Emitt warning when duplicate key with different value is found in OME metadata

Author: MSHelm

src/spatialdata_io/readers/macsima.py

@@ -354,6 +354,14 @@ def _collect_map_annotation_values(ome: OME) -> dict[str, Any]:
         for k, v in value.items():
             if k not in merged:
                 merged[k] = v
+            else:
+                # For instances where we have different values for the same repeated key, raise a warning.
+                if v != merged[k]:
+                    warnings.warn(
+                        f"Found different value for {k}: {v}. The parser will only use the first found value, which is {merged[k]}!",
+                        UserWarning,
+                        stacklevel=2,
+                    )
 
     return merged
 
@@ -382,11 +390,16 @@ def _get_software_major_version(version: str) -> int:
     if not parts:
         raise ValueError("Could not extract major software version part from version string.")
 
-    return int(parts[0])
+    major = int(parts[0])
+    logger.debug(f"Found major software version {major}")
+
+    return major
 
 
 def _parse_v0_ome_metadata(ome: OME) -> dict[str, Any]:
-    """Parse ome."""
+    """Parse Legacy Format of OME Metadata (software version 0.x.x)."""
+    logger.debug("Parsing OME metadata expecting version 0 format")
+
     metadata: dict[str, Any] = {
         "name": None,
         "clone": None,
@@ -466,7 +479,9 @@ def _parse_v0_ome_metadata(ome: OME) -> dict[str, Any]:
 
 
 def _parse_v1_ome_metadata(ome: OME) -> dict[str, Any]:
-    """Your existing MACSima-style parser (1.x.x layout)."""
+    """Parse v1 format of OME metadata (software version 1.x.x)."""
+    logger.debug("Parsing OME metadata expecting version 1 format")
+
     metadata: dict[str, Any] = {
         "name": None,
         "clone": None,

tests/test_macsima.py

@@ -258,21 +258,56 @@ def test_cli_macsima(runner: CliRunner, dataset: str) -> None:
         _ = read_zarr(output_zarr)
 
 
-def test_collect_map_annotation_values_merges_all_keys_first_wins() -> None:
+def test_collect_map_annotation_values_with_no_duplicate_keys() -> None:
     ome = OME(
         structured_annotations=StructuredAnnotations(
             map_annotations=[
                 MapAnnotation(value={"a": "1", "b": "2"}),
-                MapAnnotation(value={"b": "99", "c": "3"}),
+                MapAnnotation(value={"c": "3"}),
+            ]
+        )
+    )
+
+    result = _collect_map_annotation_values(ome)
+
+    assert result == {"a": "1", "b": "2", "c": "3"}
+
+
+def test_collect_map_annotations_values_with_duplicate_keys_identical_values() -> None:
+    ome = OME(
+        structured_annotations=StructuredAnnotations(
+            map_annotations=[
+                MapAnnotation(value={"a": "1", "b": "2"}),
+                MapAnnotation(value={"b": "2", "c": "3"}),
             ]
         )
     )
 
     result = _collect_map_annotation_values(ome)
+    # Key should only be returned once
+    assert result == {"a": "1", "b": "2", "c": "3"}
+
+
+def test_collect_map_annotations_values_with_duplicate_keys_different_values() -> None:
+    ome = OME(
+        structured_annotations=StructuredAnnotations(
+            map_annotations=[
+                MapAnnotation(value={"a": "1", "b": "2"}),
+                MapAnnotation(value={"b": "99", "c": "3"}),
+            ]
+        )
+    )
+    import re
+
+    # The parser should throw a warning when duplicate keys with different values are found
+    with pytest.warns(
+        UserWarning,
+        match=re.escape("Found different value for b: 99. The parser will only use the first found value, which is 2!"),
+    ):
+        result = _collect_map_annotation_values(ome)
 
+    # The parser should return only the first found value.
     assert result == {"a": "1", "b": "2", "c": "3"}
-    # ensure later duplicate "b" did not overwrite
-    assert result["b"] == "2"
 
 
 def test_collect_map_annotation_values_handles_missing_sa_and_empty_list() -> None: