Release v4.0.0: scikit-build-core packaging, CI gate, accuracy suites - #9
Merged
Conversation
…OpenMP DLL; add Python README
…deps, pip packages and environment prefix
…es and invalid kernel/copy cases; use cudaDeviceSynchronize in debug check
…thon module output and remove libomp DLL copy
…tion backlog Reorder the roadmap around the agreed sequence: (1) conda/mamba/pixi/uv packaging, (2) vcpkg/conan, (3) CPU fallback with warning + strict flag, (4) vRAM-resident filter pipelining with host-memory taps, (5) prioritized operation backlog. Document the develop-branch PR workflow and parallel lanes. Move C ABI/C#/NuGet, non-NVIDIA GPU, and GPU-direct I/O to later phases. Seed BACKLOG.md with a re-arrangeable priority table covering compositions, local kernels, the FFT track, and global ops with their chunking caveats. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Unify the Python build on scikit-build-core so a single `pip install .` drives CMake, compiles the Hydra CUDA extension into the hydra_image_processor package, and installs it. This is the same path used locally, by CI, and by the conda-forge feedstock, which lets the feedstock drop its source patches. - Move pyproject.toml to the repo root; version 4.0.0 is the single source of truth, and __version__ is read from installed metadata. - CMake: raise CUDA_STANDARD to 17, drop the hardcoded GPU architectures in favor of CMAKE_CUDA_ARCHITECTURES (default all-major), add a HYDRA_BUILD_TESTS option, and add package-relative install rules for the extension and shims. - Ship top-level HIP/Hydra compatibility shims so legacy `import HIP` / `import Hydra` keep working. - Remove the superseded in-repo conda-build recipe and MANIFEST.in; ignore build artifacts. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Replace the Windows-only conda-build workflow (which uploaded to a separate anaconda.org channel) with two workflows: - ci.yml: build and import the package on Windows and Linux for PRs and pushes, catching breakage before it reaches the conda-forge feedstock. Runners have no GPU, so GPU tests auto-skip; this only verifies the package builds and imports. - release.yml: on a `v*` tag, verify the tag matches pyproject's version and create a GitHub Release, which the conda-forge autotick bot picks up to open a feedstock version-bump PR. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- README: conda/mamba/pixi install, the NVIDIA GPU requirement, the unsupported pip/uv path, and the import convention. - Rewrite the Python package README for the new hydra_image_processor API and the HIP/Hydra back-compat shims. - Add CHANGELOG.md (4.0.0) and a build + release checklist in CONTRIBUTING.md. - Refresh GEMINI.md to reflect scikit-build-core and conda-forge distribution. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A standing pre-release verification checklist (TESTING.md) covering the local build, GPU accuracy suite, CI gate, feedstock dry-run, and release rehearsal. The GPU correctness steps must be run locally since CI is build-only. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The coordinate grid was shifted by vol_size/2 + 0.5, putting the ellipsoid center one voxel off (e.g. -6..+4 instead of -5..+5 for radius 5). Every morphological op using the mask diverged from the C++ createEllipsoidKernel / MATLAB results. With the fix, all eight ellipsoid-kernel ops match the MATLAB-generated ground truth exactly. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
