Merge pull request #366 from jorenham/scipy-stubs-1.14.1.5

bump `scipy-stubs` to `1.14.1.5`

.pre-commit-config.yaml

@@ -2,10 +2,6 @@ ci:
   autoupdate_branch: "master"
   autoupdate_commit_msg: "update pre-commit hooks"
   skip:
-    - markdownlint
-    - codespell
-    - ruff
-    - ruff-format
     - basedpyright
     - basedpyright-verifytypes
 
@@ -82,7 +78,7 @@ repos:
     hooks:
       - id: basedpyright
         name: basedpyright
-        entry: uv run basedpyright
+        entry: uv run basedpyright --threads=3
         language: system
         types_or: [python, pyi]
 

lmo/_poly.py

@@ -9,7 +9,7 @@
 
 from __future__ import annotations
 
-from typing import TYPE_CHECKING, Any, TypeAlias, TypeVar, cast, overload
+from typing import TYPE_CHECKING, TypeAlias, TypeVar, cast, overload
 
 import numpy as np
 import numpy.polynomial as npp
@@ -49,27 +49,31 @@ def __dir__() -> tuple[str, ...]:
 _T_shape = TypeVar("_T_shape", bound=onp.AtLeast1D)
 _T_poly = TypeVar("_T_poly", bound=PolySeries)
 
+_Float = np.float32 | np.float64
+
+###
+
 
 @overload
 def eval_sh_jacobi(
     n: int | lmt.Integer,
     a: float,
     b: float,
-    x: float | np.floating[Any],
-) -> float: ...
+    x: float | _Float,
+) -> float | _Float: ...
 @overload
 def eval_sh_jacobi(
     n: int | lmt.Integer,
     a: float,
     b: float,
-    x: onp.Array[_T_shape, lmt.Floating],
-) -> onp.Array[_T_shape, np.float64]: ...
+    x: onp.Array[_T_shape, _Float],
+) -> onp.Array[_T_shape, _Float]: ...
 def eval_sh_jacobi(  # noqa: C901
     n: int | lmt.Integer,
     a: float,
     b: float,
-    x: float | np.floating[Any] | onp.Array[_T_shape, lmt.Floating],
-) -> float | onp.Array[_T_shape, np.float64]:
+    x: float | _Float | onp.Array[_T_shape, _Float],
+) -> float | _Float | onp.Array[_T_shape, _Float]:
     """
     Fast evaluation of the n-th shifted Jacobi polynomial.
     Faster than pre-computing using np.Polynomial, and than

lmo/_utils.py

@@ -316,9 +316,7 @@ def clean_order(
         msg = f"expected {name} >= {rmin}, got {r}"
         raise TypeError(msg)
 
-    # Pyright doesn't seem to realize that its internal `Unknown` "type" is literally
-    # defined as `Any` within gradual type-systems like Python typing...
-    return int(r)  # pyright: ignore[reportUnknownArgumentType]
+    return int(r)
 
 
 @overload

lmo/diagnostic.py

@@ -59,7 +59,9 @@ def __call__(self, x: _T_x, /) -> _T_x: ...
 
 
 _Tuple2: TypeAlias = tuple[_T, _T]
-_FloatND: TypeAlias = onp.ArrayND[np.float64]
+
+_Float: TypeAlias = float | np.float32 | np.float64
+_FloatND: TypeAlias = onp.ArrayND[np.float32 | np.float64]
 
 
 _MIN_RHO: Final[float] = 1e-5
@@ -78,8 +80,8 @@ class HypothesisTestResult(NamedTuple):
             hypothesis, $H_0$.
     """
 
-    statistic: float | _FloatND
-    pvalue: float | _FloatND
+    statistic: _Float | _FloatND
+    pvalue: _Float | _FloatND
 
     @property
     def is_valid(self) -> np.bool_ | onp.ArrayND[np.bool_]:
@@ -88,7 +90,7 @@ def is_valid(self) -> np.bool_ | onp.ArrayND[np.bool_]:
 
     def is_significant(
         self,
-        level: float | np.floating[Any] = 0.05,
+        level: _Float | np.floating[Any] = 0.05,
         /,
     ) -> np.bool_ | onp.ArrayND[np.bool_]:
         """
@@ -211,7 +213,7 @@ def l_moment_gof(
     trim: lmt.ToTrim = 0,
     **kwargs: Any,
 ) -> HypothesisTestResult:
-    r"""
+    """
     Goodness-of-fit (GOF) hypothesis test for the null hypothesis that the
     observed L-moments come from a distribution with the given
     [`scipy.stats`][scipy.stats] distribution or cumulative distribution
@@ -224,14 +226,14 @@ def l_moment_gof(
     The test statistic is the squared Mahalanobis distance between the $n$
     observed L-moments, and the theoretical L-moments. It asymptically (in
     sample size) follows the
-    [$\chi^2$](https://wikipedia.org/wiki/Chi-squared_distribution)
+    [`Chi-squared`](https://wikipedia.org/wiki/Chi-squared_distribution)
     distribution, with $n$ degrees of freedom.
 
