"""Tests for the abstract Optimizer base class and dataclasses."""

from __future__ import annotations

import time

import numpy as np
import pytest

from ahdcma.algorithms.base import (
    History,
    OptimizationResult,
    Optimizer,
    SearchSpace,
)


def test_search_space_unit_cube() -> None:
    sp = SearchSpace.unit_cube(4)
    assert sp.dim == 4
    assert np.all(sp.lower == 0.0)
    assert np.all(sp.upper == 1.0)


def test_search_space_validation() -> None:
    with pytest.raises(ValueError):
        SearchSpace(dim=3, lower=np.zeros(2), upper=np.ones(3))
    with pytest.raises(ValueError):
        SearchSpace(dim=2, lower=np.array([0.0, 1.0]), upper=np.array([1.0, 0.5]))


def test_search_space_clip() -> None:
    sp = SearchSpace.unit_cube(3)
    x = np.array([-0.5, 0.3, 1.7])
    np.testing.assert_array_equal(sp.clip(x), np.array([0.0, 0.3, 1.0]))


def test_history_append_and_len() -> None:
    h = History()
    assert len(h) == 0
    h.append(0, np.zeros((4, 2)), np.array([1.0, 0.5, 2.0, 1.5]), mode="explore", entropy=2.3)
    assert len(h) == 1
    assert h.best_fitness[0] == 0.5
    assert h.mode_per_gen[0] == "explore"
    assert h.entropy_per_gen[0] == pytest.approx(2.3)


def test_optimizer_subclass_minimal() -> None:
    class _Const(Optimizer):
        def optimize(self) -> OptimizationResult:
            x = np.full(self.search_space.dim, 0.5)
            f = float(self.fitness_fn(x))
            self._history.append(0, x[None, :], np.array([f]), mode="hybrid")
            return OptimizationResult(
                best_x=x,
                best_f=f,
                history=self._history,
                config_snapshot=self.config,
                run_id=self.run_id,
                wall_time=0.0,
            )

    sp = SearchSpace.unit_cube(3)
    opt = _Const(config={"foo": 1}, fitness_fn=lambda x: float(np.sum(x**2)), search_space=sp)
    t0 = time.time()
    result = opt.optimize()
    assert result.best_f == pytest.approx(0.75)
    assert result.run_id == "anonymous"
    assert len(opt.get_history()) == 1
    assert time.time() - t0 < 1.0