from collections.abc import Iterable, Iterator, Mapping, Sequence
from typing import Callable, ClassVar, Literal
from typing_extensions import Self

from numpy import ndarray
from numpy.random import RandomState

from .._typing import ArrayLike, Float, Int, MatrixLike
from ..base import BaseEstimator
from ..svm._classes import SVC
from . import BaseCrossValidator
from ._search import BaseSearchCV

__all__ = ["HalvingGridSearchCV", "HalvingRandomSearchCV"]

class _SubsampleMetaSplitter:
    def __init__(self, *, base_cv, fraction, subsample_test, random_state) -> None: ...
    def split(self, X: ndarray, y: ndarray, groups=None) -> Iterator[tuple[ndarray, ndarray]]: ...

class BaseSuccessiveHalving(BaseSearchCV):
    def __init__(
        self,
        estimator,
        *,
        scoring=None,
        n_jobs=None,
        refit: bool = True,
        cv: int = 5,
        verbose: int = 0,
        random_state=None,
        error_score=...,
        return_train_score: bool = True,
        max_resources: str = "auto",
        min_resources: str = "exhaust",
        resource: str = "n_samples",
        factor: int = 3,
        aggressive_elimination: bool = False,
    ) -> None: ...
    def fit(
        self,
        X: MatrixLike,
        y: None | MatrixLike | ArrayLike = None,
        groups: None | ArrayLike = None,
        **fit_params,
    ) -> Self: ...

class HalvingGridSearchCV(BaseSuccessiveHalving):
    feature_names_in_: ndarray = ...
    n_features_in_: int = ...
    classes_: ndarray = ...
    multimetric_: bool = ...
    refit_time_: float = ...
    n_splits_: int = ...
    scorer_: Callable | dict = ...
    best_index_: int = ...
    best_params_: dict = ...
    best_score_: float = ...
    best_estimator_: BaseEstimator | dict = ...
    cv_results_: dict[str, ndarray] = ...
    n_required_iterations_: int = ...
    n_possible_iterations_: int = ...
    n_iterations_: int = ...
    min_resources_: int = ...
    max_resources_: int = ...
    n_remaining_candidates_: int = ...
    n_candidates_: list[int] = ...
    n_resources_: list[int] = ...

    _required_parameters: ClassVar[list] = ...

    def __init__(
        self,
        estimator: BaseEstimator | SVC,
        param_grid: Mapping | dict[str, list[int | float]] | Sequence[dict],
        *,
        factor: float = 3,
        resource: str = "n_samples",
        max_resources: str | Int = "auto",
        min_resources: int | Literal["exhaust", "smallest"] = "exhaust",
        aggressive_elimination: bool = False,
        cv: Iterable | int | BaseCrossValidator = 5,
        scoring: None | str | Callable = None,
        refit: bool = True,
        error_score: str | Float = ...,
        return_train_score: bool = True,
        random_state: RandomState | None | Int = None,
        n_jobs: None | int = None,
        verbose: Int = 0,
    ) -> None: ...

class HalvingRandomSearchCV(BaseSuccessiveHalving):
    feature_names_in_: ndarray = ...
    n_features_in_: int = ...
    classes_: ndarray = ...
    multimetric_: bool = ...
    refit_time_: float = ...
    n_splits_: int = ...
    scorer_: Callable | dict = ...
    best_index_: int = ...
    best_params_: dict = ...
    best_score_: float = ...
    best_estimator_: BaseEstimator | dict = ...
    cv_results_: dict[str, ndarray] = ...
    n_required_iterations_: int = ...
    n_possible_iterations_: int = ...
    n_iterations_: int = ...
    min_resources_: int = ...
    max_resources_: int = ...
    n_remaining_candidates_: int = ...
    n_candidates_: list[int] = ...
    n_resources_: list[int] = ...

    _required_parameters: ClassVar[list] = ...

    def __init__(
        self,
        estimator: BaseEstimator,
        param_distributions: dict,
        *,
        n_candidates: str | Int = "exhaust",
        factor: float = 3,
        resource: str = "n_samples",
        max_resources: str | Int = "auto",
        min_resources: Literal["exhaust", "smallest"] | int = "smallest",
        aggressive_elimination: bool = False,
        cv: Iterable | int | BaseCrossValidator = 5,
        scoring: None | str | Callable = None,
        refit: bool = True,
        error_score: str | Float = ...,
        return_train_score: bool = True,
        random_state: RandomState | None | Int = None,
        n_jobs: None | int = None,
        verbose: Int = 0,
    ) -> None: ...