test_accuracy.cpp never compiled: it called a nonexistent ImageOwner::getLinearAddress and passed a float[3] where the generated CudaCall_Gaussian::run stub takes Vec<double> plus an lvalue ImageView for the output. The Gaussian energy-preservation check also assumed a volume too shallow for its sigma; the filter renormalizes truncated kernels at borders, so the impulse support must stay interior for the sum to be preserved. enable_testing() now runs at the top level so `ctest --test-dir build` actually finds CppAccuracyTest. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Port of src/MATLAB/+Test/AccuracyTest.m: 15 operations compared against the ground-truth TIFFs in test_data/ (integer results within 1 LSB, float within 1e-4, EntropyFilter as a soft check since its ground truth was saved min-max scaled). Documents the numpy<->Hydra axis convention: arrays pass as tifffile loads them (chan, z, y, x) with vector parameters in (z, y, x) order. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The pinned Jimver/cuda-toolkit network install of 12.6.2 broke as the runner images moved (apt failure on ubuntu-latest). Pull cuda-nvcc + cudart from conda-forge in the existing miniforge env instead - the same channel the feedstock builds against, and one less third-party action. Linux additionally gets conda-forge c/cxx compilers as the nvcc host compiler. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The setup-miniconda-created 'test' env resolves fine right after creation but later steps on windows-latest fail activation with EnvironmentNameNotFound (seen across two runs). Installing the toolchain, python included, into base sidesteps named-env activation entirely. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Without a generator override, CMake on windows-latest picks Visual Studio, whose CUDA support needs the VS-integration files that only a system CUDA toolkit install provides - the conda-forge cuda-nvcc package does not ship them, so configure fails on the runner. Verified locally: Ninja + conda-forge nvcc 12.9 + MSVC 19.44 builds the wheel end to end. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Job logs need authentication to read, but annotations are public; on failure the build step now emits the toolchain locations and the first error lines from the pip log as ::error:: annotations. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
windows-latest now ships Visual Studio 2026 (MSVC 14.51) and CUDA 12.9 nvcc refuses it outright (C1189: unsupported Microsoft Visual Studio version). nvcc 13.x supports current MSVC. The 12.x pairing stays covered by the conda-forge feedstock CI, which pins VS 2022, and was verified locally against MSVC 19.44. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The command framework headers relied on transitive <cstdint> includes, which newer libstdc++ (pulled in by the CUDA 13 / gcc toolchain on CI) no longer provides - uint8_t and friends came up undefined in ScriptioMaps.h. Include it explicitly in every header and TU that names the fixed-width types. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
![sourcery-ai[bot]](https://avatars.githubusercontent.com/in/48477?s=60&v=4){kind=small}
There was a problem hiding this comment.
Sorry @ericwait, your pull request is larger than the review limit of 150000 diff characters
|
Important Review skippedToo many files! This PR contains 163 files, which is 13 over the limit of 150. To get a review, narrow the scope: Upgrade to a paid plan to raise the limit. ⚙️ Run configurationConfiguration used: Organization UI Review profile: ASSERTIVE Plan: Pro Run ID: ⛔ Files ignored due to path filters (31)
📒 Files selected for processing (163)
You can disable this status message by setting the Use the checkbox below for a quick retry:
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
There was a problem hiding this comment.
Pull request overview
This PR fast-forwards main to a conda-forge-oriented v4.0.0 release state, centered on scikit-build-core packaging for the CUDA Python extension, CI/release gating, and restored/added GPU accuracy suites (C++ + Python TIFF) backed by Git LFS test assets.
Changes:
- Reworks the Python build/install path around
pyproject.toml+ scikit-build-core, installing the compiledHydraextension into thehydra_image_processorpackage and addingHIP/Hydralegacy import shims. - Adds CI build+import gates (Linux/Windows) and a tag-driven release workflow that enforces tag/version parity.
- Adds/repairs accuracy test coverage and associated LFS-tracked ground-truth TIFFs + documentation for release testing.
Reviewed changes
Copilot reviewed 193 out of 194 changed files in this pull request and generated 10 comments.
