test v3_default_codecs

Author: brokkoli71

src/zarr/core/config.py

@@ -49,6 +49,11 @@ def reset(self) -> None:
                     "string": "vlen-utf8",
                     "bytes": "vlen-bytes",
                 },
+                "v3_default_codecs": {
+                    "numeric": ["bytes", "zstd"],
+                    "string": ["vlen-utf8"],
+                    "bytes": ["vlen-bytes"],
+                },
             },
             "async": {"concurrency": 10, "timeout": None},
             "threading": {"max_workers": None},

src/zarr/core/metadata/v2.py

@@ -336,7 +336,6 @@ def _default_filters_and_compressor(
 
     https://numpy.org/doc/2.1/reference/generated/numpy.dtype.kind.html
     """
-    dtype = np.dtype(dtype)
     default_compressor = config.get("array.v2_default_compressor")
     if dtype.kind in "biufcmM":
         dtype_key = "numeric"

tests/test_config.py

@@ -11,7 +11,8 @@
 from zarr import Array, zeros
 from zarr.abc.codec import CodecInput, CodecOutput, CodecPipeline
 from zarr.abc.store import ByteSetter, Store
-from zarr.codecs import BloscCodec, BytesCodec, Crc32cCodec, ShardingCodec
+from zarr.codecs import BloscCodec, BytesCodec, Crc32cCodec, ShardingCodec, TransposeCodec, GzipCodec, VLenBytesCodec, \
+    VLenUTF8Codec
 from zarr.core.array_spec import ArraySpec
 from zarr.core.buffer import NDBuffer
 from zarr.core.codec_pipeline import BatchedCodecPipeline
@@ -239,3 +240,18 @@ def test_config_buffer_implementation() -> None:
     )
     arr_Crc32c[:] = data2d
     assert np.array_equal(arr_Crc32c[:], data2d)
+
+@pytest.mark.parametrize("dtype", ["int", "bytes", "str"])
+def test_default_codecs(dtype:str) -> None:
+    with config.set({"array.v3_default_codecs": {
+        "numeric": ["bytes", "gzip"], # test setting non-standard codecs
+        "string": ["vlen-utf8"],
+        "bytes": ["vlen-bytes"],
+    }}):
+        arr = zeros(shape=(100), store=StoreExpectingTestBuffer(), dtype=dtype)
+        if dtype == "int":
+            assert arr.metadata.codecs == [BytesCodec(), GzipCodec()]
+        elif dtype == "bytes":
+            assert arr.metadata.codecs == [VLenBytesCodec()]
+        elif dtype == "str":
+            assert arr.metadata.codecs == [VLenUTF8Codec()]