self,parser):
"""
Main function in the systematics class -- does all the directing for
creating histograms and json files.
"""
## -- Configuration -- ##
p_selection = parser.get('systematics','selection') # ex. pyMiniSL/SelectionBase.py
p_make_ntuple = str2bool(parser.get('systematics','make_ntuple')) # ex. True
p_make_jsons = str2bool(parser.get('systematics','make_jsons')) # ex. True
p_make_hists = str2bool(parser.get('systematics','make_hists')) # ex. True
## ------------------- ##
selection = p_selection.split('/')[-1].split('.')[0] # pyMiniSL/SelectionBase.py -> SelectionBase
if p_make_ntuple and p_make_jsons and p_make_hists:
print
print " You have chosen to produce a new ntuples,"
print " json, and histogram outputs."
print " This will take a long time to do."
print " If this is not what you meant to do, please"
print " exit the program and modify PyMiniAna.cfg."
print " Will continue with the program."
print
if p_make_ntuple:
## Using the same miniSL code to make ntuples for the systematics
## This ensures that the systematics get the same treatment as the nominal
from miniSL import MiniSL
MakeMiniNtuple = MiniSL()
logging.info(" -- Make new ntuples ")
print " -- Make new ntuples for a given selection"
if selection == 'pre2preSelection':
nplist = nuisparams.npdict()
for np in nplist:
MakeMiniNtuple.execute(parser,treename=np)
else:
MakeMiniNtuple.execute(parser,treename='systs')
if p_make_jsons or p_make_hists:
logging.info(" -- Make histograms and json files ")
print " -- Make histograms for TXFITTER && json files for Data/Pred. plots"
from pySystematics.tree2hist import Tree2Hist
MakeHist = Tree2Hist(parser)
logging.critical(" Finished systematics.py class Systematics")
return
def initialize(self):
"""initialize some variables using the settings file"""
self.p_lepton = self.m_cfg_parser.get(self.p_cfg_name,'lepton') # ex. muel
self.p_logplot = str2bool(self.m_cfg_parser.get(self.p_cfg_name,'logplot')) # ex. False
self.p_systematics = str2bool(self.m_cfg_parser.get(self.p_cfg_name,'systematics')) # ex. True
self.p_ana_status = self.m_cfg_parser.get(self.p_cfg_name,'ana_status') # ex. "Internal"
self.p_extra_save = self.m_cfg_parser.get(self.p_cfg_name,'extra_saveAs')
self.p_underflow = str2bool(self.m_cfg_parser.get(self.p_cfg_name,'underflow')) # ex. True
self.p_overflow = str2bool(self.m_cfg_parser.get(self.p_cfg_name,'overflow')) # ex. True
self.p_scalefactor = str2bool(self.m_cfg_parser.get(self.p_cfg_name,'eventweights'))
self.p_plot1d_type = self.m_cfg_parser.get(self.p_cfg_name,'1dplot')
self.p_plot2d_type = self.m_cfg_parser.get(self.p_cfg_name,'2dplot')
self.p_2dvariables = self.m_cfg_parser.get('plot_variables','2dplots') # variables in 2-D plot
self.p_1dvariables = self.m_cfg_parser.get('plot_variables','1dplots') # variables in 1-D plot
self.eff_conditions = self.m_cfg_parser.get(self.p_cfg_name,'eff_y').split(',')
p_files = self.m_cfg_parser.get(self.p_cfg_name,'files')
p_files = open(p_files,'r').readlines()
## ------------------- ##
if self.p_plot1d_type == 'efficiency' and self.p_1dvariables:
self.file_type = 'root'
else:
self.file_type = 'json'
self.inputfilenames = ['data/'+fname.split('/')[-1].split('.')[0]+'_'+self.p_lepton+'_'+self.p_extra_save+'.'+self.file_type for fname in p_files]
if self.p_scalefactor:
if 'pb' in self.lumi['unit']:
lumi = "{0:.1f}".format(self.lumi['lumi']/1000.)
unit = 'fb'
elif 'fb' in self.lumi['unit']:
lumi = "{0:.1f}".format(self.lumi['lumi'])
unit = 'fb'
else:
lumi = "{0}".format(self.lumi['lumi'])
unit = self.lumi['unit'].strip('i')
self.plot_energy_lumi = self.lumi_label # plot the luminosity
self.lumi = {'lumi':lumi,'unit':unit} # reset for use later
else:
self.plot_energy_lumi = self.PMA_energy_label # plot only sqrt(s)
## - 1D plotting method (use function-pointer syntax to avoid if/else statements)
self.p_plot1d = [self.PMA_errorbar,self.PMA_hist_1d][(self.p_plot1d_type.startswith('hist'))]
## - 2D plotting method (use function-pointer syntax to avoid if/else statements)
self.p_plot2d = [self.PMA_scatter, self.PMA_hist_2d][(self.p_plot2d_type=='hist')]
return
def __init__(self,cfg_parser):
"""
Initialize the parameters from the config file.
