napari-mlarray 0.0.2__tar.gz → 0.0.3__tar.gz

{napari_mlarray-0.0.2 → napari_mlarray-0.0.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: napari-mlarray
-Version: 0.0.2
+Version: 0.0.3
 Summary: A reader/writer Napari plugin for MLArray images.
 Author: Karol-G
 Author-email: karol.gotkowski@dkfz.de
@@ -45,6 +45,7 @@ Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: mlarray
 Requires-Dist: numpy
+Requires-Dist: napari-bbox-fix
 Provides-Extra: all
 Requires-Dist: napari[all]; extra == "all"
 Dynamic: license-file

{napari_mlarray-0.0.2 → napari_mlarray-0.0.3}/pyproject.toml

@@ -30,7 +30,8 @@ requires-python = ">=3.10"
 # See best practices: https://napari.org/stable/plugins/building_a_plugin/best_practices.html
 dependencies = [
     "mlarray",
-    "numpy"
+    "numpy",
+    "napari-bbox-fix"
 ]
 
 [project.optional-dependencies]

{napari_mlarray-0.0.2 → napari_mlarray-0.0.3}/src/napari_mlarray/_reader.py

@@ -9,68 +9,26 @@ from mlarray import MLArray
 from pathlib import Path
 import numpy as np
 
+# Ensure napari-bbox registers its custom layer type.
+import napari_bbox  # noqa: F401
+
 
 def napari_get_reader(path):
-    """A basic implementation of a Reader contribution.
-
-    Parameters
-    ----------
-    path : str or list of str
-        Path to file, or list of paths.
-
-    Returns
-    -------
-    function or None
-        If the path is a recognized format, return a function that accepts the
-        same path or list of paths, and returns a list of layer data tuples.
-    """
+    """A basic implementation of a Reader contribution."""
     if isinstance(path, list):
-        # reader plugins may be handed single path, or a list of paths.
-        # if it is a list, it is assumed to be an image stack...
-        # so we are only going to look at the first file.
         path = path[0]
 
-    # the get_reader function should make as many checks as possible
-    # (without loading the full file) to determine if it can read
-    # the path. Here, we check the dtype of the array by loading
-    # it with memmap, so that we don't actually load the full array into memory.
-    # We pretend that this reader can only read integer arrays.
     try:
         if not str(path).endswith(".mla"):
             return None
-    # napari_get_reader should never raise an exception, because napari
-    # raises its own specific errors depending on what plugins are
-    # available for the given path, so we catch
-    # the OSError that np.load might raise if the file is malformed
     except OSError:
         return None
 
-    # otherwise we return the *function* that can read ``path``.
     return reader_function
 
 
 def reader_function(path):
-    """Take a path or list of paths and return a list of LayerData tuples.
-
-    Readers are expected to return data as a list of tuples, where each tuple
-    is (data, [add_kwargs, [layer_type]]), "add_kwargs" and "layer_type" are
-    both optional.
-
-    Parameters
-    ----------
-    path : str or list of str
-        Path to file, or list of paths.
-
-    Returns
-    -------
-    layer_data : list of tuples
-        A list of LayerData tuples where each tuple in the list contains
-        (data, metadata, layer_type), where data is a numpy array, metadata is
-        a dict of keyword arguments for the corresponding viewer.add_* method
-        in napari, and layer_type is a lower-case string naming the type of
-        layer. Both "meta", and "layer_type" are optional. napari will
-        default to layer_type=="image" if not provided
-    """
+    """Take a path or list of paths and return a list of LayerData tuples."""
     paths = [path] if isinstance(path, str) else path
     layer_data = []
     for path in paths:
@@ -82,30 +40,58 @@ def reader_function(path):
             layer_type = "labels" if mlarray.meta.is_seg.is_seg == True else "image"
             layer_data.append((data, metadata, layer_type))
         if mlarray.meta.bbox.bboxes is not None:
-            data = bboxes_minmax_to_napari_rectangles_2d(mlarray.meta.bbox.bboxes)
-            edge_color = _napari_bbox_edge_colors(
-                data,
-                labels=getattr(mlarray.meta.bbox, "labels", None),
-            )
-            text = _napari_bbox_score_text(
-                scores=getattr(mlarray.meta.bbox, "scores", None),
-                labels=getattr(mlarray.meta.bbox, "labels", None),
-                count=len(data),
-                edge_color=edge_color,
-                rectangles=data,
-            )
-            metadata = {
-                "name": f"{name} (BBoxes)",
-                "shape_type": "rectangle",
-                "affine": mlarray.affine,
-                "metadata": mlarray.meta.to_mapping(),
-                "face_color": "transparent",
-                "edge_color": edge_color,
-            }
-            if text is not None:
-                metadata["text"] = text
-            layer_type = "shapes"
-            layer_data.append((data, metadata, layer_type))
+            bboxes = np.asarray(mlarray.meta.bbox.bboxes)
+
+            # MLArray bboxes are always (N, D, 2)
+            if bboxes.ndim != 3 or bboxes.shape[2] != 2:
+                raise ValueError(f"Unsupported bbox shape: {bboxes.shape}")
+
+            dims = bboxes.shape[1]
+
+            # 2D -> keep shapes rectangles (original behavior)
+            if dims == 2:
+                data = bboxes_minmax_to_napari_rectangles_2d(bboxes)
+                edge_color = _napari_bbox_edge_colors(
+                    data,
+                    labels=getattr(mlarray.meta.bbox, "labels", None),
+                )
+                text = _napari_bbox_score_text(
+                    scores=getattr(mlarray.meta.bbox, "scores", None),
+                    labels=getattr(mlarray.meta.bbox, "labels", None),
+                    count=len(data),
+                    edge_color=edge_color,
+                    rectangles=data,
+                )
+                metadata = {
+                    "name": f"{name} (BBoxes)",
+                    "shape_type": "rectangle",
+                    "affine": mlarray.affine,
+                    "metadata": mlarray.meta.to_mapping(),
+                    "face_color": "transparent",
+                    "edge_color": edge_color,
+                }
+                if text is not None:
+                    metadata["text"] = text
+                layer_type = "shapes"
+                layer_data.append((data, metadata, layer_type))
+
+            # 3D+ -> napari-bbox layer
+            elif dims >= 3:
+                data = bboxes_minmax_to_napari_bboxes_nd(bboxes)
+                edge_color = _napari_bbox_edge_colors_count(
+                    count=len(data),
+                    labels=getattr(mlarray.meta.bbox, "labels", None),
+                )
+                metadata = {
+                    "name": f"{name} (BBoxes)",
+                    "affine": mlarray.affine,
+                    "metadata": mlarray.meta.to_mapping(),
+                    "face_color": "transparent",
+                    "edge_color": edge_color,
+                    # "edge_width": 2,
+                }
+                layer_type = "boundingboxlayer"
+                layer_data.append((data, metadata, layer_type))
     return layer_data
 
