fix scipy.stats.mstats stubtests

scipy-stubs/_typing.pyi

@@ -1,4 +1,5 @@
-# Helper types for internal use (type-check only).
+# NOTE(scipy-stubs): This ia a module only exists `if typing.TYPE_CHECKING: ...`
+
 from os import PathLike
 from collections.abc import Callable, Sequence
 from typing import IO, Any, Literal, Protocol, TypeAlias, type_check_only
@@ -10,6 +11,7 @@ import optype.numpy as onpt
 
 __all__ = [
     "RNG",
+    "Alternative",
     "AnyBool",
     "AnyChar",
     "AnyComplex",
@@ -74,6 +76,7 @@ CorrelateMode: TypeAlias = Literal["valid", "same", "full"]
 
 # scipy literals
 NanPolicy: TypeAlias = Literal["raise", "propagate", "omit"]
+Alternative: TypeAlias = Literal["two-sided", "less", "greater"]
 
 # used in `scipy.linalg.blas` and `scipy.linalg.lapack`
 @type_check_only

scipy-stubs/stats/_mstats_basic.pyi

@@ -130,15 +130,15 @@ trimdoc: str
 def argstoarray(*args) -> Untyped: ...
 def find_repeats(arr) -> Untyped: ...
 def count_tied_groups(x, use_missing: bool = False) -> Untyped: ...
-def rankdata(data, axis: Untyped | None = None, use_missing: bool = False) -> Untyped: ...
+def rankdata(data, axis: int | None = None, use_missing: bool = False) -> Untyped: ...
 def mode(a, axis: int = 0) -> Untyped: ...
 def msign(x) -> Untyped: ...
 def pearsonr(x, y) -> Untyped: ...
 def spearmanr(
     x,
     y: Untyped | None = None,
     use_ties: bool = True,
-    axis: Untyped | None = None,
+    axis: int | None = None,
     nan_policy: str = "propagate",
     alternative: str = "two-sided",
 ) -> Untyped: ...
@@ -164,80 +164,92 @@ def kruskal(*args) -> Untyped: ...
 def ks_1samp(x, cdf, args=(), alternative: str = "two-sided", method: str = "auto") -> Untyped: ...
 def ks_2samp(data1, data2, alternative: str = "two-sided", method: str = "auto") -> Untyped: ...
 def kstest(data1, data2, args=(), alternative: str = "two-sided", method: str = "auto") -> Untyped: ...
-def trima(a, limits: Untyped | None = None, inclusive=(True, True)) -> Untyped: ...
-def trimr(a, limits: Untyped | None = None, inclusive=(True, True), axis: Untyped | None = None) -> Untyped: ...
+def trima(a, limits: Untyped | None = None, inclusive: tuple[bool, bool] = (True, True)) -> Untyped: ...
+def trimr(a, limits: Untyped | None = None, inclusive: tuple[bool, bool] = (True, True), axis: int | None = None) -> Untyped: ...
 def trim(
     a,
     limits: Untyped | None = None,
-    inclusive=(True, True),
+    inclusive: tuple[bool, bool] = (True, True),
     relative: bool = False,
-    axis: Untyped | None = None,
+    axis: int | None = None,
 ) -> Untyped: ...
 def trimboth(
     data,
     proportiontocut: float = 0.2,
-    inclusive=(True, True),
-    axis: Untyped | None = None,
+    inclusive: tuple[bool, bool] = (True, True),
+    axis: int | None = None,
 ) -> Untyped: ...
 def trimtail(
     data,
     proportiontocut: float = 0.2,
     tail: str = "left",
-    inclusive=(True, True),
-    axis: Untyped | None = None,
+    inclusive: tuple[bool, bool] = (True, True),
+    axis: int | None = None,
 ) -> Untyped: ...
 def trimmed_mean(
     a,
     limits=(0.1, 0.1),
-    inclusive=(1, 1),
+    inclusive: tuple[bool, bool] = (1, 1),
     relative: bool = True,
