Skip to content

Validation status

This page preserves exact release evidence and keeps changed-domain qualification separate from carried-forward scientific and installer records. It is a claim boundary, not a certificate that every node, reader, or workflow is validated for every assay.

0.15.0a1 scope and evidence

The 0.15.0a1 release is the canonical source for its exact assets, checksums, application revision, and release qualification evidence. Do not apply the older asset hashes below to a 0.15.0a1 download.

This release changes batch execution and cancellation, configuration version 6, the inspector and diagnostic plots, source/axis presentation, and mesh/skeleton measurement providers. These changed areas need their own regression or device evidence; an unchanged CUDA installation alone does not qualify a new provider. Unaffected scientific regions retain their documented boundaries.

The hybrid mesh path performs label preparation on GPU but retains CPU marching cubes, convex hulls, and table construction. Analyze Skeleton's device path measures an already-authored Boolean skeleton, not GPU thinning. See the current operation matrix.

The new synthetic threshold gallery and exhaustive inspector example add review coverage, not independent biological validation. The ImageJ Default implementation and Fiji-related colocalization methods retain their stated experimental source-aligned status; external golden parity is not implied by the UI refresh or version change.

0.14.0a3 exact application and artifact record

The annotated v0.14.0a3 tag resolves to 3c61d0c84cc728d90537ae10c4c521e4a4f40809. The exact-main application CI, Windows installer smoke, and native macOS arm64/x86_64 smoke passed at that commit. The exact-tag unsigned-installer workflow then passed for Windows x86_64, macOS arm64, and macOS x86_64.

The GitHub prerelease contains exactly 20 qualified assets: the wheel and source archive, four Windows files, seven Apple Silicon files, and seven Intel macOS files. Independent CPython 3.12.10 builds of the exact-tag wheel and source archive matched before publication. The PyPI publication workflow passed, and freshly downloaded PyPI files matched the exact GitHub release bytes.

Public artifact SHA-256
Wheel napari_vipp-0.14.0a3-py3-none-any.whl 7231ab13c76ab8bb2926d0e1569eb527491b392d36dba6b888d4eec03ba616e2
Source archive napari_vipp-0.14.0a3.tar.gz 185bb25e19f8e04ed47df5c45993ac52153c53c5b1c14e4d7c79010d84558cad
Windows installer VIPP-Setup-0.14.0a3-Windows-x86_64-UNSIGNED.exe 8774dd14047cb31f3694a30c05878ff59e50fe2e3f5f200d09e13082eed30d73
Apple Silicon package VIPP-0.14.0a3-macOS-arm64-UNSIGNED.pkg 69b62fd576bdba9c2002457107ea46cd99b6a5cf091558b35281c99cd57f4100
Intel package VIPP-0.14.0a3-macOS-x86_64-UNSIGNED.pkg 5b97d530a58df9f1d3a873e2f22036bcfa6d62d50cbd22b960b063c76e249d73

This is application and artifact evidence. The numbered and stable documentation deployment is verified separately after this exact manual is deployed.

0.14.0a3 changed-domain qualification scope

0.14.0a3 changes core/UI behavior, workflow schema and provenance, local OME-Zarr source-window planning and memory preflight, shared execution coordination, napari/Qt compatibility, installer dependency inputs, and documentation. Qualification must cover responsive Z/Y/X Crop Stack behavior, opt-in low-RAM crop repair, exact sole-direct source-window parity and refusal cases, safe interactive and batch bypass, save/undo behavior, workflow schema 6 and batch config/manifest schema 5 migrations, napari 0.6/0.9/latest boundaries, PyQt6/PySide6 startup, and exact-tag Windows plus both native macOS packages.

GPU provider kernels and their admitted scientific regions are unchanged. Their recorded catalogue evidence may carry forward only if the a3 shared-execution checks and changed-domain declaration pass. General branch-aware region planning, remote or non-OME-Zarr exact reads, general lazy/chunked execution, Apple GPU support, signed/notarized installers, and broad assay validation are not claimed.

Previous exact record: 0.14.0a2 focused qualification boundary

0.14.0a2 changes desktop/Qt compatibility, detached-window sizing, and release packaging. It adds separate offline CPU-only macOS packages for Apple Silicon and Intel. SourceItem, reader, workflow schema 5, batch config/manifest schema 4, OME-Zarr preview, per-sample parameter, scientific operation, and GPU contracts are unchanged from 0.14.0a1 and retain that recorded evidence.

