This release is a patch to release 2.0.0. It includes :

• Python package version number fix to 2.0.1.
• Add numba as an explicit requirement in requirements.txt (FastWalkthrough have a numba dependance).
• Create create_conda_environment.sh to make installation of Acorn easier for users.
• CI pipeline improvement:New test pipeline using docker-in-docker workflow. Extend tests with InteractionGNN2 model. Factorize the code so that more models can be tested easily.
• Update of the eval plotting functions (in acorn/eval_utils.py) with the new naming scheme.

Important note: This release has been patched in release 2.0.1. Please use this patched version preferentially.

This release includes:

• Compatibility fixes for the latest version of the ATHENA RDO files dumper.
• Updates to the Dockerfile and image build in CI. Updates to tests in CI.
• PyModuleMap updates: integration of new module map features (statistical method selection for geometric cuts) and numerical compatibility with ModuleMapGraph.
• New naming scheme: The new naming scheme adds a prefix of either hit_, edge_ or track_ to each variable name in order to accurately reflect the variable type. See a complete example and instructions here. Backward compatibility is assured for the moment but please switch to the new naming scheme in the config files for new studies. The old naming scheme is deprecated.
• Fast Walkthrough: an optimized version of the Walkthrough algorithm, currently only running on CPU. Refer to D. Murnane, "High Performance Graph Segmentation for ATLAS GNN Track Reconstruction.", CHEP24, for more details.

This release contains a complete example (CTD2023 Example) to reproduce the results presented by the GNN4ITK team at CTD2023.