Fix label direct mode for installation without numba #6571
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #6571 +/- ##
==========================================
- Coverage 92.25% 92.21% -0.05%
==========================================
Files 603 603
Lines 53814 53831 +17
==========================================
- Hits 49648 49640 -8
- Misses 4166 4191 +25 ☔ View full report in Codecov by Sentry. |
| max_value = max( | ||
| (abs(x) for x in self.color_dict if x is not None), default=0 | ||
| ) | ||
| if any(x < 0 for x in self.color_dict if x is not None): |
There was a problem hiding this comment.
We really need to deprecate None and use defaultdict I think... (next PR) Anyway, like below this isn't tested — would you like it to be?
There was a problem hiding this comment.
It is tested, but not with code coverage (in pip job). I will add additional tests
There was a problem hiding this comment.
We do not want to use defaultdict globally. As defaultdict is adding elements when they are missed. So size is increasing.
There was a problem hiding this comment.
The option may be to add an additional property to DirectLAbelColormap to store default color outside color dict
There was a problem hiding this comment.
Good point @Czaki, I like this option.
When working on fix bug in npe2 I spot that napari tests in npe2 repository are failing because of lack of numba. In the context of one of the tests, it needs to be skipped when numba is not installed. However, the second failure reveals a real bug in codebase. This PR also modifies `pip` workflow to be used to test numbaless case.
* main: Fix labels mapping cache by filling it with background, not 0 (napari#6580) Simplify unused parameters of Quaternion functions. (napari#6424) Add size and ndim to LayerDataProtocol (napari#6494) Fix label direct mode for installation without numba (napari#6571) Fix test in napari_builtins to remove import from conftest (napari#6568) Remove `app-model!=0.2.4` from test constraints (napari#6577) Bump mypy version and fix errors (napari#6557) Update test to work with `app-model==0.2.4` (napari#6573) Added support for features in surface layers (napari#6515)
Description
When working on fix bug in npe2 I spot that napari tests in npe2 repository are failing because of lack of numba.
In the context of one of the tests, it needs to be skipped when numba is not installed.
However, the second failure reveals a real bug in codebase.
This PR also modifies
pipworkflow to be used to test numbaless case.