 
@@ -115,34 +101,14 @@ def bboxes_minmax_to_napari_rectangles_2d(
     dtype=np.float32,
     validate: bool = True,
 ) -> np.ndarray:
-    """
-    Convert 2D axis-aligned bounding boxes from min/max format to napari Shapes rectangles.
-
-    Accepted input formats (both mean the same thing):
-      1) (N, 2, 2): [[min_dim0, max_dim0], [min_dim1, max_dim1]]
-         Example (dim order is whatever you use, e.g. (y, x)):
-           [[[ymin, ymax], [xmin, xmax]], ...]
-
-      2) (N, 4): [min_dim0, min_dim1, max_dim0, max_dim1]
-         Example:
-           [[ymin, xmin, ymax, xmax], ...]
-
-    Output format (napari Shapes rectangle vertices):
-      (N, 4, 2) with vertices in non-twisting cyclic order:
-        (min0, min1) -> (min0, max1) -> (max0, max1) -> (max0, min1)
-
-    Raises:
-      ValueError if bboxes are not 2D (i.e., D != 2) or shapes are invalid.
-    """
+    """Convert 2D axis-aligned bounding boxes from min/max format to napari Shapes rectangles."""
     arr = np.asarray(bboxes)
 
-    # Normalize input to shape (N, 2, 2)
     if arr.ndim == 2 and arr.shape[1] == 4:
-        # (N, 4) -> (N, 2, 2)
         arr = np.stack(
             [
-                arr[:, [0, 2]],  # dim0: [min0, max0]
-                arr[:, [1, 3]],  # dim1: [min1, max1]
+                arr[:, [0, 2]],
+                arr[:, [1, 3]],
             ],
             axis=1,
         )
@@ -153,13 +119,20 @@ def bboxes_minmax_to_napari_rectangles_2d(
             f"Expected bboxes of shape (N, 2, 2) or (N, 4). Got {arr.shape}."
         )
 
-    N, D, two = arr.shape
+    # MLArray uses (N, D, 2) -> convert to (N, 2, 2)
+    if arr.shape == (arr.shape[0], 2, 2):
+        arr2 = arr
+    else:
+        arr2 = np.transpose(arr, (0, 2, 1))
+
+    N, D, two = arr2.shape
     if D != 2 or two != 2:
-        # Defensive; should never hit because of checks above.
         raise ValueError(f"Only 2D bboxes are supported. Got (N, {D}, {two}).")
 