     The sample L-moments are expected to be of consecutive orders
-    $r = 1, 2, \dots, n$.
+    `r = 1, 2, ..., n`.
     Generally, the amount of L-moments $n$ should not be less than the amount
     of parameters of the distribution, including the location and scale
-    parameters. Therefore, it is required to have $n \ge 2$.
+    parameters. Therefore, it is required to have `n <= 2`.
 
     Notes:
         The theoretical L-moments and their covariance matrix are calculated
@@ -273,7 +275,7 @@ def l_moment_gof(
 
     See Also:
         - [`l_moment_from_cdf`][lmo.theoretical.l_moment_from_cdf]
-        - ['l_moment_cov_from_cdf'][lmo.theoretical.l_moment_cov_from_cdf]
+        - [`l_moment_cov_from_cdf`][lmo.theoretical.l_moment_cov_from_cdf]
 
     """
     l_r = np.asarray_chkfinite(l_moments)
@@ -297,7 +299,7 @@ def l_moment_gof(
 
     from scipy.special import chdtrc
 
-    stat = n_obs * _gof_stat(l_r.T, lambda_r, lambda_rr).T[()]
+    stat: _Float | _FloatND = n_obs * _gof_stat(l_r.T, lambda_r, lambda_rr).T[()]
     pval = chdtrc(n, stat)
     return HypothesisTestResult(stat, pval)
 
@@ -333,7 +335,7 @@ def l_stats_gof(
 
     from scipy.special import chdtrc
 
-    stat = n_obs * _gof_stat(t_r.T, tau_r, tau_rr).T[()]
+    stat: _Float | _FloatND = n_obs * _gof_stat(t_r.T, tau_r, tau_rr).T[()]
     pval = chdtrc(n, stat)
     return HypothesisTestResult(stat, pval)
 
@@ -366,9 +368,11 @@ def _lm2_bounds_single(r: int, trim: _Tuple2[float]) -> float:
             ) / (np.pi * 2 * r**2)
 
 
-_lm2_bounds: Final = cast(
-    "Callable[[lmt.ToOrderND, _Tuple2[float]], _FloatND]",
-    np.vectorize(_lm2_bounds_single, otypes=[float], excluded={1}, signature="()->()"),
+_lm2_bounds: Final = np.vectorize(
+    _lm2_bounds_single,
+    otypes=[float],
+    excluded={1},
+    signature="()->()",
 )
 
 
@@ -377,20 +381,20 @@ def l_moment_bounds(
     r: lmt.ToOrder0D,
     /,
     trim: lmt.ToTrim = 0,
-    scale: float = 1.0,
+    scale: _Float = 1,
 ) -> float: ...
 @overload
 def l_moment_bounds(
     r: lmt.ToOrderND,
     /,
     trim: lmt.ToTrim = 0,
-    scale: float = 1.0,
+    scale: _Float = 1,
 ) -> _FloatND: ...
 def l_moment_bounds(
     r: lmt.ToOrder,
     /,
     trim: lmt.ToTrim = 0,
-    scale: float = 1.0,
+    scale: _Float = 1,
 ) -> float | _FloatND:
     r"""
     Returns the absolute upper bounds $L^{(s,t)}_r$ on L-moments

lmo/distributions/_genlambda.py

@@ -10,7 +10,7 @@
 from typing import Final, TypeAlias, TypeVar
 
 import numpy as np
-import numpy.typing as npt
+import optype.numpy as onp
 import scipy.special as sps
 from scipy.stats.distributions import rv_continuous
 
@@ -29,7 +29,7 @@
 
 
 _F8: TypeAlias = float | np.float64
-_ArrF8: TypeAlias = npt.NDArray[np.float64]
+_ArrF8: TypeAlias = onp.ArrayND[np.float64]
 
 _XT = TypeVar("_XT", bound=_F8 | _ArrF8)
 
@@ -186,7 +186,7 @@ def _stats(
         a, c = 1 + f, 1 - f
         b1, d1 = 1 + b, 1 + d
 
-        m1 = 0 if b == d and f == 0 else float(lm_genlambda(1, 0, 0, b, d, f).item())
+        m1: float = 0.0 if b == d and f == 0 else lm_genlambda(1, 0, 0, b, d, f).item()
 
         if b <= -1 / 2 or d <= -1 / 2:
             return m1, math.nan, math.nan, math.nan
@@ -209,7 +209,7 @@ def _stats(
             m2 = (
                 (a / b1) ** 2 / (b1 + b)
                 + (c / d1) ** 2 / (d1 + d)
-                + 2 * a * c / (b * d) * (1 / (b1 * d1) - sps.beta(b1, d1))
+                + 2 * a * c / (b * d) * (1 / (b1 * d1) - sps.beta(b1, d1).item())
             )
 
         # Feeling adventurous? You're welcome to contribute these missing
@@ -229,7 +229,7 @@ def _entropy(self, b: float, d: float, f: float) -> float:
 
     def _l_moment(
         self,
-        r: npt.NDArray[np.intp],
+        r: onp.ArrayND[lmt.Integer],
         b: float,
         d: float,
         f: float,