@param cfg_parser Object that parsed the configuration file
"""
logging.getLogger('share/systematics.log')
loggingLEVEL = logging.getLogger().getEffectiveLevel() # DEBUG, INFO, ERROR, etc.
logging.info("")
logging.critical(" -- In tree2hist.py")
logging.info(" ------------ ")
logging.info(" Initializing the config file.")
self.GeV = 1000.
## -- Configuration -- ##
self.p_varList = cfg_parser.get('systematics','vars') # ex. share/varNames.txt
self.p_rootfiles = cfg_parser.get('systematics','inputfile') # ex. share/systematics_ntuples.txt
self.p_selection = cfg_parser.get('systematics','selection') # ex. pre2pre
self.p_outputname = cfg_parser.get('systematics','outputname') # ex. pre
self.p_makejsons = config.str2bool(cfg_parser.get('systematics','make_jsons')) # ex. True
self.p_makehists = config.str2bool(cfg_parser.get('systematics','make_hists')) # ex. True
self.p_lepton = cfg_parser.get('systematics','lepton') # ex. muel
self.p_nEvents = cfg_parser.get('systematics','nevents') # ex. -1
## ------------------- ##
self.p_varList = open(self.p_varList,'r').readlines() # update the variable
self.p_varList = [v.rstrip('\n') for v in self.p_varList]
if not self.p_makejsons and not self.p_makehists:
print
print " You didn't specify outputs (either json, hist, or both)"
print " for the systematics output."
print
sys.exit(1)
self.json_path = info.getJsonPath()+self.p_lepton
self.hist_path = self.json_path.split('json')[0]+'hists/'+self.p_lepton
## Run the program
self.main()
return
def getConfig(cfg_parser):
"""
Determine the configuration option from the config file.
Only keeping track of 'miniSL', 'datamc', and 'systematics' by hand --
need to manually insert the configuration options.
"""
config_options = {'miniSL':False,\
'datamc':False,\
'systematics':False}
for key in config_options.keys():
config_options[key] = config.str2bool(cfg_parser.get('configuration',key.lower()))
# Now safe-guard against incorrect things in the config file
# Note: the 'str2bool' function in config.py returns '-1' if
# the value given is not related to True or False
config_booleans = []
true_configs = []
for k in config_options.keys():
config_booleans.append(config_options[k]) # all options
if config_options[k]:
true_configs.append(k) # only true options
## Make sure the configurations make sense
if any(config_booleans) == -1:
print
print " Please choose one of the configurations: "
print " >> {0}".format(config_options.keys())
print " under the [configuration] section in 'share/PyMiniAna.cfg'."
print " with either 'True' or 'False'."
print
usage()
sys.exit(1)
## Make sure at least one of them is chosen
if not any(config_booleans):
print
print " Please choose one of the configurations: "
print " >> {0}".format(config_options.keys())
print " under the [configuration] section in 'share/PyMiniAna.cfg'."
print
usage()
sys.exit(1)
## Make sure only one of them is chosen
if len(true_configs)>1:
print
print " Please choose only one of the configurations: "
print " >> {0}".format(config_options.keys())
print " under the [configuration] section in 'share/PyMiniAna.cfg'."
print
usage()
sys.exit(1)
return true_configs[0]
def initialize(self):
"""Initialize the class by setting up parameters from the config file"""
## -- Configuration -- ##
self.p_lepton = self.m_cfg_parser.get('datamc','lepton') # ex. muel
self.p_plot_signal = self.m_cfg_parser.get('datamc','plot_signal') # ex. -1; if not all, plot only the one we are sensitive to?
