Fix OME metadata (#1272)

* OME 0.5 metadata + only for non-symlinks

* add tests

* changelog

* always resolve

* cache resolved_path

* cache resolved_path

Author: normanrz

webknossos/Changelog.md

@@ -17,8 +17,10 @@ For upgrade instructions, please check the respective _Breaking Changes_ section
 ### Added
 
 ### Changed
+- For Zarr3 datasets, the OME-Zarr 0.5 metadata is now written. [#1272](https://github.com/scalableminds/webknossos-libs/pull/1272)
 
 ### Fixed
+- Fixed an issue where it was attempted to overwrite the OME-Zarr metadata for symlinked layers. [#1272](https://github.com/scalableminds/webknossos-libs/pull/1272) 
 
 
 ## [2.0.3](https://github.com/scalableminds/webknossos-libs/releases/tag/v2.0.3) - 2025-03-18

webknossos/tests/dataset/test_dataset.py

@@ -230,7 +230,7 @@ def test_create_dataset_with_layer_and_mag(
 
 
 @pytest.mark.parametrize("output_path", [TESTOUTPUT_DIR, REMOTE_TESTOUTPUT_DIR])
-def test_ome_ngff_metadata(output_path: Path) -> None:
+def test_ome_ngff_0_4_metadata(output_path: Path) -> None:
     ds_path = prepare_dataset_path(DataFormat.Zarr, output_path)
     ds = Dataset(ds_path, voxel_size=(11, 11, 28))
     layer = ds.add_layer("color", COLOR_CATEGORY, data_format=DataFormat.Zarr)
@@ -272,6 +272,84 @@ def test_ome_ngff_metadata(output_path: Path) -> None:
     )
 
 
+@pytest.mark.parametrize("output_path", [TESTOUTPUT_DIR, REMOTE_TESTOUTPUT_DIR])
+def test_ome_ngff_0_5_metadata(output_path: Path) -> None:
+    ds_path = prepare_dataset_path(DataFormat.Zarr3, output_path)
+    ds = Dataset(ds_path, voxel_size=(11, 11, 28))
+    layer = ds.add_layer("color", COLOR_CATEGORY, data_format=DataFormat.Zarr3)
+    layer.add_mag("1")
+    layer.add_mag("2-2-1")
+
+    assert (ds_path / "zarr.json").exists()
+    assert (ds_path / "color" / "zarr.json").exists()
+    assert (ds_path / "color" / "1" / "zarr.json").exists()
+    assert (ds_path / "color" / "2-2-1" / "zarr.json").exists()
+
+    zattrs = json.loads((ds_path / "color" / "zarr.json").read_bytes())["attributes"]
+    assert zattrs["ome"]["version"] == "0.5"
+    assert len(zattrs["ome"]["multiscales"][0]["datasets"]) == 2
+    assert zattrs["ome"]["multiscales"][0]["datasets"][0]["coordinateTransformations"][
+        0
+    ]["scale"] == [
+        1,
+        11,
+        11,
+        28,
+    ]
+    assert zattrs["ome"]["multiscales"][0]["datasets"][1]["coordinateTransformations"][
+        0
+    ]["scale"] == [
+        1,
+        22,
+        22,
+        28,
+    ]
+
+    validate(
+        instance=zattrs,
+        schema=json.loads(
+            UPath(
+                "https://ngff.openmicroscopy.org/0.5/schemas/image.schema"
+            ).read_bytes()
+        ),
+    )
+
+
+def test_ome_ngff_0_5_metadata_symlink() -> None:
+    def recursive_chmod(ds_path: Path, mode: int) -> None:
+        # See https://docs.python.org/3/library/os.html#os.chmod for how to use mode
+        os.chmod(ds_path, mode)
+        for root, dirs, files in os.walk(ds_path):
+            root_path = Path(root)
+            for _dir in dirs:
+                path = root_path / _dir
+                os.chmod(path, mode)
+            for file in files:
+                path = root_path / file
+                os.chmod(path, mode)
+
+    ds_path = copy_simple_dataset(DEFAULT_DATA_FORMAT, TESTOUTPUT_DIR, "original")
+    # Add an additional segmentation layer to the original dataset
+    Dataset.open(ds_path).add_layer(
+        "segmentation", SEGMENTATION_CATEGORY, largest_segment_id=999
+    ).add_mag(1)
+
+    # remove write permissions
+    recursive_chmod(ds_path, 0o555)
+    try:
+        symlink_path = prepare_dataset_path(
+            DEFAULT_DATA_FORMAT, TESTOUTPUT_DIR, "with_symlink"
+        )
+        ds = Dataset(symlink_path, voxel_size=(1, 1, 1))
+
+        # add symlink color layer
+        ds.add_symlink_layer(ds_path / "color")
+
+    finally:
+        # restore write permissions
+        recursive_chmod(ds_path, 0o777)
+
+
 def test_create_default_layer() -> None:
     ds_path = prepare_dataset_path(DEFAULT_DATA_FORMAT, TESTOUTPUT_DIR)
     ds = Dataset(ds_path, voxel_size=(1, 1, 1))

webknossos/webknossos/dataset/dataset.py

@@ -66,7 +66,7 @@
     ZATTRS_FILE_NAME,
     ZGROUP_FILE_NAME,
 )
-from .ome_metadata import write_ome_0_4_metadata
+from .ome_metadata import write_ome_metadata
 from .remote_dataset_registry import RemoteDatasetRegistry
 from .remote_folder import RemoteFolder
 from .sampling_modes import SamplingModes
@@ -329,6 +329,7 @@ def __init__(
 