Show a summary per file
| File | Description |
|---|---|
| TESTING.md | Adds a release testing checklist, including local GPU accuracy gating steps. |
| test_data/VarFilter_c2_5-5-2.tif | Adds LFS-tracked ground-truth TIFF for Python/MATLAB accuracy suite. |
| test_data/VarFilter_c1_5-5-2.tif | Adds LFS-tracked ground-truth TIFF for Python/MATLAB accuracy suite. |
| test_data/test_c2_.tif | Adds LFS-tracked input TIFF (test volume channel). |
| test_data/test_c1.tif | Adds LFS-tracked input TIFF (test volume channel). |
| test_data/test_c1_.tif | Adds LFS-tracked input TIFF (test volume channel variant). |
| test_data/test_c0.tif | Adds LFS-tracked input TIFF (test volume channel). |
| test_data/StdFilter_c2_5-5-2.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/StdFilter_c1_5-5-2.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/Opener_c2_5-5-2.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/Opener_c1_5-5-2.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/MultiplySum_c2_15-15-6.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/MultiplySum_c1_15-15-6.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/MinFilter_c2_5-5-2.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/MinFilter_c1_5-5-2.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/MedianFilter_c2_5-5-2.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/MedianFilter_c1_5-5-2.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/MeanFilter_c2_5-5-2.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/MeanFilter_c1_5-5-2.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/MaxFilter_c2_5-5-2.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/MaxFilter_c1_5-5-2.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/LoG_c2_4-4-2.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/LoG_c1_4-4-2.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/HighPassFilter_c2_19-19-9.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/HighPassFilter_c1_19-19-9.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/Gaussian_c2_19-19-9.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/Gaussian_c1_19-19-9.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/EntropyFilter_c2_5-5-2.tif | Adds LFS-tracked reference TIFF for EntropyFilter soft-check. |
| test_data/EntropyFilter_c1_5-5-2.tif | Adds LFS-tracked reference TIFF for EntropyFilter soft-check. |
| test_data/ElementWiseDifference_c2_reverse.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/ElementWiseDifference_c1_reverse.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/Closure_c2_5-5-2.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| test_data/Closure_c1_5-5-2.tif | Adds LFS-tracked ground-truth TIFF for accuracy suite. |
| src/Python/Test/test_accuracy.py | Adds a Python TIFF-based GPU accuracy gate mirroring the MATLAB suite. |
| src/Python/test-hydra.py | Adds a standalone smoke-test script for the legacy Hydra module import and basic calls. |
| src/Python/test_package.py | Adds a standalone smoke-test script for the Pythonic hydra_image_processor API. |
| src/Python/test_nonchunking.py | Updates a dev script to import Hydra (legacy shim) instead of HIP. |
| src/Python/test_import.py | Adds a script to validate portable import with bundled OpenMP DLL on Windows. |
| src/Python/README.md | Documents conda-forge installation, usage, and source builds for Python. |
| src/Python/Hydra.py | Adds legacy top-level Hydra import shim to re-export hydra_image_processor.Hydra. |
| src/Python/hydra_image_processor/utils/masks.py | Adds mask/structuring-element helpers (ball/ellipsoid) used by Python API/tests. |
| src/Python/hydra_image_processor/utils/init.py | Exposes mask helpers as part of the public utilities module. |
| src/Python/hydra_image_processor/README.md | Adds package-level README describing the Python wrapper architecture. |
| src/Python/hydra_image_processor/local/init.py | Adds CPU “local” layer API surface (currently placeholder stubs). |
| src/Python/hydra_image_processor/cuda/init.py | Adds CUDA wrapper layer exports for the Python package. |
| src/Python/hydra_image_processor/core.py | Adds GPU-first wrapper functions with fallback plumbing. |
| src/Python/hydra_image_processor/init.py | Adds package public API + metadata version sourcing from installed package metadata. |
| src/Python/HIP.so | Removes legacy LFS-tracked binary extension artifact from the repo. |
| src/Python/HIP.py | Adds legacy top-level HIP import shim to re-export hydra_image_processor.Hydra. |
