This is a recipe for skimming D3PDs on the grid. It is written with the WZ D3PDs in mind, but it can also be applied to other D3PDs, of course. The idea is to only download the skimmed datasets to save space (and time when running over them). You need a grid certificate (in case you don't have one, see here: GetGridCertificate) and you need to source pathena (for more information on pathena,DQ2 and panda and grid submission, see here: ATLASOfflineTutorial)

WZ D3PDs

script for skimming

  • can be found here: /work/msu-pc4/sarah2/D3PDSkims/D3PDSkim/slim_2e.py
  • copy this file to your work directory
  • it is currently doing the following preselection
    • GoodRunsList
    • at least 2 electrons with isEM_loose and author 1 or 3, also pt > 10 GeV (can be changed according to needs - recommend author 1 or 3 or 8, also require only 1 electron to do Tag and Probe)
  • it also throws some branches out, so that the tree gets a little slimmer (careful that you don't slim away anything that you might need in the future!)

How to use it

  • You submit it to the grid, usually you submit all the files belonging to the used GRL (GoodRunsList) together, but if they are not all on one site, you can also divide it up
  • example on how to submit it (no space between the commas and the datasets, the "-" in front of "%IN" is important for some grid sites!):
###period D4
###put the following GRL into the code
###data10_7TeV.periodD4.159041-159086_LBSUMM_DetStatus-v03-pass1-analysis-2010D_data_eg_standard_7TeV.xml
#prun --exec "python slim_2e.py - %IN" \\
--inDS group10.phys-sm.data10_7TeV.00159041.physics_L1Calo.recon.ESD.f275.WZphys.100612.01.D3PD,\\
group10.phys-sm.data10_7TeV.00159086.physics_L1Calo.recon.ESD.f275.WZphys.100612.03.D3PD \\
--outDS user10.SarahHeim.skim.periodD4.physics_L1Calo.recon.ESD.f261.WZphys.100612.02.D3PD \\
--nFilesPerJob=5 --match "*D3PD.D3PD*" --athenaTag=15.6.10.4 --outputs "skim.root" 
  • More examples can be found here: /work/msu-pc4/sarah2/D3PDSkims/D3PDSkim/submit_slim_2e.sh

When running

  • You can see the status of your job on your panda user page: Go to PANDA monitoring page: http://panda.cern.ch:25880/server/pandamon/query and type in the pandaID or go to PANDA user page: http://panda.cern.ch:25880/server/pandamon/query?dash=Quota and find your name
  • You will receive an email when the job is done and you can download the produced skimmed dataset with
     dq2-get outputfilename.bla 
    (You'll need to source dq2 for that:
    source /afs/cern.ch/atlas/offline/external/GRID/ddm/DQ2Clients/setup.sh 
    voms-proxy-init --voms atlas

-- SarahHeim - 23 Jul 2010
Topic revision: r1 - 23 Jul 2010, SarahHeim
 

This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Foswiki? Send feedback