Cleanups to implementation of create() (zarr-developers#3111)

dstansby · web-flow · commit 9d97b24d9bc2 · 2025-07-09T20:33:10.000Z
* Remove duplicate compressor handling

* Push filter parsing down the stack

* Put not implemented warnings together

* Test unimplemented kwargs warn

* pre-commit fixes

* Fix syncrhonous create() signature
diff --git a/src/zarr/api/asynchronous.py b/src/zarr/api/asynchronous.py
@@ -15,7 +15,6 @@
     Array,
     AsyncArray,
     CompressorLike,
-    _get_default_chunk_encoding_v2,
     create_array,
     from_array,
     get_array_metadata,
@@ -33,7 +32,7 @@
     _warn_order_kwarg,
     _warn_write_empty_chunks_kwarg,
 )
-from zarr.core.dtype import ZDTypeLike, get_data_type_from_native_dtype, parse_data_type
+from zarr.core.dtype import ZDTypeLike, get_data_type_from_native_dtype
 from zarr.core.group import (
     AsyncGroup,
     ConsolidatedMetadata,
@@ -48,6 +47,8 @@
 if TYPE_CHECKING:
     from collections.abc import Iterable
 
+    import numcodecs.abc
+
     from zarr.abc.codec import Codec
     from zarr.core.buffer import NDArrayLikeOrScalar
     from zarr.core.chunk_key_encodings import ChunkKeyEncoding
@@ -871,7 +872,7 @@ async def create(
     overwrite: bool = False,
     path: PathLike | None = None,
     chunk_store: StoreLike | None = None,
-    filters: list[dict[str, JSON]] | None = None,  # TODO: type has changed
+    filters: Iterable[dict[str, JSON] | numcodecs.abc.Codec] | None = None,
     cache_metadata: bool | None = None,
     cache_attrs: bool | None = None,
     read_only: bool | None = None,
@@ -1009,13 +1010,6 @@ async def create(
         _handle_zarr_version_or_format(zarr_version=zarr_version, zarr_format=zarr_format)
         or _default_zarr_format()
     )
-    zdtype = parse_data_type(dtype, zarr_format=zarr_format)
-    if zarr_format == 2:
-        default_filters, default_compressor = _get_default_chunk_encoding_v2(zdtype)
-        if not filters:
-            filters = default_filters  # type: ignore[assignment]
-        if compressor == "auto":
-            compressor = default_compressor
 
     if synchronizer is not None:
         warnings.warn("synchronizer is not yet implemented", RuntimeWarning, stacklevel=2)
@@ -1029,14 +1023,14 @@ async def create(
         warnings.warn("object_codec is not yet implemented", RuntimeWarning, stacklevel=2)
     if read_only is not None:
         warnings.warn("read_only is not yet implemented", RuntimeWarning, stacklevel=2)
+    if meta_array is not None:
+        warnings.warn("meta_array is not yet implemented", RuntimeWarning, stacklevel=2)
+
     if order is not None:
         _warn_order_kwarg()
     if write_empty_chunks is not None:
         _warn_write_empty_chunks_kwarg()
 
-    if meta_array is not None:
-        warnings.warn("meta_array is not yet implemented", RuntimeWarning, stacklevel=2)
-
     mode = kwargs.pop("mode", None)
     if mode is None:
         mode = "a"
@@ -1067,7 +1061,7 @@ async def create(
         store_path,
         shape=shape,
         chunks=chunks,
-        dtype=zdtype,
+        dtype=dtype,
         compressor=compressor,
         fill_value=fill_value,
         overwrite=overwrite,
diff --git a/src/zarr/api/synchronous.py b/src/zarr/api/synchronous.py
@@ -15,6 +15,7 @@
 if TYPE_CHECKING:
     from collections.abc import Iterable
 
+    import numcodecs.abc
     import numpy as np
     import numpy.typing as npt
 
@@ -613,7 +614,7 @@ def create(
     overwrite: bool = False,
     path: PathLike | None = None,
     chunk_store: StoreLike | None = None,
-    filters: list[dict[str, JSON]] | None = None,  # TODO: type has changed
+    filters: Iterable[dict[str, JSON] | numcodecs.abc.Codec] | None = None,
     cache_metadata: bool | None = None,
     cache_attrs: bool | None = None,
     read_only: bool | None = None,
diff --git a/src/zarr/core/array.py b/src/zarr/core/array.py
@@ -591,7 +591,7 @@ async def _create(
         chunks: ShapeLike | None = None,
         dimension_separator: Literal[".", "/"] | None = None,
         order: MemoryOrder | None = None,
-        filters: list[dict[str, JSON]] | None = None,
+        filters: Iterable[dict[str, JSON] | numcodecs.abc.Codec] | None = None,
         compressor: CompressorLike = "auto",
         # runtime
         overwrite: bool = False,
@@ -850,9 +850,10 @@ async def _create_v2(
         else:
             await ensure_no_existing_node(store_path, zarr_format=2)
 
+        default_filters, default_compressor = _get_default_chunk_encoding_v2(dtype)
         compressor_parsed: CompressorLikev2
         if compressor == "auto":
-            _, compressor_parsed = _get_default_chunk_encoding_v2(dtype)
+            compressor_parsed = default_compressor
         elif isinstance(compressor, BytesBytesCodec):
             raise ValueError(
                 "Cannot use a BytesBytesCodec as a compressor for zarr v2 arrays. "
@@ -861,6 +862,9 @@ async def _create_v2(
         else:
             compressor_parsed = compressor
 
+        if filters is None:
+            filters = default_filters
+
         metadata = cls._create_metadata_v2(
             shape=shape,
             dtype=dtype,
diff --git a/tests/test_api.py b/tests/test_api.py
@@ -1412,3 +1412,10 @@ def test_auto_chunks(f: Callable[..., Array]) -> None:
 
     a = f(**kwargs)
     assert a.chunks == (500, 500)
+
+
+@pytest.mark.parametrize("kwarg_name", ["synchronizer", "chunk_store", "cache_attrs", "meta_array"])
+def test_unimplemented_kwarg_warnings(kwarg_name: str) -> None:
+    kwargs = {kwarg_name: 1}
+    with pytest.warns(RuntimeWarning, match=".* is not yet implemented"):
+        zarr.create(shape=(1,), **kwargs)  # type: ignore[arg-type]
