Data format and image processing tools, routines, and framework for LAr TPC-derived images. Developed as bridge between LArSoft and deep learning frameworks, e.g. PyTorch, Caffe, Tensorflow.
We originally developed for MicroBooNE, but are now using this library across experiments.
MicroBooNE specific code has been moved into a new repository, ublarcvapp
dependent on this library and larlite
.
One recent big change is that the assumed row order is now in postive time order (same as LArCV2). We, however, are attempting to maintain the ability to read old "tickbackward" files created for MicroBooNE. When reading tickbackward images, the data is flipped along the row axis in order to be treated as "tickforward" data.
Planned features
Recently completed features
Dependencies to build with are determined through the presence of environment variables or executables in your PATH:
thisroot.sh
script.clone the repository
git clone https://github.com/LArbys/LArCV.git
go into the LArCV directory
run the build configuration script
source configure.sh
make a build folder somewhere, e.g. in the same folder as this README.
mkdir build
go into the folder and run cmake
cd build
cmake -DUSE_PYTHON3=ON ../
(If you require python2
, probably deprecated in the future, you can use the cmake flag -DUSE_PYTHON2=ON
instead.)
then make and install the code
make install
The output of make install
will be libaries and headers in build/installed/
.
A cmake config file is provided in build/installed/lib/larcv
in case you want to incorporate the library into other projects.
Checkout the Wiki for notes on using this code.