Problem: when I tried to compare data and bkg model, I realized that jet pt of data doesn't make sense, for example, I applied jet pT > 15 GeV (non-1st and non-2nd leading jets), why are there some jets with pT < 15 GeV (see the figure attached).

So what I checked:

• Use the simplest method to plot jet PT by Hist1D in cafe, jet pt looks making sense.
• Write a very simple cafe processor to check this, all jet pT make sense
• Add cafe.ReadPartail: false to my configure file, no effect
• At last, looked input SingleTopTopovars.cpp, and commented out a line about _JetCollection.push_back(RejectedGoodJets[i]); which may cause the problem. Yes, this is the reason. Because I specify a rejected good jet branch without any pt and eta cut, so no wonder lower jets come in. Fixed: 1. Add a printing note if the branch is speficied. 2. Remove the branch spefication in my base configure files.
  

I don't expect this configure bug will affect our TRF measurements and QCD samples. Just do btagging again now....

After meeting with Dugan and Ernest, this is a to-do list:
* MC and QCD comparison with DATA for checking MC pt cut
* Find why jet pt is not applied in data!! Fix data and do btagging of data!
* Use recaffed p17 data
* Understand why our yields of signal samples have been enhanced so much?
- Comparison between old selection eff. and new one.
* p20 data pretagging.
* Tell Christian I will give a talk then.
* Understand TRF in terms of detector eta.
- function jet physics eta to jet detector eta
* Derive p20 data luminosity information

Another error when I did MC pretagging is

"cafe: caught exception during running: No MC type for request ID 32875 specified in the file caf_mc_util/mc_sample_info/MC_reqid.list"

The reason for this is that in my script to submit jobs I forgot a break '\' where it is supposed to have one to break one command line to several parts.

cafe continues crashed due to some reasons. I spent days on this problem. Crash error at StatBranch.inputClosingFile() ... I print out the pointer of tmb_tree (in the function of treehandler->ClearIt() ), it is not null, however, I cannot get even the tree name by tmb_tree->GetName() (because it will be crashed then). So I looked into the file: bdt_tau_util/src/ApplyTauID_BDTCut.cpp and removed the line:

delete tmbtree;

in the function ApplyTauID_BDTCut::inputFileClosing, then things become working again.

d0: cafe: how to skip events

Use cafe.Skip to skip events when doing some debugging. See the file cafe/src/RunController.cpp for more options.

I realized that I reached up the limit of my quota on Fermilab account. So I have to clean to get more rooms. What I did:

Use the command

du -s * .[a-zA-Z]* | sort -n

to check home directory. Remove large files

log:

Yesterday, I gave a talk at the tau ID group meeting for showing ttbar->alljets plots.

Today I should finish things like:
-get uncertainty band with predicted plots
-consider next step
-apply METsig cut to see the comparison plots
-training....

log

Yesterday, I did:
- submit all failed skimming jobs agin
- and realized that my MC sam defintions of wjets are too tiny (each dataset includes too few files for running, in order to decrease total number of jobs submitted I have to redefine them):

- resubmit MC skimming jobs:
- resubmit failed jobs by using this:
./submit_4mc_pretag_2008-01-08.sh taujets/TauJets_MC_pretag_2008-01-08.config filelists/MC_resub_2008-01-11.list

log

Ernest checked in a New version of caf_trigger: Np18-br-21, which has been adapted by him about tau trigger stuff. I will use this version for MC skimming (pretagging).

Use the following line to create a shell script to submit all MC datasets:
cat filelists/MC_all_sig_wjets_wjets.list |awk '{print "./submit_4mc_pretag_2008-01-08.sh taujets/TauJets_MC_pretag_2008-01-08.config "$1 }' > runsubmit_4mc_btag_2008-01-09.sh

tau: skimming

Working area:
/work/yoho-clued0/zhiyil/p1811_top_cafe_v1_5_1

- compile those files
- update packages for tau stuff
- submit jobs to skim data (pretag), output dir:
/rooms/algonquin/zhiyil/Data_tau_skim/3JET_PreTag_p1811_top_cafe_v1_5_1_2008-01-08/
why use newly submit? some changes since last time:
- new packages, JES etc.
- METsig > 3.0 set in configure file (METsigAlg)
- new tau ID: NN 0.8 - 1.0

I use the command to submit the skimming jobs:
./submit_4data_pretag_2008-01-08.sh taujets/TauJets_DATA_pretag_2008-01-08.config filelists/DATA_CAF_SKIM_3JET_LOOSE_v1_2007-04-05.list &> ./submit_4data_pretag_2008-01-08.log&