| 1 |
|
#! /usr/bin/env python |
| 2 |
– |
import os,shutil,sys,pickle,subprocess,ROOT |
| 2 |
|
from optparse import OptionParser |
| 3 |
< |
from BetterConfigParser import BetterConfigParser |
| 4 |
< |
from samplesclass import sample |
| 5 |
< |
import getpass |
| 3 |
> |
import sys |
| 4 |
> |
import time |
| 5 |
> |
import os |
| 6 |
> |
import shutil |
| 7 |
|
|
| 8 |
|
parser = OptionParser() |
| 9 |
|
parser.add_option("-T", "--tag", dest="tag", default="", |
| 14 |
|
help="Mass for DC or Plots, 110...135") |
| 15 |
|
parser.add_option("-S","--samples",dest="samples",default="", |
| 16 |
|
help="samples you want to run on") |
| 17 |
< |
|
| 17 |
> |
parser.add_option("-F", "--folderTag", dest="ftag", default="", |
| 18 |
> |
help="Creats a new folder structure for outputs or uses an existing one with the given name") |
| 19 |
> |
parser.add_option("-N", "--number-of-events", dest="nevents_split", default=100000, |
| 20 |
> |
help="Number of events per file when splitting.") |
| 21 |
> |
parser.add_option("-P", "--philipp-love-progress-bars", dest="philipp_love_progress_bars", default=False, |
| 22 |
> |
help="If you share the love of Philipp...") |
| 23 |
|
|
| 24 |
|
(opts, args) = parser.parse_args(sys.argv) |
| 25 |
+ |
|
| 26 |
+ |
import os,shutil,pickle,subprocess,ROOT,re |
| 27 |
+ |
ROOT.gROOT.SetBatch(True) |
| 28 |
+ |
from myutils import BetterConfigParser, Sample, ParseInfo, sample_parser |
| 29 |
+ |
import getpass |
| 30 |
+ |
|
| 31 |
|
if opts.tag == "": |
| 32 |
|
print "Please provide tag to run the analysis with, example '-T 8TeV' uses config8TeV and pathConfig8TeV to run the analysis." |
| 33 |
|
sys.exit(123) |
| 36 |
|
print "Please provide a task.\n-J prep:\tpreparation of Trees\n-J sys:\t\twrite regression and systematics\n-J eval:\tcreate MVA output\n-J plot:\tproduce Plots\n-J dc:\t\twrite workspaces and datacards" |
| 37 |
|
sys.exit(123) |
| 38 |
|
|
| 39 |
+ |
|
| 40 |
+ |
en = opts.tag |
| 41 |
+ |
|
| 42 |
|
#create the list with the samples to run over |
| 43 |
|
samplesList=opts.samples.split(",") |
| 44 |
|
|
| 45 |
< |
en = opts.tag |
| 46 |
< |
configs = ['config%s'%(en),'pathConfig%s'%(en)] |
| 47 |
< |
|
| 45 |
> |
timestamp = time.asctime().replace(' ','_').replace(':','-') |
| 46 |
> |
|
| 47 |
> |
configs = ['%sconfig/general'%(en),'%sconfig/paths'%(en),'%sconfig/plots'%(en),'%sconfig/training'%(en),'%sconfig/datacards'%(en),'%sconfig/cuts'%(en)] |
| 48 |
> |
|
| 49 |
> |
pathconfig = BetterConfigParser() |
| 50 |
> |
pathconfig.read('%sconfig/paths'%(en)) |
| 51 |
> |
|
| 52 |
> |
if not opts.ftag == '': |
| 53 |
> |
tagDir = pathconfig.get('Directories','tagDir') |
| 54 |
> |
DirStruct={'tagDir':tagDir,'ftagdir':'%s/%s/'%(tagDir,opts.ftag),'logpath':'%s/%s/%s/'%(tagDir,opts.ftag,'Logs'),'plotpath':'%s/%s/%s/'%(tagDir,opts.ftag,'Plots'),'limitpath':'%s/%s/%s/'%(tagDir,opts.ftag,'Limits'),'confpath':'%s/%s/%s/'%(tagDir,opts.ftag,'config') } |