self.p_signal_mag = float(self.m_cfg_parser.get('datamc','signal_mag')) # ex. 10 (increase signal by factor of 10 on plot)
self.p_stack_signal = str2bool(self.m_cfg_parser.get('datamc','stack_signal')) # ex. True
self.p_blind = str2bool(self.m_cfg_parser.get('datamc','blind')) # ex. True
self.p_logplot = str2bool(self.m_cfg_parser.get('datamc','logplot')) # ex. False
self.p_systematics = str2bool(self.m_cfg_parser.get('datamc','systematics')) # ex. True
self.p_jsonfilename = self.m_cfg_parser.get('datamc','jsonfilename') # ex. 'elLHMedium_pre_1tagin'
self.p_ana_status = self.m_cfg_parser.get('datamc','ana_status') # ex. "Internal", "Preliminary", or "" (final)
self.p_merged_hists = str2bool(self.m_cfg_parser.get('datamc','merged_hists')) # ex. True
self.p_extra_saveAs = self.m_cfg_parser.get('datamc','extra_saveAs') # ex. "1btag"
self.p_new_lumi = self.m_cfg_parser.get('datamc','new_lumi') # ex. 1.307 ifb
self.p_underflow = str2bool(self.m_cfg_parser.get('datamc','underflow')) # ex. True
self.p_overflow = str2bool(self.m_cfg_parser.get('datamc','overflow')) # ex. True
self.p_format = self.m_cfg_parser.get('datamc','format') # ex. pdf
## ------------------- ##
## ##
## ---- Set up some things that are universal to all plots ---- ##
## ##
## -- Setting up the output file
self.pathSave = info.getFigurePath()+"{0}/".format(self.p_lepton)
if not os.path.isdir(self.pathSave):
os.makedirs(self.pathSave)
if self.p_extra_saveAs: # want something else in the saved filename
self.p_extra_saveAs = '_'+self.p_extra_saveAs
self.p_hatch_args = {'hatch':'',\
'color':'none',\
'edgecolor':'k',\
'linewidth':0}
if self.p_format=='png':
self.p_hatch_args['hatch'] = '/'
elif self.p_format=='eps':
self.p_hatch_args['hatch'] = None
self.p_hatch_args['color'] = '#C1E1C0'
self.p_hatch_args['edgecolor'] = '#C1E1C0'
else:
self.p_hatch_args['hatch'] = '/////'
self.plot_order = ['ttbarV','diboson','singletop',\
'zlight','zbbcc','zc','zjets',\
'wlight','wbbcc','wc','wjets',\
'ttbar'] # plot order for background samples (stacked)
return
def initialize(self):
## -- Configuration -- ##
self.p_lepton = self.m_cfg_parser.get(self.p_cfg_name,'lepton') # ex. muel
self.p_logplot = str2bool(self.m_cfg_parser.get(self.p_cfg_name,'logplot')) # ex. False
self.p_ana_status = self.m_cfg_parser.get(self.p_cfg_name,'ana_status') # ex. "Internal"
self.p_extra_save = self.m_cfg_parser.get(self.p_cfg_name,'extra_saveAs')
self.p_scalefactor = str2bool(self.m_cfg_parser.get(self.p_cfg_name,'eventweights'))
self.p_plot1d_type = self.m_cfg_parser.get(self.p_cfg_name,'1dplot')
self.p_plot2d_type = self.m_cfg_parser.get(self.p_cfg_name,'2dplot')
self.p_2dvariables = self.m_cfg_parser.get('plot_variables','2dplots')
self.p_1dvariables = self.m_cfg_parser.get('plot_variables','1dplots')
p_files = self.m_cfg_parser.get(self.p_cfg_name,'files')
p_files = open(p_files,'r').readlines()
## ------------------- ##
self.inputfilenames = ['data/'+fname.split('/')[-1].split('.')[0]+'_'+self.p_lepton+'_'+self.p_extra_save+'.json' for fname in p_files]
if self.p_scalefactor:
if 'pb' in self.lumi['unit']:
lumi = "{0:.1f}".format(self.lumi['lumi']/1000.)
unit = 'fb'
elif 'fb' in self.lumi['unit']:
lumi = "{0:.1f}".format(self.lumi['lumi'])
unit = 'fb'
else:
lumi = "{0}".format(self.lumi['lumi'])
unit = self.lumi['unit'].strip('i')
self.plot_energy_lumi = self.lumi_label
self.lumi = {'lumi':lumi,'unit':unit} # reset for use later
else:
self.plot_energy_lumi = self.PMA_energy_label
## - 1D plotting method (use this syntax to avoid if/else statements)
self.p_plot1d = self.PMA_errorbar
## - 2D plotting method (use this syntax to avoid if/else statements)
self.p_plot2d = self.PMA_hist_2d
def execute(self,parser,treename=''):
"""
Main function in the miniSL class -- does all the reading/writing
of the flat ntuples.