The annotated v0.14.0a2 tag resolves to 4aaf9961b97259c94390c859374d8a6b9f45ec6c. The exact-tag CI run passed, including the supported Windows, Linux, and macOS Python matrix and clean distribution installs. The unsigned-installer workflow passed for Windows x86_64, macOS arm64, and macOS x86_64. Its native macOS jobs inspected, installed, launched, and cleanly shut down each final package. The PyPI publication workflow also passed.

The GitHub prerelease contains 20 assets. Primary public artifact SHA-256 values are:

Artifact SHA-256
GitHub wheel napari_vipp-0.14.0a2-py3-none-any.whl a30aa4ff1f4882b06903800d60912be4f6b29f185716a1752b3f8c540690ae8c
GitHub source archive napari_vipp-0.14.0a2.tar.gz 8d1bf13b752bdb4c6777076c5e6bda5a75275073016f15e347ed8698e6b01886
Windows installer VIPP-Setup-0.14.0a2-Windows-x86_64-UNSIGNED.exe cdab2be65b83c36260ba45b3b700bc1685287e665bb9558585f9f13f9fb3f3be
Apple Silicon package VIPP-0.14.0a2-macOS-arm64-UNSIGNED.pkg 9ad6a01e614277a4abac501a27d8623da5bd22bf4184bfb7b0884c319b67f529
Intel package VIPP-0.14.0a2-macOS-x86_64-UNSIGNED.pkg 15ddfa6378af4bdf69887d35bd107b04143d89a944e68686687f350532546d14

The two macOS packages require macOS 13 or newer, are current-user-only, explicitly unsigned and unnotarized, and install CPU-only managed environments. Automated native lifecycle evidence does not establish browser-quarantine, managed-device override, signing/notarization, graphical update/uninstall, or Apple-accelerator support.

PyPI a2 is a different, earlier build

PyPI cannot replace uploaded files. Its a2 wheel (3b066887de739d600a684abf85d599832d5b92c83fb357b6c75a471ed20cd2a8) and source archive (ad72ae95a5595c8bcad6125aada8aab1c30a419b98929a91b405d3560405fd27) are the pre-resize-fix build and do not match the current GitHub release wheel/source bytes. Use a platform installer or the GitHub wheel for the detached-window fix, and record which distribution surface was used.

This is application and artifact evidence. It does not itself claim that the numbered or stable documentation deployment succeeded.

0.14.0a1 SourceItem and batch qualification boundary

The changed release domains are SourceItem identity and migration, reader-contract normalization, local OME-Zarr presentation preview, typed per-sample batch values and retained-workspace UX, source-load preflight, and installer capacity/activity presentation. Qualification uses focused core/UI, reader, schema/migration, preview, override, and installer-presentation checks, the integrated release suite, the strict verified-cache public corpus profile, and exact-tag package and installer checks.

This scope does not claim remote stores, IMS pyramid preview, HCS traversal, operation-level lazy execution, broad optional-reader equivalence, or per-sample expressions/source selectors/topology changes. The lower OME-Zarr level is presentation-only; scientific analysis remains level 0. The unchanged full GPU catalogue and installer transactional lifecycle retain their recorded 0.13 evidence rather than being silently counted as newly rerun.

Exact public packages and the immutable release assets are available from the v0.14.0a1 GitHub release and PyPI. The annotated tag resolves to e16ca87161ec7b1041a5e98c5a2bf786b11a1ec8; exact-main CI, the exact-main installer build, and the PyPI publication workflow passed. The strict public-corpus manifest SHA-256 is 3365b4cec7a220f6399f3c030eb3ac751581bde730fac60f01c9fca3b823714d. Exact distribution hashes are recorded in the 0.14.0a1 release-verification table. This application evidence does not claim a documentation deployment result.

0.13.0a9 qualification boundary

Final a9 application source is commit d476c8976c91679faaaf0fe196c90b1ab4d63458. Its exact-main CI and Windows installer build smoke passed. Focused acceptance covered immediate QYX-to-ZYX Gaussian metadata, explicit volumetric skeletonization, actionable GPU VRAM preflight, stable sibling measurement caches, and Prefer GPU planning through CPU-only nodes.

On the release workstation, the original affected workflow completed under Prefer GPU with the reviewed downstream CuPy/CuPyX operations on cuda:0 and no fallback across required CPU-only boundaries. This is bounded evidence for that environment, not a universal device or installation claim.

Installer transaction behavior, dependency/toolchain pins, and release infrastructure did not change. Their lifecycle evidence therefore carries forward from a8 rather than being replayed. The a9 tag nevertheless regenerated and checked the exact frozen payload, embedded wheel, version, NotSigned state, manifests, hashes, GitHub assets, and PyPI bytes. See the a9 release-verification table.

