A PROfessional, PROductive, PROficient and PROfound optimization and oscillations fitting framework.

General Installation Instructions

You should start from an environment with ROOT, HDF5, and BOOST already installed or setup. These can be installed on linux with apt-get, or on mac with homebrew. On the FNAL servers, you can get these setup by running:

source /cvmfs/larsoft.opensciencegrid.org/setup_larsoft.sh
setup root v6_28_12 -q e26:p3915:prof
setup cmake v3_27_4
setup hdf5 v1_12_2b -q e26:prof
setup boost v1_82_0 -q e26:prof

While we aim to keep generic to these versions, we confirm that with the above versions PROfit compiles and works on GPVMs.

Installing through CMAKE

mkdir PROfit
cd PROfit
git clone https://github.com/markrosslonergan/Elephant_Vanishes.git
cd Elephant_Vanishes/build
cmake ..
make -j 64

You can the add the PROfit executables to your path if you so wish

export PATH=$PATH:$PWD/bin

Installed dependancies

In order to insure compatability, some smaller depedancies will be installed by the CMAKE scripts. These include

• Eigen, 3.4.0 https://gitlab.com/libeigen/eigen.git
• TinyXML2, 9.0.0 https://github.com/leethomason/tinyxml2.git
• lbfgspp, our own fork of v0.3.0 https://github.com/yixuan/LBFGSpp.git

Installing through Python Interface

With global ROOT, hdf5 and boost avialble, setup a python environment:

python -m venv env
. env/bin/activate

Install the build dependencies for profit:

pip install --upgrade pip
pip install wheel setuptools pybind11 numpy==2.0.2

Install profit:

pip install git+https://github.com/gputnam/Elephant_Vanishes

Then you're done! You can now import profit in a python shell, run the python executables (PROsurf.py, e.g.), or run the PRO* binary executables with the PRO helper (PRO PROsurf, e.g.).

Installing for Development through Python Interface

If you want to develop PROfit, then it is useful to clone down a local copy of the repository and then install it. Start with the same steps to initialize your python virtual environment:

python -m venv env
. env/bin/activate
pip install --upgrade pip
pip install wheel setuptools pybind11 numpy===2.0.2

Then pull down the profit repository:

git clone https://github.com/gputnam/Elephant_Vanishes.git

Then enter the repsoitory and install it:

cd Elephant_Vanishes
pip install .

If you make some changes, you can update the installation by re-running pip install .

Usage

Basic Description of main classes and use cases

• PROlog : Fairly simple logging/verbosity wrapper class. This is the only way PROfit should print. Usage as:

  •   log<LOG_DEBUG>(L"%1% || Using a total of %2% individual files") % __func__  % num_files;
    

  • LOG ERROR = 1
  • LOG WARNING = 2
  • LOG INFO = 3 : This should be level of standard physics output that someone would plausibly use, but not create giant log files!
  • LOG DEBUG = 4
• PROconfig : Primary bookkeeping and XML loading class. Stores ALL information of mode-detector-channel-subchannel information. Must be created once and only once per PROfit executable and passed by reference to more complex functions when it is needed.
  • Contains helper functions channel<-> subchannel mapping..etc..
  • As well as functions for how to collapse subchannels to channels
• PROspec : Class for storing final spectra and errors. Barebones class as all binning,collapsing handeld by PRoconfig/
  • Essentially two Eigen::VectorXD ! Simple with helper functions
• PROpeller : The PROpeller, which moves the analysis forward. A class to keep all MC events for oscllation event-by-event.
  • Saved as a set of six std::vector floats or ints.
  • Currently only truth (energy), reco (energy) baseline, pdg, addedweight and precalculated bin incicies indicating where the reco variable fits in a given PROconfig defined binning.
• PROtocall : Currently a bit of a rogue set of functions that, given a PROconfig, dictate how to get binning and collapse Covariance matricies
• PROcess : The master weighting function that combines everything to give one final output spectra (PROspec) event-by-event.
  • FillRecoSpectra (PROconfig,PROpeller,PROsyst,PROosc, shifts, phys param)
• PROcreate : Workhorse classes for loading the set of root ntuples and systematics into PROspec CV class and a vector of SystStructs for systematics.
  • Needs a PROconfig to tell it exactly what to load and where (channel and subchannel info) to assign the variables. Also which systeatics to load.
  • The class SystStruct is also defined in here (for some reason), which stores the weights and spline information for systematics
• PROsyst : Class that groups all systematics (each with a SystStruct) and manages their formation and effect on PROspecs
  • i.e generates covariance matricies, fills splines, gets spline shifted spectrum.. etc..
  • Currently has Spline (CAFana style), Covariance (SBNfit style) and MFA (2D spline to-be-implemented) options
• PROsc : The model class (not the lack of second 'o'). Currently is a very simple hardcoded 3+1 SBL approximation. -- End goal to implement custom classes, NuSquids..etc..
• PROchi : Class that gathers the MC (PROpeller), Systematics (PROsyst) and model (PROsc) and forms a function calculating a chi^2 that can be minimized over

Some Possible PRO class names available

PROtest, PROfessional, PROspect, PROton, PROpane, PROcrastiation, PRO.file, PROtect, PROcreate, PROduce, PROgress, PROfane, PROfanity, PROlapse, PROtocoll