feat: Add zarr v3 reading support for OME-NGFF datasets#25

Merged

hongquanli merged 48 commits intomainfrom

feat/zarr-v3-support

Jan 25, 2026

Commits on Jan 25, 2026

feat: Add zarr v3 reading support for OME-NGFF datasets

hongquanli
and
claude
committed
feat: Add zarr v3 acquisition simulation script

hongquanli
and
claude
committed
feat: Support all 4 zarr structures in simulation script

hongquanli
and
claude
committed
fix: Route slider changes to correct loader for zarr mode

hongquanli
and
claude
committed
fix: Support per-FOV zarr paths and hide single-FOV slider

hongquanli
and
claude
committed
fix: Address PR review issues and change default interval to 0.05s

hongquanli
and
claude
committed
chore: Remove redundant list initialization

hongquanli
and
claude
committed
feat: Support new Squid zarr v3 format (PR #474)

hongquanli
and
claude
committed
style: Fix black formatting

hongquanli
and
claude
committed
ci: Install package in editable mode for tests

hongquanli
and
claude
committed
ci: Use python -m pytest for reliable imports

hongquanli
and
claude
committed
ci: Add debug imports step to diagnose failure

hongquanli
and
claude
committed
ci: Fix pytest import issue with pythonpath config

hongquanli
and
claude
committed
fix: Add new zarr v3 exports to root shim __init__.py

hongquanli
and
claude
committed
fix: Remove unused imports from test_zarr_push_api.py

hongquanli
and
claude
committed
fix: Address code review issues

hongquanli
and
claude
committed
fix: Remove unused n_y and n_x variables in _load_zarr_v3_6d

hongquanli
and
claude
committed
fix: Remove unused dtype variable in _load_zarr_v3_6d

hongquanli
and
claude
committed
feat: Add multi-region 6D zarr support for push-based API

hongquanli
and
claude
committed
refactor: Rename start_zarr_acquisition_6d_regions to start_zarr_acquisition_6d

hongquanli
and
claude
committed
fix: Don't cache empty planes during live zarr acquisition

hongquanli
and
claude
committed
perf: Use O(1) set lookup for cache eligibility instead of O(n) plane.max()

hongquanli
and
claude
committed
fix: Clear _zarr_written_planes in closeEvent() for consistent cleanup

hongquanli
and
claude
committed
fix: Check zarr.json exists before loading in per_fov mode

hongquanli
and
claude
committed
feat: Use tensorstore for zarr loading to support both v2 and v3 formats

hongquanli
and
claude
committed
chore: Make tensorstore a required dependency

hongquanli
and
claude
committed
refactor: Remove zarr library dependency, use tensorstore exclusively

hongquanli
and
claude
committed
fix: Write metadata to zarr.json attributes to match Squid v3 format

hongquanli
and
claude
committed
fix: Fix black images in zarr simulation by writing dual metadata format

hongquanli
and
claude
committed
refactor: Remove .zattrs legacy support, use only zarr.json

hongquanli
and
claude
committed
fix: Make 6D mode FOV slider start at 0 and increment like other modes

hongquanli
and
claude
committed
fix: Fix NameError for total_fovs in start_zarr_acquisition_6d logger

hongquanli
and
claude
committed
ci: Add ruff for static analysis to catch undefined names

hongquanli
and
claude
committed
ci: Add ruff to pre-commit hooks

hongquanli
and
claude
committed
style: Fix black formatting

hongquanli
and
claude
committed
fix: Add missing imports to zarr v3 loaders and fix outdated comment

hongquanli
and
claude
committed
docs: Add zarr viewer architecture documentation

hongquanli
and
claude
committed
test: Add tensorstore integration tests for zarr v3 loading

hongquanli
and
claude
committed
fix: Correct 6D zarr axis order to (FOV, T, C, Z, Y, X)

hongquanli
and
claude
committed
refactor: Rename 6d_single structure type to 6d

hongquanli
and
claude
committed
fix: Race condition in zarr plane caching during live acquisition

hongquanli
and
claude
committed
test: Add unit tests for cache race condition prevention

hongquanli
and
claude
committed
docs: Add plane cache rationale to architecture docs

hongquanli
and
claude
committed
refactor: Remove eager zarr store opening in start_zarr_acquisition

hongquanli
and
claude
committed
fix: Add open_zarr_tensorstore to root shim __init__.py

hongquanli
and
claude
committed
docs: Fix incorrect file reference in architecture docs

hongquanli
and
claude
committed
fix: Add thread safety for _zarr_written_planes set

hongquanli
and
claude
committed
fix: Add thread safety for zarr store dicts

hongquanli
and
claude
committed