Pixel Analysis using CRAB

<setup the Exe>

cd <release_dir/src/TrackerReco/PixelTrackVertexFinder/test/

source /opt/CRAB/crab.sh

I used the following config file:

[CRAB]
#
#   This section contains the default values for general parameters.
#   They can be set also as a command-line option, i.e.
#
#   key1 = value1
#   key2 = value2
#
#   in this section corresponds to
#
#   crab.py -key1=value1 -key2=value2
#
#   in the command line.
#
jobtype = orca_common
scheduler = condor_g
#use_boss = 0
 
 
[ORCA]
 
###############################
####### DATA TO ANALYZE #######
###############################
 
dataset = bt03_b_pt120-170
owner = bt_2x1033PU761_TkMu_g133_CMS
 
## CERN
#dataset = bt03_qcd120_170_tth
#owner = bt_2x1033PU761_TkMu_g133_CMS
 
## CNAF
#dataset = bt03_ttH120_6j1l
#owner = bt_DST871_2x1033PU_g133_CMS
 
## PIC
#dataset = hg03b_hww_2l_140
#owner = hg_DST871_2x1033PU_g133_CMS
 
## FNAL
# dataset = eg03_tt_2e2mu
# owner = eg_DST871_2x1033PU_g133_CMS
 
## LNL
#dataset = mu03_W1mu
#owner = mu_DST861_2_3_2x1033PU_g132_CMS
 
## BA
#dataset = bt03_tt_pureleptonic_etau
#owner = bt_DST871_2x1033PU_g133_CMS
 
 
### Data Tier to be accessed.
### Possible candidates are: DST, Digi, Hit (Comma separated list, mind the case!)
data_tier = Digi, Hit
#data_tier = Hit
 
### Order of the catalogs on .orcarc : do not change unless you know what you are doing!
#order_catalogs =
 
# switch to SC3 pubdb
#what = SC3
 
# use gt4 condor_g
#use_gt_4 = 1
 
 
###############################
####### ORCA EXECUTABLE #######
###############################
 
## User must provide the executable into his personal scram area
#executable = ExSimHitStatistics
#executabel = ExSimHitStatistics
executable = PixelTrackVertexFinderTest
 
## or the script that calls the executable...
#script_exe = /home/fanzago/CRAB/UserTools/src/orca_script.sh
 
## Output files produced by executable: comma separated list, can be empty but mut be present!
output_file = pvtest.root
 
## additional files to be put in InputSandBox: comma separated list
#additional_input_files =
 
###############################
########## ORCA CARD ##########
###############################
 
## orcarc card provided by user (if not in current directory, the full path
## must to be provided) NB the file must exist (could be empty)
orcarc_file = orcarc
 
#################################
######### CRAB  DIR  ############
#################################
 
## Name of UI directory where CRAB will create job to submit (with full path).
## If commented, the default directory will be "crab_0_data_time"
#ui_working_dir = /full_path/mu03_DY2mu
#ui_working_dir : /full_path/%(dataset)s
 
 
[USER]
 
#################################
#### JOB OUTPUT MANAGEMENT #####
#################################
 
### RETRIEVE ORCA OUTPUT INTO UI ###
## to have back the orca executable output into UI (return_data= 1)
return_data = 1
 
## First event to analyze: default is 0
#first_event = 0
 
## Total number of events to analyze: -1==all
#total_number_of_events = -1
total_number_of_events = -1
 
## Number of events for each job
job_number_of_events = 1000
 
### Total number of jobs to be created: incompatible with previous "job_number_of_events"
#total_number_of_jobs = 1000
 
 
### If return_data = 1 ###
## UI directory where to store the orca executable output
## FULL path is mandatory. If none /res will be used.
#outputdir=/home/fanzago/CRAB/Crab/python/out_orca
 
### If return_data = 1 ###
## UI directory where to store the stderr, stdout and .BrokerInfo of submitted jobs
## FULL path is mandatory. If none /res will be used.
#logdir=/home/fanzago/CRAB/UserTools/src/grid_job_log
 
### COPY ORCA OUTPUT INTO A SE ###
## if you want to copy the orca executable output into a SE (i:e castor)
### WARNING: if the copy fails and return_data = 0, the output is lost
#copy_data = 1
 
### if copy_data = 1 ###
## name of the SE where to copy the orca executable output.
#storage_element = castorgrid.cern.ch
## and the SE directory (or the mountpoint) that has to be writable from all
#storage_path = /castor/cern.ch/user/u/user
 
### REGISTER ORCA OUTPUT IN THE RLS CATALOG ###
## if you want also to register the orca executable output into the RLS catalog
## WARNING: to use with copy_data = 1
#register_data = 1
 
### if register_data = 1
## If you register the orca output file into the RLS catalog, this is the first part of LFN
### example LFN="lfn_dir"/"output_file"
#lfn_dir = MyDirLFN
 
#################################
####### JOB MONITORING  ### #####
#################################
 
### MonALisa Monitoring
activate_MonALisa= 0
 
### Use central BOSS DB instead of one for each task: the DB must be already been setup!
use_central_bossDB = 0
 
### Use Boss RealTime monitoring
use_boss_rt = 0
 
[EDG]
 
#
#   This section contains EDG specific stuff.
#
 
# LCG middleware version installed on testbed
lcg_version = 2
 
## to change the cms-broker RB (download these files from crab web page)
#config_vo = /home/fanzago/edg_wl_ui.conf.CMS_CERN
#config = /home/fanzago/edg_wl_ui_cmd_var.conf.CMS_CERN
 
