For a more detailed user's guide and description of the package see

1. Clone the repository git clone https://github.com/KinFit/KinFit.git
2. Build the library using cmake
   mkdir build
   cd build
   cmake .. -DCMAKE_INSTALL_PREFIX=/your_preferred_install_location
   make
   make install
3. Add library to rootlogon.C
   gSystem->Load("pathtoinstall/libKinFit.so")

KinFit depends on the ROOT libraries Core, Physics and Tree
When properly installed, KinFit can be included and linked to other CMake projects by

find_package(KinFit REQUIRED)
target_link_libraries(your_executable other_dependencies KinFit::KinFit ROOT::Core ROOT::Physics ROOT::Tree)

1. Create KFitParticles from your particle candidates
   KFitParticle particle(TLorentzVector lv, Double_t R, Double_t Z);
   FillData(particle, particle_covariance)

2. Create a vector of particles which shall be fitted
   std::vector cands;
   cands.push_back(particle);

3. Initialize the KinFitter and fit
   KinFitter fitter(cands);
   fitter.addXXConstraint(arguments);
   fitter.fit()

4. Get the fit result
   KFitParticle fcand = fitter.getDaughter(0);

The procedure is illustrated in the example macro fit_toyMC.C in QA. An example input file is provided in QA/test_files/toy_montecarlo_rz_vtx.root.

The KFitAnalyzer can be used to automatically analyze a root file event by event, performing the selected fit. The best particle combination for each event according to fit probability is written to an output file.

Input: A root file containing a TClonesArray of KFitParticles. Example: QA/test_files/input.root

Output: A root file containing a TClonesArray of KFitParticles, fit probability and chi2

1. User uses the macro analysis_user.C to define input and output and the fit that should be performed.

2. KFitAnalyzer object is created. User adds fitting tasks to it, defines PIDs of particles to be fitted.

3. KFitAnalyzer creates the output tree. The particle combination with the best probability is written there for each event.