Object Corrections: d0correct



Include your latest ID package into the next version of d0correct

You have to fill out the request page if you want your package to be included into the next version of d0correct.

List of packages requested for the next version.

How to use d0correct:

Important:

d0correct calls packages (like emreco) which need certain data files and environment variables set. If you forget to copy a file or set a variable your program might crash, but it also might not. It is strongly advised to use d0tools with options
-initscript=initchunkanalyze.sh -sourcescript=chunkanalyzesource.sh


with analysis_example as an example of analysis package or root output maker:

We have to work within a release area (e.g. p14.05.01).

get d0correct:


do: addpkg analysis_example and then:

gmake all

Before running it, you have to copy the ascii files needed for EM post-processing by calling initchunkanalyze.sh (available in d0tools).

with tmb_analyze:

We have to work within a release area (e.g. p14.05.01).
do addpkg d0correct v00-00-XX (see versioning below) then: sh d0correct/scripts/buildD0Correctp1405_with_TMBAnalyze.sh
and gmake all or copy the tarball (location in the versioning table below). This tarball contains a release structure and the binary).

What is in d0correct ?

d0correct calls for EM post-processing, MuoCandidate, applies jet ID criteria, corrects jets for energy scale and finally corrects missing Et. To do that, d0correct calls all the individual object correctors within the ID packages.

Current Strategy for Jets/JES/missing ET corrections:

  1. Create lists of good objects:
  2. muons used to apply JES and to correct missing Et are defined here.
    EM used to correct missing Et and to remove jets matching them are defined here.
  3. Apply Jet ID quality cuts (definition of good jets)
  4. Remove jets matching EM objects
  5. Perform JES correction:
    associated muon with jets (in case of few muons we'll use muon with max pTrel)
    correct jet four vector for JES and create new chunk for corrected jets
  6. Correct missing Et: loop over good muons, good EM and good Jets to corrected missing ET for them and fill additionnal variable in the missing Et chunk.

Only JCCA and JCCB jets are corrected. Chunks corrected for JES but without taking into account muons are also created.
b-ID certification is not part of d0correct.

Versioning:

The current version of the d0correct corresponds to the "analysis" release. With p17 data (RunIIa) the most recent p18 release should be used (currently p18.09.00). With p20 data (RunIIb) p21 release should be used (current version p21.02.00). Unfortunatly some of the most recent versions ID developments available only at the CAF tree level (examples: JSSR, muon smearing). Please, refer to the "Common Analysis Format" web page about the information how to use them.

Post-Winter'04 versions

 
d0correct version EM main changes Muon main changes Jets/JES main changes Missing Et main changes
  EM doc, Change log Muon doc Jets doc, JES doc Missing Et doc
setup D0RunII p16.03.00
addpkg d0correct v7
(tarballs: /work/sasani-clued0/deliot/d0correct/v7/)
new likelihood
new EM scale
T42 compatible
Certification Status: OK
small bug fixes
Certification Status: OK
ability to write certified chunk
bug fix in matching muon with jets
T42 compatible
new JES: v05-03-00
Certification Status: Moriond cuts,
will change soon

new methods
Certification Status: OK
 
d0correct v7.5
built-in version in p16.04.00
all needed packages released into p16.04.00
no need for addpkg or tarballs
 
= v7
Certification Status: OK
bug fix in chunk selector
Certification Status: OK
new Jet ID cuts
implementation of jet MC smearing
Certification Status: OK
bug fix
Certification Status: OK

Versions for Pass-2 Data

 
d0correct version EM main changes Muon main changes Jets/JES main changes Missing Et main changes
  EM doc, Change log Muon doc Jets doc, JES doc Missing Et doc
 
d0correct v8
built-in version in p16.05.01
all needed packages released into p16.05.01
no need for addpkg or tarballs
 
changes in the likelihood, add MC smearing
Certification Status: OK
add MC smearing
Certification Status: OK
changes in MC jet smearing
JES: v05-03-00
Certification Status: OK
propagate MC smearing
Certification Status: OK
D0-Note 4646 describes this version
You should call for RCP dupl = <d0correct RemoveDuplicate> before d0correct in your main framework.rcp file
Wait for the re-skim of p14.03 pass-2 data to use this version on p14.03 skims
Known bugs: Calt4_25Chunk is not copied if you use TMBStream_x to further skim the pass-2 data
Patch: addpkg cal_t42 p16-br-08, addpkg np_tmb_stream p16-br-04 (will be fixed in p16.05.02)
d0correct v8.1
built-in version in p16.05.02
same ID certification as v8
Corrections for the re-skimming bug of v8
D0-Note 4646 describes this version
d0correct v8.2
built-in version in p16.06.00
same ID certification as v8
bug fixes in muon MC smearing, muon-jet matching for muon JES corrections and ouput of electron/cps likelihood
d0correct v8.3
built-in version in p16.07.00
new EM tracking matching
unchanged wrt v8
make Jet L1conf run dependent
(apply old ID cuts where the L1 coverage is unavailable)
unchanged wrt v8

Versions for p17 Data

 
d0correct version EM main changes Muon main changes Jets/JES main changes Missing Et main changes
  EM doc, Change log Muon doc Jets doc, JES doc Missing Et doc
 
d0correct v9
built-in version in p18.01.00
This version contains the best ID codes we have for the moment to analyze p17 data.
p18 contains also all needed packages to produce CAF samples.
Certification Status:
Not certified
Certification Status:
Not certified
JES: v05-03-00
Certification Status: Not certified
Certification Status:
Not certified
d0correct v9.1
built-in version in p18.03.00
new E-loss corrections
Certification Status: OK
Certification Status: OK new jet ID cuts
JES: v05-03-00

Certification Status: Not certified
Certification Status:
Not certified

How to access corrected chunks:

Instructions how to access the certified chunks out of d0correct can be found here.
Instructions how to access the certified jet chunk out of d0correct starting from p18 versions: here.

How to remove duplicate events (for p14 analysis):

Those events come from problems occured during the skimming or from files that have been processed sereval times with different versions of d0reco.
The simpliest way to remove them is the call for RCP dupl = <d0correct RemoveDuplicate> after ReadEvent in your main framework.rcp file.
If you have already produced your root output without it, you can find in Marco's D0news how to apply the removal method in your analysis code.


Comments to CSG Conveners

Last modified: Tue Jan 9 10:05:35 CST 2007