nImage.get_layer_data_tuple for a ready-for-napari approach (#37)

# Description

This PR was built out of the need to support reading data that can be
added as an image and/or labels layer. I thought more closely about
things, and realized that I can collect the information more reliably as
a layer data tuple, and take advantage of napari's machinery for such
endeavour. This now makes `nImage` serve effectively one function
besides the `BioImage` objects it already inherits:
`layer_data_tuples = nImage(...).get_layer_data_tuples`. Everything else
is private methods to help achieve this.
Now, various things can be passed to this method, including
channel_kwargs for overrides, but otherwise sensible defaults.

The only disadvantage of this approach is that I had to remove the
ease-of-use `channel_axis` kwarg to `add_image`, but I think I finally
understand the napari discussions for why that should be removed -- it
just doesn't translate well to other Layer Data Tuples. So now, I fully
handle channel splitting myself, and make that the only way to process
channel images. No more can channels be not split.

## Copilot Summary

This pull request refactors the way multichannel images are handled and
displayed in napari, improving channel splitting and metadata handling,
and updating related tests and plugin configuration. The main changes
include splitting image channels into separate napari layers,
introducing colormap utilities, updating tests to reflect the new layer
structure, and removing unused writer plugin definitions.

**Image Layer Handling and Colormap Utilities**
* Added a new module `src/ndevio/_colormap_utils.py` to provide colormap
cycles and a utility for assigning colormaps to channels, supporting
napari's conventions for single, dual, and multi-channel images.
* Refactored image reading logic in `src/ndevio/_napari_reader.py` and
related widget code to split channels into separate napari layers,
returning lists of `LayerDataTuple` objects. This enables more accurate
visualization and metadata per channel.
[[1]](diffhunk://#diff-a1495b29f6e07c8278a427f1ad3f4b1033da9353d263c050f0c8d9ac1b2aeebdL98-L112)
[[2]](diffhunk://#diff-a1495b29f6e07c8278a427f1ad3f4b1033da9353d263c050f0c8d9ac1b2aeebdL126-L162)
[[3]](diffhunk://#diff-95cabfe66da31c6e66e52a789bb410721325aa105aacbc8a75b79194a9974351L124-R136)

**Testing Updates**
* Updated tests in `tests/test_napari_reader.py` and
`tests/test_nimage.py` to validate that multichannel images are split
into separate layers, and to check metadata and layer tuple structure
accordingly. Test names and assertions were revised to match the new
logic.
[[1]](diffhunk://#diff-0e2609b36d787eac366ca4e6829e2509c539d1615bb1a700973c91176a6879e6R37-R39)
[[2]](diffhunk://#diff-0e2609b36d787eac366ca4e6829e2509c539d1615bb1a700973c91176a6879e6L48-R63)
[[3]](diffhunk://#diff-0e2609b36d787eac366ca4e6829e2509c539d1615bb1a700973c91176a6879e6R73)
[[4]](diffhunk://#diff-0e2609b36d787eac366ca4e6829e2509c539d1615bb1a700973c91176a6879e6L83-R95)
[[5]](diffhunk://#diff-6717c11ecd6a087070b7996a7e347caff169ea4db335e1edde84909b605ffbe1L35-R35)
[[6]](diffhunk://#diff-6717c11ecd6a087070b7996a7e347caff169ea4db335e1edde84909b605ffbe1L133-R166)
[[7]](diffhunk://#diff-6717c11ecd6a087070b7996a7e347caff169ea4db335e1edde84909b605ffbe1L177)
[[8]](diffhunk://#diff-6717c11ecd6a087070b7996a7e347caff169ea4db335e1edde84909b605ffbe1L187-R213)
[[9]](diffhunk://#diff-6717c11ecd6a087070b7996a7e347caff169ea4db335e1edde84909b605ffbe1L217)
[[10]](diffhunk://#diff-6717c11ecd6a087070b7996a7e347caff169ea4db335e1edde84909b605ffbe1L229-R247)
[[11]](diffhunk://#diff-6717c11ecd6a087070b7996a7e347caff169ea4db335e1edde84909b605ffbe1L250-R263)
[[12]](diffhunk://#diff-6717c11ecd6a087070b7996a7e347caff169ea4db335e1edde84909b605ffbe1L269-R278)

**Plugin Configuration**
* Removed writer plugin definitions from `src/ndevio/napari.yaml` to
clean up unused functionality and focus the plugin on reading and
visualization.
[[1]](diffhunk://#diff-37781c656dd096700e29a43b6a5fb914d4e9f1632516709c538774e9fb0e5149L12-L17)
[[2]](diffhunk://#diff-37781c656dd096700e29a43b6a5fb914d4e9f1632516709c538774e9fb0e5149L62-L68)

Author: TimMonko

pixi.lock

@@ -0,0 +1,51 @@
+"""Colormap utilities for ndevio.
+
+This module provides colormap cycles for
+multichannel image display.
+"""
+
+SINGLE_CHANNEL_COLORMAP = "gray"
+
+TWO_CHANNEL_CYCLE = ["magenta", "green"]
+
+MULTI_CHANNEL_CYCLE = ["cyan", "magenta", "yellow", "blue", "green", "red"]
+
+RGB = ["red", "green", "blue"]
+
+
+def get_colormap_for_channel(channel_idx: int, n_channels: int) -> str:
+    """
+    Get colormap for a channel based on napari's defaults.
+
+    - 1 channel → gray
+    - 2 channels → magenta, green (TWO_CHANNEL_CYCLE)
+    - 3+ channels → cycles through MULTI_CHANNEL_CYCLE (CMYBGR)
+
+    Parameters
+    ----------
+    channel_idx : int
+        Index of the channel (0-based).
+    n_channels : int
+        Total number of channels in the image.
+
+    Returns
+    -------
+    str
+        Colormap name.
+
+    """
+    if n_channels == 1:
+        return SINGLE_CHANNEL_COLORMAP
+    elif n_channels == 2:
+        return TWO_CHANNEL_CYCLE[channel_idx % len(TWO_CHANNEL_CYCLE)]
+    else:
+        return MULTI_CHANNEL_CYCLE[channel_idx % len(MULTI_CHANNEL_CYCLE)]
+
+
+__all__ = [
+    "SINGLE_CHANNEL_COLORMAP",
+    "TWO_CHANNEL_CYCLE",
+    "MULTI_CHANNEL_CYCLE",
+    "RGB",
+    "get_colormap_for_channel",
+]