| src/Python/example_usage.py | Adds an example script demonstrating the Pythonic API and utilities. |
| src/Python/environment.yml | Updates a conda environment definition for Python build/dev tooling. |
| src/Python/download-script.py | Adds a helper script to download GitHub Actions artifacts for local testing. |
| src/MATLAB/+Test/AccuracyTest.m | Adds/repairs MATLAB accuracy test suite aligned with test_data ground truth. |
| src/MATLAB/+Hydra/WienerFilter.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/VarFilter.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/Sum.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/StdFilter.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/Opener.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/NLMeans.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/MultiplySum.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/MinFilter.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/MedianFilter.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/MeanFilter.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/MaxFilter.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/LoG.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/Info.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/IdentityFilter.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/HighPassFilter.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/Help.m | Adds MATLAB help wrapper for command help. |
| src/MATLAB/+Hydra/GetMinMax.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/Gaussian.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/EntropyFilter.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/ElementWiseDifference.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/Closure.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/CheckConfig.m | Adds/updates MATLAB wrapper using CUDA-first with local fallback. |
| src/MATLAB/+Hydra/@Cuda/WienerFilter.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/VarFilter.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/Sum.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/StdFilter.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/Opener.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/NLMeans.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/MultiplySum.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/MinFilter.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/MedianFilter.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/MeanFilter.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/MaxFilter.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/LoG.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/Info.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/IdentityFilter.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/Hydra.mexw64 | Adds LFS-tracked MATLAB mex binary for Windows. |
| src/MATLAB/+Hydra/@Cuda/HighPassFilter.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/Help.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/GetMinMax.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/Gaussian.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/EntropyFilter.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/ElementWiseDifference.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/DeviceStats.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/DeviceCount.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/Cuda.m | Adds/updates MATLAB class interface listing CUDA entrypoints. |
| src/MATLAB/+Hydra/@Cuda/Closure.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/MATLAB/+Hydra/@Cuda/CheckConfig.m | Adds/updates generated MATLAB CUDA wrapper calling the mex dispatcher. |
| src/c/test_back/test.cpp | Adds missing <cstdint> include for newer toolchains. |
| src/c/test_back/test_accuracy.cpp | Adds a minimal C++ accuracy test executable (skips if no CUDA device). |
| src/c/test_back/CMakeLists.txt | Adds test_accuracy target and registers it with CTest. |
| src/c/ScriptCmds/ScriptioMaps.h | Adds explicit <cstdint> include to avoid relying on transitive headers. |
| src/c/ScriptCmds/ScriptCommandImpl.h | Adds explicit <cstdint> include to avoid relying on transitive headers. |
| src/c/ScriptCmds/GenCommands.h | Adds explicit <cstdint> include to avoid relying on transitive headers. |
| src/c/ScriptCmds/Commands/ScrCmdDeviceCount.h | Adds explicit <cstdint> include to avoid relying on transitive headers. |
