Repository navigation
Keep geometry and mask allocation on the native array device - #8
Merged
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Mixed image-channel composition and image masks can now reshape, broadcast and allocate ones through
MemoryType, retaining the actual input/reference device instead of making an implicit NumPy roundtrip.ArrayOperationsowns shared algorithms through a nominal ABC and framework subclasses; its existing declaration constants and conversion/scaling semantics remain intact.Pyclesperanto explicitly rejects changed-shape reshape/broadcast and boolean allocation rather than silently downloading pixels. Its documented reshape implementation downloads pixels; native supported-dtype allocation uses create/set on the reference device.
Validation: 111 affected existing/new controls passed (13 unavailable-framework skips), Ruff passed, four native-geometry controls passed on a real GTX1080/CuPy environment, and JAX geometry/reference allocation passed on its installed CPU device. OpenHCS production composition receiving additionally passed mixed intensity, channels, full/channel-free/absent/shared masks and NumPy/CuPy destinations, with GPU downloads forbidden during GPU-only composition. Torch/TensorFlow/Pycles actual-device execution is not claimed. This is a residency/correctness improvement, not a claimed CPU benchmark speedup or full mixed-device pipeline profile.
Release 0.3.7 uses the repository version helper. Original release readiness gates passed 7/7 (real build and Twine metadata validation) in isolated publishing tooling; runtime benchmark dependencies were not modified. Full 239 controls passed (17 unavailable-framework skips), source/scripts Black and mypy passed, full Ruff passed after repairing one existing test import order. Sphinx warning-as-error documentation build passed after fixing two existing release-page title/toctree omissions. The existing trusted tag publisher will publish after main merge; an actual PyPI result remains required.
Closes #7