Update stateful/property tests.

Add actions to
1. overwrite data with oindex
2. read and compare a full array

Author: dcherian

src/zarr/testing/stateful.py

@@ -24,10 +24,11 @@
 from zarr.testing.strategies import (
     basic_indices,
     chunk_paths,
+    dimension_names,
     key_ranges,
     node_names,
     np_array_and_chunks,
-    numpy_arrays,
+    orthogonal_indices,
 )
 from zarr.testing.strategies import keys as zarr_keys
 
@@ -90,11 +91,7 @@ def add_group(self, name: str, data: DataObject) -> None:
         zarr.group(store=self.store, path=path)
         zarr.group(store=self.model, path=path)
 
-    @rule(
-        data=st.data(),
-        name=node_names,
-        array_and_chunks=np_array_and_chunks(arrays=numpy_arrays(zarr_formats=st.just(3))),
-    )
+    @rule(data=st.data(), name=node_names, array_and_chunks=np_array_and_chunks())
     def add_array(
         self,
         data: DataObject,
@@ -122,6 +119,10 @@ def add_array(
                 path=path,
                 store=store,
                 fill_value=fill_value,
+                zarr_format=3,
+                dimension_names=data.draw(
+                    dimension_names(ndim=array.ndim), label="dimension names"
+                ),
                 # Chose bytes codec to avoid wasting time compressing the data being written
                 codecs=[BytesCodec()],
             )
@@ -192,6 +193,14 @@ def delete_chunk(self, data: DataObject) -> None:
         self._sync(self.model.delete(path))
         self._sync(self.store.delete(path))
 
+    @precondition(lambda self: bool(self.all_arrays))
+    @rule(data=st.data())
+    def check_array(self, data: DataObject) -> None:
+        path = data.draw(st.sampled_from(sorted(self.all_arrays)))
+        actual = zarr.open_array(self.store, path=path)[:]
+        expected = zarr.open_array(self.model, path=path)[:]
+        np.testing.assert_equal(actual, expected)
+
     @precondition(lambda self: bool(self.all_arrays))
     @rule(data=st.data())
     def overwrite_array_basic_indexing(self, data: DataObject) -> None:
@@ -206,6 +215,20 @@ def overwrite_array_basic_indexing(self, data: DataObject) -> None:
         model_array[slicer] = new_data
         store_array[slicer] = new_data
 
+    @precondition(lambda self: bool(self.all_arrays))
+    @rule(data=st.data())
+    def overwrite_array_orthogonal_indexing(self, data: DataObject) -> None:
+        array = data.draw(st.sampled_from(sorted(self.all_arrays)))
+        model_array = zarr.open_array(path=array, store=self.model)
+        store_array = zarr.open_array(path=array, store=self.store)
+        indexer, _ = data.draw(orthogonal_indices(shape=model_array.shape))
+        note(f"overwriting array orthogonal {indexer=}")
+        new_data = data.draw(
+            npst.arrays(shape=model_array.oindex[indexer].shape, dtype=model_array.dtype)  # type: ignore[union-attr]
+        )
+        model_array.oindex[indexer] = new_data
+        store_array.oindex[indexer] = new_data
+
     @precondition(lambda self: bool(self.all_arrays))
     @rule(data=st.data())
     def resize_array(self, data: DataObject) -> None:

src/zarr/testing/strategies.py

@@ -43,7 +43,7 @@ def paths(draw: st.DrawFn, *, max_num_nodes: int | None = None) -> str:
     return draw(st.just("/") | keys(max_num_nodes=max_num_nodes))
 
 
-def v3_dtypes() -> st.SearchStrategy[np.dtype[Any]]:
+def dtypes() -> st.SearchStrategy[np.dtype[Any]]:
     return (
         npst.boolean_dtypes()
         | npst.integer_dtypes(endianness="=")
@@ -57,18 +57,12 @@ def v3_dtypes() -> st.SearchStrategy[np.dtype[Any]]:
     )
 
 
+def v3_dtypes() -> st.SearchStrategy[np.dtype[Any]]:
+    return dtypes()
+
+
 def v2_dtypes() -> st.SearchStrategy[np.dtype[Any]]:
-    return (
-        npst.boolean_dtypes()
-        | npst.integer_dtypes(endianness="=")
-        | npst.unsigned_integer_dtypes(endianness="=")
-        | npst.floating_dtypes(endianness="=")
-        | npst.complex_number_dtypes(endianness="=")
-        | npst.byte_string_dtypes(endianness="=")
-        | npst.unicode_string_dtypes(endianness="=")
-        | npst.datetime64_dtypes(endianness="=")
-        | npst.timedelta64_dtypes(endianness="=")
-    )
+    return dtypes()
 
