BlackHat Ntuples Analysis

• Author: Ivan Pogrebnyak, MSU, ATLAS
• Repository: https://github.com/ivankp/bh_analysis.git
• Source (zip): https://github.com/ivankp/bh_analysis/archive/master.zip

Usage

Compiled executables are created in the bin directory. Help message can be displayed by running ./bin/executable -h, or ./bin/executable --help, or, in most cases, simply ./bin/executable without any arguments.

Below is the list of executables, their purpose, and usage examples. The list is in the order in which the programs need to be run to perform the analysis and obtain plots.

reweigh

• Purpose: Reweighting of BlackHat ntuples.
• Output: A root ntuple with only new weights.
• Usage example: ./bin/reweigh --bh=born_bh.root -c weights.xml -o bort_weights.root
• XML config file: Provides new weights definitions; check the config directory for examples.

hist_foo

• Purpose: This this the analysis program. It produces plots for different weights.
• Output: A root file with histograms in directories corresponding to selected weights.
• Usage examples:
  • Minimal:
    ./bin/hist_foo --bh=born_bh.root -o bort_hist.root
    Produces histograms for the weight in the BlackHat ntuple; clustering is done by FastJet, AntiKt4 algorithm is the default.
  • Use new weights:
    ./bin/hist_foo --bh=born_bh.root --wt=born_weights.root -o bort_hist.root
    Clustering is done by FastJet
  • Use SpartyJet ntuples:
    ./bin/hist_foo --bh=born_bh.root --sj=born_sj.root --wt=born_weights.root -o bort_hist.root

Note: Numbers of entries in histograms are not numbers of events, but numbers of ntuple entries. These are not the same for real ntuples.

merge_parts

• Purpose: Merge together histograms for different kinds of ntuples (born, real, integrated-subtraction, virtual).
• Output: Root file in the same format with merged histograms.
• Usage example: ./bin/merge_parts NLO.root B.root RS.root I.root V.root

Note: This program is meant to combine histograms from different types of ntuples, e.g. born with real. For combining the same type of ntuples together use hadd program that comes with ROOT. If you have multiple ntuples of the same type, combine same types together with hadd then merge the combined histograms with merge_parts.

plot

• Purpose: Plot histograms with scale variation and PDF uncertainty bands.
• Input: A single root file with histograms in directories corresponding to scale and PDF variations.
• Output: A single pdf file with plots on multiple pages.
• Usage example: ./bin/plot foo/bar/NLO.root will output NLO.pdf in the current directory. -o flag is also supported.

---

Compilation

• Simply make
• You can safely use the -j flag

Requirements

The code is written in C++11 and requires a C++ compiler which supports the -std=c++11 flag; -std=c++0x is insufficient.

Other requirements:

• ROOT: https://root.cern.ch/drupal/ -- either version 5 or 6 is ok
• BOOST: http://www.boost.org/ -- version 1.42.0 and above
• LHAPDF6: https://lhapdf.hepforge.org/ -- version 5 is not supported
• FastJet: http://fastjet.fr/

---

Making changes

Adding a new analysis

• Make a copy of src/hist_H3j in the src directory. Make sure the new file's name also starts with hist_ for make to pick it up automatically.
• If you use GitHub, please submit a pull request, so that your code can be incorporated into the repository for the benefit of others and maintanance.