Warning, /analysis/LightFlavorRatios/README.md is written in an unsupported language. File is not indexed.
0001 # Light-Flavor Ratio Analysis
0002
0003 Repository for determination of ratios of light-flavor resonances, originally constructed for PPG 16 (Lambda/2Kshort ratio).
0004
0005 ## Input
0006
0007 The current codebase uses two KFParticle output files, one for Kshort candidates and one for Lambda candidates.
0008 The locations of these files are, for now, `/sphenix/tg/tg01/hf/mjpeters/LightFlavorResults/Kshort_3runs.root` and
0009 `/sphenix/tg/tg01/hf/mjpeters/LightFlavorResults/Lambda_3runs.root`.
0010
0011 ## Usage
0012
0013 ### Running standard analysis
0014
0015 In `yield_and_ratios`:
0016
0017 `root -l calculate_ratios_withRooFit.C`
0018
0019 This runs on the aforementioned input files and generates the file `fits.root` containing the mass peak fits, yield histograms, and ratio histograms.
0020
0021 To generate plots:
0022
0023 In `yield_and_ratios`:
0024
0025 `root -l -b plot_results.C`
0026
0027 This function extracts the histograms from `fits.root`.
0028 Inside this macros there is a variable `finalize`.
0029 If set to `true`, it saves the plots to `/sphenix/tg/tg01/hf/mjpeters/LightFlavorResults/plots`.
0030 If set to `false`, it saves the plots to `plots/` in your local directory.
0031 (Recommended best practice, if you modify the plotting macro, is to turn `finalize` to `false` until you're certain you like the results.)
0032
0033
0034 PDF versions of the plots are saved to the `pdf/` subfolder, and PNG versions are saved to `png/`.
0035
0036 ### Changing binning scheme
0037
0038 Modify the file `util/binning.h`. The static structures at the bottom of this file are the bins used in the analysis.
0039 Options for linear, logarithmic, and custom binning are provided.
0040
0041 ### Adding or modifying corrections
0042
0043 The `corrections/` folder holds the structures and settings through which various corrections are applied.
0044 Currently, this only contains the relative tracking efficiency correction, which uses Tony's 7-bin histogram, only applied to pT differential results.
0045
0046 ## Caveats/To-Dos
0047
0048 - Macros repository is almost certainly outdated and needs to be synchronized with latest light-flavor production
0049 - Unbinned fit will not work for larger datasets, needs to be supplanted by a binned-fit option.
0050 - The plots of the individual mass peaks may have very bad-looking jumps if the background function becomes negative.
0051 As far as I can tell, this is because RooFit doesn't like when that happens.
0052 Working on finding a fix -- for now, just fiddle with the background function a bit until this doesn't happen.
0053 - Background model tends to misbehave at the edges of the mass window.
