ENH: Improve logic for infering render type

Assume data with two dimensions and shape (N, 2) or (N, 3) is a point set, all
other data is assumed to be an image.

itkwidgets/_initialization_params.py

@@ -29,11 +29,3 @@ def init_params_dict(itk_viewer):
         'y_slice': itk_viewer.setYSlice,
         'z_slice': itk_viewer.setZSlice,
     }
-
-def init_key_aliases():
-    return {
-        'data': 'image',
-        'image': 'image',
-        'label_image': 'labelImage',
-        'point_set': 'pointSets',
-    }

itkwidgets/integrations/__init__.py

@@ -154,6 +154,10 @@ def _get_viewer_point_set(point_set):
 
 
 def _detect_render_type(data, input_type) -> RenderType:
+    if input_type == 'image' or input_type == 'label_image':
+        return RenderType.IMAGE
+    elif input_type == 'point_set':
+        return RenderType.POINT_SET
     if isinstance(data, itkwasm.Image):
         return RenderType.IMAGE
     elif isinstance(data, itkwasm.PointSet):
@@ -164,12 +168,7 @@ def _detect_render_type(data, input_type) -> RenderType:
         # We may need to do more introspection
         return RenderType.IMAGE
     elif isinstance(data, np.ndarray):
-        if input_type == 'point_set':
-            return RenderType.POINT_SET
-        else:
-            return RenderType.IMAGE
-    elif isinstance(data, zarr.Group):
-        if input_type == 'point_set':
+        if data.ndim == 2 and data.shape[1] < 4:
             return RenderType.POINT_SET
         else:
             return RenderType.IMAGE
@@ -190,26 +189,26 @@ def _detect_render_type(data, input_type) -> RenderType:
         elif isinstance(data, vtk.vtkPolyData):
             return RenderType.POINT_SET
     if isinstance(data, dask.array.core.Array):
-        if input_type == 'point_set':
+        if data.ndim ==2 and data.shape[1] < 4:
             return RenderType.POINT_SET
         else:
             return RenderType.IMAGE
     if HAVE_TORCH:
         import torch
         if isinstance(data, torch.Tensor):
-            if input_type == 'point_set':
+            if data.dim == 2 and data.shape[1] < 4:
                 return RenderType.POINT_SET
             else:
                 return RenderType.IMAGE
     if HAVE_XARRAY:
         import xarray as xr
         if isinstance(data, xr.DataArray):
-            if input_type == 'point_set':
+            if data.dims == 2 and data.shape[1] < 4:
                 return RenderType.POINT_SET
             else:
                 return RenderType.IMAGE
         if isinstance(data, xr.Dataset):
-            if input_type == 'point_set':
+            if data.dims == 2 and data.shape[1] < 4:
                 return RenderType.POINT_SET
             else:
                 return RenderType.IMAGE

itkwidgets/render_types.py

@@ -3,5 +3,6 @@
 class RenderType(Enum):
     """Rendered data types"""
     IMAGE = "image"
+    LABELIMAGE = "labelImage"
     GEOMETRY = "geometry"
-    POINT_SET = "point_set"
+    POINT_SET = "pointSets"

itkwidgets/viewer.py

@@ -9,7 +9,7 @@
 import uuid
 
 from ._type_aliases import Gaussians, Style, Image, Point_Set
-from ._initialization_params import init_params_dict, init_key_aliases
+from ._initialization_params import init_params_dict
 from ._method_types import deferred_methods
 from .integrations import _detect_render_type, _get_viewer_image, _get_viewer_point_set
 from .integrations.environment import ENVIRONMENT, Env
@@ -169,17 +169,17 @@ def init_data(self, input_data):
         result= None
         for (input_type, data) in input_data:
             render_type = _detect_render_type(data, input_type)
-            key = init_key_aliases()[input_type]
             if render_type is RenderType.IMAGE:
                 if input_type == 'label_image':
                     result = _get_viewer_image(data, label=True)
+                    render_type = RenderType.LABELIMAGE
                 else:
                     result = _get_viewer_image(data, label=False)
             elif render_type is RenderType.POINT_SET:
                 result = _get_viewer_point_set(data)
             if result is None:
                 raise RuntimeError(f"Could not process the viewer {input_type}")
-            _init_data[key] = result
+            _init_data[render_type.value] = result
         return _init_data
 
     async def run_queued_requests(self):