src/ndevio/_colormap_utils.py

@@ -11,7 +11,7 @@
 from .nimage import determine_reader_plugin, nImage
 
 if TYPE_CHECKING:
-    from napari.types import LayerData, PathLike, ReaderFunction
+    from napari.types import LayerDataTuple, PathLike, ReaderFunction
 
 logger = logging.getLogger(__name__)
 
@@ -95,21 +95,18 @@ def napari_reader_function(
     in_memory: bool | None = None,
     open_first_scene_only: bool = False,
     open_all_scenes: bool = False,
-    layer_type: str = "image",
-) -> list[LayerData] | None:
+) -> list[LayerDataTuple] | None:
     """
     Read a file using the given reader function.
 
     Parameters
     ----------
     path : PathLike
         Path to the file to be read
-    reader : None
-        Bioio Reader function to be used to read the file, by default None.
+    reader : Callable
+        Bioio Reader class to be used to read the file.
     in_memory : bool, optional
         Whether to read the file in memory, by default None.
-    layer_type : str, optional
-        Type of layer to be created in napari, by default 'image'.
     open_first_scene_only : bool, optional
         Whether to ignore multi-scene files and just open the first scene,
         by default False.
@@ -123,41 +120,23 @@ def napari_reader_function(
         List containing image data, metadata, and layer type
 
     """
-    if isinstance(path, list):
-        logger.info("Bioio: Expected a single path, got a list of paths.")
-        return None
-
     img = nImage(path, reader=reader)
-    in_memory = (
-        img._determine_in_memory(path) if in_memory is None else in_memory
-    )
-    # TODO: Guess layer type here (check channel names for labels?)
-    logger.info("Bioio: Reading in-memory: %s", in_memory)
+    logger.info("Bioio: Reading file with %d scenes", len(img.scenes))
 
     # open first scene only
     if len(img.scenes) == 1 or open_first_scene_only:
-        img_data = img.get_napari_image_data(in_memory=in_memory)
-        img_meta = img.get_napari_metadata(path)
-        return [(img_data.data, img_meta, layer_type)]
+        return img.get_layer_data_tuples(in_memory=in_memory)
 
-    # TODO: USE settings for open first or all scenes to set the nubmer of iterations of a for loop
-    # check napari reader settings stuff
     # open all scenes as layers
-    if len(img.scenes) > 1 and open_all_scenes:
+    if open_all_scenes:
         layer_list = []
         for scene in img.scenes:
             img.set_scene(scene)
-            img_data = img.get_napari_image_data(in_memory=in_memory)
-            img_meta = img.get_napari_metadata(path)
-            layer_list.append((img_data.data, img_meta, layer_type))
+            layer_list.extend(img.get_layer_data_tuples(in_memory=in_memory))
         return layer_list
 
-    # open scene widget
-    if len(img.scenes) > 1 and not open_all_scenes:
-        _open_scene_container(path=path, img=img, in_memory=in_memory)
-        return [(None,)]
-
-    logger.warning("Bioio: Error reading file")
+    # else: open scene widget
+    _open_scene_container(path=path, img=img, in_memory=in_memory)
     return [(None,)]
 
 

src/ndevio/_napari_reader.py

@@ -9,12 +9,6 @@ contributions:
     - id: ndevio.get_reader
       python_name: ndevio._napari_reader:napari_get_reader
       title: Open file with ndevio
-    - id: ndevio.write_multiple
-      python_name: ndevio._writer:write_multiple
-      title: Save multi-layer data with ndevio
-    - id: ndevio.write_single_image
-      python_name: ndevio._writer:write_single_image
-      title: Save image data with ndevio
     - id: ndevio.make_plugin_installer_widget
       python_name: ndevio.widgets:PluginInstallerWidget
       title: Install BioIO Reader Plugins
@@ -59,13 +53,6 @@ contributions:
         '*.wlz', '*.wmf', '*.wmv', '*.wpi', '*.xbm', '*.xdce', '*.xml', '*.xpm', '*.xqd',
         '*.xqf', '*.xv', '*.xys', '*.zfp', '*.zfr', '*.zip', '*.zpo', '*.zvi'
       ]
-  writers:
-    - command: ndevio.write_multiple
-      layer_types: ['image*','labels*']
-      filename_extensions: []
-    - command: ndevio.write_single_image
-      layer_types: ['image']
-      filename_extensions: ['.npy']
   widgets:
     - command: ndevio.make_plugin_installer_widget
       display_name: Install BioIO Reader Plugins

src/ndevio/napari.yaml

@@ -18,10 +18,6 @@ ndevio_reader:
     default: Open Scene Widget
     tooltip: How to handle files with multiple scenes
     value: Open Scene Widget
-  unpack_channels_as_layers:
-    default: true
-    tooltip: Whether to unpack multi-channel images as separate layers
-    value: true
   clear_layers_on_new_scene:
     default: false
     tooltip: Whether to clear the viewer when selecting a new scene