0.13.0a8 qualification boundary

Final a8 application source is commit 5a66ae9d1098ca5a8d409a4075c585692e3c3638. Its exact-main CI run passed all 13 jobs: quality/package checks, the full tests on Windows, Linux, and macOS with CPython 3.12 and 3.13, and clean wheel/source-archive installation lanes. The Windows installer smoke also passed against that commit.

The public GPU catalogue declares 19 implementations and 24 executable evidence owners. The full native-Windows RTX 5090 cuda:0 qualification passed all of them. Its aggregate evidence SHA-256 is 0365366dc23750e000c6e9c4f8b384cdf706afdcb338ae3a9f80cfad3d1d8506. The evidence source was candidate commit 7189cf40280d895b61b061f1468767164ccfbcf4; the only subsequent executable-code change was the packaging canonicalizer. Other changes recorded evidence and documentation and did not alter production GPU code.

Risk-based installer acceptance exercised the a8 behavior affected by this release. Updating the existing managed CUDA installation from a7 to a8 left zero cuCIM, nvimgcodec, or retired-provider residue in active a8 site-packages; pip check was clean; Compute Doctor passed 19 of 19 implementations; and the visible application reported VIPP 0.13.0a8. Both new CuPy measurement implementations passed parity, cancellation, reuse, and zero-private-pool-residue checks. The a7 rollback was deliberately preserved.

The separate cuCIM installer, source-build coordinator, private-wheel approval path, and hosted provider asset were removed. The redundant fresh CPU, installer-cancellation, same-version repair, and uninstall matrix was not repeated for this minor alpha; those unperformed paths are not claimed as a8 passes. Exact public artifacts and publication evidence are recorded in the a8 release-verification table.

Evidence available now

  • automated tests cover graph behavior, persistence, previews, export, operations, UI behavior, I/O routes, and bundled examples;
  • architecture tests enforce the Qt-free core/ dependency boundary, and focused tests cover stable source revisions, physical grids, detached snapshots, atomic persistence, typed execution, and stale-result rejection;
  • strict public-corpus v4 qualification covers 20 frozen artifacts (210.41 MiB), 97 biological fields, and the claimed OME-Zarr, TIFF, ND2, LIF, CZI/LSM, OIR/OIB/VSI, and IMS routes. All 10 strict vendor cases passed, including the IMS case under portable Temurin 21. The corpus-manifest SHA-256 is 3365b4cec7a220f6399f3c030eb3ac751581bde730fac60f01c9fca3b823714d;
  • 15 deterministic synthetic samples and 19 checked-in workflows support regression checks and inspection;
  • a generated two-source batch bundle exercises three paired items, nine exact NPY/TIFF/TSV outputs, workflow/config hashes, source identities, manifests, archives, and three finalized item sidecars;
  • an analytical phantom report exercises calibrated object/mesh morphology;
  • method notes and focused tests cover colocalization calculations;
  • tables, CSV/TSV writing, workflow JSON, and Python generation have automated behavior checks;
  • focused UI/execution tests cover isolated-tuning boundaries, actionable and waiting stale states, progressive node previews, compatible layer reuse, and rejection of stale contrast/histogram results;
  • batch tests cover attached-config validation and round-trip, guarded source- axis declarations, representative scientific preflight, background source checking on restore without calculating representative pixels, inspected multi-series expansion and identity retention, direct plan-only execution, complete-item fast skips, transient atomic-write retries, and continuing after a final item-sidecar failure. The current batch suite additionally covers individual existing-output choices, reused Run preflight, selection and reset semantics, cooperative cancellation, node progress, and the readable report;
  • compute tests cover import-safe CPU-only use, workflow-schema-6 and batch-config-6 intent, migration from workflow schemas 3/4/5 and batch versions 1/2/3/4/5, eligibility planning, exact implementation identity, resident device segments, memory admission, classified fallback, optimizer review/apply and grouped result inspection, visible dtype-repair proposals, Prefer-GPU selection/serialization/UI/durable behavior, progress, cancellation, cleanup, and atomic publication; and
  • opt-in native-Windows RTX tests exercise a real durable GPU batch and an imported generated Python workflow through the same executor.

Historical 0.13.0a7 published qualification boundary

Final a7 application source merged to main through pull request #23 at dc8a63912110a75ab1daad0e7f81c2b20e5001e6. Its exact-main CI run passed on Windows, Linux, and macOS. The complete local suite passed 5,084 tests, with 5 documented skips, 2 documented expected failures, and zero failures.