| 55 |
> |
|
| 56 |
> |
for keys in ['tagDir','ftagdir','logpath','plotpath','limitpath','confpath']: |
| 57 |
> |
try: |
| 58 |
> |
os.stat(DirStruct[keys]) |
| 59 |
> |
except: |
| 60 |
> |
os.mkdir(DirStruct[keys]) |
| 61 |
> |
|
| 62 |
> |
pathfile = open('%sconfig/paths'%(en)) |
| 63 |
> |
buffer = pathfile.readlines() |
| 64 |
> |
pathfile.close() |
| 65 |
> |
os.rename('%sconfig/paths'%(en),'%sconfig/paths.bkp'%(en)) |
| 66 |
> |
pathfile = open('%sconfig/paths'%(en),'w') |
| 67 |
> |
for line in buffer: |
| 68 |
> |
if line.startswith('plotpath'): |
| 69 |
> |
line = 'plotpath: %s\n'%DirStruct['plotpath'] |
| 70 |
> |
elif line.startswith('logpath'): |
| 71 |
> |
line = 'logpath: %s\n'%DirStruct['logpath'] |
| 72 |
> |
elif line.startswith('limits'): |
| 73 |
> |
line = 'limits: %s\n'%DirStruct['limitpath'] |
| 74 |
> |
pathfile.write(line) |
| 75 |
> |
pathfile.close() |
| 76 |
> |
|
| 77 |
> |
#copy config files |
| 78 |
> |
for item in configs: |
| 79 |
> |
shutil.copyfile(item,'%s/%s/%s'%(tagDir,opts.ftag,item.strip(en))) |
| 80 |
> |
|
| 81 |
> |
|
| 82 |
|
print configs |
| 83 |
|
config = BetterConfigParser() |
| 84 |
|
config.read(configs) |
| 85 |
+ |
|
| 86 |
|
btagLibrary = config.get('BTagReshaping','library') |
| 87 |
|
submitDir = os.getcwd() |
| 88 |
|
os.chdir(os.path.dirname(btagLibrary)) |
| 92 |
|
shutil.copyfile('/scratch/%s/%s'%(getpass.getuser(),os.path.basename(btagLibrary)),btagLibrary) |
| 93 |
|
os.chdir(submitDir) |
| 94 |
|
logPath = config.get("Directories","logpath") |
| 95 |
+ |
logo = open('%s/data/submit.txt' %config.get('Directories','vhbbpath')).readlines() |
| 96 |
+ |
counter = 0 |
| 97 |
+ |
|
| 98 |
|
#check if the logPath exist. If not exit |
| 99 |
|
if( not os.path.isdir(logPath) ): |
| 100 |
< |
print 'ERROR: ' + logPath + ': dir not found.' |
| 101 |
< |
print 'ERROR: Create it before submitting ' |
| 102 |
< |
print 'Exit' |
| 103 |
< |
sys.exit(-1) |
| 100 |
> |
print 'ERROR: ' + logPath + ': dir not found.' |
| 101 |
> |
print 'ERROR: Create it before submitting ' |
| 102 |
> |
print 'Exit' |
| 103 |
> |
sys.exit(-1) |
| 104 |
> |
|
| 105 |
|
|
| 106 |
< |
repDict = {'en':en,'logpath':logPath,'job':'','task':opts.task,'queue': 'all.q'} |
| 106 |
> |
repDict = {'en':en,'logpath':logPath,'job':'','task':opts.task,'queue': 'all.q','timestamp':timestamp,'additional':'','job_id':''} |
| 107 |
|
def submit(job,repDict): |
| 108 |
< |
repDict['job'] = job |
| 109 |
< |
command = 'qsub -V -cwd -q %(queue)s -N %(job)s_%(en)s%(task)s -o %(logpath)s/%(job)s_%(en)s_%(task)s.out -e %(logpath)s/%(job)s_%(en)s_%(task)s.err runAll.sh %(job)s %(en)s ' %(repDict) + opts.task |
| 110 |
< |
print command |
| 111 |
< |
subprocess.call([command], shell=True) |
| 108 |
> |
global counter |
| 109 |
> |
repDict['job'] = job |
| 110 |
> |
nJob = counter % len(logo) |
| 111 |
> |
counter += 1 |
| 112 |
> |
if opts.philipp_love_progress_bars: |
| 113 |
> |