| src/c/Python/PyIncludes.h | Renames NumPy unique symbol to match the Hydra module naming. |
| src/c/Python/PyArgConverter.h | Updates compile-time error strings from HIP to Hydra naming. |
| src/c/Python/HydraPyModule.cpp | Renames Python extension module from HIP to Hydra. |
| src/c/Python/CMakeLists.txt | Switches to Python3 targets, installs extension into package, installs shims. |
| src/c/Cuda/Kernel.cu | Fixes missing throw when creating offset kernel out of bounds. |
| src/c/Cuda/ImageChunk.h | Fixes missing throw in unsupported copy direction branch. |
| src/c/Cuda/CudaWienerFilter.cuh | Fixes missing throw on sendROI failure. |
| src/c/Cuda/CudaVarFilter.cuh | Fixes missing throw on sendROI failure. |
| src/c/Cuda/CudaUtilities.h | Updates debug sync macro from deprecated cudaThreadSynchronize to cudaDeviceSynchronize. |
| src/c/Cuda/CudaSum.cuh | Fixes missing throw on sendROI failure. |
| src/c/Cuda/CudaStdFilter.cuh | Fixes missing throw on sendROI failure. |
| src/c/Cuda/CudaOpener.cuh | Fixes missing throw on sendROI failure. |
| src/c/Cuda/CudaNLMeans.cuh | Fixes missing throw on sendROI failure. |
| src/c/Cuda/CudaMultiplySum.cuh | Fixes missing throw on sendROI failure. |
| src/c/Cuda/CudaMinMax.cuh | Fixes missing throw on sendROI failure. |
| src/c/Cuda/CudaMinFilter.cuh | Fixes missing throw on sendROI failure. |
| src/c/Cuda/CudaMedianFilter.cuh | Fixes missing throw on sendROI failure. |
| src/c/Cuda/CudaMeanFilter.cuh | Fixes missing throw on sendROI failure. |
| src/c/Cuda/CudaMeanAndVariance.cuh | Fixes missing throw on sendROI failure. |
| src/c/Cuda/CudaMaxFilter.cuh | Fixes missing throw on sendROI failure. |
| src/c/Cuda/CudaLoG.cuh | Fixes missing throw on sendROI failure across LoG stages. |
| src/c/Cuda/CudaIdentityFilter.cuh | Fixes missing throw on sendROI failure. |
| src/c/Cuda/CudaHighPassFilter.cuh | Fixes missing throw on sendROI failure. |
| src/c/Cuda/CudaGaussian.cuh | Fixes missing throw on sendROI failure. |
| src/c/Cuda/CudaEntropyFilter.cuh | Fixes missing throw on sendROI failure. |
| src/c/Cuda/CudaElementWiseDifference.cuh | Fixes missing throw on sendROI failure for either input ROI. |
| src/c/Cuda/CudaClosure.cuh | Fixes missing throw on sendROI failure. |
| src/c/Cuda/CudaAddTwoImages.cuh | Fixes missing throw on sendROI failure for either input ROI. |
| src/c/Cuda/_TemplateKernel.cuh | Fixes missing throw in template kernel example. |
| src/c/Cuda/CMakeLists.txt | Removes hardcoded architectures, adjusts OpenMP flagging, bumps CUDA standard to 17. |
| SECURITY.md | Adds a security policy document. |
| readme.md | Updates top-level README with conda-forge install, usage, and source build guidance. |
| pyproject.toml | Adds scikit-build-core-based Python packaging metadata and build configuration. |
| GEMINI.md | Adds repository guidance/overview document. |
| docs/images/VarFilter_c2_5-5-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/VarFilter_c1_5-5-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/test_c2__brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/test_c1__brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/StdFilter_c2_5-5-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/StdFilter_c1_5-5-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/Opener_c2_5-5-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/Opener_c1_5-5-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/MultiplySum_c2_15-15-6_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/MultiplySum_c1_15-15-6_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/MinFilter_c2_5-5-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/MinFilter_c1_5-5-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/MedianFilter_c2_5-5-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/MedianFilter_c1_5-5-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/MeanFilter_c2_5-5-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/MeanFilter_c1_5-5-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/MaxFilter_c2_5-5-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/MaxFilter_c1_5-5-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/LoG_c2_4-4-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/LoG_c1_4-4-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/HighPassFilter_c2_19-19-9_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/HighPassFilter_c1_19-19-9_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/Gaussian_c2_19-19-9_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/Gaussian_c1_19-19-9_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/EntropyFilter_c2_5-5-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/EntropyFilter_c1_5-5-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/ElementWiseDifference_c2_reverse_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/ElementWiseDifference_c1_reverse_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/Closure_c2_5-5-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| docs/images/Closure_c1_5-5-2_brightMIP.png | Adds LFS-tracked documentation image asset. |
| CONTRIBUTING.md | Updates contribution guide and adds local build/test + release process notes. |
| CODE_OF_CONDUCT.md | Adds Contributor Covenant code of conduct. |
| CMakePresets.json | Adds a Visual Studio preset for Windows builds. |
| CMakeLists.txt | Updates minimum CMake version, default CUDA arch handling, Python3 discovery, test toggles. |
| CLAUDE.md | Adds repository build/test/architecture guidance document. |
| ci-readme.md | Adds CI/CD process documentation (appears to be historical guidance). |
| check_dll_deps.ps1 | Adds a PowerShell helper for inspecting Windows DLL dependencies. |
| CHANGELOG.md | Adds a v4.0.0 changelog with packaging/CI/test changes. |
| build-win32/build.md | Removes outdated Windows build instructions file. |
| build-win32/build.bat | Removes outdated Windows build batch script. |
| build-linux/build.md | Removes outdated Linux build instructions file. |
| BACKLOG.md | Adds a prioritized backlog/roadmap of operations. |
| .gitignore | Updates ignore rules for build artifacts and compiled extensions. |
| .github/workflows/release.yml | Adds a tag-driven release workflow with tag/version guard. |
| .github/workflows/matlab-multibuild.yml | Adds multi-platform MATLAB toolbox build workflow. |
| .github/workflows/ci.yml | Adds Windows/Linux build+import CI gate using conda-forge CUDA. |
| .github/PULL_REQUEST_TEMPLATE.md | Adds a PR template. |
| .gitattributes | Updates Git LFS tracking rules (adds *.tif). |
| .condaignore | Adds conda-build ignore list to reduce source tarball noise. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
+16
to
+19
| - pip: | ||
| - e==1.4.5 | ||
| - hydra-image-processor==0.1.0 | ||
| - pefile==2024.8.26 |
Comment on lines
+4
to
+7
|
|
||
| Usage: | ||
| python scripts/download_artifacts.py --run-id <RUN_ID> | ||
| """ |
| import sys | ||
| import zipfile | ||
| from pathlib import Path | ||
| from urllib.request import urlretrieve, urlopen |
Comment on lines
+23
to
+28
| headers = {} | ||
| if token: | ||
| headers["Authorization"] = f"token {token}" | ||
|
|
||
| req = urlopen(url) | ||
| data = json.loads(req.read()) |
Comment on lines
+41
to
+50
| # Use gh CLI if available for authenticated downloads | ||
| if token or subprocess.run(["gh", "--version"], capture_output=True).returncode == 0: | ||
| subprocess.run([ | ||
| "gh", "api", url, | ||
| "-H", "Accept: application/vnd.github+json", | ||
| "-H", "X-GitHub-Api-Version: 2022-11-28", | ||
| "--output", str(output_path) | ||
| ]) | ||
| else: | ||
| urlretrieve(url, output_path) |
Comment on lines
+9
to
+12
| ```bash | ||
| cd src/Python | ||
| pip install -e . | ||
| ``` |
Comment on lines
+1
to
+4
| # Hydra Image Processor - Python Package | ||
|
|
||
| A Pythonic wrapper for the Hydra Image Processor C++/CUDA library, providing GPU-accelerated image processing operations with automatic CPU fallback. | ||
|
|
Comment on lines
+7
to
+10
| | Version | Supported | | ||
| | ------- | ------------------ | | ||
| | 3.x | :white_check_mark: | | ||
| | < 3.0 | :x: | |
| add_executable(test_back test.cpp) | ||
|
|
||
| target_link_libraries(test_back PRIVATE HydraCudaStatic) | ||
| target_include_directories(test_back PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}../Cuda) |
|
|
||
| target_link_libraries(test_accuracy PRIVATE HydraCudaStatic) | ||
|
|
||
| target_include_directories(test_accuracy PRIVATE ${CMAKE_CURRENT_SOURCE_DIR}../Cuda) |
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.
Fast-forwards main to the conda-forge branch for the v4.0.0 release.
Highlights
pip install .via scikit-build-core builds the CUDA extension and installshydra_image_processor(+HIP/Hydrashims); version single-sourced inpyproject.toml; in-repo conda recipe removed (distribution via conda-forge).src/Python/Test/test_accuracy.py) that gates releases.make_ellipsoid_maskoff-center bug; explicit<cstdint>includes for newer toolchains.Verification
🤖 Generated with Claude Code