Skip to content

Releases: DeepLearnPhysics/larcv3

Bug Fix patch

06 Jan 16:33
Compare
Choose a tag to compare

This release is very similar to 3.2.0 with some minor fixes. Compiler warnings were addressed, and the linking error on Linux is also addressed as far as I can tell.

v3.2.0: Tag larcv3 version 3.2.0, which has removed entirely the c++ API from…

12 Nov 15:48
Compare
Choose a tag to compare
… hdf5.  This ought to make larcv easier to install

v3.1.0

12 Sep 16:54
Compare
Choose a tag to compare

This is an incremental update to larcv3. It includes queue loader and many enhancements to the MPI loading compatibility. There are some very minor API changes that break compatibility of larch v3.0.0, but they are of the form of changing import files in python. larcv_interface needs to be replaced with thread loader, and the python object is now called thread_interface.

Tests for queue loader are included, but the data format is unchanged and files are 100% compatible with v3.0.0. There are no plans to modify the serialization API or data format.

larcv 3.0.0

26 Jun 19:48
05b95d9
Compare
Choose a tag to compare

This is the first official tag of larcv3.0.0. Compared to larcv2, this release has:

  • pip and make support
  • hdf5 instead of ROOT for serialization
  • swig instead of ROOT for python bindings
  • Reorganization of installation procedure to allow pip install, setup.py installation.
  • no need to source additional setup scripts or set env variables to use larcv once installed to a standard python location.
  • A full suite of tests to evaluate python bindings, numpy conversions, data format serialization and deserialization, threaded IO functionality, distributed IO functionality.
  • A complete re-write of the IO models for sparse data products to unify the 2D/3D interfaces and ImageMeta objects.

One missing piece, hopefully in the next release, is the finished porting of many image mod apps. Additionally, it is on the to-do list to switch to json for the configuration language in larcv. Neither of these are finished in this release.