-    axis: Untyped | None = None,
+    axis: int | None = None,
 ) -> Untyped: ...
 def trimmed_var(
     a,
     limits=(0.1, 0.1),
-    inclusive=(1, 1),
+    inclusive: tuple[bool, bool] = (1, 1),
     relative: bool = True,
-    axis: Untyped | None = None,
+    axis: int | None = None,
     ddof: int = 0,
 ) -> Untyped: ...
 def trimmed_std(
     a,
     limits=(0.1, 0.1),
-    inclusive=(1, 1),
+    inclusive: tuple[bool, bool] = (1, 1),
     relative: bool = True,
-    axis: Untyped | None = None,
+    axis: int | None = None,
     ddof: int = 0,
 ) -> Untyped: ...
-def trimmed_stde(a, limits=(0.1, 0.1), inclusive=(1, 1), axis: Untyped | None = None) -> Untyped: ...
-def tmean(a, limits: Untyped | None = None, inclusive=(True, True), axis: Untyped | None = None) -> Untyped: ...
-def tvar(a, limits: Untyped | None = None, inclusive=(True, True), axis: int = 0, ddof: int = 1) -> Untyped: ...
+def trimmed_stde(a, limits=(0.1, 0.1), inclusive: tuple[bool, bool] = (1, 1), axis: int | None = None) -> Untyped: ...
+def tmean(a, limits: Untyped | None = None, inclusive: tuple[bool, bool] = (True, True), axis: int | None = None) -> Untyped: ...
+def tvar(
+    a,
+    limits: Untyped | None = None,
+    inclusive: tuple[bool, bool] = (True, True),
+    axis: int = 0,
+    ddof: int = 1,
+) -> Untyped: ...
 def tmin(a, lowerlimit: Untyped | None = None, axis: int = 0, inclusive: bool = True) -> Untyped: ...
 def tmax(a, upperlimit: Untyped | None = None, axis: int = 0, inclusive: bool = True) -> Untyped: ...
-def tsem(a, limits: Untyped | None = None, inclusive=(True, True), axis: int = 0, ddof: int = 1) -> Untyped: ...
+def tsem(
+    a,
+    limits: Untyped | None = None,
+    inclusive: tuple[bool, bool] = (True, True),
+    axis: int = 0,
+    ddof: int = 1,
+) -> Untyped: ...
 def winsorize(
     a,
     limits: Untyped | None = None,
-    inclusive=(True, True),
+    inclusive: tuple[bool, bool] = (True, True),
     inplace: bool = False,
-    axis: Untyped | None = None,
+    axis: int | None = None,
     nan_policy: str = "propagate",
 ) -> Untyped: ...
 def moment(a, moment: int = 1, axis: int = 0) -> Untyped: ...
 def variation(a, axis: int = 0, ddof: int = 0) -> Untyped: ...
 def skew(a, axis: int = 0, bias: bool = True) -> Untyped: ...
 def kurtosis(a, axis: int = 0, fisher: bool = True, bias: bool = True) -> Untyped: ...
 def describe(a, axis: int = 0, ddof: int = 0, bias: bool = True) -> Untyped: ...
-def stde_median(data, axis: Untyped | None = None) -> Untyped: ...
+def stde_median(data, axis: int | None = None) -> Untyped: ...
 def skewtest(a, axis: int = 0, alternative: str = "two-sided") -> Untyped: ...
 def kurtosistest(a, axis: int = 0, alternative: str = "two-sided") -> Untyped: ...
 def normaltest(a, axis: int = 0) -> Untyped: ...
 def mquantiles(
     a,
-    prob=(0.25, 0.5, 0.75),
+    prob=[0.25, 0.5, 0.75],
     alphap: float = 0.4,
     betap: float = 0.4,
-    axis: Untyped | None = None,
+    axis: int | None = None,
     limit=(),
 ) -> Untyped: ...
 def scoreatpercentile(data, per, limit=(), alphap: float = 0.4, betap: float = 0.4) -> Untyped: ...

scipy-stubs/stats/_mstats_extras.pyi

@@ -13,19 +13,19 @@ __all__ = [
     "trimmed_mean_ci",
 ]
 
