⬆️ Bump safecheck and dev dependencies

pyproject.toml

@@ -35,7 +35,7 @@ classifiers = [
 
 [tool.poetry.dependencies]
 python = ">=3.10,<3.13"
-safecheck = "^0.3"
+safecheck = ">=0.3,<0.5"
 joblib = "^1"
 typing-extensions = ">=4"
 numpy = { version = ">=1", optional = true }
@@ -66,18 +66,18 @@ scann = ["scann"]
 jax = ["jax"]
 
 [tool.poetry.group.dev.dependencies]
-pytest = "^8.2.0"
+pytest = "^8.3.3"
 pytest-html = "^4.1.1"
-hypothesis = "^6.100.2"
-coverage = "^7.5.0"
+hypothesis = "^6.112.1"
+coverage = "^7.6.1"
 pytest-cov = "^5.0.0"
-coverage-badge = "^1.1.1"
-ruff = "^0.4.2"
-pre-commit = "^3.7.0"
-black = "^24.4.2"
-pyright = "^1.1.360"
-bandit = "^1.7.8"
-safety = "^3.1.0"
+coverage-badge = "^1.1.2"
+ruff = "^0.6.5"
+pre-commit = "^3.8.0"
+black = "^24.8.0"
+pyright = "^1.1.381"
+bandit = "^1.7.9"
+safety = "^3.2.7"
 notebook = "^7.0.4"
 pytest-benchmark = "^4.0.0"
 

src/nearness/_autofaiss.py

@@ -95,5 +95,5 @@ def query_batch(
         points: Float[NumpyArray, "m d"],
         n_neighbors: int,
     ) -> tuple[Int64[NumpyArray, "m {n_neighbors}"], Float32[NumpyArray, "m {n_neighbors}"]]:
-        dist, idx = self._index.search(points, n_neighbors)  # type: ignore[reportGeneralTypeIssues]
+        dist, idx = self._index.search(points, n_neighbors)  # type: ignore[reportCallIssue]
         return idx, dist

src/nearness/_base/_neighbors.py

@@ -46,8 +46,8 @@ def __new__(cls, name: str, bases: tuple[type, ...], attrs: dict[str, Any]) -> t
         """Check the signature of ``__init__`` to ensure keyword-only arguments."""
         if "__init__" in attrs:
             parameters = inspect.signature(attrs["__init__"]).parameters
-            for name, parameter in parameters.items():
-                if name != "self" and (kind := parameter.kind) is not inspect.Parameter.KEYWORD_ONLY:
+            for param_name, parameter in parameters.items():
+                if param_name != "self" and (kind := parameter.kind) is not inspect.Parameter.KEYWORD_ONLY:
                     msg = (
                         "Only keyword-only arguments are allowed for classes inheriting from 'nearness."
                         f"NearestNeighbors', but found parameter '{parameter}' of kind '{kind}'."
@@ -89,9 +89,9 @@ def __call__(cls, *_: Any, **kwargs: Any) -> "NearestNeighbors":
         # now we set all the relevant attributes on the ``instance``, as they should not be class-bound.
         # order is important here as ``_wrap_fit_method`` and ``_wrap_check_method`` depend on the set attributes
         obj = type.__call__(cls, **kwargs)
-        obj._parameters_, obj._config_ = cls._parameters_, cls._config_  # noqa: SLF001
+        obj._parameters_, obj._config_ = cls._parameters_, cls._config_
         # make sure that the wrapped methods are in sync when the config is changed after class instantiation
-        obj._config_.register_callback(  # noqa: SLF001
+        obj._config_.register_callback(
             "methods_require_fit",
             partial(cls._check_callback, obj=obj),
         )
@@ -144,7 +144,7 @@ def _wrap_check_method(cls, obj: "NearestNeighbors") -> None:
         """
         logger.debug("Starting to wrap methods to enable fit checking.")
         available_attributes = dir(obj)
-        methods_to_wrap = obj._config_.methods_require_fit  # noqa: SLF001
+        methods_to_wrap = obj._config_.methods_require_fit
         for attribute_name in available_attributes:
             attribute = getattr(obj, attribute_name)
             has_check = hasattr(attribute, "__check__")
@@ -186,7 +186,7 @@ def _unwrap_check_method(cls, obj: "NearestNeighbors") -> None:
         is set in ``fit``, but unsafe when ``__fitted__`` is manually set to ``False`` after ``fit``.
         """
         logger.debug("Starting to unwrap all fit checking methods.")
-        for method_name in obj._config_.methods_require_fit:  # noqa: SLF001
+        for method_name in obj._config_.methods_require_fit:
             # we set an __requires_fit__ attribute on the wrapper, because using ``__wrapped__`` alone is not
             # safe (methods also use ``__wrapped__`` starting with Python 3.10)
             if hasattr(obj, method_name) and hasattr(method := getattr(obj, method_name), "__check__"):

src/nearness/_faiss.py

@@ -50,17 +50,17 @@ def __init__(
     def fit(self, data: Float[NumpyArray, "n d"]) -> "FaissNeighbors":
         _, dim = data.shape
         self._index = self._create_index(dim)
-        self._index.train(data)  # type: ignore[reportGeneralTypeIssues]
+        self._index.train(data)  # type: ignore[reportCallIssue]
 
         if self.parameters.add_data_on_fit:
             # data might be added directly on fit, or using the ``add`` method
-            self._index.add(data)  # type: ignore[reportGeneralTypeIssues]
+            self._index.add(data)  # type: ignore[reportCallIssue]
 
         return self
 
     @typecheck
     def add(self, data: Float[NumpyArray, "n d"]) -> "FaissNeighbors":
-        self._index.add(data)  # type: ignore[reportGeneralTypeIssues]
+        self._index.add(data)  # type: ignore[reportCallIssue]
         return self
 
     def query(
@@ -76,7 +76,7 @@ def query_batch(
         points: Float[NumpyArray, "m d"],
         n_neighbors: int,
     ) -> tuple[Int64[NumpyArray, "m {n_neighbors}"], Float32[NumpyArray, "m {n_neighbors}"]]:
-        dist, idx = self._index.search(points, n_neighbors)  # type: ignore[reportGeneralTypeIssues]
+        dist, idx = self._index.search(points, n_neighbors)  # type: ignore[reportCallIssue]
         return idx, dist
 
     def _create_index(self, dim: int) -> faiss.Index:

src/nearness/_scipy.py

@@ -74,9 +74,9 @@ def query_batch(
         points: Float[NumpyArray, "m d"],
         n_neighbors: int,
     ) -> tuple[Int64[NumpyArray, "m {n_neighbors}"], Float[NumpyArray, "m {n_neighbors}"]]:
-        distance = cdist(
+        distance = cdist(  # type: ignore[reportCallIssue]
             points,
-            self._data,  # type: ignore[reportGeneralTypeIssues]
+            self._data,  # type: ignore[reportArgumentType]
             metric=self.parameters.metric,
             **self.parameters.metric_args,
         )