NGFF Schema validation

ngff_zarr/from_ngff_zarr.py

@@ -73,8 +73,15 @@ def from_ngff_zarr(
 
     metadata.pop("@type", None)
     axes = [Axis(**axis) for axis in metadata["axes"]]
+    coordinateTransformations = None
+    if "coordinateTransformations" in metadata:
+        coordinateTransformations = metadata["coordinateTransformations"]
     metadata = Metadata(
-        axes=axes, datasets=datasets, name=name, version=metadata["version"]
+        axes=axes,
+        datasets=datasets,
+        name=name,
+        version=metadata["version"],
+        coordinateTransformations=coordinateTransformations,
     )
 
     return Multiscales(images, metadata)

ngff_zarr/to_multiscales.py

@@ -363,6 +363,11 @@ def to_multiscales(
             path=path, coordinateTransformations=coordinateTransformations
         )
         datasets.append(dataset)
-    metadata = Metadata(axes=axes, datasets=datasets, name=ngff_image.name)
+    metadata = Metadata(
+        axes=axes,
+        datasets=datasets,
+        name=ngff_image.name,
+        coordinateTransformations=None,
+    )
 
     return Multiscales(images, metadata, scale_factors, method, out_chunks)

ngff_zarr/to_ngff_zarr.py

@@ -16,6 +16,17 @@
 from .to_multiscales import to_multiscales
 
 
+def _pop_metadata_optionals(metadata_dict):
+    for ax in metadata_dict["axes"]:
+        if ax["unit"] is None:
+            ax.pop("unit")
+
+    if metadata_dict["coordinateTransformations"] is None:
+        metadata_dict.pop("coordinateTransformations")
+
+    return metadata_dict
+
+
 def to_ngff_zarr(
     store: Union[MutableMapping, str, Path, BaseStore],
     multiscales: Multiscales,
@@ -55,6 +66,7 @@ def to_ngff_zarr(
     """
 
     metadata_dict = asdict(multiscales.metadata)
+    metadata_dict = _pop_metadata_optionals(metadata_dict)
     metadata_dict["@type"] = "ngff:Image"
     root = zarr.group(store, overwrite=overwrite, chunk_store=chunk_store)
     root.attrs["multiscales"] = [metadata_dict]

ngff_zarr/zarr_metadata.py

@@ -1,4 +1,4 @@
-from dataclasses import dataclass, field
+from dataclasses import dataclass
 from typing import List, Optional, Union
 
 from typing_extensions import Literal
@@ -139,6 +139,11 @@ class Axis:
     unit: Optional[Units] = None
 
 
+@dataclass
+class Identity:
+    type: str = "identity"
+
+
 @dataclass
 class Scale:
     scale: List[float]
@@ -164,6 +169,6 @@ class Dataset:
 class Metadata:
     axes: List[Axis]
     datasets: List[Dataset]
+    coordinateTransformations: Optional[List[Transform]]
     name: str = "image"
     version: str = "0.4"
-    coordinateTransformations: List[Transform] = field(default_factory=list)

pyproject.toml

@@ -76,6 +76,7 @@ test = [
     "itk-io>=5.3.0",
     "itk-filtering>=5.3.0",
     "tifffile",
+    "jsonschema",
 ]
 
 

test/_data.py

@@ -7,8 +7,8 @@
 from ngff_zarr import itk_image_to_ngff_image, to_ngff_zarr
 from zarr.storage import DirectoryStore, MemoryStore
 
-test_data_ipfs_cid = "bafybeidycdocaf7muhwt73m6vxfeaxfidbq7prvnhgn4tzrpxmpui5rila"
-test_data_sha256 = "675a2cb532cb6fbabfd2b1e19f26f05b059ac51f0b611eed1164d8dd1a22e4cb"
+test_data_ipfs_cid = "bafybeifkrfjreee5e7rfoi2ka3t2n23jlqwsrqwg4gmbyqcrc47iet27km"
+test_data_sha256 = "861dd3acb8391a3ac2b10b3487c33ee74da1415ef30b9f4c37c7f8f49e802c32"
 
 test_dir = Path(__file__).resolve().parent
 extract_dir = "data"
@@ -67,14 +67,14 @@ def store_equals(baseline_store, test_store):
     test_keys = set(test_store.keys())
     if baseline_keys != test_keys:
         sys.stderr.write("test keys != baseline keys\n")
-        sys.stderr.write("baseline - test:", baseline_keys.difference(test_keys), "\n")
-        sys.stderr.write("test - baseline:", test_keys.difference(baseline_keys), "\n")
+        sys.stderr.write(f"baseline - test: {baseline_keys.difference(test_keys)}, \n")
+        sys.stderr.write(f"test - baseline: {test_keys.difference(baseline_keys)}, \n")
         return False
     for k in baseline_keys:
         if baseline_store[k] != test_store[k]:
             sys.stderr.write(f"test value != baseline value for key {k}\n")
-            sys.stderr.write("baseline:", baseline_store[k], "\n")
-            sys.stderr.write("test:", test_store[k], "\n")
+            sys.stderr.write(f"baseline: {baseline_store[k]}, \n")
+            sys.stderr.write(f"test: {test_store[k]}, \n")
             return False
     return True
 

test/test_from_ngff_zarr.py

@@ -1,6 +1,5 @@
 from dask_image import imread
 from ngff_zarr import (
-    Methods,
     from_ngff_zarr,
     to_multiscales,
     to_ngff_image,
@@ -11,15 +10,6 @@
 from ._data import test_data_dir, verify_against_baseline
 