-    mins = arr[:, :, 0]
-    maxs = arr[:, :, 1]
+    mins = arr2[:, 0, :]
+    maxs = arr2[:, 1, :]
+    # Ensure proper min/max ordering even if input is flipped
+    mins, maxs = np.minimum(mins, maxs), np.maximum(mins, maxs)
 
     if validate and np.any(maxs < mins):
         bad = np.argwhere(maxs < mins)
@@ -171,7 +144,6 @@ def bboxes_minmax_to_napari_rectangles_2d(
     min0, min1 = mins[:, 0], mins[:, 1]
     max0, max1 = maxs[:, 0], maxs[:, 1]
 
-    # Cyclic order (no twisting):
     rects = np.stack(
         [
             np.stack([min0, min1], axis=1),
@@ -185,6 +157,40 @@ def bboxes_minmax_to_napari_rectangles_2d(
     return rects
 
 
+def bboxes_minmax_to_napari_bboxes_nd(
+    bboxes,
+    *,
+    dtype=np.float32,
+    validate: bool = True,
+):
+    """
+    Convert N-D axis-aligned bboxes from min/max to napari-bbox format.
+    Input (MLArray): (N, D, 2) where [:, :, 0] are mins and [:, :, 1] are maxs.
+    Returns:
+      - list of (2, D) arrays, one per bbox.
+    """
+    arr = np.asarray(bboxes)
+
+    if arr.ndim != 3 or arr.shape[2] != 2:
+        raise ValueError(
+            f"Expected bboxes of shape (N, D, 2). Got {arr.shape}."
+        )
+
+    mins = arr[:, :, 0]
+    maxs = arr[:, :, 1]
+    # Ensure proper min/max ordering even if input is flipped
+    mins, maxs = np.minimum(mins, maxs), np.maximum(mins, maxs)
+    if validate and np.any(maxs < mins):
+        bad = np.argwhere(maxs < mins)
+        raise ValueError(
+            "Found bbox with max < min at indices (bbox_index, dim): "
+            f"{bad[:10].tolist()}" + (" ..." if len(bad) > 10 else "")
+        )
+
+    arr2 = np.stack([mins, maxs], axis=1).astype(dtype, copy=False)
+    return [arr2[i] for i in range(arr2.shape[0])]
+
+
 def _napari_bbox_edge_colors(rectangles, labels):
     """Return RGBA edge colors for each bbox."""
     count = len(rectangles)
@@ -203,6 +209,23 @@ def _napari_bbox_edge_colors(rectangles, labels):
     return colors
 
 
+def _napari_bbox_edge_colors_count(count, labels=None):
+    """Return RGBA edge colors for each bbox (count-based)."""
+    if count == 0:
+        return np.empty((0, 4), dtype=np.float32)
+
+    if labels is not None and len(labels) == count:
+        unique_labels = list(dict.fromkeys(labels))
+        label_to_color = {
+            label: _palette_rgba(idx) for idx, label in enumerate(unique_labels)
+        }
+        colors = np.array([label_to_color[label] for label in labels], dtype=np.float32)
+    else:
+        colors = np.array([_palette_rgba(idx) for idx in range(count)], dtype=np.float32)
+
+    return colors
+
+
 def _napari_bbox_score_text(scores, labels, count, edge_color, rectangles):
     """Return napari Shapes text metadata if scores are provided."""
     have_scores = scores is not None and len(scores) == count
@@ -210,7 +233,6 @@ def _napari_bbox_score_text(scores, labels, count, edge_color, rectangles):
     if not have_scores and not have_labels:
         return None
 
-    # Place text at the top-left corner of each rectangle.
     top_left = rectangles[:, 0, :]
     top_left = np.maximum(top_left - np.array([4.0, 0.0], dtype=top_left.dtype), 0)
 
@@ -221,7 +243,6 @@ def _napari_bbox_score_text(scores, labels, count, edge_color, rectangles):
             parts.append(f"Label: {labels[idx]}")
         if have_scores:
             parts.append(f"Score: {scores[idx]:.3f}")
-        # Add a trailing empty line to create spacing below the score.
         parts.append("\n")
         strings.append("\n".join(parts))
 

{napari_mlarray-0.0.2 → napari_mlarray-0.0.3}/src/napari_mlarray/_version.py

@@ -28,7 +28,7 @@ version_tuple: VERSION_TUPLE
 commit_id: COMMIT_ID
 __commit_id__: COMMIT_ID
 
-__version__ = version = '0.0.2'
-__version_tuple__ = version_tuple = (0, 0, 2)
+__version__ = version = '0.0.3'
+__version_tuple__ = version_tuple = (0, 0, 3)
 
-__commit_id__ = commit_id = 'gb43d6a1af'
+__commit_id__ = commit_id = 'g5aeee1d3c'

{napari_mlarray-0.0.2 → napari_mlarray-0.0.3}/src/napari_mlarray.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: napari-mlarray
-Version: 0.0.2
+Version: 0.0.3
 Summary: A reader/writer Napari plugin for MLArray images.
 Author: Karol-G
 Author-email: karol.gotkowski@dkfz.de
@@ -45,6 +45,7 @@ Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: mlarray
 Requires-Dist: numpy
+Requires-Dist: napari-bbox-fix
 Provides-Extra: all
 Requires-Dist: napari[all]; extra == "all"
 Dynamic: license-file

{napari_mlarray-0.0.2 → napari_mlarray-0.0.3}/src/napari_mlarray.egg-info/requires.txt

@@ -1,5 +1,6 @@
 mlarray
 numpy
+napari-bbox-fix
 
 [all]
 napari[all]