ENH: Add support for itkVectorImage

[bnmajor]

Addresses #613

[bnmajor]

@thewtex Testing this out with the notebook that Paul shared and I get back the following:

---------------------------------------------------------------------------
KeyError                                  Traceback (most recent call last)
Input In [6], in <cell line: 1>()
----> 1 viewer = view(image=outputImage)

File ~/itkwidgets/itkwidgets/viewer.py:451, in view(data, **kwargs)
    337 def view(data=None, **kwargs):
    338     """View the image and/or point set.
    339 
    340     Creates and returns an ImJoy plugin ipywidget to visualize an image, and/or
   (...)
    449         properties on the object to change the visualization.
    450     """
--> 451     viewer = Viewer(data=data, **kwargs)
    453     return viewer

File ~/itkwidgets/itkwidgets/viewer.py:147, in Viewer.__init__(self, ui_collapsed, rotate, ui, **add_data_kwargs)
    143 def __init__(
    144     self, ui_collapsed=True, rotate=False, ui="pydata-sphinx", **add_data_kwargs
    145 ):
    146     input_data = self.input_data(add_data_kwargs)
--> 147     data = self.init_data(input_data)
    148     """Create a viewer."""
    149     self.viewer_rpc = ViewerRPC(
    150         ui_collapsed=ui_collapsed, rotate=rotate, ui=ui, data=data
    151     )

File ~/itkwidgets/itkwidgets/viewer.py:186, in Viewer.init_data(self, input_data)
    184         render_type = RenderType.LABELIMAGE
    185     else:
--> 186         result = _get_viewer_image(data, label=False)
    187 elif render_type is RenderType.POINT_SET:
    188     result = _get_viewer_point_set(data)

File ~/itkwidgets/itkwidgets/integrations/__init__.py:78, in _get_viewer_image(image, label)
     76 if isinstance(image, itk.Image) or isinstance(image, itk.VectorImage):
     77     ngff_image = itk_image_to_ngff_image(image)
---> 78     multiscales = to_multiscales(ngff_image, method=method)
     79     to_ngff_zarr(store, multiscales, chunk_store=chunk_store)
     80     return store

File ~/.virtualenvs/test-itkwidgets/lib/python3.8/site-packages/ngff_zarr/to_multiscales.py:118, in to_multiscales(data, scale_factors, method, chunks)
    115     method = Methods.DASK_IMAGE_GAUSSIAN
    117 if method is Methods.DASK_IMAGE_GAUSSIAN:
--> 118     images = _downsample_dask_image(
    119         ngff_image, default_chunks, out_chunks, scale_factors, label=False
    120     )
    121 elif method is Methods.DASK_IMAGE_NEAREST:
    122     images = _downsample_dask_image(
    123         ngff_image, default_chunks, out_chunks, scale_factors, label="nearest"
    124     )

File ~/.virtualenvs/test-itkwidgets/lib/python3.8/site-packages/ngff_zarr/methods/_dask_image.py:128, in _downsample_dask_image(ngff_image, default_chunks, out_chunks, scale_factors, label)
    121 # Compute output shape and metadata
    122 output_shape = [
    123     int(image_len / shrink_factor)
    124     for image_len, shrink_factor in zip(
    125         previous_image.data.shape, shrink_factors
    126     )
    127 ]
--> 128 output_scale = _compute_next_scale(previous_image, dim_factors)
    129 output_translation = _compute_next_translation(previous_image, dim_factors)
    131 if label == "mode":

File ~/.virtualenvs/test-itkwidgets/lib/python3.8/site-packages/ngff_zarr/methods/_dask_image.py:21, in _compute_next_scale(previous_image, dim_factors)
      8 """Helper method to manually compute output image spacing.
      9 
     10 previous_image: _NgffImage
   (...)
     18     Example {'x': 2.0, 'y': 1.0}
     19 """
     20 input_scale = previous_image.scale
---> 21 return {dim: input_scale[dim] * dim_factors[dim] for dim in previous_image.dims}

File ~/.virtualenvs/test-itkwidgets/lib/python3.8/site-packages/ngff_zarr/methods/_dask_image.py:21, in <dictcomp>(.0)
      8 """Helper method to manually compute output image spacing.
      9 
     10 previous_image: _NgffImage
   (...)
     18     Example {'x': 2.0, 'y': 1.0}
     19 """
     20 input_scale = previous_image.scale
---> 21 return {dim: input_scale[dim] * dim_factors[dim] for dim in previous_image.dims}

KeyError: 'c'

Any ideas?

[thewtex]

@bnmajor I rebased, added a commit to your branch that bumps ngff-zarr -- it brings in a fix for the backtrace but also performance improvements.