## to add other requirements to jdl file, as example the Operating System
#requirements = (other.GlueHostOperatingSystemName == "RedHat")
 
## cpu time and wall_clock_time(=real time) in minutes. Written into the jdl file
#max_cpu_time = 60
max_wall_clock_time =2880
 
## CE Black List: all the CE whose name contains the following strings (comma
## separated list) will not be considered for submission.
## Use the dns domain (eg fnal, cern, ifae, fzk, cnaf, lnl,....)
#CE_black_list = cern
 
## CE White List: only the CE whose name contains the following strings (comma
## separated list) will be considered for submission.
## Use the dns domain (eg fnal, cern, ifae, fzk, cnaf, lnl,....)
CE_white_list = unl.edu
 
## fields written into jdl
virtual_organization = cms
 
## number or retry count
retry_count = 0
 
## LFC catalog parameters
lcg_catalog_type = lfc
lfc_host = lfc-cms-test.cern.ch
lfc_home = /grid/cms

Here is a plot of the vertex resolution for the Divisive Method in ORCA (875) versus the number of tracks used in the vertex. This is in bb100 events. I don't understand why it has this shape.

After making pixel rechits, I was able to run the little example that spins over the pixel triplets

First, make the file with the 6 muons per event with the pixel recHits, as described in the previous log entry.
Then, compile and run the pixel triplet executable from Marcin Konecki:

scramv1 project CMSSW CMSSW_0_6_0_pre5
cd CMSSW_0_6_0_pre5/src
cmscvsroot CMSSW
cvs co -A RecoPixelVertexing
eval $(scramv1 runtime -sh)
scramv1 b

Then edit the config file in the test area so that it loads the root file with the pixel recHits in it from the previous step. Then run it:

cd RecoPixelVertexing/PixelTriplets/test
cmsRun trip.cfg

You should see some printout about the number of triplets found per event. I see like 4-7 triplets per event with my 6 muon file.

Its been tough to get the older Pixel Vertexing code to run.  To get Susanna's code to compile correctly, do the following:

cmscvsroot ORCA
cvs co -r ORCA_8_7_3 TrackerReco/PixelTrackVertexFinder
cvs co -A TrackerReco/PixelTrackVertexFinder/test/
cd TrackerReco/PixelTrackVertexFinder/src
cvs update -A
ll PTVFDivisiveVertexGenerator.cc
cd ../interface
cvs update -A
pwd
cd ..
scram b
cd test
scram b bin 

This should allow you to run PixelTrackVertexTest to make histograms.

The orcarc file that I used to run over one of Susanna's sets is:

FirstEvent =0
MaxEvents = 10
 InputFileCatalogURL = @{
xmlcatalog_file:/opt/METADATA/Prod/bt03_b_pt120-170/POOL_Catalog.META.bt03_b_pt120-170.xml
xmlcatalog_file:/opt/METADATA/Prod/bt03_b_pt120-170/POOL_Catalog.EVD.bt_2x1033PU761_TkMu_g133_CMS.bt03_b_pt120-170.xml
xmlcatalog_file:/opt/METADATA/Prod/bt03_b_pt120-170/POOL_Catalog.EVD.bt_Hit3653_g133.bt03_b_pt120-170.xml
xmlcatalog_file:/opt/METADATA/Prod/bt03_b_pt120-170/POOL_Catalog.EVD.bt_Hit752_g133.bt03_b_pt120-170.xml
 
}@
 
InputCollections = /System/bt_2x1033PU761_TkMu_g133_CMS/bt03_b_pt120-170
 
TFileAdaptor=true

Difference in the cfg files that Aaron needs.

The config files from Aarons post are 404 (have I mentioned lately how much I dispise this interface?).  I'll try and show the changes he made in the files inline so that you can make cfg files on your own from the checked out code:

---

sim10mum.cfg:
Change:
vint32  PartID = {13}
to
vint32  PartID = {13, -13, 13, -13, 13, -13}

---

digitize_strips_from_flatptgun.cfg:
Add
 #include Pixel Digitizer
 include "SimTracker/SiPixelDigitizer/data/PixelDigi.cfi"
and add pixdigi to the line with stripdigi:
pixdigi,stripdigi

---

DigisToClusters.cfg
Add:
#include  pixClust
include "RecoLocalTracker/SiPixelClusterizer/data/SiPixelClusterizer.cfi"
and PixClust to the line with ThreeThresholdClusterizer
 pixClust,ThreeThresholdClusterizer

---

ClustersToHits.cfg:
Add:
#include pixRecHitConverter
include "RecoLocalTracker/SiPixelRecHits/data/SiPixelRecHits.cfi"
and add pixRecHitConverter to the line with LocalMeasurementConverter
pixRecHitConverter,LocalMeasurementConverter

---

Basically all Aaron is doing here is turning on the Pixel layers in the run.

Here is the recipe and the config files to make pixel recHits to be used in triplet finding under CMSSW_0_6_0_pre5

I followed the general recipe from Oliver at https://twiki.cern.ch/twiki/bin/view/CMS/RS_0_6_0_pre5
But I had to modify most of the early config files to add in the pixel stuff. See the attached files:

• sim_10mum.cfg: make 6 muons per event
• digitize_strips_from_flatptgun.cfg
• DigisToClusters.cfg
• ClustersToHits.cfg

Finally, you should get an output file like this: ClustersToHits.root