Producing baby ntuple

Download the code

git clone --recursive [email protected]:cmstas/VBSHWWBabyLooper.git
cd VBSHWWBabyLooper

Download JECs

cd NanoTools/NanoCORE/Tools/jetcorr/data/
sh download_jecs.sh
cd -

Compile the code

source setup.sh
make clean
make -j

Run the code

sh run.sh -a v2       -t v2.5_SS [-S VBSWWHToLNuLNubb_C2V]  # v2 is the tag provided by the user
sh run.sh -a v2_jecUp -t v2.5_SS [-S VBSWWHToLNuLNubb_C2V]  # _jecUp recognizes that it is to run jecUp variation
sh run.sh -a v2_jecDn -t v2.5_SS [-S VBSWWHToLNuLNubb_C2V]  # _jecDn recognizes that it is to run jecDn variation

Save the ntuples to nfs area

sh scripts/publish_baby.sh v2          # Saving v2 to nfs area
sh scripts/publish_baby.sh v2_jecUp    # Saving v2_jecUp to nfs area
sh scripts/publish_baby.sh v2_jecDn    # Saving v2_jecDn to nfs area

In case you change something, here is how one can run a test job where each job is limited to 50000 events only

sh testrun.sh

Running MiniLooper

First download the code

git clone --recursive [email protected]:cmstas/VBSHWWBabyLooper.git
cd VBSHWWBabyLooper

Set up the environment

source setup.sh

Then compile the code

cd minilooper
make clean
make -j

Run the minilooper

sh scripts/run.sh Nominal    v2.4_SS    v2    Run2    jguiang # Nominal is the tag you provide

                  ^^^^^^^    ^^^^^^^    ^^    ^^^^    ^^^^^^^
                  |||||||    |||||||    ||    ||||    |||||||
                  Yourtag    NanoSkim  Baby   Year    Username of the nfs-area where the baby is saved

Copy the histogram outputs from the minilooper to your nfs area

sh scripts/publish_result.sh Nominal v2.4_SS v2 Run2 # Nominal is the tag you provide

Run the extrapolation factor study on the output

python scripts/alpha_exp.py Nominal    v2.4_SS    v2    Run2    yxiang  # Nominal is the tag you provide

                            ^^^^^^^    ^^^^^^^    ^^    ^^^^    ^^^^^^
                            |||||||    |||||||    ||    ||||    ||||||
                            Yourtag    NanoSkim  Baby   Year    Username of the nfs-area where the histograms are saved

Run the plotter to create a bunch of plots

sh scripts/make_figures.sh Nominal    v2.4_SS    v2    Run2    yxiang   

                           ^^^^^^^    ^^^^^^^    ^^    ^^^^    ^^^^^^
                           |||||||    |||||||    ||    ||||    ||||||
                           Yourtag    NanoSkim  Baby   Year    Username of the nfs-area where the histograms are saved

Running HiggsCombine

Start with a different terminal

cd combine
source /cvmfs/cms.cern.ch/cmsset_default.sh

Following line has to be done once. Once setup next time one can skip.

source install_higgs_combine.sh # Only have to do this once

Next time, just run the following to setup it up:

source setup.sh

Now write the data cards

sh make_cards.sh

The output of the data cards are located in datacards/
For example, datacards/bdt_c2v4p5_combined.txt contains the 4 bin fit of the BDT anlysis for c2v = 4.5.
datacards/bdt_c2v4p5/ directory contains the individual bin's data cards.

Now to run the asymptotic limits, run the following:

sh run_asymptotic_limits.sh run_1

The script is not parallelized although it is easily doable.
Take a look at the script to see how they are configured. Currently it is configured to run over bdt analysis. It will take about 5 mins to run all.

Then to print out the limits in each c2v coupling points, use the following script.
Also take a look at the script to see how they are ocnfigured. Currently it is configured to print the limit value obtained from running bdt limits from previous step.

sh print_asymptotic_limits.sh run_1 # Same tag value. e.g. 'run_1'

This will parse and write the results into a python script in limits/run_1_asymptotic_results.py

To plot the result

python plot.py run_1 # Same tag value e.g. 'run_1'

The output will be in limit.pdf/png

To run toys

To run toys, instead of running sh run_asymptotic_limits.sh command one runs sh run_toy_limits.sh.
The script is provided, but have not obtained a stable results yet.