to extract signal lifetime ----
-
Branching fraction fit
-
sPlot lifetime fit ----- For the YR study, we only include statistical uncertainty.
- gcc 4.8 or newer
- ROOT 6.10/02
cd your_working_area
git clone https://github.com/kfjack/bmm4uml
cd bmm4uml
mkdir fig input/bmm4
make -j8
Please modify the Makefile to match your system (generally it should work out of box, but just in case!)
There are 3 tools at this moment:
- bmm4fitter: the main fitter and for UML model construction.
- bmm4toystudy: the toy generator and fitter
- bmm4validation: for sPlot lifetime systematic studies
The wspace_prefit.root contains all of the PDFs, parameters, and constraints.
Please take a look at the full command list: % ./bmm4toystudy
commands:
- genfit : gen & fit to toy
- minos : call MINOS for BFs
- tausplot : enable lifetime sPlot study
- freq : produce frequentist toy (w/ postfit nuisances)
- bdstat : calculate profile likelihood test statistics (F&C study or upper limit for Bd)
- make_plots : produce resulting plots from toy
- tausyst : enable lifetime systematic study
- make_demo : produce demo projection plots
- seed=[1234] : set the random seed
- iterations=[10] : set # of iterations
- yield_scale=[1.0] : set scaling factor for the yields
- ws_gen=[wspace_prefit.root] : set generation workspace
- ws_fit=[n/a] : fit model workspace
- ws_res=[wspace_toyresult.root] : resulting workspace file
- set_genpar [name]=[value]=[float/fix] : set parameter value & state (for gen workspace)
- set_fitpar [name]=[value]=[float/fix] : set parameter value & state (for fit workspace)
A typical call:
% ./bmm4toystudy genfit minos tausplot seed=1234 iterations=10 ws_gen=wspace_prefit.root ws_res=wspace_toyresult.root
This will generate 10 sets of toy, with initial seed of 1234, based on the wspace_prefit.root UML model. The fitting model is not set, so it will be the same as the gen model (=wspace_prefit.root). The minos option means the MINOS command will be called for the two branching fractions. The tausplot option means the gen & fit also involes sPlot lifetime fit.
The results will be stored in wspace_toyresult.root.
One can excute the following commend to analyize the results and produce the figures.
% ./bmm4toystudy make_plots ws_res=wspace_toyresult.root
This command also supports for multiple result files, for example, if all of the wspace_toyresult*.root
are placed under toy/
, one can analyize them in a single call:
% ./bmm4toystudy make_plots ws_res=toy/wspace_toyresult\*.root
The resulting figures (mean/error distributions/pulls/etc) will be stored under fig/
.
A demonstration of generator mass PDF + overlapping toy events can be generated by
% ./bmm4toystudy make_demo
For advanced users:
- set_genpar/set_fitpar command can change any of the generating/fitting PDF parameters on the fly.
- yield_scale commmand can scale the overall yields by the given factor in each toy.
- Please copy all of the files under
bmm4/small/180428/s01/
toinput/bmm4/
% ./bmm4fitter create_base
This should re-build the full UML model based on the input data card and small trees. Note this will take a long time.
The default outputs are wspace_base.root and wspace_prefit.root. The contents in wspace_base.root and wspace_prefit.root
should be exactly the same at this stage.
% ./bmm4fitter update_pdf
This will only update the PDF of the UML model, reads from wspace_base.root, and saves to wspace_prefit.root.
This will take a long time.
% ./bmm4fitter update_config
This will only update the configuration of the UML model (yields and contraints):
reads from wspace_base.root, and saves to wspace_prefit.root.
% ./bmm4fitter update_data
This will only update the data events storaged in the workspace:
reads from wspace_base.root, and saves to wspace_prefit.root.