dev(narugo): add test and docs for sigmoid function

imgutils.utils.func

====================================

.. currentmodule:: imgutils.utils.func

.. automodule:: imgutils.utils.func

sigmoid

-------------------------

.. autofunction:: sigmoid

.. toctree::

    :maxdepth: 3

    func

    onnxruntime

"""

This module provides mathematical functions related to neural networks.

It includes the sigmoid activation function, which is commonly used in various

machine learning and deep learning models. The sigmoid function maps any input

value to a value between 0 and 1, making it useful for binary classification

problems and as an activation function in neural network layers.

Usage:

    >>> from imgutils.utils import sigmoid

    >>> result = sigmoid(input_value)

"""

import numpy as np

__all__ = ['sigmoid']

def sigmoid(x):

    """

    Compute the sigmoid function for the input.

    The sigmoid function is defined as:

    :math:`f\\left(x\\right) = \\frac{1}{1 + e^{-x}}`

    This function applies the sigmoid activation to either a single number

    or an array of numbers using NumPy for efficient computation.

    :param x: Input value or array of values.

    :type x: float or numpy.ndarray

    :return: Sigmoid of the input.

    :rtype: float or numpy.ndarray

    :example:

        >>> import numpy as np

        >>> sigmoid(0)

        0.5

        >>> sigmoid(np.array([-1, 0, 1]))

        array([0.26894142, 0.5       , 0.73105858])

    """

    return 1 / (1 + np.exp(-x))

import numpy as np

import pytest

from imgutils.utils import sigmoid

@pytest.fixture

def sample_input():

    return np.array([-1, 0, 1])

@pytest.fixture

def expected_output():

    return np.array([0.26894142, 0.5, 0.73105858])

@pytest.mark.unittest

class TestUtilsFuncSigmoid:

    def test_sigmoid_scalar(self):

        assert np.isclose(sigmoid(0), 0.5)

    def test_sigmoid_array(self, sample_input, expected_output):

        result = sigmoid(sample_input)

        np.testing.assert_array_almost_equal(result, expected_output)

    def test_sigmoid_large_positive(self):

        assert np.isclose(sigmoid(100), 1.0)

    def test_sigmoid_large_negative(self):

        assert np.isclose(sigmoid(-100), 0.0)

    def test_sigmoid_zero(self):

        assert sigmoid(0) == 0.5

    def test_sigmoid_type(self):

        assert isinstance(sigmoid(1), float)

        assert isinstance(sigmoid(np.array([1])), np.ndarray)

    def test_sigmoid_shape(self):

        input_array = np.array([[1, 2], [3, 4]])

        result = sigmoid(input_array)

        assert result.shape == input_array.shape