Skip to content

PandoraPFA/Documentation

BranchesTags

Repository files navigation

Pandora build instructions / suggestions

As with any multi-library application, there are a number of different ways in which one can choose to structure and build your Pandora application.

A few detailed examples are presented below, working with either the CMake files or the simple Makefiles provided alongside the Pandora source code.

  1. Using CMake and the PandoraPFA metadata package
  2. Using CMake for each individual package
  3. Using simple Makefiles and the PandoraPFA metadata package
  4. Using simple Makefiles for each individual package

This document describes how to build the core Pandora packages (PandoraSDK and PandoraMonitoring), the library of LAr TPC algorithms and tools (LArContent) and a simple client application for developing algorithms in the Pandora LAr TPC standalone environment (LArReco).

If it is at all unclear how to adapt these instructions to your circumstances, please get in touch: john.marshall AT warwick.ac.uk

Build notes

  1. These instructions have been tested with: -CentOS Linux release 7.7.1908, gcc 8.2.0, ROOT 6.18.04 -macOS Catalina, 10.15.5, Apple LLVM version 10.0.0 (clang-1000.11.45.5), ROOT 6.18.99

  2. Please note that some configurations of the LArReco application now require access to files maintained in the PandoraPFA/LArMachineLearningData repository. Users requiring these files should clone this repository, checkout the tag matching the PANDORA_LAR_RECO_VERSION below, run the download.sh script (see examples below) then add the directory path to the colon-separated list stored in the FW_SEARCH_PATH environment variable.

  3. c++17 is now the default option for most Pandora packages, and is mandatory for some. In cases where it is not yet mandatory, using a version of e.g. ROOT built when demanding a more recent standard may still force you to demand that standard. This may then require specification of e.g. the CMake arguments -DCMAKE_CXX_FLAGS=-std=c++17 or similar.

  4. Using CMake, the user may find that the identified C and C++ compilers revert to the system default compilers. If the intention is to pick-up alternative compilers, either specify CC and CXX environment variables or provide the additional CMake arguments: -DCMAKE_C_COMPILER=/your/cc/path and -DCMAKE_CXX_COMPILER=/your/c++/path

  5. PandoraMonitoring functionality is optional, but its usage is assumed throughout the following examples. It is assumed that the user has already built ROOT, including the EVE libraries. The environment variable ROOTSYS is used below and may need to be specified/replaced carefully on some systems.

  6. LArContent v03_23_00 adds process information to LArMCParticles by default, for pndr files created with earlier versions of LArContent it is necessary to specify <LArMCParticleVersion>1</LArMCParticleVersion> in the LArEventReading algorithm XML configuration. For pndr files created since v03_23_00, this version number should be removed, or updated to version 2.

Deep Learning updates

LArContent v04_00_00 introduces vertex finding via a deep neural network. The DUNEFD horizontal drift beam configuration makes use of this feature by default. This adds a requirement for LibTorch for this use case, necessitating updates to the build. In particular, access to a LibTorch installation is required. For the current use case this can be achieved via

source /cvmfs/dune.opensciencegrid.org/products/dune/setup_dune.sh
setup libtorch v1_6_0d -q e20

Alternatively, you can install LibTorch locally. The cmake commands must then be updated, in particular, the LArReco and LArContent (or PandoraPFA in the version 1 build below) cmake commands should now each include

-DPANDORA_LIBTORCH=ON

along with corresponding additions to the CMAKE_PREFIX_PATH for LArContent

-DCMAKE_PREFIX_PATH=/cvmfs/larsoft.opensciencegrid.org/products/libtorch/v1_6_0d/Linux64bit+3.10-2.17-e20/share/cmake

and finally the LArReco cmake command should also include

-DLArDLContent_DIR=$MY_TEST_AREA/LArContent/

Recommended library/application versions

Use 'git tag' to check the list of available tags. Current recommended versions are as defined below:

