Fix fill_value serialization of NaN; add property-based tests

Author: moradology

src/zarr/core/metadata/v2.py

@@ -144,15 +144,26 @@ def _json_convert(
                 return o.name
             raise TypeError
 
+        def _sanitize_fill_value(fv: Any):
+            if isinstance(fv, (float, np.floating)):
+                if np.isnan(fv):
+                    fv = "NaN"
+                elif np.isinf(fv):
+                    fv = "Infinity" if fv > 0 else "-Infinity"
+            elif isinstance(fv, (np.complex64, np.complexfloating)):
+                fv = [_sanitize_fill_value(fv.real), _sanitize_fill_value(fv.imag)]
+            return fv        
+
         zarray_dict = self.to_dict()
+        zarray_dict["fill_value"] = _sanitize_fill_value(zarray_dict["fill_value"])
         zattrs_dict = zarray_dict.pop("attributes", {})
         json_indent = config.get("json_indent")
         return {
             ZARRAY_JSON: prototype.buffer.from_bytes(
-                json.dumps(zarray_dict, default=_json_convert, indent=json_indent).encode()
+                json.dumps(zarray_dict, default=_json_convert, indent=json_indent, allow_nan=False).encode()
             ),
             ZATTRS_JSON: prototype.buffer.from_bytes(
-                json.dumps(zattrs_dict, indent=json_indent).encode()
+                json.dumps(zattrs_dict, indent=json_indent, allow_nan=False).encode()
             ),
         }
 
@@ -300,7 +311,6 @@ def parse_fill_value(fill_value: object, dtype: np.dtype[Any]) -> Any:
     -------
         An instance of `dtype`, or `None`, or any python object (in the case of an object dtype)
     """
-
     if fill_value is None or dtype.hasobject:
         # no fill value
         pass

src/zarr/testing/v2metadata.py

@@ -0,0 +1,22 @@
+from hypothesis import strategies as st
+import hypothesis.extra.numpy as npst
+
+from zarr.testing.strategies import v2_dtypes
+
+@st.composite
+def array_metadata_v2_inputs(draw):
+    dims = draw(st.integers(min_value=1, max_value=10))
+    shape = draw(st.lists(st.integers(min_value=1, max_value=100), min_size=dims, max_size=dims))
+    chunks = draw(st.lists(st.integers(min_value=1, max_value=100), min_size=dims, max_size=dims))
+    dtype = draw(v2_dtypes())
+    fill_value = draw(st.one_of([st.none(), npst.from_dtype(dtype)]))
+    order = draw(st.sampled_from(["C", "F"]))
+    dimension_separator = draw(st.sampled_from([".", "/"]))
+    return {
+        "shape": shape,
+        "dtype": dtype,
+        "chunks": chunks,
+        "fill_value": fill_value,
+        "order": order,
+        "dimension_separator": dimension_separator,
+    }

tests/test_properties.py

@@ -1,3 +1,4 @@
+import json
 import numpy as np
 import pytest
 from numpy.testing import assert_array_equal
@@ -8,7 +9,11 @@
 import hypothesis.strategies as st
 from hypothesis import given
 
+from zarr.core.buffer import default_buffer_prototype
+from zarr.core.common import ZARRAY_JSON, parse_shapelike
+from zarr.core.metadata.v2 import ArrayV2Metadata, parse_fill_value, parse_dtype, parse_shapelike
 from zarr.testing.strategies import arrays, basic_indices, numpy_arrays, zarr_formats
+from zarr.testing.v2metadata import array_metadata_v2_inputs
 
 
 @given(data=st.data(), zarr_format=zarr_formats)
@@ -69,3 +74,38 @@ def test_vindex(data: st.DataObject) -> None:
 #     nparray = data.draw(np_arrays)
 #     zarray = data.draw(arrays(arrays=st.just(nparray)))
 #     assert_array_equal(nparray, zarray[:])
+
+
+@given(array_metadata_v2_inputs())
+def test_v2meta_fill_value_serialization(inputs):
+    metadata = ArrayV2Metadata(**inputs)
+    buffer_dict = metadata.to_buffer_dict(prototype=default_buffer_prototype())
+    zarray_dict = json.loads(buffer_dict[ZARRAY_JSON].to_bytes().decode())
+
+    if isinstance(inputs["fill_value"], (float, np.floating)) and np.isnan(inputs["fill_value"]):
+        assert zarray_dict["fill_value"] == "NaN"
+    else:
+        assert zarray_dict["fill_value"] == inputs["fill_value"]
+
+
+@given(npst.from_dtype(dtype=np.dtype("float64"), allow_nan=True, allow_infinity=True))
+def test_v2meta_nan_and_infinity(fill_value):
+    metadata = ArrayV2Metadata(
+        shape=[10],
+        dtype=np.dtype("float64"),
+        chunks=[5],
+        fill_value=fill_value,
+        order="C",
+    )
+
+    buffer_dict = metadata.to_buffer_dict(prototype=default_buffer_prototype())
+    zarray_dict = json.loads(buffer_dict[ZARRAY_JSON].to_bytes().decode())
+
+    if np.isnan(fill_value):
+        assert zarray_dict["fill_value"] == "NaN"
+    elif np.isinf(fill_value) and fill_value > 0:
+        assert zarray_dict["fill_value"] == "Infinity"
+    elif np.isinf(fill_value):
+        assert zarray_dict["fill_value"] == "-Infinity"
+    else:
+        assert zarray_dict["fill_value"] == fill_value