Ignore extra keys in v2 metadata

Ignore unexpected keys in Zarr V2 metadata, to enable reading zarr
files written by other systems, which might store additional data
in the top level of the `.zgroup` and `.zarray` files`

Closes zarr-developers#2296

Author: TomAugspurger

src/zarr/core/group.py

@@ -3,7 +3,7 @@
 import asyncio
 import json
 import logging
-from dataclasses import asdict, dataclass, field, replace
+from dataclasses import asdict, dataclass, field, fields, replace
 from typing import TYPE_CHECKING, Literal, cast, overload
 
 import numpy as np
@@ -116,6 +116,15 @@ def __init__(
     @classmethod
     def from_dict(cls, data: dict[str, Any]) -> GroupMetadata:
         assert data.pop("node_type", None) in ("group", None)
+
+        zarr_format = data.get("zarr_format")
+        if zarr_format == 2 or zarr_format is None:
+            # zarr v2 allowed arbitrary keys here.
+            # We don't want the GroupMetadata constructor to fail just because someone put an
+            # extra key in the metadata.
+            expected = {x.name for x in fields(cls)}
+            data = {k: v for k, v in data.items() if k in expected}
+
         return cls(**data)
 
     def to_dict(self) -> dict[str, Any]:

src/zarr/core/metadata/v2.py

@@ -13,7 +13,7 @@
     from zarr.core.common import JSON, ChunkCoords
 
 import json
-from dataclasses import dataclass, field, replace
+from dataclasses import dataclass, field, fields, replace
 
 import numcodecs
 import numpy as np
@@ -140,6 +140,13 @@ def from_dict(cls, data: dict[str, Any]) -> ArrayV2Metadata:
         _data = data.copy()
         # check that the zarr_format attribute is correct
         _ = parse_zarr_format(_data.pop("zarr_format"))
+
+        # zarr v2 allowed arbitrary keys here.
+        # We don't want the ArrayV2Metadata constructor to fail just because someone put an
+        # extra key in the metadata.
+        expected = {x.name for x in fields(cls)}
+        _data = {k: v for k, v in _data.items() if k in expected}
+
         return cls(**_data)
 
     def to_dict(self) -> dict[str, JSON]:

tests/v3/test_array.py

@@ -6,6 +6,7 @@
 import pytest
 
 import zarr.api.asynchronous
+import zarr.storage
 from zarr import Array, AsyncArray, Group
 from zarr.codecs.bytes import BytesCodec
 from zarr.core.array import chunks_initialized

tests/v3/test_group.py

@@ -963,3 +963,15 @@ async def test_open_mutable_mapping():
 def test_open_mutable_mapping_sync():
     group = zarr.open_group(store={}, mode="w")
     assert isinstance(group.store_path.store, MemoryStore)
+
+
+class TestGroupMetadata:
+    def test_from_dict_extra_fields(self):
+        data = {
+            "attributes": {"key": "value"},
+            "_nczarr_superblock": {"version": "2.0.0"},
+            "zarr_format": 2,
+        }
+        result = GroupMetadata.from_dict(data)
+        expected = GroupMetadata(attributes={"key": "value"}, zarr_format=2)
+        assert result == expected

tests/v3/test_metadata/test_v2.py

@@ -72,3 +72,29 @@ def test_metadata_to_dict(
         observed.pop("dimension_separator")
 
     assert observed == expected
+
+
+def test_from_dict_extra_fields() -> None:
+    data = {
+        "_nczarr_array": {"dimrefs": ["/dim1", "/dim2"], "storage": "chunked"},
+        "attributes": {"key": "value"},
+        "chunks": [8],
+        "compressor": None,
+        "dtype": "<f8",
+        "fill_value": 0.0,
+        "filters": None,
+        "order": "C",
+        "shape": [8],
+        "zarr_format": 2,
+    }
+
+    result = ArrayV2Metadata.from_dict(data)
+    expected = ArrayV2Metadata(
+        attributes={"key": "value"},
+        shape=(8,),
+        data_type="float64",
+        chunk_grid=(8,),
+        fill_value=0.0,
+        order="C",
+    )
+    assert result == expected