The clean-source native-Windows RTX 5090 full GPU admission profile passed all 23 executable evidence owners across 18 public implementations. The aggregate evidence SHA-256 is 3ad655f7d3e36055449bda3e8bb41c914e010fd7607ced26763e23045dcee7ae, and its admission-manifest SHA-256 is 3b6081b0aec45f81227bd86d86bc0f2df1aa4fe6b28752aa4e96aaee3d8e0ce7. The new evidence covers exact dtype conversion, Binary Threshold, Extract Channel, Boolean Remove Small Objects, and Boolean Fill Holes across parity, difficult inputs, metadata, unchanged inputs, memory, cancellation, cleanup, fallback, provenance, and end-to-end timing contracts.

Source-current native-Windows RTX 5090 checks exercised a single resident path from the visible dtype conversion through filtering, thresholding, Boolean cleanup, and Connected Components. One retained terminal output used one upload and one final download; retaining an intermediate deliberately added another download. These are bounded development/reference-system results, not a support claim for every GPU or a downloaded public artifact.

The exact-source wheel, source archive, unsigned Windows installer and sidecars, and no-wheel cuCIM bundle were finalized and hash-locked locally. Their filenames and SHA-256 values are recorded in the a7 release-verification table. Twine accepted both Python archives; deterministic-wheel and byte-reproducible cuCIM-bundle checks passed; isolated wheel and source-archive installations passed package, resource, entry-point, manifest, and headless checks; and unsigned installer finalization plus native NotSigned inspection passed.

A display-independent exact-artifact acceptance exercised the production Windows installer engine and registration/removal services from the final a7 wheel, bound to the finalized unsigned EXE. CPU and CUDA each completed a new install, installed-package scientific checks, repair, and ownership-safe uninstall. The CUDA route additionally passed Compute Doctor with 14 of 18 public regions admitted, the portable segmentation corridor, 3D RL and RL-TV, and the non-ASCII effective-TEMP compiler regression. Both routes removed their owned environments, registry entries, shortcuts, cached a7 setup, and transaction residue.

That result is production-backend lifecycle evidence; by itself, it did not operate the frozen EXE's setup window. Separately, on 2026-08-15, the operator attested that the exact locally finalized EXE's visible setup and installed-UI checks passed. The attestation covered visible shortcut launch, the GPU-tip Add conversion/Undo and post-calculation persistence checks, and grouped Find-fastest result readability.

The local operator pass was not a fresh-account or public-download installation. The field checklist therefore leaves public-download SmartScreen, novice comprehension, unusual account paths, cancellation, network rollback, and update from an older release as not run.

The public v0.13.0a7 tag peels to the exact qualified commit. The official GitHub prerelease contains exactly seven assets; a fresh download of every asset matched the recorded SHA-256. The protected PyPI workflow published only the already-qualified wheel and source archive, and both public PyPI digests matched. The numbered-manual workflow succeeded and all required 0.13.0a7 URLs returned HTTP 200. Stable-alias workflow 31871941473 succeeded; at that deployment, public versions.json mapped stable to 0.13.0a7, and the required stable URLs returned HTTP 200 with no stale prepublication wording.

Fresh-account Unicode Known Folder, public-download SmartScreen, novice-pilot, RTX 40-series Windows, and native Linux CUDA field evidence remain not run. No a6 artifact, hash, URL, or qualification record is reused as a7 evidence.

Historical 0.13.0a6 qualification and field boundary

The source candidate's complete local suite passed 4,572 tests, with 5 documented skips, 2 documented expected failures, and zero failures. Clean wheel and source-archive installations passed in fresh environments, and the cross-platform workflow now checks both distribution forms across Windows, Linux, and macOS on the supported Python versions.

On the native-Windows RTX 5090 reference environment, Compute Doctor 2.0 admitted all 13 of 13 current public GPU regions. The strict quick admission profile passed 130 of 130 mapped checks across all 16 executable implementation owners, including parity, difficult inputs, metadata, unchanged inputs, memory, cancellation, cleanup, fallback, provenance, and transfer-inclusive timing evidence.

The final clean-tag RTX 5090 run also passed the strict full admission profile for 16 of 16 executable owners, 13 public implementations, and all 10 contract facets. Its aggregate evidence SHA-256 is 44bd66033afedcbece8d8746e1779d833af2897ac230adc4e8a98d7847f7f56c.

This is strong engineering evidence for the source candidate, not a substitute for testing the downloaded installer elsewhere. The Windows field checklist therefore retains fresh-account CPU/CUDA installation, spaces and non-ASCII account paths, cancellation and network rollback, repair/update/uninstall, an RTX 40-series machine, and a novice first workflow as explicit field checks. Anything not performed remains not run.

The immutable v0.13.0a6 tag resolves to application commit 859738a28354981ba784d9e49a04cb6e1158a79f. Its exact-source CI run passed on Windows, Linux, and macOS. The tagged wheel, source archive, cuCIM local-build bundle, and Windows sidecars passed the release qualification gates. Independent Windows inspection confirmed that the explicitly unsigned installer is NotSigned, with no signer or timestamp certificate. Exact artifact hashes are recorded in 0.13.0a6 release verification.

The public GitHub pre-release contains exactly seven assets whose published SHA-256 digests match the release table. The public PyPI release has matching wheel and source-archive hashes. The numbered-manual deployment workflow succeeded, and its required numbered pages returned HTTP 200.

Publication and clean-tag GPU qualification do not replace the downloaded installer field checklist. External fresh-machine and novice acceptance remain not run unless a tester records them explicitly.

Historical 0.13.0a5 installer acceptance and release boundary

The release application commit is 067c89559072fbbb101e9d63b91514345e5896e6, merged through release PR #16.

The development installer passed fresh managed CPU and CUDA installation, real Auto and Prefer-GPU execution with CPU parity and clean accelerator release, optional cuCIM installation and execution, update and repair, and independent CPU/CUDA removal on the Windows reference system. The DEVELOPMENT file is not a release artifact. The final tagged release route created only the explicit -UNSIGNED filename from the immutable tag's exact wheel after build bytes, frozen payload, NotSigned status, release manifest, and SHA-256 checks passed.

The exact tagged release EXE then passed a clean installer-owned CPU lifecycle: reviewed/hash-locked resolution, install, package health, shortcuts, responsive first launch, transactional same-version repair, safe refusal while one DLL was temporarily locked, and complete removal after the application closed. A local build has no browser Mark-of-the-Web, so acceptance did not synthesize a SmartScreen page. The user guide documents the expected Unknown publisher and Windows protected your PC warning for browser downloads, the safe More info → Run anyway path, checksum verification, and managed-device fallback. The filename without -UNSIGNED remains reserved for a future valid Authenticode-signed and timestamped installer.

Final CI, tag, signing status, and artifact hashes are recorded in 0.13.0a5 release verification.

Historical 0.13.0a4 release source and artifacts

The immutable annotated v0.13.0a4 tag resolves to application commit 4bec1e8145b31e161beaf44a290bff24aea36f5e on main. Its exact-source CI run passed package, manifest, lint, wheel smoke, and CPython 3.12 and 3.13 tests on Windows, Linux, and macOS.

The final local source suite completed with 4,018 passed, 109 expected skips, 2 documented expected failures, and zero failures. The clean tagged wheel and source distribution passed Twine and content inspection. The publication artifacts are:

Artifact SHA-256
napari_vipp-0.13.0a4-py3-none-any.whl 5FA75FC48955E2CA9AD7D5BC13218AD83EF14C23F0B72F74D8A7486CA8453085
napari_vipp-0.13.0a4.tar.gz B99B829C45BE734B705DEE17829886EF91DEAABC3C3D3C5F76C164DB87A617AF

The exact tagged wheel was installed into an isolated overlay over the pinned native-Windows CUDA 13 environment. On an NVIDIA GeForce RTX 4050 Laptop GPU (compute capability 8.9), Auto and Prefer GPU selected eligible CuPy/CuPyX/cuCIM implementations. All 13 bundled examples completed as fresh CPU, Auto, and Prefer-GPU graphs; every selected GPU node passed its declared production parity contract, no fallback record was emitted, and cleanup succeeded. Provider probes also passed for CuPy, CuPyX, and the approved local cuCIM build.

This complements the retained RTX 5090 evidence rather than turning either machine into a model allowlist. Public a4 admission accepts a probed NVIDIA CUDA device with compute capability 7.5 or newer and driver API 13.3 or newer while retaining the exact supported native-Windows, CPython 3.12, CUDA runtime 13.2, scientific-stack, provider-provenance, workload, memory, fallback, and cleanup gates. Minor floating-point differences can occur across GPU models, drivers, compiler paths, and reduction order within a provider's declared tolerance; record the full environment and validate consequential work against CPU.

Historical 0.13.0a1 release source and artifacts

The immutable v0.13.0a1 tag resolves to application commit 7520a5bb3ea9fe296bb231c63d1598b833ac10f6 on main. Its exact-source CI run passed package, manifest, lint, distribution-metadata, installed-wheel smoke, and CPython 3.12 and 3.13 tests on Windows, Linux, and macOS. The CI-built distributions are qualification outputs, not publication artifacts.

On native Windows, the exact selected source's complete local suite completed with 4,077 passed, 2 skipped, 2 documented expected failures, 83 warnings, and zero failures. The skips are the opt-in real-CUDA tests and the expected failures are the documented CuPy integer-parity gaps. The exact source was then covered by the cross-platform CI matrix above.

The final tagged artifacts published through PyPI and attached to the GitHub pre-release are:

Artifact SHA-256
napari_vipp-0.13.0a1-py3-none-any.whl C157D2D0E5909A76A1A6093493AB17A245682DF53D83DD0A23FAA0FDF7A3BE02
napari_vipp-0.13.0a1.tar.gz E3231CF22EA2907C3FE05F73477E5E3CD3B10FF2A053BBA92A8045D140E7F7E0

The bundled synthetic-volume example was launched from a clean, non-editable, release-style Windows environment, and the operator explicitly accepted its appearance. A separate headless Prefer-GPU smoke selected cucim-subtract_background-v2 for Subtract Background without fallback, returned a (31, 37) native-uint16 result, and reported clean accelerator cleanup. This is bounded evidence for one example and one declared cuCIM operation region, not broad manual GUI, reader, filesystem, hardware, or assay qualification.

The pinned private Windows cuCIM route was exercised against upstream tag v26.06.00 at commit 3c15781c207eab93a317dd9803a6e726fe01f7c4. Two clean builds from the exact 41-distribution no-dependency lock produced byte-identical wheel files and the policy-pinned canonical payload d640d1e17bcce15d32d03841997252bf915b63da855e406c35f0d70c5a5ea667. Metadata, licenses, exact file inventory, install-helper admission, pip check, and a real GPU import/runtime probe passed. This locally built wheel remains private: VIPP neither ships nor hosts it, and each user must repeat the pinned build for their own environment.

These are the release files. The historical artifacts below must not be uploaded or substituted for them.

The earlier pre-Prefer-GPU candidate at application commit e024409 passed package/manifest, lint, and CPython 3.12 and 3.13 CI on Linux, Windows, and macOS. The candidate wheel and source distribution built and passed Twine metadata checks; a clean CPU-only wheel environment passed installation/import checks, and the installed CUDA wheel environment passed compute-doctor plus the two opt-in real-CUDA durable batch/generated-Python tests. A green CI/package matrix is not equivalent to a manual GUI smoke pass on every operating system, Qt/display environment, filesystem, microscope reader, or GPU. See the candidate CI run. Those results remain evidence for that exact tree, but the Prefer-GPU source change invalidates it as the final release candidate.

The later automated checkpoint is application commit 444f68290fe4359b05c68a027d3ae0a413412fe5 on codex/gpu-cross-platform-support. Its full local suite completed in 299.65 seconds with 3,754 passed, 2 skipped, 2 xfailed, 83 warnings, and zero failures. The skips are the opt-in real-CUDA durable batch smokes. The xfails are the documented CuPy uint8/uint16 integer-parity gaps. Ruff, source and installed-wheel npe2 manifest validation, package build, Twine, and the installed-wheel resource smoke passed. These counts describe that exact historical tree only. Subsequent compute-lifecycle, optimizer, source-loading, generated-CLI, and cleanup-quarantine hardening changed application behavior, so 444f682 is superseded and is not the current release candidate.

The checkpoint's real RTX 5090 Prefer-GPU integration test passed and selected CPU Extract Channel, cuCIM Subtract Background, CPU native-uint16 Gaussian, and CuPyX Median without per-node overrides. The complete pipeline had exact parity, no fallback, and clean accelerator cleanup. This verifies the automated placement contract for that environment; manual Prefer-GPU UI acceptance is still pending.

The exact historical checkpoint artifact hashes are:

  • wheel: 58b08cbb8396c9fe27d28a69d52b06e160817e58987e6dc3b8c5059f3dae9804;
  • source distribution: 00fcb15452d3ed71d344859d4306cbcdcc458686959169a2c47485d8f7abec9b.

These artifacts must not be uploaded for 0.13.0a1. They are superseded by the tagged 7520a5b release source and final hashes above. See the release notes for the historical artifact table.

A bounded source-candidate smoke on 4 August 2026 used application commit e024409 on an Apple M1 Max (arm64) running macOS 26.5.2. Manual checks covered launch/basic CPU processing, collection-batch progress and cooperative cancellation, and memory presentation as one system-RAM budget without a fabricated VRAM total. Four focused tests additionally covered the CPU-safe cancel path, worker-token propagation, single-shot UI cancellation, retained cancelled state, manifest, and cleanup evidence. The follow-up operator record is commit ff21040. This is bounded source-checkout evidence, not a raw test log, clean-wheel Mac qualification, broad reader/display/filesystem coverage, or Apple GPU evidence.

A bounded manual Windows acceptance pass inherited from 0.12.0a3 covered direct unpreviewed batch execution, complete Skip items, continued processing after an item failure, attached-config save/reload, and a representative 3D deconvolution batch. It does not establish cross-platform behavior, broad filesystem interoperability, large-collection scalability, or restoration quality for arbitrary samples.

On 4 August 2026, the operator also reported a bounded native-Windows napari UI smoke pass on the later ff21040 development checkout. A local schema-4 Custom workflow loaded the representative private ND2 acquisition and exercised the intended channel, slice navigation/display behavior, and backend-badge presentation. The retained last-run JSON was subsequently overwritten by an Auto run, so it does not independently preserve the exact mixed-backend assignment or cleanup result. This is operator-attested UI regression evidence only; it is not Prefer-GPU, final-wheel, durable-replay, or broad Windows/GPU qualification.

The e024409 full local suite passed 3,722 tests, with 2 skipped, 2 documented expected failures, and 83 warnings. The two final real-CUDA durable execution tests passed when enabled. These numbers describe that earlier candidate checkout. Preserve them as historical evidence, but do not use them as final-release counts. Use the release-source and final-artifact results above for 0.13.0a1 qualification.

The recorded native-Windows RTX 5090 evidence records operation-level scientific parity, memory, progress/cancellation, cleanup, and timing for the declared public GPU regions. The source-candidate timing refresh measured Richardson-Lucy at 24.898/0.414 seconds CPU/GPU (60.20x) on the private ND2 volume, 35.997/0.455 seconds (79.14x) on the medium synthetic volume, and 137.820/1.517 seconds (90.87x) on the large synthetic volume. Richardson-Lucy TV measured 34.921/0.599 seconds (58.34x) on the private volume and 55.936/0.564 seconds (98.61x) on the medium volume; both RL-TV workloads reported parity and clean cleanup. These are descriptive machine-local results, not portable performance guarantees or durable Auto assignments.

The release's generated calibrated-morphology report records 28/28 checks passed. Its own scope excludes broad numerical equivalence, biological interpretation, and all data conditions. The application test suite checks that the report remains synchronized with the validation script.

Do not generalize this evidence into

  • usability superiority over other tools;
  • broad equivalence to Fiji, CellProfiler, scikit-image, or another package;
  • scalability to whole-slide or high-content datasets;
  • complete OME/acquisition metadata fidelity;
  • biological validity of a segmentation, restoration, or measurement workflow;
  • user-study evidence beyond explicitly described pilot observations.

High-priority evidence gaps

Passing deterministic tests is valuable internal evidence, but it is not the same as an external comparison or assay validation. The distinction matters:

Area Current in-repository evidence Next evidence needed
Watershed/object separation Touching-disk split tests, exported-workflow execution, and 3D-default behavior tests Broader 3D phantoms, split/merge metrics, external comparison, representative real images
Colocalization/association Deterministic metric, overlap, distance, and association tests plus synthetic examples External numerical comparisons and assay-specific positive/negative controls
Skeleton networks Synthetic network workflows and focused operation tests Prespecified topology and calibrated-length packs, perturbation tests, external comparison
I/O and metadata Focused format, dtype, validation, and round-trip tests plus strict public-corpus v4 qualification across 20 frozen artifacts, 97 biological fields, and the claimed microscope-reader routes Broader independent facility files, negative controls for unusual vendor dimensions, network/remote filesystems, and cross-reader comparisons outside the frozen corpus
PSF/deconvolution Deterministic 2D/3D synthetic images, measured-PSF samples, and operation tests Real bead PSFs, representative microscopy images, artifact/noise analysis, performance characterization
Compute/GPU execution Exact operation-region tests, immutable policy v10, regenerated a3 full-catalogue qualification for 19 CuPy/CuPyX implementations and 24 evidence owners on native Windows RTX 5090, focused a9 and 0.14 shared-planner/source-axis real-GPU evidence, OOM/cancellation/cleanup coverage, bounded M1 Max CPU and Windows UI smokes, a3 exact-main Windows and native macOS smokes, and exact-tag three-platform installer qualification Qualify native Linux GPU, more NVIDIA architectures and compatible drivers, an Apple provider if pursued, and broader cross-platform manual GUI acceptance
Sources and physical grids Revision-change, owned-snapshot, stale-worker, semantic-axis, scale/unit/origin, mask-broadcast, and image/PSF grid tests Independent corpus covering live readers, network filesystems, registration histories, and heterogeneous microscope metadata
Large data/batch Functional cache/path/memory tests plus deterministic attached/standalone config, planner, direct plan-only execution, source verification, complete-item fast skips, staging, retry, manifest/archive, sidecar, collision, replay, continuation, exact-output bundle, a bounded Windows acceptance pass, and bounded M1 Max CPU progress/cancellation evidence Representative memory/time benchmarks, forced-process interruption studies, large collection stress tests, broader cross-platform/cloud-filesystem studies, semantic-axis iteration, and HCS traversal
Workflow/export architecture Workflow schema 6 with explicit schema-3/4/5 migration, batch config 6 and manifest 5 with supported earlier-version migration, canonical SourceItems, typed per-sample overrides, exact-item output choices, safe authored and batch bypass intent, guarded source-axis declarations, attachment validation, snapshot materialization, atomic-write failure, shared-executor compute provenance, multi-source binding, cancellation, and runtime-version tests Independent reproducibility exercises across archived environments and long-lived release migrations
Usability No release-pinned public usability study Ethics-reviewed, preregistered task study with a controlled comparator and neutral outcomes

