fix: handle generator-based MDA sequences without crashing#364
fix: handle generator-based MDA sequences without crashing#364tlambert03 merged 3 commits intopymmcore-plus:mainfrom
Conversation
Generator sequences (e.g. from `mmc.run_mda(my_generator())`) have unknown shape and cannot pre-allocate zarr layers. The handler now detects GeneratorMDASequence in _on_mda_started and skips layer creation, and _on_mda_frame forwards generator frames to the preview layer so they remain visible. Also adds a PixelSizeWidget deprecation warning filter for tests.
hinderling
force-pushed
the
fix/generator-mda
branch
from
e6b0a58 to
85dc348
Compare
|
any changes you want to make here are fine with me @hinderling ... as you know, we don't focus on napari too much anymore. But I do want to keep it working and I appreciate your eyes on it. I just sent you an invite to be a maintainer of this repo (write+ access), and would welcome any modifications you want to make (still through PRs though) |
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #364 +/- ##
==========================================
+ Coverage 78.25% 78.72% +0.47%
==========================================
Files 13 13
Lines 630 644 +14
==========================================
+ Hits 493 507 +14
Misses 137 137 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
|
can you change all the |
hinderling
force-pushed
the
fix/generator-mda
branch
from
598bb11 to
9ebdd09
Compare
The mypy pre-commit environment with pymmcore_plus exceeds the pre-commit.ci 250MiB tier limit due to tensorstore (~46MB). Skip mypy on pre-commit.ci and run it as a dedicated CI job instead, where it has access to the full dependency tree with no size limit.
hinderling
force-pushed
the
fix/generator-mda
branch
from
9ebdd09 to
a0872c3
Compare
|
skipped mypy in pre-commit because it exceeded the 250MB limit (mainly numpy/tensorstore deps from pymmcore-plus), instead added it to the workflow. |
hinderling
force-pushed
the
fix/generator-mda
branch
from
99642ed to
a0872c3
Compare
|
ready for merge right? (that failing test is probably unrelated... but we'll need to fix/ignore it before release) |
|
yes seems unrelated! ready to merge from my side |
|
just pushed tag for v0.2.0 ... let me know if it doesn't make it to PyPI :) |
|
thank you talley! |
2 participants
Summary
mmc.run_mda(my_generator())) have unknown shape and previously crashed the MDA handler when it tried to pre-allocate zarr layers_on_mda_startednow detectsGeneratorMDASequenceand skips layer creation_on_mda_frameforwards generator frames to the preview layer so acquired images remain visiblePixelSizeWidgetdeprecation warning filter for tests