export PANDORA_PFA_VERSION=v04-11-02
export PANDORA_SDK_VERSION=v03-04-02
export PANDORA_MONITORING_VERSION=v03-06-00
export PANDORA_LAR_CONTENT_VERSION=v04_11_02
export PANDORA_LC_CONTENT_VERSION=v03-01-06
export PANDORA_EXAMPLE_CONTENT_VERSION=v03-01-00
export PANDORA_LAR_RECO_VERSION=v04-11-02
export PANDORA_LC_RECO_VERSION=v03-01-05

export MY_TEST_AREA=/path/to/your/test/area

1. Using CMake and the PandoraPFA metadata package

cd $MY_TEST_AREA
git clone /~https://github.com/PandoraPFA/PandoraPFA.git
cd PandoraPFA
git checkout $PANDORA_PFA_VERSION # select your chosen tag, defining all daughter library tags
mkdir build
cd build
cmake -DCMAKE_MODULE_PATH=$ROOTSYS/etc/cmake -DPANDORA_MONITORING=ON -DPANDORA_EXAMPLE_CONTENT=OFF -DPANDORA_LAR_CONTENT=ON -DPANDORA_LC_CONTENT=OFF ..
make -j4 install

cd $MY_TEST_AREA
git clone /~https://github.com/PandoraPFA/LArReco.git
cd LArReco
git checkout $PANDORA_LAR_RECO_VERSION
mkdir build
cd build
cmake -DCMAKE_MODULE_PATH="$MY_TEST_AREA/PandoraPFA/cmakemodules;$ROOTSYS/etc/cmake" -DPANDORA_MONITORING=ON -DPandoraSDK_DIR=$MY_TEST_AREA/PandoraPFA/ -DPandoraMonitoring_DIR=$MY_TEST_AREA/PandoraPFA/ -DLArContent_DIR=$MY_TEST_AREA/PandoraPFA/ ..
make -j4 install

$MY_TEST_AREA/LArReco/bin/PandoraInterface -h

./bin/PandoraInterface
    -r RecoOption          (required) [Full, AllHitsCR, AllHitsNu, CRRemHitsSliceCR, CRRemHitsSliceNu, AllHitsSliceCR, AllHitsSliceNu]
    -i Settings            (required) [algorithm description: xml]
    -e EventFileList       (optional) [colon-separated list of files: xml/pndr]
    -g GeometryFile        (optional) [detector geometry description: xml/pndr]
    -n NEventsToProcess    (optional) [no. of events to process]
    -s NEventsToSkip       (optional) [no. of events to skip in first file]
    -p                     (optional) [print status]
    -N                     (optional) [print event numbers]

Note: In this configuration, you will need to read in events via e.g. the EventReading algorithm, configured in the specified settings file.

2. Using CMake for each individual package

cd $MY_TEST_AREA
git clone /~https://github.com/PandoraPFA/PandoraPFA.git # still need this for the CMake modules
cd PandoraPFA
git checkout $PANDORA_PFA_VERSION

cd $MY_TEST_AREA
git clone /~https://github.com/PandoraPFA/PandoraSDK.git
cd PandoraSDK
git checkout $PANDORA_SDK_VERSION # now need to select all package tags manually
mkdir build
cd build
cmake -DCMAKE_MODULE_PATH=$MY_TEST_AREA/PandoraPFA/cmakemodules ..
make -j4 install

cd $MY_TEST_AREA
git clone /~https://github.com/PandoraPFA/PandoraMonitoring.git
cd PandoraMonitoring
git checkout $PANDORA_MONITORING_VERSION
mkdir build
cd build
cmake -DCMAKE_MODULE_PATH="$MY_TEST_AREA/PandoraPFA/cmakemodules;$ROOTSYS/etc/cmake" -DPandoraSDK_DIR=$MY_TEST_AREA/PandoraSDK ..
make -j4 install

