TP07_Regularisation.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NsKhV-BRUT_z"
      },
      "source": [
        "# 2CSSID-TP07. Régularisation et sélection d'attributs\n",
        "\n",
        "Dans ce TP, nous allons voir l'effet de la régularisation et la sélaction d'attributs sur l'apprentissage.\n",
        "Premierement, nous allons implémenter les deux fonctions de réglage L1 et L2, ainsi que la méthode de filtrage ANOVA.\n",
        "Ensuite, nous allons tester l'effet de ces techniques sur la performance en utilisant plusieurs métriques."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AUh9an5eUT_1"
      },
      "source": [
        "Binômes : \n",
        "- **Binôme 1 :** MAHMAHI Anis \n",
        "- **Binôme 2 :** BEKKAR Ilhem"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DOvjA0JlUT_2",
        "outputId": "0f8c741f-f97c-4516-b327-e879652c76a4"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "('1.21.6', '1.3.5', '3.2.2')"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ],
      "source": [
        "import matplotlib\n",
        "import numpy             as np\n",
        "import pandas            as pd \n",
        "import matplotlib.pyplot as plt \n",
        "from matplotlib      import colors \n",
        "%matplotlib inline\n",
        "\n",
        "np.__version__, pd.__version__, matplotlib.__version__"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 22,
      "metadata": {
        "id": "guChuLVCUT_3"
      },
      "outputs": [],
      "source": [
        "from typing import Tuple, List, Dict, Callable"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DeIuvoV4UT_4"
      },
      "source": [
        "## I. Réalisation des algorithmes\n",
        "\n",
        "Dans cette partie, nous allons implémenter deux fonctions de régularisation et une fonction de filtrage. \n",
        "La régularisation sera appliquée sur la régression logistique binaire dont le code est fourni ici."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 33,
      "metadata": {
        "id": "H2fX-tZqUT_4"
      },
      "outputs": [],
      "source": [
        "# Le code de la réression logistique \n",
        "# Ne modifier pas ceci\n",
        "\n",
        "# Somme pondérée des caractéristiques\n",
        "def zf(X: np.ndarray, theta: np.ndarray) -> np.ndarray:\n",
        "    return np.dot(X, theta)\n",
        "\n",
        "# Transformation d'une valeur à une probabilité\n",
        "def sg(Z: np.ndarray) -> np.ndarray:\n",
        "    return  1 / (1 + np.exp(-Z))\n",
        "\n",
        "# probabilité d'un échantillons\n",
        "def hf(X: np.ndarray, theta: np.ndarray) -> np.ndarray:\n",
        "    return sg(zf(X, theta))\n",
        "\n",
        "# Fonction objective\n",
        "# utilisation d'un petit nombre pour éviter log(0)\n",
        "def BCE(Y: np.ndarray, H: np.ndarray, e: float=1e-8) -> float: \n",
        "    return - np.mean(Y * np.log(H) + (1-Y) * np.log(1-H))  \n",
        "\n",
        "# Fonction des gradients\n",
        "def dBCE(X: np.ndarray, Y: np.ndarray, H: np.ndarray) -> np.ndarray:\n",
        "    return  np.dot(X.T, H-Y)/X.shape[0]\n",
        "\n",
        "# Fonction de mise à jour des paramètres\n",
        "def majTheta(Theta: np.ndarray, Gradient: np.ndarray, alpha: np.ndarray) -> np.ndarray: \n",
        "    return Theta - alpha * Gradient\n",
        "\n",
        "\n",
        "def normaliser(X: np.ndarray, mean: np.ndarray=None, std: np.ndarray=None) -> np.ndarray: \n",
        "    if (mean is None) or (std is None): \n",
        "        mean = np.mean(X, axis=0)\n",
        "        std  = np.std(X, axis=0)\n",
        "    X_norm = np.where(std==0, X, (X - mean)/std)\n",
        "    return X_norm, mean, std\n",
        "\n",
        "def preparer(X: np.ndarray, norm: np.ndarray=True, \n",
        "             const: np.ndarray=True, mean: np.ndarray=None, std: np.ndarray=None) -> np.ndarray: \n",
        "    X_pre = X.copy()\n",
        "    if norm: \n",
        "        X_pre, mean, std = normaliser(X_pre,mean=mean, std=std)\n",
        "    if const:\n",
        "        X_pre = np.append(np.ones((X_pre.shape[0],1)), X_pre ,axis=1)\n",
        "    return X_pre, mean, std\n",
        "\n",
        "def generer_zeros_1(nbr: int) -> np.ndarray:\n",
        "    return np.zeros(nbr)\n",
        "\n",
        "def generer_uns_1(nbr: int) -> np.ndarray:\n",
        "    return np.ones(nbr)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VmS7J1NnUT_5"
      },
      "source": [
        "### I.1.  Régularisation L2 (Ridge)\n",
        "\n",
        "\n",
        "#### I.2.1. Fonction du coût de la régularisation L2\n",
        "\n",
        "La fonction du coût de la régularisation **L2** est calculée somme : \n",
        "\n",
        "$$J_{L2}(\\theta) = \\frac{\\lambda }{2M} \\sum\\limits_{j=1}^{N} \\theta_j^2$$\n",
        "\n",
        "- **M** : nombre des échantiollons (observations)\n",
        "- **N** : nombre des caractéristiques\n",
        "\n",
        "\n",
        "**Remarque** Dans les énoncés, la régularisation n'est pas divisée par le nombre des échantillons $M$. Mais, en pratique, on la divise. [Voici des réponses sur stackexchange](https://stats.stackexchange.com/questions/287920/regularisation-why-multiply-by-1-2m) expliquant pourquoi."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 34,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hLPuzdVjUT_5",
        "outputId": "023847b7-75d2-4880-bd91-9cae9ce81b19"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.0, 0.0005050000000000001)"
            ]
          },
          "metadata": {},
          "execution_count": 34
        }
      ],
      "source": [
        "# TODO: Coût L2\n",
        "def J_L2(Theta: np.ndarray, M: int, Lambda: float=0.) -> float:\n",
        "    return (Lambda / (2* M) ) * sum (list(map(lambda t : t**2 , Theta[1:])),) \n",
        "\n",
        "#=====================================================================\n",
        "# TEST UNITAIRE\n",
        "#=====================================================================\n",
        "# Resultat : (0.0, 0.0005050000000000001)\n",
        "#---------------------------------------------------------------------\n",
        "\n",
        "Theta_t = np.array([1, 0.01, 0.1]) # Theta avec le premier élément pour le biais\n",
        "\n",
        "\n",
        "J_L2(Theta_t, 4), J_L2(Theta_t, 4, Lambda=.4)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "95nRIJ0qUT_6"
      },
      "source": [
        "#### 1.2.2. Gradients de la régularisation L2\n",
        "\n",
        "La fonction des gradients de la régularisation **L2** est définie comme suit :\n",
        " \n",
        "$$\\frac{\\partial J_{L2}(\\theta)}{\\partial \\theta_j} = \\frac{\\lambda }{M} \\theta_j \\text{ si } j > 0$$\n",
        "\n",
        "$$\\frac{\\partial J_{L2}(\\theta)}{\\partial \\theta_0} = 0$$\n",
        "\n",
        "- **X** : matrice (échantillons X caractéristiques)\n",
        "- **Y** : vecteur des sorties réelles (échantillons)\n",
        "- **H** : vecteur des estimations (échantillons)\n",
        "- **M** : nombre des échantillons\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 35,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Pp0d4i4SUT_7",
        "outputId": "8ca00cb8-85a8-4c48-f41d-76199e664eea"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([0., 0., 0.]), array([0.   , 0.001, 0.01 ]))"
            ]
          },
          "metadata": {},
          "execution_count": 35
        }
      ],
      "source": [
        "# TODO: Gradient L2\n",
        "def dJ_L2(Theta: np.ndarray, M: int, Lambda: float=0.) -> np.ndarray:\n",
        "    dj =np.array(list(map(lambda t: (Lambda/M)* t , Theta)))\n",
        "    dj[0] = 0\n",
        "    return dj\n",
        "\n",
        "#=====================================================================\n",
        "# TEST UNITAIRE\n",
        "#=====================================================================\n",
        "# Resultat : (array([0., 0., 0.]), array([0.   , 0.001, 0.01 ]))\n",
        "#---------------------------------------------------------------------\n",
        "\n",
        "Theta_t = np.array([1, 0.01, 0.1]) # Theta avec le premier élément pour le biais\n",
        "\n",
        "dJ_L2(Theta_t, 4), dJ_L2(Theta_t, 4, Lambda=0.4)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "o2vIr6PqUT_7"
      },
      "source": [
        "### I.2.  Régularisation L1 (Lasso)\n",
        "\n",
        "#### I.2.1. Fonction du coût de la régularisation L1\n",
        "\n",
        "La fonction du coût de la régularisation **L1** sur cette fonction, on aura : \n",
        "\n",
        "$$J_{L1}(\\theta) = \\frac{\\lambda}{M} \\sum\\limits_{j=1}^{N} |\\theta_j|$$\n",
        "\n",
        "- **M** : nombre des échantiollons (observations)\n",
        "- **N** : nombre des caractéristiques"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 36,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "363WnkSuUT_7",
        "outputId": "8c08620c-546a-4d59-ac2f-535aab316132"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.0, 0.011000000000000001)"
            ]
          },
          "metadata": {},
          "execution_count": 36
        }
      ],
      "source": [
        "# TODO: Coût L1\n",
        "def J_L1(Theta: np.ndarray, M: int, Lambda: float=0.) -> float:\n",
        "    return (Lambda / (M) ) * sum (list(map(lambda t : abs(t) , Theta[1:])),)\n",
        "\n",
        "#=====================================================================\n",
        "# TEST UNITAIRE\n",
        "#=====================================================================\n",
        "# Resultat : (0.0, 0.011000000000000001)\n",
        "#---------------------------------------------------------------------\n",
        "\n",
        "Theta_t = np.array([1, 0.01, 0.1]) # Theta avec le premier élément pour le biais\n",
        "\n",
        "J_L1(Theta_t, 4), J_L1(Theta_t, 4, Lambda=.4)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Iin3LiG_UT_8"
      },
      "source": [
        "#### I.2.2. Mise à jour des paramètres \n",
        "\n",
        "La dérivée de $L1$ est :\n",
        "$$\\frac{\\partial L1}{\\partial \\theta_j} = \\frac{\\partial \\frac{\\lambda }{M} \\sum\\limits_{j=1}^{N} |\\theta_j|}{\\partial \\theta_j} = \\frac{\\lambda }{M} \\frac{\\theta_j}{|\\theta_j|}$$\n",
        "\n",
        "Il est clair que le gradient n'est pas définit dans le cas $\\theta_j = 0$. \n",
        "En appliquant [le sous-différentiel](https://fr.wikipedia.org/wiki/Sous-diff%C3%A9rentiel), nous aurons : \n",
        "$$\n",
        "\\frac{\\partial L1}{\\partial \\theta_j} = \n",
        "\\begin{cases}\n",
        "-\\frac{\\lambda}{M} & \\text{si } \\theta_j < 0 \\\\\n",
        "0 & \\text{si } \\theta_j = 0 \\\\\n",
        "\\frac{\\lambda}{M} & \\text{si } \\theta_j > 0 \\\\\n",
        "\\end{cases}\n",
        "$$\n",
        "\n",
        "Nous ne pouvons pas utiliser la descente du gradient pour résoudre ça. \n",
        "Nous allons utiliser un algorithme de descente du gradient approximative, qui s'appelle **Iterative soft-thresholding algorithm (ISTA)**. \n",
        "$$prox(\\theta) = S_{\\lambda}(\\theta) = \n",
        "\\begin{cases}\n",
        "\\theta_j + \\lambda & \\text{si } \\theta_j < - \\lambda \\\\\n",
        "0 & \\text{si } -\\lambda \\le \\theta_j \\le \\lambda \\\\\n",
        "\\theta_j - \\lambda & \\text{si } \\theta_j > \\lambda \\\\\n",
        "\\end{cases}\n",
        "$$\n",
        "\n",
        "**Vous devez implémenter cette fonction**\n",
        "\n",
        "Les thétas sont mises à jour comme suite : \n",
        "$$ \\theta = S_{\\lambda}(\\theta - \\frac{\\partial J_0}{\\partial \\theta})$$\n",
        "Où, $J_0$ est la fonction du coût sans régularisation\n",
        "\n",
        "Source : [Ryan Tibshirani (2015), Lecture 08: Proximal Gradient Descent](https://www.stat.cmu.edu/~ryantibs/convexopt-S15/scribes/08-prox-grad-scribed.pdf)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 37,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WENGxc_uUT_8",
        "outputId": "1675bf2b-040d-443d-f5f6-cde8d180d6e4"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([ 0.5 , -0.15,  0.  ])"
            ]
          },
          "metadata": {},
          "execution_count": 37
        }
      ],
      "source": [
        "# TODO: Soft-thresholding : prox(theta)\n",
        "def S(Theta: np.ndarray, Lambda: float) -> np.ndarray: \n",
        "    o = Theta\n",
        "    for i in range(1,len(Theta)) :\n",
        "        if (Theta[i]<-Lambda) : \n",
        "            o[i] = Theta[i]+Lambda\n",
        "        else :\n",
        "            if (Theta[i]>Lambda) : \n",
        "                o[i] = Theta[i]-Lambda\n",
        "            else : \n",
        "                o[i] = 0.0\n",
        "    return o\n",
        "\n",
        "#=====================================================================\n",
        "# TEST UNITAIRE\n",
        "#=====================================================================\n",
        "# Resultat : array([ 0.5 , -0.15,  0.  ])\n",
        "#---------------------------------------------------------------------\n",
        "\n",
        "                     \n",
        "Theta_t = np.array([0.5, -0.25, 0.05])\n",
        "\n",
        "S(Theta_t, 0.1)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Zxc7JKERUT_9"
      },
      "source": [
        "### I.3. Régression avec réglularisation\n",
        "\n",
        "**Rien à programmer ici**\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 38,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PYB3pv-uUT_9",
        "outputId": "e58a3f28-dfec-42cc-ed5e-84a3c9727347"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([ 0.94839532, -0.56840255,  1.2813832 ]),\n",
              " array([ 0.46328109, -0.15750488,  0.        ]),\n",
              " array([ 0.93431089, -0.53451115,  1.03711393]))"
            ]
          },
          "metadata": {},
          "execution_count": 38
        }
      ],
      "source": [
        "def descente(X: np.ndarray, Y: np.ndarray, Theta: np.ndarray, approx: Callable=None, it_max: int=100, alpha: float=0.1, Lambda: float=0.): \n",
        "\n",
        "    M = len(Y)\n",
        "    \n",
        "    if approx :\n",
        "        def maj(Theta, J, dJ):\n",
        "            J    += J_L1(Theta, M, Lambda=Lambda)\n",
        "            Theta = majTheta(Theta, dJ, alpha)\n",
        "            Theta = S(Theta, Lambda/M) # Lamdba doit être divisé sur le nombre des échantillons\n",
        "            return J, Theta\n",
        "    else:\n",
        "        def maj(Theta, J, dJ):\n",
        "            J    += J_L2(Theta, M, Lambda=Lambda)\n",
        "            dJ   += dJ_L2(Theta, M, Lambda=Lambda)\n",
        "            Theta = majTheta(Theta, dJ, alpha)\n",
        "            return J, Theta\n",
        "            \n",
        "    couts  = []\n",
        "    Thetas = []\n",
        "    \n",
        "    #pour ne pas modifier le contenu de Theta en entrée\n",
        "    Theta  = Theta.copy()\n",
        "    \n",
        "    # ici, on boucle jusqu'à arriver aux maximum des itérations\n",
        "    for it in range(it_max):\n",
        "        H        = hf(X, Theta)\n",
        "        J        = BCE(Y, H)\n",
        "        dJ       = dBCE(X, Y, H)\n",
        "        J, Theta = maj(Theta, J, dJ)\n",
        "        couts.append(J)\n",
        "        Thetas.append(Theta.copy())\n",
        "    \n",
        "    return Thetas, couts\n",
        "\n",
        "#=====================================================================\n",
        "# TEST UNITAIRE\n",
        "#=====================================================================\n",
        "# Resultat : \n",
        "# (array([ 0.94839532, -0.56840255,  1.2813832 ]),\n",
        "#  array([ 0.46328109, -0.15750488,  0.        ]),\n",
        "#  array([ 0.93431089, -0.53451115,  1.03711393]))\n",
        "#\n",
        "#---------------------------------------------------------------------\n",
        "\n",
        "X_t = np.array([\n",
        "    [1, 1, 0.2],\n",
        "    [1, 2, 0.1],\n",
        "    [1, 3, 0.4],\n",
        "    [1, 4, 0.3]\n",
        "])\n",
        "\n",
        "Theta_t = np.array([1., 1., 1.])\n",
        "Y_t     = np.array([1, 0, 1, 0])\n",
        "\n",
        "# pas de régularisation : Lambda = 0\n",
        "Theta_P, J_P = descente(X_t, Y_t, Theta_t)\n",
        "# régularisation L1 : approximation avec lambda\n",
        "Theta_1, J_1 = descente(X_t, Y_t, Theta_t, approx=S, Lambda=0.1)\n",
        "# régularisation L2 : lambda sans approximation\n",
        "Theta_2, J_2 = descente(X_t, Y_t, Theta_t, Lambda=0.1)\n",
        "\n",
        "Theta_P[-1], Theta_1[-1], Theta_2[-1]\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 39,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WCf4SzBUUT_9",
        "outputId": "43bd647b-354f-4f0a-fc02-780d9b8d7585"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([0.86383643, 0.5273866 , 0.96123235]),\n",
              " array([0.79241836, 0.        , 0.        ]),\n",
              " array([0.8515839 , 0.38351525, 0.7672672 ]),\n",
              " array([0, 0, 1, 1]))"
            ]
          },
          "metadata": {},
          "execution_count": 39
        }
      ],
      "source": [
        "class RegLogReg(object): # Régression logistique régularisée\n",
        "    \n",
        "    # par défaut : pas de régularisation\n",
        "    def __init__(self, nbr_iter: int=100, \n",
        "                 alpha: float=.01, Lambda: float=1.0, norm: bool=True, const: bool=True, reg: str=None): \n",
        "        self.nbr_iter = nbr_iter\n",
        "        self.alpha    = alpha\n",
        "        self.norm     = norm\n",
        "        self.const    = const\n",
        "        self.approx   = None\n",
        "        self.Lambda   = Lambda\n",
        "        if reg == \"l1\":\n",
        "            self.approx = S\n",
        "        elif reg != \"l2\":\n",
        "            self.Lambda = 0.\n",
        "    \n",
        "    def entrainer(self, X: np.ndarray, Y: np.ndarray): \n",
        "        X_pre, self.mean, self.std = preparer(X, norm=self.norm, const=self.const)\n",
        "        Theta = generer_uns_1(X_pre.shape[1])#\n",
        "        self.Thetas, self.couts = descente(X_pre, Y, Theta, \n",
        "                                          approx=self.approx, \n",
        "                                          it_max=self.nbr_iter, \n",
        "                                          alpha =self.alpha, \n",
        "                                          Lambda=self.Lambda)\n",
        "        \n",
        "    # La prédiction\n",
        "    # si classes=True elle rend une vecteur de 1 et 0\n",
        "    # sinon elle rend un vecteur de probabilités\n",
        "    def predire(self, X: np.ndarray, classes: bool=True, seuil: float=0.5, it: int=-1) -> np.ndarray:\n",
        "        X_pre, self.mean, self.std = preparer(X, norm=self.norm, const=self.const, mean=self.mean, std=self.std)\n",
        "        H = hf(X_pre, self.Thetas[it])\n",
        "        if classes:\n",
        "            return (H >= seuil).astype(int)\n",
        "        return H\n",
        "\n",
        "#=====================================================================\n",
        "# TEST UNITAIRE\n",
        "#=====================================================================\n",
        "# Resultat : \n",
        "# (array([0.86383643, 0.5273866 , 0.96123235]),\n",
        "#  array([0.79153709, 0.        , 0.        ]),\n",
        "#  array([0.8515839 , 0.38351525, 0.7672672 ]),\n",
        "#  array([0, 0, 1, 1]))\n",
        "#---------------------------------------------------------------------\n",
        "\n",
        "X_t = np.array([\n",
        "    [100, 20],\n",
        "    [200, 10],\n",
        "    [300, 40],\n",
        "    [400, 30]\n",
        "])\n",
        "\n",
        "Y_t = np.array([1, 0, 1, 0])\n",
        "\n",
        "# pas de régularisation\n",
        "reglog = RegLogReg()\n",
        "reglog.entrainer(X_t, Y_t)\n",
        "# régularisation L1\n",
        "reglog1 = RegLogReg(reg=\"l1\")\n",
        "reglog1.entrainer(X_t, Y_t)\n",
        "# régularisation L2\n",
        "reglog2 = RegLogReg(reg=\"l2\")\n",
        "reglog2.entrainer(X_t, Y_t)\n",
        "\n",
        "reglog.Thetas[-1], reglog1.Thetas[-1], reglog2.Thetas[-1], reglog.predire(X_t)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jfv0gDHXUT_-"
      },
      "source": [
        "### I.4. Filtrage des attributs (ANOVA : Analysis of Variance)\n",
        "\n",
        "Le nombre des échantillons pour chaque classe est différent. \n",
        "Pour cette raison, nous allons implémenter ANOVA for a CRD (COMPLETELY RANDOM DESIGN). \n",
        "L'algorithme est expliqué dans [Richard Horsly (2012) COMPLETELY RANDOM DESIGN (CRD)](https://www.ndsu.edu/faculty/horsley/CRD.pdf). \n",
        "Pour valider l'implémentation, nous allons utiliser les mêmes données de la page 10.\n",
        "\n",
        "#### I.4.1. One way F value pour un seul attribue\n",
        "\n",
        "Pour un attribut donné ($A$), l'ensemble des valeurs des échantillons sera divisé sur $N$ ensembles, où $N$ est le nombre des classes. \n",
        "\n",
        "$$CF = \\frac{(\\sum_{ij} A_{ij})^2}{\\sum_j |A_j|} $$\n",
        "Où $\\sum_{ij} A_{ij}$ est la somme de toutes les valeurs de toutes les classes, \n",
        "$\\sum_j |A_j|$ est le nombre total des échantillons (de toutes les classes)\n",
        "\n",
        "$$ TotalSS = \\sum_{ij} A_{ij}^2 - CF$$\n",
        "Où $\\sum_{ij} A_{ij}^2$ est les somme des carrées de toutes les valeurs\n",
        "\n",
        "$$ TreatmentSS = \\sum_j \\frac{(\\sum_{i} A_{ij})^2}{|A_j|} - CF$$\n",
        "Où $\\sum_{i} A_{ij}$ est la somme des valeurs de la classe $j$ et $|A_j|$ est le nombre des éléments dans la classe $j$\n",
        "\n",
        "$$ ErrorSS = TotalSS - TreatmentSS $$\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JI7pukIOUT_-",
        "outputId": "01e222af-a100-4fc8-de5a-fba61f9c3fe4"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "6.418894830659841"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ],
      "source": [
        "# TODO Compléter la fonction one-way ANOVA pour un attribut\n",
        "def f_oneway(A): \n",
        "    # A est une liste de N listes où N est le nombre des classes\n",
        "    # Ces listes n'ont pas la même taille\n",
        "    nbrs = np.array([len(a) for a in A]) # nombre des échantillons de chaque classe\n",
        "    sommes = np.array([np.sum(a) for a in A]) # somme des valeurs de chaque classe\n",
        "    sommes_caree = np.sum([np.sum(np.array(a)**2) for a in A]) # somme des valeurs carrées de toutes les valeurs\n",
        "    \n",
        "    # Compléter ici\n",
        "    #====================================\n",
        "    CF = np.sum(sommes) ** 2 / np.sum(nbrs)\n",
        "    TotalSS = sommes_caree - CF\n",
        "    TreatmentSS = np.sum((sommes**2)/nbrs) - CF\n",
        "    ErrorSS = TotalSS - TreatmentSS\n",
        "    #====================================\n",
        "    \n",
        "    MSB = TreatmentSS / (len(A) - 1) \n",
        "    MSW =  ErrorSS/(nbrs.sum() - len(A))\n",
        "    return MSB/MSW\n",
        "\n",
        "\n",
        "#=====================================================================\n",
        "# TEST UNITAIRE\n",
        "#=====================================================================\n",
        "# Resultat : 6.418894830659841\n",
        "#---------------------------------------------------------------------\n",
        "\n",
        "AA = [[2., 2.2, 1.8, 2.3, 1.7], [1.7, 1.9, 1.5], [2., 2.4, 2.7, 2.5, 2.4], [2.1, 2.2, 2.2, 1.9]]\n",
        "# Résultat : 6.418894830659841 (voir page 11 step 6)\n",
        "f_oneway(AA)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "tags": [],
        "id": "X-CBoZhkUT__"
      },
      "source": [
        "#### I.4.2. One way F value pour tous les attribues\n",
        "\n",
        "**Rien à programmer ou analyser ici**\n",
        "\n",
        "Ici, nous allons calculer ANOVA pour chaque attribut. \n",
        "Les attributs avec des grandes valeurs ANOVA sont meilleurs. "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "093ZtbeCUT__",
        "outputId": "ce5addbf-0091-4db7-b911-1959c4c86c56"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([  6.41889483,   0.40595497, 105.2745098 ]),\n",
              " array([  6.41889483,   0.40595497, 105.2745098 ]))"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ],
      "source": [
        "def anova_f(X: np.ndarray, Y: np.ndarray) -> np.ndarray: \n",
        "    \n",
        "    C = np.unique(Y)\n",
        "    F = []\n",
        "    for j in range(X.shape[1]): \n",
        "        A = X[:, j]\n",
        "        F.append(f_oneway([A[Y==c] for c in np.unique(Y)]))\n",
        "    return np.array(F)\n",
        "\n",
        "\n",
        "#=====================================================================\n",
        "# TEST UNITAIRE\n",
