def merge_files_by_process(root_files):
global input_folder, output_folder
electron_qcd_samples = [ 'QCD_Pt-20to30_BCtoE',
'QCD_Pt-30to80_BCtoE',
'QCD_Pt-80to170_BCtoE',
'QCD_Pt-20to30_EMEnriched',
'QCD_Pt-30to80_EMEnriched',
'QCD_Pt-80to170_EMEnriched',
'GJets_HT-40To100',
'GJets_HT-100To200',
'GJets_HT-200']
singleTop_samples = [ 'T_tW-channel',
'T_t-channel',
'T_s-channel',
'Tbar_tW-channel',
'Tbar_t-channel',
'Tbar_s-channel']
wplusjets_samples = [ 'W1Jet', 'W2Jets', 'W3Jets', 'W4Jets']
vplusjets_samples = wplusjets_samples
vplusjets_samples.append('DYJetsToLL')
diboson_samples = [ 'WWtoAnything', 'WZtoAnything', 'ZZtoAnything']
signal_samples = [ 'TTJet', 'SingleTop']
summations = {
'QCD_Electron':electron_qcd_samples,
'SingleTop' : singleTop_samples,
'WPlusJets' : wplusjets_samples,
'VPlusJets' : vplusjets_samples,
'DiBoson': diboson_samples,
'Signal': signal_samples
}
summation_files = {}
file_template = ''
template_token = '<temp>'
for summation, samples in summations.iteritems():
summation_files[summation] = []
for file_in_path in root_files:
process_name = get_process_from_file(file_in_path)
if not file_template:
file_template = file_in_path.replace(process_name, template_token)
file_template = file_template.replace(input_folder, output_folder)
if process_name in samples:
summation_files[summation].append(file_in_path)
for summation, files in summation_files.iteritems():
output_file = file_template.replace(template_token, summation)
merge_ROOT_files(files, output_file)
if not sample in ['WJets', 'DYJets', 'VJets-matchingup',
'VJets-matchingdown', 'VJets-scaleup',
'VJets-scaledown']: #
continue
print "Merging"
current_working_directory = os.getcwd() #find current working directory
output_file = config_8TeV.central_general_template % sample
output_file = output_file.replace("/hdfs/TopQuarkGroup/results/histogramfiles", current_working_directory)
input_files = [config_8TeV.central_general_template % input_sample for input_sample in input_samples]
print output_file
for input_file in input_files:
print input_file
if not os.path.exists( output_file ):
merge_ROOT_files( input_files, output_file, compression = 7 )
print "merging ", sample
new_files.append( output_file )
print '=' * 120
# if 8 concurrent processes, wait until they are finished before starting the next set to avoid overloading the machine
while ( int( subprocess.check_output( "ps ax | grep 'hadd' | wc -l", shell = True ) ) - 2 ) >= 8:
time.sleep( 30 ) # sleep for 30 seconds
# merge all other histogram files
for category in config_8TeV.categories_and_prefixes.keys():
for sample, input_samples in sample_summations.iteritems():
if not sample in ['QCD_Electron', 'QCD_Muon', 'VJets',
'SingleTop']: #
continue
print "Merging"
# Make folder
make_folder_if_not_exists(path_to_AN_folder + "/" + category)
current_working_directory = os.getcwd() #find current working directory
output_file_hdfs = config.general_category_templates[category] % sample
output_file = output_file_hdfs.replace(
"/hdfs/TopQuarkGroup/results/histogramfiles", current_working_directory)
input_files = [
config.general_category_templates[category] % input_sample
for input_sample in input_samples
]
if not os.path.exists(output_file):
merge_ROOT_files(input_files,
output_file,
compression=7,
waitToFinish=True)
print "merging ", sample
else:
print 'Not merging ', sample, 'as', output_file, 'already exists'
# Now move output file to hdfs
# Check if file already exists on hdfs
if os.path.exists(output_file_hdfs):
print "Output file on hdfs already exists. Removing and replacing with new version."
command = 'hadoop fs -rm -skipTrash ' + output_file_hdfs.split('/hdfs')[-1]
p = subprocess.Popen(command, shell=True)
p.wait()
print '\nStarting rsync'
output_log_file = output_file.replace(".root", ".log")
current_working_directory = os.getcwd() #find current working directory
output_file = config_7TeV.central_general_template % sample
output_file = output_file.replace(
"/hdfs/TopQuarkGroup/results/histogramfiles",
current_working_directory)
input_files = [
config_7TeV.central_general_template % input_sample
for input_sample in input_samples
]
print output_file
for input_file in input_files:
print input_file
if not os.path.exists(output_file):
merge_ROOT_files(input_files, output_file, compression=7)
print "merging ", sample
new_files.append(output_file)
print '=' * 120
# if 8 concurrent processes, wait until they are finished before starting the next set to avoid overloading the machine
while (int(
subprocess.check_output("ps ax | grep 'hadd' | wc -l", shell=True))
- 2) >= 8:
time.sleep(30) # sleep for 30 seconds
# merge all other histogram files
for category in config_7TeV.categories_and_prefixes.keys():
for sample, input_samples in sample_summations.iteritems():
if not sample in ['QCD_Electron', 'QCD_Muon', 'VJets', 'SingleTop']: #
continue
category = job[0]
sample = job[1]
input_samples = job[2]
# print 'Test with :',sample, category, input_samples
# Make folder
make_folder_if_not_exists( path_to_AN_folder + "/" + category)
current_working_directory = os.getcwd() #find current working directory
output_file_hdfs = config.general_category_templates[category] % sample
output_file = output_file_hdfs.replace("/hdfs/TopQuarkGroup/results/histogramfiles", current_working_directory)
input_files = [config.general_category_templates[category] % input_sample for input_sample in input_samples]
if not os.path.exists( output_file ):
merge_ROOT_files( input_files, output_file, compression = 7, waitToFinish=True )
print "merging ", sample
else :
print 'Not merging ',sample,'as',output_file,'already exists'
# Now move output file to hdfs
# Check if file already exists on hdfs
if os.path.exists( output_file_hdfs ):
print "Output file on hdfs already exists. Removing and replacing with new version."
command = 'hadoop fs -rm -skipTrash ' + output_file_hdfs.split('/hdfs')[-1]
p = subprocess.Popen(command, shell=True)
p.wait()
print '\nStarting rsync'
output_log_file = output_file.replace(".root", ".log")
command = 'rsync --verbose --progress --stats --compress --recursive --times --update %s %s >> %s' % (output_file,output_file_hdfs,output_log_file)
