|
|
Documentation of Appmodule QaBetaTools/QaElecIdTest
$Id: QaElecIdTest.html,v 1.3 1998/05/31 00:59:36 dyaeb Exp $
Benchmark
- 040402 040404 040405 for muons of 500 MeV, 1.2 GeV and 3 GeV respectively
- 040474 040476 040477 for electrons of 500 MeV, 1.2 GeV and 3 GeV
- 040488 040490 040491 for pions of 500 MeV, 1.2 GeV and 3 GeV
- 040510 for J/psi(mumu/ee)K0s tape
Input
- three BtaCandidate lists from the event, "trackCandidates", "calorCandidates" and "truthCandidates"
- the defaults for the parameters are set to "ChargedTracks", "CalorObjects" and "MCTruth" respectively.
Method
- accumulate an ntuple of xxxPidInfo consistencies
- truth is judged according to the "Default" BtaMcAssoc list
- using a .kumac, derive misID vs efficiency plots at three energies for e/pi and mu/pi separation in each subsystem
Output
- ElecIdTest_RELEASE.ps and MuonIdTest_RELEASE.ps
Detailed Information
This is still useful, but no longer correct -- to be rewritten.
Here is a step-by-step guide on how to run it.
- Checkout the test release you want to QA, with any recommended tags.
- Build QaBetaToolsApp (you can of course use any speedup script you are familiar with):
- addpkg QaBetaTools
- gmake lib
- gmake QaBetaToolsApp.bin
- Make sure you have workdir checked-out and do gmake setup therein.
- Go into QaBetaTools directory
- copy qarun into workdir directory
- setenv VER 4.3.9a
(or whatever version of beast tapes you want to read - this is not necessarily the same number as the release version you are using)
- source submitElecID
(this will submit all the jobs for you). Check that they are running with:
- bjobs -u yourid
(there should be 10 of them)
- When they finish, the log files are "reco$RUN.out" and the ntuple files are "elecid$RUN.hbook" in the workdir directory where $RUN is one of the following run numbers:
- 040402 040404 040405 for muons of 500 MeV, 1.2 GeV and 3 GeV respectively
- 040474 040476 040477 for electrons of 500 MeV, 1.2 GeV and 3 GeV
- 040488 040490 040491 for pions of 500 MeV, 1.2 GeV and 3 GeV
- 040510 for J/psi(mumu/ee)K0s tape
- Still from the QaBetaTools directory, when all (except the last which will take longer) are finished, type:
- Start paw and type
- exec ElecIdTest
when that finishes, type
- exec MuIdTest
when that finishes, type
- exec multih
- The .ps files should be in:
- elecidtest.ps
- muonidtest.ps
- multih.ps
- That's it.
To Interpret:
- For the multih.ps, just check that they make reasonable sense (ie. some semblance to flatness for the right hypothesis and some semblance of a peak near zero for the wrong one. Check that there are not too many entries of -1 for any system and any
hypothesis (for example > 25% or -1's is pretty bad).
- That's it.
Maintained by Paul Harrison.
Please send corrections, additions and (polite) suggestions to:
pfh@SLAC.Stanford.Edu
|
|
