# -*- coding: utf-8 -*-
import numpy as np
from numpy import (abs, asarray, cos, exp, floor, pi, sign, sin, sqrt, sum,
                   size, tril, isnan, atleast_2d, repeat)
from numpy.testing import assert_almost_equal

from .go_benchmark import Benchmark


class CarromTable(Benchmark):
    r"""
    CarromTable objective function.

    The CarromTable [1]_ global optimization problem is a multimodal
    minimization problem defined as follows:

    .. math::

        f_{\text{CarromTable}}(x) = - \frac{1}{30}\left(\cos(x_1)
        cos(x_2) e^{\left|1 - \frac{\sqrt{x_1^2 + x_2^2}}{\pi}\right|}\right)^2


    with :math:`x_i \in [-10, 10]` for :math:`i = 1, 2`.

    *Global optimum*: :math:`f(x) = -24.15681551650653` for :math:`x_i = \pm
    9.646157266348881` for :math:`i = 1, 2`

    .. [1] Jamil, M. & Yang, X.-S. A Literature Survey of Benchmark Functions
    For Global Optimization Problems Int. Journal of Mathematical Modelling
    and Numerical Optimisation, 2013, 4, 150-194.

    """

    def __init__(self, dimensions=2):
        Benchmark.__init__(self, dimensions)

        self._bounds = list(zip([-10.0] * self.N, [10.0] * self.N))
        self.global_optimum = [(9.646157266348881, 9.646134286497169),
                               (-9.646157266348881, 9.646134286497169),
                               (9.646157266348881, -9.646134286497169),
                               (-9.646157266348881, -9.646134286497169)]
        self.fglob = -24.15681551650653

    def fun(self, x, *args):
        self.nfev += 1

        u = cos(x[0]) * cos(x[1])
        v = sqrt(x[0] ** 2 + x[1] ** 2)
        return -((u * exp(abs(1 - v / pi))) ** 2) / 30.


class Chichinadze(Benchmark):
    r"""
    Chichinadze objective function.

    This class defines the Chichinadze [1]_ global optimization problem. This is a
    multimodal minimization problem defined as follows:

    .. math::

        f_{\text{Chichinadze}}(x) = x_{1}^{2} - 12 x_{1}
        + 8 \sin\left(\frac{5}{2} \pi x_{1}\right)
        + 10 \cos\left(\frac{1}{2} \pi x_{1}\right) + 11
        - 0.2 \frac{\sqrt{5}}{e^{\frac{1}{2} \left(x_{2} -0.5 \right)^{2}}}


    with :math:`x_i \in [-30, 30]` for :math:`i = 1, 2`.

    *Global optimum*: :math:`f(x) = -42.94438701899098` for :math:`x =
    [6.189866586965680, 0.5]`

    .. [1] Gavana, A. Global Optimization Benchmarks and AMPGO retrieved 2015

    TODO: Jamil#33 has a dividing factor of 2 in the sin term.  However, f(x)
    for the given solution does not give the global minimum. i.e. the equation
    is at odds with the solution.
    Only by removing the dividing factor of 2, i.e. `8 * sin(5 * pi * x[0])`
    does the given solution result in the given global minimum.
    Do we keep the result or equation?
    """

    def __init__(self, dimensions=2):
        Benchmark.__init__(self, dimensions)

        self._bounds = list(zip([-30.0] * self.N, [30.0] * self.N))
        self.custom_bounds = [(-10, 10), (-10, 10)]

        self.global_optimum = [[6.189866586965680, 0.5]]
        self.fglob = -42.94438701899098

    def fun(self, x, *args):
        self.nfev += 1

        return (x[0] ** 2 - 12 * x[0] + 11 + 10 * cos(pi * x[0] / 2)
                + 8 * sin(5 * pi * x[0] / 2)
                - 1.0 / sqrt(5) * exp(-((x[1] - 0.5) ** 2) / 2))


class Cigar(Benchmark):
    r"""
    Cigar objective function.

    This class defines the Cigar [1]_ global optimization problem. This
    is a multimodal minimization problem defined as follows:

    .. math::

        f_{\text{Cigar}}(x) = x_1^2 + 10^6\sum_{i=2}^{n} x_i^2


    Here, :math:`n` represents the number of dimensions and :math:`x_i \in
    [-100, 100]` for :math:`i = 1, ..., n`.