Just added the systematics functionality, so we've made this
a hybrid class that can run over the nominal tree and
trees for each systematics.
@param treename The treename that we need to access for data
This is now an option because of the systematics
"""
self.timeStamp = strftime("%d%b%Y",localtime())
cfg_name = 'miniSL' if not treename else 'systematics'
## -- Configuration -- ##
p_sel_script = parser.get(cfg_name,'selection') # ex. pyMiniSL/SelectionBase.py
p_nEvents = int(parser.get(cfg_name,'nEvents')) # ex. -1
p_firstEvent = int(parser.get(cfg_name,'firstevent')) # ex. 0
p_cutsfile = parser.get(cfg_name,'cutsfile') # ex. share/pre2loose_cuts.txt
p_inputfile = parser.get(cfg_name,'inputfile') # ex. share/miniSL_ntuples.txt
p_outputname = parser.get(cfg_name,'outputname') # ex. loose
p_isLoose = config.str2bool(parser.get(cfg_name,'isLoose')) # ex. False
## ------------------- ##
files = open(p_inputfile,'r').readlines()
sel_path,selection = p_sel_script.split('.py')[0].split('/')
self.selection = selection # for access later
selClassName = selection[0].upper()+selection[1:]
self.LUMI = info.LUMI()
signalIDs = [i for key in info.dsids()['signal'].keys() for i in info.dsids()['signal'][key]] # TTS and XX signals
ttbarIDs = [i for i in info.dsids()['background']['ttbar'].keys()]
# Dynamic import the file without needing any hard-coding
# Assumes that the class is the same as the filename just capitalized!
pyEventFile = importlib.import_module(sel_path+"."+selection)
pyEventSel = getattr(pyEventFile,selClassName)
loggingLEVEL = logging.getLogger().getEffectiveLevel()
logging.info(" -- In file {0}".format(os.path.basename(__file__)))
logging.info(" -- Files: ")
for i in files:
logging.info(" "+i.rstrip('\n'))
## -- Load the selection script
evtSel = pyEventSel(parser) # imported using 'importlib' above
evtSel.initialize(cfg_name) # setup some of the configurations
# ----------------------------------------- #
# File Loop #
# ----------------------------------------- #
total_num_files = len(files)
for ff,file in enumerate(files):
## -- Open the current file
f = ROOT.TFile.Open("{0}".format(file.strip()))
if not f or f.IsZombie():
print
print " FILE {0} DOES NOT EXIST".format(f.GetName())
print " Continuing to next file in the loop. "
print
continue
## -- Load the TTree
# For nominal, treename is an empty string
# For systematics, pre2preSelection: it's the treename for a systematic
# SelectionBase: it's a dummy string
if cfg_name != 'miniSL':
if treename == 'systs':
treename = f.GetListOfKeys()[0].GetName() # only 1 tree and it's the systematic
systname = '_'+treename
new_treename = treename
else:
if p_isLoose:
treename = 'nominal_Loose'
else:
treename = info.treename() # nominal
new_treename = 'nominal'
systname = ''
t = f.Get(treename)
try:
t.GetEntry(0) # to get the mcChannelNumber
except AttributeError:
print " File {0} does not have the GetEntry() attribute.".format(f.GetName())
print " Skipping and going to the next file.\n"
continue
## New File Stuff (the one to which we're writing)
name,isMC = self.getName(t)
newfilename = "data/{0}{1}_{2}.root".format(name,systname,p_outputname)
newfile = ROOT.TFile(newfilename,"RECREATE")
self.newtree = ROOT.TTree(new_treename,"Event Information")
## Metadata
self.setMetaData(p_outputname)
## Log and print to the console before doing anything
print " ++ Running over file ({0}/{1}): {2}".format(ff+1,total_num_files,file.strip())
print " ++ Producing new file: {0}{1}_{2}.root".format(name,systname,p_outputname)
logging.info(" ----")
logging.info(" {0} Selection".format(p_outputname.title()))
logging.info(" File: {0}".format(file.strip()))
logging.info(" {0}".format(self.timeStamp))
logging.info(" ---- \n")
# ----------------------------------------- #
# Initialize the branches #
# ----------------------------------------- #
self.initBranches()
