"""Test methods for faster fishers"""
import numpy as np
import pytest
from faster_fishers import exact, exact_with_odds_ratios
from scipy import stats

np.random.seed(42)
values = np.random.randint(1000, size=(10000, 4)).astype(dtype=np.uint64)


def test_exact():
    """Test the exact method with numpy arrays as input"""
    a_values = np.array([1, 3], dtype=np.uint64)
    b_values = np.array([2, 5], dtype=np.uint64)
    c_values = np.array([1, 4], dtype=np.uint64)
    d_values = np.array([5, 50], dtype=np.uint64)

    lesses = exact(a_values, b_values, c_values, d_values, "less")
    np.testing.assert_array_almost_equal(lesses, np.array([0.916666, 0.996303]))

    greaters = exact(a_values, b_values, c_values, d_values, "greater")
    np.testing.assert_array_almost_equal(greaters, np.array([0.583333, 0.039707]))

    two_tails = exact(a_values, b_values, c_values, d_values, "two-sided")
    np.testing.assert_array_almost_equal(two_tails, np.array([1.0, 0.039707]))

    odds_greaters = exact_with_odds_ratios(a_values, b_values, c_values, d_values, "greater")
    np.testing.assert_array_almost_equal(odds_greaters[0], np.array([2.5, 7.5]))
    np.testing.assert_array_almost_equal(odds_greaters[1], greaters)


@pytest.mark.benchmark
def test_benchmark_faster_fischer(benchmark):
    """Benchmark faster fishers."""
    _ = benchmark(exact_with_odds_ratios, values[:, 0], values[:, 1], values[:, 2], values[:, 3], "greater")


@pytest.mark.benchmark
def test_benchmark_scipy(benchmark):
    """Benchmark scipy fishers."""

    def scipy_p_values(a, b, c, d, alternative) -> float:
        """Return scipy's fishers's exact test."""
        odds_ratio, p_value = stats.fisher_exact([[a, b], [c, d]], alternative=alternative)
        return p_value

    function = np.vectorize(scipy_p_values)
    _ = benchmark(function, values[:, 0], values[:, 1], values[:, 2], values[:, 3], "greater")