Redefine SummaryImage to allow for 3d summary images (#42)

* Add neurodata types for core and ndx-ophys-devices in namespace

* Add 3d SummaryImage

* Add description attribute to imaging space and segmentation classes

* Update mock segmentation functions to use ndx_microscopy.SummaryImage type for summary_images

Author: alessandratrapani

spec/ndx-microscopy.extensions.yaml

@@ -58,9 +58,6 @@ groups:
     neurodata_type_inc: NWBContainer
     doc: Abstract class to contain metadata about the region of physical space that imaging data was recorded from. Extended by PlanarImagingSpace and VolumetricImagingSpace.
     datasets:
-      - name: description
-        dtype: text
-        doc: Description of the imaging space.
       - name: origin_coordinates
         dtype: float64
         dims:
@@ -77,6 +74,9 @@ groups:
             default_value: micrometers
             doc: Measurement units for origin coordinates. The default value is 'micrometers'.
     attributes:
+      - name: description
+        dtype: text
+        doc: Description of the imaging space.
       - name: location
         dtype: text
         doc:
@@ -220,15 +220,39 @@ groups:
         doc: Results from image segmentation of a specific imaging space.
         quantity: "+"
 
-  - neurodata_type_def: SummaryImages
-    neurodata_type_inc: Images
+  - neurodata_type_def: SummaryImage
+    neurodata_type_inc: NWBContainer
     doc: Summary images that are related to the segmentation, e.g., mean, correlation, maximum projection.
+    datasets:
+      - name: data
+        doc: Summary image data.
+        dtype: numeric
+        dims:
+          - - height
+            - width
+          - - height
+            - width
+            - depth
+        shape:
+          - - null
+            - null
+          - - null
+            - null
+            - null
+    attributes:
+      - name: description
+        dtype: text
+        doc: Description of the summary image.
 
   - neurodata_type_def: Segmentation
     neurodata_type_inc: DynamicTable
     doc: Abstract class to contain the results from image segmentation of a specific imaging space.
+    attributes:
+      - name: description
+        dtype: text
+        doc: Description of the segmentation method used.
     groups:
-      - neurodata_type_inc: SummaryImages
+      - neurodata_type_inc: SummaryImage
         doc: Summary images that are related to the segmentation, e.g., mean, correlation, maximum projection.
         quantity: "*"
 

spec/ndx-microscopy.namespace.yaml

@@ -1,14 +1,30 @@
 namespaces:
-- name: ndx-microscopy
-  doc: Microscopy extension to NWB standard.
-  author:
-  - Cody Baker
-  - Alessandra Trapani
-  contact:
-  - cody.baker@catalystneuro.com
-  - alessandra.trapani@catalystneuro.com
-  schema:
-  - namespace: core
-  - namespace: ndx-ophys-devices
-  - source: ndx-microscopy.extensions.yaml
-  version: 0.1.0
+  - name: ndx-microscopy
+    doc: Microscopy extension to NWB standard.
+    author:
+      - Cody Baker
+      - Alessandra Trapani
+    contact:
+      - cody.baker@catalystneuro.com
+      - alessandra.trapani@catalystneuro.com
+    schema:
+      - namespace: core
+        neurodata_types:
+          - Device
+          - LabMetaData
+          - NWBContainer
+          - TimeSeries
+          - NWBDataInterface
+          - DynamicTable
+          - VectorData
+          - VectorIndex
+          - DynamicTableRegion
+      - namespace: ndx-ophys-devices
+        neurodata_types:
+          - ExcitationSource
+          - Indicator
+          - OpticalFilter
+          - Photodetector
+          - DichroicMirror
+      - source: ndx-microscopy.extensions.yaml
+    version: 0.1.0

src/pynwb/ndx_microscopy/__init__.py

@@ -35,7 +35,7 @@
 )
 
 Microscope = get_class("Microscope", extension_name)
-
+SummaryImage = get_class("SummaryImage", extension_name)
 PlanarImagingSpace = get_class("PlanarImagingSpace", extension_name)
 VolumetricImagingSpace = get_class("VolumetricImagingSpace", extension_name)
 

src/pynwb/ndx_microscopy/testing/_mock.py

@@ -114,23 +114,32 @@ def mock_VolumetricImagingSpace(
     return volumetric_imaging_space
 