# ----------------------------------------- #
# Event Loop #
# ----------------------------------------- #
maxEntries = t.GetEntries()
if p_nEvents < 0 or p_nEvents > maxEntries:
nEvents = maxEntries
logging.info(" Selected nEvents = {0}; running {1}".format(nEvents,maxEntries))
else:
nEvents = p_nEvents
## -- While-loop (no list generation with for loop)
evtSel.execute(t,f) # setup the details for this particular file in selection
entry = p_firstEvent # in case the user has split the file into pieces
while entry < nEvents:
#### -- APPLY THE SELECTION -- ####
results = evtSel.event_loop(entry)
#### ------------------------- ####
if not results['result']:
logging.info(" ---- ")
logging.info(" miniSL::Event {0} Failed Selection ".format(entry))
logging.info(" ---- ")
entry+=1 # iterate entry to go to the next one
continue
self.saveEvent(results['objects'])
# ----------------------------------------- #
# Clear the ROOT Vectors #
# ----------------------------------------- #
self.clearVectors()
## increment the while loop
entry+=1
# ----------------------------------------- #
# Write the new file #
# ----------------------------------------- #
newfile.Write()
print
print " New output file: {0}".format(newfilename)
print " Ended at: {0}".format(strftime("%c",localtime()))
print
newfile.Close()
return
def main(self,parser):
"""
Main function in the datamc class -- does all the directing for
reading/writing of data from ROOT to json files, and plotting histograms.
"""
## -- Configuration -- ##
self.p_vars_file = parser.get('datamc','vars') # ex. share/varNames.txt
self.p_jsonoutput = str2bool(parser.get('datamc','makejsonfile')) # ex. False
self.p_plotoutput = str2bool(parser.get('datamc','makeplot')) # ex. True
self.p_mergejson = str2bool(parser.get('datamc','mergejson')) # ex. True
self.p_outfile = parser.get('datamc','jsonfilename') # ex. pre
self.p_lepton = parser.get('datamc','lepton') # ex. muel
## ------------------- ##
## -- Sanity check -- ##
if not any([self.p_plotoutput,self.p_jsonoutput,self.p_mergejson]):
print
print " You have specified that you don't want to "
print " make json outputs, don't want to make "
print " plots, and don't want to merge json files. "
print " There's nothing left to do. "
print " Exiting. "
print
sys.exit(1)
## ------------------ ##
self.p_varlist = info.read_txt_file(self.p_vars_file) # variables from text file
# for variables that may have [N], e.g., jet_pt[0], make a list
# that just contains the name, e.g., jet_pt.
# Plan to is to make a single json file for jet_pt, but
# only plot [N], or [N+1] (e.g., the user specifies both jet_pt[0] and jet_pt[1]
# in the text file)
self.p_varlist_nolead = list(set([p_var.split('[')[0] for p_var in self.p_varlist]))
loggingLEVEL = logging.getLogger().getEffectiveLevel()
logging.info(" -- In file dataMC.py")
logging.info(" -- Make json output: {0} ".format(self.p_jsonoutput))
## -- Conver ROOT to JSON -- ##
if self.p_jsonoutput:
import pyDataMC.root2json as root2json
logging.info(" > Specified json output ")
logging.info(" Will produce json files and then plots automatically ")
logging.info(" -- Making json output")
print "\n -- Converting ROOT to json output -- \n"
## Making json output, and then making plots (one step is easier...)
root2json.ROOT2json(parser)
## -- Merge json files -- ##
if self.p_mergejson:
## merge before plotting!!
self.merge_json_files()
## -- Plot Histograms -- ##
if self.p_plotoutput:
from pyDataMC.json2hist import DataMCPlotter
logging.info(" -- Making plots from json output")
print "\n -- Producing figures -- \n"
plotter = DataMCPlotter(parser)
plotter.initialize()
for var in self.p_varlist:
print " ++ Plotting {0} ++\n".format(var)
logging.info(" ++ Plotting {0} ++".format(var))
plotter.datamcplotter(var)
logging.info(" Finished datamc.py class DataMC")
return