    *Global optimum*: :math:`f(x) = 0` for :math:`x_i = 0` for
    :math:`i = 1, ..., n`

    .. [1] Gavana, A. Global Optimization Benchmarks and AMPGO retrieved 2015
    """

    def __init__(self, dimensions=2):
        Benchmark.__init__(self, dimensions)

        self._bounds = list(zip([-100.0] * self.N,
                                [100.0] * self.N))
        self.custom_bounds = [(-5, 5), (-5, 5)]

        self.global_optimum = [[0 for _ in range(self.N)]]
        self.fglob = 0.0
        self.change_dimensionality = True

    def fun(self, x, *args):
        self.nfev += 1

        return x[0] ** 2 + 1e6 * sum(x[1:] ** 2)


class Cola(Benchmark):
    r"""
    Cola objective function.

    This class defines the Cola global optimization problem. The 17-dimensional
    function computes indirectly the formula :math:`f(n, u)` by setting
    :math:`x_0 = y_0, x_1 = u_0, x_i = u_{2(i2)}, y_i = u_{2(i2)+1}` :

    .. math::

        f_{\text{Cola}}(x) = \sum_{i<j}^{n} \left (r_{i,j} - d_{i,j} \right )^2


    Where :math:`r_{i, j}` is given by:

    .. math::

        r_{i, j} = \sqrt{(x_i - x_j)^2 + (y_i - y_j)^2}

    And :math:`d` is a symmetric matrix given by:

    .. math::

        \{d} = \left [ d_{ij} \right ] = \begin{pmatrix}
        1.27 &  &  &  &  &  &  &  & \\
        1.69 & 1.43 &  &  &  &  &  &  & \\
        2.04 & 2.35 & 2.43 &  &  &  &  &  & \\
        3.09 & 3.18 & 3.26 & 2.85  &  &  &  &  & \\
        3.20 & 3.22 & 3.27 & 2.88 & 1.55 &  &  &  & \\
        2.86 & 2.56 & 2.58 & 2.59 & 3.12 & 3.06  &  &  & \\
        3.17 & 3.18 & 3.18 & 3.12 & 1.31 & 1.64 & 3.00  & \\
        3.21 & 3.18 & 3.18 & 3.17 & 1.70 & 1.36 & 2.95 & 1.32  & \\
        2.38 & 2.31 & 2.42 & 1.94 & 2.85 & 2.81 & 2.56 & 2.91 & 2.97
        \end{pmatrix}


    This function has bounds :math:`x_0 \in [0, 4]` and :math:`x_i \in [-4, 4]`
    for :math:`i = 1, ..., n-1`.
    *Global optimum* 11.7464.

    .. [1] Jamil, M. & Yang, X.-S. A Literature Survey of Benchmark Functions
    For Global Optimization Problems Int. Journal of Mathematical Modelling
    and Numerical Optimisation, 2013, 4, 150-194.
    """

    def __init__(self, dimensions=17):
        Benchmark.__init__(self, dimensions)

        self._bounds = [[0.0, 4.0]] + list(zip([-4.0] * (self.N - 1),
                                               [4.0] * (self.N - 1)))

        self.global_optimum = [[0.651906, 1.30194, 0.099242, -0.883791,
                                -0.8796, 0.204651, -3.28414, 0.851188,
                                -3.46245, 2.53245, -0.895246, 1.40992,
                                -3.07367, 1.96257, -2.97872, -0.807849,
                                -1.68978]]
        self.fglob = 11.7464

        self.d = asarray([[0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
                 [1.27, 0, 0, 0, 0, 0, 0, 0, 0, 0],
                 [1.69, 1.43, 0, 0, 0, 0, 0, 0, 0, 0],
                 [2.04, 2.35, 2.43, 0, 0, 0, 0, 0, 0, 0],
                 [3.09, 3.18, 3.26, 2.85, 0, 0, 0, 0, 0, 0],
                 [3.20, 3.22, 3.27, 2.88, 1.55, 0, 0, 0, 0, 0],
                 [2.86, 2.56, 2.58, 2.59, 3.12, 3.06, 0, 0, 0, 0],
                 [3.17, 3.18, 3.18, 3.12, 1.31, 1.64, 3.00, 0, 0, 0],
                 [3.21, 3.18, 3.18, 3.17, 1.70, 1.36, 2.95, 1.32, 0, 0],
                 [2.38, 2.31, 2.42, 1.94, 2.85, 2.81, 2.56, 2.91, 2.97, 0.]])

    def fun(self, x, *args):
        self.nfev += 1

        xi = atleast_2d(asarray([0.0, x[0]] + list(x[1::2])))
        xj = repeat(xi, size(xi, 1), axis=0)
        xi = xi.T

        yi = atleast_2d(asarray([0.0, 0.0] + list(x[2::2])))
        yj = repeat(yi, size(yi, 1), axis=0)
        yi = yi.T

        inner = (sqrt(((xi - xj) ** 2 + (yi - yj) ** 2)) - self.d) ** 2
        inner = tril(inner, -1)
        return sum(sum(inner, axis=1))


class Colville(Benchmark):
    r"""
    Colville objective function.