         """
         self._read_only = read_only
+        self._resolved_path: Optional[Path] = None
         self.path: Path = strip_trailing_slash(UPath(dataset_path))
 
         if count_defined_values((voxel_size, voxel_size_with_unit)) > 1:
@@ -472,8 +473,15 @@ def open(
         dataset = cls.__new__(cls)
         dataset.path = dataset_path
         dataset._read_only = read_only
+        dataset._resolved_path = None
         return dataset._init_from_properties(dataset_properties)
 
+    @property
+    def resolved_path(self) -> Path:
+        if self._resolved_path is None:
+            self._resolved_path = self.path.resolve()
+        return self._resolved_path
+
     @classmethod
     def announce_manual_upload(
         cls,
@@ -2399,12 +2407,12 @@ def add_remote_layer(
             mag.path = str(foreign_layer.mags[mag.mag].path)
         layer_properties.name = new_layer_name
         self._properties.data_layers += [layer_properties]
-        self._layers[new_layer_name] = self._initialize_layer_from_properties(
-            layer_properties
-        )
+        new_layer = self._initialize_layer_from_properties(layer_properties)
+        new_layer._resolved_path = foreign_layer_path
+        self._layers[new_layer_name] = new_layer
 
         self._export_as_json()
-        return self.layers[new_layer_name]
+        return new_layer
 
     def add_fs_copy_layer(
         self,
@@ -2794,7 +2802,9 @@ def _export_as_json(self) -> None:
                 json.dump({"zarr_format": 3, "node_type": "group"}, outfile, indent=4)
 
         for layer in self.layers.values():
-            write_ome_0_4_metadata(self, layer)
+            # Only write out OME metadata if the layer is a child of the dataset
+            if layer.resolved_path.parent == self.resolved_path:
+                write_ome_metadata(self, layer)
 
     def _initialize_layer_from_properties(self, properties: LayerProperties) -> Layer:
         if properties.category == COLOR_CATEGORY:

webknossos/webknossos/dataset/layer.py

@@ -216,6 +216,7 @@ def __init__(self, dataset: "Dataset", properties: LayerProperties) -> None:
             properties.element_class, properties.num_channels
         )
         self._mags: Dict[Mag, MagView] = {}
+        self._resolved_path: Optional[Path] = None
 
         for mag in properties.mags:
             self._setup_mag(Mag(mag.mag), mag.path)
@@ -260,6 +261,12 @@ def path(self) -> Path:
             else self.dataset.path / self.name
         )
 
+    @property
+    def resolved_path(self) -> Path:
+        if self._resolved_path is None:
+            self._resolved_path = self.path.resolve()
+        return self._resolved_path
+
     @property
     def is_remote_to_dataset(self) -> bool:
         """Whether this layer's data is stored remotely relative to its dataset.

webknossos/webknossos/dataset/ome_metadata.py

@@ -12,6 +12,53 @@
     from .layer import Layer
 
 
+def get_ome_0_5_multiscale_metadata(
+    dataset: "Dataset", layer: "Layer"
+) -> Dict[str, Any]:
+    return {
+        "ome": {
+            "version": "0.5",
+            "multiscales": [
+                {
+                    "axes": [
+                        {"name": "c", "type": "channel"},
+                        {
+                            "name": "x",
+                            "type": "space",
+                            "unit": "nanometer",
+                        },
+                        {
+                            "name": "y",
+                            "type": "space",
+                            "unit": "nanometer",
+                        },
+                        {
+                            "name": "z",
+                            "type": "space",
+                            "unit": "nanometer",
+                        },
+                    ],
+                    "datasets": [
+                        {
+                            "path": mag.path.name,
+                            "coordinateTransformations": [
+                                {
+                                    "type": "scale",
+                                    "scale": [1.0]
+                                    + (
+                                        np.array(dataset.voxel_size) * mag.mag.to_np()
+                                    ).tolist(),
+                                }
+                            ],
+                        }
+                        for mag in layer.mags.values()
+                    ],
+                }
+            ],
+        }
+    }
+
+
 def get_ome_0_4_multiscale_metadata(
     dataset: "Dataset", layer: "Layer"
 ) -> Dict[str, Any]:
@@ -57,7 +104,7 @@ def get_ome_0_4_multiscale_metadata(
     }
 
 
-def write_ome_0_4_metadata(dataset: "Dataset", layer: "Layer") -> None:
+def write_ome_metadata(dataset: "Dataset", layer: "Layer") -> None:
     if not is_writable_path(layer.path):
         return
     if layer.data_format == DataFormat.Zarr3:
@@ -66,7 +113,7 @@ def write_ome_0_4_metadata(dataset: "Dataset", layer: "Layer") -> None:
                 {
                     "zarr_format": 3,
                     "node_type": "group",
-                    "attributes": get_ome_0_4_multiscale_metadata(dataset, layer),
+                    "attributes": get_ome_0_5_multiscale_metadata(dataset, layer),
                 },
                 outfile,
                 indent=4,