 
-def test_gaussian_isotropic_scale_factors(input_images):
-    dataset_name = "cthead1"
-    image = input_images[dataset_name]
-    baseline_name = "2_4/DASK_IMAGE_GAUSSIAN.zarr"
-    multiscales = to_multiscales(image, [2, 4], method=Methods.DASK_IMAGE_GAUSSIAN)
-    # store_new_multiscales(dataset_name, f'{baseline_name}', multiscales)
-    verify_against_baseline(dataset_name, baseline_name, multiscales)
-
-
 def test_from_ngff_zarr(input_images):
     dataset_name = "lung_series"
     data = imread.imread(input_images[dataset_name])

test/test_large_serialization.py

@@ -4,6 +4,7 @@
 
 
 def test_large_image_serialization(input_images):
+    default_mem_target = config.memory_target
     config.memory_target = int(1e6)
 
     dataset_name = "lung_series"
@@ -21,3 +22,5 @@ def test_large_image_serialization(input_images):
     test_store = MemoryStore(dimension_separator="/")
     to_ngff_zarr(test_store, multiscales)
     # verify_against_baseline(dataset_name, baseline_name, multiscales)
+
+    config.memory_target = default_mem_target

test/test_ngff_validation.py

@@ -0,0 +1,76 @@
+import json
+from typing import Dict
+
+import numpy as np
+import urllib3
+import zarr
+from jsonschema import Draft202012Validator
+from ngff_zarr import Multiscales, to_multiscales, to_ngff_zarr
+from referencing import Registry, Resource
+
+HTTP = urllib3.PoolManager()
+
+NGFF_URI = "https://ngff.openmicroscopy.org"
+
+
+def load_schema(version: str = "0.4", strict: bool = False) -> Dict:
+    strict_str = ""
+    if strict:
+        strict_str = "strict_"
+    response = HTTP.request(
+        "GET", f"{NGFF_URI}/{version}/schemas/{strict_str}image.schema"
+    )
+    return json.loads(response.data.decode())
+
+
+def check_valid_ngff(multiscale: Multiscales):
+    store = zarr.storage.MemoryStore(dimension_separator="/")
+    store = zarr.storage.DirectoryStore("/tmp/test.zarr", dimension_separator="/")
+    to_ngff_zarr(store, multiscale)
+    zarr.convenience.consolidate_metadata(store)
+    metadata = json.loads(store.get(".zmetadata"))["metadata"]
+    ngff = metadata[".zattrs"]
+
+    image_schema = load_schema(version="0.4", strict=False)
+    # strict_image_schema = load_schema(version="0.4", strict=True)
+    registry = Registry().with_resource(
+        NGFF_URI, resource=Resource.from_contents(image_schema)
+    )
+    validator = Draft202012Validator(image_schema, registry=registry)
+    # registry_strict = Registry().with_resource(NGFF_URI, resource=Resource.from_contents(strict_image_schema))
+    # strict_validator = Draft202012Validator(strict_schema, registry=registry_strict)
+
+    validator.validate(ngff)
+    # Need to add NGFF metadata property
+    # strict_validator.validate(ngff)
+
+
+def test_y_x_valid_ngff():
+    array = np.random.random((32, 16))
+    multiscale = to_multiscales(array, [2, 4])
+
+    check_valid_ngff(multiscale)
+
+
+# def test_z_y_x_valid_ngff():
+#     array = np.random.random((32, 32, 16))
+#     image = to_spatial_image(array)
+#     multiscale = to_multiscale(image, [2, 4])
+
+#     check_valid_ngff(multiscale)
+
+
+# def test_z_y_x_c_valid_ngff():
+#     array = np.random.random((32, 32, 16, 3))
+#     image = to_spatial_image(array)
+#     multiscale = to_multiscale(image, [2, 4])
+
+#     check_valid_ngff(multiscale)
+
+
+# def test_t_z_y_x_c_valid_ngff():
+#     array = np.random.random((2, 32, 32, 16, 3))
+#     image = to_spatial_image(array)
+#     multiscale = to_multiscale(image, [2, 4])
+
+#     check_valid_ngff(multiscale)

test/test_to_ngff_zarr_dask_image.py

@@ -15,10 +15,13 @@ def test_gaussian_isotropic_scale_factors(input_images):
     multiscales = to_multiscales(image, method=Methods.DASK_IMAGE_GAUSSIAN)
     verify_against_baseline(dataset_name, baseline_name, multiscales)
 
+
+def test_gaussian_isotropic_scale_factors_two_components(input_images):
     dataset_name = "brain_two_components"
     image = input_images[dataset_name]
     baseline_name = "2_4/DASK_IMAGE_GAUSSIAN.zarr"
     multiscales = to_multiscales(image, [2, 4], method=Methods.DASK_IMAGE_GAUSSIAN)
+    # store_new_multiscales(dataset_name, baseline_name, multiscales)
     verify_against_baseline(dataset_name, baseline_name, multiscales)