    This class defines the Colville global optimization problem. This
    is a multimodal minimization problem defined as follows:

    .. math::

        f_{\text{Colville}}(x) = \left(x_{1} -1\right)^{2}
        + 100 \left(x_{1}^{2} - x_{2}\right)^{2}
        + 10.1 \left(x_{2} -1\right)^{2} + \left(x_{3} -1\right)^{2}
        + 90 \left(x_{3}^{2} - x_{4}\right)^{2}
        + 10.1 \left(x_{4} -1\right)^{2} + 19.8 \frac{x_{4} -1}{x_{2}}


    with :math:`x_i \in [-10, 10]` for :math:`i = 1, ..., 4`.

    *Global optimum*: :math:`f(x) = 0` for :math:`x_i = 1` for
    :math:`i = 1, ..., 4`

    .. [1] Jamil, M. & Yang, X.-S. A Literature Survey of Benchmark Functions
    For Global Optimization Problems Int. Journal of Mathematical Modelling
    and Numerical Optimisation, 2013, 4, 150-194.

    TODO docstring equation is wrong use Jamil#36
    """

    def __init__(self, dimensions=4):
        Benchmark.__init__(self, dimensions)

        self._bounds = list(zip([-10.0] * self.N, [10.0] * self.N))

        self.global_optimum = [[1 for _ in range(self.N)]]
        self.fglob = 0.0

    def fun(self, x, *args):

        self.nfev += 1
        return (100 * (x[0] - x[1] ** 2) ** 2
                + (1 - x[0]) ** 2 + (1 - x[2]) ** 2
                + 90 * (x[3] - x[2] ** 2) ** 2
                + 10.1 * ((x[1] - 1) ** 2 + (x[3] - 1) ** 2)
                + 19.8 * (x[1] - 1) * (x[3] - 1))


class Corana(Benchmark):
    r"""
    Corana objective function.

    This class defines the Corana [1]_ global optimization problem. This
    is a multimodal minimization problem defined as follows:

    .. math::

        f_{\text{Corana}}(x) = \begin{cases} \sum_{i=1}^n 0.15 d_i
        [z_i - 0.05\textrm{sgn}(z_i)]^2 & \textrm{if }|x_i-z_i| < 0.05 \\
        d_ix_i^2 & \textrm{otherwise}\end{cases}


    Where, in this exercise:

    .. math::

        z_i = 0.2 \lfloor |x_i/s_i|+0.49999\rfloor\textrm{sgn}(x_i),
        d_i=(1,1000,10,100, ...)


    with :math:`x_i \in [-5, 5]` for :math:`i = 1, ..., 4`.

    *Global optimum*: :math:`f(x) = 0` for :math:`x_i = 0` for
    :math:`i = 1, ..., 4`

    ..[1] Gavana, A. Global Optimization Benchmarks and AMPGO retrieved 2015
    """

    def __init__(self, dimensions=4):
        Benchmark.__init__(self, dimensions)

        self._bounds = list(zip([-5.0] * self.N, [5.0] * self.N))

        self.global_optimum = [[0 for _ in range(self.N)]]
        self.fglob = 0.0

    def fun(self, x, *args):
        self.nfev += 1

        d = [1., 1000., 10., 100.]
        r = 0
        for j in range(4):
            zj = floor(abs(x[j] / 0.2) + 0.49999) * sign(x[j]) * 0.2
            if abs(x[j] - zj) < 0.05:
                r += 0.15 * ((zj - 0.05 * sign(zj)) ** 2) * d[j]
            else:
                r += d[j] * x[j] * x[j]
        return r


class CosineMixture(Benchmark):
    r"""
    Cosine Mixture objective function.

    This class defines the Cosine Mixture global optimization problem. This
    is a multimodal minimization problem defined as follows:

    .. math::

        f_{\text{CosineMixture}}(x) = -0.1 \sum_{i=1}^n \cos(5 \pi x_i)
        - \sum_{i=1}^n x_i^2


    Here, :math:`n` represents the number of dimensions and :math:`x_i \in
    [-1, 1]` for :math:`i = 1, ..., N`.

