Labels properties are a dict of lists

[will-moore]

Fixes #6.

This updates napari-ome-zarr to transform label properties into the format that napari is now expecting, using index as the label ID as described at https://forum.image.sc/t/napari-labels-layer-properties/57649

e.g.

    {
        "index": [1381342, 1381343...]
        "key1": [1381342, 1381343...],
        "key2": [1682567, 1682567...]
    }

To test, using data exported from omero-cli-zarr at PR: ome/omero-cli-zarr#82

napari https://minio-dev.openmicroscopy.org/idr/v0.3/idr0062-blin-nuclearsegmentation/6001247.zarr

Mousing over a label should show associated values in the footer:

[joshmoore]

Where does pooch get used, @will-moore ?

[will-moore]

See previous run failure: https://github.com/ome/napari-ome-zarr/runs/3651193426

.tox/py37-linux/lib/python3.7/site-packages/napari/conftest.py:5

>>> import pooch
ModuleNotFoundError

pooch is listed as a test dependency of napari.

[sbesson]

As pooch is a testing dependency of napari (https://github.com/napari/napari/blob/95a5d06fe36f10fb26f5798f1327a3ea082d581f/setup.cfg#L105), updating https://github.com/ome/napari-ome-zarr/blob/main/tox.ini#L33 with

-    napari
+   napari[testing]

would be another option. Does it bring too many extra dependencies?

[will-moore]

Just trying this locally into my existing napari & ome-zarr-py env:

$ pip install -U napari[testing]

This download 8 packages, mostly very small (few MB or less) except torch (218 MB). 🤷

[joshmoore]

With origin/main, I still see the error from #6: TypeError: len() of unsized object

With this PR, the image & label layers load.

I note that I see a separate error when viewing the labels in 3D:

/usr/local/anaconda3/envs/z/lib/python3.9/site-packages/napari/layers/labels/labels.py in _get_value_ray(self=<Labels layer '0'>, start_point=<class 'numpy.ndarray'> (3,) float64, end_point=<class 'numpy.ndarray'> (3,) float64, dims_displayed=[2, 3, 4])
    972                 sample_points, self._display_bounding_box(dims_displayed)
    973             ).astype(int)
--> 974             values = im_slice[tuple(clamped.T)]
        values = undefined
        im_slice = <class 'numpy.ndarray'> (257, 26, 31) int64
        global tuple = undefined
        clamped.T = <class 'numpy.ndarray'> (3, 504) int64
    975             nonzero_indices = np.flatnonzero(values)
    976             if len(nonzero_indices > 0):

IndexError: index 144 is out of bounds for axis 1 with size 26

[will-moore]

@joshmoore No idea what's up with the 3D labels, but I don't think it's related to this PR.
Again, shame we don't know when it last worked.

[joshmoore]

I created #14. Merging.

[imagesc-bot]

This pull request has been mentioned on Image.sc Forum. There might be relevant details there:

https://forum.image.sc/t/tool-for-adding-labels-group-to-ome-ngff/72839/4