Implement upstream changes from scikit-image 0.23 (part 1 of 2) #727
Conversation
This file contains 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
grlee77
added
improvement
also attempt broadcast_to of provided alpha value
use __init__.pyi
* replace remove_kwarg and deprecate_kwarg with deprecate_parameter * rename gaussian parameter output->out
* add new intensity_std property * fix bug in cache
* now warns if any image contained zero intensity everywhere
grlee77
force-pushed
the
skimage023_updates
branch
from
c41fe3b
to
17110a6
Compare
* need check based on CuPy, not SciPy
|
/merge |
rapids-bot bot
pushed a commit
that referenced
this pull request
May 16, 2024
…ology) (#728) There were a number of non-trivial changes to the morphology module, so I broke those out from the other changes in #727. Please review and merge that MR first before reviewing this one. Highlights from upstream are: - binary morphology functions have a new `mode` argument that controls how values outside the image boundaries are interpreted - grayscale morphology functions have new `mode` and `cval` arguments that control how boundaries are extended (these were already available in `scipy.ndimage`/`cupyx.scipy.ndimage`, they just weren't exposed via the `skimage`/`cuCIM` APIs) - binary and grayscale morphology functions have bug fixes in the case of even-sized/non-symmetric footprints - additional corresponding test cases were added Aside from the upstream changes, novel changes in this MR are: - refactored utility functions to mirror and pad the footprints to allow use with the cuCIM-specific optimization of passing a tuple for rectangular footprints instead of explicitly allocating a GPU footprint array - refactored some test cases to better use `pytest.mark.parametrize` - some grayscale tests now compare directly to `skimage` CPU outputs instead fetching previously saved values - bumped our version pinning for scikit-image to allow 0.23.x to be installed I marked as "non-breaking" as the existing behavior has not changed except in the case of the bug fixes for even-sized footprints. Authors: - Gregory Lee (https://github.com/grlee77) Approvers: - Gigon Bae (https://github.com/gigony) - Jake Awe (https://github.com/AyodeAwe) URL: #728
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.
This MR updates cuCIM for consistency with version scikit-image
0.23.2rc1. There were a relatively large number of changes in themorphologymodule, so I am going to submit those in a subsequent MR. I will also bump the scikit-image version pinning as part of that second MR.The commits in this MR can be reviewed individually. Much of the new code is private functions in
cucim.skimage._sharedthat are copied from upstream and don't need to be reviewed in detail.The highlights are:
cucim.skimage.measure.regionprops(andregionprops_table) support one new region property:intensity_stdcucim.skimage.segmentation.expand_labelsnow supports aspacingkeyword argument to take a pixel's physical dimensions into account.__init__.pyiinstead of just__init__.py_shared/compat.pyfor future NumPy 2.0/SciPy compatibilityassert_stacklevelcucim.skimage.filters.gaussian