    *Global optimum*: :math:`f(x) = -0.1N` for :math:`x_i = 0` for
    :math:`i = 1, ..., N`

    .. [1] Jamil, M. & Yang, X.-S. A Literature Survey of Benchmark Functions
    For Global Optimization Problems Int. Journal of Mathematical Modelling
    and Numerical Optimisation, 2013, 4, 150-194.

    TODO, Jamil #38 has wrong minimum and wrong fglob. I plotted it.
    -(x**2) term is always negative if x is negative.
     cos(5 * pi * x) is equal to -1 for x=-1.
    """

    def __init__(self, dimensions=2):
        Benchmark.__init__(self, dimensions)

        self.change_dimensionality = True
        self._bounds = list(zip([-1.0] * self.N, [1.0] * self.N))

        self.global_optimum = [[-1. for _ in range(self.N)]]
        self.fglob = -0.9 * self.N

    def fun(self, x, *args):
        self.nfev += 1

        return -0.1 * sum(cos(5.0 * pi * x)) - sum(x ** 2.0)


class CrossInTray(Benchmark):
    r"""
    Cross-in-Tray objective function.

    This class defines the Cross-in-Tray [1]_ global optimization problem. This is a
    multimodal minimization problem defined as follows:

    .. math::

        f_{\text{CrossInTray}}(x) = - 0.0001 \left(\left|{e^{\left|{100
        - \frac{\sqrt{x_{1}^{2} + x_{2}^{2}}}{\pi}}\right|}
        \sin\left(x_{1}\right) \sin\left(x_{2}\right)}\right| + 1\right)^{0.1}


    with :math:`x_i \in [-15, 15]` for :math:`i = 1, 2`.

    *Global optimum*: :math:`f(x) = -2.062611870822739` for :math:`x_i =
    \pm 1.349406608602084` for :math:`i = 1, 2`

    .. [1] Jamil, M. & Yang, X.-S. A Literature Survey of Benchmark Functions
    For Global Optimization Problems Int. Journal of Mathematical Modelling
    and Numerical Optimisation, 2013, 4, 150-194.
    """

    def __init__(self, dimensions=2):
        Benchmark.__init__(self, dimensions)

        self._bounds = list(zip([-10.0] * self.N, [10.0] * self.N))

        self.global_optimum = [(1.349406685353340, 1.349406608602084),
                               (-1.349406685353340, 1.349406608602084),
                               (1.349406685353340, -1.349406608602084),
                               (-1.349406685353340, -1.349406608602084)]
        self.fglob = -2.062611870822739

    def fun(self, x, *args):

        self.nfev += 1
        return (-0.0001 * (abs(sin(x[0]) * sin(x[1])
                           * exp(abs(100 - sqrt(x[0] ** 2 + x[1] ** 2) / pi)))
                           + 1) ** (0.1))


class CrossLegTable(Benchmark):
    r"""
    Cross-Leg-Table objective function.

    This class defines the Cross-Leg-Table [1]_ global optimization problem. This
    is a multimodal minimization problem defined as follows:

    .. math::

        f_{\text{CrossLegTable}}(x) = - \frac{1}{\left(\left|{e^{\left|{100
        - \frac{\sqrt{x_{1}^{2} + x_{2}^{2}}}{\pi}}\right|}
        \sin\left(x_{1}\right) \sin\left(x_{2}\right)}\right| + 1\right)^{0.1}}


    with :math:`x_i \in [-10, 10]` for :math:`i = 1, 2`.

    *Global optimum*: :math:`f(x) = -1`. The global minimum is found on the
    planes :math:`x_1 = 0` and :math:`x_2 = 0`

    ..[1] Mishra, S. Global Optimization by Differential Evolution and Particle
    Swarm Methods: Evaluation on Some Benchmark Functions Munich University,
    2006
    """

    def __init__(self, dimensions=2):
        Benchmark.__init__(self, dimensions)

        self._bounds = list(zip([-10.0] * self.N, [10.0] * self.N))

        self.global_optimum = [[0., 0.]]
        self.fglob = -1.0

    def fun(self, x, *args):
        self.nfev += 1

        u = 100 - sqrt(x[0] ** 2 + x[1] ** 2) / pi
        v = sin(x[0]) * sin(x[1])
        return -(abs(v * exp(abs(u))) + 1) ** (-0.1)


class CrownedCross(Benchmark):
    r"""
    Crowned Cross objective function.

