Merge branch 'main' into dependabot/github_actions/actions-31ef6db2d7

Author: d-v-b

changes/4219.bugfix.md

@@ -0,0 +1,3 @@
+`DefaultChunkKeyEncoding.decode_chunk_key` now validates that a chunk key
+starts with the configured `c<separator>` prefix and raises `ValueError` for
+malformed keys, instead of silently decoding them incorrectly.

pyproject.toml

@@ -94,18 +94,18 @@ homepage = "https://github.com/zarr-developers/zarr-python"
 # pins deliberately, e.g. via dependabot or `uv lock --upgrade`.
 [dependency-groups]
 test = [
-    "coverage==7.14.3",
+    "coverage==7.15.2",
     "pytest==9.1.1",
     "pytest-asyncio==1.4.0",
     "pytest-cov==7.1.0",
     "pytest-accept==0.3.0",
     "numpydoc==1.10.0",
-    "hypothesis==6.155.7",
+    "hypothesis==6.160.0",
     "pytest-xdist==3.8.0",
     "pytest-benchmark==5.2.3",
     "pytest-codspeed==5.0.3",
-    "tomlkit==0.15.0",
-    "uv==0.11.26",
+    "tomlkit==0.15.1",
+    "uv==0.11.31",
 ]
 remote-tests = [
     {include-group = "test"},
@@ -121,15 +121,15 @@ release = [
 ]
 docs = [
     # Doc building
-    "mkdocs-material[imaging]==9.7.6",
+    "mkdocs-material[imaging]==9.7.7",
     "mkdocs==1.6.1",
-    "mkdocstrings==1.0.4",
+    "mkdocstrings==1.0.6",
     "mkdocstrings-python==2.0.5",
     "mike==2.2.0",
     "mkdocs-redirects==1.2.3",
-    "markdown-exec[ansi]==1.12.1",
+    "markdown-exec[ansi]==1.12.3",
     "griffe-inherited-docstrings==1.1.3",
-    "ruff==0.15.20",
+    "ruff==0.15.22",
     # Changelog generation
     {include-group = "release"},
     # Optional dependencies to run examples
@@ -143,7 +143,7 @@ dev = [
     {include-group = "remote-tests"},
     {include-group = "docs"},
     "universal-pathlib",
-    "mypy==2.1.0",
+    "mypy==2.3.0",
 ]
 
 [tool.coverage.report]

src/zarr/core/chunk_key_encodings.py

@@ -79,7 +79,11 @@ def __post_init__(self) -> None:
     def decode_chunk_key(self, chunk_key: str) -> tuple[int, ...]:
         if chunk_key == "c":
             return ()
-        return tuple(map(int, chunk_key[1:].split(self.separator)))
+        # Strip the "c<sep>" prefix (e.g. "c/" or "c.") before splitting.
+        prefix = "c" + self.separator
+        if chunk_key.startswith(prefix):
+            return tuple(map(int, chunk_key[len(prefix) :].split(self.separator)))
+        raise ValueError(f"Invalid chunk key for default encoding: {chunk_key!r}")
 
     def encode_chunk_key(self, chunk_coords: tuple[int, ...]) -> str:
         return self.separator.join(map(str, ("c",) + chunk_coords))

src/zarr/core/indexing.py

@@ -512,7 +512,8 @@ def replace_ellipsis(selection: Any, shape: tuple[int, ...]) -> SelectionNormali
 
 def replace_lists(selection: SelectionNormalized) -> SelectionNormalized:
     return tuple(
-        np.asarray(dim_sel) if isinstance(dim_sel, list) else dim_sel for dim_sel in selection
+        cast("ArrayOfIntOrBool", np.asarray(dim_sel)) if isinstance(dim_sel, list) else dim_sel
+        for dim_sel in selection
     )
 
 
@@ -1193,7 +1194,7 @@ def __init__(
         # some initial normalization
         selection_normalized = cast("CoordinateSelectionNormalized", ensure_tuple(selection))
         selection_normalized = tuple(
-            np.asarray([i]) if is_integer(i) else i for i in selection_normalized
+            np.asarray([i], dtype=np.intp) if is_integer(i) else i for i in selection_normalized
         )
         selection_normalized = cast(
             "CoordinateSelectionNormalized", replace_lists(selection_normalized)

tests/test_chunk_key_encodings.py

@@ -0,0 +1,65 @@
+from __future__ import annotations
+
+import pytest
+
+from zarr.core.chunk_key_encodings import DefaultChunkKeyEncoding, V2ChunkKeyEncoding
+
+
+@pytest.mark.parametrize("separator", ["/", "."])
+@pytest.mark.parametrize(
+    "coords",
+    [(), (0,), (1, 2), (10, 0, 3)],
+)
+def test_default_encoding_round_trips(separator: str, coords: tuple[int, ...]) -> None:
+    """Encoding coordinates and decoding the result returns the coordinates."""
+    encoding = DefaultChunkKeyEncoding(separator=separator)  # type: ignore[arg-type]
+
+    key = encoding.encode_chunk_key(coords)
+    assert encoding.decode_chunk_key(key) == coords
+
+
+@pytest.mark.parametrize("separator", ["/", "."])
+@pytest.mark.parametrize("coords", [(0,), (1, 2), (10, 0, 3)])
+def test_v2_encoding_round_trips(separator: str, coords: tuple[int, ...]) -> None:
+    """The v2 encoding round-trips coordinates for either separator."""
+    encoding = V2ChunkKeyEncoding(separator=separator)  # type: ignore[arg-type]
+
+    key = encoding.encode_chunk_key(coords)
+    assert encoding.decode_chunk_key(key) == coords
+
+
+@pytest.mark.parametrize("separator", ["/", "."])
+def test_v2_zero_dimensional_key_is_ambiguous(separator: str) -> None:
+    """A 0-d v2 array stores its sole chunk under `"0"`, the same key a 1-d
+    array uses for chunk 0, so decoding cannot recover the empty tuple on its
+    own -- the array's dimensionality is what disambiguates it."""
+    encoding = V2ChunkKeyEncoding(separator=separator)  # type: ignore[arg-type]
+
+    assert encoding.encode_chunk_key(()) == "0"
+    assert encoding.decode_chunk_key("0") == (0,)
+
+
+@pytest.mark.parametrize(
+    "chunk_key",
+    [
+        "0/1",  # no "c" prefix at all
+        "c0/1",  # "c" not followed by the separator
+        "x/0/1",  # wrong prefix character
+        "",
+    ],
+)
+def test_default_encoding_rejects_key_without_prefix(chunk_key: str) -> None:
+    """A key that does not carry the `c<separator>` prefix is not a chunk key
+    for this encoding, and must be rejected rather than silently decoded."""
+    encoding = DefaultChunkKeyEncoding(separator="/")
+
+    with pytest.raises(ValueError, match="Invalid chunk key"):
+        encoding.decode_chunk_key(chunk_key)
+
+
+def test_default_encoding_rejects_key_using_the_other_separator() -> None:
+    """A key encoded with `.` is not valid for a `/`-separated encoding."""
+    encoding = DefaultChunkKeyEncoding(separator="/")
+
+    with pytest.raises(ValueError, match="Invalid chunk key"):
+        encoding.decode_chunk_key("c.0.1")