Skip to content

cmstas/MT2Analysis

Repository files navigation

MT2 Analysis

Setup:

To simply setup on UAF, do

git clone [email protected]:cmstas/MT2Analysis.git

and then

cd MT2Analysis
source setup.sh

Babymaker

For a test run

cd babymaker
source doTest.sh

Babymaking in batch mode

See further instructions in the README.md in folder

babymaker/batchsubmit

Laptop setup to run on MT2babies

(this assumes you already have root setup)

git clone [email protected]:cmstas/MT2Analysis.git
cd MT2Analysis/MT2Looper
make

Executing main analysis

Starting from the (skimmed) babies:

Run QCDLooper to produce rphi estimate files

Go to QCDLooper and modify do.sh to point towards the data and mc babies you wish to use. The standard skims will not work here, as the rphi estimate needs low MT2 events. Worst case, you can use completely unskimmed babies, but this takes a long time.

. do.sh

TODO: Modify QCDLooper to run on any of 2016, 2017, or 2018 intelligently. It currently (9 Aug 2018) runs with 2017 settings with 2016 commented out.

Now, follow the instructions in QCDLooper and scripts/qcdEstimate, then edit string rphi_file_name in MT2Looper/MT2Looper.cc point to your qcdHistos.root output.

Run MT2Looper to produce root histogram files

Go to MT2Looper and modify INDIR, OUTDIR, and Samples in do.sh, then do:

. do.sh

Make the data driven backgrounds

Go to scripts and modify INDIR in makeDataDrivenEstimates.sh to point to the output dir from MT2Looper. Then do:

. makeDataDrivenEstimates.sh

Merge histograms for signal scans, if necessary

Go to scripts and modify INDIR in processSignal.sh to point to the output dir from MT2Looper. Also comment out any signal scans that you did not process. Then do:

. processSignal.sh

Note that this also creates histogram files translated to ETH's format, for sharing with them.

Create datacards

Go to scripts and modify INDIR, OUTDIR, DOFULLSCAN, and samples in makeAllCards.sh. Then do:

root -b -q cardMaker.C+  # to compile
. makeAllCards.sh

Run limits

See instructions in README inside the limits/batchsubmit subdirectory.