        "#=====================================================================\n",
        "# Resultat : \n",
        "# (array([  6.41889483,   0.40595497, 105.2745098 ]),\n",
        "#  array([  6.41889483,   0.40595497, 105.2745098 ]))\n",
        "#---------------------------------------------------------------------\n",
        "\n",
        "X_t = np.array(\n",
        "    [[2. , 1. , 1. ],\n",
        "     [2.2, 2. , 1. ],\n",
        "     [1.8, 3. , 1. ],\n",
        "     [2.3, 4. , 1. ],\n",
        "     [1.7, 5. , 2. ],\n",
        "     [1.7, 1. , 2. ],\n",
        "     [1.9, 2. , 2. ],\n",
        "     [1.5, 3. , 2. ],\n",
        "     [2. , 1. , 3. ],\n",
        "     [2.4, 2. , 3. ],\n",
        "     [2.7, 3. , 3. ],\n",
        "     [2.5, 4. , 3. ],\n",
        "     [2.4, 5. , 3. ],\n",
        "     [2.1, 1. , 4. ],\n",
        "     [2.2, 2. , 4. ],\n",
        "     [2.2, 3. , 4. ],\n",
        "     [1.9, 4. , 4. ]])\n",
        "\n",
        "Y_t = np.array([\"A\", \"A\", \"A\", \"A\", \"A\", \"B\", \"B\", \"B\", \"C\", \"C\", \"C\", \"C\", \"C\", \"D\", \"D\", \"D\", \"D\"])\n",
        "\n",
        "# pour valider notre implémentation, nous allons utiliser f_classif de scikit-learn \n",
        "# https://github.com/scikit-learn/scikit-learn/blob/95d4f0841d57e8b5f6b2a570312e9d832e69debc/sklearn/feature_selection/_univariate_selection.py#L42\n",
        "from sklearn.feature_selection import f_classif\n",
        "f_sklearn, a = f_classif(X_t, Y_t)\n",
        "f_notre      = anova_f  (X_t, Y_t)\n",
        "\n",
        "# la première liste est celle de scikit-learn et la deuxième est la notre\n",
        "f_sklearn, f_notre"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NhzJo7vrUUAA"
      },
      "source": [
        "## II. Application et analyse\n",
        "\n",
        "Nous allons utiliser le dataset [Diabetics prediction using logistic regression](https://www.kaggle.com/kandij/diabetes-dataset). "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "jEbc6kliUUAA",
        "outputId": "23bd496a-acd4-4a84-9f81-903a9b768ebd"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   Pregnancies  Glucose  BloodPressure  SkinThickness  Insulin   BMI  \\\n",
              "0            6      148             72             35        0  33.6   \n",
              "1            1       85             66             29        0  26.6   \n",
              "2            8      183             64              0        0  23.3   \n",
              "3            1       89             66             23       94  28.1   \n",
              "4            0      137             40             35      168  43.1   \n",
              "\n",
              "   DiabetesPedigreeFunction  Age  Outcome  \n",
              "0                     0.627   50        1  \n",
              "1                     0.351   31        0  \n",
              "2                     0.672   32        1  \n",
              "3                     0.167   21        0  \n",
              "4                     2.288   33        1  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-d5af2989-5a1c-450f-a26e-96329577e95a\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Pregnancies</th>\n",
              "      <th>Glucose</th>\n",
              "      <th>BloodPressure</th>\n",
              "      <th>SkinThickness</th>\n",
              "      <th>Insulin</th>\n",
              "      <th>BMI</th>\n",
              "      <th>DiabetesPedigreeFunction</th>\n",
              "      <th>Age</th>\n",
              "      <th>Outcome</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>6</td>\n",
              "      <td>148</td>\n",
              "      <td>72</td>\n",
              "      <td>35</td>\n",
              "      <td>0</td>\n",
              "      <td>33.6</td>\n",
              "      <td>0.627</td>\n",
              "      <td>50</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>85</td>\n",
              "      <td>66</td>\n",
              "      <td>29</td>\n",
              "      <td>0</td>\n",
              "      <td>26.6</td>\n",
              "      <td>0.351</td>\n",
              "      <td>31</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>8</td>\n",
              "      <td>183</td>\n",
              "      <td>64</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>23.3</td>\n",
              "      <td>0.672</td>\n",
              "      <td>32</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1</td>\n",
              "      <td>89</td>\n",
              "      <td>66</td>\n",
              "      <td>23</td>\n",
              "      <td>94</td>\n",
              "      <td>28.1</td>\n",
              "      <td>0.167</td>\n",
              "      <td>21</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0</td>\n",
              "      <td>137</td>\n",
              "      <td>40</td>\n",
              "      <td>35</td>\n",
              "      <td>168</td>\n",
              "      <td>43.1</td>\n",
              "      <td>2.288</td>\n",
              "      <td>33</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-d5af2989-5a1c-450f-a26e-96329577e95a')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-d5af2989-5a1c-450f-a26e-96329577e95a button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-d5af2989-5a1c-450f-a26e-96329577e95a');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ],
      "source": [
        "diabetes   = pd.read_csv(\"/content/diabetes2.csv\") \n",
        "X_diabetes = diabetes.iloc[:, :-1].values  \n",
        "Y_diabetes = diabetes.iloc[:,  -1].values\n",
        "\n",
        "# Cette configuration est mise en place comme ceci exprès\n",
        "# C'est pour tester le cas où la régression est difavorisée\n",
        "NBR_TEST = 240\n",
        "# Supposant que les 30% premières lignes sont pour le test et le reste pour l'entraînement\n",
        "X_test  = X_diabetes[-NBR_TEST:, :] # 30% ou plus\n",
        "Y_test  = Y_diabetes[-NBR_TEST:   ]\n",
        "\n",
        "X_train = X_diabetes[:-NBR_TEST, :] \n",
        "Y_train = Y_diabetes[:-NBR_TEST   ]\n",
        "\n",
        "diabetes.head()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CWCFoY9IUUAA"
      },
      "source": [
        "### II.1. Régularisation\n",
        "\n",
        "Nous avons entrainé 3 modèles :\n",
        "- **reglog** : régression logistique sans régularisation\n",
        "- **reglog_l1** : régression logistique avec régularisation L1\n",
        "- **reglog_l2** : régression logistique avec régularisation L2\n",
        "\n",
        "Nous avons testé ces modèles en utilisant deux datasets (avec la métrique F1 dans chaque itération) :\n",
        "- **entrainement** : pour tester la convergence (fitting)\n",
        "- **test** : pour tester la généralisation\n",
        "\n",
        "Afin de capturer les F1 scores et les paramètres, nous avons entrainé chaque modèle avec une seule itération et utiliser une boucle globale pour contrôler les traitements de chaque itération. Pour continuer l'entrainement à partir des thétas précédents, nous avons activé l'option **warm_start**. Ceci va générer des graphiques non lisses (puisque l'algorithme d'optimisation est lancé à chaque itération).\n",
        "\n",
        "#### II.1.1.  Convergence et généralisation sans et avec régularisation par itérations\n",
        "\n",
        "Ici, nous supposons que l'algorithme a convergé lorsqu'il atteint le maximum de F1 score dans l'antrainement (normalement, il doit sortir s'il y a une chutte de F1 score et retourne les thétats les plus optimaux). Donc, au delà du point de la convergence, nous avons continuer l'itération pour voir ce qui va se passer. \n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 663
        },
        "id": "5BTe6-4rUUAB",
        "outputId": "64ba9c46-c6fe-4452-f8c0-6a6c148e25ed"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1440x720 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "import warnings\n",
        "from sklearn.exceptions   import ConvergenceWarning\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "from sklearn.metrics      import f1_score\n",
        "warnings.filterwarnings(\"ignore\", category=ConvergenceWarning)\n",
        "\n",
        "MAX_IT     = 500\n",
        "CC         = 0.002\n",
        "COEF_      = np.array([[1.] * X_train.shape[1]])\n",
        "INTERCEPT_ = np.array([1.])\n",
        "\n",
        "reglog = LogisticRegression(penalty='none', solver=\"saga\", max_iter=1, warm_start=True, random_state=0)\n",
        "reglog.coef_      = COEF_.copy()\n",
        "reglog.intercept_ = INTERCEPT_.copy()\n",
        "\n",
        "reglog_l1 = LogisticRegression(penalty='l1', solver=\"saga\", max_iter=1, warm_start=True, random_state=0, C=CC)\n",
        "reglog_l1.coef_      = COEF_.copy()\n",
        "reglog_l1.intercept_ = INTERCEPT_.copy()\n",
        "\n",
        "reglog_l2 = LogisticRegression(penalty='l2', solver=\"saga\", max_iter=1, warm_start=True, random_state=0, C=CC)\n",
        "reglog_l2.coef_      = COEF_.copy()\n",
        "reglog_l2.intercept_ = INTERCEPT_.copy()\n",
        "\n",
        "coefs         = []\n",
        "coefs0        = []\n",
        "f1train       = []\n",
        "f1test        = []\n",
        "\n",
        "coefs_ridge   = []\n",
        "coefs0_ridge  = []\n",
        "f1train_ridge = []\n",
        "f1test_ridge  = []\n",
        "\n",
        "coefs_lasso   = []\n",
        "coefs0_lasso  = []\n",
        "f1train_lasso = []\n",
        "f1test_lasso  = []\n",
        "\n",
        "for it in range(MAX_IT):\n",
        "    reglog.fit(X_train, Y_train)\n",
        "    coefs.append(reglog.coef_.flatten().tolist())\n",
        "    coefs0.append(reglog.intercept_[0])\n",
        "    f1train.append(f1_score(Y_train, reglog.predict(X_train)))\n",
        "    f1test.append(f1_score(Y_test, reglog.predict(X_test)))\n",
        "    \n",
        "    reglog_l1.fit(X_train, Y_train)\n",
        "    coefs_lasso.append(reglog_l1.coef_.flatten().tolist())\n",
        "    coefs0_lasso.append(reglog_l1.intercept_[0])\n",
        "    f1train_lasso.append(f1_score(Y_train, reglog_l1.predict(X_train)))\n",
        "    f1test_lasso.append(f1_score(Y_test, reglog_l1.predict(X_test)))\n",
        "    \n",
        "    reglog_l2.fit(X_train, Y_train)\n",
        "    coefs_ridge.append(reglog_l2.coef_.flatten().tolist())\n",
        "    coefs0_ridge.append(reglog_l2.intercept_[0])\n",
        "    f1train_ridge.append(f1_score(Y_train, reglog_l2.predict(X_train)))\n",
        "    f1test_ridge.append(f1_score(Y_test, reglog_l2.predict(X_test)))\n",
        "\n",
        "\n",
        "coefs       = np.array(coefs)\n",
        "coefs_lasso = np.array(coefs_lasso)\n",
        "coefs_ridge = np.array(coefs_ridge)\n",
        "\n",
        "fig, (ax1, ax2) = plt.subplots(1,2)\n",
        "fig.suptitle(\"F1 score par nombre des iterations\")\n",
        "fig.set_figwidth(20)\n",
        "fig.set_figheight(10)\n",
        "\n",
        "ax1.plot(range(MAX_IT), f1train, color=\"blue\", label=\"Reg\")\n",
        "ax1.plot(range(MAX_IT), f1train_lasso, color=\"red\", label=\"Reg-L1\")\n",
        "ax1.plot(range(MAX_IT), f1train_ridge, color=\"green\", label=\"Reg-L2\")\n",
        "ax1.legend()\n",
        "ax1.set_title(\"Entrainement\")\n",
        "ax1.grid()\n",
        "\n",
        "ax2.plot(range(MAX_IT), f1test, color=\"blue\", label=\"Reg\")\n",
        "ax2.plot(range(MAX_IT), f1test_lasso, color=\"red\", label=\"Reg-L1\")\n",
        "ax2.plot(range(MAX_IT), f1test_ridge, color=\"green\", label=\"Reg-L2\")\n",
        "ax2.legend()\n",
        "ax2.set_title(\"Test\")\n",
        "ax2.grid()\n",
        "\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4qTzdhjnUUAB"
      },
      "source": [
        "**TODO: Analyser les résultats**\n",
        "\n",
        "- Que remarquez-vous ?\n",
        "- Analyser les cas de convergence (qui converge plus rappidement, avec quel prix (quel est le gain ou la perte)) en précisant dans ce cas l'intérêt de la régularisation. \n",
        "- Pourquoi les modèles avec régularisation perdent-ils de performance avec plus d'térations ?\n",
        "\n",
        "**Réponse**\n",
        "\n",
        "- On remarque que les modeles avec régularisation perdent de la performance avec plus d'iteration, et le modele sans regularisation donne un meilleur F1 score. On remarque aussi que L1 subit une chute importante avec plus d'iteration.\n",
        "- On remarque que L1 converge plus rapidement mais on va perdre en terme de performance. La regression sans regularisation donne un meilleur F1 score mais on perd en terme de temps d'entrainement. \n",
        "L'interet ici c'est d'éviter le sur-apprentissage et de réduire le temps d'entrainement. \n",
        "- la régularisation a pour effet de réduire l'amplitude des coefficients du modèle. Avec les iterations la force de régularisation devient élevée, certains des coefficients du modèle peuvent devenir nuls, ce qui peut entraîner un sous-apprentissage et donc un modèle incapable de capturer la complexité des données."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OYImQ5hCUUAB"
      },
      "source": [
        "#### II.1.2. Convergence des paramètres par itérations\n",
        "\n",
        "Ici, nous avons afficher les valeurs des paramètres thétas pour chaque itération. \n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "scrolled": true,
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "TD0oncuGUUAC",
        "outputId": "2d312c07-a7e7-448e-bd14-71d10729b524"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x1440 with 3 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlsAAATuCAYAAAD6EuaNAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeXxU1eH//9eZmWSybyTsS0BZJew7oihVUKl8pHWpttbWWj/WpauWLp9q/dj+tNJ+6opbLVJrW2uFH8WNKkYEFzYRgSBhCQRElpCEhGSSTOZ8/7g3k0lI2MyQQN7Ph/OYueeee86ZmZi8OXcz1lpEREREJDo8rT0AERERkTOZwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYgAxphcY8x3TkE/1hhzdrT7EZG2Q2FL5DRjjLnWGLPKGFNujNljjHnNGHNua49LjtTSAU5BTeT0pLAlchoxxvwI+CPwW6AT0BN4HJjRmuOKZIzxtfYYRETaEoUtkdOEMSYVuBe41Vr7srX2sLW2xlr7b2vtnW4dvzHmj8aYz9zHH40xfnfdZGPMLmPMj40x+9xZsW+568YaYz43xngj+rvCGLPOfe0xxswyxmw1xhQZY140xmS467LdGZcbjTE7gSXGGK8x5vfGmAPGmO3GmNvcOr6692KM+ZM7ht3GmPvq+jbG3GCMWWaMmW2MKXa3vyRiXBnGmD+776/YGLMgYt10Y8xaY0yJMeY9Y8yQo3yeFxljNhljSo0xjwKm0fpvG2Py3D7eMMb0csuNMeb/3M/wkDHmE2PM4Cba/w0wCXjUnYV8NGL1l4wx+e44HzPGmIjtmut3qVvlY7e9q40x6caYRcaY/W79RcaY7hFt3WCM2WaMKXM/x+ua+zxEJIqstXroocdp8ACmAUHAd5Q69wIfAB2BLOA94H/ddZPd7e8FYoBLgQog3V2/Fbgooq1/ArPc19932+0O+IEngb+567IBC8wDEoF44L+BjW79dOBNt47P3Wa+20aiO9YVwM3uuhuAGuAmwAvcAnwGGHf9K8A/3HZjgPPd8uHAPmCsu903gQLA38TnlAmUAV912/ih+9l8x10/A9gCDAR8wC+B99x1U4HVQBpOQBsIdGnm+8itazOizAKL3O17AvuBacfqN2LbsyOWOwBfARKAZPc7W+CuSwQOAf3d5S7AOa39c6yHHu3x0eoD0EMPPY7vAVwHfH6MOluBSyOWpwIF7uvJQCURYc0NJ+Pc1/cBz7qvk4HDQC93OQ+YErFdFzcQ+agPW30i1i+pC0/u8pfcOj6c3Z9VQHzE+q8Bb7uvbwC2RKxLcLft7PYbwg2Ijd77HNxgGVH2KW4Ya1R+PfBBxLIBdkWErdeAGyPWe3CCaS/gQmAzMA7wHOP7aC5snRux/CL1obbZfiO2Pfso/Q0Dit3XiUCJG8bijzZOPfTQI7oP7UYUOX0UAZnHOCaqK7AjYnmHWxZuw1objFiuAJLc1y8AM93djjOBNdbaurZ6AfPd3V4lOOGrFic41SlsNI7CZtb1wplN2hPR3pM4M1x1Pq97Ya2tcF8mAT2Ag9ba4ibeey/gx3Vtuu32aPT+mxyftdY2McaHIto5iBPIullrlwCPAo8B+4wxTxljUpro42g+j3gd+R00229TjRhjEowxTxpjdhhjDgFLgTRjjNdaexi4GmeWcY8x5hVjzIATHKeItACFLZHTx/s4M0L/dZQ6n+H8wa7T0y07JmvtRpxwdglwLU74qlMIXGKtTYt4xFlrd0c2EfF6D84uxDo9GrVVBWRGtJVirT3nOIZZCGQYY9KaWfebRmNMsNb+rYm6eyLH5B4z1XiMNzdqK95a+x6AtfZha+1IYBDQD7izmfHaZsqP9v6a7bcJPwb6A2OttSnAeXVvyR3nG9bai3BmBDcBT5/geESkBShsiZwmrLWlwK+Ax4wx/+XOasQYYy4xxvzOrfY34JfGmCxjTKZb//kT6OYFnOOzzsM5/qfOE8BvIg7WzjLGHO0MyBeB7xtjurnB6KcR72MPsBj4vTEmxT34/ixjzPnHGpy77WvA4+7B4THGmLqA8TTw38Y52N8YYxKNMZcZY5KbaOoV4BxjzEx3pvAOnN2Uke/3Z8aYc9z3m2qMudJ9PdrtIwZnV2sAZ9dmU/YCfY71vo6n32baS8bZNVxinBMW7q5bYYzpZIyZYYxJxAm35UcZp4hEkcKWyGnEWvt74Ec4B07vx5kJuQ2oOyPvPmAVsA74BFjjlh2vvwHnA0ustQciyh8CFgKLjTFlOAfLjz1KO0/jBKp1wEfAqzgHoNe6668HYnEOoi8GXsKZfTke38A5XmwTzjFnPwCw1q7COaj+UbfNLTjHfx3BfW9XAvfj7J7tCyyPWD8feAD4u7t7bj3OjB9Aivv+inFmAouAB5sZ60PAV90zBR8+1hs7Rr8A9wDPubsZr8K5DEg8cADnO3k9oq4H52flM5zdkefjnGwgIqdY3dk9IiJRY5xLNzxhre11zMoiImcYzWyJSIszxsQbYy41xviMMd1wdm/Nb+1xiYi0Bs1siUiLM8YkAO8AA3COKXoF+L619lCrDkxEpBUobImIiIhEkXYjioiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYiIiESRwpaIiIhIFClsiYgcJ2PMJGPMp609DhE5vRhrbWuPQUTaGWNMAdAJqAXKgdeB26y15a05LhGRaNDMloi0li9ba5OAYcBw4Gct2bgxxteS7YmInCyFLRFpVdbaz4E3cEIXxphxxpj3jDElxpiPjTGT6+oaY3obY5YaY8qMMW8aYx4zxjzvrss2xlhjzI3GmJ3AErf828aYPGNMsTHmDWNML7fcGGP+zxizzxhzyBjziTFmsLvuUmPMRref3caYn7jlk40xuyLGM9AYk+uOdYMx5vKIdXPd8b3itvOhMeasaH+eItL2KGyJSKsyxnQHLgG2GGO6Aa8A9wEZwE+AfxljstzqLwArgA7APcA3mmjyfGAgMNUYMwP4OTATyALeBf7m1rsYOA/oB6QCVwFF7ro/ATdba5OBwbjBrdG4Y4B/A4uBjsDtwF+NMf0jql0D/BpIB7YAvznOj0VEziAKWyLSWhYYY8qAQmAfcDfwdeBVa+2r1tqQtfY/wCrgUmNMT2A08CtrbbW1dhmwsIl277HWHrbWVgL/Dfx/1to8a20Q+C0wzJ3dqgGSgQE4x6/mWWv3uG3UAIOMMSnW2mJr7Zom+hkHJAH3u+NZAiwCvhZRZ761doXb919xZ+9EpH1R2BKR1vJf7szRZJzAkwn0Aq50d8uVGGNKgHOBLkBX4KC1tiKijcIm2o0s6wU8FNHWQcAA3dxw9CjwGLDPGPOUMSbF3e4rwKXADmPMO8aY8U300xUotNaGIsp2AN0ilj+PeF2BE85EpJ1R2BKRVmWtfQeYC8zGCUp/sdamRTwSrbX3A3uADGNMQsTmPZpqMuJ1Ic7uwMj24q2177l9P2ytHQkMwtmdeKdbvtJaOwNn9+AC4MUm+vkM6GGMifw92hPYfcIfgoic0RS2RKQt+CNwEfAe8GVjzFRjjNcYE+celN7dWrsDZ5fiPcaYWHe26cvHaPcJ4GfGmHMAjDGpxpgr3dejjTFj3WOvDgMBIOS2fZ0xJtVaWwMcAkJNtP0hzmzVXcaYGPdA/i8Df/+Cn4WInGEUtkSk1Vlr9wPzgDuAuoPa9+PMTN1J/e+q64DxOAey3wf8A6g6SrvzgQeAvxtjDgHrcQ7GB0gBngaKcXb/FQEPuuu+ARS42/y322/jtqtxwtUlwAHgceB6a+2mE/4AROSMpouaishpyxjzD2CTtfbu1h6LiEhzNLMlIqcNd9ffWcYYjzFmGs4s2ILWHpeIyNHoCssicjrpDLyMc52tXcAt1tqPWndIIiJHp92IIiIiIlGk3YgiIiIiUaSwJSIiIhJFbfaYrczMTJudnR31fg4fPkxiYmLU+5Hjp++kbdL30jbpe2l79J20TdH+XlavXn3AWpvV1Lo2G7ays7NZtWpV1PvJzc1l8uTJUe9Hjp++k7ZJ30vbpO+l7dF30jZF+3sxxuxobp12I4qIiIhEkcKWiIiISBQpbImIiIhEUZs9ZktERKSmpoZdu3YRCARaeyjHLTU1lby8vNYehjTSUt9LXFwc3bt3JyYm5ri3UdgSEZE2a9euXSQnJ5OdnY0xprWHc1zKyspITk5u7WFIIy3xvVhrKSoqYteuXfTu3fu4t9NuRBERabMCgQAdOnQ4bYKWnNmMMXTo0OGEZ1oVtkREpE1T0JK25GR+HhW2REREjsLr9TJs2DAGDx7MlVdeSUVFRWsP6bgsXLiQ+++/v7WHIShsiYiIHFV8fDxr165l/fr1xMbG8sQTTzRYHwwGW2lkR3f55Zcza9as1h6G0EJhyxjzrDFmnzFmfTPrjTHmYWPMFmPMOmPMiJboV0RE5FSaNGkSW7ZsITc3l0mTJnH55ZczaNAgamtrufPOOxk9ejTjx4/nySefBCAUCvG9732PAQMGcNFFF3HppZfy0ksvAc6dUu6++25GjBhBTk4OmzZtAmDFihWMHz+e4cOHM2HCBD799FMA5s6dy8yZM5k2bRp9+/blrrvuCo/r9ddfZ8SIEQwdOpQpU6aE6992220A7N+/n6985SuMHj2a0aNHs3z5cgDeeecdhg0bxrBhwxg+fDhlZWWn5oNsZ1rqbMS5wKPAvGbWXwL0dR9jgTnus4iIyHH59b83sPGzQy3a5qCuKdz95XOOq24wGOS1115j2rRpAKxZs4b169fTu3dvnnrqKVJTU1m5ciUHDhxg2rRpXHzxxaxevZqCggI2btzIvn37GDhwIN/+9rfDbWZmZrJmzRoef/xxZs+ezTPPPMOAAQN499138fl8vPnmm/z85z/nX//6FwBr167lo48+wu/3079/f26//Xbi4uK46aabWLp0Kb179+bgwYNHjP373/8+P/zhDzn33HPZuXMnU6dOJS8vj9mzZ/PYY48xceJEysvLiYuLa4FPVRprkbBlrV1qjMk+SpUZwDxrrQU+MMakGWO6WGv3tET/J2vTK78j6aCf0uXdiU1Px5vmx5fmx8T7dECmiIgAUFlZybBhwwBnZuvGG2/kvffeY8yYMeHT/xcvXsy6det46aWXCIVClJWVkZ+fz7Jly7jyyivxeDx07tyZCy64oEHbM2fOBGDkyJG8/PLLAJSWlvLNb36T/Px8jDHU1NSE60+ZMoXU1FQABg0axI4dOyguLua8884LjyUjI+OI9/Dmm2+ycePG8PKhQ4coLy9n4sSJ/OhHP+K6665j5syZdO/evaU+Nolwqq6z1Q0ojFje5ZY1CFvGmO8C3wXo1KkTubm5URtQVbAan+8v0K2CVZWPEb+nL4kHBpNYlENMRXeC8R6C8RCMs9TGAG72ssZ5Xfd8RBmNyjlynTW2yfKjbRPZT5OvT7C9o20Tfm4l5eXlUf3u5eToe2mbzvTvJTU1Nbxr60eTe0alj2PtOouPj+fdd98NL1dVVVFRUYHf7w9vW1NTwwMPPMCXvvQlamtr8Xq9ACxYsIBAIBCuFwwGqayspKysDGstNTU1lJWVEQgEqKqqoqysjFmzZjF+/HjmzZvHjh07uOyyy8J1jDHhtqy1HDp0iMrKynA7kQKBANXV1ZSVlVFbW8t//vOfBjNX1lpuvfVWJk+ezOLFi5kwYQLz58+nX79+X/xDbYNqa2tbbDdpIBA4of/v2tRFTa21TwFPAYwaNcpG8+7cVcFafvLiM2zZtZxv5HxG787rOZDxEgd4CV8ojeTKYSQezCFhb0+8h1Lx1MaCjdpw2h4DeAwY9zRXA7jPxtNwGWMwdctHbFO33jRos8E2kcseQ3Gxh/QOGc2ub277psbWYOyextu47Rxtm4gAanD7r/t8qF9vGtUNr6vrI1x+ZLtQN66GbUb2Zeq2q9umuf4j+6irErndEf2bcNXwe46oG/m+PvjwQ8YNGdfws2rq/Ub032CCuPE4Grw2DcvD1Y+xTTPr2tPMdG5uLtH8Xdna8vLy2sQFQhuPISEhAZ/PFy6/7LLLeO6555g+fTqBQIA9e/bQrVs3LrjgAp577jluvvlm9u/fz7Jly7j++utJTk7GGENSUhLJyckkJibi9XpJTk6moqKCs846i+TkZF566SWMMSQnJxMXF0dsbGy4T5/PR0JCAqNHj+bHP/4xBw4cCO9GzMjIaFB/6tSpzJ07lzvvvBNwdkcOGzaMrVu3Mm7cOMaNG8e6desoLCxk5MiRp/bDPUVa8mKzcXFxDB8+/Ljrn6qwtRvoEbHc3S1rNX6fl4euGcNNT5Tw83f6M3P417n3y50pK11O0cF3OXhwOcWJueFRezx+PJ449znWeW38eD1+PMbvlBt3nYnDY2KdMvx4TCzGxDrPOM8eE4OxMc42xOAxsUAMHmLx2Fg8+DD4McTgIQaDF2sthKwT+qzzbN3nuvLwcuT6UBP13fLG9bEWG9FeeDminyO34YixHblNE300Wm9DFmotphZClcEG65t9b0esb1Qn/N6bbqNdBegvKBsvny9d2drDOHEnGNTC4feINo4v+H2hbZta11R7EZ30qvSwZ+XKiE1Mw7rHMRZzPJ9Lc+3WBeujrGu6zaYC+ZHb1A4JUbO/ookP9jg18ZaOd5vIheDBhhexrD1Uja0JhctvmPl1tuVtYfjQ4dhQiKysLP71/IvMuPAy/vPqYgYNGEj3rt0ZPmQYSZ54giUBCFmCpQGCvgDBsipsMESwJMCPbvk+N37vJv731//LpRdPc+qVVFFbUUOoupZgaRUANhiitrya9NgU5vzfY1wx4wpCoRAds7J4fcGrDer/4TcPcsePf0DO4Bxqg0EmTTiXx//4KP/3u9+T++47eDweBg0YyMUTL6T2UBVN/uyd9Gd4Ituc4IbHWd14vsib+eKMcxhVCzRkTDawyFo7uIl1lwG3AZfiHBj/sLV2zNHaGzVqlF21alWLjO1o3n77bTbY7sxevJnJ/bN4/LoRJMT6sDZEWdkGDldspSqwh5pgKaFQlfOodZ5rQwG3zHmudcsjy0KhqhYZpzE+N+T5w0HOeGLry5p6bWLxeOvqR65ratv6esYTg8f4MCYGjycGY2IwHp8TEMNlPvfR8j/Ap/Jf6s2GzciQWv9UXw/cZ1v335EBzjZaV7dN5Ou6xQbt1q+3jeqG14WcQhs5libq2sh1TdS1Ee+hqfdnwQmzwKa8TQwY0D9i3E1/FuHPMXIsNKrXeJ1t+MI2t842LmtcufG2EQvHMZaGY478zhpv23S7R/w+PYGxnOy2ez//nI6dOh19m+bW2SPf7nFv20S1ZscZ+bPSoJql6Z+D+kr7Rhn6Z/elSbbZheadbLVmtztyhQ1Z54+7u6r8cDlJiUkUHSxi4vQLyF3wHzp37PSFxtdkZf3jsVnG5yGQGGqxma28vDwGDhzYsA9jVltrRzVVv0VmtowxfwMmA5nGmF3A3UAMgLX2CeBVnKC1BagAvtUS/bYEYwy3XdCXDkl+fjH/E659+kP+fMNo0hNjSUnJISUl5wu1b22IUKgGa6upDVVhQ9VuCGvqOeK1PbLchiLbcNfbaif82WqCwTJCoQPNtFmF+xe6RTUMX/Wvw88mFuNx1xmfG+RiMA3qRWxnYgiF9rBt21qnTU8MxnjD4c8Yb0Tw84aDoFOnrixyu7qxed1+6l/XBUaPxwfGEw6O7WcH1IkpK8sjcVTn1h6GNLIudw8DJw9o7WFETVFeHjEdE1p7GCfE2V2VFF6eOXk6JSUlVFdX86t7fkXPYWe12tiOe4LlZILbF5m8Oan+Ijc7VgMGAodPopOW0VJnI37tGOstcGtL9NVSbG0t2796JYl9z4bJk/namJ5kJMZy+98+4qtPvMe8G8fSLS3+C/djjAev1w/48dG6xx2EQkGsrT560LM12FAN1tYQskFsqCZcVrdsbY0bIINuWbXzbGuwoSAhW+0+u3VC1VgbxIaC1AYDDbdv1J61QSxVbC849RcJNBHBrD5ARoY0X8OyutcRdT0R4a++jq+JsqaCYn3d8HYRfTQoc8dzzLJwX4qQIq2lLZ3AcNy/C07qV8ap/z1zQj2e2O0MW1SbOkD+VDJeL7ayEt/O+pMkp57Tmb98ewzfmbeKrzz+HvNuHEO/Tq1/YGZL8Xh8gA+vt23/KzE3N5fzzz8fa2vrA5j7CLmhrcFyU2WhmvD2Tuirdes4r+uCoA3VRpTVr68Pi7XhENlgO1tbH0CDVe52bp1QxHYNyoLh8lPNNApgDcNj02WeRiGuNlTM+vXzw8sN63kbBMPIZU+D0BcZIOvLmq0T8Vw/Pm/E7KS3UV2FShFpe9pt2LLWYs/uQ3D9hgblY/t04MWbx/PNZ1fw1Tnv8fh1Izm3b2YrjbL9MsZgjBMOzzTWWjd0RQayEwiSTYbLyOWmwuVRguQRZfVhtTZUhQ06r6GUQ2VFDcYdCgWBWjdQOv1HY3f18aoPYr5GgSwyxPnwhHc3HxnqGgc9T2QbnsbtNhMojwiGxwiUnka7wZsJlB5Pw/ejO66JnB7OvL9kx8lay19TU8js04uRgQCeiGuPDOySwr9umcA3n13B1//0IX2yEumZkUCX1HiykmLJTPbTIdFPpvs6M8lPSpz+VS3Hp2GQ9Lf2cI5bbm4uE8ZPPmY9a0ONwmTdDGV9IDty1rKW+pnIRnVCwYZtNGqzbpaxYVldO5FlNQ36sRFjCYV3d1ceOb5QU303HH/r8vJ2bkyD8OY5Ivg1HSgb7go/Vug8iUDpadzGiQdK5+epLsAb/Z6V01K7DVsejweP8VCZkEDVlq3ED254u4YeGQm8csckXlixk5XbD1JYXMH63aUUHa6mqWMAY30eMhOd8NUxOY7u6fHuI4Hu6fH0SE8gRVeml3bAGA/GeHDOkTnzb/3hHJIaOrlAGapttP74A2VdGzsKttG9e7eG7YSaC50N+6kNVTaqe2RAbmo7Z6bz1MhIf5KyskazpcZgGl7YLaLMXQ5fwsPUPwwR2zVVp+m2jTmevuqXrQ1QEwxf3KL58cARbTe9LKe7dhu2APxeD9VxcVTl5x8RtgDiY73ceG5vbjy3d7isNmQ5eLiaA+VV9Y8yZ3l/eRUHyqspPFjB+1sPcLi64S+k5DgffbKSOCsrkbOykjgrK4mzOybSMyORWJ92B4icjpw/hs5MDIB74fBTpnBnLn3PnnxK+2w4e3lkoAw1MdN41EAZahzm6gNpaUkKfn9n6q7BUX82nbtcf32ViDLqrqnSsI61WOeaLo3quCURy5Hb7dt3gJ/97HesXLmOtLQUYmNj+P73v0VaWgqPPPIcL7746BGfUWVFC37gTYQ76i6WfLQwd0S4jAx3xwqXx+qrUZ0m2m643L7DZbsOW4nx8RTVBKnKzz/ubbweQ1ayn6zko+/+sdZSWlnDruJKdhVXsKu4koKiw2zbf5jlWw7w8pr6a7p6PYZeGQkM6JLMwM4pDOySwsCuKXRNjTtjf/BE5PTVcPYyuvLK8/D7s6LeT3OstUydOoHrr7+eF19cBFgKCgr497//TZcuPfF6E0lK6oeNCHyHDx8mISGB+gDorLONw1xEIHRLIi6fcPLhsm7c9eHyyDB5tHDZ+hfsOt7A13C5qXBZFyaNiQEST+WbaKBdh63k5GT2V1ZRvnkzzVxe7qQZY0hLiCUtIZbB3VKPWF8WqGH7gcNs3V/O1n2Hyd9XxobPDvHqJ5+H66TE+RjQJYVB7iOneyp9Oybh82oWTETkVFiyZAmxsbHccsst4bLevc/ijjt+QG5uLsZ48Hj83HPPPSQlJfGTn/wEY4IMGzaGRYsWkZ2dzbx585g9ezbGGIYMGcJf/vIXCgoK+Pa3v82BAwfIysriz3/+Mz179uSf//wnv/71r/F6vaSmprJ06VJqa2uZNWsWubm5VFVVceutt3LzzTdH7T0fEfAahbumyhouw/GEyybDZJTCpccTi8JWK0nPyID9RezdXclO/FUAACAASURBVMipvsRcclwMQ7qnMaR7WoPy8qogn35+iI17ysjbc4hNew7x4qpCKtxdkvExXgZ3S3G3TWVYjzR6ZiRoBkxEznyvzYLPP2nZNjvnwCX3N7t6w4YNjBgx4qSb37BhA/fddx/vvfcemZmZHDx4EIDbb7+db37zm3zzm9/k2Wef5Y477mDBggXce++9vPHGG3Tr1o2SkhIA/vSnP5GamsrKlSupqqpi4sSJXHzxxfTu3ftoXZ+0hrNKcKb8eWmpm1CfjHYdtjI7doRP8ykKBqktLcWbeuQM1KmW5PcxslcGI3tlhMtCIUtB0WE+2V3K2sIS1u0q5fkPdlAVdBJ8WoIT3IZ2T2VEz3RG9EwnNSH60/siIu3NrbfeyrJly4iNjeXBBx88Zv0lS5Zw5ZVXkpnpXEIoI8P53f7+++/z8ssvA/CNb3yDu+66C4CJEydyww03cNVVVzFz5kwAFi9ezLp163jppZcAKC0tJT8/P2phS1peuw5bnbp1B+BQciJVW7aQ0EbvdO7xGPpkJdEnK4kZw7oBUFMbYvPeMj4uLGXdrhI+3lXK47lbqQ1ZjIF+HZMZmZ3OqF7pjOqVQY+MeM1+icjp7SgzUNFyzjnn8K9//Su8/Nhjj3HgwAFGjWp4Czyfz0coVH/WZCBwcpcrf+KJJ/jwww955ZVXGDlyJKtXr8ZayyOPPMLUqVNP7k1Iq2vXB/907tETgMOJiSd0kHxbEOP1cE7XVK4d25P7vzKE174/iU/uuZgXbhrLj77Uj86pcfx77Wf86MWPOe/Btxnz27e45fnVPPPuNj4uLKGmtvUuPCkicrq48MILCQQCzJkzJ1xWUXHkqYbZ2dmsWbMGgLVr17J9+/bw9v/85z8pKioCCO9GnDBhAn//+98B+Otf/8qkSZMA2Lp1K2PHjuXee+8lKyuLwsJCpk6dypw5c6ipce4+sXnzZg4fbr37/MmJa9czWwmJiZjaWiqTkqjafHqFraYkxPqYcFYmE85ypqtrQ5bNe8tYtaOY1QUHWbWjmNfWOwfgx8d4Gd4zjbG9OzCuTwZDe6QRF3OKz1kXEWnjjDEsWLCAH/7wh/zud78jKyuLxMREHnjggQb1vvKVrzBv3jzOOeccRowYQb9+/QBnZuwXv/gF559/Pl6vl+HDhzN37lweeeQRvvWtb/Hggw+GD5AHuPPOO8nPz8day5QpUxg6dChDhgyhoKCAESNGYK0lKyuLBQsWnPLPQk6eOe47gJ9io0aNsqtWrYpqH4u2LeKjZ5fjrwxw9d4D9Hr+L1Htry3YeyjAqoJiVhYcZMX2g+R9fghrnYuyjnDD19g+GYzomd5q4Ss3N5fJkye3St/SPH0vbdOZ/r3k5eUxcODA1h7GCSkrKyM5+cy5r+6ZoiW/l6Z+Lo0xq621o5qq365ntt7461oya2upjo2lyv2XxJl+XFOnlDguG9KFy4Z0AaC0ooYVBQf5cFsRH24/yCNL8nnoLYj1ehjWI41xfTIY26cDI3qmEx+rmS8REZET1a7DVu99wyhPWEYw1kewtJTg/v3EdOzY2sM6pVITYrhoUCcuGuRcaexQoIZVBQf5YJsTwB59ewsPL9lCjNcwtHsaY/tkMLZ3B0Zlp5MQ265/fERERI5Lu/5rGfLVgPWB10d5nJ+q/Px2F7YaS4mL4cIBnbhwgBO+ygI1rNpRzAfbivhw20GeeGcbj729FZ/HMKR7KuP6dGBsnw6M6pVOor9d/ziJiIg0qV3/dQz5arHEAiFKUpOp2pxP0sSJrT2sNiU5LoYL+nfkgv5OCC2vCrJ6RzEfbivig21FPLV0G4/nOuErpy589c5gVHYGSQpfIiIi7TtsERPC1CRgKaesY8fT7vIPrSHJ7+P8flmc38+5V9nhuvC1vYgPth3k6aXbmJO7Fa/HkNOtbuYrg9EKXyIi0k61679+JjYE1SlAOYezMhW2TkKi38d5/bI4zw1fFdV1M18H+WBbEX9ato0n3nHC1+BuqYzrk8E495iv5Dhd5V5ERM587TpseWLBG0omZC0ViYlUvb8SGwphPO36Wq9fSEKsj0l9s5jUtz58rdlR4s58FfHssu08+c42PIYGM1+jsjNIUfgSkTbI6/WSk5ODtRav18ujjz7KhAkTKCgoYPr06axfv/4L9zF58mRmz57NqFGjyM7OJjk5GWMMnTt3Zt68eXTu3LkF3om0lnYdtrx+gy8Yi7UhKmNisJWV1OzaRWzPnq09tDNGQqyPc/tmcm5f50KrldW1rNlZd8zXQZ5dvp0nlzrha7AbvhLKg4wI1Ch8iUibEB8fz9q1awF44403+NnPfsY777wT1T7ffvttMjMz+fnPf85vf/tbHn744fA6ay3WWjynYGIgGAzi87XrqNAi2vUUjs/vIaY2Fr/PS8A4H4V2JUZXfKyXiWdn8qOL+/Pif49n3d1TeeE7Y7ntwr7ExXiZu7yAP66pYtivF/PlR5bxm1c28lbeXkora1p76CIiHDp0iPT09CPKA4EA3/rWt8jJyeHcc8/l7bffPqJ8+PDh4fLKykquueYaBg4cyBVXXEFlZWWT/Z133nls2bKFgoIC+vfvz/XXX8/gwYMpLCzkwQcfZPTo0QwZMoS7774bgMOHD3PZZZcxdOhQBg8ezD/+8Q8AZs2axaBBgxgyZAg/+clPALjhhhvCN7cGSEpKApwL5U6aNInLL7+cQYMGUVtby5133hnu68knn2yhT7P9aNdxNcbvxdb6ifHHUlxVDThhK3nKlFYeWfsRH+tlwtmZTDjbmfkK1NTy54W5BFJ68MG2Ip57bwdPv7sdY+CcrimM692BcX06MLp3BqnxmvkSaU8eWPEAmw5uatE2B2QM4KdjfnrUOpWVlQwbNoxAIMCePXtYsmTJEXUee+wxjDF88sknrF69miuuuILNmzc3KN+0aRMXX3wxmzdvZs6cOSQkJJCXl8e6desYMWJEk30vWrSInJwcAPLz83nuuecYN24cixcvJj8/nxUrVmCt5fLLL2fp0qXs37+frl278sorrwBQWlpKUVER8+fPZ9OmTRhjKCkpOebnsmbNGtavX0/v3r156qmnSE1NZeXKlVRVVTFx4kQuvvhievfufcx2xNGuw5Y/LoYgXvwJ8RRVBAh173ZG3CPxdBYX42VgBy+TJzv3FQvU1PLRzvpjvuZ9sINnljnha1CXFMb1ccLXmOwMUhMUvkSk5UXuRnz//fe5/vrrjzhOa9myZdx+++0A9OvXj169erF58+YG5QMGDAiXL126lDvuuAOAIUOGMGTIkAbtXXDBBXi9XoYMGcJ9991HSUkJvXr1Yty4cQAsXryYxYsXM3z4cADKy8vJz89n0qRJ/PjHP+anP/0p06dPZ9KkSQSDQeLi4rjxxhuZPn0606dPP+Z7HjNmTDhMLV68mHXr1oVnwUpLS8nPz1fYOgHtO2ztX81hziMuMR6KSqk4qw/x+Ztbe1gSIS7Gy/izOjD+rA6AE77WFpaEz3Z8/oMd/MkNX/07JTO2dwaje2cwJjuDjilxrTx6EWlJx5qBOhXGjx/PgQMH2L9/f1T7qTtmq05JSQmJiYnhZWstP/vZz7j55puP2HbNmjW8+uqr/PKXv2TKlCn86le/YsWKFbz11lu89NJLPProoyxZsgSfz0coFAIgFApRXV0dbqNxX4888ghTp06NxlttF9p12Iov2wKchz/R2U99qFNHkpa/j62uxsTGtu7gpElxMd7wbNb36UtVsJaPC0t5f2sRKwsO8s/Vu3ju/R0AZHdIYHR2BmN6O4+eGQln/L0vRSS6Nm3aRG1tLR06dKCioiJcPmnSJP76179y4YUXkp+fz86dO+nfv3+D8s2bN4fLzzvvPF544QUuvPBC1q9fz7p1605oHFOnTuV//ud/uO6660hKSmL37t3ExMQQDAbJyMjg61//OmlpaTzzzDOUl5dTUVHBpZdeysSJE+nTpw8A2dnZrF69mquuuoqFCxdSU9P0sbFTp05lzpw5XHjhhcTExLB582a6devWIJDJ0bXvsOUNAuBNTgDgUEI8XYNBqgoKiOvXrzWHJsfJ7/OGwxRATW2IDZ8dYuX2g6woOMh/8vbyz9W7AOiU4m8Qvvp1TMbjUfgSkaOrO2YLnFme5557Dq/X26DO9773PW655RZycnLweDzMnTsXv9/foNzn84XLb7nlFr71rW8xcOBABg4cyMiRI09oTBdffDF5eXmMHz8ecA5uf/7559myZQt33nknHo+HmJgY5syZQ1lZGTNmzCAQCGCt5Q9/+AMAN910EzNmzGDo0KFMmzat2fD0ne98h4KCAkaMGIG1lqysLBYsWHCiH2O7Zqy1rT2GJo0aNcquWrUqqn0su/+rfFzwPbpeVcPHS9+nT4Kf0c/Oo+vvZ5N62WVR7Vual5uby+TJk1ukrVDIsmV/OR9uP+gEsO0H+fxQAIDU+BhGZ6czOtvZ9ZjTLZUYb7s+QfeoWvJ7kZZzpn8veXl5DBw4sLWHcULKyspITk5u7WFIIy35vTT1c2mMWW2tHdVU/XY9s5Xkc4JmdTCICYUoDwbB69XlH84gHo+hX6dk+nVK5hvjemGtZVdxJSvc4LWy4CBv5u0DID7Gy/Ceac7MV3YGw3umEx/rPUYPIiIiR9euw1ayG7YCgWpiPVBRVUVs72ydkXgGM8bQIyOBHhkJfGVkdwD2lQVYVVAcDmAPvZWPtYRvrj0m27m342id8SgiIiehRcKWMWYa8BDgBZ6x1t7faH1P4Dkgza0zy1r7akv0/UWk+J3jdQKV1cTFxlJx+DD+vn0JrN/QyiOTU6ljchyX5nTh0pwuABwK1LB6hxO+Vm6vv8q9MdCvYzIjs9MZ1SudUb0y6JERr4PuRUTkqL5w2DLGeIHHgIuAXcBKY8xCa+3GiGq/BF601s4xxgwCXgWyv2jfX1RSrHN8TnUgSFJiAocqKok56yzKXnudUEUFnoSEVh6htIaUuBgu6N+RC/p3BOovN1G323Hh2s944cOdAGQl+xnVK52RvdIZlZ3BOV1TdNyXiIg00BIzW2OALdbabQDGmL8DM4DIsGWBFPd1KvBZC/T7hfnjnONxqgNBOqWksrv4EIHOnQCo2rqVePeqvdK+RV5uAqA2ZNm8t4xVO4pZXXCQVTuKeW39525dD0O7pzEq25n5GtEzXbseRUTauZYIW92AwojlXcDYRnXuARYbY24HEoEvNdWQMea7wHcBOnXqRG5ubgsMr3mDSg5Ra6opPVhGRqpTtnL3bs4BPl74bwJFRVHtX5pWXl4e9e++JfQAenSG/+rsoTgQT35JiPziWrYUlTCn4CAhuxWArkmGvmle+qZ76JvmpWOCOS13PZ4u30t7c6Z/L6mpqZSVlbX2ME5IbW3taTfm9qAlv5dAIHBC/9+dqgPkvwbMtdb+3hgzHviLMWawtTYUWcla+xTwFDiXfoj66cwH/0ZtXhV+XwKjxuaw49+vkNGzOyYujt4eD53O4NOp27Iz4VT2iuogawtLWF1QzKodxazZWcw7u5yrM2cm+RnZK41RvTIYmZ3O4K6pxPra/q7HM+F7OROd6d9LXl5em7iMwm9+8xteeOEFvF4vHo+HJ598kquvvppVq1Y1uNI7wNixY/nwww+bbeuKK65g+/btlJeXs3///vBtbx5//HGuvfbaJttcuHAhGzduZNasWU22WVBQwPTp04+4jZDUa8lLP8TFxYVvlXQ8WiJs7cb5R36d7m5ZpBuBaQDW2veNMXFAJrCvBfo/eT4/IU+AULWlc4+eABwsKqL3WWfp8g/yhSTE+phwViYTznJ+YYZClvx95azacTAcwN7YsBcAv8/Z9TgyO52RPdMZ0SudjETdwUCkrXj//fdZtGgRa9aswe/3c+DAgQa3tmnszTffPGp78+fPB5ygPHv2bBYtWnTMMVx++eVcfvnlJzZwaTNa4p/TK4G+xpjexphY4BpgYaM6O4EpAMaYgUAcEN0bSx2PmHisJ4Ct9pDRIROspbSkBH/fvgpb0qI8HkP/zslcN7YXf7h6GEvvuoAVP5/CnOtG8PVxvaiqDfH00m18Z94qRvzvf5j84Nv86B9r+csHO9jwWSnB2tCxOxGRqNizZw+ZmZn4/X4AMjMz6dq1a3h9ZWUll1xyCU8//TQAXbo4ZzbXzTp+9atfZcCAAVx33XUcz4XEH3nkEUaMGEFOTg6bNm0CYO7cudx2220A7N27lyuuuIKhQ4cydOhQ3nvvvQbbb9u2jeHDh7Ny5Urmzp3LzJkzmTZtGn379uWuu+4K11u8eDHjx49nxIgRXHnllZSXlwMwa9YsBg0axJAhQ/jJT34CwD//+U8GDx7M0KFDOe+8807qc2zPvvDMlrU2aIy5DXgD57IOz1prNxhj7gVWWWsXAj8GnjbG/BDnYPkbbFu4dL3PD54A1MTh9XrxYimvqMDfrx+lCxYQLC7Gl57e2qOUM1THlDguyenCJe4lJyqra1m3q4Q1O0tYs7OYpfn7efkjZ5I4IdbLkO6pjOiZzvCe6YzomUaHJH9rDl/klPv8t7+lKm9Ti7bpHziAzj//+VHrXHzxxdx7773069ePL33pS1x99dWcf/75gHPM3DXXXMP111/P9ddff8S2H330ERs2bKBr165MnDiR5cuXc+655x61v8zMTNasWcPjjz/O7NmzeeaZZxqsv+OOOzj//POZP38+tbW1lJeXU1xcDMCnn37KNddcw9y5cxk6dCgbNmxg7dq1fPTRR/j9fvr378/tt99OfHw89913H2+++SaJiYk88MAD/OEPf+DWW29l/vz5bNq0CWMMJSUlANx777288cYbdOvWLVwmx69Fjtlyr5n1aqOyX0W83ghMbIm+WpQvHo+nEk+Ncz+oOK+XyuoA/r59AajKz8c3ZkxrjlDakfhYL2P7dGCse9Zj3dXu1+wsZs2OYj4qLOGppdsIhpx/p/TqkMDwHmmM6JXOiJ7p9O+crMtOiERBUlISq1ev5t133+Xtt9/m6quv5v77nctJzpgxg7vuuovrrruuyW3HjBlD9+7OBZSHDRtGQUHBMcPWzJkzARg5ciQvv/zyEeuXLFnCvHnzAPB6vaSmplJcXMz+/fuZMWMGL7/8MoMGDQrXnzJlCqmpzllggwYNYseOHZSUlLBx40YmTnT+NFdXVzN+/HhSU1OJi4vjxhtvZPr06UyfPh2AiRMncsMNN3DVVVeFxyfHr11fQR6fH48ngKfGOTU/IT6OokCA2L5nA07YSlTYklYSebX7GcO6Ac7s1ye7S/lop3PQ/fKtRSxY61xJJT7GS447+zWiZxrDe6aTlazZLzlzHGsGKpq8Xi+TJ09m8uTJ5OTk8NxzzwFOCHn99de59tprmzzLuG7XY10bwWDwmH3VbXO89eukpqbSs2dPli1b1iBsNTUGay0XXXQRf/vb345oZ8WKFbz11lu89NJLPProoyxZsoQnnniCDz/8kFdeeYWRI0eyevVqOnTocNxja+/aediKw2dK8QVjCNkQKcnJ7C87TE18HJ6UFB23JW1OfKzXuXdj7wygfvbro8ISZ/ZrZzHPvFs/+9UjI97Z9ejOgA3soouuipyoTz/9FI/HQ193r8fatWvp1asXn3zyCffeey/33nsvt956K48//vgpGc+UKVOYM2cOP/jBD8K7EQFiY2OZP38+U6dOJSkpiWuvvbbZNsaNG8ett97Kli1bOPvsszl8+DC7d++ma9euVFRUcOmllzJx4kT69OkDwNatWxk7dixjx47ltddeo7CwUGHrBLTvsBUTR4ynkphQHJXBStIzMuDzfezbvQt/v75U5W9p7RGKHFXk7NflQ50DdgM1tazfXerufizh/a1F/P/u7FdcjIch3dIY7s58jeiZRseUuNZ8CyJtXnl5ObfffjslJSX4fD7OPvtsnnrqqfBZhA899BDf/va3ueuuu/jd734X9fE89NBDfPe73+VPf/oTXq+XOXPmhA/KT0xMZNGiRVx00UUkJSU120ZWVhZz587la1/7GlVVVQDcd999JCcnM2PGDAKBANZa/vCHPwBw5513kp+fj7WWKVOmMHTo0Ki/zzOJaQvHqTdl1KhRdtWqVdHtZO0L/PnF9zl8cAZX/j6H7Ss38nruO0ybOI5e733IoUWv0G/Fh6flBShPZ2f6dYNONWstn5UG3Jkv5+D7DZ+VUlPr/L/fJTWOYT3Swo+c7qkkxB757zB9L23Tmf695OXlMXDgwNYexglpyes5Sctpye+lqZ9LY8xqa+2opuq375ktnx8/lVTgobSijE7uQYwH9u2lX9++hMrKCO7dS0znzq08UJGTZ4yhW1o83dLi+XLE7NeGz0r5aGcJawtL+HhXSfiWQx4D/TolM7ynE76G9kijb0f94RAROVntPGzFE2cqASgvP0z/Ls6B8cUHDxI3zjlbpCo/X2FLzjhxMV5G9spgZK+McNmB8io+Lizh48ISPios4ZV1e/jbCudOXImxXnokWd6vzGN4jzSG9Uinc6p2P4qIHI92Hrb8xLth69DhwyQkJGBsiLKy8vrLP2zOJ2nSpNYcpcgpkZnkZ8rATkwZ6NyMPRSybC86zMeFzuzXuxsKeXbZ9vDux84pcQztkcqwHukM65HGkO6pJPrb968UEZGmtO/fjDHxJFIBwOGKSowxxBpDRSCANy0NX8eOVG3e3MqDFGkdHo/hrKwkzspKYuaI7uSmHmDcxEls3HOItTudXY9rC0vCtx3yGOjbMdk59qtnGkO7p9GvUxI+nf0oIu1c+w5bPj9JODNbhyuc5/jYGMrLnTMz/AP6E9i4sdWGJ9LWxMV43et41d9Z4eDh6vDs19rCEt7Y+Dn/WOXsfqy79tdw99ivYT3S6JIap5NORKRdaedhK44kDgNQWeEErKTEREorKghWVxM/OIcDy5YTqqjAk5DQmiMVabMyEmO5YEBHLhjQEXDOftxRVBEOXx8VlvDn5QVUu/d37JjsZ0j3NIZ2T2VIjzSGdEslXTfeFpEzWLsPWynuMVuBSucO7qlpaewqLqV43+fE5QyGUIjAxo0kjGrybE4RacQYQ3ZmItmZifzXcOfK91XBWvL2lLF2ZzEf7yrl410lvJm3N7xNj4z4+gDWPY3B3VJJ0vFf0kYkJSWFLxzaEgoKCpg+fTrr169n1apVzJs3j4cffrjF2pe2p33/NvPFhY/ZqgrUANAhKwu27+DznYUMzMkBoPKT9QpbIl+A3+cNX8erzqFADet3lfLxrlI+2V3C2p3OGZAAxsDZWUlOAOvhBLABnZOJi/G21lsQiYpRo0YxSn9fznjtO2zFxON3Z7aqK537T2V1dq7Cu+/zz8iZMBFfly5Urvu41YYocqZKiYthwtmZTDg7M1x2oLyKT9yZr3W7Snln8z7+tWYXADFeQ//OyQ1mwPp21AH4curk5uZyzz33kJmZyfr16xk5ciTPP/88xhhmzZrFwoUL8fl8TJ48mYcffpgbbriB6dOn89WvfhVoeoYsNzeX2bNns2jRIu655x527tzJtm3b2LlzJz/4wQ+44447WuOtSgtr32HL5yfGBAAIVjnHk3Tu3gOAg0UHAEgcP56yxYsJVVfjidVxJSLRlJnkP+L4r89KA6wrLGHd7lLW7Srh3x9/xgsf7gScA/DP6ZrSYAYsu0OCDsA/Q7374mYOFLbc7jyAzB5JTLqq33HX/+ijj9iwYQNdu3Zl4sSJLF++nIEDBzJ//nw2bdqEMYbCwsKTHs+mTZt4++23KSsro3///txyyy3ExMScdHvSNrTzsBWHMZagp4oaN2ylpTtnWZWWlAKQMm0qpS+/zOHly0m+4IJWG6pIexR59ftLcpxZ51DIUlB0mHURM2AvrNjBs8ud/4dT4nzkuDNfdTNgOgNSWsqYMWPo7t5tZNiwYRQUFDBu3Dji4uK48cYbmT59Oueff/5Jt3/ZZZfh9/vx+/107NiRvXv3hvuT01f7DlseLyHjI+SrJuQcH09MTAxeG6K8wjlLMXHcODwpKZS9/rrClkgb4PEY+mQl0ScrKXwAfrA2xOa95azbVT8D9vTSbQRDzgVYM5P84eA1pHsqg7ulkpXsb823ISfhRGagosXvr/+58Xq9BINBfD4fK1as4K233uKll17ioYce4p133sHn8xEKOf8ICIVCVFdXn1T7cvpr32ELCHlisN5qbMT/A36fj0DAOXDexMaSMm0apQsX0ukXv8CbktJKIxWR5vi8HgZ1TWFQ1xSuccsCNbXk7TnUYAZsyaf7sE7+onNKHIO7pZLTLZWc7ikM7pZKx2TdgkhOXHl5ORUVFVx66aVMnDiR3r17A5Cdnc3q1au56qqrWLhwITU1Na08UmktClueWPBWY2vqD7JNjI+jqLICGwphPB7SrrqKkhdfpPTf/ybjuutacbQicrziYrwM75nO8IgLsJZXBdmwu5RPdpey3n1+a9NeBTD5QsrKypgxYwaBQABrLb/97W8BuOmmm5gxYwZDhw5l2rRpJCYmtvJIpbUobHliMd4aTE39KeUpycnsP1RGeUkxyRkdiB98DnHnnEPJP14k/dprdeyHyGkqye9jbJ8OjO3TIVx2rADWKcVPTrfU+hDWLZWOKQpg7UndGYSTJ09m8uTJ4fJHH300/HrFihXh12VlZQB06tSJDz74IFz+wAMPAM6M1/r1649o85577mnQb10dOf0pbHli8Hir8FTXXyE+PaMD7NnLgc92k5zh/FJOu+oqPr/7bgIff0z8sGGtNVwRaWHNBbCNnx1qFMDqd0F2TG4UwLqn0kkBTESaobDl8ePzVOMNphKyITzGQ2anTrBhI3t376L34P/H3n2HR1Xlfxx/3zt90nuFJJCEkBAIEDpCEJGmCIo06RZc26or9r6uoui6Kv5WUYoiHRQQZVGEgIJGOtKkhpaEkISQMjPJlPv7Y5IhkdAnBTiv5+EJVI5IlAAAIABJREFUmbnlzNzAfHLOud/TEgDv/v3JffttTs9fIMKWIFznPHVq2sf40z7G3/VYaZmN3dlF/HHcGb7+OFF9DlhQRQBLFgFMEIS/EGFL1qBWWdDYdZhtZjw0HoRGOG+zzcs9u5yIytMD79tuc06Uf+5ZMVFeEG4wHjo17aL9aRddcwCr7AFL/zMXx18CWNUhyBBvnZiKIAg3GBG2ZC0aqQytXU9JeQkeGg9nzxZwuqCg2ra+Q8VEeUEQzqopgJnKzw5BVg5DVg1gAR5akiJ8SAr3JincmxbhPjT2NyLLIoAJwvXqhg9bdpUWnWxB49BRUl5CiEcIHh4eSIpC8V+WVTAkVUyUnzdfTJQXBKFGRq2a1Gh/UmsIYDtPnGFXVhE7s4qq1QHz1KlJDHOWrkgKd94FGRvsiUYsRSQI14UbPmw5ZA16nOsjFpWUgJ+zarVGlig1W87Z3m/EcLJfeBHzpk0Y27Wr6+YKgnANqimAldns7D9Zwq6sM+w8UcSurDPM33gMs9UOgFYt0yzEixYR3iSGO3vCmod6Y9CKxbgF4VojwpasxSA5q8UXV1SNBzBotZRWFDatyrtfP06+M5mCOXNE2BIE4Yrp1CpaVMznGlrxX4ndoXA4r5RdWc4esF1ZZ1ixM4e5vzvX2pMlaBrkWTEE6QxgpValHl/FjUGlUpGcnIyiKKhUKqZMmULnzp3JzMwkJiaGF154gTfeeAOAvLw8wsLCmDBhAlOmTOHVV1/F09OTp556qp5fhVCfRNiStRgVZ8gqKTW7Hvf0MFJUUkK52YTWcLYshGww4DtoEAVffYU1NxdNcHCdt1kQhOuTSpaIDfYkNtiTO1KcSxFVLsZdOQS5O+sMvx0qYMm2LNd+b21Z7QpgLSKcX4O9xER8dzEYDGzbtg2AlStX8txzz7F27VoAYmJi+O6771xha+HChTRv3rze2io0TCJsyVo8FOfcrNLSsz1Zvr6+nMg/zenck4RExVTbx2/4MApmzqRwwUKCHnm4TtsrCMKNpepi3L2TQl2P55eUsSuriG9/2YpJ78vurCJW7jp7B3Wgp5bEcB9aVOkFExPxr15RURF+fmdXJTAajTRv3pxNmzaRmprK/PnzGTRoEPn5+fXYSqGhEWFL1uKlnALAZC5zPR4QFASHMsk9fuycsKWNisLjppsonD+fwAkPIGk0ddpmQRCEAE8d3eKDcGRpSUtrAziLse7JPjsRf1dWEVNrmIjfPMyLxHBvmod5Ex/ihV5zbcwDWzNzKrlHDrn1mMFRTegx9oELbmM2m0lJScFisZCdnc3q1aurPT9s2DDmzZtHSEgIKpWKsLAwEbaEakTYkjV42QoBsFQJW0FhYQDkZmfXuJ/fiOEc/9tDFP/0E959+tR+QwVBEC7Cs4ZSFJUT8V3DkNlFLNp8nNJfnRPxVbJEk0APV/hyhjFvgrx09fUyGpyqw4i//voro0ePrraUTp8+fXjppZcICQlh6NChOByO+mqq0EC5JWxJktQH+ABQAZ8rijKphm2GAK8CCrBdUZQR7jj31XLIWow452xZzGdXZA8NdxY2LcjPq3E/z27d0EREcHr2HBG2BEFosKpOxK/kcCgcLTCxJ9sZvvZkF7HxcAFLq8wDC/TUVQQwL2dZijBvYgI9UNdjOYqL9UDVhU6dOpGXl8epU6dcj2m1Wtq2bct7773H7t27WbBgQT22UGiIrjpsSZKkAj4GegHHgY2SJC1TFGV3lW3igOeALoqinJYkqcHMKnfIWrSSs8RDmeVs2PLx9QWgsLCwxv0klQq/4cPIffc9yvbvRxcXV/uNFQRBcANZlogO9CA60IO+yWGuxwtN5ezJLnYFsN1ZRUw/mIfV7hyG1KllmoV6uXq/mod5kxDmhbf+xplKsXfvXux2OwEBAZhMZ+f5/uMf/6B79+74+/tfYG/hRuWOnq32wAFFUQ4BSJI0D7gD2F1lm/uBjxVFOQ2gKEquG87rFs4K8s67EG1lZ7t+tVotKhSKiorOu6/PXXdx6sOPOD13LqEvv1zrbRUEQahNvkYtnZoG0Knp2UW5y20ODp4qcYWvPTlFrNyVw7yNx1zbNPI3uAJY5ddIP8N1czdk5ZwtcN4d+sUXX6BSVZ/nlpSURFJSUn00T7gGSIpydTVaJEkaDPRRFOW+iu9HAR0URXmkyjZLgH1AF5xDja8qivK/Go71APAAQEhISNt58+ZdVdsuhd/hb2mZ+Tkfn/ya49Fb6dMx1fXcr+vSsRUX0aV3X2R1zb+5ec/8At3WreRNegvFYKj19t4ISkpK8PT0rO9mCH8hrkvDVB/XRVEUCssUjhY7OFrk4FjF15MmhcpPFIMaGnvJNPKSaewt09hLJtxTRqu6vADm4+NDbGys+19ELbLb7eeEMaH+ufO6HDhwgDNnzlR7rEePHpsVRUmtafu6miCvBuKANCASWCdJUrKiKNXG6BRFmQpMBUhNTVXS0tJqvWG7T6YjSeBQW9GpjFQ95/GDBziQmUmzRpFENKu5boo5IIDMu4eQXFAg1kt0k/T0dOri2guXR1yXhqkhXRdTuY29OcVne8Gyi9iQU8yqo+WAczJ+0yAPEkKdw48JoV4khHoT5qM/by/Ynj178PLyqsuXcdWKi4uvuTbfCNx5XfR6Pa1bt77k7d0Rtk4Ajap8H1nxWFXHgQxFUazAYUmS9uEMXxvdcP6r4pC1zq9qK47y6s/FxMVx4NhxDu3Zed6wZUhORp+czOk5c8V6iYIg3NCMWjVtGvvRpvHZOlQOh8KRysn4FXdDbj5ymmXbz07G99arqwQwb5qFetEs1AtP3Q1/w7xwnXDHT/JGIE6SpBicIWsY8Nc7DZcAw4EZkiQFAvGAe4ulXKHKsKWobSjW6kGpSVw8rF7D4QMHuekCx/AbPpzs55/HlPE7Hh071GJrBUEQri2yLBET6EFMoAf9qkzGP2O2su9kMXuzi9ibU8zenGK+3nKCkrIjrm0a+Rt4q4c/OWcs6DUyeo0KnVoWv9QK15yrDluKotgkSXoEWIlzPtZ0RVF2SZL0OrBJUZRlFc/dKknSbsAOTFQUpUFUfKsMW5LajmStPpYbHByMDJw4fozi/Dy8AgJrPIZ3v77kvv02p+fMEWFLEAThEvgYNOfUBFMUheOnzc7wlV3E3pPF2BwKp4otrrlgsiShUzuDl16jwlARwuqzJIUgXIxb+mgVRfke+P4vj71c5e8K8GTFnwbFITsnvssaG1JZ9bClUqmIjIzkmM3Kmtkzuf3Rf9T4G5Ws1+Mz+C4KZn6BNScHTWjoOdsIgiAIFyZJEo38jTTyN9IrMQRwztlqFu5Dmc2O2erAYrVjsdopttg4bTo790Mty+g1MgaNCl1FCNOpVWJ5IqFBuOF/FXDIzirJKrUdlU2DQ6le+bd3374oag07DmWSsWTheY/jN3w4OBwUimJ2giAIbiXLEgatGn8PLeG+BpoEeboq3scEehDmY8BLr8buUMgrLef4aRP7c0vYlVXEnznFHM0vJbfIQpHZSrnNztXehS8Il0uErYqeLbXGjsaux2Q1VXs+IiKCrl27YvMNZNXadWxYNLfG42gjI/Hs1o3TCxbiKC+vcRtBEATBfTQqGS+9hiAvHY38jcSFeNEi3LneY2N/I0FeOnRqGVO5nZwiC5n5pezNKWZ3VhEHcks4ftpEXnEZJRYrVrvjvCFMpVKRkpJCUlISrVq14r333nMtybNp0yYee+yxC7Zz5syZPPLIIxfc5q/efPPNy9q+qrFjxxITE0NKSgpt2rTh119/veR9q7b1k08+4csvv7zidlyKzMxMDAYDKSkprj/lbvwMnTlzJllZZ2/GuO+++9i9e/cF9qgdN/ytHpVztjRqB1q7jhJrCZ7a6jVrevbsiSzLrFu3jlUbt2Aym+k5ctw5Q4p+o0ZRct99FK9Ygc8dd9TZaxAEQRCcJElyzeeqyu5wYKkchrQ5vxaZrRQ4qg9F6irmgOnVZyfkV10bMTc3lxEjRlBUVMRrr71Gamoqqak1lla6Km+++SbPP//8Fe8/efJkBg8ezA8//MCECRPYsWPHZR/jwQcfvKztbTYbavXlx4qmTZu63l93mzlzJi1atCA8PByAzz//vFbOczGiZ6sibGnVNjR23Tk9W+D8x3vzzTczZMgQJKMn6/88yDefTDnntyCPLp3RNm1KwZezRDe1IAhCA6KSZTx0agI8dUT4Gmga5OlacqhJoAfhPga8DWoUBU6XlnOi0MzBUyXszi7CocDhvFKyCs2oPXz5YMr/MWWK8zMgPT2d2267DYDff/+dTp060bVrVzp37syff/7pOv+xY8dIS0sjLi6O1157zfX4V199Rfv27UlJSWHChAnY7XaeffZZV9X6eyrqN9a0nd1uZ+zYsbRo0YLk5GTef//9c153t27dOHDgwHmPATBjxgzi4+Np374969evd+376quv8u677wKwceNGWrZsSUpKChMnTqRFixaAM8wMGDCAm2++mZ49e1JaWsr48eNp3749rVu3ZunSpYCzoOjEiRNp164dLVu25NNPP73g9apaqHfRokWMHTsWcPbaPfbYY3Tu3JkmTZqwaNEi13Zvv/02ycnJtGrVimeffZZFixaxadMm7rnnHlJSUjCbzaSlpbFp0yYA5s6dS3JyMi1atOCZZ56pdu4XXniBVq1a0bFjR06ePHnBtl4K0bNVMYyoU9tRK1qKLcXn3TYxMZGghx5ixtRP2ZGTx5l/T2b040+iUjnfRkmS8B81kpxXX8O8dSvGNm3q5DUIgiDcCAq/PUh5Vqlbj6kN98D39qZ46s8+pigKVntFT5jNjiSBze6gpMzm/EXaMxirzcb6nQfJOWOmzGan0FROVJM41q5bR5nFQkZGBs8//zyLFy8GnEFs586dGI1G2rVrR//+/fHw8GD+/PmsX78ejUbDQw89xOzZs5k0aRJTpkxx9fbs2bOnxu2SkpI4ceIEO3fudL4/Nazl++2335KcnHzeY/Tq1YtXXnmFzZs34+PjQ48ePWos1jlu3Dg+++wzOnXqxLPPPlvtuS1btrBjxw78/f15/vnnufnmm5k+fTqFhYW0b9+eW265hdmzZ+Pj48PGjRspKyujS5cu3HrrrUiSxMGDB13LIXXp0oWPP/74gtcsOzubX375hb179zJgwAAGDx7MihUrWLp0KRkZGRiNRgoKCvD392fKlCm8++67pKamUlx89vM9KyuLZ555hs2bN+Pn58ett97KkiVLGDhwIKWlpXTs2JF//etfPP3003z22We8+OKLl/DTdH43fNiyq5wT5A1qGwBFpgv/Qw4KCuLvT01k2pQPOVJcypS33+KBx/+BwWgEwGfAAHL//T4FX84SYUsQBOEaJEkSWrUKrVqFNxokIC7EC0VRKLc5sNgczuFKtYzVrlBuUzhaYCIn6wRvv/wsRzMPoZIl7DYbZ0zlWO0OevXqRUCAc83JO++8k19++QW1Ws3mzZtp164d4FyDMTg4+Jz2/PTTTzVud/vtt3Po0CEeffRR+vfvz6233uraZ+LEibzxxhsEBQUxbdq08x4jIyODtLQ0goKCABg6dCj79u2rdv7CwkKKi4vp1KkTACNGjGD58uWu53v16uVagPuHH35g2bJlrh4xi8XC0aNH+eGHH9ixY4erJ+rMmTPs37+f+Pj4yx5GHDhwILIsk5iY6Op1WrVqFePGjcNY8Vl8sQXBN27cWO1133PPPaxbt46BAwei1WpdvZVt27blxx9/vOS2nc8NH7YUyfkWGGUrACWl5w4j/pVOp+NvTz7F/KmfsDcrh/+8+w7j7p9AaFgYstGI792DnWUgsrPRhIVd9HiCIAjCxfne3rRezy9JEjqNihPHjqBWqWibEEPJyaN46tXEh3gx6dl36NGjB8PGzObIsaOMvrM/RwpM5JyxUGi2sv9kMTqNitIyG0abAx12Ro8ezaRJky54XkVRGDNmDG+99dY5z23fvp2VK1fyySefsGDBAqZPnw6cnbNVac2aNTUeY8mSJVf9vnh4eFRr6+LFi2nWrNk5r+Gjjz6id+/e1R7PzMys8ZhV50RbLJZqz+l0umrHdTeNRuM6v0qlwmazXfUxb/g5W0gSqPUYVc6wVVpqvsTdJIZN+Budm8dRXm5l6qefsOsP5wRE/xEjQFE4PWdOrTVbEARBqHunTp3iwQcf5JFHHnF9IEuAXqPCXFpM89hoQjxkfv5uMRqVTGywJ34eWn7/JZ2SokLyC4v5dtlSopq3JrplB+bOX8ivOw9xNN/E3sws/ti7H4vVjkajwWp1fi717NmTRYsWkZubC0BBQQFHjhwhLy8Ph8PBXXfdxRtvvMGWLVvO2+7zHaNDhw6sXbuW/Px8rFYrCxeeW+LI19cXLy8vMjIyAJg3b955z9O7d28++ugjVwjaunWr6/H//ve/rte0b98+SkvPP5IUEhLCnj17cDgcfPPNN+fdrlKvXr2YMWMGJpPJ9foAvLy8qg0fVmrfvj1r164lLy8Pu93O3Llz6d69+0XPc6Vu+J4tANR6PCt6tkrNlxa2Kt06bCR+K79nxZo1LFy0mJMnT9Kj5y143XILpxcsJPBvf0Ou6NYUBEEQrj2Vk9WtVitqtZpRo0bx5JPn1uh++umnGTNmDK+//jq333474Fwv0lOnpmOHDjz5wGiOHz/OiHvu4Y5eN2GxOnju5Ve5d9hA7A4HKrWa59+YjOIZxKDhY0hIbEFyqxQ+mf4Fz730Krf0uhUUBxqNho8//hiDwcC4ceNcZShq6vmqlJiYyBtvvMGtt96Kw3H2GB07duTVV1+lU6dO+Pr6uuZO/dW0adO4//77kWWZ7t274+PjU+N2L730Eo8//jgtW7bE4XAQExPD8uXLue+++8jMzKRNmzYoikJQUNAFe9UmTZrEbbfdRlBQEKmpqZSUlJx3W4A+ffqwbds2UlNT0Wq19OvXjzfffJOxY8fy4IMPYjAY+OGHH1zbh4WFMWnSJHr06IGiKPTv3587arGKgNRQ75pLTU1VKu8YqE3p6emkbZrAkYChLM9IQ317FhP6j7zs4+zN2MDiBfOxevnRLC6OfrGxZI0ZS+irr+I3bGgttPz6lZ6eTlpaWn03Q/gLcV0apuv9uuzZs4fmzZvXdzMuS3FxMV5eXpe9n92hUG5zlqYos9qxWB2U2RzOQqwV20iAVq2qWLJIRldRpqK2q+WXlJS47hCcNGkS2dnZfPDBB7V2vtpwpdelJjX9XEqStFlRlBrrgIieLQC1DqPkHBM2m8uu6BAJHTozytOTeVP/y5/7FApOF9C9VSsKvpqF79AhYuFUQRAE4YJUFZXyDdrqjzscCmU2B2W2ygDm/FpssaFwtsNEq5bRq1XoKpYq0ld8VbkhhH333Xe89dZb2Gw2oqKimDlz5lUf80YiwhaAxoAWZxdlueXKJ8I1TmrJmMefYu6/3yZPcbA8KZGbVvyPkA0b8OzSxV2tFQRBEG4gzuWKVBi01Qu1OirujiyzVukNszkorixRUUGjOlugVaepCGRq+bIW7x46dChDh4pRmislwhaAWodGufqwBRAS05Qxz7/C3EmvkecZyOpbeiLPn08nEbYEQRAEN5KrVMuvOoOqaomKqkGstMyGo0oIU8sSOvXZEKar6BXTqmQxGuNmImwBqA1oFefdCrYyx0U2vji/sAhGvfIm8998lSy1kZW+vqh++IH2VWqgCIIgCEJtqCxRodOowKBxPe4q1mpzUFYxHFlmc1BksWEzOarvr5Yr/lQOS7pvSPJGJMIWgFqHylqKIjmwl7nnhgEv/0DuefUtFv3zBTJNVr7fsAGTVkv37t3FbwyCIAhCnatarBV99edsdkfFvLCKEGZ1VtAvMlup+qmoUckVk/NVZwOZRoValsRn2wWIsAWg1iNZCnGobdjdt9g4Bk8vhr72NgsmjCXTZiU9PZ2ioiL69++PSqW6+AEEQRAEoQ6oVc45XB666o+75oVVBLAym3MJo4LS8mpDkipJOjsUWRHAdGoZrVpGFiFMhC0ANHqwlaFo7Cjl7v2h0OoNDHrqRZY89ShH4uLZssVZDffOO++8otXRBUEQhLq3ZMkSBg0axJ49e0hISKjv5tSZqvPCMJx9XFEUbHbFeWdkZY+Y1U5JmY3TVYckkdC6esCqhDG1jOoG6g0Tn/YAaj1YzaBxgNX9RfU9kpPpGh6DLjOT/XY7u4GysjKGDh2KVqu92O6CIAhCPZs7dy5du3Zl7ty5vPbaa/XdnHonSRIatYRGLeP5l+fsDsU1H6zq3LC/3iWpqjJB3xXI1DLa63BumFiuB0CtA1sZklZBstbO8F7A+HtJPHCMdqFh6LIyOXjwILNmzTpnzSdBEAShYSkpKeGXX35h2rRprqVqHA4HDz30EAkJCfTq1Yt+/fq5FlneunUr3bt3p23btvTu3Zvs7Oz6bH6dU8kSRq0aP6OWUB89UQEexId40SLcm2ahXkQHehDuY8DXoEGWoKTMxskiC0cLTOzPLWFX1hn2ZBdx6FQJJ06bOFVcRpHZSpnVXitrIdYF0bMFoDaAzYxKK6GyqLE77Khk94Yuj65d0MfH02TTDvwfup8fvl7IMUVh5owZjB4zxrVSuSAIglCzFStWkJOT49ZjhoaG0rdv3wtus3TpUvr06UN8fDwBAQFs3ryZw4cPk5mZye7du8nNzaV58+aMHz8eq9XKxIkTWb58OUFBQcyfP58XXnjBtUD0jcx5l6MKXQ0T9J3V8529YOW2s5P1C81W7I7yasfQqqoOS8quivoNeZK+CFvg6tlS6SQ0dj0mmwkvrXtK+leSJAn/8ePIfvY54j198Rj/AMumTyUH+GLGDMaOH4/BYLjocQRBEIS6NXfuXP7+978DMGzYMObOnYvNZuPuu+9GlmVCQ0Pp0aMHAH/++Sd79uyhV69eANjtdsLCwuqt7dcK1XkKtyqKUjEsefZOycowds6wpCSh/cu8sIYyLCnCFoDGADYLGi8VGruOUmup28MWgE+/fpz69/vkT59OsxkzGObtzYIp/+EkMGPa54y79z4RuARBEM7jYj1QtaGgoIDVq1fzxx9/IEkSdrsdSZIYNGhQjdsrikJCQgK///57Hbf0+iRJEmqVVOOdkpV1w8qq9ISVWe2YymwUmqrXzDRoVYTW48ermLMFzp4tQKuTXGGrNkhaLf6jR2H69Tcsu3fTKKklo599Cd/Ck+SeymP6Z1PFHC5BEIQGZNGiRYwaNYojR46QmZnJsWPHiImJwd/fn8WLF+NwODh58iTp6ekANGvWjLy8PH799VcArFYru3btqsdXcP2qrBvmpdcQ6KkjwtdAkyBPEsK8aRHuQ3yIF1H+RkK99fhUKe5aH0TYAufdiIBOJ6O16ymxltTaqXyHDEE2GsmfPgOAoKgYxr/4OoGmQk7l5/P5J/8VgUsQBKGBmDt37jm9WHfddRc5OTlERkaSmJjIyJEjadOmDT4+Pmi1WmbNmsUzzzxDq1atSElJYcOGDfXU+huXLDtLVvgYtQR76wn20l98p1okhhHBFbb0WhmVIlNsqb2wpfL2xnfIEApmzSL4ySfQhIfjHRTMvS//k1nv/Ius04VM/XgKEx55FJ1Od/EDCoIgCLVmzZo15zz22GOPAc67FD09PcnPz6d9+/YkJycD0LJlS9atW1en7RQaNtGzBa6wZdA5346SElOtns5/9CgACr6c5XrM4OXNuBdfI0onUVBUzCcf/IfycjeWsxcEQRDc6rbbbiMlJYWbbrqJl156idDQ0PpuktBAibAFrjlbxsqwZardsKUJD8e7b18KFyzAXlR09nGdnjHPvESsrwenS03833uTsYrAJQiC0CClp6ezbds2du/ezdixY+u7OUIDJsIWOO9GBDw1zltDTabanzMVMH4cDpOJwgULqj0uq1Tc8/hEmocFU1hmZcrbb1Jmrt3wJwiCIAhC7RFhC1w9W57ayrBVVuun1CcmYuzUkYIvZ6H8pfdKkiSGPvgwSVGNOGOHj9/6F6aiM7XeJkEQBEEQ3M8tYUuSpD6SJP0pSdIBSZKevcB2d0mSpEiSlOqO87qN2tmzZdQ463KYzbUftsC5hI8tN5fCpUtrfP7ucfeSHNuEIlnDf996g9PZWXXSLkEQBEEQ3Oeqw5YkSSrgY6AvkAgMlyQpsYbtvIC/AxlXe063q5ggr5WdIavcYquT03p07YK+RQvyp36GYqv5nHfeM4oW8XEU6zyY+s6bHNqysU7aJgiCIAiCe7ijZ6s9cEBRlEOKopQD84A7atjun8DbQMMrIlUxjKhRVYYta52cVpIkAv/2INZjxyj6/vvzbnPX8BEkJzbH7BPIvM8/ZcPC2Tgc9jppoyAIwo0sPz+flJQUUlJSCA0NJSIigpSUFHx9fUlMPKdf4YKWLFnC7t27L7pdWVkZQ4cOJTY2lg4dOpCZmXmFrRcaCneErQjgWJXvj1c85iJJUhugkaIo37nhfO5XMUFeIznDlrWs7oKMZ48e6OLjyfvkUxSHo8ZtJEli0OC7SUpKpDw4krU/rWLRP1+kKC+3ztopCIJwIwoICGDbtm1s27aNBx98kCeeeML1vSxf3kfopYatadOm4efnx4EDB3jiiSd45plnrrT5QgNR60VNJUmSgX8DYy9h2weABwBCQkJcyx/UppKSEn7duIVOQOb+bUBPigtNdXLuSrpuN+H7+TR+e/8/lLVtc97tAgICCQgIIB84nHOU6U/8jcbdeuEfl1Bnba0LJSUldfr+C5dGXJeG6Xq/Lj4+PhQXF9d3MwBnj5NGo6G4uJiSkhKsVitjx44lIyODsLAw5s2bh8Fg4MCBA0ycOJH8/HwMBgMfffQRp0+fZunSpaSnp/P6668za9Ys1q1bx4wZM7BarTRp0oSpU6diNBpZvHgxzz33HMXFxfTu3ZuHH36YoqIiJKl+F1O+1tntdrf9LFn8zLqfAAAgAElEQVQslsv6d+eOsHUCaFTl+8iKxyp5AS2A9IoflFBgmSRJAxRF2VT1QIqiTAWmAqSmpippaWluaN6Fpaen06ldR/gNmsdFsWq9A62spy7OXUm56SYOrfqJkJ9/JubJJy74D6pbt27Mnj2bTEnCz2bi8Krl6C0l3DzuQfSennXW5tqUnp5ep++/cGnEdWmYrvfrsmfPHry8vADYt++fFJfscevxvTybEx//0iVtq9Pp0Ol0eHl54enpycGDB5k/fz4pKSkMGTKEH374gZEjR/LEE0/w2WefERcXR0ZGBhMnTmT16tXccccd3HbbbQwePBiAyMhIHn30UQBefPFFFixYwKOPPsrJkydJSEhwvW5fX1/Ky8sJDAx062u/0RQXF7ve06ul1+tp3br1JW/vjmHEjUCcJEkxkiRpgWHAssonFUU5oyhKoKIo0YqiRAO/AecErXpVMUEemxlFY0Mpr9vfHiSVioAJEyjbu5eSiyRltVrNsGHDCAsL46TWk4R+g9i7YR0znnyQvevXoihK3TRaEAThBhcTE0NKSgoAbdu2JTMzk5KSEjIyMrj77rtJSUlhwoQJZGdn17j/zp07uemmm0hOTmb27Nliwerr2FX3bCmKYpMk6RFgJaACpiuKskuSpNeBTYqiLLvwERoAV9gqA40DxVr3XbU+t/Unb8oU8j75BM+0tAv2bul0OkaOHMn06dPZfuIkA558kW1fz+G7Dyeza91qbrn3b/gEi2UjBEG4vlxqD1Rdqbp+rUqlwmw243A48PHxYdu2bRfdf+zYsSxZsoRWrVoxc+ZM17BUREQEx44dIzIyEpvNxpkzZwgICKitlyHUAbfU2VIU5XtFUeIVRWmqKMq/Kh57uaagpShKWoPq1QJQqUFWg82CpFWQrKo6b4Kk0RBw//1Ytu+g9OefL7q90Whk1KhRGAwGvl+TTu8nn6fHmPs5sXc3M//xML8umovV0vBu/BQEQbieeXt7ExUVxcKFCwFQFIXt27cD4OXlVW3OUHFxMWFhYVitVmbPnu16fMCAAXzxxRcALFq0iJtvvlnM17rGiQryldR6sFqQtaCyabDXQ2kF3zsHoYmI4NQHH17ScKCPjw+jRjkXtf7qq9k07ZLG2Pf+j5jWbdmwcDbTn5jArrU/nfcuR0EQBMH9Pv/8c6ZNm0arVq1ISkpiaUXh6mHDhjF58mRat27NwYMH+ec//0mHDh3o0qULCQlnb3S69957yc/PJzY2ln//+99MmjSpvl6K4CZSQ53jk5qaqmzaVPsdYK7Jpe80gcSBfLLjFrIKcnjqzcF4a71r/fx/Vfj1N2Q//zwRH32Id69el7RPVlYWM2fOxNvbm3HjxuHh4cHxPTtZO2saOQf3ExzdlJuGjyaqVZtr4rej633C77VKXJeG6Xq/Lnv27KF58+b13YzL4s6J2IL7uPO61PRzKUnSZkVRalwhR/RsVVIbwFaGWiejseswWetn8WefAbejjYkh78MPUeyX1rsWHh7OiBEjKCwsZPbs2ZSVlRHZvAUj3niPfo8+haW0mMVvvcLclyeSuX2LmEQvCIIgCHVIhK1Kah3YzGj1KjR2PSXlJfXSDEmtJujRRyjbf4Ci71dc8n7R0dHcfffdZGdnM2/ePGw2G5Is07xrGuPe/5Rb7nuYkvx8Fr/5sghdgiAIglCHRNiqpNaDrQytXoPWrqPUVlpvTfHq0wddQgKnpnyEYr30pYOaNWvGHXfcweHDh1m8eDH2ip4xtUZDq159Gf/B1Gqha/bzT7J3/Vrs51mXURAEQRCEqyfCViWNHqxm9AYtGruOkrL66dkCkGSZoMcew3rkKIVLllzWvikpKfTu3Zs9e/awfPnyar1XVUNXr/sfodxs4rsPJzPt7/ez6duvKTPVz9CpIAjChYheeKEhuZKfx1pfrueaUdGzZfDUImOhxFK/wcOzRxr6Vi3J+7//4nPHHcha7SXv26lTJ8xmM+vWrcNoNNLrLxPt1RoNLW/pQ/LNt3Jo60Y2Lf+GtV9N59fFc0lKu4WWPfsQ2CjK3S9JEAThsun1evLz8wkICLgmbvARrm+KopCfn49er7+s/UTYqqTWg+UMRoMeKKK4tP6GEcG5+HTw449zdNx4Tn81m4Dx4y5r/x49emAymVi/fj1Go5EuXbqcew5ZpmnbDjRt24Gcg/vZ/N0Stv+wgq0rviW8WSIte/YmvmMXNLrL+6ESBEFwl8jISI4fP86pU6fquymXzGKxXPaHsVD73HVd9Ho9kZGRl7WPCFuV1HqwncTDaADAZK7/gqAenTrh0e0m8j75BJ9BA1H7+V3yvpIk0a9fP8xmMz/++CNGo/GC6ziFNo2j/2MT6TH2AXav/YkdP63kf//3PmtmTiWhSzcSuqYREd8c6TJXuRcEQbgaGo2GmJiY+m7GZUlPT7+sdfOEulGf10WErUpqHdgseHoYATCVmuu5QU4hEydy6I6B5P33v4Q+//xl7SvLMoMGDcJsNrNs2TIMBkO1wnk1MXr7kHr7nbS9bRDH9+xkx6r/sWvtarb/uAKvwCASunSneZfuBEVdW//5CYIgCEJ9EWGrksZZZ8tgcK51ZbaU13ODnHRxcfgOHszpOXPxHzECbXT0Ze2vVqsZOnQoX375JQsXLmTUqFFEX8IxJEmiUWIyjRKTKTebOLApg72/pLPp26/ZuHQRgY2iaNbpJmLbdyIgsrGYSyEIgiAI5yHGhCqpdWA1o9E710UsMzWMsAUQ9OgjyFotue+9d0X763Q6RowYgZ+fH3Pnzj3vCvTnozUYSbypB3c+9xoPfjqLnuP/htbowfoFX/HFUw8z/e8PkD5rGsf37sJRD8scCYIgCEJDJnq2KlXcjegKW5ZLr29V29RBQQTcfx+nPvgQ08aNGNu1u+xjeHh4MGrUKKZNm8ZXX33F+PHjr2gVeaO3Dym9+5PSuz8lpws4uCmDA5t+Y9v/vmXz8m8wePvQtG17mrbtQOMWLdEajJd9DkEQBEG4noiwVUmtr6gg73xLrGUNa/Fm/7FjOT1vPiffmUz0/HlXNFG9cuHq6dOnM2vWLO69996rWifK08+fVr360qpXX8pMJjK3b+bAxt/Y99t6dq75EVmlJiIhkehWbYhJaUtg42gx3CgIgiDccETYqqTWg8OGRuMsVma1NKzhMNlgIOiJx8l+9jnOLFuG78CBV3ScoKAgRo4cycyZM5k1axbjxo3DYDBcdft0RiPNOt1Es043YbdZOb5nF5nbt3Bk+xZ+njOTn+fMxMPP3xW8GienYPAUC7UKgiAI1z8RtippnLU3tGpnyLKXN7yKxT4DBnB67lxyJ7+L1803o/L2vqLjREREMGzYMObMmcOcOXMYNWoU2ssomnoxKrWGqOQUopJTYOR4igvyOLJ9K4e3b+Hgxt/Ylb4KSZIJaRpLdMvWRCW3Jiy+GSq1xm1tEARBEISGQoStSmpn2JIdZThUdhzlDW+4S5JlQl9+mcy7h3Dqw48IffGFKz5W06ZNufPOO1m4cCELFy5k2LBhqFQqN7b2LC//QFr06EWLHr1w2O3kHNzn7PXasY2MJQv57ev5aPQGGiUlE5XcGovFiqIoYshREARBuC6IsFVJ7Sz5gM0CGjs0wLAFYEhKwm/YME7PmYPvnYPQJyZe8bGSkpIwm80sX76cpUuXMnDgQORaLloqq1SExzcnPL45ne++B0tpCcd27eDIjm0c2bGVQ5t/B+Doj98S1bI10a1a07hFKwxeV9aLJwiCIAj1TYStSuqKeUs2C2gUsDbcqhhBf3+Mov/9j5zX/0nUnNlXVdU9NTWV0tJS1qxZg9FopHfv3nXao6T38CSufWfi2ncGoPBkDj8umo/OUsr+jPXsXPMDSBIhMbFEtUwhumVrwps1F0OOgiAIwjVDhK1KVXq2ZK2CbNNgc9hQyw3vLVL5+BD81FNkP/88Z75Zgu9dd17V8bp164bJZOK3337DaDTSrVs3N7X08vmGhBKU1Iq0tLSKIcf9HNmxlSN/bGXjssX8vmQhGp2eyMQWzvleLdvgHxEphhwFQRCEBqvhJYn6UjFnC5sFlU5CW6Kj1FqKj86nftt1Hj4D76Bw4UJy330Xr543o/L1veJjSZJE7969MZlMrF69GqPRSGpqqhtbe2WcQ44JhMcn0GnwcMpMJo7t/oMjO7ZwZMdW1mzdBICnfwBRLVs7/ySnYPRumNdMEARBuDGJsFWp4m5ErBbUWhmNXYfJamqwYUuSZUJfeZnDdw3m5OTJhP/rX1d1PFmWGThwIGazme+++w6j0UjiVcwHqw06o5HY1A7EpnYA4EzuSY78sZUjO7a57nIECI5uSlSr1hVDjomoNWLIURAEQag/ImxVqtKzpdGr0Dh0lFhL6rdNF6FPSCBg/DjyP/sc73798OzS5aqOp1KpGDJkCLNmzWLx4sXo9XqaNGnipta6n09wCC179qFlzz44HHZOHjrgmmi/efk3bFy6CLVWR2RiC6KSnfO9AhpFiSFHQRAEoU6JsFXJFbbK0OqNaOx6Sq2l9dumSxD48MMU/7iKnJdepsm3y5A9PK7qeFqtlhEjRjBjxgzmzZvHmDFjiIiIcFNra48sqwiLbUZYbDM63jmUcrOJY7t3Oud77djK2lnTWAt4+PkTlZxCTEpbYlqnojNe3fslCIIgCBcjwlYlV9gyo9P7oLHrromwJev1hP3rDY6MHEXu+/+5qtpblQwGAyNHjmT69OnMnj2b8ePHExgY6IbW1h2twVixRmN7AIryTjmHHLdv5dDWTexetxpZpSIyMZnY1A40Te2Ad2BwPbdaEARBuB6JsFXJdTdiGXqDFo1dR0l5ww9bAMa2bfEbMYLTs2fj3a8vxjZtrvqY3t7e1dZRHD9+PD4+DXP+2qXwDgwiucetJPe4FYfDTva+Pzm4OYMDmzJYPeNTVs/4lKDoJjRt65wTFhzTVAw3CoIgCG7RcItJ1TXN2TpbRqMeGZkS87URtgCCn3wCTVgY2S+8iKOszC3HDAgIYOTIkZjNZmbNmoXJZHLLceubLKuISEik2z3jGP/+J4x7/xO63TMOrV5Pxtfz+eq5x5n68DhWTfsvmds2Y7Na67vJgiAIwjVM9GxVquzZslrwMDqDV2mpuR4bdHlkDw9C//k6x+69j1MffEjI0xPdctywsDCGDx/OV199xezZsxk9ejQ6nc4tx24o/MMj8R8QSbsBd2EqOsOhLRs5uCmDXWtXsf2H71xDknEdOhPdqg0anb6+mywIgiBcQ0TYqlTlbkRX2DJdO2ELwLNLF3yHDaVgxgw8u3XDo2MHtxw3JiaGwYMHs2DBAubPn8/w4cPRXKflFIzePrRIu4UWabdgLS/j6B/bObDxVw5symDPL+modTqapKQS16EzMa3boTMa67vJgiAIQgMnwlalKmFL76UFwGIur8cGXZmQp5/G9FsGWc8+S5OlS1C5aZ5V8+bNGTBgAEuXLmXBggUMHToUtfr6/vHRaHWuSfa97HaO7f6D/RkbOLDxV/ZlrEel0RCVnEJchy40Te2AwdOrvpssCIIgNEBu+bSUJKkP8AGgAj5XFGXSX55/ErgPsAGngPGKohxxx7ndRpJApQObBa1eBYDZ7J65T3VJNhoJnzyZzOHDyXntNcLfe89tE71bt26N3W5n+fLlLFy4kCFDhqBSqdxy7IZOVqmISk4hKjmFnuMf5MS+PezP2MD+3zdwaMtGZJWKRkktiWvfmdh2HfHw9avvJguCIAgNxFWHLUmSVMDHQC/gOLBRkqRliqLsrrLZViBVURSTJEl/A94Bhl7tud1OowdbGRq9820ps9jquUFXxpDcgqBHHuHUf/6DZ1oaPgMGuO3Yqamp2O12VqxYwaJFixg8ePANE7gqSbJMZEISkQlJpI2+j5MH97Pv9w3sz1jPqs8/5qdp/yWieaJrgW2vgGurbIYgCILgXu7o2WoPHFAU5RCAJEnzgDsAV9hSFGVNle1/A0a64bzup9aD1ezq2bJeo2ELIOD++yj5+WdyXv8nhjZt0Ua6rzBphw4dcDgcrFy5km+++YY777wTWb4xb2yVJInQ2HhCY+O5afgY8o5msi/DGbzWzJzKmplTCYttRlyHzsR16IJvSGh9N1kQBEGoY+4IWxHAsSrfHwcuNDP7XmCFG87rfmqds2dL53xbrGWOem7QlZNUKsLffpvDAwdy4h9PEj1rFpJW67bjd+rUCbvdzqpVq1zrKt6ogauSJEkERcUQFBVDlyH3UJB13DXUuG72DNbNnkFQVAxxHToT36EL/hGNRC0vQRCEG4CkKMrVHUCSBgN9FEW5r+L7UUAHRVEeqWHbkcAjQHdFUc6ZECVJ0gPAAwAhISFt582bd1VtuxQlJSV4enoC0O73hyn1aMwfcU+z92uF3U3WcHf7nrXehtqk27IF36mfYerRg+KhQ9x+/MzMTDIzMwkNDaVZs2ZuCQ9Vr8n1oqzoDIWH93P60H5Kc04AoPP1x69JHL5N4jAGhjT44HU9XpfrgbguDY+4Jg1TbV+XHj16bFYUJbWm59zRs3UCaFTl+8iKx6qRJOkW4AXOE7QAFEWZCkwFSE1NVdLS0tzQvAtLT0/HdZ69/nh4e9PjljT2fr0GtaSjLtpQq9LSOFlWBl98SdwdA/Du08ftp1izZg1r164lJCSEO+6446p7uKpdk+tQyekCDmz8jf0Z6zm2bSM5WzLwDgomrn0n4tp3ITw+AakB9hJe79flWiWuS8MjrknDVJ/XxR1hayMQJ0lSDM6QNQwYUXUDSZJaA5/i7AHLdcM5a4daDzYLsiyhqO0o117lhxoF/+MfmLdtJ/uFF9E1a4YuJsatx+/RoweyLLNmzRrsdjuDBg264SbNXw5PP39Sbu1Hyq39nEVUN//O/t83sG3ld2z+bikevn7EtutEXIfONEpMRhbvpSAIwjXtqsOWoig2SZIeAVbiLP0wXVGUXZIkvQ5sUhRlGTAZ8AQWVgyVHFUUxX23yLmLRg9WS8XfHWC9Pj7kJK2WiP+8z+FBd3Li748TPX8essHg1nN0794dlUrFqlWrsNvtN+RdilfC6O1Dix69aNGjF2UmE4e2bmR/xnp2rfuJ7T9+j97Ti6apHYjv0IXGySmor9NisoIgCNczt9TZUhTle+D7vzz2cpW/3+KO89Q6tR4sRQBIWgXZqsbmsKGWr/3inZqwMMInT+bYAw+Q/corhL/9ttvnCHXt2hWVSsXKlStZsGABd99993Vf+NSddEYjzbt0p3mX7ljLLGRu3+KcYJ+xgV3pq9AaDDRp41w2KKZVWzR6sWyQIAjCtUB8ElaldtbZApC1oLHpKLWW4qNzTxX2+uZ5U1eC/v4Yp/7zAfr4eALuu8/t5+jUqRMqlYrvv/+e+fPnM2TIkOt2aZ/apNHpXXW67DYrR//Yzr6MDRzY9Bt7169FrdUR1TKFpm070KRNO1FEVRAEoQETYasqtR5szvUQ1ToZbdn1FbYAAiZMoGzffnLf+zfaJk3xurmH28/Rvn17ZFlm+fLlzJ07l2HDhqF1Y9mJG41KrSGmdSoxrVPpZX+Y43t2sf/3DRzcnMHBTRkAhMU2o2lqB5q2bU9Ao6gGf2ejIAjCjUSEraoq6mw5/yqjsesptZbWc6PcS5Ikwt78F+VHj5L11FNEzZuLPj7e7edJTU1FpVKxbNkyvvzyS0aMGIFRLNp81WSVisYtWtK4RUtuHjeBU0cOc2jz7xzcnMEv877kl3lf4h0UQtPU9jRt24HI5kmo1KJnURAEoT6JsFWVxgBWZ8+WVq9GY9ddd2ELQNbrifx4CpmD7+b43x4ieuEC1P7+bj9P69at0el0LF68mJkzZzJq1Ci8vMRize4iSRLB0U0Ijm5Cx7uGUVKQz6EtGzm4OYM/Vq1k64pv0RqMxKS0pWlqB6JT2orFsgVBEOqBCFtVVenZ0uk1aOw6Sqwl9dyo2qEJCSHy4ykcGTWa4w8/QuPp09x+hyJAYmIier2euXPnMm3aNEaPHo1/LQQ7ATz9A2h5Sx9a3tIHq8XCkT+2cXDz7xza8jt//vozkiQTGhdPTKu2xKS0JaRJbIOs5yUIgnC9EWGrqoo6WygKeoMG7XU4jFiVoWVLwie/w4m/P86JJ54k8qMPkWphMnuTJk0YM2YMs2fPZvr06YwcOZLQULFGYG3S6PXEtutIbLuOKA4H2Qf2cXjbZjK3b2bDojlsWDgbg5c30a3aEJ3SluiWrTH6+NZ3swVBEK5LImxVpdYDCtjLMRj0aBw6SsoL67tVtcr71luxv/IyOa++RvbLrxD25r9qZXJ1ZGQk48aNY9asWcycOZMRI0bQuHFjt59HOJcky4THJxAen0CXIfdgKjrDkR1bydy2mcPbt7Dnl3SQJEJiYolJaUN0SiphsfGimKogCIKbiLBVlbqibpHNgtHo/HtpqakeG1Q3/IYNw5aXT96UKagDAwn+x5O1cp7g4GDuvfdevvzyS7788ksGDx5MQkJCrZxLOD+jtw/Nu6bRvGsaisNBbuYhDm/bzOFtm8n4ZiG/fT0fndGDyMQWNE5qSeMWrQhoFFXfzRYEQbhmibBVlaYybJXhURG2SkzmemxQ3Ql8+CFseafI/+wzVH5+BIwfVyvn8fX1Zfz48cyZM4f58+fTt29f2rdvXyvnEi5OkmVCmsQS0iSWjncOxVJSwpE/tnH0j20c3bndVVrC6OOLLigUf7uFxi1S8A0Rw8CCIAiXSoStqip7tqxm9AYdABZzjWtmX3ckSSL0pZewF54h9513QJYIGDu2Vs7l6enJ2LFjWbRoEd9//z1nzpyhZ8+eV72AtXD19J6eNOvUlWadugJQdCqXozu3c3TXDg5s/p0fp04BwDsohMYtWtIoqSWRCUl4BwXXZ7MFQRAaNBG2qlKf7dnS6D0AMN8gYQtAUqmImPwOJxSF3ElvgwIB48bWyrm0Wi1Dhw5lxYoVrF+/nsLCQgYOHCiqzTcw3kHBrrUb16xZQ8u4WI7u2s7RP7az//cN7FzzIwBeAUFEJCQSkZBEZEIiAZGNxZ2OgiAIFUTYqsoVtsxodc7JwRaLtR4bVPckjYaIdydzQpLIffttUJRaG1JUqVT0798fX19fVq1aRUlJCUOHDq2VcwlXT5IkAiIbERDZiNa9b8PhsHPqSCYn9u7mxJ+7Obb7D/auXwuA3sOT8GbNiUhIIiIhidCmsaK4qiAINywRtqqq1rPlDFtl5vJ6bFD9kDQaZw8XOIcUHfZaWUcRnB/gXbt2xcfHhyVLljB9+nSaNm1aK+cS3EuWVYTENCUkpilt+t6OoiicOZnDiT93c3zPLk7s3cWhLRsBUGk0BEc3ISy2GaGx8YTFNsMnJFQsKyQIwg1BhK2q1M55WtgsaI3Ot8ZsuvHCFpzt4cqSZXLffQ9bXj7BT0+staGh5ORkvLy8mDdvHps3byYhIYGYmJhaOZdQOyRJwjc0DN/QMJK69wTAdKbQ1fOVc3AfO35ayZYVywDQe3pVBK94QmPjCW0aj9H7+lmHVBAEoZIIW1VpKiqoWy1ne7YsN2bYApDUasInv4PK35+CmTOxnTpF2FtvItfSotLR0dHcf//9TJs2jVmzZtG3b1/atWtXK+cS6obRx5e4Dp2J69AZAIfdTt6xI+Qc2Ef2gX3kHNzHb4vnoygOAHxCQiuWIGpKcHQTgqJj8PQLED1ggiBc00TYqqpKz5amYs5WucWOoig37H/2kkpFyAvPow4J5tR7/8ZWkE/kRx+h8vSslfMFBATQpk0bTp48yXfffcfJkyfp27cvKlFg87ogq1Su9Rxb3tIHgHKLmZOHDpBzYB85B/aRe+QQ+zM2uPYxeHkTVLFP5R+/sAhRdFUQhGuGCFtVqSt6tmxlFWFLQbKqMNlMeGg86rVp9UmSJALvvx91UBDZL77EkRH3EPl/H6ONjKyV86nVaoYPH85PP/3E+vXrycvLY8iQIRiNxlo5n1C/tHoDjRKTaZSY7HqszGTi1NHDnMo8RG7mYXIzD7J1xTLsNhsAao0W/4hGBDRqTEBEIwIaRREQ2Qif4BBkWYQwQRAaFhG2qnL1bJmRJAlJC1qHngJzwQ0dtir5DhyIOjCIE08+Sebgu4n4z/t4dOxYK+eSZZlevXoRHBzMsmXL+Oyzzxg2bBghISG1cj6hYdEZjUQmJBGZkOR6zG6zUZB1nNzDBzl15BB5x45ybPcf7Pl5jWsbtUaLX0QkARGNCGwUhX9kI/zDI/ENCRV3QwqCUG9E2Kqqyt2IACqdhMauI9+STyPvRvXYsIbDs2sXYhbM59gjj3D03vsIeXoifqNH19owa6tWrfD392f+/Pl89tln3H777bRq1apWziU0bCq1mqDG0QQ1jgZ6uh4vM5WSf/wY+SeOOr8eP8qJP3e7ylAASJKMd1AQvqHh+IWF4xcajm/FV++gEFRq8V+hIAi1R/wPU5XmbAV5AK1O5Qpbwlna6Gii580n69lnOPnWJCy7dxP68svIHrXT+9eoUSMmTJjAokWL+Oabbzh27Bh9+vRBLT4gBUBn9HAttF1VudlE/oljnM7O4nR2FoU5zq97fk6nzFTq2k6SZXyCQvANDcM7KBjvoBB8goJdf/fw8RUFWgVBuCri06qqv/Rs6QwatMV6CiwF9diohknl6UHkhx+S98kn5E35GPO27YT/+z0MSUkX3/kKeHl5MXr0aFavXs369evJyspiyJAh+Pr61sr5hGuf1mAkLLYZYbHNqj2uKArmojPOEJZzNoSdyT3JyUMHMBcXVdtepdHgHegMXz5BIc4QFhiEp3+A649Gp6/LlyYIwjVGhK2qZDVIMtgsABg99ehOGykwi7BVE0mWCXroIYypqWQ9/QyZw4YT/MQT+I8dUys9ASqVil69ehEZGcmSJUv49NNPGThwIM2aNbv4zoJQQZIkjD6+GH18iUhI/H/27js8joSn2NcAACAASURBVOpc/Pj3zHZVq7nJsuWKK+4GYwM2JZhOICSUhBIICSUh3F/ITbk35CYkIeUmJAFCSy4QAoSQgMEQErAtmsEF94Jx77ZsS7Lqasuc3x8zu16tVs3a1UrW+3meeWbmzJk5Rzur3XfPnDnTbHvA30D14fLodOzwIXv5EFt2bqeh+lizfTyZmWTl2cFXXkGTQCwrL5+s/AIycnLlDkoheikJtmIpZd2RaAdbvkw3vnAmFf7taa5Y95Y5YwbDXnmZA//935T/4hfUvf8+A378I1zFxSkpb8yYMfTt25cXX3yR559/nhkzZnD++efLcxVFUri9PgpLhlBYMiTh9qDfT/XRw9RWHD0+VVbY86Mc3bubusrK6NhhsbzZOWTm9iEjJxdfbp/ociT4y8jNJSOnDxl9+uDyeHvtkDNCnGwk2Irn9ESDLU+GC084kwNyGbFNjj59KP7d76j664sc+sUv2HbpZfT9j/8g77prU9LKVVBQwK233srbb7/N0qVL2bVrF1dddRV9+/ZNellCxHJ5vdZwE8Ut3zRjmmHqq6qorThKTeVR6ioqqK+uov5YFfXHjlFfXcXhndvZdayqSf+xWE63B292Nr7MLLzZOXizsvBmZePNysZnz71ZWdTs38OR3Tuj25wpGnRYCHHiJNiK5/LFBFtO3EEvO4/tTG+degilFHnXfIGsM2dz4L4fcuj++6l+4w0G3P9jPMOGJb08l8vFhRdeyPDhw3nllVd4/PHHmTdvHlOnTpUWAZFWhuGIXkbs30beUDBIQ/UxOxCror76GHVVldRXH8NfWxOdKvbtpaGmGn9tLWY41OQYn87/a3TZ6fHgzcjE7cvAk5GJOyMDjy8Dd0YmnoyMaLonI8Pelok7w2enZeL2+XC6PfI/JEQSSbAVz+mB4PFgS2mDnUd3Ux+sJ8Mlg2q2h6u4mJInHqf61Vc59NOfsf3yKyi48QYKvva1lIw8P2rUKG6//XZefvllFixYwKeffsoll1xCTk5O0ssSItmcLhfZBYVkFxS2K7/WmmCjH39tDQ01NSz94H1OGT48GpQ11FTTWF9PoL6OxoZ6AvX11Bw5TKChnsb6eoKN/jbLUIaBy+PF7fXi8npxeXy4vB5cXh8ujwe314czst0TyWOtO71e3JH8Hi9Otxun22PNXW6cbrfc3Sl6HQm24jm90ZYtb4bVB8gZ9LDx6Eam9Z+Wzpr1KEopci+/nMxZsyj/9W84+uQfqXplPn3v+Sa5n/1s0j9ss7Oz+eIXv8jSpUtZuHAhjzzyCBdeeCGnnnqq/EIXJxWlFG6vD7fXZ90luWsPp8yc3e79zXCYQEMDjfV1NNbXRYMwKzhriKYFG/0E/Y0E/Q3WcmMj/ppqao40EvA3EGy0toWDwQ7/DQ6nMxqAOewALBqUuVwJA7RImsPe7nA6cThd1tzlwnC6cDqdGE4XDtfx9Ph80fwOp3w2iC4jwVa8mGDLk2G9PJ6Qj7VH1kqwdQKchYUM/OlPyLv2Gg795Kcc+P5/UfmX5yi65x4yZ89K6oedYRjMnDmTkSNHMn/+fF5++WU2btzIJZdcQnZ2dtLKEaInMxwOu/9XclqZzXA4GngFG/0E/H6CjX5Cfj+BRj+hQCBmarTmwQDhBGmRfP7aWis9GGiyf/zl085yxAdncUGZ4XTicDgxHI6YyUo3DMPKE5tuL+/bu48lh/dH1x1OJ8pw4HA2zWctGxgOJw6HA+VwWOU1y+dAGQaGYaAMhz03mqY7jqcrZUgg2c1IsBXP6T0+zlam1bJV4i5lVfmqdNaqx/NNmMCQ55+jesHrHP7Nb9jzla/gmzaVvnffTcb06Uktq7CwkJtvvpmPPvqIRYsW8fDDD3PBBRcwadIk+QASIskMhwNPRgaeLnh2qWmGCQUChINBwqHQ8XkoiBkKEQoGMUNBKz0cSpiv+X5BQsHQ8f3i8kXKNM0wZiiMGQ5hhsOY4TDhcAgdDhMOR9JNzHCIcCjEwZUfpfz1aE2iIMwwEgdnLaa3EeRF0qLbHUaTdKWUvR4JAlXTgDAuT5N9VMx+hrLqGN2mUIYDZRzPaxhxx7PTMKwy0z0WngRb8ZweCNQCx1u2JmRP4m/7/yT9tjpJKUXupZeQfcFnqHrpJY7+4VF2fekGMs84g4KvfpWMGdOTFgwZhsEZZ5zBqFGjmD9/PvPnz2fVqlVcfPHF8nxFIXoow3Dg9vrA60t3VVpVVlbG2WeddTxAM8OEQ5Eg7XhQZoZCmKaJGQoRDoftwO14MBcN7EwTHZmbJmY4bM1NE212MD32OGai9OPbtRnGDJuEgiG02YgZjkmPLSdBOpHtWke3aa2tbVp3+TnJLy5h6GVf6PJyI5ISbCml5gG/BRzAk1rrB+K2e4BngKnAUeALWuudySg76Vw+qDsCHA+2xmdP5M+H/by7913mDZ2XztqdFAy3m/zrrqPPZz9L5fMvcPTJJ9l94414J0yg4NZbwZW83wCRVq7Vq1fz1ltv8dhjj3H66adz9tln4/F4klaO6MZME8xQG1O49fVwzLIOgzbtSccs25MZjktLkKdD21ufTtm/H6r+BuiYL7FEy/Z6R5ZjvxSTeuwEEv7QUj0sj7U+/sgR1IHHcQCO9vyAbOE4x9NbW4/bppT1TezsyL4dKVeBcnWwjk3Xtf220UqhTY3GuvFDozC1tt7a2Hmw3+qRbVrZea3tpv3WMiPHiBzb3te0h7tzZuWxlfTp9LeaUsoBPAycD+wFliulXtVab4zJdgtQqbUeoZS6Bvg5kL4QszUx42xFOsj3dw6kyFfEP3f8U4KtJDJ8Pgq+fDN5113LsVde4eif/o99d99NQVERR3fvps8VV+BIwuN4DMNgypQpnHLKKbz99tssWbKE9evXc8EFFzB27NjecWnRNK0gIRo8BNsOMJqsB8mrWAmfNrZ/n4RltF1Ox+oVFwgl2k4LX+7dkrKeYpFoMhKn5zcGoN5u6Wn2BRr3xRr7xdfu5Za+nDt57CYSnKNmQVk3zxOz6vXXQlV9gjyJ3ost5YkPThOsdyRvk6JOZN/Olnt8m/WO6Ui5J1DHeIWnsDX/9ObpXSQZTQgzgK1a6+0ASqkXgMuB2GDrcuCH9vJLwENKKaV1GtoS2+L0RftsubzWNeFAfZiLh13Msxuf5UjDEQp97btFW7SP4fWSd8019Ln6amreepsdv/sd5Q/8nMO//g3Z8y4k9/OfxzPhVOtXTFijTY1pasywPr4ettKaLIet5vnYaXzf8RROLWDZJ8v529/+RkF2PqcOGkueN8fKH7ab3CPLpokO62izeuRnkjatVgVtWh882m6hsJrMIy0V1jqRNGLWo/k4ni+6nWi+Ji0HkVaQSFrC7cRsN2O2t669/4j73vk4QWp7glUXKLf9ZWs0/YJOuGzlUfHbmnyx252AlWp63NjJefxYYB9PxeZrYRvKCm6Iy6fsu2iVQsUGI5HgI6Z1ocPb44P+uPVmr7KdcODAAQYUDUiwXTXL2/Kx2ldWe/dvO3/8ahs7dKD8Nsvq4B/X9vGav/937drJkOLSdh5PtbbaTKv1b/Nvjd/eVob41TZeqy783druH8n2558nwwmBT1NYo9YlI9gqBvbErO8FTmspj9Y6pJQ6BhQAR5JQfnI5PRBqAKyT6fE5aawPceXIK3lqw1PM3zqfWybckuZKJhYOmgQbwwQDYUKBMKGASShoYoZMKzgxtR046LigxWwexMQFLmY4Zl+zheUEx2lp+XhAFHscEzPsITT0HihVmKaGagOerALeTepr5eJUsnwHqQzvYnHN+3gaisisLcURbtoXRGFiELZ+hSnT/mzRKDSxH+8Kw0pX2t6u7C3a3jeSE1D6+HdvwvXYhgMVn3D8Czq6bG8zImUmOlD8csx+Ku64LezT4G/A58tIUH5bx2jvh2LrCW3FjE1+1Gpa3aHtY7XcatF0NTbgbUfdWjxe6+W1drzGRth99GjLu7d16DTWPVFCG2+D1uvfZt741TZ26OBrF9twdHjjzvb/ihEpl9c/g4Fz0ld+t+ogr5S6DbgNoF+/fpSVlaW8zNra2ibljDh4hP7+Ot6300xlsnvnPvQqgxGeETy79lmGHhmKoVI/KJ/WGjMI4YDV2Bby21MDhBv18XV7Mjs+3E37xf+4V3HLsY0VkeWYBgQMcOgADt2I0/TjMRtwmvW4zHocZgNO04+TRpQRxvCEMZSJIowywwSqDBorHITrFEqHceVo3EXgLnKg3FbrI4Z9mcVhzbXhAPvuFAzQ9h0tGFhpytonBGytdrBNHaYq4zBD89yMKPLidTujx9DKQCunPXdE5xBZd8Skx+XpgvdJV6mtNbFGC4hp1hdpV1tbT1bWyfM+61qdbYpJvH9tbS1Z7Rhao82LOx34N+vwdaIO5k/Jdaiu/BhRDdTW1nVJXJFIMoKtfUDsQ8IG2WmJ8uxVSjmBXKyO8k1orR8HHgeYNm2anjNnThKq17qysjKalBMqgwNvRtOqVq6kpsLPmbNOp25PHd9997sYfb2UhkdRW9VIY32QQEOIUKD5Q2eBxP+LGquVR4MOm4TDmkBDiMb6EI0NIQL21NgQavHN6MlwkpHjJquPm4wcN74cNxnZLlxeJy63A6fbwOl24HQZGE4Dw6EwDIXhUPZttM2XDYeBYShUJK+9bAVMHfhQCofg0HrYvxIOrodDG6B8IzRWW6+HYb8wuYMgtwSy+0FmEWT2g6wi1u04xITTzgFvLniywZsDrgwad+6k+vU3qH7tNQLv7gKXi8wZM8iaO5fsuXNO+MHXM4Dq6moWLVrEmjVr2FVtMn36dGbNmtWuD8zeotn/iugW5Lx0P3JOuqd0npdkBFvLgZFKqaFYQdU1wHVxeV4FbgQ+BD4HLOqW/bXAGmcr0knXcDD5giEs+P0a/vXkerQu5uYNP2PDR342sDa6i+FQON2OZldLWvsLrf6udrDjUHh8LjwZTrLzvXgynHh8Ttz23JPhxJdtB1XZbjKy3Thc3eiXbDgEe5fB9ndgz0ewd0V0+Aw8OdBvHJz6eSgaDfnDIK/UCrKciR+Ye7SmDEqaj73lGTqUorvupPDOO/Bv2Ej1G29Qu2gRh+6/n0P334/nlFPImjuHrFmz8E6ciNGBB/Lm5ORwxRVXcOaZZ/LOO+/w0UcfsWLFimjQlZmZeQIvjBBdQ5smhELokDXuE/at/gnnYeuOyebzMNhDALQ4b3HfmLndX9Dq14jdz1G3sK6t28m0NUQAZkx/w2i/Rh2zj70c008ysk/8emyfyib7x65HPqSjl6BjLgnHzKOXG9uTH02figp2Pf10B45Ls23x+TW6c/ni89uZEn4VJ/ruaukLLeH+7UyL1LcLyncPGQw33JD4GF2g08GW3QfrLuBfWDec/klrvUEp9SNghdb6VeCPwJ+VUluBCqyArHty2sMBhBrBncGQcQVMOLuYzUsP4s1yoYbX8q75Kj+6+PsMGtAPT4YTh6sXjtbbUAVb3oJP34Stb4O/yoog+42DidfC4NNh0HToMzjpvSaVUvjGj8M3fhz9vn0vjdt3ULt4MbWLF3P0iSc5+uhjKI8H3+TJZMyYTuZpp+EdPx6jHUM9FBQUcOWVV3LWWWfxzjvv8OGHH7Js2TImT57M6aefTkFBQVL/FtE9aa2t4CUQwAwE0IEgOtCIDgSaTGZ0ufn2ptsCViAUClrHDdqBUTQt3HS9yfbQ8UAqNi0YRIdC9A0G+aSb/nbtkJjL+6qldSPmxgl7W4vr0f3tmxFi15vdlEHTebSPZWxagnzYh4nrQ6kCQXRjoPV8SqGifS6a5rNWE/TNbFIPomnx5SeubxvHjZcwrXnS8WO2I3NL3wVdUL6zb9/EO3eRpPTZ0lq/AbwRl/aDmGU/cHUyyko5p91BOuQHtzWA6VnXnsJZ154CwPaq7fxx/v0sCS/k5j43p6uW6REOwtaFsOZ52PxPCDdCRiGcchGMugCGz7Uu/XUxz7CheIYNpeCWLxM+doz6jz+mfulS6pYt58jvH+LI734PLhfekSPxTpiAb8J4vBMm4Bk+HOVM/C9QWFjIVVddxVlnncWSJUtYuXIly5cvZ8yYMZxxxhmUlJQk3E+kng6HMRv8aH8DZoM16YYGzAY/ZkN9zHKDlae+AdMf2Walm/4GdHx6YyO68XjAlLROKg4Hyu1GOZ3RCZcT5XQdX3c6mqyrTE/TPA4HyuUEZ9P9lMvJrn37KB0+AuV0WMdyOFEOAwxHm3Mc1ojhGG3NWziGPZJ47BylmvSLBKsFv83g6CRSVlbGRLmM2O1sTFN/LehmHeS7hWjLlj/h5mF9hjGl7xRe3PwiN4y9AYfh6MLKpUnlLlj+BKx+HuqPQEYBTL0JJnwOiqfZH6jdgyM3l+xzziH7nHMACFdVUb9iBQ1r1tKwfh3Vb7xB1V//amV2ufCUluIeMRzP8BF4RgzHXVqKq7gYh/0sxaKiIi6//HLOOeccli1bxvLly9m0aRODBg1i+vTpjB07FpfLla4/t1vSgYAdxESCoJhAyO+3gx97W+xyGwFS5Fg6EOhwnZTPh+H1Yvh8qAwfhteH4fPhyOuDa+BADK8X5fVaQZHbhXK7MTwelMttp7mbbmuS5olua5puT47UfkZsLCujSL7YhejWJNiK54pp2WrBtaOv5d537+W9fe8xp2RO19QrHfYsgyW/h08WAApGXwQTr4MR57XY36q7cfTpQ/Z555F93nmA1bclsGsX/vXrafx0C41bt+LfsJGaN//VpCXDyMnBNXAgruJiXMUDcfXrz9SCfCafdhobjh1j1Y4dvPzyy7z55ptMmTKFqVOnkp+fn64/s1201uhg0Ap4Ii1D/ka7hciPbrTSTX9DNE9ssJOzYzv7Xn0tQctQTFDl90Oogw8LNgwrCPJZAZC17MXwZeDK7oeR4UN5I9u8duDks9Jjl73WPobP2+R4yuOxWlqEEN2StvuvNZvby6a2xzeMrGNG+5pF8prajPb/iuwTu57uhhEJtuJFWraCLQdb5w45l74ZfXl207MnZ7C1eymU/Qy2LwZvHzjjGzDjK9bdgz2cMgw8Q4fiGTq0Sbrp9xPYvp3A7j0E9+2zpv37Ce7eTd2HH6Lr66N5C4HzgEP9+rFt9Cksqavjg/ffZ2BtLSMb/JQC7sxMjIwMlMu+5ONyHb8M5HJFLwXFj8ocGdTUSrLTQiErSGoyBdDBSP8fOy1gzc3GRiv4icz9fit48vuPP7uiI1wuDJ8Pt1L4+/RBZWRYrUSZmTgKCxMHQZHAJ1EQZOcxvF5U5DU6yS4jieNfeKY2CekQpjYJa+vZeWEdPr5uz8Nm87TYecgMNVmPnSJfwNFlbUa/kKN54r6oY9M7epy2tu89upeyJWVNjg20u5zYQCM+8OjQPEFaNHCJX48JbJoELXa9Ih3vm6zbf1Ps8VvM00qdusKw3GHc0+eeLikrEQm24jntJ4O30rLlMlxcO/pafrvyt2yp3MLIvJFdVLkUO7QR/v1fsG2hNRTD+T+G6beA++S/E8/wevGOHYt37Nhm27TWmHV1hI8eJVRRQbiigtDRo/StqGBsZRU19XVsCAT41OfjnexsloTDDDl8mNKNGyg8fMRq6bEDok6JBmqR4C3B5HSivF4cRUVW8OPxWK1EXh/K62kyN3xelMdrzb0+DK/HbkGy0yMBkX2ZVG5nT72QGSIQDhA0gwTCARrDjQTMAMGwtR4wA022B8IB1tWuo2JLBSEzRNAMEjJDhHWYkBlqOulQu9Ji9w2awWiesBluFgAlCoJit51MFApDWTdDGRjHl5WBQcyyMggGgny691OU3Xk9kt5k2T5O7H4K1WSfyADFTdYhuk8kL4rosRLtkyhPs3lkOX49rk6R8pust5Gnzb+jo3Vsbd7C35HjzoEdaXjj2CTYiteOYAvgcyM/x6NrHuUvm/7CD8/4YerrlUq15bD4p7DyaWtcq14UZLWHUgpHVhaOrCzcQ4Y0294PGAGYpsnOnTtZs2YNGzduZGv//uTl5XHqqacybtw4ioqKrFvvI61R9uW2aKtOpNNw7J1EKKufkNMpl8LSIGyGqQ/VUx+sxx/24w/5aQg1RJcTrocbosv+kP/4trDfCp7igqVoMGUGTjxAWdLyJody4DScOA1nk2WX4WqW5jScOJU19zq9TdYdhgOHcmAoo8k8UbqhDByGtexUzqb7GImPEZ/WrjyG0TTwaSEIaimwiQ98WgqgIl/Y7SU/TLqnsh1laStbgq147Qy2+nj7cMmwS3h9++t8c8o36ePt/AOTu5xpWgHWW/dBsA5m3AZn/ydkdO++R92VYRgMGzaMYcOGcfHFF7Np0ybWrFnDO++8wzvvvENhYSFjx45l7Nix9OvXTy6dpZDWmoZQA9WBao41HqM6UE11YzU1wRrqgnXUBeuoDdZSH6ynNlgbTasL1jVJa7Af3dURbsON1+nF6/Tic/rwOqxlr8NLpisTt+HG4/DgcrhwO9y4DTduhxuXYa+3kOZxeI6vR7Y7XHy87GNmz5zdJGByKAcuwxUNUIQQ6SXBVrzYcbbacN2Y6/j7lr/z9y1/77bPS2zRkS3w6jdg9xIoPRMu/jUUjUp3rU4abrebiRMnMnHiRGpqavjkk0/YuHEj7733Hu+++y75+fmMHTuWUaNGMWjQIAxptWqR1pqaYA0VDRVU+I9PR/1HqW48HkzFz4NtPL/KaTjJcmWR6cok05VJliuLfG8+g7MHR9MiU4Yro1ng5HP6okFVZN3j8HR5R9zdrt0MyBrQpWUKITpGgq14kbsRg23/oh2VN4oZ/WfwwuYXuHHcjTiNHvByag3LHoe3fmC14l32EEz+YtIHHhXHZWdnM336dKZPn05tbW008Prggw94//338fl8jBgxglGjRjF8+HAyMjLSXeUuEQgHKK8vj06H6g9xuP5ws4Cqwm/1R0oky5VFjjuHXE8uOe4chvcZ3mQ9fp7lzooGWG5Hz7ijVgjR8/WA6KCLdaBlC+D6Mddz9+K7Wbh7IReUXpDCiiVBzUF45Q6rA/zIC+Cy31vPJRRdJisri2nTpjFt2jQaGhrYtm0bn376KVu3bmXdunUopSgpKWHYsGEMHTqU4uJinC0MvNqdhc0w5fXl7Kvdx77afeyv28+hukNNgqvKxspm+3kcHgp9heR78+mb0ZfR+aPJ9+aT780nz5tHgbeAfJ+97snD5ZAxzoQQ3V/P+xRPNWfb42zFOnvQ2RRnFfPsxme7d7C1vQxe+jIE6uHi/4Vpt0hrVpr5fD7Gjx/P+PHjMU2Tffv2sWXLFrZs2UJZWRllZWW4XC4GDx4cDb769+/fLS45aq2pbKxkd/Xu4wFV7X721u5lX80+DtYdJKSbtkZFAqh+mf2YUDSBvhl96Z/Rn74ZfaNTjjtH+rIJIU46EmzFa2ME+XgOw8EXx3yRny//OavLVzOp76QUVu4EmCZ88BtYdD8UjISb/wxFp6S7ViKOYRiUlJRQUlLCOeecQ0NDAzt37mTHjh3s2LGDt956CwCv1xvNV1JSwsCBA/G045mPJ6oh1MDu6t3srN7J4qrF/Pu9f7Ozeic7q3dSE6hpkrfAW0BxVjETCidwQekFFGcXU5xZTHF2MQMyB8hlOyFEryXBVrx23o0Y68qRV/KHNX/gqQ1P8WDfB1NUsRPQWAsvf9UaAX7cldZlQ09Wumsl2sHn8zFmzBjGjBkDQE1NTTT42rNnD1u2bAGsYSP69+/PoEGDKCkpobi4mLy8vA63fvlDfrYd28aWyi1srdzKlqotbD+2nYN1B5vk6xfsR2luKRcNvYjSnFIG5wxmUNYgBmQNwBdpFRZCCNGEBFvxIsFWKyPIx8twZXDN6Gt4Yu0T7Di2g6G5Q9veKdWO7YPnv2ANVHrBz+D02+WyYQ+WnZ3NhAkTmDBhAgANDQ3s3buXPXv2sGfPHtasWcPy5csB607IAQMGNJkKCgpwOByY2mRvzV42V25mS+UWa6rawp6aPdExntyGm+F9hjO131RKc0opzS2lNKeUXat3ccE53fhSuRBC2LTWhEKh6BR5vE+6SLAVzzDA4e5QyxZYz0t8av1TPL3h6fQPcrp/NTx/jdWydd2LMPK89NZHJJ3P52PkyJGMHGk9vcA0TcrLy9m/fz8HDhzgwIEDrFixglDkOYUGBHwBjjiOcNRxlGp3NTWuGgrzCxmRN4ILh17IyD4jGZE3gsHZgxPeWXvQONgsTQghEgmHw4RCoeg8WVN7jxcOh5vUp6ioiHHjxqXp1ZBgKzGnt913I0YU+gq5fMTlzN86n7sm30WhrzBFlWvDtsXwwvXWwKS3/Av6pe/NJbqOYRgYOQZH/EfYrDez0bWRjRkboQ76BPpQECigv9mfgY0DGdg4MLqf65CLwsJCioqK8BZ6qa+tpzy/nLy8PHw+uSwoRE+kte5UkJOMACkZLUlOp7PVyefzRZcdDkereTMzMykvL0/Cq3uCf0vaSu7OnF44gZGjbxx3Iy99+hLPbXqOb0z5Rgoq1oZNr1l3HBaOgi/+Q4Z1OImFzBBbKrew+vBqVpdb0/66/YA1WOeovFFcMPQCxhaMZVzBOEb2GRkdJqGhoYHDhw9Hp/Lycnbs2MHatWublOHz+cjPzycvL4/8/HwOHTrEtm3byM3NJScnB7dbOrwLkYjWmkAgkNJgpq3jdZZhGE2ClUTBjNfrbTMgamlqKzhyOBxJvzNZgq3u5gRatgCG5Azh3MHn8sLmF7h1wq1kuLpwcMrVz8P8O6F4Clz/N/DldV3ZIuUaw42sPbyWZQeXsap8FesOr6M+VA9Aka+ISX0ncf2Y65nUdxKj80e3euefz+dj8ODBDB48uGkZjY1UVlZSWVlJRUUFFRUVVFZWsnfvXjZs2IDWms2bN0fze73eaOCVk5NDbm4u2dnZZGVlkZmZGZ164jhhoufSWmOaZlouXUUm0zR55513OvV3OY1AzQAAIABJREFUtBWMtBXotLV/W/s6HF37JISTnXwKJuLydrjPVsRN42/i7d1v8/ctf+dLY7+U5Iq1YNVfYP4dMPRsuOY5uePwJBA0g2w4soFlB5ex7MAyVh9eTWO4EUMZjMobxaXDL2Vy38lM6juJgZkDk/IL0OPx0L9/f/r3799sWygU4q233mLMmDEcO3aM6upqqquro8v79u2jvr4+4XG9Xm808IoEYl6vF5/PF53HLnu9Xtxut4y31cPEX7pqabmj6x3dNxmtOkqpNgMSt9vd4ra9e/cyYsSIE27pcTgc3WI8PZE8Emwl4vR06G7EWBOLJjKl7xSe2fgM15xyTepHuF77N6tFa9hcuPYFK1AUPY7Wmi1VW1iybwlLDy7l40MfRx+CPCpvFFePuprTBpzGlH5TyHHndHn9Iv0jSktLW8wTDAapqamhrq6O2tpa6urqolNkvby8nLq6Ovx+f6t9OgzDwOv14vF4cLlcuN3uNqdEX1gtpTmdTqufm2GglIrO0y3SKmOaZpPl+PXY5draWvbv3084HG5xMk2z1e0dydta0JMMkctXLZ23+Fad9rbgdKSlp7Pvh7KyMmbPnp2U10OcHCTYSsR54i1bAF859Svc/vbtvLLtFa4edXUSKxZn43xrHK0hs6wWLQm0epTaQC0fHfiI9/e9z3v73qO83upPMDR3KJcNv4wZ/Wcwvf908rw945Kwy+UiPz+f/Pz8NvOapkkgEKChoQG/39/iPBAINJlqa2ujy8FgkGCw9YdNd1R88NXSHGgWLHZkvaWA6kStWLHihPaLBJyRy0atTS6XC6/X2yxoaW39RLZJi444GUmwlUgng61ZA2cxvmA8T659kitGXIHLSEHr1rZFVmf44qlw3Qvg7h0PL+7ptldtZ/Gexby/731Wl68mpENkubKYOXAms4tnM2vgLPplnvw3NkRarrzezv1AME2TYDDYrDNyotvO49PC4XA02GlrHp8W3+rR0fVIkBMbxMUvt3fbpk2bmDBhQrsCptgpchwhROpJsJWI0wv1R054d6UUt0+6nTsX3slr217jypFXJrFywIE18NcvQeEpVmd4T3Zyjy+SxtQm64+sZ+HuhSzavYid1TsBOCXvFG4cdyOzi2czse/E1ATkvYBhGHg8npQ+sqi7O3LkCKNHj053NYQQrZBgKxGn54TuRox1ZvGZjCsYx+NrH+fS4Zcm78u0chf85Wrw9oEvvgS+Psk5rkiaoBlk+YHlLNqziMW7F1PeUI5TOZnWfxrXj7meOSVz6J/ZvBO6EEKIk5MEW4m4fBDs+DhbsZRSfG3i1/j6oq+zYNsCPjvys52vV0MlPHuVdYnzy69CzsC29xFdwtQmKw+t5J87/sm/d/2bqsYqfE4fswbO4pzB53DWoLPI9eSmu5pCCCHSQIKtRJLQsgVw9qCzGZM/hifWPcGlwy9N+AiUdguHrD5aVbvghvnQVy4bpJvWmk0Vm3hj+xu8ufNNDtUfwuvwMrdkLvOGzuOMgWfgdcpNC0II0dtJsJWI09epDvIRkdatuxffzRs73uCy4Zed+MEW/tDqFH/Z72HIGZ2umzhxB+sOMn/rfBZsX8DO6p04lZNZxbO4Z+o9zC2Z27WD2QohhOj2JNhKxOlJSrAFMLdkLqPzR/P42se5aOhFJ9a6tfZFWPJ7mP4VmHJDUuolOiYQDlC2p4x/bP0HH+7/EFObTOs3jRvH3cj5Q86XS4RCCCFaJMFWIpGhH7SGTg50qJTi9om3c/fiu3l126sdvzPx4Dp49eswZDbM+1mn6iI67tPKT3l5y8ss2L6AqsYq+mX049YJt3LFiCsoyS5Jd/WEEEL0ABJsJeLygjYhHARn5x+2O7dkLqcWncojqx/h4mEX43G08zb1QJ3VT8vbB65+ClI9Gr0AIBgO8q9d/+L5T55n7eG1OA0n55Scw2dHfpaZA2biMOSZYUIIIdpPgq1EIp2aQ/6kBFtKKe6efDe3/PsWXvjkBW4cd2P7dnzzu3BkC9zwCmQVdboeonXl9eW8uPlFXvr0JY76jzIkZwj3TruXS4df2mNGcRdCCNH9dCrYUkrlA38FSoGdwOe11pVxeSYBfwBygDDwE631XztTbspFg63O35EYMWPADGYOmMmT657kqpFXkeVu42HRn7wBK5+GWd+EYXOSVg/RlNaaVeWreO6T51i4ayFhHeasQWdx7ehrmTlwJoaSEbaFEEJ0Tme/Sb4DLNRajwQW2uvx6oEbtNbjgHnAg0qp7j0SZzTY6txYW/G+MeUbVDVW8czGZ1rP2FAFC+6BvuNg7veTWgdhCZkh/rnjn3xhwRe48c0bWbJ/CdePuZ7XP/s6D537ELOKZ0mgJYQQIik6exnxcmCOvfw0UAb8Z2wGrfWnMcv7lVLlQBFQ1cmyUycFLVsA4wvHc97g83h6w9NcM/oa8r0tPLD33/8FdYetZx4m4TKmOK4h1MArW1/h6Q1Ps692H6U5pfxg5g+4eOjFMmSDEEKIlOhssNVPa33AXj4ItPoEXaXUDMANbOtkuanliumzlWRfn/x1Fu1ZxJPrnuTb07/dPMO2xbDqz9blw4GTk15+b1Xpr+SFT17guU+eo6qxilOLTuXe6fcyt2SutGAJIYRIKaW1bj2DUm8DiR7k9n3gaa11n5i8lVrrhD2JlVIDsFq+btRaf9RCntuA2wD69es39YUXXmjP39AptbW1ZGU17T+Vf/RjTl33I1ZO/gXVuackvcxnjzzLx3Uf8/2B36fQVRhNd4QamLbiG2jlZMW0BzHbe9fiSSbROTlRx0LHeLv6bZbULiGgA4z3jee8nPMY5hmG6uSwHr1NMs+LSB45L92PnJPuKdXnZe7cuR9rracl2tZmy5bW+ryWtimlDimlBmitD9jBVHkL+XKA14HvtxRo2WU9DjwOMG3aND1nzpy2qtdpZWVlNCtnhwHrYMqpY2HomUkvc3TdaC59+VKWuJbw6zm/Pr7hrfvAXw43v8lZQ2YmvdyeIuE56aAKfwX/t/7/eP6T5wmZIS4edjE3jbuJkXkjk1PJXigZ50Ukn5yX7kfOSfeUzvPS2cuIrwI3Ag/Y8/nxGZRSbuBl4Bmt9UudLK9rOH3WPAWXEQH6Z/bny+O/zCNrHmH5weVM7z8djm6Djx6BiddBLw60OutY4zGe3vA0z256Fn/IzyXDLuFrE7/G4JzB6a6aEEKIXqqznVUeAM5XSm0BzrPXUUpNU0o9aef5PHAWcJNSarU9TepkuanltC/fpSjYArhp/E30z+zPL5f/krAZtjrFO9xw3n0pK/NkVhuo5dE1j3Lh3y/kiXVPcNags3jl8lf46Zk/lUBLCCFEWnWqZUtrfRQ4N0H6CuBWe/lZ4NnOlNPlXJGWreTejRjL5/Rxz5R7+M/3/pP5Hz7AlZvfgPN+CNmJuseJltQH63lh8wv8af2fONZ4jLklc7lz0p2ckp/8vnZCCCHEiZAR5BOJtGwFkzvOVrwLh17Ic5v+wm8/fYHz8kvJOf2OlJZ3MmkMN/K3zX/jiXVPUOGvYFbxLO6adBfjC8enu2pCCCFEExJsJeJM3dAPsZRSfDdrLNcdXsODI6fzA2fvvPuwI4LhIC9vfZnH1j5GeX050/tP58HJDzK5rwyTIYQQonuSYCuRFA1q2kygjnHLnuKLRYU8U76Uiw6uYFr/hHeN9nohM8Rr217jsbWPsa92H5OKJvHT2T/ltAGnpbtqQgghRKtkNMdEUvS4nmaWPQ515dx51s8ozirmfz78H/wpbk3racJmmNe3v84V86/gB0t+QK4nl0fOfYRnLnxGAi0hhBA9grRsJeJwASq1LVv+Y/D+gzDyM2QMm8N9Xg+3vXUbv/7413zvtO+lrtwewtQmC3cv5OFVD7Pt2DZG5o3kwbkPck7JOTIYqRBCiB5Fgq1ElLLuSExlB/klD4G/Cs75LwBmDpzJl8Z+iT9v/DOzBs7i7JKzU1d2N6a1Zn39eh5Z8AibKjZRmlPKL8/6JZ8p/Yw8VkcIIUSPJMFWS5ye1LVs1R2xBjAdewUMmBhN/uaUb7LswDL++4P/5sVLX6R/Zu8ZBkJrzYcHPuThVQ+z9shaBmUN4iezf8JFQy/CacjbVAghRM8lTQUtcfpSdzfikt9DsB7mfr9Jstvh5hdn/4LGcCN3L767V/Tf0lrz0YGPuOnNm/jqW1+lvKGca/Kv4dXPvsplwy+TQEsIIUSPJ8FWS5ye1ARbDZWw/I8w7kooGtVs87DcYTxw5gNsOrqJ+5bcR1sPCu+ptNYsPbCUm968ia/8+yvsrd3Ld2d8l9c/+zqzsmfhMlzprqIQQgiRFNJs0BKnNzV9tpY9CYEamH1Pi1nmDp7LXZPv4verfs+AzAF8c+o3k1+PNNFas+zgMh5Z/Qgry1fS19eX7532Pa4ceSUeh4wzJoQQ4uQjwVZLXCm4jBios/pqjZoH/Vsf6fwrE77CobpD/HH9H8lyZ3HrhFuTW5cuprVmyf4lPLHuCT4+9DF9fX357ozvctWoqyTIEkIIcVKTYKslLh8Ekxxsffw0NFTAmf+vzaxKKb5/+vepC9Xx25W/JRAOcPvE23vcsAdBM8ibO97k6Q1Ps7lyswRZQggheh0Jtlri8kF9RfKOF2q0OsYPmQ0lM9q1i6EMfjzrxziVkz+s+QMH6g7wg9N/gMvR/fsz1QXr+MeWf/DMxmc4WHeQ4bnD+dEZP+LiYRfjdrjTXT0hhBCiy0iw1RKnN7mXEde8ADX74fKHOrSby3Dx41k/ZkDWAB5d8yifVn7KL876BUNyhiSvbkm0uWIzL25+kQXbF1Afqmdqv6n89+n/zezi2TJOlhBCiF5Jgq2WuDKs4RmSwQzDBw/CgEkw/JwO766U4s5JdzI6bzT3fXgfV792NXdOupPrxlzXLe7aqw3U8vbut/nHln+wqnwVHoeHeaXzuGb0NYwvbL1vmhBCCHGyk2CrJS5v8vpsbXwFKrbD5/9sjU5/gs4dci7jCsfxow9/xK9W/Ip/bPkHd02+i3MHn9vlrUZBM8iH+z9kwbYFLN6zGH/Yz5CcIXxr2re4YsQV5Hpyu7Q+QgghRHclwVZLXBnJGfpBa6uvVsEIGH1Jpw/XP7M/j5z3CO/seYdfrfgV/1H2HwzPHc61o6/lomEXke3O7nydW1Dpr+T9fe/z7t53+WDfB9QEa+jj6cPlIy7n0uGXcmrhqT2uA78QQgiRahJstcTlg1ASgq3dH8H+VXDx/4KRvNans0vOZnbxbN7a9RZ/XP9H7l96P79a8SvOGHgGc0rmMLXfVEqyS044+NFac7DuIOuOrGNV+SpWl69mY8VGTG1S6Cvk/NLzmVsyl1kDZ/WIDvtCCCFEukiw1RKnD8IBq7+V4Tjx43z4EPjyYOK1yaubzWE4mDd0HheUXsDGio28suUVFu9ZzKI9iwDIdmczNn8sJTkl9M3oS7+MfmS5svA4PLgcLkxt0hhqxB/2U9VYRXl9OYfrD7OrZhfbqrZRF6wDwOPwMKFwAl899aucPehsxhSMkc7uQgghRDtJsNUSl8+aBxvAk3Vix6jYDp+8Dmf+B7gzk1e3OEopxhWMY1zBOL532vfYUrWFdYfXseHoBjYe3cjCXQupbKxs8zhOw0mRr4hB2YO4dNiljMwbyej80YzJHyOtV0IIIcQJkmCrJckItpY+BoYTpn8lefVqg1KKUXmjGJU3iqu4KpoeCAc43HCY+mA9gXCAxnAjhjLwOr14HB76ePqQ68mVFishhBAiySTYakk02DrB4R8aqmDln2H8VZAzIHn1OkFuh5virOJ0V0MIIYTodaQZoyVOrzU/0YFNVz4DwTqYeUfy6iSEEEKIHkeCrZa4Mqz5ibRshUPWJcTSM2HAxOTWSwghhBA9igRbLYleRjyBlq1N86F6L8y8M7l1EkIIIUSPI8FWS060z5bW8OHDkD8cRl6Q/HoJIYQQokeRYKslkWCro3229iyDfR/D6bcndRBTIYQQQvRMEg20xBkz9ENHfPgQePvApOuSXychhBBC9DgSbLXEdQLBVuVO+GQBTLs5pYOYCiGEEKLnkGCrJScSbC19DJQBM25LTZ2EEEII0eN0KthSSuUrpd5SSm2x53mt5M1RSu1VSj3UmTK7TLTPVjuDLf8xaxDTcVdCzsDU1UsIIYQQPUpnW7a+AyzUWo8EFtrrLfkx8G4ny+s6He2ztfLPEKiRQUyFEEII0URng63Lgaft5aeBKxJlUkpNBfoB/+5keV3HMMDhaV+wFQ7B0kdhyCwYODn1dRNCCCFEj9HZYKuf1vqAvXwQK6BqQillAP8LfKuTZXU9l699wdYnr8GxPTKIqRBCCCGaafNB1Eqpt4H+CTZ9P3ZFa62VUjpBvjuAN7TWe5VSbZV1G3AbQL9+/SgrK2urep1WW1vbYjkzTYOK3dvY3EY9Jq/8GW5vf5Ye8MLB1vOKtrV2TkT6yHnpnuS8dD9yTrqndJ6XNoMtrfV5LW1TSh1SSg3QWh9QSg0AyhNkmwmcqZS6A8gC3EqpWq11s/5dWuvHgccBpk2bpufMmdPOP+PElZWV0WI5a/swoCiPAa3VY88yKNsMF/6SOaedm4oq9jqtnhORNnJeuic5L92PnJPuKZ3npc1gqw2vAjcCD9jz+fEZtNbXR5aVUjcB0xIFWt2Ssx2XET98GLy5MoipEEIIIRLqbJ+tB4DzlVJbgPPsdZRS05RST3a2cmnnzbGGdGhJ5S7Y9CpMvQk8WV1WLSGEEEL0HJ1q2dJaHwWaXTvTWq8Abk2Q/hTwVGfK7FK+fKja1fL2ZY/bg5h+tevqJIQQQogeRUaQb40vD+orEm/zV8PKZ2DsFZBb3LX1EkIIIUSPIcFWazLyoKEy8bZVz0JjtQz3IIQQQohWSbDVGl++9bie+E7yZhiW/gEGz4TiKempmxBCCCF6BAm2WuOzH/UYfynxkwVQtVtatYQQQgjRJgm2WpORb83jLyV++DDklcIpF3V5lYQQQgjRs0iw1ZpIy1ZDTMvW7qWwZymcficYjvTUSwghhBA9hgRbrfHZLVuxlxE//D14+8Dk6xPvI4QQQggRQ4Kt1kRbtuzLiEe3waYFMP1WcGemr15CCCGE6DEk2GpNtM+W3bL10SPgcMGM29JXJyGEEEL0KBJstcbls56P2FAJdUdh1V/g1C9Adr9010wIIYQQPYQEW23x5UF9JSx/0hpza+Zd6a6REEIIIXqQTj0bsVfIyIf9q2DjK9ZQD31Hp7tGQgghhOhBpGWrLdNvgfINoE2Y90C6ayOEEEKIHkZattoy7cuQUQjuDMgbku7aCCGEEKKHkWCrPcZelu4aCCGEEKKHksuIQgghhBApJMGWEEIIIUQKSbAlhBBCCJFCEmwJIYQQQqSQBFtCCCGEECkkwZYQQgghRApJsCWEEEIIkUISbAkhhBBCpJAEW0IIIYQQKSTBlhBCCCFECkmwJYQQQgiRQhJsCSGEEEKkkNJap7sOCSmlDgO7uqCoQuBIF5Qj2k/OSfck56V7kvPS/cg56Z5SfV6GaK2LEm3otsFWV1FKrdBaT0t3PcRxck66Jzkv3ZOcl+5Hzkn3lM7zIpcRhRBCCCFSSIItIYQQQogUkmALHk93BUQzck66Jzkv3ZOcl+5Hzkn3lLbz0uv7bAkhhBBCpJK0bAkhhBBCpJAEW0IIIYQQKdRrgy2l1Dyl1Gal1Fal1HfSXZ/eRCn1J6VUuVJqfUxavlLqLaXUFnueZ6crpdTv7PO0Vik1JX01P3kppUqUUouVUhuVUhuUUnfb6XJe0kgp5VVKLVNKrbHPy//Y6UOVUkvt1/+vSim3ne6x17fa20vTWf+TnVLKoZRapZRaYK/LeUkjpdROpdQ6pdRqpdQKO61bfIb1ymBLKeUAHgYuBMYC1yqlxqa3Vr3KU8C8uLTvAAu11iOBhfY6WOdopD3dBvyhi+rY24SA/6e1HgucDtxp/0/IeUmvRuAcrfVEYBIwTyl1OvBz4Dda6xFAJXCLnf8WoNJO/42dT6TO3cCmmHU5L+k3V2s9KWY8rW7xGdYrgy1gBrBVa71dax0AXgAuT3Odeg2t9btARVzy5cDT9vLTwBUx6c9oy0dAH6XUgK6pae+htT6gtV5pL9dgfYEUI+clrezXt9ZeddmTBs4BXrLT489L5Hy9BJyrlFJdVN1eRSk1CLgYeNJeV8h56Y66xWdYbw22ioE9Met77TSRPv201gfs5YNAP3tZzlUXsy9xTAaWIucl7exLVauBcuAtYBtQpbUO2VliX/voebG3HwMKurbGvcaDwLcB014vQM5Lumng30qpj5VSt9lp3eIzzJmqAwtxorTWWiklY5KkgVIqC/g78E2tdXXsj285L+mhtQ4Dk5RSfYCXgdFprlKvp5S6BCjXWn+slJqT7vqIqNla631Kqb7AW0qpT2I3pvMzrLe2bO0DSmLWB9lpIn0ORZpw7Xm5nS7nqosopVxYgdZftNb/sJPlvHQTWusqYDEwE+uSR+THcuxrHz0v9vZc4GgXV7U3mAVcppTaidUN5Rzgt8h5SSut9T57Xo71w2QG3eQzrLcGW8uBkfadI27gGuDVNNept3sVuNFevhGYH5N+g33nyOnAsZgmYZEkdv+RPwKbtNa/jtkk5yWNlFJFdosWSikfcD5Wf7rFwOfsbPHnJXK+Pgcs0jJyddJprb+rtR6ktS7F+v5YpLW+HjkvaaOUylRKZUeWgc8A6+kmn2G9dgR5pdRFWNfcHcCftNY/SXOVeg2l1PPAHKAQOATcB7wCvAgMBnYBn9daV9hBwENYdy/WAzdrrVeko94nM6XUbOA9YB3H+6B8D6vflpyXNFFKnYrVqdeB9eP4Ra31j5RSw7BaVPKBVcAXtdaNSikv8GesPncVwDVa6+3pqX3vYF9G/JbW+hI5L+ljv/Yv26tO4Dmt9U+UUgV0g8+wXhtsCSGEEEJ0hd56GVEIIYQQoktIsCWEEEIIkUISbAkhhBBCpJAEW0IIIYQQKSTBlhBCCCFECkmwJYQQQgiRQhJsCSGEEEKkkARbQgghhBApJMGWEEIIIUQKSbAlhBBCCJFCEmwJIYQQQqSQBFtCCCGEECkkwZYQQgghRApJsCWEEEIIkUISbAkhhBBCpJAEW0IIIYQQKSTBlhBCCCFECkmwJYQQQgiRQhJsCSGEEEKkkARbQgghhBApJMGWEEIIIUQKSbAlhBBCCJFCEmwJIYQQQqSQBFtCCCGEECkkwZYQQgghRApJsCWEEEIIkUISbAkhhBBCpJAEW0IIIYQQKSTBlhBCCCFECkmwJYQQQgiRQhJsCSGEEEKkkARbQgghhBApJMGWEKLXUkqdqZTanO56CCFObhJsCSE6TSm1UynVoJSqVUodVEo9pZTKSne92qK1fk9rfUpXlqmU0kqpEQnSL1ZKva+UqrJfwyeVUtldWTchRGpIsCWESJZLtdZZwCRgMvDdZB5cKeVM5vFSyQ42b+rgbrnA/cBAYAxQDPwyyVUTQqSBBFtCiKTSWh8E/oUVdKGUOl0ptcRusVmjlJoTyauUGqqUelcpVaOUelsp9bBS6ll7W6ndCnSLUmo3sMhO/7JSapNSqlIp9S+l1BA7XSmlfqOUKldKVSul1imlxtvbLlJKbbTL2aeU+padPkcptTemPmOUUmV2XTcopS6L2faUXb/X7eMsVUoNT+Lr9pzW+k2tdb3WuhJ4ApiVrOMLIdJHgi0hRFIppQYBFwJblVLFwOtYLTb5wLeAvyuliuzszwHLgALgh8CXEhzybKyWnguUUpcD3wOuBIqA94Dn7XyfAc4CRmG1En0eOGpv+yPwVa11NjAeO3CLq7cLeA34N9AX+DrwF6VU7GXGa4D/AfKArcBP2vmynIizgA0pPL4QootIsCWESJZXlFI1wB6gHLgP+CLwhtb6Da21qbV+C1gBXKSUGgxMB36gtQ5ord8HXk1w3B9qreu01g3A14Cfaa03aa1DwE+BSXbrVhDIBkYDys5zwD5GEBirlMrRWldqrVcmKOd0IAt4wK7PImABcG1Mnpe11svssv+C3XqXbEqp84EbgR+k4vhCiK4lwZYQIlmusFuO5mAFPIXAEOBq+7JclVKqCpgNDMDqm1Shta6POcaeBMeNTRsC/DbmWBWAAort4Ogh4GGgXCn1uFIqx97vKuAiYJdS6h2l1MwE5QwE9mitzZi0XVh9pyIOxizXYwVnACil1sbU6zrgkZi/+5EE5SWklDodq8Xvc1rrT9u7nxCi+5JgSwiRVFrrd4CngF9hBUp/1lr3iZkytdYPAAeAfKVURszuJYkOGbO8B+tyYOzxfFrrJXbZv9NaTwXGYl1OvNdOX661vhzr8uArwIsJytkPlCilYj8XBwP72vl3nxqpE1awdEdMHe9ozzGUUpOxWve+rLVe2J59hBDdnwRbQohUeBA4H1gCXKqUukAp5VBKee1O6YO01ruwLin+UCnltlubLm3juI8C31VKjQNQSuUqpa62l6crpU6z+17VAX7AtI99vVIqV2sdBKoBM8Gxl2K1Vn1bKeWyO/JfCrzQydciEbf9WkQmh92Z/03g61rr11JQphAiTSTYEkIkndb6MPAM8A0g0qn9MFbL1L0c/+y5HpiJ1ZH9fuCvQGMrx30Z+DnwglKqGliP1RkfIAfrDr5KrMt/Rzk+dMKXgJ32Pl+zy40/dgAruLoQOAI8Atygtf6kwy9A2zYADTHTzcD/w+r0/0d7vLJapZR0kBfiJKC01m3nEkKILqCU+ivwidb6vnTXRQghkkVatoQQaWNf+huulDKUUvOwWsFeSXe9hBAimXrMiMxCiJNSf+AfWONs7QVu11qvSm+VhBAiueQyohBCCCFECsllRCGEEEKIFOq2lxELCwt1aWlpysupq6sjMzMz5eWI9pNz0j3Jeeme5Lx0P3JOuqdUn5ePP/74iNa6KNG2bhtSVwOUAAAgAElEQVRslZaWsmLFipSXU1ZWxpw5c1Jejmg/OSfdk5yX7knOS/cj56R7SvV5UUrtammbXEYUQgghhEghCbaEEEIIIVJIgi0hhBBCiBTqtn22hBBCiGAwyN69e/H7/emuSrvl5uayadOmdFdDxEnWefF6vQwaNAiXy9XufSTYEkII0W3t3buX7OxsSktLUUqluzrtUlNTQ3Z2drqrIeIk47xorTl69Ch79+5l6NCh7d5PLiMKIYTotvx+PwUFBT0m0BInN6UUBQUFHW5plWBLCCFEtyaBluhOTuT9KMGWEEIIIUQKSbAlhBBCtMLhcDBp0iTGjx/P1VdfTX19fbqr1C6vvvoqDzzwQLqrIUhSsKWU+pNSqlwptb6F7Uop9Tul1Fal1Fql1JRklCuEEEKkms/nY/Xq1axfvx63282jjz7aZHsoFEpTzVp32WWX8Z3vfCfd1RAkr2XrKWBeK9svBEba023AH5JUrhBCCNFlzjzzTLZu3UpZWRlnnnkml112GWPHjiUcDnPvvfcyffp0Zs6cyWOPPQaAaZrccccdjB49mvPPP5+LLrqIl156CbAeS3ffffcxZcoUJkyYwCeffALAsmXLmDlzJpMnT+aMM85g8+bNADz11FNceeWVzJs3j5EjR/Ltb387Wq8333yTKVOmMHHiRM4999xo/rvuuguAw4cPc9VVVzF9+nSmT5/OBx98wP9n78zjo6rO//++y2zJTDayJ2RhD1tYwhqRIAooVAoVRa3gbi1qrVvR1lb9+fWr1fqtimKptkCtGwiIuACyCYKyiyxhD1uAsCWZZDL7/f1xJ5MJCXtCQnLeL+Y1c88995wnc8PkM8/znOcALFu2jG7dutGtWze6d++O3W6/PG9kM6NOSj9omvadJEkZZ+kyEpiuaZoG/CBJUpQkSUmaph2ui/kvFr+/cX4bEQgEAkFNnv9iC1sLS+t0zI7JEfzlF53Oq6/X6+Xrr79m2DDdt7B+/Xo2b95MZmYmU6ZMITIykjVr1nD8+HGGDRvGkCFDWLduHQUFBWzdupWioiKysrK4++67g2PGxsayfv163nnnHV577TXee+89OnTowPLly1FVlW+//ZZnnnmGzz77DICNGzeyYcMGTCYT7du35+GHH8ZsNnPffffx3XffkZmZycmTJ2vY/rvf/Y7f//73XHXVVezfv5+hQ4eybds2XnvtNd5++21yc3MpKyvDbDbXwbsqOJ3LVWcrBTgQcnww0NZgYsvv9/P9/EEo3kh2b9xEXOtrsVmzxKoXgUAgEFSjoqKCbt26Abpn65577mHlypX07t07WGtpwYIFbNq0iZkzZ+L3+7Hb7ezcuZMVK1YwZswYZFkmMTGRQYMGVRt79OjRAPTs2ZNZs2YBUFJSwvjx49m5cyeSJOHxeIL9Bw8eTGRkJAAdO3Zk3759nDp1iquvvjpoS0xMTI2f4dtvv2Xr1q3B49LSUsrKysjNzeWxxx7j9ttvZ/To0aSmptbV2yYIoVEVNZUk6X70MCMJCQksXbq03uZyON14DncmusUWCk6+QcHJN5C8kUhSJzS1ExKdkCRRlK4hKCsrq9d7L7g4xH1pnDT1+xIZGRkMbT2Wl1Yvc5wrdGaxWFi+fHnw2OVy4XA4MJlMwWs9Hg+vvPIK1157LT6fD0VRAJgzZw5OpzPYz+v1UlFRgd1uR9M0PB4Pdrsdp9OJy+XCbrczceJE+vXrx/Tp09m3bx/Dhw8P9pEkKTiWpmmUlpZSUVERHCcUp9OJ2+3Gbrfj8/lYuHBhNc+VpmlMmDCBvLw8FixYQP/+/Zk9ezbt2rW79De1EeLz+eosTOp0Oi/o/93lEluHgJYhx6mBtmpomjYFmAKQk5Oj5eXl1atR//tVS6Z/t4e/dQuntbaBEv9qylv8hKasREPCFtaJiKhszKZEVEMUimxClk3IsjnwbEJRql6ffixJSr3a31RZunQp9X3vBReOuC+Nk6Z+X7Zt29YoqrGfbkNYWBiqqgbbhw8fzrRp0xgxYgROp5PDhw+TkpLCoEGDmDZtGg888ADHjh1jxYoVjBs3DpvNhiRJWK1WbDYb4eHhKIqCzWbD4XDQunVrbDYbM2fORJIkbDYbZrMZo9EYnFNVVcLCwujVqxePP/44x48fD4YRY2JiqvUfOnQoU6dO5cknnwT0cGS3bt3YvXs3ffv2pW/fvmzatIkDBw7Qs2fPy/vmXibqsrK/2Wyme/fu593/comtucBDkiR9DPQBSho6XwvgD8M6sG77Pib8VM47t43lutQHKd9wmBP5P1KqrsURu5kj9rn45ItTwpJkqBJhsglJNgafZdkYOBfyWqreXmt/6QzXnt5fCmmXVBEeFQgEgnrk3nvvpaCggB49euDz+UhISGDOnDn86le/YtGiRXTs2JGWLVvSo0ePYBjwTDz11FOMHz+eF198keHDh59z7ri4OKZMmcLo0aPx+/3Ex8ezcOHCan3efPNNJkyYQNeuXfF6vVx99dW8++67/P3vf2fJkiXIskynTp24/vrrL+l9ENSOpOesX+IgkvQRkAfEAkeBvwAGAE3T3pX0v/ST0FcsOoC7NE1be7Yxc3JytLVrz9qlTliwaAnvbjeypbCUj+7vS4+0aDRNw73fjmP9URw/HcfnLsdvdKAkGVHTzRhSjciJKsge/H4XPr8Lv8+J3+/C73cG2tzB11Xn3Pqz5q567a/9tRboUzfINYRZqHirLuoMukiUDEiyGvIcEG2VbdXOB/oHzxsDz2qg32njSQZkOfScWq1t2bLvmvQ39SuVpu5BuVJp6vdl27ZtZGVlNbQZF8TpHpSysjKsVisnTpygd+/efP/99yQmJjaghc2TuvRs1fZ7KUnSOk3TcmrrX1erEW89x3kNmFAXc9UlmteLyePin+P6M+qdldw3bS2zf5tLWoswTOkRmNIjiPpFa9z77Th3F+PaXYx7uR77RvFjbGnD1DoVa6tITGkRSIa6rRGraVpQdNUqzkJEm3amPn53oF/Ndi3ktc/nwO8txu/3oGnewHPoay+a5sHv9wCXLtDPjsySpZWiLCDiJCVEkFV/yAGRV/3YELimUuApgT7666AQlJQQwaeGCMbqx8Fxq/WpmuOMtlVrEzWEBYLmyIgRIyguLsbtdvPss88KodUMaVQJ8pcTzedjR5++hOfm0mLYMKbe1YvRk1dy59TVzHqwP1FhRgAkVcbUKhJTq0i4Lh2/24e7oBTX7mKcu4uxL96PfRGgSpjSIoJ9jWkRSOql/XGVJAlJ0sOQjQlN89Uqxirb/JoXrfK1PyDSNA+a3xt4DvQPtFU/76Vg705apqYEx/FrHjTNp8/n9+rza5UC0Kf38XvQNGeITd7A+apHbW2XF/kcYrFSCBoCxwGBWSk2Q4Rn0JMY0i5XE4cG5EohWJvwDLTLgTlCxzqT8NS0k7jcx6vGrSY2RZhaIDgTTXkBg+D8aLZiS1IUDCkpqIcLAWgVZ2XKHTn8+r0fuf8/6/jPPb0xqTUT3GWjgrldNOZ20UQC/govroISXLtLcO0toXTRfvgWUGVMabaA+IrCmGa7ZPHVWJAkJbDSpn7qsezft5Q2bfLqZexQdM+hL+i10wWgLygG9WNvQOCFHnuqXVf9mppCMHhdyFhVx1VjVYnKSpGpj+HzV9QiMr1BcVqboAR/vbxnK1bU3h7qfazp0avNy3e6ODyTlzLUK3mO64KPWkRjtbZKO5UantPa24SQFAgEl0azFVt+v5+itGRc+duDbb0zY3h1TFd+9/FGnpq5ib/f0u2cH7SyRcWS1QJLVgt93Aovrr0luPaU4NpTHBBf+3XxlW7D1CpK93y1bDri60pF9xyq6P8NmlYhP03zBwSkL+hdDIq6EDFZmxCsXWR6yM/fQru2baq8l7UJ0xqCMjCvv6aY1IWkC80bIlRPs616m+7dvPzIVTmGQU9eTWFYe1uop7K2thCRd1pou2ZbqLCsatO0rZw6ZT5LiLuW0HioCBViUiCod5qt2JIkiW9i4miR4WSAw4EcFgbAyG4pHDxVwavzt5MeE8ZjQ9pf0LiyRcXSsQWWjgHx5fDgCoQdXXtKKP12n57yFCq+WkdiTBXiS1B3SJKMJOnhZ6WOKpDs2B5Lampe3Qx2kejpn/7TBGTtorJ2T2RNIXn+bSGeRX8tbUHx6cOvefH7nUGhWbv4rBnW1jTfRb0v6zdc/Ht6dg9kaK5klTAMeh3lWtrO4Yk8n3zI0FC53x+B11sesFUCQh8123TtWPlaCElB46BZiy3V78VlseDasxdL56rtGn6b15r9Jxy8uXgXsTYT4/plXPQ8cpihpvjaW4prT4j4WgiSQcaYHoEpM1KIL4HgDOh/PCvD2I0rl7Eu0EPbp4u+2jyIVW3r168lu1vn0zyNvipPYogIrb0tZL7Tw9xnafP5y/F7qzyOeruvugezmhC+uPB2TPQ/cDgu5RuDBJJUKc1CjqXTzlcJuOrHNUVcdSEn1ZhD09y43K7qcwSvOT+hWNscVf057bygsdNsxRaASVFwmcy4du2sJrYkSeLFUZ0psjv58+dbWLStiJtzWjI4Kx6z4dLcBHKYAUunFlg6hYqvyrBjCaULTxNflQn3QnwJBE0ePbRtIFA55zyvKSEmul/9GVWH1Axvn3sRy8GDMmFhmYAW8GxWroYOPdYfmj5JyLEW6B56HHL9acfBMTUNLSAMNTSKjh5j4sSXWLPmJ6KiIjAYDDz66N1ERdl4882pfPrp25y+StvlrKc38XROF4rnIer0f6HH5xCf5xxT96YHr7/gMc82R9OgWYut8DALFV4f7t27a5wzKDLvje/FP77bzbSVBUz4cD3hRoXBWQnc0CWRge3isRgvPT6ji69YLJ1iAfCVe3CHiq8F+4CA+MqIqEq4T7EK8SUQCK4oLia8XVi4DVW11qNVZ0fTNG6/fRzjx4/n00/nAbBv3z7mzp1LcnImqmojIqIzWohos9vtWK3h1C4MA8dBoacFW6gmHnVxWtm/aozTj89/zKD4DIrJCxmzITiTMOTMx2cQbfqXmIb7PWrWYstms1HkcGHfsYP4Ws4rssRv89rwwNWtWbX7BF/+XMj8LUeZ+1MhFoPCoA5xXN85iUEd4rGa6uatVMINWDrHYulcm/gqpnT+PmCfLr7SbBgzIjFlRmJMsyHXgfgTCAQCQRWLFy/GaDTym9/8JtiWnp7Oww8/XK2kw/PPP4/VauWJJ55AkmS6du3GvHnzyMjIYPr06bz22mtIkkTXrl35z3/+Q0FBAXfffTfHjx8nLi6Of//736SlpTFjxgyef/55FEUhMjKS7777Dp/Px8SJE1m6dCkul4sJEybwwAMPXNb34exeRWp6GUM8imcXdWcTmxc4R2CMyn6hY8qyDyG2Goio6GgoOsHRwkO0Pks/RZa4qm0sV7WN5f+N9LN670m+2nyYbzYf5aufj2BUZQa2i+OGLokMzkogwnz+IYBzUZv4cu0pwV2gl5qwL96PXQMUCWOKVRdemZGYMiKQzc369goEgqbG1xPhyM91O2ZiF7j+5TOe3rJlCz169Ljo4bds2cKLL77IypUriY2N5eTJkwA8/PDDjB8/nvHjx/Ovf/2LRx55hDlz5vDCCy8wf/58UlJSKC4uBuD9998nMjKSNWvW4HK5yM3NZciQIWRmZl60XRdK9ZAfXIkRvrrahPpiaNZ/jVvExcP2XZxwVeCvqEC2WM55jarI9G8TS/82sTx/Y2fW7TvFVz8f5pvNR1i49SgGReKqNrHc0CWJ67sk1ZnHqxIl3EBYl1jCuujiy1/hxbWvVPd+FZRiX3EIlh0ECQyJ4VXiKzMCxWqsU1sEAoGguTFhwgRWrFiB0Wjk1VdfPWf/xYsXM2bMGGJj9c/smJgYAFatWsWsWbMAuOOOO3jqqacAyM3N5c477+Tmm29m9OjRACxYsIBNmzYxc+ZMAEpKSti5c+dlFVuCS6NZi634pGQA7OFWXLv3VEuSPx8UWaJ3Zgy9M2P484iObDxYzNc/H+arn4+wZPsmnv18M8M6JTKqRypXtYlFkev+q4BsUbF0iMHSQf8P7Hf7cB+w6+Jrbwnla45QtlIv3KrGWfTVjgEBpkY1vdVcAoGgCXMWD1R90alTJz777LPg8dtvv83x48fJyam+BZ6qqvj9Vastnc6Ly5B/9913+fHHH/nyyy/p2bMn69atQ9M03nrrLYYOHXpxP4SgwWnWGdaJaekAlIeF4d6965LGkmWJHmnR/HF4R1b8YRCfPdifX/VIZXF+EeP/tZp+/7uIl77aRv6R0row/cx2GBXMraOIuDaduPu6kvyXfsQ9mE3k9RmoLSw4Nh3j5CfbOfLyag6/spqTn2ynfPURPMcc1MWm5AKBQNCUuOaaa3A6nUyePDnY5nA4avTLyMhg/fr1AGzcuJG9e/cGr58xYwYnTpwACIYR+/fvz8cffwzAf//7XwYMGADA7t276dOnDy+88AJxcXEcOHCAoUOHMnnyZDwevajvjh07KC8vr6efWFAfNGvPVrjVCn4fFeHhuHZdmtgKRZIkeqZH0zM9mj//oiNL8ov4bP0h/rViL1O+20PHpAjG9m7JyG4pRFrqLr+rVltUObiptm0gaH4Nz5FyXHtLcO8twbnjFI4NRQDIVkOV5ysjAkNiOFI9eOMEAoHgSkGSJObMmcPvf/97/vrXvxIXF0d4eDivvPJKtX6/+tWvmD59Op06daJHjx60a9cO0D1jf/zjHxk4cCCKotC9e3emTp3KW2+9xV133cWrr74aTJAHePLJJ9m5cyeapjF48GCys7Pp2rUrBQUF9OjRA03TiIuLY86cOZf9vRBcPFJj9Wbk5ORoa9eurdc53v3pXYpm7MNS7mCs00PLye/U63wnylzM23SYT9YcYOvhUswGmRFdk7m1d0t6pEU3SE0RTdPwHqvQxVdBKa49JfhKXABIJgVjmg1TRiTG9IjLtuJx6dKl5OXl1fs8ggtD3JfGSVO/L9u2bSMrK6uhzbgg7HY7Nputoc0QnEZd3pfafi8lSVqnaVpObf2btWdr36c+zH4/bpMJ1+Zt9T5fC6uJ8f0zGN8/g58PlvDh6v3M3XiImesO0i7BytheaYzukUJU2OVLZJckCUN8GIb4MOiTBID3pBP3vlJcBSW495VWbTEkgyHZiik9Qq/5lR6JEiGS7gUCgUAgOBvNWmzFlaZTFnYAr0HFffDgea9IrAu6pEbyv6ld+NPwLL74qZCP1hzghXlbefmbfG7onMitvdPonRnTIN4uNcaMGmMmrLtefcxf4cW1v1T3fBWUUr76CGXf60n3Soy5SnxlRKDGhYnQo0AgEAgEITRrseVXPaApaIqKW5Fx7dmDpdOFrUi8VMJNKmN7pzG2dxpbC0v5eM1+Zm84xJyNhbRPsDGufzqjuqcQZmy4WyVbVCztY7C011c8al4/7sKygPerFOfOqrwvyaKGeL4i9G2GDM16HYZAIBAImjnNW2wZfKAZQIJTETbcu3dfdrEVSsfkCF4Y2Zmnr8/ii02FTFtZwB9nb+blr/O5Oacld/RNJyM2vMHsq0RSZUxpEZjSIrANCOR9nXAGPF966NGZr6+4qSy2asyIDIowJbx+FwUIBAKBQNCYaNZiC4MPXGbASWlkBK6ddbci8VKwGBVuzmnJmJ6prN9fzLSVBUxbWcD7K/aS1z6O8f0zGNg2DrmRhOskScIQa8EQayE8JwEIbDO0r1QvuFpQStn3hyj77iCg1/syputhR2NaBGqcpUltOCoQCAQCQSjNWmxJRg3JaQWcOBIT6rT8Q10QWkLiT8Oz+HD1fv77437u+vca0luEcUffdMbktKz38hEXgxJuwNKxBZaOLQDQPH7ch+y4Ckp1z9fWEzjWHgUCocc0G8a0CIzpNiRvQ1ouEAgEAkHd0qzFlmyUkH1R+DhOeUQkrm07GtqkMxIfYebRa9vx27w2zN9yhGkrC3jxy238bcEObuqZyt1XZZLZCEKMZ0IyyJgyIjFlRAJ6vS/v8Qrc+0px77fj2leKc/s+AFohc3TLen2j7XQ9XKm0MAvvl0AgaBAURaFLly5omoaiKEyaNIn+/ftTUFDAiBEj2Lx58yXPkZeXx2uvvUZOTg4ZGRnYbDYkSSIxMZHp06eTmJhYBz+JoKFo1mJLMYLBH4bX76fCYsZz4MBlXZF4MRhVmV9kJ/OL7GQ2Hyph2soCPllzgA9+3MfgDgncOyCTPg20ivFCkOSqkhPhvfQPEX+FF/cBOzuWbSJJNuDYeIzyH48AIIcbQsSXDUPq5an5JRAIBBaLhY0bNwIwf/58nn76aZYtW1avcy5ZsoTY2FieeeYZXnrpJd58883gOU3T0DQNWa7/xUderxdVbdZSoU5o1svEVJOM6jNhlMApy6BpuANbLFwJdE6J5NUx2Xw/8RoevqYt6/efYuyUH/jFpBXM3nAQt9d/7kEaEbJFxdwumpNtNeLu6ULyX/qR8GgPoka1wdwhBu/xCkq/KeDYlJ8pfG4lR9/awKnPd+HYWIT3pFNsNyQQCOqd0tJSoqOja7Q7nU7uuusuunTpwlVXXcWSJUtqtHfv3j3YXlFRwdixY8nKymLUqFFUVFTUOt/VV1/Nrl27KCgooH379owbN47OnTtz4MABXn31VXr16kXXrl35y1/+AkB5eTnDhw8nOzubzp0788knnwAwceJEOnbsSNeuXXniiScAuPPOO4ObWwNYrVZAL5Q7YMAAbrzxRjp27IjP5+PJJ58MzvWPf/yjjt7N5kOzlqtGs4rmM4FBpcKr7znl2rULc8eODWzZhRFnM/HYde34bV5rZm84xPsr9vL7T37i5a/zGd8/g9t6p13WQql1hSRLGBLDMSSGBwuu+so9+kbb+0px7y/Fse4o5asOA/p2Q5VhR2O6DWOKFckgvF8CQVPhldWvkH8yv07H7BDTgT/0/sNZ+1RUVNCtWzecTieHDx9m8eLFNfq8/fbbSJLEzz//zLp16xg1ahQ7duyo1p6fn8+QIUPYsWMHkydPJiwsjG3btrFp0yZ69OhR69zz5s2jS5cuAOzcuZNp06bRt29fFixYwM6dO1m9ejWapnHjjTfy3XffcezYMZKTk/nyyy8BKCkp4cSJE8yePZv8/HwkSaK4uPic78v69evZvHkzmZmZTJkyhcjISNasWYPL5SI3N5chQ4aQmZl5znEEOs1ebHmQMZuN2J0uNFXFtWt3Q5t10ZgNCrf2TuOWnJZ8t/MY76/Yy1+/2c5bi3ZdEXld54MSbsDSIQZLh0DNL5+G52g57v2luPfZce0vxbnlRKCzpFe8b2nTQ5AtbSgxIvdLIBBcGKFhxFWrVjFu3LgaeVorVqzg4YcfBqBdu3akp6ezY8eOau0dOnQItn/33Xc88sgjAHTt2pWuXbtWG2/QoEEoikLXrl158cUXKS4uJj09nb59+wKwYMECFixYQPfu3QEoKytj586dDBgwgMcff5w//OEPjBgxggEDBuD1ejGbzdxzzz2MGDGCESNGnPNn7t27d1BMLViwgE2bNgW9YCUlJezcuVOIrQugWYst86EVlDMUo8WMVubE2yqj0a1IvBhkWSKvfTx57ePJP1LK+8v3XpF5XeeDpEgYk60Yk62gfwbhK3Pj3m/HvV8vPVG+5ghlK/WK93K4ijFVF17GtAiMqVbksMa3mlMgENTkXB6oy0G/fv04fvw4x44dq9d5KnO2KikuLiY8vOrLsqZpPP300zzwwAM1rl2/fj1fffUVf/rTnxg8eDB//vOfWb16NYsWLWLmzJlMmjSJxYsXo6oqfr+ebuL3+3G73cExTp/rrbfeYujQofXxozYLmrfYch4CwGgxA2BPSyOskdTaqis6JEbw6phsnhzWng9+2M8HP+xj7JSjdE6J4N6rWjG8axIGpWml7ilWY/WyE5XerwP24MO545S+3yOgxlp08RV4GJLCkdSm9Z4IBIK6IT8/H5/PR4sWLXA4HMH2AQMG8N///pdrrrmGnTt3sn//ftq3b1+tfceOHcH2q6++mg8//JBrrrmGzZs3s2nTpguyY+jQoTz77LPcfvvtWK1WDh06hMFgwOv1EhMTw69//WuioqJ47733KCsrw+FwcMMNN5Cbm0urVq0AyMjIYN26ddx8883MnTsXj8dzxrkmT57MNddcg8FgYMeOHaSkpFQTZIKz06zFlkXWCzoZwsMAsLeIIWbRAfxOJ7LZ3JCm1TnxNnONvK5HP9nIy1/nc2duBrf2TmuU9brqgmrer0Dul9/pxX2wrEp87aracgg10D9EgInwo0DQfKnM2QLdyzNt2jQUpXo+6G9/+1sefPBBunTpgizLTJ06FZPJVK1dVdVg+4MPPshdd91FVlYWWVlZ9OzZ84JsGjJkCNu2baNfv36Antz+wQcfsGvXLp588klkWcZgMDB58mTsdjsjR47E6dQXEr3++usA3HfffYwcOZLs7GyGDRt2RvF07733UlBQQI8ePdA0jbi4OObMmXOhb2OzRmqsK7hycnK0tWvX1usc3788ho0FD9Ji+Cny1/1MJ5NK52n/JXPWZ1dckvyF4vdrLNtxjH8u38PK3ScINyrc3Ksld+dm0jImrEFtW7p0KXl5eZd1Tk3T8JW49PBjQIB5DpWheXQXuxxuqCa+jC1tyJbm9V2lIe6L4Nw09fuybds2srKyGtqMC8Jut2Oz2RraDMFp1OV9qe33UpKkdZqm5dTWv3n9tTgNq6oLTU3VPTplAc+Fa/fuJi+2ZFliUId4BnWIZ0thCe8v38t/Vu1j2soChnVO5N4BreiRVnN5c1NFkiTUKDNqlJmwrnEAaD4/niOOkPBjyJ6PBLYdOj382MRCsgKBQCC4dOpEbEmSNAx4A1CA9zRNe/m082nANCAq0Geipmlf1cXcl4LNoIstt8eH7PdT7vWCqjaaPRIvF52SI3n9lm48NawD00NS63MAACAASURBVFYV8N8f9vHVz0fomR7NfQMyua5jIkoj2YfxciIpsr6JdooV+oaEH0/L/XKsrww/yhiTwzGm2jCk6GFINdaC1AzfO4FAIBBUccliS5IkBXgbuA44CKyRJGmupmlbQ7r9CfhU07TJkiR1BL4CMi517kslwqj/EXRWuDHKEg63G2N6Oq7dV275h0shMdLMH4Z14KFBbZix9gD/+r6A33ywnrSYMO7OzWBMTkvCTc3aGYpsVjG3jcbcVvf6aZqG75SrSoAdtFO+9gjaSj38KJkUjClWDKk2jKlWjKk2lGiTyP8SCASCZkRd/OXsDezSNG0PgCRJHwMjgVCxpQERgdeRQGEdzHvJhBv1kI/b6cViMmIvc2Nq0wZn/rYGtqxhCTep3JmbyR39Mli49Qj/XL6X577YyusLd3B733TG98sgMbJpLSC4WCRJQo0xo8aYCcsOhB/9Gt4iB+6Ddj0J/6Cdsu8PgU/3pMrhKoaUKvFlbGlDsV15RWcFAoFAcH7UhdhKAQ6EHB8E+pzW5zlggSRJDwPhwLV1MO8lYzLrq0ncTi+RYWGcclSgZmbiWbiwSa5IvFAUWWJY5ySGdU5i/f5TvLd8D/9Ytpv3lu/hF12TuWdAJp2SIxvazEZHaOX78ECqpOb14zlSrguwA2V4DtmxL6kqP6FEGHXvV8uAAEsR9b8EAoGgqXC5YkK3AlM1TfubJEn9gP9IktRZ07Rqm/dJknQ/cD9AQkICS5curVejOhaX4JPclJy0E2bwgKKyuayUFL+flTNm4G3Zsl7nv9K4OQUGRVtYsM/Dl5sOMWvDIbJiZIZlGugSqyDXUWisrKys3u99gxKnPyQvmOxgKpEwlzgxFbgwbj0R7OYO03BFaLgiwRmp4YoArQGjuE3+vlyhNPX7EhkZid1ub2gzLgifz3fF2dwcqMv74nQ6L+j/XV18dB8CQlVJaqAtlHuAYQCapq2SJMkMxAJFoZ00TZsCTAG99EO9L2c+9TG+bS7MahjtOqRxdO16YjrrqxC7REUT2YSXU18KY4CSCg8frd7P1O8L+L91TtrEW7n3qkx+2T0F8yXuR9jUl7KfDb/Dg/tQWTD86Dlox3ckUNVZAjU+TPd8pVoxpFgxJIYjGy/P/o/N+b40Zpr6fdm2bVujKKPwP//zP3z44YcoioIsy/zjH//glltuYe3atdUqvQP06dOHH3/88YxjjRo1ir1791JWVsaxY8eC296888473HbbbbWOOXfuXLZu3crEiRNrHbOgoIARI0bU2EZIUEVdln4wm83BrZLOh7oQW2uAtpIkZaKLrLHAbaf12Q8MBqZKkpQFmIH63evgfFBN+GUXPrdGXGIiAMV+PxZFaRLb9tQnkRYDvxnYmrtzM/nq58P8c/keJs76mVfnb+eOfunc0TedFlZTQ5t5xSGHGaol4AP47O5g/pfnoB1n/gkc644GLgA1Liy4atKQYsWQbL1sAkwgaA6sWrWKefPmsX79ekwmE8ePH6+2tc3pfPvtt2cdb/bs2YAulF977TXmzZt3ThtuvPFGbrzxxgszXNBouOSiQJqmeYGHgPnANvRVh1skSXpBkqTK34zHgfskSfoJ+Ai4U2sM1VRVM5rsRHPLJKbqzrlTp07pKxKF2DovjKrML7unMO/hq/jovr50axnF37/dSf+XF/P0rJ/ZVVTW0CZe8Sg2I5asFkRel07sXZ1J+lNfEif2osUdWdjyWqJGmXDuOEXxF3s49u4mCv+ykiOvr+PkJ9uxrziEa28Jfpe3oX8MgeCK5fDhw8TGxmIy6V8gY2NjSU5ODp6vqKjg+uuv55///CcASUl6qZhKr+NNN91Ehw4duP322zmfP31vvfUWPXr0oEuXLuTn5wMwdepUHnroIQCOHj3KqFGjyM7OJjs7m5UrV1a7fs+ePXTv3p01a9YwdepURo8ezbBhw2jbti1PPfVUsN+CBQvo168fPXr0YMyYMZSV6Z/XEydOpGPHjnTt2pUnnngCgBkzZtC5c2eys7O5+uqrL+p9bM7USQZIoGbWV6e1/Tnk9VYgty7mqlNUM8hO8JiIidVXkpWUFGNq0wbX9u0NbNyVhSRJ9Gvdgn6tW7CrqIz3V+xl1vqDfLR6P9d0iOfeAZn0a9VClDyoA0ILsFo66aEGTdPwl7r1EOShMjyHynDuKq7agkjS94A0BDxgxkoPmLl5l/IQXFkceeklXNvy63RMU1YHEp955qx9hgwZwgsvvEC7du249tprueWWWxg4cCCg58yNHTuWcePGMW7cuBrXbtiwgS1btpCcnExubi7ff/89V1111Vnni42NZf369bzzzju89tprvPfee9XOP/LIIwwcOJDZs2fj8/koKyvj1KlTAGzfvp2xY8cydepUsrOz2bJlCxs3bmTDhg2YTCbat2/Pww8/jMVi4cUXX+Tbb78lPDycV155hddff50JEyYwe/Zs8vPzkSSJ4uJiAF544QXmz59PSkpKsE1w/jTvT1rVjCxXIHnDUFUVRfNTVu7A1KYN9m+/xe9yIZtEKOxCaRNv5X9Hd+GJIe344If9TF9VwG3//JFOyRHcN6Bpbn7d0EiShBJpwhJpCm7ADeArdeMu1MOP7kNluPeUULGxKoIfFGDJ1qAQa27bEAkE58JqtbJu3TqWL1/OkiVLuOWWW3j5Zb1298iRI3nqqae4/fbba722d+/epKamAtCtWzcKCgrOKbZGjx4NQM+ePZk1a1aN84sXL2b69OkAKIpCZGQkp06d4tixY4wcOZJZs2bRMWQXlMGDBxMZqa8c79ixI/v27aO4uJitW7eSm6v7QdxuN/369SMyMhKz2cw999zDiBEjGDFiBAC5ubnceeed3HzzzUH7BOdP8/5UVU3IcgWKR38bjLJMhcuNqU1r8Ptx792LuUOHBjbyyqWF1cTvrm3LAwNbMWfDId4L2fz6jn7pjO3VUuR11TNKhBFLRAyWDjHBNp+9UoCV4S4sw72vlIqfqgSY0sJcTXyJMhSCxsK5PFD1iaIo5OXlkZeXR5cuXZg2bRqgi5BvvvmG2267rVbPvSnkC7uiKHi95w7pV15zvv0riYyMJC0tjRUrVlQTW7XZoGka1113HR999FGNcVavXs2iRYuYOXMmkyZNYvHixbz77rv8+OOPfPnll/Ts2ZN169bRokWLGtcKaqd5iy2DBVUqRfEZ0DSNMLOJYpcTY5s2ALh27hJiqw4wGxTG9k7j5pyWLNt5jH+t2Mur87fzxqKd3JidzJ39M+icIup1XS4UmxFL+xgs7UMEWJkbT2F5IASpV8Kv+Pl41TVRJhKNMiWefboQSw5HiRKV8AXNg+3btyPLMm3btgVg48aNpKen8/PPP/PCCy/wwgsvMGHCBN55553LYs/gwYOZPHkyjz76aDCMCGA0Gpk9ezZDhw7FarVy222nr1Wrom/fvkyYMIFdu3bRpk0bysvLOXToEMnJyTgcDm644QZyc3Np1aoVALt376ZPnz706dOHr7/+mgMHDgixdQE0b7GlmlClCgw+ExXeCqzh4ZwoK4eEBFAUXLtFknxdIssSg9rHM6h9PLuK7ExbuY/P1h9k5rqD9MqIZnz/DIZ2SmxoM5slitWI0s6IuV3VKsjKMhSewjLcheUYdzmxL94fLMQqh6kYksIxJOv5X8bkcNTYMCRFCDBB06KsrIyHH36Y4uJiVFWlTZs2TJkyJbiK8I033uDuu+/mqaee4q9//Wu92/PGG29w//338/7776MoCpMnTw4m5YeHhzNv3jyuu+46rFbrGceIi4tj6tSp3HrrrbhcLgBefPFFbDYbI0eOxOl0omkar7/+OgBPPvkkO3fuRNM0Bg8eTHZ2dr3/nE0JqTEsCqyNnJwcbe3atfU7ycYP+fcnP1J26heMfT2b5bO+ZtOuPTww7tc4HvodxlaZtJw0qX5taOaUVHiYue4g01YWsP+kg8QIM7kJPp65ZaAIMTYyli5dytX9B+A5Uo6nsEz3hBWW4TlSDt7A54gqY0gMC3i/dA/Y5awF1hxpDnW2srKyGtqMC6Iu6zkJ6o66vC+1/V5KkrRO07Sc2vo3c8+WGaPkQEamtNxOi7g42LWHowcPEtu6Na6dOxvawiZPpMXAPVdlcmf/DJZuL2LqygI+23mcL15eLEKMjRDZqGBKi8CUFhFs03wa3mMO3IfL8QQ8YY5Nx9FWH9E7SKDGWXTvV1JAgCVbUcJFHphAIGgeNHuxZZEqKAZKHWXEJepu2GNFR0hp2wb7okViReJlQpElBmclMDgrgQ/nLWabN77WEKNYxdj4kJSqvSDpHg/opSh8xa5gCNJTWIZ7b2m1lZBKpLEqBBkIRyrRIg9MIBA0PZq52DJhkSoAsJc5SEvRE+NPnjiBsbVYkdhQJFtlbsvrzBND2zNj7QGmr9rHQx9uIDHCzB390rmlV0tiRYixUSNJEmq0GTW6qhYYgK/cg+dwSAiysBxn/slgHphkVjEmh+u5YEnhGJKsGOLDkAxCZAsEgiuX5i22DBYsOAAodziIio4GTaO0pBRTnr7qxLVrtxBbDUSkxcC9A1pxV25mMMT46vzt/P3bHdzQJYk7+qbTMz1aeEKuIJRwA0qbaMxtQhLx3T49D+xwedATVr76CJonsE+9DGpsWFCAGQPPss0o7r1AILgiaN5iSzVhRfdslTucKIqCgkZ5hQNjZoa+InGXyNtqaEJDjLuK7Hzww34+W3eQzzcW0iHRxq/7pvPL7ilYTc371/lKpdY8ML+G90SFLsAOl+M5Ul6jHpgcruqer8QQT1h8GJIqvGACgaBx0bz/OqlmbJLu2apw6EtfTaqC01mBbDRiTEvDtVOUf2hMtIm38dyNnXhqWHvmbixk+qp9/GnOZl7+Op/RPVL4dd902iWIVUBXOpIsYYgLwxAXBl3jgu1+hyfoBXMHRFjZD4fBW+kFkzDEWwICzBoUYYrN2EA/iUAgEAixhU0LiC2nLrbCzGZOVlSgaRrmrCwcGzY0pIWCMxBmVBnbO41berVkw4FiPli1j49X6/ldvTNjuKNvOkM7JWIUXo4mhRxmwNQqClOrqGCb5jvNC3a4DNfeEhwhyfiy1VDl/aoUYnEW4QUTnBdWqzVYOLQuKCgoYMSIEWzevJm1a9cyffp03nzzzTobX9D4aPZiyxrwbLkqPABEWK0ct5fhKC3B3LkzpV99hffECVRRKbdRIkkSPdKi6ZEWzZ9GdGTG2gN88OM+Hv5oA7FWE2N7teTWPmmkRFka2lRBPSEpEob4MAzxYZBd5QXzlVd5wSpDkWUrC6tqglVeFxqGTApHsQovmODykZOTQ05OraWZBE2IZi+2jJITALdL338qKjoajh7jROEhWnTuBIBzyxasV1/dYGYKzo+YcCMPDGzNfQNasWznMf77wz7eXrqLd5buYnBWArf3SWNA2zgUWSRVNweUcANK6yjMrU/zgh13BAWY+3A5zl3FODYUBfvIVoMuwBLCgiUt1IQwUZhVwNKlS3nuueeIjY1l8+bN9OzZkw8++ABJkpg4cSJz585FVVXy8vJ48803ufPOOxkxYgQ33XQTULuHbOnSpbz22mvMmzeP5557jv3797Nnzx7279/Po48+yiOPPNIQP6qgjmneYstgxhgo/eBx+gCIiYuH/B0cKTxESp/+IElUbN4sxNYVROi2QAdOOvho9X4+WXOAhVuPkhJl4ZZeLRmTk0pSpPB2NTckRcKQEI4hIRy6VbX7ytwBL5hDfz562opICZQYs35tYogIa2ER2xNdRpZ/uoPjB+ounAcQ29LKgJvbnXf/DRs2sGXLFpKTk8nNzeX7778nKyuL2bNnk5+fjyRJHDhw4KLtyc/PZ8mSJdjtdtq3b8+DDz6IwSAKAF/pNG+xpZiQJT8+2YPXpYcWEpKTAThRVIRitWLMzKRi48aGtFJwCbSMCeOpYR149Np2LNx6lI9W7+f1hTv4+7c7GNQ+nlt7p5HXPg5VFEtt1ihWI0obY7WSFJpfw3fSqYuvI+V4jupCzLntRLAuGGogkT8gvgyJYaiJ4SgRoixFU6V3796kpqYC0K1bNwoKCujbty9ms5l77rmHESNGMHDgwIsef/jw4ZhMJkwmE/Hx8Rw9ejQ4n+DKpZmLLRW/pOBX3PgDYisuSRdbp06eACC8f3+KP/0Uf3k5cnh4g5kquDSMqszwrkkM75rEvhPlfLLmADPWHWTR9LUkRJi4JaclN/dqSWp0WEObKmgkSLKEGmtBjbVg6VxVmFXz+PEUVXnAPEccOHdXD0VKFrVaGLLSGyabm/dH7qVyIR6o+sIUsqOIoih4vV5UVWX16tUsWrSImTNn8sYbb7Bs2TJUVcXv172jfr8ft9t9UeMLrnya/f98v2zEr7jwB/4P2Gw20DTspXb9+LrrOPXBB5St+J6IoUMa0FJBXZHeIpynhnXg99e1Y3F+ER+t3s9bS3bx1pJdDGgbx229WzI4K0FsDSSoFckgY0yxYkyxVmsPlqUIeMA8Rxw4NhShuXzBPkqkqXoYMjFcrIpsApSVleFwOLjhhhvIzc0lMzMTgIyMDNatW8fNN9/M3Llz8Xg8DWypoKEQYks2guJG8+gfdoqioEpQXqHncoX17IESFYV94UIhtpoYBkVmaKdEhnZK5OApB5+uPciMtQf4zQfribWauKlnKmN7tSQjVng0Beem1rIUlXtEBsSX50g53qN6Uj6+QCwy4EEzJIYF8snCUBPC9HwwsZjjisButzNy5EicTieapvHSSy8BcN999zFy5Eiys7MZNmwY4SI60mwRYks2ICkeJE/VN0uzquJ06KsUJVXFNnQoJZ9/jq+sDMVqPdNQgiuY1OgwHruuHb8b3JZlO4r48McD/HP5Ht5dtpt+rVpwc69UhnVKwiJWpAkugGp7RGZVlY/RvH68xyuqiTD3ATsVm45XXRzIB1MTwvSQZLwuxJQYsxBhl5nKFYR5eXnk5eUF2ydNmhR8vXr16uBru12PjCQkJPDDDz8E21955RVA93ht3ry5xpjPPfdctXkr+wiufITYko3IihvJWxUnD7eYcTjK8ft8yIpC1KhfUvzJJ9jnzyfqV79qQGsF9Y0iS1zTIYFrOiRwtNTJp4Hcrt9/8hN/Nm1hRHYSN/VsSY+0KJEALbhoJFUOhhJD8bt8eIsceiiySBdi7r2lVIQUaEWV9Sr5CXpJirAi8J6oQIkWIkwgaKwIsSUbkWUXiteKpmlIkoTNFsHREjv2k8eJjEvAnJ2NMTOT4tmzhdhqRiREmHl4cFsmDGrD6oKTzFh7kDkbCvlo9QFax4VzU8+W/KpHCvER5oY2VdBEkE0KxpY2jC2rbznld3rxFDnwHg0IsaPluAJJ+ckoHFm/Fskgo8YHvGAJYagJ+l6RSpRJiDCBoIERYks2YFBcGHwmnD4nFtVCdEw0FB7meGEhkXEJSJJE5KhRHHv9ddz79mFMT29oswWXEVmW6NuqBX1bteD5kZ34clMhM9Ye5JVv8nltwXYGtotjTM9UBmcliO2BBPWCbFZrbNYN4K/wsvqbFXROaR8UYs6dxTjWh6yMNFaKsPBqQkyJFOUpBILLhRBbsgmD7MLgM1LuKceiWmgRnwBspajwEK2zuwMQOfJGjv3975R8/jlxoqJvs8VqUrmlVxq39Epjz7EyZq47yGfrD7I4v4joMAMju6UwJieVTsmRDW2qoBkgW1Sc0WDtnVSt3e/w6OUpjlZ6w8pxbj+JY93RYB/JpOjCK0SIqfFhQoQJBPWAEFuyAaPkxOAzU+4uJ9YSS0KKXkDu+LGqb4eGhAS95tacOcQ+9BCSLDwYzZ1WcVaeGtaBx4e0Z/nOY8xYe5APf9zP1JUFdEyKYExOKjdmJ9PCajr3YAJBHSKHGTBlRGLKqC76feUeXXwVlQeFmHPbSRxrQ0SYUUGNt1Ql58eFocZbUGNEtXyB4GIRYks2YpIqkJEpdZRBJMTGxwNQfOpUtb6Ro35J4eNP4PjxR8L79WsIcwWNEEWWyGsfT177eE6Vu5n7UyEz1h3g+S+28j9fbmNguzhG9Ujh2qwEzAaxmlHQcCjhBpRWkZhanSbCytx6Yn5RReDZEcwJq7o4UKIiPuANi7egxodjiLUgGcSXT4HgbDR7seVTjFhwAGAv15f3hoeH64VNA8t3K7ENHoxss1E8a7YQW4JaiQ43Mr5/BuP7Z7D9iJ1ZGw4yZ8MhFuUXYTOp3NAliVE9UuidEYMskpYFjQTFakSxGqvVCAM9Md97rEJPyj/m0IVYYRkVm49XbVlUuW/kaZ4wQ3xYk6mYrygKXbp0QdM0FEVh0qRJ9O/fn4KCAjIzM/njH//Iiy++CMDx48dJSkrigQceYNKkSTz33HNYrVaeeOKJBv4pBA1J0/ifcAn4ZQNhlANgL9dFlyzLGGQJh9NZra9sNhMx/AZKZs/B9+yfUCIiaownEFTSPtHG09dn8dTQDqzafYJZGw7yxaZCPll7gJQoC7/snsyo7qm0iRe12wSNE9ms1ro6UvP48Ryv8oJVPjt3nqoq1gooEcaAF6xKgKnxYcjhhisqL8xisbAxsEfu/Pnzefrpp1m2bBkAmZmZfPnll0GxNWPGDLKyshrMVkHjRIgt2UiYpnu0yh0VwXaL0YCjwhEsB1FJ1JgxFH/8CSXz5hFz222X3V7BlYciS1zVNpar2sby4i+9LNhylFkbDjF56W7eXrKbrqmRjOqewi+yk4kV+V2CKwDJIGNMCseYVL1OmObT8J5y6nlhx6pEWPnao2juqm2L5DAVNS6wMjKuMiQZhhLZ+MtUlJaWEh1dtWF5WFgYWVlZrF27lpycHD755BNGjRrFiRMnGtBKQWNDiC3ZiM2v5yU4HFWerAirFXtZOa7ycswhVeMtnTph6phF8cyZQmwJLpgwo8ovu6fwy+4pFJU6mftTIbPWH+L5L7by4pfbyBP5XYIrGEmRMMRaMMRasBBSMV/T8JW4a3jCKrYcx19etdGyZJD1zb/jwwL5YRY02Y/m15BkiSVTp1C0b0+d2hyf3opBd95/1j4VFRV069YNp9PJ4cOHWbx4cbXzY8eO5eOPPyYhIQFFUUhKShJiS1ANIbZkIzZ/CQDOClewPTo6moPHT3LySCHJbarvNB91000cfeH/UbFlC5ZOnS6rvYKmQ3yEmXsHtOLeAa1qze8a2jmRG7OT6d+6BarYFFtwBSNJEmqUCTXKhLlddLVzenJ+he4JO1aB95gjsHXRMdDAd6MVT2EZKDL+Ci+a1w+ShCQBlVGHenaGhYYRV61axbhx46ptpTNs2DCeffZZEhISuOWWW/D7/fVrkOCKo07EliRJw4A3AAV4T9O0l2vpczPwHHpa5U+apjUKt5BfNhCh6YnwzoqqHdlj4xNg1x6O7N9fQ2xFjhhB0St/pXjmTCG2BHVCaH7XD3tOMGv9IeZvPsLMdQeJtRq5oUsSN2Yn0yMtWiTWC5oUVcn51VdIah4/3hMVlBzdixJhRPNqXP2rO9E8ftCq8sKQJCRVRjLoz6iyfqzK9RKS7NevH8ePH+fYsaotlIxGIz179uRvf/sbW7du5dNPP63zeQVXNpcstiRJUoC3geuAg8AaSZLmapq2NaRPW+BpIFfTtFOSJMVf6rx1hV82YpT0XC2Xs0psJaa2BFZRdKSwxjVKRAQRw4ZS+sU8Ep58Ejks7HKZK2jiKLJEbptYctvE4vR0Zun2Iub+VMgnaw4wfdU+UqIsjMjWhVfHpIgrKslYILgQJIO+f6R8SkGJqMpl1DQN/Bqax4/mrXxo+F1+cHirD6LISAYZSZWqBJhBBlm66P87+fn5+Hw+WrRogcPhCLY//vjjDBw4kJiYmIsaV9C0qQvPVm9gl6ZpewAkSfoYGAlsDelzH/C2pmmnADRNK6oxSgMRKrY8zqoEzvgkvSLzieO1x92jbrqJks/nUjp/AVGjfln/hgqaHWaDwrDOSQzrnITd6WHh1qPM/amQ95fv5R/L9tA6Lpwbs1O4sVsymbHh5x5QIGgCSJIEioRUS2hd82tBAUalEPNo+F2+mt4wQ6UHTNK9YQYZSandG1aZswW62Js2bRqKUj2nslOnTnQSkQ7BGZC00F/AixlAkm4Chmmadm/g+A6gj6ZpD4X0mQPsAHLRQ43PaZr2TS1j3Q/cD5CQkNDz448/viTbzoeYvXPpXPBvJh/9jAOt1jOsdw4Afr+f75Ytw+apoOeQG2peqGm0eO45/DYbp0T9lDqlrKwMq1WUQzgTdrfGmiNefjzsZccpPxqQGSHTJ0mlT5JCtLl+8rvEfWmcNPX7EhkZSZs2bS5tEE1/SD6Q/SCd9qjWVa758CvoeWHn6Qzz+Xw1xJig4anL+7Jr1y5KSkqqtQ0aNGidpmk5tfW/XAnyKtAWyANSge8kSeqiaVpxaCdN06YAUwBycnK0vLy8ejds25GlyJIfn+zFrIYROucPK5bjcfoYOHBgrS7nE3v2UPTqa/RLS8PUqlW929pcWLp0KZfj3l/J/CLwfLikgnk/HWbuT4V8vL2ET3ZA74wYRmQnM6xTInG2uislIe5L46Sp35dt27Zhs9nO3fEiCfWGaV4/ePSwZM3cMHRvWqUXTK0KT54elrTb7fVqs+DiqMv7Yjab6d69+3n3rwuxdQhoGXKcGmgL5SDwo6ZpHmCvJEk70MXXmjqY/5Lwywb9WfXgd1U/F2mzccJRTtmpE9hiYmtcGzlyJEX/93eKZ35GwlNPXg5zBYJqJEVauO/qVtx3dSv2HCvji58OM/enQzw7ZzN/+XwzvTNjGN4liaGdE4m3mRvaXIGg0SHJEpJRAWN1j0cwNyyQE6aHJPXwpN/lraqgD1VJ+qoEBhnZA363r96S9AVXHnUhttYAbSVJykQXWWOB01cazgFuBf4tSVIs0A6o22IpF4lfNgKgqV78nurnUlNTOVZcwqGdO+jQp6bYUmNjYO3NXAAAIABJREFUsQ0aRMmcOcQ/+jsko/FymCwQ1EqrOCu/u7Ytjwxuw46jZXz582G++vkwz36+hT/P3ULvjBiGd01imBBeAsE5qZYbdpqDWNM08FX3iGleDb/HDxVeDIC3IpA8L1dPzg/miJ0hP0zQNLlksaVpmleSpIeA+ej5WP/SNG2LJEkvAGs1TZsbODdEkqStgA94UtO0RlHxzafoAklSveCu/s2mTVYWG7ZsZfe2LXTo07/W66PG3IR94ULsS5YSMXRIvdsrEJwLSZJon2ijfaKNx65rx46jduZt0oXXnz/fwl+E8BIILglJkqAyhHgamqZRXlJGmMlSTYz5XV5wnJYjrZy2UjLwQL341ZKCxkmd5GxpmvYV8NVpbX8Oea0BjwUejYpKz5Zk8CF5q4uttPQMAAr2nNkJF56bi5qURPHMmUJsCRol7RJsPHZdlfD6ctNhvgwRXr0y9FDj9Z0TiY8QwksguBQkSUJTQLbU/PNac7Wkfuyv8IK/uhCrqhlWXYyhCCF2JSIqyAdytmTVh+xSq+2FaLPZsBgNnDp+igNbNtGyU9ca10uKQtTo0Rx/5x08hw5hSEm5rPYLBBdCuwQb7a6z8fsQ4fXVz4f5y9wtPPfFFnqlx3BDl0SGdEokOcrS0OYKBE2KC8kPI+gR06on6lOLEFPkoKdNCLHGSbPfA6TSs6WqXgw+I26/u9r57O498Nqi+Gb6v/A4nbUNQdToUQAUz5pdv8YKBHVIuwRddC18bCALfn81j1zTllMON899sZX+Ly/mxkkreHvJLnYVlTW0qQJBg6IoCt26daNTp05kZ2fzt7/9Lbglz9q1a3nkkUfOev3UqVN56KGHznhekvTcMNmkooQbUCNN/PUf/4chIRxDcjiGpHDUOAtKtBnZZtQLs/o0/OVefMUuvCcq9M2/D5XhPlLOuLG/JjM9g25ds+nRrTvfL1+B5j+/Mk+htr777rtMnz79PN+li6OgoACLxUK3bt2CD7fbfe4Lz5OpU6dSWFhVnPzee+9l69atZ7mifhCerUqxZfBj8IVT5i7DZKnKhhyYl8fGDRs47DLxxaS/Meqxp5Hk6hrVkJJCeG4uxbNmEfvbB5FEfRXBFUaox2tXURnztxxhwZYjvDp/O6/O306beCsdrG5i2hTTJSVSfHsWNCtC90YsKiritttuo7S0lOeff56cnBxycmotrXRJvPTSSzzzzDPnTtT3a1WrJQMeMYD//eOLjL5hJAuXLeI39/+GdQtX1ZIjdvZk/d/85jcXZLPX60VVL1xWtG7dOvj+1jVTp06lc+fOJCcnA/Dee/+fvfsOj6pK/D/+vtNn0ie9kR5CgARCIJQIoQgo0nsTxLK66q7uT8S1rOKyrru63107IE0ghKqgogICoUqkiAiETuikJ6Rnyv39MckQBKVNCnBez5NnyMy5557JhJlPzj1lVr2c53pEz1bNZUSNyozaoqXcVH7F43q9nuEjR4LewP68ItLTrp3y3UeOwHzhAqWbNtd7mwWhPkX6OPN090hWPZPM9pd6MHVAS7ydtXxz0sSAD7eR/K+NvPHlAXacyMdyg38tC8LdwsfHh5kzZ/Lhhx8iyzLp6ek89NBDAPz444906tSJ5ORkOnfuzOHDh+3HnTlzhpSUFKKiopg6dar9/oULF9KhQwfatGnDH/7wBywWCy+99JJ91fqxY8f+Zjmr1cojj04irl08bTom8MGcj1F56lHoVKiMOtT+TnR/6H6OnzqB0lVD2qqldO7dlXbJHfjDU09SlVOGObucWf+dTlREJO0TEtmyYTNytQVLuYnXX/sb77zzDgA7d+4kLi6ONm3aMHnyZFq1agXYwsyAAQPo0aMHPXv2pKysjEmTJtGhQwfatm3LqlWrANuCopMnT6Z9+/bExcUxY8aM3/05112od/ny5UycOBGAiRMn8qc//YnOnTsTHh7O8uXL7eX+9a9/0bp1a+Lj43nppZdYvnw5u3btYuzYsbRp04aKigpSUlLYtWsXAGlpabRu3ZpWrVoxZcqUK879yiuvEB8fT8eOHcnOzr65X5JrED1bCtufClqVBbVFS6np6ksmERERjB4zhkWpqWzZdwC96xd07D/4ijIuKSmofHwoXLIYlx7dG6TtglDfAtz1TOgcyoTOoXy9diMVxkjWHLjIoh9PM297FkYnDb1a+NC3lR+dI7zQqUWvrlB/ir46TvX5MofWqQlwwr1/xE0dEx4ejsViISfnyp3nYmJi2LJlCxUVFWRkZPDyyy+zYsUKwBbE9u/fj8FgoH379vTr1w8nJyeWLFnCtm3bUKvV/PGPfyQ1NZW3336bDz/80N7bk5mZec1yLVu25Ny5c+zfvx+AoqIr1glHUipYvfYbWrduzZFzJ1i++nO2//gDarWap556iiXrV9KrWw/+/r+3yFi/HVeDC/cP7kt8yzgsBZVYy0xY5GqqL5QycfwEpr/3MZ07deKvU18FGWSLFVmW2bNnD/v27cNoNPLyyy/To0cP5syZQ1FRER06dKBXr16kpqbi5ubGzp07qaqqokuXLvTu3RtJkjh+/Lh9O6QuXbrw0Ucf/e7P/8KFC2zdupVDhw4xYMAAhg0bxrfffsuqVavIyMjAYDBQUFCA0Wjkww8/5N133yUxMZGSkhJ7HefPn2fKlCns3r0bDw8PevfuzcqVKxk0aBBlZWV07NiRf/zjH7z44ot8+umnvPrqqzf1O/JrImzV9GzplGYUKLhUVgqeV5eLiopi1OjRLE5LY+32Haj1Btr16mN/XFKrcR82jLxPPqH67Dk0QWKgvHB3cdZIPJQYzPDEYMqqzKQfzmXNgYt8+8tFlu46i5NGSUqMD31a+pHS3BtXnbqxmywIDaq4uJgJEyZw+PBhlEolJtPlxRvvv/9+PD1tHy5Dhgxh69atqFQqdu/eTfv27QHbHow+Pj5X1bt+/fprluvfvz8nTpzg2WefpV+/fvTufXlG/OTJk5k2bRre3t7Mnj37mnX4+vqy65c9pHTvjn+0bW3yUePHcPjwYVQ+BiS9CoVOyaXKUkrLSkmKa4/lUjUj+w5h9derMV0ow1JURc/k7rhY9ZiLqlj73Rq+XPUl777zLkhQWVnJ6dOnWbt2Lfv27bP3RBUXF3P06FGio6Nv+jLioEGDUCgUxMbG2nudvv/+ex555BEMBgPAdTcE37lzJykpKXh7ewMwduxYNm/ezKBBg9BoNPbeynbt2rFu3bobbttvEWGrJmzplbbd4kvLy3+zbPPmzRk3bhypCxfw9cZ0VDot8ckp9sfdhw8jb/p0ipYtw+f55+q13YLQmJy0KvrF+dMvzp8qs4Xtx/NZe+Ai6w5ms3rfBVQKiaRwI71a+NKrhS/BRkNjN1m4C9xsD1R9OXHiBEqlEh8fHzIzM+33v/baa3Tv3p358+eTn59/xTZKvx7nKEkSsiwzYcIE/vnPf/7u+X6v3M8//8yaNWuYPn06S5cuZc6cOQC88847DBs2zF5u48aN16xj5cqVV9UpSRIKjRKFWmm7JOmhA4WExt8JWZZRZuuQVAqU7lokrRInJyewWLFWWbCaraR9PJ/mEVG2ymrGiVmrzPzv7f+jT98+NTMobePQsrKyrvmc6/68Kn81OU2rvTx47Xb3d74WtVptP79SqcRsNt92nff8mC0kBSi1GJS2v0BKy347bAFEREby8MMTkFRqVn63jl92/GB/TO3vj3NKCkUrViA7cDaFIDRlWpWS7s19+OeQODJe7sXyJzvx2H3hZF+qYupXB7nv3xvp+7/NvLPmED+dLsQqxnkJd7Dc3FyefPJJnnnmmasCVHFxMYE1y//MmzfvisfWrVtHQUEBFRUVrFy5ki5dutCzZ0+WL19uvxxZUFDAqVOnANsHfm3P2G+Vy8vLw2q1MnToUKZNm8aePXt+s92/VUdSUhKbNm0iPz8fk8nEsmXLrjrW3d0dFxcXMjIykCSJpSuWgQRKZw1KgxqFXmWfOdnnwb5MXzwLhbsWhauGnw//AjL0uq8H02dMpzK7BHNOOQe27aXwWDam3AqwyJiLq7CUVWOtNCObrfj6+pKZmYnVauWLL64/0//+++9n7ty5lNd0mBQUFAC2JZzqXj6s1aFDBzZt2kReXh4Wi4W0tDS6det23fPcqnu+ZwsAlQ4npS0clVVUXLd4aHg4EydM4LN58/h89WoklZJWiR0A8Bg1ktINGyjZsAHXvn3rtdmC0NQoFRKJoUYSQ4289EAMJ/PKWJ+ZzfeZ2UzfdIKPNh7Hy1lLzxgfesX6khzphV4jxnkJTVvtYHWTyYRKpWL8+PH85S9Xr9H94osvMmHCBN5880369+9/xWMdOnRg6NChnD17lnHjxtlnME6bNo3evXtjtVpRq9V89NFHhISE8MQTTxAXF0dCQgKpqanXLKfX63nkkUfsy1D8Xg9ZbGzsNevo2LEjb7zxBp06dcLd3d0+durXZs+ezeOPP45CoaBbt264ubldVUaSJP72+t947rnnaNupHVarlbCwML7++muefOEZzhReoGP/bsiyjJenFysWLAGlrYfPWlp9xX6T0ya/zkMP9MPLy4t2bRMoKy/HWmGyLQx7jT/Y+vbty969e0lMTESj0fDggw/y1ltvMXHiRJ588kn0ej1r1661l/f39+ftt9+me/fuyLJMv379GDhw4G/+/G6XVB9dcI6QmJgo184YqE/p6emk7Hyco55jWftjV/RDcpjUe9QNHXvm+DHmzZ2LVaFkyKCBtE5oh2yxcLx3H9TBwYTMm1vPrb87paenX9H9LjQNt/u6FJVXk344l+8zs9l0OJeSKjNalYLkSC96xfrSM8ZHrGB/C+72/y+ZmZm0aNGisZtxU0pKSnBxcWnsZjhUaWmpfYbg22+/zYULF3jvvfccVv8V+01a6izsWvvvXwes2j0nldLlBV6Vv7/KviNfl2v9XkqStFuW5WuuAyJ6tgDUOpylKgAqKq69cOm1BEdE8vD48cyf/xkrVn2JFYhPaIf7iBHk/ve/VJ04iTY8rJ4aLQh3FneDhkFtAxnUNpBqs5UfTxbwfU2v1/pDtksb8UFu9GrhS/cYH1oGuIr1vAShiVi9ejX//Oc/MZvNhISEXHWZ9Hb93n6TUDPzsTZ41d0E3GRFrrjGmCpl3QBmC2SNSYQtAJUOLbbpxJWVpusUvlJI8xjGjR7NwoUL+WLVl5gtFuKHDiH3gw8oWroU35emXL8SQbjHaFQKkqO8SI7y4vX+sRzOLmF9Zg7rDmbzn3VH+M+6I/i4aOne3IfuMd50ifTCRcxuFIRGM3LkSEaOHNlo55eUCiQlV211BL/RK2axbX1UuwG4pFZAI87TEWELQKVFLdvCVnXlzc86CGsVx6jhw1m8ZAlffb2a6t69Cb6/F8VffIH3c39GoROXRgTht0iSRIyfKzF+rjzdPZLckio2Hcll4+Ecvtl/gSW7zqBSSLQPNdI9xpvuzX2I9HEWvV6CIAA30Ctmta20T4Vj12i7GSJsAaj0aLDNVjBVWm6piqh27RluNrF82TLWrFtHl4R2BH37HSVr1uBWj4PuBOFu4+2iZVi7IIa1C8JssbLndBEbD+ew8VAOb31ziLe+OUSgu94evDpFeGLQiLcyQRCuTVJIcI3tiBqSeIcCW8+WxRa2zFXWW64mJqkzQ0zVfP7552w7fozY+5LRpS0WYUsQbpFKqaBDmJEOYUam9I3hQnEF6Ydz2Xgoh8/3nGPhjtNoVAo6hnvSvbktfIV6OTV2swVBEK4gwhaASoeiOg+r0oyl+vZmZ8Ymp2CuruKrVV9yMDCQqiNH8TtwAH3Llg5qrCDcu/zd9Izu0IzRHZpRZbaw82ShrdfrcA5TvzrI1K8OEublRLdob7pGe5EU5omTVrzNCYLQuMS7EIBaB+YqrCoL1urb72qM69EHc1UV337zDcejo/hq8WKGv/mmGGMiCA6kVSntg+xfeyiWU/lltl6vwzks3mnbu1GtlEgMMXJftBddo7yJ9XdF0ciXE4Q708qVKxk8eDCZmZnExMQ0dnOEO4wIWwAqHZgqQG0BB4QtgIQHBmCqquL79es56OnHd19+Sd8BA0TgEoR6EuLpxITOTkzoHEqlycKurEK2HM1l05Fc/v3dYf793WG8nDUkR3rRNdqb5CgvfFzE5BXhxqSlpZGcnExaWhpTp05t7OYIdxgRtgBUWjBXgVoGk+NWs04aNJyKc+fYdvgIGT/9hMbZmZ49ezqsfkEQrk2nvtzr9dcHW5BzqZItR/PYcjSXLUfzWLn3PAAt/F3pGmULX+1CPNCpxWr2wtVKS0vZunUrGzdupH///kydOhWr1cozzzzDhg0bCA4ORq1WM2nSJIYNG8ZPP/3Ea6+9RmlpKV5eXsybNw9/f//GfhpCIxJhC0ClB3MlCg0oKh37I+n2xz9TNGYYBySJLVu2oFKp6nX/JUEQrubjqmNouyCGtgvCapU5eOESm4/msuVIHnO2nWTG5hPo1LaB9vdFedMt2osIb7G8RFPz7bffcvHiRYfW6efnxwMPPPC7ZVatWkXfvn2Jjo7G09OT3bt3c/LkSbKysjh48CA5OTm0aNGCSZMmYTKZmDx5Ml9//TXe3t4sWbKEV155xb5BtHBvEmELanq2KlFqQVWqodpSjUapcUjVkiTRfcITVL07jaOt4tm4cSNKpZLk5GSH1C8Iws1RKCRaBbrRKtCNP6ZEUlZlZseJfLYczWPzkVz+fvggfwf83XR0ivAkOdKLLpFe+IqthO5ZaWlp/PnPfwZg1KhRpKWlYTabGT58OAqFAj8/P7p37w7A4cOHyczM5P777wfAYrGIXi1BhC3ANmbLXInKSYHaqqHMVOawsAXg2qsnbd9+GynnAkckie+//x6VSkXHjh0ddg5BEG6Nk1ZFzxa+9GzhC8CZgnK2HM1j27E8+xITABHeTnSJ9KJzhBedwj1xM4gV7Rva9Xqg6kNBQQEbNmzgl19+QZIkLBYLkiQxePDga5aXZZmYmBh+/PHHBm6p0JQ17mZBTYVaB7IVtVaB2qKj1FTq0OollQrjqJG03HOQVkZ3VJcK+e6779i7d69DzyMIwu0LNhoYk9SMj8YmsPvV+1n9p2RefjCGIA8Dy3ad5cmFu2n797UM+HArb397iC1Hc6movrXFkIWmb/ny5YwfP55Tp06RlZXFmTNnCAsLw2g0smLFCqxWK9nZ2aSnpwPQvHlz8vLy+OGHHwAwmUwcOHCgEZ+B0BSIni2w9WwBWq2E2qKl3FTu8FO4Dx9O3kcf00FWQ6APv1zIZ9XKlRgMBqKjox1+PkEQbp9CIdEywI2WAW480TWCarOVvWeK2HYsj+3H85i15QTTNx1Ho1SQEOJOlwgvOkd6ER/khkop/pa9G6SlpTFlypV73A4dOpTMzEyCgoKIjY0lODiYhIQE3Nzc0Gg0LFiwgClTplBcXIzZbOa5556jpVhr8Z4mwhbYw5ZGq0BtUVFa7dieLQCVpyeuDz5AycpVPLhhPdZPP2B/3iWWLF7MxEceITg42OHnFATBsTSqyyvaP39/NGVVZn7MKmD7sTy2Hcu3b6LtrFWRFGakc6QXHcONtPAT63vdqTZu3HjVfX/6058A2yxFZ2dn8vPz6dChA61btwYgLi6OzZs3N2g7haZNhC2why29RoECBSXljg9bAB5jx1K86ktKv17NwOdfwvyft8gsLmfBZ5/x+B/+gLe3d72cVxCE+uGkVdG9uQ/dm/sAUFBWzQ/H89l2PI/tx/JYfygHADe9mg5hRpLCjHQM96SFvytKEb7ueA899BBFRUVUV1fz2muv4efn19hNEpooEbbAHrYMWlu3f2l5Rb2cRh8Xh651awpTU/EYM5ohf3mZZe/8gyPlFcyZ9SlPPf0Mrq6u9XJuQRDqn9FJQ784f/rF2WafnS+qIONkPjuOF7DjZD7rDmYD4KJT2YNXUpgnsQEifN2JasdpCcL1iLAFtqUfAINGAmRKyxw/ZquWcfw4zr84hbKtW3Hu2pXhL7zM4n9P41h5BbNmTOepZ55Fr9fX2/kFQWg4Ae56BrcNYnDbIAAuFFeQcaKAHSfyyThZwPeZtp4vF52KDqFGksJtASzW31WM+RKEu4gIWwBqW7hxrglb5RWV9XYq1759yXnnXQrmfYZz166oNBpGvvgKqf+aRlZpKXNnzuCJp59BpRIvjSDcbfzd9AxqG8igtoEAXCyutPV8nSgg40S+/bKjs1ZF+1APW89XuCetAkT4EoQ7mfhEB3vPlnPNsjmV5dX1dipJo8Fj3Dhy//tfKg8fQdc8GrVGy9gXX2He29M4V1jEwtmfMuGJJ8Xq1YJwl/Nz0zGwTSAD29jCV/alSjJO2nq+dpzIZ+PhXAAMGiVtm7nTPtRI+1AjbZu5Y9CIt29BuFM45H+rJEl9gfcAJTBLluW3f6PcUGA50F6W5V2OOLdD1IzZ0inNgIKKiqp6PZ3HyBHkTZ9OwWefEfDWPwBQa3VMnPIqM97+O1kXslmxcD7Dxk+o13YIgtC0+LrqGBAfwID4AABySirJOFHArqwCfswq5L31R5FlUCokWgW40j7UiL7MTKvSKryctY3cekEQfstthy1JkpTAR8D9wFlgpyRJX8qyfPBX5VyAPwMZt3tOh6sJW2plFaCnutJcr6dTurvjPngQRcuW4/P8c6hqZiGqdToen/IKH739FvuPn8R5xVL6Dh1Rr20RBKHp8nHR0T8+gP414etSpYndpwrZlVXAzqxC5u84RbXZygc/fU+4txPtQ4y0DzPSPtSDZkaD6B13gPz8fHr27AnAxYsXUSqVeHt7k5WVRUBAAAcPHrxODZetXLmS6OhoYmNjf7dcVVUVDz/8MLt378bT05MlS5YQGhp6O09DaGSOGATQATgmy/IJWZargcXAwGuU+zvwL6D+BkTdqtqwJdl6tKqr6jdsAXiMH49sNlOYlnbF/Vq9gacmT8EgW9ixbz9bvv263tsiCMKdwVWnpntzHyb3iWHpHzrxyxu9eSVJx0sPxBDm6cR3By7ywrKf6fZOOklvrefp1D3M3XaS/eeKsVjlxm7+HcnT05O9e/eyd+9ennzySZ5//nn79wrFzX2Erly58obC2ezZs/Hw8ODYsWM8//zzVy2qKtx5HHEZMRA4U+f7s0BS3QKSJCUAwbIsr5YkafJvVSRJ0hPAEwC+vr4NMq22tLSUHbt+oiOQdXwf0JXigpIGObdb69bkzF/AgZgY0Fy5F2N8Ukd2bt/O+u07OHMyi4DYVvXenqaitLRUTKlugsTr0jT5qytwls8QEwpjQ9ScL1VxpNDC0UIzO45eZPUvFwDQKSHMTUGku5IId9utbVJQ0+bm5kZJSUljNwOw9Tip1WpKSkooLS3FZDIxceJEMjIy8Pf3Z/Hixej1eo4dO8bkyZPJz89Hr9fzwQcfUFhYyKpVq0hPT+fNN99kwYIFbN68mblz52IymQgPD2fmzJkYDAZWrFjBX//6V0pKSujTpw9PP/00ly5dEj2Vt8lisTjsd6mysvKm3g/rfYSlJEkK4P+AidcrK8vyTGAmQGJiopySklKvbQPbOikdEztABrSMDmLzNtCq9DTEucv0Bk5PmEB8UREeI66+XNimbVtmTp/B0bPnCA30p/PAofXepqYgPT29QX7+ws0Rr0vTdL3X5VxRBbuyCth9qpA9pwv5JqvE3ssV7uVEQogHCc08SAhxJ8rHpcmt95WZmYmLiwsAR478nZLSTIfW7+Lcgujo126orFarRavV4uLigrOzM8ePH2fJkiW0adOGESNGsHbtWsaNG8fzzz/Pp59+SlRUFBkZGUyePJkNGzYwcOBAHnroIYYNGwZAUFAQzz77LACvvvoqS5cu5dlnnyU7O5uYmBj783Z3d6e6uhovLy+HPvd7TUlJif1nert0Oh1t27a94fKOCFvngLp7zQTV3FfLBWgFpNekcj/gS0mSBjSZQfI1sxEVlkqsSjPW+h0fb2fo0B5tbAsK5n2G+7BhSL/qkvb1D+DhSY/w2dx5fL99B9Wll+g2ZqL460YQhBsW6K4nsM6Mx/JqM/vOFrPndCF7ThWy4VAOy3efBWxLTrQJdiehmTttQzxICPbAzaBuzOY3aWFhYbRp0waAdu3akZWVRWlpKRkZGQwfPtxerqrq2h8q+/fv59VXX6WoqIjS0lL69OnTIO0WGp4jwtZOIEqSpDBsIWsUMKb2QVmWiwF7HJckKR14ockELbCP2cJciay2IFc3TJiRJAnPiRM5/+IUStM34dKj+1VlQkJCGT5iBEuWLmXrnr2U5r1Dn6f+jFojZh4JgnDzDBoVHcM96RjuCYAsy5zKL7eFr9OF7DlVxIcbj1E7xCvSx5mEZu41vV8eRHo7N9o+jzfaA9VQtNrL78NKpZKKigqsVitubm7s3bv3usdPnDiRlStXEh8fz7x58+yXpQIDAzlz5gxBQUGYzWaKi4vx9PSsr6chNIDbDluyLJslSXoGWINt6Yc5siwfkCTpTWCXLMtf3u456p1SA0hgrgK1jGxquMUDXR94gNz33idvxnScu6dcs9eqRWwsffv25bvvvuPn4ycpfuMlBr7wKs5G8Z9PEITbI0kSoV5OhHo5MSTBttJ9WZWZn88U1QSwItYezGbpLlvvl4tWRVywG/FB7sQHu9Mm2B1fV11jPoUmxdXVlZCQEJYtW8bw4cORZZl9+/YRHx+Pi4vLFWOGSkpK8Pf3x2QykZqaSmCgrfdxwIABfPbZZ3Tq1Inly5fTo0cPcUXjDueQMVuyLH8DfPOr+/72G2VTHHFOh5IkW++WuQJJLSOZlQ13arUaz8ce5eLUNynP+BGnjknXLNexY0eKiorYsWMH5/LOk/ry8/T/y8sERMc0WFsFQbg3OGlVdI70onOk7aKELMuczCtjz+kifjpdyL6zxczcfAJzTfeXn6uO+GA3W/gKcqfIiu3gAAAgAElEQVRVkBuuunv38uOsWbOYPHky06ZNw2QyMWrUKOLj4xk1ahSPP/4477//PsuXL+fvf/87SUlJeHt7k5SUZA9ijz76KOPHjycyMhKj0cjixYsb+RkJt0ssQVxLpQVzFQoNqMrVmCwm1MqGebNwGzKE3I8+Jn/mzN8MWwC9e/emqKiIQ4dAfymXJW9M4b7RE2j30GDxV48gCPVGkiTCvZ0J93ZmWDtb71elycLBC5f4+UyR7etsMWsOZNeUhwhvZ+KC3GgT7E58kDsx/i5oVQ33h2x9eOONN+z/Dg0NZf/+/fbvX3jhhSse++677646vkuXLlcs/fDUU0/x1FNPXVVOp9OxbNkyB7VaaApE2Kql1oO5EpVWQl2io8xUhrvSvUFOrdBq8XxkIjnvvEvFL7+gb9362uUUCoYMGcK8efPIVasI9/Vj08I5nMncT98/Po/e2TGzLARBEK5Hp1baxnE187DfV1Rezb6zxTXhq4jNR/L4fI9tvpRGqaBFgCttgmw9YHFB7oR7OTXa+C9BaEgibNVSacFUiUqrQG3RUmYuw52GCVsA7iNHkTdjJvkzZxL0wQe/WU6j0TBmzBhmzZrFWZOJTqMnsHNpKgum/IkHn/5/BN1D63EJgtC0uBs0dI32pmu0bVcMWZY5X1xpD18/nyli+e6zfPbDKcA2+zE2wJVWAW60DnKldaAbYV7OTW75CUG4XSJs1VLpwFyJRq9CbdVSWl3aoKdXOjthHDeWvI8/oerYMbSRkb9Z1tnZmTFjxjB79mx+PpfDkFensW76/1gy9SXaPtCf+0ZNQK0TA1YFQWhckiTZlp5w1/Nga38ALFaZ47ml7D1TxP5zxfxyrphFP56icpsVsG263TLAlZYBbrQOdCNKa8Vqtd70au2CUF9k+eZ3YxBhq5ZKB+YqtFo1aouWcnN5gzfBY/x48ufOI//TWQT865p7edv5+PgwcuRIFi5cSPrO3Yz95//YvmQBP337FSd/2kWfp54jKKZlA7VcEAThxigVEtG+LkT7ujAi0bZEo9li5XhuGb+cK7YHsCU7zzBvexavdPXErD6Di7sRg1aFXq1Er1GiUynEWFWhwcmyTH5+Prqb7NAQYatWzWxErd4Wthq6ZwtA5eGBx4gRFCxciNczT6MJDv7d8uHh4QwYMICVK1eyZt33DHzkSaI6dGHN9PdY8sZLxN//IMkjx6Nzdm6gZyAIgnDzVEoFzf1caO7nYh+Ab7HKnMgt5cDZQnILiygpLsRssdrX/1JItuM0Sgm1UlHzJTWJAFZZWXnTH8ZC/XPU66LT6QgKCrqpY0TYqqXSgrkSg5sOBZWUlJc1SjOMkyZRuHgxeR9/QsA/37pu+TZt2lBYWMimTZvw8PCgW7duTHj3Q7amzWfvmtUc2bGVbuMmEdtVrNMiCMKdQ6mQiPJ1IcrXBWgG2ALYybwye+/XL+eKOXj+EqVVZvsxEd5OtPB3JdbfldgAV1r4u+Ll3LCLQKenp9/UVi5Cw2jM10WErVpqPVQWYTDYUm9ZeUXjNMPXB49RoyhYsADPJx5HGxZ23WNSUlIoLCxk48aNeHh4EBcXR49H/kDLlF6sn/0x3338X37ZsIYejzyJT2h4AzwLQRAEx1MqJCJ9nIn0cWZQW9sCoFarzOmCcjIvXOLghUtkXrjEzpMFrNp73n6cj4vWHrxqQ1iop5MYiC80GBG2atXMRnRy0gNQ3khhC8DziccpXLqUvI8+JvDdd65bXpIkBgwYQHFxMatWrcLV1ZXQ0FB8wyIY/eY77N/0PVtS57HgpT8Tm5xCl1HjcfXyqf8nIgiCUM8Uissr4D9QMwgfbMtQHLxwiYPnL5F5oYSDFy6x7dgJTBbbdUi9WklzP5crQliMnwtOWvGxKDie+K2qVTMb0UlvC1tl5ZWN1xRPT4zjxpE/axaeTzyOLjr6+seoVIwaNYpZs2axePFiHnvsMby8vJAUClp3701Uh878uGo5e75ZxeEdW0l4YAAdBg1H5yTGcwmCcPdxN2joHOFF5wj71rxUm60cyym194AdPH+J1fsusCjjNGBbjDXEaLCNH/N1obmfK839XAj1NKBSitmQwq0TYatWzWxEnUEDQGVFdaM2xzjpEQoXLSLvw48Iev+9GzpGr9czduxYZs2aRWpqKo899hhOTk4A6Jyc6TpmIm16P8i2JQvZ+dXn/LJhLe0HDKVNn35odPr6fDqCIAiNTqNSEBtgu4xYS5ZlLhRX1vSAXSLz4iUOXyxh3cFs+2B8jUpBpLezfRC/LYi54O+mE2NhhRsiwlatmtmImpou5MrKxg1bKg8PjBMmkPfxx1QePIguNvaGjjMajYwZM4Z58+aRlpbGhAkTUKsvbzvk6uXDA0//hXb9BrFl0Ty2LJrHzq8+J7HfINr0eQitwVBfT0kQBKHJkSSJAHc9Ae56esX62u+vNFk4llPKkewSDl8s4XB2CTtO5PPFT+fsZVx0KnvwqhvCBOHXRNiqVbM3olpn27urusLcyA0C48QJFCxcSO77HxA8/ZMbPi4oKIghQ4awdOlSvvjiC4YNG3bVgoA+oeEMfflNzh85RMYXS9i6eD47v1pBwgMDaPvAALH1jyAI9zSdWkmrQDdaBbpdcX9xuYkjOSUculjC4YuXOHKxlK9+Pk9qxuXPDHetROvjGcT4uRDl40JEzaB+N/29uzn3vU6ErVo1eyNqtLZQYq6yNnKDQOnqiuekSeT+73+U//QThpuYshobG0vv3r1Zu3Yt69ev5/77779muYDoGAZPeZ3sE8fY8flifliexs6vPqdlt14kPDAAY0Cgo56OIAjCHc/NoKZ9qJH2oUb7fbIsk32pikMXL3Eku4RNe49RUFbNZz+cotp8+bPE11VLpI8zUT4u9lmVUT7OeDbw0hRCwxNhq5bK9suuVloAMFVZGrM1dsbx4yhIXUjOv98hZFHqTY0P6NSpEwUFBWzbtg0PDw8SExN/s6xveCQDX3iV3NNZ7P56Jfs3rOHntasJT2hPwoMDadYqXoxNEARBuAZJkvBz0+HnpiOluQ/R1jOkpNyHxSpztrCco9mlHM0p5VhOKcdySli26wxl1Zc/Y4xOGiK9nYn0tYWv2kDm66oV77t3CRG2aqls62uplSYArI07ZMtO4eSE9zPPcvH11ylZtw7X3r1v+FhJknjggQcoLi5m9erVuLm5ERUV9bvHeDcLpe8fn+O+MRP4ed23/LzuG5ZPexXPoGbE9exDi649xCVGQRCEG6BUSIR4OhHi6XTFeLDaQfl1A9jR7FJW77tAcYXJXs5FqyLC53IAi/B2JtzbiWCjAbWYHXlHEWGrVk3YUsjVWJUWLFWN3J463IcOoWD+fHL/83+4dO+OpL7x6/5KpZJhw4Yxd+5cli1bxqRJk/Dz87vucU7uHnQePoYOA4dxaPtm9q37lo2ffcrmRfOITupC6559CGrRSvzVJQiCcJPqDsrvFu1tv1+WZfJKqy8HsJowln4kl2W7z9rLqRQSzYwGwr2dCPd2JszLiXAv27+9nDXifbkJEmGrVk3YwlwBaguYms4vq6RS4TP5Bc4++RSFS5ZiHDf2po7XarWMGTOGTz/91L4khJub2/UPBFQaDa1SetEqpRe5p06yb/0aMrdsJHNrOh7+gbTqfj8tklNw8fS6fmWCIAjCb5IkCW8XLd4uWjpFeF7xWHG5ieN5pZzILeNkze2J3DI2H827YlyYi05lD17hXk6EeTsR7mULZHqNsqGfklBDhK1aNWO2MFchaWQwNa1fSudu3TAkJZH34Ye4DRyA0uXmLuW5uroyduxY5syZw6JFi5g0aRJa7c0NyvQOCaPnpCfpOnYiR3ZsY9/6NbblI9I+o1nL1rRI7k5UUhexfIQgCIKDuRnUJDTzIKGZxxX3W6wy54sqOJFXxonc2jBWRsavlqkACHTX23rBvJ1qgpgzoZ4GAt31YtHWeibCVi11zaKe5kokjYzCrMJsNaNSNI0fkSRJ+Lw4maxhw8n76GN8X5py03X4+fkxYsQIUlNTWbZsGaNHj0apvPlQqdbqaNmtJy279aTw4nkyt6STuXUja6a/x/rZnxCemETsfSmExiegVImpzoIgCPVFqZAINhoINhquuCQJUF5t5mSeLXzZesJKOZlXxhd7zlFSdXmpCpVCIshDT4inE6GeBtutl4FmRieCjXq0qqbV+XAnahpJoimo7dkyVaLUSGhKtZSZynDT3tjltoagb9kS92HDKFiwAPehQ9BeZ7D7tURGRtKvXz++/vprvv32W/r163db1/c9/ALoPHwMnYaN5uKxIxzcspHD2zdz5Ict6JxdiGzfieYduxDcKh6lSvy6CYIgNBSDRkXLADdaBlz5OSbLMrmlVZzMLeNUQTmn8svIyrfd7jlVeEUQkyQIcNMT6mW4Mox5OtHMaBCXJm+Q+PSrZR+zVYlKq0BdrGtyYQvA+y/Pc2ntWi7+4y2azZ1zS0EpMTGRwsJC+5IQXbp0ue12SZKEf1Rz/KOak/LwY2T9vMcWunZsYf/GtTXBqyPNOyaL4CUIgtCIJEnCx0WHj4uOpPArx4bJskxBWTVZ+eWcLigjK+9yGPv2lwsUlpuuKO/nqiPE01DzZZsp2cxoINhDj9FJDNavJT7xaqlqLyNWodYqUVtsPVtNjcrDA5/n/szFqW9SsmYNrn373lI9PXv2pKioiHXr1uHu7k7Lli0d1kalSkVEuw5EtOuAubqarJ/3cGTHVo7s2Mr+jevswSu6YzLNRPASBEFoMiRJwtNZi6ezlnYhHlc9Xlxu4lRBTU9Y3uUesQ2HcskrPXtFWYNGSbCHgWCjniCPmhBmtH0f7GHASXvvvPffO8/0euwD5CvQ6NzQNNGwBeA+YgSFS5eR/fa/cO7aFcUtDEhXKBQMGjSI4uJivvjiC1xdXQkODnZ4W1UaDZHtOxLZvqMteO37iSM/bLkcvJyciajp8WrWOl6M8RIEQWjC3Axq4gzuxAW5X/VYWZWZM4XlnCmo4ExBec2/bd9vP55PefWVi4UbnTS28OWhr7m9HMQC3PVoVHfPoH0RtmrZLyNWodWrUVmbbtiSlEr8XnuVU2PGkjdjJj7PP3dL9ajVakaPHs2sWbNIS0vjsccew2g0Xv/AW6TSaIhMTCIyMely8NqxlaMZ2ziQ/j1aJyciEztS5eSK2dQF1U2sJyYIgiA0Lietihg/V2L8XK96rPby5JnCq4PYL+eK+W7/RcxW2V5eIYG/m54gD1uvWKC7jkAPPYHuBgI99Pi76dCp75zxYiJs1VJfHrOl02lQW7SUVjfNsAVgSEjAbeBACubMwW3QQLRhYbdUj5OTE2PHjmX27Nmkpqby6KOPYmiApRuuCF4mE6f27eHIjm0c27mDqvIypqevJaJdB6I7JRPSui0qjabe2yQIgiDUj7qXJ9sEX90rZrHKXLxUyZmCck4XlHO2oNwezLYfzyP7UiV1shgAXs5aAj30BLnra4KYbaHYwJrvm9LG3yJs1art2TJVYjDoUFBCaUVJ47bpOnxe+H+UbNjAxb+9TrPP5iEpbq3L1cvLi1GjRjF//nyWLFnC+PHjUTXgOCqVWk1EuyQi2tmC1+pFC9CWXeLYrh84uGUjGr2eiHZJRHdMJjQ+QQQvQRCEu4xSIdlCkruejr8atA9gsli5WFzJ2cIKzhdVcK6ognOFttuDFy6xLjP7isVdwbbdUaCHLYDF+LnQQddQz+ZqImzVso/ZqsTJYBssX1ZW0YgNuj6Vtze+U17kwquvUbRsOR4jR9xyXSEhIQwaNIgVK1awatUqhgwZ0iizSFRqNW4h4aSkpGAxP83p/fs4smMrx3buIHNrOmqd3tbj1bELoW3aodbc3MKsgiAIwp1HrVTY1xO7ltqtjmpDWG0gO1sTyKrMFjpENnCj6xBhq1ad2Yi1Yau8vLIRG3Rj3IYOpfjr1eS88w7OKd1Q+/pe/6Df0Lp1awoLC9mwYQMeHh706NHDgS29eUqVmrA27Qhr045ejz3NmQO24HV05w4ObduEWqsjPKE90R27ENY2EbW2Ef9sEQRBEBpN3a2OrnWZEiA9Pb1hG1WHCFu1lCqQlGCuQOdmu0xVUdGEdqP+DZIk4f/mVE4MGMjFN6YS9PFHt9Ujdd9991FYWMjmzZvx8PCgbdu2DmztrVOqVITGJxAan1ATvH6xBa8ft3P4hy2otFrC29qCV3jb9qh1IngJgiAITYNDwpYkSX2B9wAlMEuW5bd/9fhfgMcAM5ALTJJl+ZQjzu1QKp1tnS2dbYZDZUV1IzfoxmiaNcP7T38i59//puTbb3F98MFbrkuSJB566CGKi4v56quvcHNzIzw83IGtvX0KpZKQuDaExLWh56NPcTZzP0d2bOPoj9s5smMraq2OiMQkWiSnEBLXVqzjJQiCIDSq217EQpIkJfAR8AAQC4yWJCn2V8V+AhJlWY4DlgP/vt3z1gu1DsyVaGoWWqusvDPCFoDx4fHoWrfm4rR/YC4svK26lEolI0aMwNPTkyVLlpCTk+OgVjqeQqmkWat4ej32R/4w/TOGv/YWLZJTyNq7my/+NZXpTz7Muk8/5OzB/chW6/UrFARBEAQHc8SKYR2AY7Isn5BluRpYDAysW0CW5Y2yLJfXfLsDCHLAeR1PpQNTpb1ny1Rpuc4BTYekUuE/bRqWS5e4OPVNZFm+/kG/Q6fTMXbsWNRqNQsXLqSoqMhBLa0/CoWSZq3iuP+JZ3hy5gIGvfg3QuPacnDLRpZMfYmZz0xi08I5ZJ88fts/H0EQBEG4UdLtfuhIkjQM6CvL8mM1348HkmRZfuY3yn8IXJRledo1HnsCeALA19e33eLFi2+rbTeitLQUZ2dnADpkPEmJSxT7Qv/CkS9lMqM3MCyhV723wZEM363BZeVKiidOoLJjx9uur7S0lL1796JWq2nTpg1abf3P/qv7mjiCxVRNcdZxCo5mUnwmC6xWtO5GjFExGCNboHO/eksK4WqOfl0ExxCvS9MjXpOmqb5fl+7du++WZTnxWo816GAWSZLGAYlAt2s9LsvyTGAmQGJiopySklLvbUpPT8d+noNGDEZXuna/jyNfbkaj1NEQbXAk+b77OH32LIplywkbNw5N0O13IsbHxzN//nyOHz/OxIkT633R0yteEwerKLnE0YztZG5L5+yuH7iwczu+4VG0SO5G80734Wy8en0XwaY+Xxfh1onXpekRr0nT1JiviyMuI54D6m6qF1Rz3xUkSeoFvAIMkGW5aU7zU2nBXIlaY/uxWO+cIVt2klJJwL/eBoWC85NfRDabb7vO4OBgRo8eTX5+PqmpqVRVNc2X70boXVyJ69WXka+/zRMfzaXbuEnIspX0+bOY8ceJLPv7y+xbv4bK0tLGbqogCIJwl3BE2NoJREmSFCZJkgYYBXxZt4AkSW2BGdiCVtMdbV0zG1GhVCArLXdk2AJQBwbi97e/UfHTT+R/+qlD6gwPD2f48OGcP3+etLQ0TCaTQ+ptTC6eXiT2H8L4t9/jkf9Op9PQUVzKy2XdzA/45IlxrHxnGod/2IKpqumvtyYIgiA0Xbd9GVGWZbMkSc8Aa7At/TBHluUDkiS9CeySZflL4B3AGVhWswbUaVmWB9zuuR1OrYOqmi161FYw3bk7jrv1f4jS9HRyP/wIp86d0cfH33adMTExDB48mM8//5ylS5cyatQolMo7ZyPQ32MMCKLz8LF0GjaG7BPHOLQtnUPbt3B81w7UOj1R7TsSk5xCs1bxYikJQRAE4aY45FNDluVvgG9+dd/f6vz7zhhlrtJBaS4AkkZGYVZhsVpQKu7MQOH3uq136+zzzxO2YgUqj9sfCB4XF0d1dTVff/01n3/+OUOGDLlrAhfY1hnzi4jCLyKKruMmcfbgfjK3buJoxjYObtmI3tWN5p2SiemSQkB0TKNsaSQIgiDcWcSf6HXVjNkCUGhAbdFSZi7DVePayA27NUpXVwLfe49TY8Zw/sUpBM+YfsubVdeVmJhIVVUV69atQ5IkBg8efFcFrlq2pSTiadYqnp6PPsXJvbs4tHUT+zesY++a1bh6+xDTuSsxySl4Nwtt7OYKgiAITZQIW3Wp9GC2Df5WaiXUJTrKTeV3bNgC0Lduhe8rr3DxjTfI++QTvJ9+2iH1dunSBVmW+f777wHu2sBVS6VWE9W+E1HtO1FVXs7xXTvI3LaJnV99zo+rluMVHEJMl27EdOmGm8+t708pCIIg3H1E2KpLpQVzBQBqrRJ1kZbS6lJwauR23Sb3kSOo+Okn8j78CH1cPM73JTuk3uRkWz33SuCqpTUYiO3ag9iuPSgvLuLwjq0c2rqJrYvns3XxfAKiWxCT3I3mHZMxuF17Q1RBEATh3iHCVl01sxEB1Dql/TLinU6SJPzeeJ3KzEzOv/ACYZ+vQB0Y6JC679XAVcvg5k7bPg/Rts9DFOdc5NC2zRzatokNc6azcd5MQlq3IaZLNyISk9A5iUUOBUEQ7kUibNVVszcigFanRmPRUma688MWgEKvJ+j99zg5bDhnnnmW0IULUDg5psvuXg9ctdx8/EgaPIKkwSPIPZ3FoW2bOLRtE999/F8UShUhreOJ6tiFyMSO6F3u3EvTgiAIws0RYasulQ4s1WC1oNWrUVt1d03YAtCEhhL4f//hzJNPcW7KFILef98hA+bhysAly/JdN0vxZnk3C8W7WSjJox7mwtHDHP1xO0d2bOPk9PdZp/iQ4JZxRCd1IbJ9R5zEdkGCIAh3NRG26lLV7PtnrkKv16K2aGxjtu4izl274vvSS2S/9Ra5//0vPv/v/zms7uTkZCRJYt26dZhMJoYPH45arXZY/XciSZIIiI4hIDqGrmMfIefkcY5kbOPIjq18P+sj1s/+hMAWsUQndSGqQ2exXZAgCMJdSIStulR62625EoNeh4SCsoryxm1TPfAYP46qE8fJ/3QWmtBQ3IcOdVjdXbp0QaPRsHr1alJTUxk9enSDbF59J5AkCd/wSHzDI0ke9TB5p7M4krGNoxnb2TB3BhvmziAgugURiUlEJCZhDAgS63gJgiDcBUTYqsves1WJs8EWvMrKKxqxQfVDkiT8XnkF0+kzXPjb6yiNRly6d3dY/e3bt0ej0bBy5Urmz5/P2LFj633z6juNJEl4h4ThHRJGlxHjyD972h68tiyax5ZF8/DwDyA8oQMRiUkENo9FcQ9flhUEQbiTibBVl0pnuzVXoqsJB+Xld+e+eJJaTeD773N64kTOPfc8zebOwZCQ4LD64+Pj0Wq1LFu2jHnz5jF+/HhcXFwcVv/dxjOoGZ2CmtFp6Ggu5eVwYvdOju/OYO+ar9m9eiU6J2fC2iYSkZhEaHwCWsMdvh6JIAjCPUSErbrUtWGrCrXWFgwqK+7Q3ahvgNLZieCZMzg1egxnnnyKkIUL0EVHO6z+mJgYxo4dS1paGnPnzmXcuHEYjUaH1X+3cvXyoU2ffrTp04/qinKy9v3Eid0/cnzPTjK3pqNQqghoHkNoXAKh8Qn4hIY7bKKDIAiC4HgibNVV27NlqkCjs12yqawwNWKD6p/KaCR49mxOjR7N6UmPEjL/M7Th4Q6rPzw8nIcffphFixYxe/ZsxowZQ6CD1vi6F2j0BqKTuhCd1AWr1cKFI4c5vudHsn7eY19E1eDmTkjrNoTGJxAS11bMbhQEQWhiRNiqq85sRLXW9qOpqrx7e7ZqaYICaTZvLqcmTOTUhAmEfDYfbXiYw+oPDg7m0UcfZeHChcybN48RI0YQFRXlsPrvFQqFksCYWAJjYuk6ZiJlRYVk/bzH/pW5NR0A79BwW/Bq1YaA5jGotbrGbbggCMI9ToStuurMRlQbbD1bFRVVjdighqONiCCkJnCdnjCBZp995tDA5eXlxaOPPkpqaiqLFi2if//+JDhwjNi9yMndg5bdetKyW09kq5WcrBP24LX76y/YuWo5CqUK/6hogmNbExTbmoBoEb4EQRAamghbddWZjWi/jFh+9/ds1dJGRtoD16mHH6bZrE/RxcQ4rH4XFxceeeQRli5dypdffsmlS5fo1q2bWN7AASSFwr6sRNLgEVSVl3P+8EHOHPyFMwd/IeOLZez4fMkV4Su4ZRz+Uc1F+BIEQahnImzVVWc2olpXexnx7h6z9WvaqChCFszn9KOPcWr8wwRP/wRDu3aOq1+rZcyYMXz55Zekp6dTUFBA//797/nFTx1NazAQ1jaRsLaJAFSVl3Pu8AHOHPiFs1eELyXeIWH4R8XYF1919fYVAVgQBMGBRNiq64rZiDVrGpmUlJvKMajvnXWitBERhC5K5fSkRzn96GME/u+/uKSkOKx+pVLJoEGD8PT0ZMOGDRQUFDBy5EixNEQ90hoMhLdtT3jb9kBN+Dp0gPNHMjl/OJP96evYu+ZrwLa5dkB0jD2A+UZEodaIhWkFQRBulQhbddWZjahQSEgqGbVFS0FlwT0VtgDUAQGELErlzONPcPaPT+P7179iHD/OYfVLkkTXrl3x8vLiiy++4NNPP2X06NH4+/s77BzCb9MaDIQntCc8wRa+rBYLuaezuHDkEOePHuLCkUMc27kDsF2i9AwMxqp3wrXiEj5hEfiEhKHR31v/JwRBEG6VCFt1qS73bAEotRJqq5b8ynyCXIIasWGNQ2U0EjL/M85NfpHsf/yD6pMn8X35r0gqx/3axMbG4uHhQVpaGnPmzGHw4MEOq1u4cQqlEt+wCHzDImjTpx8A5cVFnD96mOzjR8g+eZwzhw6ycd5B2wGShId/IL5hEbbwFRqOV3AIBjd3cQlSEAThV0TYqqvOmC0AlVaJxqKloKKgERvVuBROTgR98D45//k/CubMoTori4D/vIvKw3FrOfn7+/P444+zZMkSli5dSrNmzejatSsKsVBnozK4uROZmERkYhIA6enpJMbHkXPyONknjpF98jjnDk4fhQsAACAASURBVB3k0LZN9mP0Lq54BYfgGdys5jYEr6AQdM7OjfU0BEEQGp0IW3XVmY0IoNOrUZfryK/Mb8RGNT5JqcT3xclow8O4OPVNTg4ZStB7/0MfF+ewc7i4uDBhwgS+/fZb9uzZw4IFCxg6dCjO4kO6SXH2MOLsYbRffgQov1RMbtZJ8s+eIu+M7evg5g1UV1RccZxncAge/oE1XwF4+Afi6u2NQiH2fBQE4e4mwlZdCiUo1PawpXfSor2kJ7/i3g5btdyHDUMb04Jzf/4zWWPH4fvii3iMG+uwy0ZqtZoBAwZQVlbG8ePHmTFjBsOHD6dZs2YOqV+oHwZXN0Li2hAS18Z+nyzLlOTnknfmFPlnTttuz56pCWHl9nJKlQo3X397+PLwC8DNxw9XHx9cvbxRqsQsVUEQ7nwibP2aWm8fs2Vw0qC3OFNQmdW4bWpC9K1aErZiOeenvET2P/5B6ebN+E+bhtrXx2Hn8Pf3JyUlhaVLlzJv3jx69+5NUlKSGAt0B5EkCVcvH1y9fOwzIMEWwsqLiyi8eJ7CC+covHCewvPnKLp4nqyf92AxmepWgrPREzdvH1y9fS/f+vji6u2Ls9ETlVgyRBCEO4AIW7+m0oLJdvlDa1ChMztx4R6/jPhrSnd3gqZ/QuGiReS88y4nBwzAb+obuPbt67Bz+Pv788QTT7By5Uq+++47Tp48yYABA3BycnLYOYSGJ0kSTu4eOLl7EBTT8orHZKuVkvw8inOzuZSbQ3HORdttbjbnDh3g0NZNyLL1imP0Lq44Gz0vf3nYbl3q3KdzdhFBXRCERiXC1q+pdPaeLa2TGo1Zz8mik43cqKZHkiSMY8fi1Kkz56dM4dxzz1Py0Hp8//oSKk9Ph5xDr9czcuRIMjIy+P777/nkk08YOHCg2FfxLiUpFLh6++Dqfe1eUovZTGlBHsU5ORTnXqS0IL/OVwHZJ45RXlx01XFKtRondw8Mrm4Y3NzR19warnGrd3VD6cDZtoIgCCDC1tVUussD5J3UKKxKThacosxUhpNa9Kr8mjY8jNBFqeTNmEnejBmUbt6Mz/PP4T5iBJLy9gc+KxQKOnXqRFhYGCtWrCA1NZWkpCR69eolVp2/xyhVKtx8/HDz8QOuPTnDYjZRVlhISU0IKyvMp6Qgn/LiIsovFVNaWEDOqZNUFBdhMZv/P3v3HR9HcT5+/PNcv1O15SbLxnID417BDhhMszFgm9ACgcQQCBBIIOUL34TkG1J/gTRKgNBCSwgETG+hi44rxgXjbty7rWK1K/P7Y/dOp9OdLFt3Psl63q/XvnZ3dnZmdGOfHs3O7iYtw5eTiz+/AF9OLr7cXLw5ufhy86zFTmvYzrP3c3V+mVIqJQ22EsUFW96A9fF4gj4W71zM+OLx2WxZmyVuN12/fy35Z0xl629+y9Zf/4a9zz5Hj1/+Ev+woWmpo0ePHlx55ZW89dZbzJ49mzVr1vD1r3+dnj17pqV8dXhwutzNjo5FGWOor6m2grDycqorGq9rKsqp3VdFTWUFe7ZspraqktrqfWBMyjJdXi9efwCPvXgDfty+AF6/H0/ATvdZ2/H5PH4/Hn8At8+L2+PD7fPicLr00qdShxENthK5vHHBlvWXqjcUYOH2hRps7Ye3Xz+OePghKl5+hW233sq6888nf9o0ul5/HZ5erX8orNvtZurUqQwcOJDnn3+eBx54gOOOO44TTzxRR7nUARERvIEcvIEcOhWXtOgcE4lQV11tBV77qqx1VSW1+/bFtutra6ivrrbWNdXUbN9KXY21XV9TTSQcbln7HA7cXh9urxe314fL67W3vbi8PtweL26fD5fHy5bt2/lk5xZcXi8ujweX24PL7cbpduN0W/tOt8taezw4XW5cHredbuVxujS4UyqTNNhKFHc3ojfH+nj6+QYyf9v8bLaq3RARCqadRe6kE9n1wIPsfuwxKv77Xzp94xt0+d7VaZnPNWDAAK699lreeOMNPvzwQ5YtW8b06dPp06dPGn4CpZIThyN2yfBgGGMIB4PU11RTV1NtBWU1VmBWV11NqK6OYF0dwbpaQvV1BGtrCdrrUH30WB01VTvtvLWE6uqoq6lh22dzWv3zNQRfblzRoMzedzhdOFxOnC43DqcTh9OF0+XC4XTG1g6XG6fTicPVcMzhjOZxW+c7XdZxO5+1HZ/XgcPpRBwOHA577WxYOxwOxOFs2LbXiXk1cFRtjQZbiVxeqLaeGO+zR7ZGFIzm0a13sbd2L4W+wmy2rt1w5uXR7cc/otPFF7Pz7rvZ88QT7H32WTpdcAGdL7sUd/furSrf7/czY8YMhg4dyksvvcTDDz/M6NGjOeWUU/SORdUmiYg18uTxEChI3/dIWVkZJ0ycSLCujnCwnlCwnnAwSCgYJFxfTygUJFwftNJDdlowaOeJyxt3bjgYJBTNFwoSCYWIhMOx0blIKEQ4traOhUOhWL6InZYtIo6EQM0O0hyOxkGa04lI44AuGqyJw2mvxSpPxDoWPd5Mnu07dlCzdEGsHdZxO4/DLidpmY6GslLWm6yMhrxE2xbdBnA4rHWTYwIOey2ClewAAaEhXzSvOASw8pFQTsMxsc6XhqBX7LeBNGlf0jId1vb+yowrq0mZcT9H4/ZlT1qCLRE5HbgDcAIPGmNuSTjuBR4DxgC7gG8YY9alo+60i78b0Z6zNSx3JKGqEG+vf5tzjzw3m61rd9zdu1H8m1/T+dJL2Xnv39n9z3+y+/HHKTx7BkWXX46ntLRV5ffv35/vfe97lJWV8emnn/LFF19wyimnMGbMGH3dT7ZFIhAJgQlb60gIIuGEdcK2CSccix5PSDORhDx2PhOx51WZZrZNwnakYftgz0tZRuYdtWULjvKn8R5sAQ7Aay8tJoATxJWQBvZvSoyx/wkYCEcMkYghbKxuCgORiCESsY+Zhn1jIGKsNGvbOhbbNsbqftOQZoyJy2Pi6m5Is8q28xEhEgljTLDxeaFoXQZDQ5dHyzGYWJqJpsWvsdbB+hAbt6yxj6XOl6oclX6du3Wm77nfyVr9rQ62RMQJ3A2cBmwE5orIi8aYL+KyXQ7sMcYMEJELgVuBb7S27oxIuBsRoJN04Yi8I3h17asabB0kb7++lPzxj3S97jp2P/QQe2c9w95Zz5BzwkQ6XXQRuRMnHvTdi16vlylTpjBq1CheffVVXnnlFebPn8/UqVMP/aXFSJIgoElgkCKgMMmCjBQBikkVtCSWcQD176fccZXlsNjb8qCIdvhLQxzYf0I33sbeb7RNkvQk52VY57o6qPZlvJ6YWDBgUqcZY4djBifgTpFn/+XQgjz7KUfsdezrJcV5jYKcZso+BKzADgxib4sdoMWvpSEgRGJ5sc+LttqYhm3sc4grP1pf7BgN9RFXd0PbEvLFf+wpzjNIrM5oPVa++Dz2scR8dttiP09cWSbu/1fiz9q4TvDnV5DNtxynY2TrGGCVMWYNgIg8CcwA4oOtGcCv7O1ZwF0iIqYthvBxwZbb50QcQn11iOn9p3PXwrvYULGB3vm9s9zItiH2V2LYEAlb25HodjhiHWuUZohE8ohc+AN8U75FxdtlbPzgI9b94l4cnZ/GP+E4fKPHsGdjmCXh1UTqazHBesL1dZhgPZFQPZFgiEgoTCQcwYQjVj1x9RWH/LgdJXy1dRsPP/wwBSaXnqGueCI++y9kIRIRjLHWERP/ZUOj7cbr+C+y+H2JfWlYxxrW1rY0+kLZn6a/mh32EncDgOyvtOjxVL/oBZHm8khcHXFfZsbEhukhNoDR5FTrAkRiWtxGwr7E7zd3LK78JtvRyxAJ7YjtpmxDknMTSfympDzWuJ5k5TQ+mPwzimZNXVDioerqagKBQPI6mmlP02OSdHP/5TQfUB7QZ5KY9yBj1dZN2WrpyU3/H4pYyeUV5RTkF7TonAOto2kW01Bxi4ttyJvyszIpd1pywn7ztLyLEgLvpCeapJvx8jq5cVLb4lrTLR3BVgmwIW5/I3BsqjzGmJCIlANFwM74TCJyJXAlQPfu3SkrK0tD85pXVVXVqJ6B23fRtaaSj+00h9uwZtVXdM/tjgMHt719G9M7Tc94u1KJhA2hWgjVWEs4aA8kBCEcNLGBDCINQ+DJrow0uiqSJG/SKyhJrpq0Th+IH3n6CvhqEyBsnvtVkvwOwIODIA4iOCSEAA4J2fthHIRxiIc+FLPXX8EufwXLXFV0qvdTXOfHgyASwSERxGkSfl/HXQZp9Is7ftTCPkES8selmVheSShTMDSMehji0qQhODP2yIixy210PJbP0XA8FgA5GurOgPr6IB6PFfQd0J9JLczbfJkJow4tjTdbUn1LfxbT7O7BlZmi3JSnJzngzAkRdlXvN+8Btbe5/Ux8tgdabrrqzJAIESr2VbT8hCy1ub1/zgeiygQprKo+JHFFMm1qgrwx5n7gfoCxY8eaSZMmZbzOsrIyGtVT9ybs/CCWtundTykqzGXKqSfzzjvvMH/HfG6deCtuZ9NHDZhI3OiOPeITDkcaj+7Y2w1pEXuugrUfDkaorQpSU1VPTWXDurq8jn1766ndF2xSb5TT7cDjc1p39DgEcYq1dthrJzgcgsMZl5awHz1PxEqPlSMN5Tni106xJ58KDongqN6Go3ITjpptOKq24qjahFRvxRGpswMhOzBygMOXg8OXi/hzcfhzCde42bdiL+WffQXl1TicQs6QPuQfM5i8Y0fi6d4d8fgRtw+cbnB6rMXhath2uhsFG/v27eP9999n7ty5VATqGTVqFMcffzyFhXqjw4Fq8n9FtQnaL22P9knblM1+SUewtQmIv67Wy05LlmejiLiAAqyJ8m1P3LsRAfKLfGxasZfafUHO7n0uXy3cw9OPvUd+dVeqdtdRVx2kriZEsC6c9kjf5XXiz3Xjz3WT38VPj/6F5BR4yCn0klPgJZDvwRtw4fG7cPucOJ2HeEJ49W5YUwbrP4UNC2Dr4tglWBxuKOoPAwdC5zFQ0AvySyC/p7UdKEo5AlP27rscU1hIxSuvUvn22+y850l23vsUgTFjyDnuOHImjMc3pD/Sgteq5OTkMHXqVMaPH8+HH37IggULWLBgASNGjGDixIl07tw5jR+IUi1jjIFwGBOJWOtwBCJhTDicsB+x58FZf7SBsbcj9hBYBNf6DdQsXWrNFE9ynEjEqi/SMFxtorPPTTSvvZ80b6Rx2cmOQ3TSDA071r5pcl0+fru5/DTkaWm+2Lpx+Sa+jGbyN2prS8tNInf9BrbNmZv0WLNDuAd1LPU5zc7Uae73VarzDrbtKSprvn3pbbura1dI00O2D0Y6gq25wEAR6YsVVF0IfDMhz4vATOAT4DzgnTY5XwusOVsmDOEQOF2M/3p