This script allows to produce BoostHistograms in N dimensions in MT mode, unroll them in 1D and run combineTF. (credits J. Bendavid, J. Eysermans). After unrolling it is possible to remove empty bins and require either minimum total yield per bin or minimum total MC statistics per bin.

Installation

Make sure singularity is installed by typing

singularity version

Install CMSSW and combineTF (do not source env with cmsenv)

cmsrel CMSSW_10_6_20
cd CMSSW_10_6_20/src/
cmsenv
git clone -b tensorflowfit https://github.com/bendavid/HiggsAnalysis-CombinedLimit.git HiggsAnalysis/CombinedLimit
scram b -j 8
cd ../../

Run instructions

In a new shell source environment:

To run the templates.py script to produce unrolled 1D template histograms with HistoBoost within the singularity image:

python template.py  --datadir /eos/experiment/fcc/ee/analyses/case-studies/higgs/flat_trees/zh_vvjj_var_v3 --cfg config/config_zhvvjj_13poi.py --min-yield 0.0 --min-mcstat 1.0 --tag vvjj_13poi

arguments:

• --datadir: where the processes flat trees are located
• --cfg: analysis config file, where processes, cuts, and axes for the Ndim boosthist are defined
• --min-yield: minimum yield per bin when unrolling in 1D
• --min-mcstat: minimum mc stat per bin when unrolling in 1D

This script will also generate a combine datacard in the job directory.

Now run limits using combineTF:

python limits.py  --datacard job_vvjj_13poi/datacard.txt --tag vvjj_13poi

arguments:

• --datacard: where the datacard is located (by default will take the one present in the job dir)
• --tag: a string to indentify the job (has to be the same as in the template.py script)