cd $MY_TEST_AREA
wget https://gitlab.com/libeigen/eigen/-/archive/3.4.0/eigen-3.4.0.tar.gz
tar -xf eigen-3.4.0.tar.gz
mv eigen-3.4.0 Eigen3
cd Eigen3
mkdir build
cd build
cmake -DCMAKE_INSTALL_PREFIX=$MY_TEST_AREA/Eigen3/ ..
make -j4 install

cd $MY_TEST_AREA
git clone /~https://github.com/PandoraPFA/LArContent.git
cd LArContent
git checkout $PANDORA_LAR_CONTENT_VERSION
mkdir build
cd build
cmake -DCMAKE_MODULE_PATH="$MY_TEST_AREA/PandoraPFA/cmakemodules;$ROOTSYS/etc/cmake" -DPANDORA_MONITORING=ON -DPandoraSDK_DIR=$MY_TEST_AREA/PandoraSDK -DPandoraMonitoring_DIR=$MY_TEST_AREA/PandoraMonitoring -DEigen3_DIR=$MY_TEST_AREA/Eigen3/share/eigen3/cmake/ ..
make -j4 install

cd $MY_TEST_AREA
git clone /~https://github.com/PandoraPFA/LArReco.git
cd LArReco
git checkout $PANDORA_LAR_RECO_VERSION
mkdir build
cd build
cmake -DCMAKE_MODULE_PATH="$MY_TEST_AREA/PandoraPFA/cmakemodules;$ROOTSYS/etc/cmake" -DPANDORA_MONITORING=ON -DPandoraSDK_DIR=$MY_TEST_AREA/PandoraSDK/ -DPandoraMonitoring_DIR=$MY_TEST_AREA/PandoraMonitoring/ -DLArContent_DIR=$MY_TEST_AREA/LArContent/ ..
make -j4 install

### LArMachineLearningData: If your configuration depends on LArMachineLearningData
cd $MY_TEST_AREA
git clone /~https://github.com/PandoraPFA/LArMachineLearningData.git
cd LArMachineLearningData
git checkout $PANDORA_LAR_RECO_VERSION
source download.sh <experiment>
## End LArMachineLearningData

$MY_TEST_AREA/LArReco/bin/PandoraInterface -h # as for example 1.

3. Using simple Makefiles and the PandoraPFA metadata package

cd $MY_TEST_AREA
git clone /~https://github.com/PandoraPFA/PandoraPFA.git
cd PandoraPFA
git checkout $PANDORA_PFA_VERSION

cd $MY_TEST_AREA/PandoraPFA
git clone /~https://github.com/PandoraPFA/PandoraSDK.git PandoraSDK
cd PandoraSDK
git checkout $PANDORA_SDK_VERSION # without cmake, need to select all package tags manually

cd $MY_TEST_AREA/PandoraPFA
git clone /~https://github.com/PandoraPFA/PandoraMonitoring.git PandoraMonitoring
cd PandoraMonitoring
git checkout $PANDORA_MONITORING_VERSION

cd $MY_TEST_AREA/PandoraPFA
wget https://gitlab.com/libeigen/eigen/-/archive/3.4.0/eigen-3.4.0.tar.gz
tar -xf eigen-3.4.0.tar.gz
mv eigen-3.4.0 Eigen3

cd $MY_TEST_AREA/PandoraPFA
git clone /~https://github.com/PandoraPFA/LArContent.git LArContent
cd LArContent
git checkout $PANDORA_LAR_CONTENT_VERSION

cd $MY_TEST_AREA/PandoraPFA
mkdir include lib
make -j4 MONITORING=1 PANDORA_DIR=$MY_TEST_AREA/PandoraPFA
make install MONITORING=1 PANDORA_DIR=$MY_TEST_AREA/PandoraPFA INCLUDE_TARGET=$MY_TEST_AREA/PandoraPFA/include # optional

cd $MY_TEST_AREA
git clone /~https://github.com/PandoraPFA/LArReco.git
cd LArReco
git checkout $PANDORA_LAR_RECO_VERSION
mkdir bin
make -j4 MONITORING=1 PROJECT_DIR=$MY_TEST_AREA/LArReco PANDORA_DIR=$MY_TEST_AREA/PandoraPFA