    This class defines the Crowned Cross [1]_ global optimization problem. This
    is a multimodal minimization problem defined as follows:

    .. math::

        f_{\text{CrownedCross}}(x) = 0.0001 \left(\left|{e^{\left|{100
        - \frac{\sqrt{x_{1}^{2} + x_{2}^{2}}}{\pi}}\right|}
        \sin\left(x_{1}\right) \sin\left(x_{2}\right)}\right| + 1\right)^{0.1}


    with :math:`x_i \in [-10, 10]` for :math:`i = 1, 2`.

    *Global optimum*: :math:`f(x_i) = 0.0001`. The global minimum is found on
    the planes :math:`x_1 = 0` and :math:`x_2 = 0`

    ..[1] Mishra, S. Global Optimization by Differential Evolution and Particle
    Swarm Methods: Evaluation on Some Benchmark Functions Munich University,
    2006
    """

    def __init__(self, dimensions=2):
        Benchmark.__init__(self, dimensions)

        self._bounds = list(zip([-10.0] * self.N, [10.0] * self.N))

        self.global_optimum = [[0, 0]]
        self.fglob = 0.0001

    def fun(self, x, *args):

        self.nfev += 1
        u = 100 - sqrt(x[0] ** 2 + x[1] ** 2) / pi
        v = sin(x[0]) * sin(x[1])
        return 0.0001 * (abs(v * exp(abs(u))) + 1) ** (0.1)


class Csendes(Benchmark):
    r"""
    Csendes objective function.

    This class defines the Csendes [1]_ global optimization problem. This is a
    multimodal minimization problem defined as follows:

    .. math::

        f_{\text{Csendes}}(x) = \sum_{i=1}^n x_i^6 \left[ 2 + \sin
        \left( \frac{1}{x_i} \right ) \right]


    Here, :math:`n` represents the number of dimensions and :math:`x_i \in
    [-1, 1]` for :math:`i = 1, ..., N`.

    *Global optimum*: :math:`f(x) = 0.0` for :math:`x_i = 0` for
    :math:`i = 1, ..., N`

    .. [1] Jamil, M. & Yang, X.-S. A Literature Survey of Benchmark Functions
    For Global Optimization Problems Int. Journal of Mathematical Modelling
    and Numerical Optimisation, 2013, 4, 150-194.
    """

    def __init__(self, dimensions=2):
        Benchmark.__init__(self, dimensions)

        self.change_dimensionality = True
        self._bounds = list(zip([-1.0] * self.N, [1.0] * self.N))

        self.global_optimum = [[0 for _ in range(self.N)]]
        self.fglob = np.nan

    def fun(self, x, *args):
        self.nfev += 1

        try:
            return sum((x ** 6.0) * (2.0 + sin(1.0 / x)))
        except ZeroDivisionError:
            return np.nan
        except FloatingPointError:
            return np.nan

    def success(self, x):
        """Is a candidate solution at the global minimum"""
        val = self.fun(asarray(x))
        if isnan(val):
            return True
        try:
            assert_almost_equal(val, 0., 4)
            return True
        except AssertionError:
            return False

        return False


class Cube(Benchmark):
    r"""
    Cube objective function.

    This class defines the Cube global optimization problem. This
    is a multimodal minimization problem defined as follows:

    .. math::

        f_{\text{Cube}}(x) = 100(x_2 - x_1^3)^2 + (1 - x1)^2


    Here, :math:`n` represents the number of dimensions and :math:`x_i \in [-10, 10]` for :math:`i=1,...,N`.

    *Global optimum*: :math:`f(x_i) = 0.0` for :math:`x = [1, 1]`

    .. [1] Jamil, M. & Yang, X.-S. A Literature Survey of Benchmark Functions
    For Global Optimization Problems Int. Journal of Mathematical Modelling
    and Numerical Optimisation, 2013, 4, 150-194.

    TODO: jamil#41 has the wrong solution.
    """

    def __init__(self, dimensions=2):
        Benchmark.__init__(self, dimensions)

        self._bounds = list(zip([-10.0] * self.N, [10.0] * self.N))
        self.custom_bounds = ([0, 2], [0, 2])

        self.global_optimum = [[1.0, 1.0]]
        self.fglob = 0.0

    def fun(self, x, *args):

        self.nfev += 1
        return 100.0 * (x[1] - x[0] ** 3.0) ** 2.0 + (1.0 - x[0]) ** 2.0