Release-specific limitations

  • Workflow schemas 1 and 2 are intentionally rejected. Valid schema-3 workflows load as explicit CPU, while schema-4 and schema-5 workflows retain authored compute intent. Schema 5 also retains canonical SourceItems; schema-4 sources acquire them when they resolve. Supported earlier files save as schema 6, and nodes without authored bypass intent default to Run. Cached pixels/tables are not serialized and exported Python is runtime-version pinned. Recalculate, regenerate exports, and validate after upgrading.
  • Version-1 batch configs load as explicit CPU; version-2 configs retain their saved compute request; and version-3 configs retain guarded source-axis declarations and acquire SourceItems when resolved; version 4 adds SourceItems and typed numeric overrides. Earlier versions have no batch execution-profile override and are written as version 5 only after review and save. A saved Auto, Prefer GPU, or Custom request is intent; actual implementation provenance must be retained from each run. Auto uses reviewed GPU defaults without compatible history; accelerated-only history schedules one same-surface CPU measurement before a later matching run applies the 1.20x/20-ms gate. Prefer GPU instead requests every reviewed eligible accelerator regardless of speed; Custom owns per-node choices and benchmarking.
  • GPU candidates cover only declared operation/dtype/parameter/shape/memory and environment regions. Public admission requires native Windows, CPython 3.12, the pinned CUDA 13.2/scientific/provider stack, driver API 13.3 or newer, and a probed NVIDIA CUDA device with compute capability 7.5 or newer. macOS is CPU-only in this release; the bounded M1 Max CPU smoke above does not admit an Apple accelerator. The CPU/GPU matrix is a readable summary; the runtime policy/decision remains authoritative.
  • The current public GPU catalogue uses CuPy/CuPyX. See the Windows CUDA guide.
  • Batch processing remains local-file and pairs source lists positionally, although each resolved item is pinned to a canonical SourceItem rather than being silently reassigned by a changed series order. It does not iterate selected T/C/Z combinations or discover plate/well/field structure.
  • Most operations are eager even when a source format supports lazy/chunked access. One strictly eligible direct local OME-Zarr Crop Stack can read only its exact retained level-0 window; other graphs retain full-read memory preflight. Background execution improves responsiveness, not total work.
  • Declared-grid validation cannot prove biological registration or metadata truth. It can only enforce the axes/calibration supplied to VIPP.
  • Richardson-Lucy/TV controls and synthetic tests do not establish a validated restoration parameter range for a real microscope or assay.
  • Manifest and item sidecar writes improve recovery evidence but are not one transaction across all outputs and provenance files.
  • Cooperative progress/cancellation cannot split an opaque library call or file writer into truthful internal percentages.
  • Revised native-intensity colocalization is a source-aligned compatibility implementation targeting Fiji Coloc 2 3.1.0, but independent numerical parity validation is pending. The ImageJ threshold path similarly targets ImageJ 1.54p for scalar uint8, uint16, and float32; Boolean and RGB/RGBA handling are VIPP extensions and are not claimed as ImageJ-exact. Treat both paths as experimental and validate externally before consequential use.

For your workflow

Use validate a workflow to choose assay-specific evidence. If a public claim depends on a gap above, label it as a limitation or produce the required evidence before making the claim.