 
+def mock_SummaryImage(
+    *,
+    name: Optional[str] = None,
+    description: str = "A mock instance of a SummaryImage type to be used for rapid testing.",
+    image_shape: Tuple[int, int] = (10, 10),
+    data: Optional[np.ndarray] = None,
+) -> ndx_microscopy.SummaryImage:
+    name = name or name_generator("SummaryImage")
+    data = data if data is not None else np.ones(image_shape)
+    summary_image = ndx_microscopy.SummaryImage(name=name, description=description, data=data)
+    return summary_image
+
+
 def mock_Segmentation(
     *,
     name: Optional[str] = None,
     description: str = "A mock instance of a Segmentation type to be used for rapid testing.",
-    summary_images: Optional[List[pynwb.base.Images]] = None,
+    summary_images: Optional[List[ndx_microscopy.SummaryImage]] = None,
 ) -> ndx_microscopy.Segmentation:
     """Base abstract class with summary images."""
     name = name or name_generator("Segmentation")
 
     # Create default summary images if none provided
     if summary_images is None:
-        mean_image = pynwb.base.Image(
-            name="mean", data=np.ones((10, 10)), description="Mean intensity projection"  # Example shape
-        )
-        max_image = pynwb.base.Image(
-            name="max", data=np.ones((10, 10)), description="Maximum intensity projection"  # Example shape
-        )
+        mean_image = mock_SummaryImage(name="mean", description="Mean intensity projection")
+        max_image = mock_SummaryImage(name="max", description="Maximum intensity projection")
         summary_images = [mean_image, max_image]
 
     segmentation = ndx_microscopy.Segmentation(name=name, description=description, summary_images=summary_images)
@@ -145,15 +154,15 @@ def mock_Segmentation2D(
     description: str = "A mock instance of a Segmentation2D type to be used for rapid testing.",
     number_of_rois: int = 5,
     image_shape: Tuple[int, int] = (10, 10),
-    summary_images: Optional[List[pynwb.base.Images]] = None,
+    summary_images: Optional[List[ndx_microscopy.SummaryImage]] = None,
 ) -> ndx_microscopy.Segmentation2D:
     """2D segmentation with image_mask/pixel_mask."""
     name = name or name_generator("Segmentation2D")
 
     # Create default summary images if none provided
     if summary_images is None:
-        mean_image = pynwb.base.Image(name="mean", data=np.ones(image_shape), description="Mean intensity projection")
-        max_image = pynwb.base.Image(name="max", data=np.ones(image_shape), description="Maximum intensity projection")
+        mean_image = mock_SummaryImage(name="mean", description="Mean intensity projection", image_shape=image_shape)
+        max_image = mock_SummaryImage(name="max", description="Maximum intensity projection", image_shape=image_shape)
         summary_images = [mean_image, max_image]
 
     segmentation_2D = ndx_microscopy.Segmentation2D(
@@ -181,23 +190,15 @@ def mock_Segmentation3D(
     description: str = "A mock instance of a Segmentation3D type to be used for rapid testing.",
     number_of_rois: int = 5,
     image_shape: Tuple[int, int, int] = (10, 10, 10),
-    summary_images: Optional[List[pynwb.base.Images]] = None,
+    summary_images: Optional[List[ndx_microscopy.SummaryImage]] = None,
 ) -> ndx_microscopy.Segmentation3D:
     """3D segmentation with image_mask/voxel_mask."""
     name = name or name_generator("Segmentation3D")
 
     # Create default summary images if none provided
     if summary_images is None:
-        mean_image = pynwb.base.Image(
-            name="mean",
-            data=np.ones((image_shape[0], image_shape[1])),  # Project along Z
-            description="Mean intensity projection",
-        )
-        max_image = pynwb.base.Image(
-            name="max",
-            data=np.ones((image_shape[0], image_shape[1])),  # Project along Z
-            description="Maximum intensity projection",
-        )
+        mean_image = mock_SummaryImage(name="mean", description="Mean intensity projection", image_shape=image_shape)
+        max_image = mock_SummaryImage(name="max", description="Maximum intensity projection", image_shape=image_shape)
         summary_images = [mean_image, max_image]
 
     volumetric_segmentation = ndx_microscopy.Segmentation3D(