v2.51.0

[kba]

Changed:

• core cuda Docker: CUDA base image working again, based on ocrd/core not nvidia/cuda in a separate Dockerfile.cuda, Reduce ocrd/core-cuda #1041
• core-cuda Docker: adopt Dockerfile: Use virtual environment #1008 (venv under /usr/local, as in ocrd_all, instead of dist-packages), Reduce ocrd/core-cuda #1041
• core-cuda Docker: use conda (micromamba) for CUDA toolkit, and nvidia-pyindex for CUDA libs – instead of nvidia/cuda base image, Reduce ocrd/core-cuda #1041
• more robust workaround for shapely#1598, Reduce ocrd/core-cuda #1041

Removed:

• Revert No more pkg_resources #882 (fastentrypoints) as it enforces deps versions at runtime
• Drop ocrd_utils.package_resources and use pkg_resources.* directly, Reduce ocrd/core-cuda #1041
• ocrd resmgr: Drop redundant (processor-provided) entries in the central resource_list.yml.

---

This discussion was created from the release v2.51.0.