 
 def safe_unicode_for_dtype(dtype: np.dtype[np.str_]) -> st.SearchStrategy[str]:
@@ -144,7 +138,7 @@ def array_metadata(
     shape = draw(array_shapes())
     ndim = len(shape)
     chunk_shape = draw(array_shapes(min_dims=ndim, max_dims=ndim))
-    np_dtype = draw(v3_dtypes())
+    np_dtype = draw(dtypes())
     dtype = get_data_type_from_native_dtype(np_dtype)
     fill_value = draw(npst.from_dtype(np_dtype))
     if zarr_format == 2:
@@ -179,14 +173,12 @@ def numpy_arrays(
     *,
     shapes: st.SearchStrategy[tuple[int, ...]] = array_shapes,
     dtype: np.dtype[Any] | None = None,
-    zarr_formats: st.SearchStrategy[ZarrFormat] = zarr_formats,
 ) -> npt.NDArray[Any]:
     """
     Generate numpy arrays that can be saved in the provided Zarr format.
     """
-    zarr_format = draw(zarr_formats)
     if dtype is None:
-        dtype = draw(v3_dtypes() if zarr_format == 3 else v2_dtypes())
+        dtype = draw(dtypes())
     if np.issubdtype(dtype, np.str_):
         safe_unicode_strings = safe_unicode_for_dtype(dtype)
         return draw(npst.arrays(dtype=dtype, shape=shapes, elements=safe_unicode_strings))
@@ -255,17 +247,24 @@ def arrays(
     attrs: st.SearchStrategy = attrs,
     zarr_formats: st.SearchStrategy = zarr_formats,
 ) -> Array:
-    store = draw(stores)
-    path = draw(paths)
-    name = draw(array_names)
-    attributes = draw(attrs)
-    zarr_format = draw(zarr_formats)
+    store = draw(stores, label="store")
+    path = draw(paths, label="array parent")
+    name = draw(array_names, label="array name")
+    attributes = draw(attrs, label="attributes")
+    zarr_format = draw(zarr_formats, label="zarr format")
     if arrays is None:
-        arrays = numpy_arrays(shapes=shapes, zarr_formats=st.just(zarr_format))
-    nparray = draw(arrays)
-    chunk_shape = draw(chunk_shapes(shape=nparray.shape))
+        arrays = numpy_arrays(shapes=shapes)
+    nparray = draw(arrays, label="array data")
+    chunk_shape = draw(chunk_shapes(shape=nparray.shape), label="chunk shape")
+    extra_kwargs = {}
     if zarr_format == 3 and all(c > 0 for c in chunk_shape):
-        shard_shape = draw(st.none() | shard_shapes(shape=nparray.shape, chunk_shape=chunk_shape))
+        shard_shape = draw(
+            st.none() | shard_shapes(shape=nparray.shape, chunk_shape=chunk_shape),
+            label="shard shape",
+        )
+        extra_kwargs["dimension_names"] = draw(
+            dimension_names(ndim=nparray.ndim), label="dimension names"
+        )
     else:
         shard_shape = None
     # test that None works too.
@@ -286,6 +285,7 @@ def arrays(
         attributes=attributes,
         # compressor=compressor,  # FIXME
         fill_value=fill_value,
+        **extra_kwargs,
     )
 
     assert isinstance(a, Array)
@@ -385,13 +385,19 @@ def orthogonal_indices(
     npindexer = []
     ndim = len(shape)
     for axis, size in enumerate(shape):
-        val = draw(
-            npst.integer_array_indices(
+        if size != 0:
+            strategy = npst.integer_array_indices(
                 shape=(size,), result_shape=npst.array_shapes(min_side=1, max_side=size, max_dims=1)
-            )
-            | basic_indices(min_dims=1, shape=(size,), allow_ellipsis=False)
-            .map(lambda x: (x,) if not isinstance(x, tuple) else x)  # bare ints, slices
-            .filter(bool)  # skip empty tuple
+            ) | basic_indices(min_dims=1, shape=(size,), allow_ellipsis=False)
+        else:
+            strategy = basic_indices(min_dims=1, shape=(size,), allow_ellipsis=False)
+
+        val = draw(
+            strategy
+            # bare ints, slices
+            .map(lambda x: (x,) if not isinstance(x, tuple) else x)
+            # skip empty tuple
+            .filter(bool)
         )
         (idxr,) = val
         if isinstance(idxr, int):

tests/test_properties.py

@@ -76,10 +76,10 @@ def deep_equal(a: Any, b: Any) -> bool:
 
 
 @pytest.mark.filterwarnings("ignore::zarr.core.dtype.common.UnstableSpecificationWarning")
-@given(data=st.data(), zarr_format=zarr_formats)
-def test_array_roundtrip(data: st.DataObject, zarr_format: int) -> None:
-    nparray = data.draw(numpy_arrays(zarr_formats=st.just(zarr_format)))
-    zarray = data.draw(arrays(arrays=st.just(nparray), zarr_formats=st.just(zarr_format)))
+@given(data=st.data())
+def test_array_roundtrip(data: st.DataObject) -> None:
+    nparray = data.draw(numpy_arrays())
+    zarray = data.draw(arrays(arrays=st.just(nparray)))
     assert_array_equal(nparray, zarray[:])