repDict['name'] = '"%s"' %logo[nJob].strip() |
| 114 |
> |
else: |
| 115 |
> |
repDict['name'] = '%(job)s_%(en)s%(task)s' %repDict |
| 116 |
> |
command = 'qsub -V -cwd -q %(queue)s -l h_vmem=6G -N %(name)s -o %(logpath)s/%(timestamp)s_%(job)s_%(en)s_%(task)s.out -e %(logpath)s/%(timestamp)s_%(job)s_%(en)s_%(task)s.err runAll.sh %(job)s %(en)s ' %(repDict) + opts.task + ' ' + repDict['job_id'] + ' ' + repDict['additional'] |
| 117 |
> |
print command |
| 118 |
> |
subprocess.call([command], shell=True) |
| 119 |
> |
|
| 120 |
> |
if opts.task == 'train': |
| 121 |
> |
train_list = (config.get('MVALists','List_for_submitscript')).split(',') |
| 122 |
> |
print train_list |
| 123 |
> |
for item in train_list: |
| 124 |
> |
submit(item,repDict) |
| 125 |
> |
|
| 126 |
|
|
| 127 |
|
if opts.task == 'dc': |
| 128 |
|
#DC_vars = config.items('Limit') |
| 135 |
|
if not opts.task == 'prep': |
| 136 |
|
path = config.get("Directories","samplepath") |
| 137 |
|
samplesinfo = config.get("Directories","samplesinfo") |
| 138 |
< |
infofile = open(samplesinfo,'r') |
| 72 |
< |
info = pickle.load(infofile) |
| 73 |
< |
infofile.close() |
| 74 |
< |
|
| 138 |
> |
info = ParseInfo(samplesinfo,path) |
| 139 |
|
|
| 140 |
|
if opts.task == 'plot': |
| 141 |
< |
repDict['queue'] = 'short.q' |
| 141 |
> |
repDict['queue'] = 'all.q' |
| 142 |
|
for item in Plot_vars: |
| 143 |
|
submit(item,repDict) |
| 144 |
|
|
| 145 |
+ |
|
| 146 |
|
elif opts.task == 'dc': |
| 147 |
|
repDict['queue'] = 'short.q' |
| 148 |
|
for item in DC_vars: |
| 152 |
|
submit(item,repDict) |
| 153 |
|
|
| 154 |
|
elif opts.task == 'prep': |
| 155 |
< |
submit('prepare',repDict) |
| 155 |
> |
if ( opts.samples == ""): |
| 156 |
> |
path = config.get("Directories","PREPin") |
| 157 |
> |
samplesinfo = config.get("Directories","samplesinfo") |
| 158 |
> |
info = ParseInfo(samplesinfo,path) |
| 159 |
> |
for job in info: |
| 160 |
> |
submit(job.name,repDict) |
| 161 |
|
|
| 162 |
< |
elif opts.task == 'eval' or opts.task == 'sys' or opts.task == 'syseval': |
| 162 |
> |
else: |
| 163 |
> |
for sample in samplesList: |
| 164 |
> |
submit(sample,repDict) |
| 165 |
> |
elif opts.task == 'sys' or opts.task == 'syseval': |
| 166 |
> |
path = config.get("Directories","SYSin") |
| 167 |
> |
samplesinfo = config.get("Directories","samplesinfo") |
| 168 |
> |
info = ParseInfo(samplesinfo,path) |
| 169 |
|
if ( opts.samples == ""): |
| 170 |
|
for job in info: |
| 171 |
+ |
if (job.subsample): |
| 172 |
+ |
continue #avoid multiple submissions form subsamples |
| 173 |
+ |
# TO FIX FOR SPLITTED SAMPLE |
| 174 |
|
submit(job.name,repDict) |
| 175 |
|
else: |
| 176 |
|
for sample in samplesList: |
| 177 |
|
submit(sample,repDict) |
| 178 |
|
|
| 179 |
< |
os.system('qstat') |
| 179 |
> |
elif opts.task == 'eval': |
| 180 |
> |
path = config.get("Directories","MVAin") |
| 181 |
> |
samplesinfo = config.get("Directories","samplesinfo") |
| 182 |
> |
info = ParseInfo(samplesinfo,path) |
| 183 |
> |
if ( opts.samples == ""): |
| 184 |
> |
for job in info: |
| 185 |
> |
if (job.subsample): |
| 186 |
> |
continue #avoid multiple submissions from subsamples |
| 187 |
> |
if(info.checkSplittedSampleName(job.identifier)): # if multiple entries for one name (splitted samples) use the identifier to submit |
| 188 |
> |
print '@INFO: Splitted samples: submit through identifier' |
| 189 |
> |
submit(job.identifier,repDict) |
| 190 |
> |
else: submit(job.name,repDict) |
| 191 |
> |
else: |
| 192 |
> |
for sample in samplesList: |
| 193 |
> |
submit(sample,repDict) |
| 194 |
> |
|
| 195 |
> |
|
| 196 |
> |
elif( opts.task == 'split' ): |
| 197 |
> |
path = config.get("Directories","SPLITin") |
| 198 |
> |
samplesinfo = config.get("Directories","samplesinfo") |
| 199 |
> |
repDict['job_id']=opts.nevents_split |
| 200 |
> |
info = ParseInfo(samplesinfo,path) |
| 201 |
> |
if ( opts.samples == "" ): |
| 202 |
> |
for job in info: |
| 203 |
> |
if (job.subsample): continue #avoid multiple submissions from subsamples |
| 204 |
> |
submit(job.name,repDict) |
| 205 |
> |
else: |
| 206 |
> |
for sample in samplesList: |
| 207 |
> |
submit(sample,repDict) |
| 208 |
> |
|
| 209 |
> |
#BDT optimisation |
| 210 |
> |
elif opts.task == 'mva_opt': |
| 211 |
> |
total_number_of_steps=1 |
| 212 |
> |
setting = '' |
| 213 |
> |
for par in (config.get('Optimisation','parameters').split(',')): |
| 214 |
> |
scan_par=eval(config.get('Optimisation',par)) |
| 215 |
> |
setting+=par+'='+str(scan_par[0])+':' |
| 216 |
> |
if len(scan_par) > 1 and scan_par[2] != 0: |
| 217 |
> |
total_number_of_steps+=scan_par[2] |
| 218 |
> |
setting=setting[:-1] # eliminate last column at the end of the setting string |
| 219 |
> |
print setting |
| 220 |
> |
repDict['additional']=setting |
| 221 |
> |
repDict['job_id']=config.get('Optimisation','training') |
| 222 |
> |
submit('OPT_main_set',repDict) |
| 223 |
> |
main_setting=setting |
| 224 |
> |
|
| 225 |
> |
#Scanning all the parameters found in the training config in the Optimisation sector |
| 226 |
> |
for par in (config.get('Optimisation','parameters').split(',')): |
| 227 |
> |
scan_par=eval(config.get('Optimisation',par)) |
| 228 |
> |
print par |
| 229 |
> |
if len(scan_par) > 1 and scan_par[2] != 0: |
| 230 |
> |
for step in range(scan_par[2]): |
| 231 |
> |
value = (scan_par[0])+((1+step)*(scan_par[1]-scan_par[0])/scan_par[2]) |
| 232 |
> |
print value |
| 233 |
> |
setting=re.sub(par+'.*?:',par+'='+str(value)+':',main_setting) |
| 234 |
> |
repDict['additional']=setting |
| 235 |
> |
# repDict['job_id']=config.get('Optimisation','training') |
| 236 |
> |
submit('OPT_'+par+str(value),repDict) |
| 237 |
> |
# submit(config.get('Optimisation','training'),repDict) |
| 238 |
> |
print setting |
| 239 |
> |
|
| 240 |
> |
|
| 241 |
> |
os.system('qstat') |
| 242 |
> |
if (opts.philipp_love_progress_bars): |
| 243 |
> |
os.system('./qstat.py') |
