Skip to content
forked from angerami/MLTree

Athena package to save cell images of clusters for ML training

Notifications You must be signed in to change notification settings

mswiatlo/MLTree

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

41 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MLTree

ATLAS Athena package to save calorimeter clusters as images, with normalized cell energies as pixel values. Six images are saved for each cluster, corresponding to the barrels layers of the EM (EMB1, EMB2, EMB3) and HAD calorimeters (TileBar0, TileBar2, TileBar3). The image size is 0.4x0.4 in eta-phi space. Images are generated from ESD (Event Summary Data).

For questions please contact: joakim.olsson[at]cern.ch

Setup

Using release 20
mkdir MLTreeAthenaAnalysis; cd MLTreeAthenaAnalysis
git clone https://github.com/jmrolsson/MLTree.git 
setupATLAS
#asetup 20.7.7.9,AtlasProduction,here
asetup 20.1.0.3,AtlasProduction,here
lsetup panda
cmt find_packages && cmt compile

Using release 21

Follow instructions for sparse checkout; this is a clunky way to get the athena/Projects directory structure.

mkdir MLTreeAthenaAnalysis; cd MLTreeAthenaAnalysis
setupATLAS
lsetup git
git atlas init-workdir https://:@gitlab.cern.ch:8443/atlas/athena.git

Clone this git repository and create a package filter so athena knows to compile it

git clone https://github.com/jmrolsson/MLTree.git athena/MLTree
echo "+ MLTree" > package_filters.txt
echo "- .*" >> package_filters.txt

Now setup for an out-of-source build

mkdir build; cd build
asetup 21.3,latest,Athena
cmake -DATLAS_PACKAGE_FILTER_FILE=../package_filters.txt ../athena/Projects/WorkDir
make

Test run

This requires that input test files exist, which is specified in MLTreeMaker.py

mkdir run; cd run
athena MLTree/MLTreeMaker.py

Running on the grid

A script for launching grid jobs with different input files is available here:

python python/launch_job.py --user <user> 

About

Athena package to save cell images of clusters for ML training

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages

  • C++ 87.7%
  • Python 8.8%
  • CMake 3.4%
  • Makefile 0.1%