Skip to content

Releases: kpedro88/Analysis

v4.2

19 Sep 04:38
Compare
Choose a tag to compare

This release updates the skimming and plotting code for TreeMaker ntuples through 2015B v2.0. There are also various code improvements, including:

  • use of Selectors and Variators with KBuilder
  • fully generalized RA2bin selector (including non-exclusive binning option)
  • option to disable branches when plotting (significant speedup)
  • allow bin labels
  • add sync selectors (for synchronization of cuts on specific objects)
  • new samples, selections, datacards
  • simplify MakeClass dependencies (both skimming and plotting use NtupleClass)
  • improvements to legends and plotting
  • some code cleanup

v4.1

14 May 16:37
Compare
Choose a tag to compare

This release updates the skimming and plotting code for the latest version of the ntuples produced by TreeMaker. A class to handle configurable RA2 bin assignments was added. Some style improvements for skimming and plotting output were made.

v4.0

06 May 16:12
Compare
Choose a tag to compare

This is the first release that replaces support for the LQ2012 analysis with support for the SUSY2015 (RA2/b hadronic SUSY) analysis.

Major changes were made to the selectors, KBuilder, underlying MakeClass headers, and input files. Support for TChains was added and the KBuilder Loop() was optimized. Some LQ2012-specific files (corrections, drivers) were removed.

v3.5

23 Apr 21:42
Compare
Choose a tag to compare

This release adds a laundry list of features, mostly style and plotting options. The important style classes KPlot and KLegend can now be used in a standalone manner, to create publication-level plots in any macro.

KPlot features:

  • can be initialized from a pre-built histogram
  • most text and sizes are configurable via input options
  • can add dotted cut lines on a given histo (in x or y), specified via input options

KLegend features:

  • can place a legend in a specified quadrant of the plot, without any smart resizing of the y axis
  • multiple lines of extra text at the top of the legend
  • sets can have their own extra_text that goes under their name in the legend
  • multiple panels now allowed: entries must either be assigned panels, or can be auto-balanced between the specified number of panels

plotting options:

  • ROC curve option: for each defined histo, plot efficiency (sig vs. bkg); lists of sig and bkg provided in input options
  • (sets calculate their own efficiencies for each requested histo and cache the results)
  • Profile option: doesn't work with stacks or data-driven bkg reweighting

Various parts of KPlot, KLegend, and KPlotManager have been streamlined, and the condor submission scripts for skimming have been updated.

This is (provisionally) the last release that supports the LQ2012 (third-generation leptoquark) analysis. Future releases will support the SUSY2015 (RA2/b hadronic SUSY) analysis.

v3.0

30 Mar 14:55
Compare
Choose a tag to compare

This version unifies the skimming and plotting code into a single framework. The selections and systematic variations for skimming are now performed in an object-oriented way, using (respectively) KSelection + KSelector and KVariation + KVariator.

In addition, some updates to the processing of input files have been made. KParser is now a namespace for various basic text parsing functions that act on individual lines or elements of the text input files. This namespace is extended in the headers of certain object classes that can be created by KParser functions. The KManager classes handle the structure of the text input files and the process-dependent creation of objects from text input. Text input files can now point to other text input files in the "INPUT" section.

v2.8

05 Feb 18:01
Compare
Choose a tag to compare

This version adds support for 2D histograms. Several new classes were introduced to support 2D histograms:

  1. KValue, a small class used to keep track of value/weight pairs for filling histograms in KBuilder
  2. KPlot2D, an extension of KPlot for 2D histograms.

The 2D histograms are drawn using the "colz" option. For a given set of quantities x vs. y, a separate histogram is drawn for each top-level set in the input file. If the ratio option is enabled, another histogram is drawn, showing (numer - denom)/sigma, i.e. the discrepancy between numer and denom in number of standard deviations.

The drawing of 1D histograms in KManager was also streamlined in the process of this upgrade.

An example configuration for a 2D histogram is provided in Plotting/input_final_dd_fig.txt.

v2.71

03 Feb 18:20
Compare
Choose a tag to compare

This version includes the reorganization of various helper functions into two new namespaces, KMath for math functions and KText for text processing functions. The math functions were previously member functions of KBuilder and the text functions were previously member functions of KParser, but their use cases are more general, so separate helper namespaces make more sense.

v2.7

02 Feb 19:47
Compare
Choose a tag to compare

This version introduces a new format for input options, type:name[value], which can be used for global options, histos, and sets. Histos/plots and sets now have a separate set of local options used to store member variables and other object-specific settings. The broad use of this new option format significantly improves the flexibility of the code and streamlines the parsing of input files. The input files themselves contain slightly more text, but gain in readability and adaptability.

v2.6

27 Jan 21:22
Compare
Choose a tag to compare

This version includes mostly cosmetic, but important, changes. All plotting styles and legend positioning are done algorithmically now. Elements of the plotting style have been updated to the latest CMS PubComm specifications (https://ghm.web.cern.ch/ghm/plots/). Some borrowed "database" code to apply physics object data/MC corrections has been removed and replaced with a simpler process.

v2.5

01 Dec 20:44
Compare
Choose a tag to compare

This is the version of the code that was used for results publicly approved by CMS and published in arXiv:1408.0806.

The code in Selection runs over ROOT files (ntuples) produced by the LQ RootTupleMaker. The code in Plotting runs over ntuples produced by the code in Selection.

Batch submission scripts are formatted for use with Condor at the UMD Tier3 computing cluster.