### LArMachineLearningData: If your configuration depends on LArMachineLearningData
cd $MY_TEST_AREA
git clone /~https://github.com/PandoraPFA/LArMachineLearningData.git
cd LArMachineLearningData
git checkout $PANDORA_LAR_RECO_VERSION
source download.sh <experiment>
## End LArMachineLearningData

export LD_LIBRARY_PATH=$MY_TEST_AREA/PandoraPFA/lib:$LD_LIBRARY_PATH

$MY_TEST_AREA/LArReco/bin/PandoraInterface -h # as for example 1.

4. Using simple Makefiles for each individual package

cd $MY_TEST_AREA
mkdir lib
export LD_LIBRARY_PATH=$MY_TEST_AREA/lib:$LD_LIBRARY_PATH

git clone /~https://github.com/PandoraPFA/PandoraSDK.git PandoraSDK
cd PandoraSDK
git checkout $PANDORA_SDK_VERSION # without cmake, need to select all package tags manually
mkdir lib
make -j4 PROJECT_DIR=$MY_TEST_AREA/PandoraSDK
make install PROJECT_DIR=$MY_TEST_AREA/PandoraSDK LIB_TARGET=$MY_TEST_AREA/lib

cd $MY_TEST_AREA
git clone /~https://github.com/PandoraPFA/PandoraMonitoring.git
cd PandoraMonitoring
git checkout $PANDORA_MONITORING_VERSION
mkdir lib
make -j4 MONITORING=1 PROJECT_DIR=$MY_TEST_AREA/PandoraMonitoring PANDORA_DIR=$MY_TEST_AREA
make install MONITORING=1 PROJECT_DIR=$MY_TEST_AREA/PandoraMonitoring PANDORA_DIR=$MY_TEST_AREA LIB_TARGET=$MY_TEST_AREA/lib

cd $MY_TEST_AREA
wget https://gitlab.com/libeigen/eigen/-/archive/3.4.0/eigen-3.4.0.tar.gz
tar -xf eigen-3.4.0.tar.gz
mv eigen-3.4.0 Eigen3

cd $MY_TEST_AREA
git clone /~https://github.com/PandoraPFA/LArContent.git
cd LArContent
git checkout $PANDORA_LAR_CONTENT_VERSION
mkdir lib
make -j4 MONITORING=1 PROJECT_DIR=$MY_TEST_AREA/LArContent PANDORA_DIR=$MY_TEST_AREA EIGEN_INC=$MY_TEST_AREA/Eigen3/
make install MONITORING=1 PROJECT_DIR=$MY_TEST_AREA/LArContent PANDORA_DIR=$MY_TEST_AREA LIB_TARGET=$MY_TEST_AREA/lib

cd $MY_TEST_AREA
git clone /~https://github.com/PandoraPFA/LArReco.git
cd LArReco
git checkout $PANDORA_LAR_RECO_VERSION
mkdir bin
make -j4 MONITORING=1 PROJECT_DIR=$MY_TEST_AREA/LArReco PANDORA_DIR=$MY_TEST_AREA

### LArMachineLearningData: If your configuration depends on LArMachineLearningData
cd $MY_TEST_AREA
git clone /~https://github.com/PandoraPFA/LArMachineLearningData.git
cd LArMachineLearningData
git checkout $PANDORA_LAR_RECO_VERSION
source download.sh <experiment>
## End LArMachineLearningData

$MY_TEST_AREA/LArReco/bin/PandoraInterface -h # as for example 1.

About

Useful documents describing the Pandora project

Resources

Readme
Activity

Stars

4 stars

Watchers

11 watching

Forks

9 forks
Report repository

Releases

16 tags

Packages

No packages published

Contributors 7

Languages