This is an alternative implementation of the RecoZoR algorithm currently used in LiquidO that doesn't require ROOT as a dependency, exists as a standalone tool separate to other LiquidO analysis software, and also can be run on multiple cores. Combining the improved single thread performance with multi-core processing, RecoMore can very easily be 1-2 orders of magnitude faster than the original software.

The dependencies are all easily installable on OSX and Linux systems, however Ceres-solver is not as easy to install for Windows. The current workaround I'm using is to install WSL2 and then run RecoMore on there - note that if try to run RecoMore on data that is outside of your WSL virtual disk it will be very slow.

Dependencies:

• Boost
• Ceres-solver (≥ 2.1)
• CMake
• pybind11 (for CReader)
• pybind11-stubgen (for CReader type hints)

sudo apt-get update
sudo apt-get install -y libceres-dev

sudo apt-get update
sudo apt-get install -y libceres-dev

sudo dnf install ceres-solver

sudo pacman -S ceres-solver

For openSUSE, you can install Ceres Solver directly from the Science repository:

sudo zypper addrepo https://download.opensuse.org/repositories/science/openSUSE_Tumbleweed/science.repo
sudo zypper refresh
sudo zypper install ceres-solver

sudo zypper addrepo https://download.opensuse.org/repositories/science/openSUSE_Factory_PowerPC/science.repo
sudo zypper refresh
sudo zypper install ceres-solver

sudo zypper addrepo https://download.opensuse.org/repositories/science/15.6/science.repo
sudo zypper refresh
sudo zypper install ceres-solver

brew install ceres-solver

For other Linux distributions, you may need to build Ceres Solver from source.
Please refer to the official installation guide for detailed instructions.

To build RecoMore:

• Make a build directory
• cd into build directory
• cmake -DCMAKE_BUILD_TYPE=Release <pathToProjectDir>
• make

To run RecoMore:

./PEFinder -i <pathToDataFile> --template_dir <pathToIdealPDFsDir> --n_threads <numberOFThreads> --n_batches <numberOfBatches>

For more info run ./PEFinder -h to get the help output.

Note: Make sure to include the trailing slash of the path to the PDF dir

CReader is a python wrapper around C++ file parsers to combine high performance with easy usage. To initialise pybind11 simply run git submodule update --init from the repo root directory.

To build and install CReader to be a usable python module:

• Navigate to the CReader_ directory.
• Make sure you are in the same python environment that you intend to run CReader in.
• python3 setup.py build_ext --inplace install
• To rebuild the stubs (used for IDE type hints etc):

  pybind11-stubgen CReader --output-dir="./stubs/generated" --root-module-suffix="" --ignore-invalid=all --no-setup-py
  

Import with from CReader import *

DebugUtils contains some utilities for debugging/analysing RecoMore output along with some examples of using CReader.