-def hdquantiles(data, prob=(0.25, 0.5, 0.75), axis: Untyped | None = None, var: bool = False) -> Untyped: ...
+def hdquantiles(data, prob=[0.25, 0.5, 0.75], axis: int | None = None, var: bool = False) -> Untyped: ...
 def hdmedian(data, axis: int = -1, var: bool = False) -> Untyped: ...
-def hdquantiles_sd(data, prob=(0.25, 0.5, 0.75), axis: Untyped | None = None) -> Untyped: ...
+def hdquantiles_sd(data, prob=[0.25, 0.5, 0.75], axis: int | None = None) -> Untyped: ...
 def trimmed_mean_ci(
     data,
     limits=(0.2, 0.2),
     inclusive=(True, True),
     alpha: float = 0.05,
-    axis: Untyped | None = None,
+    axis: int | None = None,
 ) -> Untyped: ...
-def mjci(data, prob=(0.25, 0.5, 0.75), axis: Untyped | None = None) -> Untyped: ...
-def mquantiles_cimj(data, prob=(0.25, 0.5, 0.75), alpha: float = 0.05, axis: Untyped | None = None) -> Untyped: ...
-def median_cihs(data, alpha: float = 0.05, axis: Untyped | None = None) -> Untyped: ...
-def compare_medians_ms(group_1, group_2, axis: Untyped | None = None) -> Untyped: ...
-def idealfourths(data, axis: Untyped | None = None) -> Untyped: ...
+def mjci(data, prob=[0.25, 0.5, 0.75], axis: int | None = None) -> Untyped: ...
+def mquantiles_cimj(data, prob=[0.25, 0.5, 0.75], alpha: float = 0.05, axis: int | None = None) -> Untyped: ...
+def median_cihs(data, alpha: float = 0.05, axis: int | None = None) -> Untyped: ...
+def compare_medians_ms(group_1, group_2, axis: int | None = None) -> Untyped: ...
+def idealfourths(data, axis: int | None = None) -> Untyped: ...
 def rsh(data, points: Untyped | None = None) -> Untyped: ...

scipy-stubs/stats/_stats_mstats_common.pyi

@@ -1,8 +1,57 @@
-from scipy._typing import Untyped
+from typing import Any, Literal, TypeAlias
+from typing_extensions import Self
+
+import numpy as np
+import numpy.typing as npt
 from . import distributions as distributions
+from ._typing import BaseBunch
+
+__all__ = ["_find_repeats", "siegelslopes", "theilslopes"]
+
+_Method: TypeAlias = Literal["hierarchical", "separate"]
+
+class SiegelslopesResult(BaseBunch[float, float]):
+    def __new__(_cls, slope: float, intercept: float) -> Self: ...
+    def __init__(self, /, slope: float, intercept: float) -> None: ...
+    @property
+    def slope(self, /) -> float: ...
+    @property
+    def intercept(self, /) -> float: ...
 
-TheilslopesResult: Untyped
-SiegelslopesResult: Untyped
+class TheilslopesResult(BaseBunch[np.float64, np.float64, np.float64, np.float64]):
+    def __new__(
+        _cls,
+        slope: np.float64,
+        intercept: np.float64,
+        low_slope: np.float64,
+        high_slope: np.float64,
+    ) -> Self: ...
+    def __init__(
+        self,
+        /,
+        slope: np.float64,
+        intercept: np.float64,
+        low_slope: np.float64,
+        high_slope: np.float64,
+    ) -> None: ...
+    @property
+    def slope(self, /) -> np.float64: ...
+    @property
+    def intercept(self, /) -> np.float64: ...
+    @property
+    def low_slope(self, /) -> np.float64: ...
+    @property
+    def high_slope(self, /) -> np.float64: ...
 
-def theilslopes(y, x: Untyped | None = None, alpha: float = 0.95, method: str = "separate") -> Untyped: ...
-def siegelslopes(y, x: Untyped | None = None, method: str = "hierarchical") -> Untyped: ...
+def _find_repeats(arr: npt.NDArray[np.number[Any]]) -> tuple[npt.NDArray[np.float64], npt.NDArray[np.intp]]: ...
+def siegelslopes(
+    y: npt.ArrayLike,
+    x: npt.ArrayLike | None = None,
+    method: _Method = "hierarchical",
+) -> SiegelslopesResult: ...
+def theilslopes(
+    y: npt.ArrayLike,
+    x: npt.ArrayLike | None = None,
+    alpha: float | np.floating[Any] = 0.95,
+    method: _Method = "separate",
+) -> TheilslopesResult: ...