/Zv1hHv/65SfUV4eZbC5jlyOMo7iWvCI/XXvn4gm4cHvtEaXEUZ/Y6JEjSVr8KJF13Oly4M1x4c/z4Pa0/nU3aRWJwMY5sOJ1WP0ObPkcMODOgeIRMO4K6DkKikdCp1JwHuQ/LxECo0YRGDWK7j+/idovvqDq7bepfOdddtx+OztuB0duLoFjjiFn/Hj8o0biPeooHB5PyiI7derEtGnTOOGEE/joo4+YP38+CxcuZMiQIYwfP55eaXjoqsosEwph6uqI1Ndj6uow9jpSV4+pt/fr6618wRAmGMSEghAKWWn1QWsdCmGCVj4a5Q1hQkFMMNiQHmp8jGCoIUCKRCAcsgKiRoFTGBMJQ1x6LIiy8xANUNKgCFiXttI6sEbTAmg6DYDoLIIk+RIEwmH2OJ2pL+k3NyevJW08gPLSfeyg2tfcsYy0vWmSp0+f9h1s2XOwvg+8jnW/x0PGmKUi8htgnjHmReAfwD9FZBWwGysga5tc9l09oVpw5tK1dx6nXT6Er5buIq+Tl3nu93hg6+28eu4r9MrrAL+gjYHNC2DR0/DFC1C52bps12scTPoZ9D8Jeo4++MBqP0QE/5Ah+IcMoet11xHatYvq2bPZ98mn7Js9m6p33rHyud14Bw3CP2wYvmHD8B11JJ6+fXH4/Y3KKygo4IwzzmDixIl8/PHHLFiwgCVLllBSUsKxxx7L4MGDcemLiFvEBINEamuJVNdgamuI1FiLqalpnF5t79dUY2pqidTWYOrsQCkaJMXv19XHAqn44CqdAQoAIojLhbjd4HZb2/a+tXbZ6XHHcrxgP4sJl8taR/edTuuOWvuZTg3rFOkuZ0I+B+J0JTnfYd+pK9YzkRz23Y/iAIewdOlShg4fnvS4OKRRXhGB2PGmZTUct5+1lKws+7gIjcqyP9SGexCaCVrif1FKsmNN8sXdxNCCclMGRYn5MjS/US8jtk3rsjRfC9I0Z8sY8yrwakLaL+O2a4Hz01FXxsUHW17rSdEDxnRjwBjrXWtH7Atw3zO38+zKZ7lu9HXZamXm1ZbDoqdg/qOwbbE1H2rAaTDkN3DkFPDlZ6VZrqIi8s84g/wzzgAguHkzNYsWU7N4EbWLl1D+/PPs+fe/rcwiuEtK8PTvh7f/ADxH9Mbdsyfunj3J6dmTKVOmMGnSJBYuXMjs2bN59tlneeONNxgzZgyjRo1q9/O6TChkBUDVNZia6lgwFKmusQOf6HaKoCghvWjXLlaKWGXU1kIw9fzBpEQQvx+Hz4f4vDjcHsTrtRcPjoAfZ2Fhw77Xi8Ty2PseD+KJ249P87it0U2XG3HHBU3RgKpRIOVOy4vS24I6t5s8/cWuVJumf8InctlP9gslv0W0R04PJpZM5JmVz3D1iKvxOFNfumqXKjbDp/fAvEegvhJ6DIcz/wrDzgNfsluZsysaPOWfPgUAEw5Tv3YtdatWUbd6NfWrV1O3eg3Vn3yKqa9vdK6zoABXt2706NyZr3fqxOZOhSytq+O9997jvffe44jCQob16sWRfUrx5uXiCARw+P2I1wsOJ+JqGNEgxStCYnNzEi4jmXA47jJYdDSnrsl+bLuuzg6WGgdJkZpqItXVmOqGkaVITQ2mutq6HHZAH6Ybh89n/Yx+Hw5/wNoP+HF27kwwL5eiPn0apYvPj8Pvx+H3WYGUvYjPjyNgB1bRNK83YyMJSmWL9ZBSQ8REMPadskETpC5cFzvWZE3DnLP4tIiJNCoXaJw3brJSYlrSfKbx8SblJjxWIVkZ8WmxulPkiy8nVV3xZSXN18KfNf6za/IzJCk34AqQTRpsJXLbl53sp8gn882jv8lVb17Fa2tfY8aAGYeoYRlWvgneuxUW/tuaszbkHJhwLZSMznbLDog4nXgHDMA7YECjdBMOE9qxg+DmzQQ3b7HXmwjt3El49x7qly+n0+7dTCgvZ3hOgLWlfVnbry+v7N3LGwsWcMT69fRev4GuO3bgSDXd0H4NCCINc3PSPDVRvF4rcAn4raAnGuB07YI7fj/gt4OcgBX0+JvuOwIBJLrv81mjP81YXVbGaB1BOWAREyEYCRKKhAhFQkRMhLAJE46ErbW9HTERQsY+nnAsWb7o9uKqxVSsrmjIF5c/WlfEWBPbIyZChLhtO0AIm3CTtNi2MU3OiZYTX26yc6LbsfITzknWlsSAJGmAkiyASTzP/uXfKBCCpPkS143yNVd+c/6V0X9W6gD1K+jHjwp/lLX6NdhKFB3ZirsjMdGE4gn0L+jP48seZ3r/6e37r/Xq3fDhbTDnfuvuojGXwte+b01wP4yI04m7Rw/cPXpAM/GjCYUI793LsMpKItXVrNuwgUVr17LS52P1gAEEnE4GBAIM8PspdroQE538HLEnRdsBVtycG3E5rZEwp6Nh7XTFXRpLvCyWsG9fTnP4fYfNpa9MMMZQH6mnLlxHfbie2lAt9WFrP35JTIvmjW7XR+pjwVFsHQ4SMta6ybFm1sFIsNFoRcZ8eOCnCILDftecAwcOaVhErGMOUmzH50Uan0fybac4G50f3Y7mi66taV/SqH3R9opI7Fg0X7QNSfMlrpPki54PNMkXrSfavib5UpS/du1a+vfr3/hzbqaNQEOeuLRYX8XXndCGlPmkcf5kaYltsjearavF+RJ/hiR1RT/f/dUVX1ayn68lP6vX6WXLoi1kiwZbiWJztlKPbIkI3zz6m/z209+yYPsCxnQfc4gal0aRCHz2T3jrZqjZCyMuhJNugsIjst2yrBKXC1eXLri6dAHg6CFDOBqor69nxYoVLF26lC9WrmRRZSW5ubkMHDiQI488kn79+uH1HvTb6TqkiIlQFayisr6SqvoqKuorqKqvoipYRXWwmupQNTWhGqqD9no/+zWhmv2PNjRDEHwuH26H21qcblzistYOF25Hw9rr8pLryG2UlrhutG2X5XQ4cYgjtu0UZ6M0hzga0iXhuMPVJJ9LXMybO48J4yc0yu8U65z48+ODpWS/0FX6lO0uY9KwSdluxmEpEokQDocbLSH7xejNLQ53dt+Vq8FWovgJ8s2Y1n8adyy4g3998a/2F2xtXQwv/wg2zoU+x8HUP0KP7N0S2x54PB6GDh3K0KFDqaurY/ny5SxfvpwvvviCzz77DKfTSWlpKUceeSQDBw6kU6dOHeaXWSgSoryunL11e9ldu5u9dXvZU7uHPbV72Fu3l711e6msr7SWYGUsuKoKVrWofJ/TR8AdwO/y43f5CbgC+N1+Cn2FBFyBRsf8Lj8ehwefy4fH6cHr9OJxevA5G/bjF4+zIa9LXO2yz9a719M7T18hptLHehVbhFAoFAtkWrqdmNbccjB5DvapUV26dGHo0Hb86IfDTizYSn0ZEcDv8nPekefxyNJH2FS1iZLckkPQuFYKh+DDv1pzs3yF8PX7YPg3mn9eiWrC6/UyfPhwhg8fTjgcZv369axYsYIVK1bw2muv8dprr5Gfn0/fvn0pLS2lb9++7e7ORmMMFfUV7Kjewfaa7eyo3sHs8tl8PPtjdlTvYGfNTiuoqttDRV1FyhGlHHcOhd5C8jx55Hny6JXbK7ad58kj151LviefXE+ulebOI9eTS447B7/Lj8/pw+nQS6eqYzDGxAKMYDAYC1oSlwMNeA5mO10cDgdOpzO2uFyuRvvx6V6vt9k8LSknVR63282iRYvS9nMdKA22EsXuRkx9GTHqokEX8ejSR3nyyyf5ydifZLhhrbRzJTx3FWyaD8POt0azAvr09NZyOp307duXvn37MmXKFHbt2sXq1atZt24dK1eu5PPPPwegsLCQ0tJSevXqRa9evejWrRsOR/aGtevCdWyp2sLmqs1s2reJTZWb2Fy1mW3V29hevZ0dNTuoCzf9P5BXnUc3fze6+LswqPMgCr2FdPJ1otBbSGdfZwp9hXTydoqlH3Z366oOITqyEx/wJAY/qY4Fg0HWrFlDbW1ts/lSHUuHaKARDTaSbXu93ibpLTlvf9uJQU42v+faEg22EsXuRtz/28F75PTg1D6n8syKZ/jeiO8RcGf31tKUFj0FL11vPSvrvIdh6DnZbtFhq6ioiKKiIo455hgikQg7duxg7dq1rFu3juXLl7Nw4UIA3G43PXv2pFevXpSUlNCjRw8KCwvT+sVUUV/BV+Vfsa5iHesq1rGxciObqqygakfNjkZ5XQ4XxTnF9MjpwYhuI+jm70bXQFe6Brpa2/6uLJu3jMknT05b+5Q6UNFLW8FgMOVSX1/f7PH4pblgqDUvORERtm7ditvtjgUjLpcrth8IBFIeS7UfvzQX8DidznZ5Ofxwp8FWotjdiPsPtgAuOfoSXl/3Oi+ufpELB7WxB+OH6uD1n8PcB+CICXDeQ9a7CdUh4XA46N69O927d2f8+PEYY9i9ezebNm1i48aNbNy4kU8++YSI/WR0j8cTyx9dunXrhs/nS1mHMYYt+7awcs9K1pavZV3Futh6d+3uWD6nOCnOKaYkt4TjS46nZ25PSnJLKMktoWduT7r6u+73ct1qx+r0fDDqsBQdCUpXIJQs/8GM/IgIHo8Ht9vdaHG5XPj9fvLy8poEN80FPi0Jij744AN9grxqRIOtRC2cIB81ousIhhYN5fFlj3PBURfgkDYyZFq5DZ78JmyaB1/7AZxyMzibf46SyiwRiY18DR8+HIBgMMi2bdsaLUuWLGHevHmx83JzcykqKqKgUwFhf5hyVzlbIltYXb+aFeUrqAxWxvJ29nWmNL+USb0nUZpfSp/8PpQWlNI7tzdu7f8OKxwOH1RwcyD5w+HwAbfL4XAkDYTcbjc+ny9peqr8qRZ9/ZZqC/RfYaIWPPohnogwc8hMbnj/Bt5d/y6n9Dklg41roW1fwL8vgOpdcMFjMPgwefDqYcjtdsfmcUUZY9i2exvzV81n1cZVbN+xnS07tuDe4MYXaRjl6kMfSr2l+HP9dCrsRHGXYrp26kpBQQEFBQXk5eURCARw6rO52ixjTMrLYsnSkwVDmzZtYvPmzc0GQ5GDeK+k0+lMGdhEL4O1NhDSf5uqo9BgK1EL70aMd2qfU+mV24uHljzEyUecnN3r5avegqcutd7reNlr0HNk9tqiWiRiIqzau4qF2xeyeOdiluxcwpryNbGHYfbs0pMhg4bQr3M/+gb60tV0xVHjYNeuXZSXl1vLnnIWrluYdHQhEAiQk5NDTk4Oubm5se1AIIDP50u6dOTRgFS3vMfP6zmQ+UD7Ww5G9PKVx+MhGAzG9r1eL7m5uSmDmwMJhnRis1Lp03G/UVM5gLsRY6c4XFw65FJ+N/t3zNs2j3E9xmWocfux6Cl47mroPhgu+g8UtIPHUXRAwXCQpbuWsmD7AhZsW8Bn2z+jor4CgE7eTgzpMoTT+pzG0C5DGVI0hCJ/UYvKNcZQXV0dC8Cqqqqoqqpi3759sfXmzZupqqqiPuE9kYlcLhc+n6/RL+eqqio2bdrU6PJM/LbD4WjREv/HSKP3pDWzHYlEYks4HD6g7VRBU6qA6mBGgaKit5i35LJY/Od3oJfF4gOhsrIynR+kVBunwVYiEWt0q4VztqJma/l+PAAAIABJREFUDJjBPZ/fw0NLHspOsDXvIXj5x1B6PFz0BHjzDn0bVFLhSJilu5byyeZPmL11Not3LKY2bP37Ks0v5dQ+pzK622hGdxtNr7xeBz0yKiKxUauePZu/ESIYDFJTU0NtbW2Tpa6urtF2dJSmsrKSqqqqpKM2rblzK12it5lHn+sT3U5211Z09C7Zsebu/NpfgKSjQUqpZDTYSsblbfHdiFE+l4+Lj76Yv332N5bvXs5RnY/KUOOS+OgOePOXcOTpcP6j4E5995o6NDZVbeLjzR9bAdaW2bGRq0GdB3HekecxuvtoRnUbRRd/l6y0Lxoc5Ofnt/ic5kZQoiNJiaNQqRZI8u40afpes+h2siAqcV8ppdoqDbaSOYiRLYBvHPUN/rH4Hzy89GFumXhLBhqWxEd3WoHWkK/DOQ/oHYdZEowEWbBtAWUbyvhg0wd8VfEVAN0D3Tn5iJP5Ws+vcWzxsXT2HZ4Pko0+30cppVRTGmwl4/Id0JytqAJvAecdeR6PL3ucH4z6QeZf4TP3QXjz/2Dw2XDOg+DU7jyUKuor+GjTR7y74V0+3PghlcFKvE4vx/Q4hosGXcSE4gn0LeirDxhUSqkOTn87J+PyHdDdiPG+Nfhb/PvLf/PY0sf42bE/S3PD4ix8Al75iXXp8JwHNNA6RMrrynln/Tu8vu51Zm+ZTciE6OzrzCl9TuGk3icxvnh8232TgFJKqazQ39DJuLwHNbIF1it8zup3Fs+ufJbvDv9uZubkLP8vvHAN9D3RmqPl0vfPZVJVfRXvbniX19e9zkebPyIUCVGSW8K3Bn+Lk484mWFdhunLkpVSSqWkwVYybv9BzdmK+u6w7/LS6pd4aMlD3DjuxjQ2DNj8Gcy6DIpHwIX/1snwGRKKhPh488e8sOoFyjaUUR+pp0dODy4edDGn9z2dIUVD9PKgUkqpFtFgK5mDuBsx3hH5RzCt/zSeWv4Ulw25jK6Brulp19718O9vQKCL9Rwtb256ylUxq/eu5oVVL/DSmpfYWbOTTt5OnHfkeUztO5XhXYe3ndcxKaWUajc02ErG5YPailYVceXwK3l59cv8Y8k/+OkxP219m2rL4fHzrSDw2y9CXvfWl6kAqA5W8/Kal3l+1fMs3rkYl7iY2GsiMwbM4ISSE/SdgkoppVpFg61kDvJuxHi983ozY8AMnl7+NJcNuYzuOa0IjiIR68nwu1bBt56DboNa1TZlWVO+hv98+R9eXP0iVcEqBnYayA1jb+DMfme2+KntSiml1P5osJVMK+5GjPfd4d/lhVUv8ODiB/n5+J8ffEEf/gWWvwqn3wp9T2h1uzqyUCRE2YYynvzySWZvnY3b4WZy6WQuPOpCRnQdofOwlFJKpZ0GW8m04m7EeCW5JZw98GyeWfkMlw+7nB45PQ68kFVvwTu/h2Hnw7FXtbpNHdW+4D5mrZjFv5b9i637tlKcU8z1o6/n6wO+rqNYSimlMkqDrWRaeTdivCuHXcnzq57ngUUP8H8T/u/ATt67AWZdDt0Gw7Q7rPc2qgOyo3oHjy97nKeWP0VlsJKx3cfys2N+xom9TtTHNSillDokNNhKppV3I8Yrzi3m3IHn8syKZ7h0yKX0zu/dshMjEetZWpEQfOOf4MlJS3s6ijXla3h06aO8tPolwibMKUecwmVDLmNY12HZbppSSqkORoOtZKLvRjQmLaNJVw2/ihdXv8idn93Jn078U8tOmvsArH0fpt0JRf1b3YaOYtWeVdy76F7eWPcGXqeXcwaew8zBM1se5CqllFJppsFWMi4fYCAcTMvT2bsGuvKtwd/i/kX3M3PITIZ2Gdr8CTtXwZs3w4DTYPS3W11/R7Byz0ru/fxe3vzqTfwuP1cMu4JLBl9y2L74WSmlVPuhwVYyLvup7KGatL0K57Ihl/H08qe5bf5tPDj5wdR3vYVD8PzV1qXM6X/TeVr7sWLPiliQlePO4YphV/Dtwd+m0FeY7aYppZRSQCuDLRHpDPwHKAXWARcYY/Yk5BkJ/B3IB8LA740x/2lNvRnn8lrrNNyRGJXryeWqEVdxy5xb+HDTh0zsNTF5xjn3wca5cO4/IL84bfUfbrZUbeFvn/2Nl9e8TMAd4MrhV/Ltwd+mwFuQ7aYppZRSjbT23SM/Bd42xgwE3rb3E1UD3zbGDAFOB24XkbY97OD2W+s03ZEYdcGRF9A7rze3LbiNcCTcNEPFFnj3DzBwMgw9N611Hy4q6yu5ff7tnPXcWby+7nUuHXopr5/7Oj8Y9QMNtJRSSrVJrQ22ZgCP2tuPAmcnZjDGrDDGrLS3NwPbgTS9LDBDopcR03RHYpTb6ea60dexcs9KXl7zctMMb/wCwvUw9Va9fJggGA7y+LLHOfPZM/nHkn8wuXQyL339JX485scaZCmllGrTWjtnq7sxZou9vRVo9p00InIM4AFWt7LezIpdRkxvsAUwpc8UHi16lLsW3sXk0sn4XfYo2tr3YcksOPF/oXO/tNfbXhljeGv9W9w+/3bWV67n2B7H8uOxP2Zw0eBsN00ppZRqETHGNJ9B5C0g2aPPfw48aowpjMu7xxjTKUU5xUAZMNMY82mKPFcCVwJ07959zJNPPtmSn6FVqqqqyM3NbZTWedcChi/+NQtG/ZGKgqPSXueq2lXcse0OphZM5YzCM5BIkLHzfogjEmTuuL8RcXrTXmd7Eu2TtXVreW7Pc6ytW0uxu5gZnWYw2DdYX6mTJcn+r6js035pe7RP2qZM98tJJ5003xgzNtmx/Y5sGWNOTXVMRLaJSLExZosdTG1PkS8feAX4eapAy67rfuB+gLFjx5pJkybtr3mtVlZWRpN61jphMYwefnRG3kU4iUmseG8Fb69/m+tOuY5eS16E6o1w0X844agpaa+vvXn6zad5h3d4c+ubdPV35VcTfsWMATNwOfTm2WxK+n9FZZ32S9ujfdI2ZbNfWvvb60VgJnCLvX4hMYOIeIDngMeMMbNaWd+hEXv0Q/ruRkz047E/pmxjGX+a/QfuWPAa9JsER52esfragz21e7hv0X08sfkJvC4v14y8hpmDZxJwB7LdNKWUUuqgtTbYugV4SkQuB74CLgAQkbHA1caYK+y0E4AiEbnUPu9SY8zCVtadOe5osJX+OVtRPXJ6cOXwK7ljwR28SzUnnfqrjNXV1tWGanl82eM8uPhBqkPVTMidwO/O+B1d/F2y3TSllFKq1VoVbBljdgGnJEmfB1xhb/8L+Fdr6jnkMnQ3YqKZfc7gtU//wm979GRMl/7kZ7S2tidiIryy5hXu/OxOtu7byom9TuRHY37EhoUbNNBSSil12Gjtox8OTxm8GzGe+4O/8ptde9hNmL/M+0tG62prPt3yKRe+fCE3fXgTnX2deWjKQ9x1yl30L9T3QCqllDq86IzjZFyZeahpI7vXwPyHGTL628wsPoKHljzESb1PYlLvSZmrsw1YuWclf53/Vz7c9CE9c3pyy8RbmNp3Kg7RuF8ppdThSYOtZA7FyNb7fwGHC078X64NdOaTzZ/wi49+waxps+iRk+xJG+3b1n1b+fvnf+f5Vc+T487hJ2N+wkVHX4S3gz/mQiml1OFPhxOScWV4gvze9bDoSRhzKeT1wOP08KcT/0QwHOR/3/9fQpFQZurNgor6Cm6bfxtnPXcWL61+iYuPvphXv/4qlw69VAMtpZRSHYIGW8k43SCOzD364aM7AYGvXRdL6pPfh/+b8H8s2L6AW+bcwv4eNtvW1YXreHTpo0x9ZioPL3mYyX2s1+vcOO5GCn1t+9WYSimlVDrpZcRkRKzRrWBN+suu3AoLHoORF0FBSaNDZ/U7ixW7V/Dw0ocpzS/lksGXpL/+DAtGgry0+iXu/fxetuzbwnElx/Gj0T/iqM7pfxK/Ukop1R5osJWKy5uZka1P7oJIEI77YdLDPxzzQ76q+Io/zv0jnXydOLPfmelvQwYEw0FeXP0iDyx+gE1VmxhaNJTfHvdbji0+NttNU0oppbJKg61UXP70z9mq3g1zH4Kh50JR8kccOMTBHyb+gWvfvpabPrwJoE0HXMFwkBdWv8ADix5g877NDC0ayk3H3sTEkon6DkOllFIKDbZSc3nTH2zNuR+C++D4HzebLeAOcPcpd8cCrj21e7j46IvbVPBSXlfO0yue5ollT7C9ZjvDuwznF+N/wfElx7epdiqllFLZpsFWKumesxWsgTkPwJGnQ/fB+80eDbh+9sHPuHXurXy5+0tuOvamrL8ncE35Gp5Y9gQvrH6BmlANE4on8JvjfsPXen5NgyyllFIqCQ22UnGn+TLi509C9U6Y8P0WnxJwB7jtpNu47/P7uOfze5i3bR6/nPBLvtbza+lrVwtUB6t5fd3rPLfqOT7b/hluh5uz+p3FJYMv4chORx7StiillFLtjQZbqbj96Xs3YiQCn9wNxSOg9PgDOtUhDr438nuM6zGOX3/ya6568yomlkzk2pHXMqTLkPS0L4naUC2fbP6Et9a/xVtfvUV1qJrS/FJ+PObHTOs/Td9dqJRSSrWQBlupuP3WhPZ0WPkG7FoJ5zxoPVbiIIztMZZZ02fxzy/+ycNLHubCVy5kdLfRnDPwHE4+4mTyPHmtaqIxhrXla5m9dTazt8zm480fUxOqIc+Tx5TSKZwz8BxGdB2hlwqVUkqpA6TBVirpvIz4yV2QXwJDzm5VMV6nlyuGXcGFR13IrBWzmLVyFr/46Be4PnYxqvsoRncbzeCiwZTml9Ijp0eT+V3GGOrCdeyo2cHWfVvZXLWZ1XtXs2z3MpbvXs6euj0AFOcUc2a/MzntiNMYVzwOt8PdqnYrpZRSHZkGW6m4/BCsbn05mxfCug/gtN9aT6ZPg1xPLpcOvZSZQ2by+Y7PKdtQxoebPuSBxQ8QMZFYPpe48Lq8eJ1egpEgNcEaQqbxq4DcDjcDCgdw0hEnMazLMI4tPpZeub10BEsppZRKEw22UnH703M34id3gScPxsxsfVkJRISR3UYysttIfjjmh9SEalixZwUbKzeyZd8W9gX3URuqpT5cj9vpJuAK4Hf56eLvQnFuMT0CPSjJK9GRK6WUUiqDNNhKxR1o/QT58o2w5FkY/z3wFaSnXc3wu/yM6DqCEV1HZLwupZRSSrWMvog6Fbev9ZcRP/27tT726ta3RymllFLtkgZbqbgD1jsMw6H9502mtsJ64fSQs6Gwd3rbppRSSql2Q4OtVNx+ax06yHlbn/0T6ioO6CGmSimllDr8aLCVistnrQ9mknwkYr0H8YgJUDI6ve1SSimlVLuiwVYq0WdUHcy8rVVvwZ51cMx309okpZRSSrU/GmylEr2MeDB3JM59AHJ7wKBp6W2TUkoppdodDbZSiQVbBziytWs1rHwTxlwKLk/am6WUUkqp9kWDrVRiwdYBztma9xA4nFawpZRSSqkOT4OtVKJztg7kbsT6ausuxKOnQ35xZtqllFJKqXZFg61UDuZuxMVPQ225ToxXSimlVIwGW6nE7kZs4QR5Y2DOA9B9qPXIB6WUUkopNNhK7UAnyG+YDdsWw7grQCRz7VJKKaVUu6LBVioHOkF+zv3gLYDhF2SuTUoppZRqd1oVbIlIZxF5U0RW2utOzeTNF5GNInJXa+o8ZA7kdT2VW+GLF2DUJeDJyWy7lFJKKdWutHZk66fA28aYgcDb9n4qvwXeb2V9h86BTJCf/yhEQjDu8sy2SSmllFLtTmuDrRnAo/b2o8DZyTKJyBigO/BGK+s7dETA5d//nK1w0Hq21oBToaj/oWmbUkoppdqN1gZb3Y0xW+ztrVgBVSMi4gD+AvxPK+s69Nz+/d+NuOwlqNoKx1x5aNqklFJKqXbFtb8MIvIW0CPJoZ/H7xhjjIiYJPmuAV41xmyU/dylJyJXAlcCdO/enbKysv01r9WqqqpS1jM+4mDPhjUsb6YdIz/7M15fd2ZvcsHm1PlUyzXXJyp7tF/aJu2Xtkf7pG3KZr/sN9gyxpya6piIbBORYmPMFhEpBrYnyTYBmCgi1wC5gEdEqowxTeZ3GWPuB+4HGDt2rJk0aVILf4yDV1ZWRsp6FhdSXFRAcarjW5dA2VI47bdMOu6UTDWxw2m2T1TWaL+0TdovbY/2SduUzX7Zb7C1Hy8CM4Fb7PULiRmMMRdHt0XkUmBsskCrTXIHmp+zNfcBayL9qEsOXZuUUkop1a60ds7WLcBpIrISONXeR0TGisiDrW1c1vkLoWZv8mM1e2HRUzDsPAh0PrTtUkoppVS70aqRLWPMLqDJ9TNjzDzgiiTpjwCPtKbOQypQZF0qTGbhv61Rr3H6HkSllFJKpaZPkG+OvzPU7G6aHolYlxB7HQM9Rx76dimllFKq3dBgqzmBIqjZA5Fw4/RVb8HuNXDsVdlpl1JKKaXaDQ22mhPoDCYCteWN0+fcB7k94Ojp2WmXUkoppdoNDbaaEyiy1tVxlxJ3rrRGtsZ+B1ye7LRLKaWUUu2GBlvN8dt3GVbvakib8wA43DDm0qw0SSmllFLtiwZbzYk+0iE6Sb6u0roLceg5kNfkzURKKaWUUk1osNWcQMLI1sInoL4SjtGJ8UoppZRqGQ22mhObs7XLetzDnPugZAz0GpPddimllFKq3dBgqzmeXGt+VvVuWPMO7FoFx16d7VYppZRSqh3RYKs5ItboVvUu+PguyOkGg8/OdquUUkop1Y609kXUh79AEax4HfZth8m/18c9KKWUUuqA6MjW/oy/Gmr3QtEAOObKbLdGKaWUUu2Mjmztz+hvQ7+TwOnRUS2llFJKHTANtlqisHe2W6CUUkqpdkovIyqllFJKZZAGW0oppZRSGaTBllJKKaVUBmmwpZRSSimVQRpsKaWUUkplkAZbSimllFIZpMGWUkoppVQGabCllFJKKZVBGmwppZRSSmWQBltKKaWUUhmkwZZSSimlVAZpsKWUUkoplUFijMl2G5ISkR3AV4egqi7AzkNQj2o57ZO2SfulbdJ+aXu0T9qmTPdLH2NM12QH2mywdaiIyDxjzNhst0M10D5pm7Rf2ibtl7ZH+6Rtyma/6GVEpZRSSqkM0mBLKaWUUiqDNNiC+7PdANWE9knbpP3SNmm/tD3aJ21T1vqlw8/ZUkoppZTKJB3ZUkoppZTKIA22lFJKKaUyqMMGWyJyuogsF5FVIvLTbLenIxGRh0Rku4gsiUvrLCJvishKe93JThcRudPup0UiMjp7LT98iUhvEXlXRL4QkaUicr2drv2SRSLiE5E5IvK53S+/ttP7ishs+/P/j4h47HSvvb/KPl6azfYf7kTEKSKficjL9r72SxaJyDoRWSwiC0Vknp3WJr7DOmSwJSJO4G5gKjAYuEhEBme3VR3KI8DpCWk/Bd42xgwE3rb3weqjgfZyJfD3Q9TGjiYE/MQYMxgYD1xr/5/QfsmuOuBkY8wIYCRwuoiMB24FbjPGDAD2AJfb+S8H9tjpt9n5VOZcDyyL29d+yb6TjDEj456n1Sa+wzpksAUcA6wyxqwxxtQDTwIzstymDsMY8z6wOyF5BvCovf0ocHZc+mPG8ilQKCLFh6alHYcxZosxZoG9XYn1C6QE7Zessj/fKnvXbS8GOBmYZacn9ku0v2YBp4iIHKLmdigi0gs4E3jQ3he0X9qiNvEd1lGDrRJgQ9z+RjtNZU93Y8wWe3sr0N3e1r46xOxLHKOA2Wi/ZJ19qWohsB14E1gN7DXGhOws8Z99rF/s4+VA0aFtcYdxO3AjELH3i9B+yTYDvCEi80XkSjutTXyHuTJVsFIHyxhjRESfSZIFIpILPAP80BhTEf/Ht/ZLdhhjwsBIESkEngMGZblJHZ6InAVsN8bMF5FJ2W6PijneGLNJRLoBb4rIl/EHs/kd1lFHtjYBveP2e9lpKnu2RYdw7fV2O1376hARETdWoPW4MeZZO1n7pY0wxuwF3gUmYF3yiP6xHP/Zx/rFPl4A7DrETe0IjgOmi8g6rGkoJwN3oP2SVcaYTfZ6O9YfJsfQRr7DOmqwNRcYaN854gEuBF7Mcps6uheBmfb2TOCFuPRv23eOjAfK44aEVZrY80f+ASwzxvw17pD2SxaJSFd7RAsR8QOnYc2nexc4z86W2C/R/joPeMfok6vTzhjzM2NML2NMKdbvj3eMMRej/ZI1IpIjInnRbWAysIQ28h3WYZ8gLyJnYF1zdwIPGWN+n+UmdRgi8gQwCegCbANuBp4HngKOAL4CLjDG7LaDgLuw7l6sBi4zxszLRrsPZyJyPPABsJiGOSg3Yc3b0n7JEhEZjjWp14n1x/FTxpjfiEg/rBGVzsBnwCXGmDoR8QH/xJpztxu40BizJjut7xjsy4j/Y4w5S/sle+zP/jl71wX82xjzexEpog18h3XYYEsppZRS6lDoqJcRlVJKKaUOCQ22lFJKKaUySIMtpZRSSqkM0mBLKaWUUiqDNNhSSimllMogDbaUUkoppTJIgy2llFJKqQzSYEsppZRSKoM02FJKKaWUyiANtpRSSimlMkiDLaWUUkqpDNJgSymllFIqgzTYUkoppZTKIA22lFJKKaUySIMtpZRSSqkM0mBLKaWUUiqDNNhSSimllMogDbaUUkoppTJIgy2llFJKqQzSYEsppZRSKoM02FJKKaWUyiANtpRSSimlMkiDLaWUUkqpDNJgSymllFIqgzTYUkoppZTKIA22lFJKKaUySIMtpZRSSqkM0mBLKaWUUiqDNNhSSimllMogDbaUUkoppTJIgy2llFJKqQzSYEsppZRSKoM02FJKdVgiMlFElme7HUqpw5sGW0qpVhORdSJSIyJVIrJVRB4Rkdxst2t/jDEfGGOOOpR1iogRkQFJ0otF5EUR2WznKT2U7VJKZY4GW0qpdJlmjMkFRgKjgJ+ls3ARcaWzvEyyg81LD/C0CPBf4Nz0t0gplU0abCml0soYsxV4HSvoQkTGi8jHIrJXRD4XkUnRvCLSV0TeF5FKEXlLRO4WkX/Zx0rtEZ7LRWQ98I6d/h0RWSYie0TkdRHpY6eLiNwmIttFpEJEFovIUPvYGSLyhV3PJhH5Hzt9kohsjGvP0SJSZrd1qYhMjzv2iN2+V+xyZotI/zR+btuMMfcAc9NVplKqbdBgSymVViLSC5gKrBKREuAV4HdAZ+B/gGdEpKud/d/AHKAI+BXwrSRFnggcDUwRkRnATcA5QFfgA+AJO99k4ATgSKAAuADYZR/7B3CVMSYPGIoduCW02w28BLwBdAN+ADwuIvGXGS8Efg10AlYBv2/hx6KU6sA02FJKpcvzIlIJbAC2AzcDlwCvGmNeNcZEjDFvAvOAM0TkCGAc8EtjTL0x5kPgxSTl/soYs88YUwNcDfzBGLPMGBMC/h8w0h7dCgJ5wCBA7Dxb7DKCwGARyTfG7DHGLEhSz3ggF7jFbs87wMvARXF5njPGzLHrfhx79E4ppZqjwZZSKl3OtkeOJmEFPF2APsD59mW5vSKyFzgeKAZ6AruNMdVxZWxIUm58Wh/gjriydgMClNjB0V3A3cB2EblfRPLt884FzgC+EpH3RGRCknp6AhuMMZG4tK+Akrj9rXHb1VjBGQAisiiuXd8E7on7ue9JUp9SqoPQYEsplVbGmPeAR4A/YwVK/zTGFMYtOcaYW4AtQGcRCcSd3jtZkXHbG7AuB8aX5zfGfGzXfacxZgwwGOty4g12+lxjzAysy4PPA08lqWcz0FtE4r8XjwA2tfDnHh5tE9bl0Wvi2nhNS8pQSh2eNNhSSmXC7cBpwMfANBGZIiJOEfHZk9J7GWO+wrqk+CsR8dijTdP2U+69wM9EZAiAiBSIyPn29jgROdaee7UPqAUidtkXi0iBMSYIVGDd+ZdoNtZo1Y0i4rYn8k8DnmzlZ5GMx/4soovT/hl8gNfO47X3lVLtnAZbSqm0M8bsAB4DrgOik9p3YI1M3UDDd8/FwASsiey/A/4D1DVT7nPArcCTIlIBLMGajA+QDzwA7MG6/LcL+JN97FvAOvucq+16E8uuxwqupgI7gXuAbxtjvjzgD2D/lgI1cctldnoNUGVvf2nvK6XaOTHG7D+XUkodAiLyH+BLY8zN2W6LUkqli45sKaWyxr70119EHCJyOtYo2PPZbpdSSqVTu3kis1LqsNQDeBbrOVsbge8ZYz7LbpOUUiq99DKiUkoppVQG6WVEpZRSSqkMarOXEbt06WJKS0szXs++ffvIycnJeD2q5bRP2ibtl7ZJ+6Xt0T5pmzLdL/Pnz99pjOma7FibDbZKS0uZN29exuspKytj0qRJGa9HtZz2Sduk/dI2ab+0PdonbVOm+0VEvkp1TC8jKqWUUkplkAZbSimllFIZpMGWUkoppVQGtdk5W0oppVQwGGTjxo3U1tZmuyktVlBQwLJly7LdDJUgXf3i8/no1asXbre7xedosKWUUqrN2rhxI3l5eZSWliIi2W5Oi1RWVpKXl5ftZqgE6egXYwy7du1i48aN9O3bt8Xn6WVEpZRSbVZtbS1FRUXtJtBShzcRoaio6IBHWjXYUkop1aZpoKXakoP596jBllJKKaVUBmmwpZRSSjXD6XQycuRIhg4dyvnnn091dXW2m9QiL774Irfccku2m6FIU7AlIg+JyHYRWZLiuIjInSKySkQWicjodNSrlFJKZZrf72fhwoUsWbIEj8fDvffe2+h4KBTKUsuaN336dH76059muxmK9I1sPQKc3szxqcBAe7kS+Hua6lVKKaUOmYkTJ7Jq1SrKysqYOHEi06dPZ/DgwYTDYW644QbGjRvHhAkTuO+++wCIRCJcc801DBo0iNNOO40zzjiDWbNmAdZr6W6++WZGjx7NsGHD+PLLLwGYM2cOEyZMYNSoUXzta19j+fLlADzyyCOcc845nH766QwcOJAbb7wx1q7//ve/jB49mhEjRnDKKafE8n//+98HYMeOHZx77rmMGzfqX1X5AAAgAElEQVSOcePG8dFHHwHw3nvvMXLkSEaOHMmoUaOorKw8NB9kB5OWRz8YY94XkdJmsswAHjPGGOBTESkUkWJjzJZ01H+w9mych4m0n2e3KKVUR/brl5byxeaKtJY5uGc+N08b0qK8oVCI1157jdNPt8YWFixYwJIlS+jbty/3338/BQUFzJ07l507d3L66aczefJk5s+fz7p16/jiiy/Yvn07Rx99NN/5zndiZXbp0oUFCxZwzz338Oc//5kHH3yQQYMG8cEHH+ByuXjrrbe46aabeOaZZwBYuHAhn332GV6vl6OOOoof/OAH+Hw+vvvd7/L+++/Tt29fdu/e3aTt119/PT/60Y84/vjjWb9+PVOmTGHZsmX8+c9/5u677+a4446jqqoKn8+Xhk9VJTpUz9kqATbE7W+00xoFWyJyJdbIF927d6esrCxjDTKRMCZ4PcZZz/uvDCTiGorxDAV66Z0vWVZVVZXRvlcHR/ulbTrc+6WgoCA22hKsDxIOh9NafrA+uN/RnJqaGoYPHw7AhAkTuOCCC5g9ezZjxoyhS5cuVFZW8uqrr7JkyRKeeuopjDFUVlby+eef88477zBt2jT27dtHTk4OEydOpKamhsrKSowxTJ48mcrKSgYNGsTTTz9NZWUlmzZt4sYbb2T16tWICMGg1cba2lpOOOEEHA4HwWCQI488kmXLlrF3714mTJgQa4vb7Y7lr6+vp7KykjfffJMlSxpm+pSXl7NlyxbGjh3L9ddfzwUXXMD06dMpKSlJ6+fbloTD4bSN3NXW1h7Q/7s29VBTY8z9wP0AY8eONZl8O/eeymruvP9qJnT+Am/nLwj6Z4GZhVuKKOpyIkVdT6Co8/G43Z0y1gaVXKbfzK4OjvZL23S498uyZctiD6L83bkjs9IGv9/PokWLGqUFAgHy8/NjbXM6ndx9991MmTKl0cMzy8rK8Pl8sX2Xy4Xf7ycvLy/2zKa8vDzy8/MxxpCXl8ett97KaaedxksvvcS6deuYNGkSeXl5+Hw+cnNzY2V5vV48Hg9+vx+3293kgZ0+nw+Px0NeXh7GGObMmdNk5Ormm2/mnHPO4dVXX2XKlCm8/vrrDBo0KCOfY7al82GzPp+PUaNGtTj/oQq2NgG94/Z72WlZ0ykvwPjTz+Oqf87nG0d9h5/kOdi57h0qvfPZXv86W7c/Cwh+X2+8vh64XQU4nD4cDi8Ohw+nw2tt22lOhw+Hw4fD4cHpjG7HH/fG0pxOHyIeHI42FesqpZQ6SFOmTOHvf/87J598MgArVqygpKSE4447jkcffZSZM2eyY8cOysrK+OY3v9lsWeXl5bERpkceeWS/dY8fP55rrrmGtWvXxi4jdu7cuVGeyZMn87e//Y0bbrgBsC5Hjhw5ktWrVzNs2DCGDRvG3Llz+fLLLw/bYCubDtVv+xeB74vIk8CxQHm252sBTBnSgwuO8vDk8m10Ku7PjZf9hPr1leybv5nda+dQlbeIYKethAqrqPftxjhDREwtkUgdkUgd4XAtEDno+kWcdtDlxeFoydqDQzw4nF5r3YLzxOHBaa8dDm/S80VceulUKaVa4YorrmDdunWMHj2acDhM9+7def755zn33HN5++23GTx4ML1792b06NEUFBQ0W9aNN97IzJkz+d3vfseZZ56537q7du3K/fffzznnnEMkEqFbt268+eabjfLceeedXHvttQwfPpxQKMQJJ5zAvffey+233867776Lw+FgyJAhTJ06tVWfg0pOrDnrrSxE5AlgEtAF2Ab8f/bOOz6qKv3/7zt9JmXSE5JASKWmACGUCASQokYQu7Iq6toQ/a6rIOq6q/5Y14K6CiyKDVEUBQWxg3QQgYQmEEgjEAk1kJ6ZJDP398edTAqhJyRhzvv1youZe88955m5ZOaT5zznc/4FaAFkWX5XUr7JZ6KsWKwA7pVlOe1sfSYmJsppaWdt0iysWrWK5UV+fL7pIK/dFMetfZUEnFxto3JPIeXpx7DmFIFNBrWEPswTfZQX+igvtMHuoKpRhJfdit1mxW63OH4UMeYUZnYLdudzC3Z7lePHil2uUq6VqxoeP+e/Vi5F7NWhcogzLZKkRSVpkVQahwjTOo5rUEkaJFX9847jzvNapKbaOh9r6p2vf63Sn0pSxtyxYxe9e/d1jl93ri6u2rZKfBokSd0M74PgbFzp01XtlSv9vmRkZNCtW7fWDuOCaDxdVVZWhru7O4WFhSQlJbFhwwaCgoJaMULXpDmnEZv6fylJUrosy4lNtW+u1Yh3nOO8DDzaHGM1F7LdTuX27WgOH+HF24aQf7KCZxf/QYi3keQoPyStGlN8AKb4AOxVNqrySrBkn8KaVUTJsgOw7ACSQY0+wgtDtCK+dH6Xf/8uu70GWT5fcdbo3wYiz4rdXo0s1yDbq7HLtY9rHI+rHY9rsNkrsdc4zss1TV/nPF59Ua8rLf1Cr5DqCS9NPfGncWQQa8VePZHYQPw1euxsq6knGDWN+ql3znFMdZpA1DSKq/G5ekJUiEeB4IokNTWVoqIiqqqqeP7554XQckFct2hIksh/4EFMiYlo77yDWeN7c/Ps33j4s3QWTxxIVECd+lXp1BhivDHEKMXytrIqrDnFWLOLsGSdwrKnEAC1p86Z9TJEeaH21Lf4y1DqvjSo1aYWH+tikGUZWbY5BFh1nTCrJ8bsDqEmy4qY274tjbi4Ho7nishr8rFsczx39OEUfHWPa/s87ZxDSNpqrIqQlGscwrWxyKypF28NzZNJPB9OF491orCx6DtPwXge4rHu2tMzirK8h5MntU3HpdIK8SgQnIErebWo4PxwWbElSRK6qEg0R5TSMU+Dlo8m9OWGWb9x79wtLJ6YjJ9702JJ7a7DFO+PKd4fWZaxnbRgySrCmlOEZe9JKrYeA0ATYMQQ5Y0+0gt9pBmVwfXebkmSkCRFEML5+bdIUjm+voNbNK6LRZbtDtFYUycgHeKs8WNF6Nmcou/c4rGm3vMLFI92RXyer3isPQcXVkawbfvFvnNSPWHWWCCeQTw2mJ4+m1DUnlEgnmvK+sxZR12D4w3aiPpGgUBwgbjet78Du93O5iB/TNk5xDuOhXqb+OCeRG6fs5EH56Xx+QP9MWjP/he5JElofI24+xpx798B2S5TfbhcyXpln6J8yxHKfisACXShHs7Mlz7ME0kjtqZsb0iSCklSVqJeCdRmHc9HPKanbyYhIc4h4GznLR7tjn6bEpJnEo+1tYuyXNFQKJ4mGqvrBKXcvP5LZ0KZlm5Yz9ikKKyfLXQsRGk6+3iOOsfTsoUNaydleQ+nThkdQrFePKrGj2tFZ61oFJ8/AsHlwmXFlkqlYq+nN14hwdScOoXGW5kiTOjoxVu3JvDI/K08tXAH79zeC5Xq/P+SlVQSuhB3dCHueAwJRa6xYz1QgjWnCGt2EaVr8ildlY+kVaHr7Ikhygt9pFJsL13AOAJBc6AIB/V5iUdJOom3d9JliOriULKODTN39R/XicHTM38NBKZTBNbPNjYlQqubEJJNiEJ7DTU1pQ3qHhs+rh23Lp4LZeu2i3nHVA2yjPVFWYOsXoNMorZO1DUWho0W2CjnHMedglJX73rH89pxnf01vF6WbdjtNY6MovIjsouC9obLii0AjWzDajBSlZuLpk8f5/FrYjsw9ZquvPLTXnzcdDyf2h2t+uL+CpQ0KgyRXhgivWAk2C01WHMd9V45RRT/lAeAyqRBH2FGH+WNPsoLja9BfKAIBBeAknXUATrU7bhETKlzrF8r2MRiFEf2T5Zr2Lp1C/HxPc8yVX2GqegGmcV69ZSN6xXrXV9jK0euqa6Lxbkopr5YVPq60CnqM+Hj/R5lZU30JUlINBZgjp+LPsdFXdegHVK996CpdgJXxKXFlkGjwaI3YM3OwVRPbAE8NDiCE6VWPli/nzWZx7mlTyjjeocS4mW8pDFVBg3G7r4Yu/sCYCupUmq9spXMV+UuR7G9lx59pGOlY6QXag/dJY0rEAjaB0qdoxaHe855tC/Gx2dgywZ1ESgZqep6Qq5WkFXVywzWE2wNMpB12cPCQi8Mhg4oi9qVHxkZnM+VrKZj1EbnZId4tdedszfVR/3np3PsWCHPPPMaW7bsxMvLE51Oy//93714eXkyY8YnfPXVzNOuKSs7wxvTSLQ1eH5RIvH0Pi9WJJ7pnBCJl45Liy03NzfKamxYc7JPOydJEv9I7U6/CF8+XJ/L9GWZTF+WSd/O3lwfH8y1sR3OWEB/Iag9dZh6BWDqFYAsy9ScqFSmHLOKqNxTSEX6UQA0gSbnlKM+wjWL7QUCQftBktSo1WrOd2HMmSguzkCn82ueoM4DuQmhdtddg7nrrr+wYME3gMyBAwdYuvR7goI6oFYbMRrDqC/aLJWVGAx6Zx9NC7oznWs+kdisNLtIPNM5HI9Vde2g3nMc7VT1Hrd9kejS39ienh4cKa+kJCuLM7mejOgeyIjugeSfrGDpjgKWbi/gn9/u5sXv9jAw0pcx8cGM6hmEp+H8/go9G5IkofU3ofU34d4/WCm2LyhTsl45RZRtOkLZhgJQ1RXbG6K80HUSxfYCgUDQHNTPGEkSrFixAp1Oz8SJk5xtIiK68re/dWX16tVIkgat1pMXXngBd3d3nnrqKawWDb17D+D777+nc+fOzJs3j+nTpyNJEnFxcXz66afk5eVx3333ceLECfz9/fn444/p1CmMhQsX8uKLL6JWqzGbzaxduxabzcbUqVNZvXo1VquVRx99lIceeqhB3I1FYlOPz3auKbHXoiKxXmay5agTdEpdakALj3dmXFpsefn4wOFjHD16lJhztO3oY+LRoVE8OjSKvUdKWLq9gO92FjB50U6eW7KLoV38GRMfwvBuAedcwXi+SCoJXagHulAPSOmIXG3HerAEq2PKsXRVPqUrHcX24ea6YvsObqLYXiAQXHn8NBWO/NG8fQbFwjWvnPH07t276d2790V3v3v3bqZNm8Zvv/2Gn58fJ0+eBOCxxx7jnnvu4Z577uGjjz7i8ccfZ8mSJbz00kv88ssvhISEUFRUBMCHH36I2Wxmy5YtWK1WkpOTGTlyJOHh4c5xGovE9oBcT8CdVQg2g2CUJDVVVZf15TXApcWWr38gsJeTtmpspaWoz9PGv2uQJ11HezJ5VBe25RexdHsBP/xxmF92H8VDryE1PphbEkPp1dGrWdOYkrZesf0osFcqxfaW7FNYc4oo/nE/4Ci2j6xnruojiu0FAoGgOXj00UdZv349Op2O119//ZztV65cyS233IKfnzIVWrtB9MaNG/nmm28AuOuuu5gyZQoAycnJTJgwgVtvvZUbb7wRgGXLlrFz504WLVoEKBtVZ2VlNRBb7ZH604bK85Ydr6qqtGUHOAsuLbYCHbuql7q5UZWTgzEh4YKulySJ3p286d3Jm+dTu/N7biFfb/2TJdsO8cXmg0QFuHNrYijjeoXi79H8vkwqowZjD1+MPRzF9sVWLA6LCWt2EZV/nABA7d2o2N5dFNsLBIJ2yFkyUC1Fjx49+Prrr53PZ82axYkTJ0hMbLgFnkajwW6v22HCYrFc1HjvvvsumzZt4ocffqBPnz6kp6cjyzIzZsxg1KhRF/ciBK2OSxf6BIaGAlDu5oY1J+eS+lKrJJKj/Hjz1gQ2PzecV26MxWzU8vKPe+n/nxU8MC+NNZnHsdtbbo5abdbj1jsQn1u7EPRMEoF/74PX2Ei0we5U7jrByS/2cXjaJo7+N52i73Op3HsSu/XCPX0EAoHAVRg2bBgWi4XZs2c7j1VUVJzWrnPnzmzduhWA7du3s3//fuf1CxcupLBQWWleO404cOBAFixYAMD8+fMZNGgQADk5OfTr14+XXnoJf39/8vPzGTVqFLNnz6a6WtlrNjMzk/Ly8hZ6xYKWwKUzW0ajCcluo9Jkwpp9aWKrPh4GLbcndeL2pE5kHytjYXo+X6f/yfI9Rwn3c2N8v07c0qcjZtOlF9WfCUmS0AaY0AaYcB/gKLY/VK/Y/vcCytYfApWErmO9YvuOHqLYXiAQCBxIksSSJUt44okneO211/D398fNzY1XX321QbubbrqJefPm0aNHD3r37k1MjFIJ3KNHD5577jmGDBmCWq2mV69ezJ07lxkzZnDvvffy+uuvOwvkASZPnkxWVhayLDN8+HDi4+OJi4sjLy+P3r17I8sy/v7+LFmy5LK/F4KLR5Ivx5LRiyAxMVFOS0tr0TGeW/8cup/tGCsqub2sgk5z5rTYWNYaGz/vOsK8jQdIP3AKg1bFDQkh3DUgjB7B5hYb90zI1TbF2T5b8fiqPlQGMkg6Ffpws7KlUKQX2qDLX2y/evVqUlJSLuuYgnMj7kvb5Eq/LxkZGXTr1q21w7ggSktL8TjPGmDB5aM570tT/y8lSUqXZTmxqfYundkyLe5JjW4H1TodVdnNvMKlEXqNmrEJIYxNCGF3QTGfbjzAku2HWLAlnz5h3tx/VTijegShvkzCRtKqMUR5Y4jyxgzYK6qVYntHzZflB0exvZsWfaRZ2VA7yguNz6V55ggEAoFA4Gq4tNgyVrtTppGo0WqpKijAXl6Oys2txcftEWzmlZvieOaabixMz2fexgNMnL+Vjj5G7k8O55bEjrjpL++tUZm0GHv6YeyprJipKbY6C+0t2UVU7nQU2/sYFIuJKMVcVRTbCwQCgUBwdlxabNk0NSCrkTUaKnVarLn7Mcb2vGzjm01a/joognuTw1m2+wjvr8vlhe/28ObyTMb3D2PCwM4EerZOJklj1qPpE4hbn0DF2f54JdasU1hyiqnYcZzyzUcA0HZwq6v3Cjej0rXjTekEAoFAIGgBXFpsyRobyFpApsjDHWtO9mUVW7WoVRLXxHbgmtgOpB84xQfrcnlvTQ4frMvl+vhgHhgUQbcOnpc9rloaFNsnhyDbZKoOlTozX2W/FVC27hCoJXSdPDBEeqGP9kYX6o50kRt4CwQCgUBwpeDSYgutDaxGoIISL0+qLtH+oTnoE+ZNn7A+HCys4KMN+/kqLZ9vth5iaBd/Jg2Lok+YT2uHiKSW0HfyRN/JE4Z1wl5lo6pesX3JioPw60EknRp9hNmZ+dIEmoS5qkAgEAhcDpcWW5JORrJ4AhVUBAQ0q/3DpdLJ18QLY3rwxNUxfPp7Hh9tyOOm2RvpF+7DpGFRXBXl12aEi0qnxhDtjSG6rtjeklOsbKidXYRl70mKAZW7VjFXddR8abxFsb1AIBAIrnxcWmypdKC2eWDjCBWenlgzslo7pNMwm7RMGhbNfVeF88XmfOaszeGuDzcTH2pm4tAoRnQLRNXG9kFUmbSYYv0wxTqK7YssdcX2OUVU7jgOgMbXoBTaR3mhj/BC7dZyvmMCgUBwsajVamJjY5FlGbVazcyZMxk4cCB5eXmkpqaya9euSx4jJSWF6dOnk5iYSOfOnfHw8ECSJIKCgpg3bx5BQUHN8EoErYVLiy21XkJjN1Et26k0GKjOz8dusaAytL2Mi0mn4f6rwvlL/058s/UQs1fn8NCn6cQEujMxJYrUuA5o2mh9lMbLgCYxCLfEIKXY/liFYq6aXUTF9uOUbzoCUl2xvakc7FYbKr0othcIBK2P0Whk+/btAPzyyy8888wzrFmzpkXHXLVqFX5+fjz77LO8/PLLvPPOO85zsqxssqxStfxnfk1NDRqNS0uFZqFtfjtfJrR6NRqbDp0kYVGpQJapcmyx0FbRa9TckdSJlU8O4e3blb0c//bldoa9sYYvtxyk2mY/Rw+tiyRJaAPd8EgOwe+eHgT/cwD+j8TjeXUYkl5D2YYCgtPVFLy4kWOzd1C8/ADW3CLkmrb9ugQCgWtQUlKCt7f3acctFgv33nsvsbGxXHXVVaxateq047169XIer6ys5Pbbb6dbt26MGzeOysrKJscbPHgw2dnZ5OXl0aVLF+6++2569uxJfn4+r7/+On379iUuLo5//etfAJSXl3PdddcRHx9Pz549+fLLLwGYOnUq3bt3Jy4ujqeeegqACRMmODe3BnB3dwcUo9xBgwYxZswYunfvjs1mY/Lkyc6x3nvvvWZ6N10Hl5arWr0au00PWg0VVYrutGbnYGgHbsUatYqxCSFcHxfMrxlHmbkqm6e//oNZq3KYNCyKG3uFtNlMV30ktYQ+zBN9mCeew5Vi+y1L19HVPQxLTjGlKw9SugIkrQpdZ0+l5ivSC22I+2V3thcIBK3Lq5tfZe/Jvc3aZ1efrjyd9PRZ21RWVpKQkIDFYuHw4cOsXLnytDazZs1CkiT++OMP0tPTGTduHJmZmQ2O7927l5EjR5KZmcns2bMxmUxkZGSwc+dOevfu3eTY33//PbGxsQBkZWXxySef0L9/f5YtW0ZWVhabN29GlmXGjBnD2rVrOX78OMHBwfzwww8AFBcXU1hYyOLFi9m7dy+SJFFUVHTO92Xr1q3s2rWL8PBw5syZg9lsZsuWLVitVpKTkxk5ciTh4eHn7Eeg4NJiS2fQUI0Kg0FHaaUFu1qNNSe7tcO6IFQqiZE9ghjRPZAVGcf474pMpizayaxV2Tw2LJobEoLbheiqRaVTU+kH5pRwpdjeUoM111Fsn1NEyc95lACSQa1sKxQpVjoKBIKWpf404saNG7n77rtPq9Nav349jz32GAAxMTGEhYWRmZnZ4HjXrl2dx9euXcvjjz8OQFxcHHFxcQ36Gzp0KGq1mri4OKZNm0ZRURFhYWH0798fgGXLlrFs2TJ69eoFQFlZGVlZWQwaNIgnn3ySp59+mtTUVAYNGkRNTQ0Gg4H777+f1NRUUlNTz/mak5KSnGJq2bJl7Ny505kFKy4uJisrS4itC8ClxZYhfyXlXI/eaEAuraS6c1ibsH+4GCRJ4urugQzvFsCvGcf476+ZPLVwh0N0RTEmvn2JrlpUBg3G7r4Yu/sCYCutcoovS04RloyGKx31kWYMkV6ofQxCfAkEVxjnykBdDgYMGMCJEyc4fvx4i45TW7NVS1FREW71djiRZZlnnnmGhx566LRrt27dyo8//sg//vEPhg8fzj//+U82b97MihUrWLRoETNnzmTlypVoNBrsdqVEw263U1VV5eyj8VgzZsxg1KhRLfFSXYL29+3bjBiqlF8WnckIQFlYxzZl/3AxSJLEiO6BfP/YVbx3Vx8MWjV//2oHI99ay5Jth7DZ2+bG4+eL2kOHKd4f7xuj6TC5L0FP98X75mgM0d5Yc4sp+iabI6+nceTVLZxcmEn5tmPYSqytHbZAILhC2Lt3LzabDV9f3wbHBw0axPz58wFluu/gwYN06dKlwfHMzEzn8cGDB/P5558DsGvXLnbu3HlBcYwaNYqPPvqIsrIyAA4dOsSxY8coKCjAZDLxl7/8hcmTJ7N161bKysooLi7m2muv5a233mLHjh0AdO7cmfT0dACWLl1KdXX1GceaPXu283xmZibl5eUXFK+r49KZLaOqBgCNmwmAUh9vfNesx15VhUrXvvf8kySJUT2CGNEtkGV7jvDfX7P425fbmbEyi8eHR5MaF3zZNr1uSTTejVY6Hq90TjlaMgqpSD+qtPM31nl8RZhRmYTNhEAgOD9qa7ZAyfJ88sknqNUNV0tPnDiRRx55hNjYWFQqFXPnzkWv1zc4rtFonMcfeeQR7r33Xrp160a3bt3o06fPBcU0cuRIMjIyGDBgAKAUt3/22WdkZ2czefJkVCoVWq2W2bNnU1paytixY7FYLMiyzJtvvgnAAw88wNixY4mPj2f06NENsln1+etf/0peXh69e/dGlmX8/f1ZsmTJhb6NLo0ky20z05GYmCinpaW16BjrX7mFHXmP4JdaTEbaDrprJWI/XUD40m8xxMS06NiXG7td5ufdR/jvr5lkHi0jOsCdJ0bEMLpHUJvz6Vq9ejUpKSmX3I9sl6k+XO4UX9b9xchV9gY2E/pIL/SdzcJm4jxorvsiaF6u9PuSkZFBt3awaKk+paWleHh4tHYYgkY0531p6v+lJEnpsiwnNtXepTNbHhpFaNrVyttQJimzqlU5OVec2FKpJK6N7cDoHkH88Mdh/vtrJhPnb6VbB0/+PiKGq7sFXHE1TpJKQhfiji7EHY/Bocg2O1V/likGqzlFlG0ooGztIVBJ6Dp6KPVeUV7oOnkiaVx6hl0gEAgEzUiziC1JkkYDbwNq4ANZll9pdL4T8Ang5WgzVZblH5tj7EvBXauIrarqGlSynfIaQJLafd3W2VCpJK6PD+ba2A58u/0Qb6/I4oF5acSHmnliRAxDYvyvONFVi6RWOW0mGF5vT8ecYiw5RZSuyqd0ZT5oVOg7eyp7OkZ6oQ12R1Jfme+JQCAQCFqeSxZbkiSpgVnACOBPYIskSUtlWd5Tr9k/gK9kWZ4tSVJ34Eeg86WOfal46pQvUEtlFTqVisrqKrQdO2JtpysSLwS1SuLG3qFcHx/MN1v/5J0V2Uz4eAt9wrx5ckQMA6P8zt1JO+e0PR3PZDOhd2yoLWwmBAKBQHARNEdmKwnIlmU5F0CSpAXAWKC+2JIBT8djM1DQDONeMu5ax7ShpQajTkep1YI+MpKqdua1dSlo1Spu69uJcb1C+TItn1krs7nzg030j/DhyZFd6NvZp7VDvGycZjNRVoU1pwmbCTct+sg68SVsJgQCgUBwNppDbIUA+fWe/wn0a9TmBWCZJEmPAW7A1c0w7iWjMygvv9pqw+xmoqi8HHV4Z8rWr0eurkbSus6KNZ1GxV39w7ilTyifbzrI/1bncMu7Gxkc48/fR8SQ0NGrtUO87KjdFZsJU7w/ADWnLA3EV+XOE0o7L71SaB/lhSHSjNpT35phCwQCgaCNcbkK5O8A5sqy/IYkSQOATyVJ6inLcoMN7yRJehB4ECAwMJDVq1e3aFDdi4uwSTUUFZZg0FUja7TsqSgjpLqa9V9/jc1Fd1mPADcJSbUAACAASURBVP7dX82KfC0/5h7nhlnHSfBXMy5aS5hny6/aKysra/F7f0kEAgGgLQfjSQlToQXjziNOm4kqN5kKX5lKH5lKH7C3bxcRJ23+vrgoV/p9MZvNlJaWtnYYF4TNZmt3MbsCzXlfLBbLBf3eNYfYOgR0rPc81HGsPvcDowFkWd4oSZIB8AOO1W8ky/IcYA4o1g8tvpz51AJsGRYMWhNdu3bm2JY0fLp3ByDexwfPK3g59fkwCvintYaP1+/n/XW5/Os3C9f0DOKJETHEBLbcsub2uJS9sc2Efn8xXgfr2Uw4Ml/t2WaiPd4XV+BKvy8ZGRltwkbh3//+N59//jlqtRqVSsV7773HbbfdRlpaWgOnd4B+/fqxadOmM/Y1btw49u/fT1lZGcePH3due/O///2PO++8s8k+ly5dyp49e5g6dWqTfebl5ZGamnraNkKCOprT+sFgMDi3SjofmkNsbQGiJUkKRxFZtwN3NmpzEBgOzJUkqRtgAFp2r4PzQaPHrrJgrwK/oEAAipAxQrvdtqe5cddreGx4NHcP7MyH63L5aEMeP+8+wvVxwfzt6mgi/N1bO8Q2wTltJn4roGxdQ5sJfaQX+k6eSFphMyEQtGU2btzI999/z9atW9Hr9Zw4caLB1jaN+fXXX8/a3+LFiwFFKE+fPp3vv//+nDGMGTOGMWPGXFjggjbDJX/Ky7JcA0wCfgEyUFYd7pYk6SVJkmr/ZzwJPCBJ0g7gC2CC3BbcVDUGZJUVuUoiMCQUgFPFxWhDQq5o+4eLwWzU8veRXVg3ZSgPDY5k+Z6jXP3mGp78agcHCytaO7w2R63NhOfwTvg/GEfICwPwu78nHoNDwS5TuiqfE+//waEXN3L8gz8oWXUQ64ESZJv93J0LBILLyuHDh/Hz80OvV+ox/fz8CA4Odp6vrKzkmmuu4f333wegQ4cOQF3W8eabb6Zr166MHz+e8/nqmzFjBr179yY2Npa9e/cCMHfuXCZNmgTA0aNHGTduHPHx8cTHx/Pbb781uD43N5devXqxZcsW5s6dy4033sjo0aOJjo5mypQpznbLli1jwIAB9O7dm1tuucW59c/UqVPp3r07cXFxPPXUUwAsXLiQnj17Eh8fz+DBgy/qfXRlmqVmy+GZ9WOjY/+s93gPkNwcYzUrGj2oLMjVOry8lVV3xcXF6KIiXcL+4WLwdtMx9Zqu3H9VOO+uyeHT3w/w7fZD3JLYkceGRRHsZWztENskkrbOZgIa20wUU/LLAeAAkk6NPtzTsam2F9oObkhtzOFfIGgtjrz8MtaMvc3ap75bV4KeffasbUaOHMlLL71ETEwMV199NbfddhtDhgwBlJq522+/nbvvvpu77777tGu3bdvG7t27CQ4OJjk5mQ0bNnDVVVeddTw/Pz+2bt3K//73P6ZPn84HH3zQ4Pzjjz/OkCFDWLx4MTabjbKyMk6dOgXAvn37uP3225k7dy7x8fHs3r2b7du3s23bNvR6PV26dOGxxx7DaDQybdo0fv31V9zc3Hj11Vd58803efTRR1m8eDF79+5FkiSKiooAeOmll/jll18ICQlxHhOcPy7tII/GiKSqRKo2oNfrUckyZRXl6COjqNj4O7LNhqRun/U1LY2/h57nU7vz4OAIZq3K5ovNB/k6/U/uSOrIo0OjCPA0tHaIbZombSb2F9etdty3HwDJoEEfYcbgmHYUHl8CweXH3d2d9PR01q1bx6pVq7jtttt45RXFu3vs2LFMmTKF8ePHN3ltUlISoaHKzElCQgJ5eXnnFFs33ngjAH369OGbb7457fzKlSuZN28eAGq1GrPZzKlTpzh+/Dhjx47lm2++obuj/hhg+PDhmM1mALp3786BAwcoKipiz549JCcreZCqqioGDBiA2WzGYDBw//33k5qaSmpqKgDJyclMmDCBW2+91Rmf4PxxcbGlRy1ZUNUoFg86tYpKayX6HpHIVVVU//knurCwVg6ybRPoaeClsT15cHAEM1dm89mmgyzYks9d/cN4OCUSP3dhg3A+qN11mGL9McUqNhO2EqvT2d6aW4xlTyEAKnet02BVH+mFxld4fAlch3NloFoStVpNSkoKKSkpxMbG8sknnwCKCPn555+58847m/xdrJ16rO2jpqbmnGPVXnO+7Wsxm8106tSJ9evXNxBbTcUgyzIjRozgiy++OK2fzZs3s2LFChYtWsTMmTNZuXIl7777Lps2beKHH36gT58+pKen4+vre96xuTquXZmrMaBWWdDUaJFlGZNeR5VdRhcZASCmEi+AUG8Tr9wUx8onh3BdXAc+2rCfwa+t4tWf91JUceZCUkHTqD31mHoF4HNzDB2m9CVoSl+8b47GEO2NNa+EosXZHJ2expH/bObkl/soTztKzSlLa4ctEFyR7Nu3j6ysLOfz7du3E+b4Q/yll17C29ubRx999LLFM3z4cGbPng0odgbFxcUA6HQ6Fi9ezLx58/j888/P2kf//v3ZsGED2dmKiXd5eTmZmZmUlZVRXFzMtddey1tvvcWOHTsAyMnJoV+/frz00kv4+/uTn59/tu4FjXD5zJZGqkRj02G1WXF3d+dUaRmyw1/Lmp2Dx7BhrRxk+yLM1403b01gYkoUb6/IUuq6Nh7g/qvCuX9QOJ4G1zGKbU40PgY0PkG4JQYhyzI1JyqVKcfcIiyZp6jYprioqH0MGCK9lNWOEV6oPa8Qky+BoBUpKyvjscceo6ioCI1GQ1RUFHPmzHGuInz77be57777mDJlCq+99lqLx/P222/z4IMP8uGHH6JWq5k9e7azKN/NzY3vv/+eESNG4O5+5tXi/v7+zJ07lzvuuAOr1QrAtGnT8PDwYOzYsVgsFmRZ5s033wRg8uTJZGVlIcsyw4cPJz4+vsVf55WE1BYWBTZFYmKinJaW1rKDbP+cD7/cQnnRtdz5Ri/WfPMjf2Tn8tBd4ym7/0Hc+iUR/OqrLRvDFc7eIyW8tTyTX3YfxWzU8uDgCCYM7Iyb/sw6/0r3DWpuZLtMzbEKZcoxpxhrbjGyRZl60PgbnVOO+ggzareLF7vivrRNrvT7kpGRQbdu3Vo7jAuiOf2cBM1Hc96Xpv5fSpKULstyYlPtXT6zpZcqschqyizl+Pj7Q3YuRw/9iU9kpLB/aAa6Bnny3l2J7DpUzJvLM3n9l318uH4/Dw+J4K7+nTHqxAKES0VSSWiD3NAGueGRHKIYrBaUOTNfFVuPUv77YcBhsFpb8xVhRmVw7Y8AgUAguBy49ietxoBeUjyiSsrL8A9U0rAnjhyhQ1Qkp75aiGy3I6lcu7StOegZYuajCX3ZevAUby3P5OUf9/L+uv1MTInkjqROGLRCdDUXkkpCF+qBLtQDjyH1DFYd7vZlm45QtqFAcbcPcXdMO3qh6+yJSohfgUAgaHZcXGzpMVIJQElZOV2COwNw8mQhushI5MpKqgsOowsNacUgryx6d/Lm0/v7sSm3kDeWZ/Lid3uYszaXScOiuKVPR3QaIWybm1qDVX2YJwzrhFxtx3qwRBFfucWUrjtE6Zo/Qe1wt48Q7vYCgUDQnLi42DJikhSxVVFZiafDh6SkuBh90kAAqnKyhdhqAfpF+PLlg/3ZkF3IG8v38dziXcxencPjw6PxtbfNOsIrBUmrwhDphSHSCwB7lY2qPEV8WXKKKF2VT+nKfNCo0Id5OGu+EOb2AoFAcFG4uNgy4IYyjVheXqkYmyJTVlGBPsJh/5Cdg7vDKVjQvEiSxFXRfiRH+bJ633HeXJ7JlEU7CTRJPON9iOvjg1EL9/QWR6VTY4jxxhDjjZlG7va5xZQsU9ztI9QqThzYhT5CWe2oDXYX7vYCgUBwHri42NLjRm1mS1n6qlersVRUovbyQu3vh9XhQSJoOSRJYmjXAFK6+LNsz1H+3+Kt/O3L7cxalc0TI2IY3SMIlfhSv2yc5m5fXo01t5jcdbvRnbSc5m6vjzRjiPRCE2AS4ksgEAiawLXFltaIpyOzZXGILZNBz6nyMux2G4boGCz7mncfLsGZkSSJUT2C0B4zUuHbhbeWZzJx/la6dfDk7yNiuLpbgHBLbwXUblpMsX6cKJTpmZKIraRK8ffKrnO3LwZUblqnv5c+0ozGzyjul+CKwN3d3blJc3OQl5dHamoqu3btIi0tjXnz5vHOO+80W/+Ctodriy2Nvp7YqgaUX6qTxSWUF53C0LMnhR99hN1qRaUX285cLlSSRGpcMNf07MDSHYf4769ZPDAvjfhQM0+MiGFIjL/4Em9F1J46TAkBmBICAKg5ZXHu6WjNKaJy5wkAVJ66BgarGh+xX6ZA0JjExEQSE5u0ZhJcQbj2UiONAb1K2eKkyqKILS8vb2SNlqKjRzH07AE1NVj37WvNKF0WtUpiXK9Qfv37EF69KZYTZVVM+HgLN7+7kQ3ZJ2irhryuhsbbgFtiID63dSHomSQCn0rEa1wU+s6eWDJPcWpRFkde28Lh17ZwclEmFduOYSuxtnbYAsEFU2sge/PNN9O1a1fGjx/v/ByaOnUq3bt3Jy4ujueeew6ACRMmsGjRIuf1TTm6r1692rnZ8wsvvMB9991HSkoKERERItt1BeHymS2tYzVildUGgK+/H+zL5FjBIYJ69ADAsns3xri4VgvT1dGqVdzWtxPjeoXyZVo+s1ZmM/6DTSSF+/D3ETH0jxCbobYVJElC62dE62fEvV8HZWuhoxWOlY7FVO4qpCLtKFDf3d6xtdAluNsLXIN1X2VyIr/5pvMA/Dq6M+jWmPNuv23bNnbv3k1wcDDJycls2LCBbt26sXjxYvbu3YskSZe0b+DevXtZtWoVpaWldOnShUceeQStVvxutHdcXGwZUVONXbJRY1XWtft3CAbgxNEjaIaNQO3tTeUfu/C+ozUDFQDoNCru6h/GLX1CWbD5ILNW53D7nN8ZGOnLEyNi6NvZp7VDFDRCkurc7d1r3e0PlzunHCu2Hqtztw9yU4RXpBf6cDMqo2t/PAnaJklJSYSGhgKQkJBAXl4e/fv3x2AwcP/995OamsqQS1jBft1116HX69Hr9QQEBHD06FHneIL2i2t/mqk1IKmwqauwWZVUsF9AIACnTp5EkiRMiX0oX79eOMm3IQxaNROSw7k9qRPzNx1k9upsbnl3I4Oi/fjb1TH0CfNu7RAFZ0BSSehC3NGFuOMxuJ67fa6yr2Njd3u9ww9MuNsLgAvKQLUU+nr1u2q1mpqaGjQaDZs3b2bFihUsWrSIt99+mzVr1qDRaLDblT/k7XY7VVVVF9W/oP3j2mILsKt0yOoq7I7fAU9PTwCKS0oA8Bg5itLlv1K5fTum3r1bK0xBExi0au6/Kpw7kjry2e8HeHdNLjfN/o0hMf48MSKGhI5erR2i4Bw0cLcfCnKNnaqDJVgcBfdl6w9RtuZPUDnc7SOFu72g7VFWVkZFRQXXXnstycnJhIeHA9C5c2fS09O59dZbWbp0KdXV1a0cqaC1EGJLpUVWV0G1srqt1ti0olKp5XIfmoKk01Hy889CbLVRTDoNDw6OZHy/MOZtPMB7a3O4YdYGhncN4IkRMfQMMbd2iILzRNKoFOuICC8YEVbnbp+r1HzVudtL6Dt5KsIrygtdqDuSWogvQetQWlrK2LFjsVgsyLLMyy+/DMADDzzA2LFjiY+PZ/To0bi5ubVypILWwuXFlk2tA3UVVNd9UBs0GiyV5QCo3d1xu+oqSpctJ3DqVDGV2IZx02t4JCWSuwaE8clvecxZm0vqjPWM6B7I366OpkewEF3tjSbd7fcXO60mSpYfgOUHkHQqdJ3NGByZL+FuL2hOaj22UlJSSElJcR6fOXOm8/HmzZudj0tLSwEIDAzk999/dx5/9dVXASXjtWvXrtP6fOGFFxqMW9tG0P5xebFlV+lQqauQquvqQUwGAydLS6iprkaj1eI5ehRlK1dSuWMHpl69WjFawfngrtfw6NAo7hoQxsfr8/hgfS7XvXOUa3oG8X9XR9M1yLO1QxRcJCqDBmM3X4zd6tztq/YXY3EU3Bf/dAoAyaBGH64IL0OUcLcXCAStixBbKh0qVRWqqrovYA8PdwqLiigrPIFXUAfchw5F0mop/fkXIbbaEZ4GLf93dTQTkjvz4fr9fLR+Pz/tOsJ1cR342/BoogM9WjtEwSWidtNi7OmHsacfgNPd3ppTjCW3CEvGSYe7vcbhbC/c7QUCweVHiC2VFo3aisamo8Zeg0alwcvbG/nwUU4dO4JXUAfUHh64DRpEyU8/ETBlMpJarIpqT5iNWv4+Iob7kjvzwbr9fLxhPz/+cZgx8cE8PjyaSP/TjQYF7ZMzutvnOtzt/6jnbh/hKLaPFO72AoGgZRFiS6VDq7Kitekpry7HrDfj6x8A7OVYwSHC45RMlnnMGMpWrqT8999xT05u3aAFF4WXScdTo7pw31XhzFmbyye/5fHdjgJuSAjhseHRhPuJ4tUrDY23AU2i4nAvyzK2QotzytGSXUTF9uMAqL31TuFliDCjNovtuQQCQfMhxJZKh1ayNBBb/sGKsWnhsWPOdu5DU1B5elL87bdCbLVzfNx0TL2mK38dpIiueRvz+HZHAeN6hfD4sGg6+ZpaO0RBCyBJEho/I+713e2PVShTjjlFVO5u5G5fm/mKMKN217Vy9AKBoD0jxJZKi16yoJG1lFrKwB18fJTi21MnTzrbqfR6PK+5huKlS7GVlaN2F1mQ9o6fu55nr+3GXweF896aXD77/QBLth3i5j6hPDo0io4+QnRdyUiShDbQDW2gG+4Dg093t992nPJNRwDQBpkcwksRX8LdXiAQXAgu72NgV+kwSBUAlJYry3trjU1rl+/WYh47FrmyktLlyy9vkIIWJcDDwPOp3Vk7ZSh/6R/GN1sPMeyN1Ty7+A8OFVW2dniCy0Stu73H4FD87u1J8L/64z8xHs9RnVG56yjbdITCT/dQ8NJGjs7cRtGP+7HsO4ndsa+q4MpFrVaTkJBAfHw8vXv35rfffgMgLy8PSZL4xz/+4Wx74sQJfHx8mDRpEqDYOUyfPr1V4ha0HVz+zzO7SocRRWyVlStfrHq9HjVQXmlp0NbYKwFtp04Uf/stXuNuuNyhClqYQE8DL4zpwUNDIvjfqhwWbDnIorQ/ua1vRx4dGkWQWRRRuxKSWqUYp3byhKEdHe72pUrNV24RZRsOUba2kbt9hBf6MA8krVhEcyVhNBrZvn07AL/88gvPPPMMa9asASA8PJwffviBadOmAbBw4UK6devWarEK2iZCbKl0mGQlo1VaUe48rtdqsFQ03JNKkiTMY8dwYuYsqgsK0DpquwRXFh3MRv7fDT15OCWSmSuz+WLzQb5My+fOpE5MTIkkwFOILldEcbc3o48wAw53+wMlToPV0tVNuNtHmtGFeiBpXH4S4YqhpKQEb++6/VdNJhPdunUjLS2NxMREvvzyS8aNG0dhYWErRiloawixpdLiJiv7IFZU1GWy3I1GTpSWYK2oQG+qq90xjxnDiRkzKf7ue/weevCyxyu4fIR4GfnPjbFMdIiuT38/wBebD/KX/mE8PCQSfw+xYs2VUenUGKK9MUQrX7x2Sw3WvBKs2Urmq+TXA7AcJK0KXbjD3T7CC22IcLe/WFbNncOxA7nN2mdAWARDJ5z9s7yyspKEhAQsFguHDx9m5cqVDc7ffvvtLFiwgMDAQNRqNR06dBBiS9AAIbZUOtxtitiqrLQ6j5vNnhw7dYqSE8fw79TZeVzXsSPGxD4Uf/stvg8+IIwRXYCOPiZevTmOiUMjmbEym4837Gf+pgPcM6AzDw6OwNddiC6Bw92+qw/Grj5AY3f7Yop/ygMautvrI73QBgp3+7ZO/WnEjRs3cvfddzfYSmf06NE8//zzBAYGctttt2G321srVEEbpVnEliRJo4G3ATXwgSzLrzTR5lbgBUAGdsiyfGdzjH2p2FU6PO3FAFgqq5zHfXz9IP8QhYcLGogtUArljzz/Tyy7dmGMjb2c4QpakTBfN6bfEs+jQ6OYsSKL99fl8unvB7jbIbp83IQ9gKCO09ztS+vc7a05irs9gMqkcVpM6CO90PgLd/szca4M1OVgwIABnDhxguPHjzuP6XQ6+vTpwxtvvMGePXv46quvWjFCQVvkksWWJElqYBYwAvgT2CJJ0lJZlvfUaxMNPAMky7J8SpKkgEsdt7mwq7QYJaUw3mKpE1v+QR2AHRw7fJiuja7xHD2ao/9vGsVLvhViywUJ93PjzdsSmDg0indWZPHe2hzmbczjrv5hPDA4Aj+R6RI0gdpDhyk+AFO8w92+yOIUXtac4jp3ew8d+kgzBuFu3ybZu3cvNpsNX19fKioqnMeffPJJhgwZgo+PTytGJ2irNEdmKwnIlmU5F0CSpAXAWGBPvTYPALNkWT4FIMvysdN6aSVqTU0Bqi11S7iDQkMBOHH89FDVHh54XD2ckh9+IPDpKUg6kdFwRaIC3Hnnjl48PjyKGSuzeX9dLvM2HuAv/Tvx4GBR0yU4OxovA5o+Btz6ONztT1qcU47W7CIqa93tvRzu9lHC3b61qK3ZApBlmU8++QR1o23bevToQY8ePVojPEE7QJJl+dI6kKSbgdGyLP/V8fwuoJ8sy5PqtVkCZALJKFONL8iy/HMTfT0IPAgQGBjYZ8GCBZcU2/ngs/9buu2fz5xjC/gzaiujEhMBqK6uZsOGDXjbq4gfNvK063R/7MJ71iyKHn4Ya0J8i8fpSpSVleHu3v72KzxcZue73Go2FtSgVcHQjhquCdfiZbgyVqK11/vSLpFBWw6mQgnjSQnjSVBXK1OLVSaZSl+ZSh+o9JEprr6y74vZbCYqKqq1w7ggbDbbaWJM0Po0533Jzs6muLi4wbGhQ4emy7Kc2FT7y1UgrwGigRQgFFgrSVKsLMtF9RvJsjwHmAOQmJgop6SktHhgGUdWopGsyMgYtG7UjinLMhs3rEeWZZqKQ77qKrIWLKBjTg6hf/u/Fo/TlVi9enWT73l74A5g/4lyZq7MZsn2Q6w+ZOeOpE48khJJYDu3jGjP96W9I9tlqo+UO6cc9fuLMecrmXiruye+cSGKz1e4GZVJ28rRNi8ZGRl4eHi0dhgXRGlpabuL2RVozvtiMBjo1avXebdvDrF1COhY73mo41h9/gQ2ybJcDeyXJCkTRXxtaYbxLwm7SoskgV1Tjb2qLssnSRJGnY6yUyXYaqpRaxp+gEkaDebUVE5+/jm2oiLUXl6XO3RBGyXcz403bo3n8eFRzFqlWEZ8vvkgd/TtyMMpkXQwG1s7REE7Q1JJ6ILd0QW74zEoFNkmU3WoFGtuMUe27Kd8yxHKfisACbTB7orwivRC39kTld7lF50LBK1Oc8xvbAGiJUkKlyRJB9wOLG3UZglKVgtJkvyAGKB5zVIuErtKqbeyq6uxVzU85+/ri01v4viBvCavNd8wFqqrKfnppxaOUtAeCfN147Wb41n9VAo39gph/qaDDHltNc8v2UWB2AZIcAlIasU41TOlIwV97QT/awD+D8XhObwTkk5F2YYCCj/eTcGLv3Psf9sp/iUPS/Yp5GqxtZBA0Bpc8p88sizXSJI0CfgFpR7rI1mWd0uS9BKQJsvyUse5kZIk7QFswGRZltuE41ut2JK1NcjVDZdbh0dGknf4CHl79xAUGX3atfquXdHHxFC85Fu877jjssQraH909DHxyk1xPDo0iv+tVrYBWrDlILckdmRiSiSh3mLDa8GlIWlUindXuBnPq+u52+c63O3X5FO6Kh/UEvowT8VmIspLuNsLBJeJZskvy7L8I/Bjo2P/rPdYBv7u+GlT1IotSWNDqm5YOBfZtRur1m8gNyuL/k1cq2zfM5Zjr7+Odf9+9OHhlyFiQXulo4+J/9wYy6RhUcxenc1XW/5kYVo+N/cJZWJKFB19hOgSNA9ndLfPKcKaW0zJioPw60HF3b6zsrWQIdILbbA7klp4fAkEzY3LT+bbVUotlkpTg1TVUGwFBQUhAfkHD5y2bU8tnqmpHHvjDYqXLiXg/0ShvODchHgZmXZDLBNTonh3TQ4LNuezMO1PbuwdwqSh0XTyFaJL0Lw0dre3V1QrWa9cxeG+5Oc8SgBJr3bs/6js66gNchPu9gJBM+Dy+ePazJZaa0Ndo8Uu122zoNFoCAzwx2JwY+tP3zV5vTYwALeBAyn5dimy2KJBcAEEexl5aWxP1k4Zyl/6h7FkewFD31jNUwt3kHei/NwdCAQXicqkuNt7jYkk6Ik+dHiuHz53dMUU70/N8UqKf8jl2DvbODztdwo/20PZxgKqj1VwqVZB7RW1Wk1CQgI9evQgPj6eN954w7klT1paGo8//vhZr587dy6TJk06a5vGvPzyyxcd74QJEwgPDychIYHevXuzcePG8762fqzvvvsu8+bNu+g4zoe8vDyMRiMJCQnOn6qqqnNfeJ7MnTuXgoIC5/O//vWv7Nmz5yxXtAwis1UrtjQ2tHY9lTWVuGndnOeHjxjJ/PnzWbtqBVGJSfiHnT5VaB47loLJk6lIS8MtKemyxS64MggyG3hhTA8mpkTy7ppc5m86wOJthxibEMykoVFE+F+5HkqCtoHibu+PKd4fgJoiq3PK0ZpTROUupcRW5aFVphwdmS+1j8ElthaqvzfisWPHuPPOOykpKeHFF18kMTGRxMQmrZUuiZdffplnn332oq9//fXXufnmm1m2bBkPPfQQO3fuvOA+Hn744QtqX1NTg0Zz4bIiMjLS+f42N3PnzqVnz54EBwcD8MEHH7TIOOdCZLYc04hajQ2tTU95dcOMQnR0NJHh4VT6dmDhjDepqqw4rQ+Pq4ejMpko/vbbyxKz4MokwNPAP6/vzrqnh3LvwM78+Mdhrn5zDY9/sY19R0pbOzyBC6Hx0uPWJxCfW2IIerovQZMT8b4xGn2EF9acIk59k8WRLfPKHgAAIABJREFU19M48uoWTi7MpHzrUWqKra0d9mUhICCAOXPmMHPmTGRZZvXq1aSmpgKwefNmBgwYwFVXXcXAgQPZt2+f87r8/HxSUlKIjo7mxRdfdB7/7LPPSEpKIiEhgYceegibzcbUqVOdrvXjx48/YzubzcaECRPo2bMnsbGxvPXWW6fFO3jwYLKzs8/YB8DHH39MTEwMSUlJbNiwwXntCy+8wPTp0wHYsmULcXFxJCQkMHnyZHr27AkoYmbMmDEMGzaM4cOHU15ezn333UdSUhK9evXiW8f3os1mY/LkyfTt25e4uDjee++9s77P9Y16Fy1axIQJEwAla/f4448zcOBAIiIiWLRokbPdq6++SmxsLPHx8UydOpVFixaRlpbG+PHjSUhIoLKykpSUFNLS0gD44osviI2NpWfPnjz99NMNxn7uueeIj4+nf//+HD169Kyxng8is+XIbOkcYqusuowAGm7dePOttzJn9myO2e0s+d/b3PL3qQ3+mlMZjXiMHk3pTz9jf+45VE3UdgkE50uAh4F/pHbn4ZRI3l+by2e/H2DpjgJGdA9k0tAo4jsKTzfB5UOSJDS+RjS+RtySgpBlmZrjlQ6D1SIsGYVUpCtfRho/o+LxVZv5cm/ercyKvsuhqqB5p9h1wW54XR95QddERERgs9k4dqzhdm5du3Zl3bp1VFZWsmnTJp599lm+/vprQBFiu3btwmQy0bdvX6677jrc3Nz48ssv2bBhA1qtlokTJzJ//nxeeeUVZs6c6cz2ZGRkNNmuR48eHDp0iF27dgFQVFREY7777jtiY2PP2MeIESP417/+RXp6OmazmaFDhzZp1nnvvffy/vvvM2DAAKZOndrg3NatW9m5cyc+Pj48++yzDBs2jI8++oiioiKSkpK4+uqrmT9/PmazmS1btmC1WklOTmbkyJFIkkROTo5zO6Tk5GRmzZp11vf/8OHDrF+/nr179zJmzBhuvvlmfvrpJ7799ls2bdqEyWTi5MmT+Pj4MHPmTKZPn05iYiKlpXV/tBYUFPD000+Tnp6Ot7c3I0eOZMmSJdxwww2Ul5fTv39//v3vfzNlyhTef/99/vGPf5zrv8VZEWLLIbb06hq0dj3l1tN/kY1GI/fcdx+zZ80k42Qp6775ksE33d6gjde4Gyj+5htKli3D64YbLkvsgisbP3c9z1zbjUdSIvl4Qx5zf8tj7J4NDIr2Y9LQKPpF+LZ2iAIXRJIktAEmtAEm3AcE13O3V6YcK7Yfp3zTEQA0gSbHhtpXprt9Y4qLi7nnnnvYt28farWa6upq57kRI0bg66v8zt54442sX78ejUZDeno6ffv2BZQ9GAMCAk7rd8WKFU22u/7668nNzeWxxx7juuuuY+TIuq3lJk+ezLRp0/D39+fDDz88Yx+bNm0iJSUFf39lCvm2224jMzOzwfhFRUWUlpYyYMAAAO68806+//77Bq+tdgPuZcuWsXTpUmdGzGKxcPDgQZYtW8bOnTudmaji4mKysrKIiYm54GnEG264AZVKRffu3Z1Zp19//ZV7770XkyPZca4Nwbds2dLgdY8fP561a9dyww03oNPpnNnKPn36sHz58vOO7UwIseUQWwZNDQClFU3/1eTl5cW99/+V9999l9Vp2wnoFE7Xvv2c542JiWjDOlH8zWIhtgTNipdJxxMjYnhgcASf/X6AD9blctuc30nq7MOjw6IYHO3nEnUzgrZJQ3f7EGSbTHVBmWNT7aLT3e0jHO724Rfubn+hGaiWIjc3F7VaTUBAABkZGc7jzz//PEOHDmXevHkUFhY22N6q8e+oJEnIssw999zDf/7zn7OOd7Z2O3bs4JdffuHdd9/lq6++4qOPPgLqarZqWbVqVZN9LFmy5Lxf95lwc6urc5Zlma+//pouXbqc9hpmzJjBqFGjGhzPy8trss/675fFYmlwTq+v24y9JRZtaLVa5/hqtZqamppL7tPla7ZklRokNUa1Q2yVn16TVUtQUBB33Hknsk7Pom++4cShfOc5SZLwGncjFZs3U3XwYIvHLXA93PUaHh4Sybopw3jh+u7kn6rgno82M2bmBn7edQS73TVXignaFpJaQtfRA8+UjvjfH6u42z9c626vpuy3Agrn7qbgxY3t0t3++PHjPPzww0yaNOk0AVVcXExISAig1DLVZ/ny5Zw8eZLKykqWLFlCcnIyw4cPZ9GiRc7pyJMnT3LgwAFA+cKvzYydqd2JEyew2+3cdNNNTJs2ja1bt54x7jP10a9fP9asWUNhYSHV1dUsXLjwtGu9vLzw8PBg06ZNACxYsOCM44waNYoZM2Y4RdC2bducx2fPnu18TZmZmZSXn3lKODAwkIyMDOx2O4sX/3/27js8yipt4PDvTJ/03hNSKEloAULoTTqE3hUUO66d1VVX3PWzLa5l7RUEBQm9iSAthCa9SZXeW0ggpM1kyvv9MUkIGJBIkknCuS+5nMy85Zn+zCnPmXfT7Yp07dqVSZMmkZeXV3z/ANzd3a/rPiySlJTE6tWruXTpEjabjZSUFDp06PCn5/mr7vqWLQA0BlxVjqmmuXk3T7YAatetS+8e3Vn0y1Imfvklz77yKobCrN6zfz/SP/mEK/PmyZpbUoUx6tSMbhPFvS1qMXf7ab5cfYQxU7dRN9CNJzvVpnfDYDTqu/53lFRFCI0KfaQn+khPPLqAYrFhPnG1uNuxZHV7XYQHhsJ1HXXhVae6fdFgdYvFgkajYdSoUYwd+8ca3f/4xz944IEHeOONN+jTp891tyUlJTFo0CBOnz7NyJEji2cwvvXWW3Tr1g273Y5Wq+Xzzz+nVq1aPPbYYzRq1IimTZvy448/lrqd0WjkwQcfLC5DcasWsvj4+FKP0bJlS15//XVatWqFl5dX8dipG02cOJFHH30UlUpFhw4d8PT0LHW71157jeeee45GjRpht9uJiopi0aJFPPLIIxw/fpymTZuiKAr+/v63bFUbP348ycnJ+Pv7k5iYSE5Ozk23BejRowc7d+4kMTERnU5Hr169eOeddxg9ejRjxozBaDSybNmy4u2Dg4MZP348nTp1QlEUevfuTb9+/W55jjshqmrdlMTERKVoxkBFSktLo+Pmh9jl8xDrtrbGc0QmIzsM/tP9ls6fy4adv+GrU/Pky6+iUjk+FE4++hjmw4epvWI5Qq3+k6NIpUlLS7uu+V26NavNzs+7z/FZ6mEOXcyhlq8LT3SIYWDTMHTl+GUln5eqqbo/L3Zzier2R7KwnM0BheLq9hebQlx8PEKnqjbd5dnZ2bi7uzs7jHKVk5NTPENw/PjxnDt3jo8//tjJUZVNeT4v+/fvJy4u7rrrhBDbFEUptQ5I1fjZ4GwaA+44pi3n5d3e9OXu/QdSJyiAjAIbU7/+ovh6r4EDsJ47R+6GjRUSqiTdSKNW0S8hlKXPteerkc3wMGh5ee5uOr63isnrj2GqJt0z0t1JpddgrOeDV69oAp9uQshrLfEdFYdr8yBsVwuw51uxpudhOZuL5VI+tuwC7AW2u7bAqrP8/PPPJCQk0KBBA9auXXvHs/PuNjLZAtDocRX5AJhMt1+5dsRjY/DXqjh64RKLZjv6ud06d0bt6UnW3LkVEqok3YxKJejRIIiFT7Vh8oPNCfU28vpP+2j7bipfrT5CjvnOB3lKUkVTuWgx1r9W3V7tqUftY0DlogGrHVuWGevFPCzncrFm5GPLKcBukclXRRs2bBg7d+5kz549/Pzzz8Wz+KTbI5MtAI0RnXCM1Sow3f4Xkkql4uHn/46rxcTWPXvZsGY1Kp0Oj+RksleswJaVVVERS9JNCSHoWC+AWWNaM+OxlsQFezB+yQHajE/lw+UHycwtv6UwJKmiCZVA7aJF421AG+SKNsgVtbcBlUGDvcCO7YoZ64U8LOcLk6/cAhSrXSZfUpUiky0AjR6t4pgVUZBv+ZONr2dwcWX0355Ea8pl6cpU9u3dg9eggSgFBWT9/HNFRCtJt61FtC9THm7BgifbkBTlwycrD9F6/EpeX7iX05dvPRlEkqoioVGhdtWi8TGgDXK5lnzp1dgLbNgum7Gcz8VyPg9rpglbrgXFKtetlZxLJlsAGgM6xTE11FJQ9jelf2g4QwcNRmXOZ9as2aS7uaGPjSVrjuxKlKqGxuFefHt/Isufb09yoxCmbjxBh/fSeG76Dg6cv+rs8CTpLxFClEi+jGiDXNEEuqD20qPSqbCbrNgumwqTr1ysl03Y8iwoNpl8SZVLJlsAGj2awmTLZv5rb8I6zVvQuUUiFJj44fvJ2JOTMe3di6nE2liS5Gx1At15f0hj1vyjE6NbR7Js3wV6fLSWBydtZvOxTNn1IlVrQghUWjVqNx0aXyPaYFc0AS6oPfWgUWHPs2LLNGE5l4vlQi7WKybs+TL5kiqeTLYAtEZUtnxsaiu2O1hLtXX/ITQO9sdqNjHn4gXyPDzkQHmpSgrxMvJacjy/vnwPf+9al12nsxj69QYGffkry/ddkAVSpRpBCIFKp0btrkPrZ0QbUiL5Uquw51qxZpRMvszY860opbz+58+fjxCCAwcOOOGeSNWdTLYANHqwmFC0VuxlG7J1HSEEyWOeIVxYyTflk9atK+mLl6AUyAHJUtXk5aLj6c51WP/SPbzRrz4Xs808+sNWun+0htnbTlMgx7pINUipyZe/EbWHDlQCe24B1ox8LGdzsFzMw5plxm5yJF8pKSm0bduWlJQUZ98NqRqSyRaAxgBWM2jtUHBnRfM0Wi3D/v4KPlnpZKvVrKkfT9aqVeUUqCRVDKNOzf2tIkl7oSMfD09ArRK8MGsXHd5bxYS1R8mVZSOkGkgIgUqvQe2hR+vvgjbYDY2fEZW7Y81ce3YB1kv5XD58nnVr1vLNx18yPWU6iqJgt9v529/+RmxsLF27dqVXr17Fiyzv2LGDDh060KxZM7p37865c+eceTelKkAu1wOOli2rCaFVEJY7r/ru6uXNsGdf4Pv3x3MxKIxflq9g+A2Lb0pSVVRUILVv4xDSDqbzVdoR3vp5P5+mHqZDCDRMNOPrpv/zA0lSBViyZAnnz58v12MGBQXRs2dPwFFmQhg0qAyOr0bFrqAU2Fg0ZQ7dOnUlJqAWPh5ebPplHcfPneTY4aPs2bGb9CuXiI+P56GHHsJisfDiiy+yaNEi/P39mTFjBq+++mrxAtHS3UkmWwAaI1jNqHQg8jQoinLHy0IERtem78j7mZMyjQO+QWxYsYJWXbqUU8CSVLGEEHSqF0CnegFsP3mZr9KOsHDfBZaOT2VwszAebhtFtL+bs8OUpApVlHzNXDCbZ599Fm2wK8NHDGfm4jlYzRYGdu+HPcOEr3CnQ5v22POt7N+zj/3799O1a1cAbDYbwcHBTr4nkrPJZAsKW7byUesFmmwdZpsZg8Zwx4eNbd2eNrt3sfbgEZatXUtAVBQxMTHlELAkVZ6mEd58c38i0xal8pvZj1nbTjNt80k6xwbyWPtomkd6V5s166TqragFqjJlZmaSmprK7t27EUJgs9kQQjBgwADUXjrUPgYUsw3sCvY8C9ZL+cTXiWXtstWo9GqEXl1lFtSWnEe+AsAxZstWgEYn0Nr05Fpyy+3QHR59kujzpxHmfKanpJCenl5ux5akyhTipmL8oEasf+kenu5Um20nMhn69Qb6f76en3adxSqnz0s10OzZsxk1ahQnTpzg+PHjnDp1iqioKHx8fJg7bx7CoCajIIs1G9eh9tIT17wh6ZmX2PDrBmxXzOSfzmJX2tZr1e0tsrr93UgmW+Bo2QK0eoHOpifPUn6VtYVKRffh9+N7eB9Wk4mpU6aQlycrd0vVl7+7nrHd6vHry515q38DrpqsPJ2ygw7vpTFx3TG5BqNUo6SkpDBgwIDrrhs0aBDnz58nLCyM+Ph4Ro4cSdOmTfHy8cbo5coPU6cw7r3Xad6rLUm92rJx1+Zr1e0vyOr2dyPZjQigNQKg06nQ2g3kWsuvZQvAt3sP2vz3PVLdjpClVjN71ixGjhqFSiVzXan6MurUjGxZi3uTIlix/wIT1h7jzUX7+GjFQe5tEcHo1pEEexqdHaYk3ZFVpcwmf+aZZwDIycnBzc2NjIwMkpKSaNiwIQCNGjVizZo11+2jKIpjzUazDcVsw26yQp4FG4BGVdzlqNKrEWr53VDTyGQLilu29DoVWpuGbHN2uR5e6HSE9OtH0szprDcYOSoEqakr6dKla7meR5KcQaUSdKsfRLf6Qew8dYVv1x7l2zVHmbj2GH0ah/BIuyjqh3g6O0xJKnfJyclcuXKFgoICXnvtNYKCgm66rRACoVWDVg1uhcmXxZF82c027PlWyHUkX0KjQhjUCJ1MvmoKmWyBY8wWYNQJBCpy88u/m89r8GB8J0ykrX8gaZfTWbduPSEhocTHx5f7uSTJWRLCvfj83qacyszju/XHmLHlFPN2nKFNbV8eaRdNx7r+cjC9VGOkpaX95X2FEAidGnRq1O6lJF+5VsgpTL60KoRec23AvUq+h6obmS7DtWRL73gBZ+eWf7Kli4zEpWVLAn/dQvP6cajyc5gzZzYXL14s93NJkrOF+7jw7z712fBKZ17uGcuRi7k8OGkL3f63hpTNJ8kvsDk7REmqUkqvbu9Sorq95abV7aWqTyZbUJxsuegcD0duXn6FnMZ76BAsZ87Qol4jarvqsJvNTJ3yAyaTqULOJ0nO5mnUMqZDDGv+0YkPhzZGq1bxytzdtBq/kv/+coBzWRXzXpOk6s5R3V5dorq9a6nV7S3nZPJVHchkC4rHbBW+hsnLr5jkx61LF9Te3mTPmcOAsa/gn5/F1avZzJk1U04Flmo0nUbFwKZh/PxMW2Y81pKWUb58tfoIbd9dxVPTtrP95GVnhyhJVZpQCVQGDRpPPdoAF7QhhUsLuZWSfKXnYbtqxm62yu+WKqJcki0hRA8hxO9CiMNCiJdvsd0gIYQihEgsj/OWm8KWLbfCEWz5eeYKOY1Kp8Ozf3+yU1PRmswM//vLuFy5yKEjR9mwfn2FnFOSqhIhBC2ifflqVDNWv9iJh9pEsvpgOgO/+JV+n69nwc4zcvFrSboNpSZfvkZUrjqwK9iuFmBNz8dyNhfLpXxs2QXYC2wy+XKSO062hBBq4HOgJxAPjBBC/GHUtxDCHXgW2HSn5yx3WkeypdM46gOZTAUVdiqvIUPAauXKvHn4hIQx9KHH0ORksXzFCk6fPlVh55Wkqibcx4VXe8ez8ZXOvNGvPtn5Fp6dvpN2/03ls9RDZOZW3PtQkm5XRkYGCQkJJCQkEBQURGhoKAkJCXh5eZV5gtP8+fPZt2/fn25nNpsZNmwYtWvXpkWLFhw/fvxP9xEqgcqoQeOlRxvo6uh29DWgctWA1Y4ty4z1Yh6WczL5cobyaNlKAg4rinJUUZQCYDrQr5Tt3gTeBareAKXCli2dyhFaQX7FFWXUR0fh0rw5V2bNRrHbiWzchG6dOoLFzNTJk8nPl2NYpLuLq17D/a0iWTG2A5NGN6duoDvvLztIq/+s5KXZv3Hg/FVnhyjdxXx9fdm5cyc7d+5kzJgxPP/888V/l7VW4u0mWxMnTsTb25vDhw/z/PPP89JLL5U5bqFWoTJq0XgZ0AY5ki+1jwGV8Y/JlzUjH1tOAXaLTL4qSnmUfggFSjbJnAZalNxACNEUCFcU5WchxIs3O5AQ4jHgMYDAwMA7mlZ7u3Jycti07SwtgBNH9gItyEi/XKHnNjRsgOeWLWz45hsKYmPB1ZNgvYazFgtffvQ/mrRsdVdPj8/JyamU514qm8p4XgTwcAz0CDSy/ISFedtPMWPrKeJ8VHSL1NLYX43qLn5vlKamv188PT3Jzi7f2od/ldlsRqvVkp2dTU5ODhaLhdGjR7Np0yaCg4OZPn06RqORw4cP8+KLL5KRkYHRaOTTTz/l8uXLLFiwgLS0NN544w2mTJnCmjVrmDRpEhaLhejoaL755htcXFyYM2cOr7zyCtnZ2XTv3p0nn3ySq1evls/3gqbwnx1UVlBZFRSzFVHYyKAIsGuu/UMU/qsBbDZbub2WTCZTmd53FV5nSwihAj4ERv/ZtoqifAN8A5CYmKh07NixQmMDR52UFk0SYDPE1Q3l141g0LtQkee2t2rF4Tlzifj9d8LGjAGgfbt2fPP2/3HeXIAp6zI9+w+ssPNXdWlpaRX6+Et/TWU/L/cBV/IKSNl8ih82HOfj7SZq+bowqmUEQ5qF4+mirbRYqrKa/n7Zv38/7u7uABw8+CbZOfvL9fjubnHUrfvabW2r1+vR6/W4u7vj5ubGkSNHmDFjBgkJCQwdOpRly5YxcuRInn/+eb799lvq1KnDpk2bePHFF0lNTaVfv34kJyczePBgAMLCwnj66acBGDduHDNnzuTpp5/mwoULxMbGFt9vLy8vCgoK8PPzK9f7XkRRFLApjgH1JhvCbENtKWzhUt9Q3b4aL6qdnZ1d/JjeKYPBQJMmTW57+/JIts4A4SX+Diu8rog70ABIK8zKg4CFQoi+iqJsLYfz3zmNY0kRrXB0I9rMFduMqtLr8ezfn8xp07BeuoTGzw+VWs0DY1/kk/+8zaYdu4iMiiaucUKFxiFJVZ2Xi44nOsbwSLsolu49z6T1x3nr5/18sOwg/ZuEMKplJPEhHs4OU7pLRUVFkZDg+Jxu1qwZx48fd/SWbNrEkCFDirczm0ufdLVnzx7GjRvHlStXyMnJoXv37pUS942EEKARqDU6cL310kJCo3IUVpVLC5VJeSRbW4A6QogoHEnWcODeohsVRckCitNxIUQa8EKVSbSguPSD2p6PXWXDXgnjcr2GDyPz+++5MmsWfk88AYDRzZ37H3ucb7/9ltlz5vB0WBhevhXzS0aSqhOtWkVyoxCSG4Ww50wWUzacYN6OM6RsPkVSpA+jWtWiR4MgtPKDv0a73RaoyqLX64svq9Vq8vPzsdvteHp6snPnzj/df/To0cyfP5/GjRszefLk4m6p0NBQTp06RVhYGFarlaysLHx9fSvqbvzBny4tlFdiaSGt6lripZPJ183c8aOiKIoVeApYCuwHZiqKslcI8YYQou+dHr9SFA6Qx2pG0VZOsqWPisK1bVsup0xHsViKrw+uFUVy927YhOC7zz7FZq24wfqSVB01CPXk3cGN2PhKZ17tFcf5qyaeTtlBm/Gp/G/5QS5crXpzcKS7h4eHB7Vq1WLWrFmAI1HZtWsXAO7u7teNGcrOziY4OBiLxcKPP/5YfH3fvn35/vvvAZg9ezb33HOPU8fxllrdPsAFtace1CrsuVasGSbHTMcLuVivmLHnywKrJZVLCqooymJFUeoqihKjKMrbhdf9S1GUhaVs27FKtWoBqDUg1GA1gdYOlsrJzL3vuxfrxYtkr1x53fVN23WkQXQUVxXB1I8/kLNDJKkUXi46Hm0fTdoLHZk0ujnxIR58vPIQbcan8tS07Ww+linfO5JTTJgwgYkTJ9K4cWPq16/PggULABg+fDjvvfceTZo04ciRI7z55pu0aNGCNm3aEBsbW7z/ww8/TEZGBrVr1+bDDz9k/PjxzrorpSp9aSFjiaWFCuTSQjeQC1EX0RrBakZoFYRFXSmndGvfHm1YGJlTp+LRo8d1tw0c9QDnPnyfY1k5LP7ua3o/PKZSYpKk6kalEnSKDaBTbADHL+UydeMJZm49xaLfzhEb5M79rSLp3yQEF538uJPuzOuvv158OTIykj179hT//cILL1x32y+//PKH/du0aXNd6YcnnniCJwqHkZRkMBiKW8aqAyEEQq8BvQY1oNgVlAJHl6NitmHPLsCeDQgQWjXCUKLb8S6ZXSw7V4to9GDJR6UHtU2LxW75833ukFCr8b73XvK3bsN04MB1t6lUKkY//gRajYZth4+xcX71eeNJkrNE+rkyLjmeTf/swviBDRFC8M95u2nxzkreXLSP45dynR2iJNV4f7q0UMnq9sVLC9XsGl8y2SqiMYDVjFon0Nr05FnyKuW0XgMHIAwGLv847Q+3ubu7M2zECOx6IytXpbF71bJKiUmSqjujTs3wpAgWP9OW2WNa0bFeAN//epyO76dx/3ebWbr3PFabXBZIkirDzZcW0pRYWiivRi8tJNvVi2gMYDWh1avR2vTkWnLx1HtW+GnVXl549kkm66efCPj7WNReXtfdXrtOHVq1asWGDRtYPG0qWr2B2NbtKzwuSaoJhBAkRvqQGOnDxd5xpGw+xfQtJ3l8yjYCPfQMax7B8ObhhHgZnR2qdAuKotw13U13A6ESCKPGUc0eUGz24i5HxWzDZiqcGKYSCJ0alcFRakJoVFXidfBXkkDZslWkKNkyqNHaDORaKq+7wfu++1BMJq7MnVfq7V26dCEkOBhzSBSLvviE/evSKi02SaopAjwMPNulDmv/0Ylv708kLtiDT1MP0fbdVB75fgurDlzEdhcP4K2qDAYDGRkZNaqVQ7qeUKtQu2jReP9xaSHFasd2xYz1QtVYWkhRFDIyMjAYDGXaT7ZsFdHowWpCb9CgK2zZqiyG2FiMic24PG0aPg/cj1BfP0BfrVYzeMgQvv76K2y141n82YcoikJ8u06VFqMk1RQatYqu8YF0jQ/kVGYe07ecZMaW06zYv4VQLyMjksIZmhhOgEfZPkylihEWFsbp06dJT093dii3zWQylfnLWLo5xe6o84XV7ii2WvijSKiEo6K9RuWo96W6datXeT0vBoOBsLCwMu0jk60ihWO29AYdGrue3ILKHUjrM3IkZ557npxVq3Dv0uWPt/v4kJzchzlz5uAX15gln3+IYrdTv0PnSo1TkmqScB8XXuwey7Od67Ji/wV+3HSC95cd5KMVh+gaH8h9LWrROsYX1Z98iEsVR6vVEhUV5ewwyiQtLa1MS7lIt09RFKwZJsxHrjj+Hc3CnuOY0Kb21qOP9kJf2wu1IJ3KAAAgAElEQVRDtKejDlgJznxeZLJVRGsAczZGTx1qxUS2qXKTLfcuXdCGhpLx3aRSky2Ahg0bcuTIEXbu3ElkfAK/fPkRFpOJhO69KzVWSappdBoVvRoG06thMMcu5ZKy+SSztp5iyZ7zRPq6MCIpgsHNwvB10//5wSRJqjBCCEdtLz8jbi2CHcnXxTzMR7IwHblC/r4M8rZdAEDjZ0Qf44k+xgt9dMWPwb4VmWwV0RggJx0XowG4Sk5u5cxGLCI0GnweeIAL77xD/s6dGBNKXxexV69enDp1iktmM7WaJLLyuy/JzrxE2+H3V4mBg5JU3UX5ufLPXnGM7VqXX/acZ9qmk/xnyQE+WHaQHg2CuLdFBC2ifOT7TZKqACEE2kBXtIGuuLUOQbErWM7lYj56BfORLPJ2ppO76TxqLz20dF6cMtkqUjhmy8XomJWUl1/5S354DRpI+mefkTFpMmEff1TqNjqdjsGDBzNhwgQKomvT0MuHzfNnkZOZQbfHn0GtkU+pJJUHg1ZN/yah9G8SysEL2UzbdJI520+zcNdZov1cGdo8nIFNQwlwl2NzJKmqECqBLtQNXagb7u3CUGwKBWeyseda4MJup8UlZyMWKRyz5ebqAkB+XuUnWypXV7yHDSN7+XIKTp686XbBwcF069aNQ4cO4dG4Oa2H3se+NanMe/f/KMiv3BY5Sbob1A105/W+9dn8zy68P6Qxfm56xi85QKv/pPLYD1tJPXBB1u2SpCpIqAX6CA+McZW3kHdpZLJVpLD0g9HFUeE2L9/slDC8R44EtZrM73+45XZJSUnUrVuX5cuXE9m6I93GPMPJPbuYNu4FLp8/W0nRStLdxahTM7hZGDPHtGLF2A480jaK7Scv89DkrbR9dxXvL/2dkxnyB48kSdeTyVaR4qKmjm44s6nil+spjTYwAM/kZK7MnYv18uWbbieEoF+/fri4uDB79mzqtenIoH++Qe6Vy/z4z+c5vmt7JUYtSXef2gFuvNIrjg2vdOarkc2IC3bni7TDtH9vFfdN2MiCnWcwWWzODlOSpCpAJltFCsds6QyOGlcFTkq2AHweHI2Sn8+VGTNuuZ2rqysDBw4kIyODJUuWUKthAve98z/cff2Z+5/X2fLTXFkIUJIqmFatokeDICY9mMT6l+/h713rciIjj2en76TFOyt5feFe9p+76uwwJUlyIplsFdEYwFaAVueYYVRgct4vUkPduri2bUvm1B+xm2/dnRkVFUW7du3YsWMHe/bswSswiBFvvkedpFasmfodiz56F3OeXHxXkipDsKeRpzvXYc2LnfjxkRa0r+vPtE0n6fnxWvp9to5pm06S7cQfcpIkOYdMtopoHTOKtBpHkmU1O7f53/fhh7BdukTWvNKX8CmpY8eOhIWF8dNPP3H58mV0BiPJz79Mu3tHc2jzr0x5+VkuHD1cCVFLkgSgUgna1Pbj0xFN2PTPzvwrOR6Txc4/5+0m6e2VjJ2xk18PX8IulweSpLuCTLaKaBzJlk5dAIDN7NwPQZeWLTE0bkTGN9+iWG79S1itVjNo0CAA5syZg81mQwhBUr/BDPv3eGxWKymvvcD2JT/JbkVJqmTerjoeahvFL8+1Y/6TbejfJJTl+y9w74RNtPvvKj5c9jvHL8nWZ0mqyWSyVUTjqAytUgpQhB17gXPDEULg98QTWM6eJWvhT3+6vbe3N8nJyZw+fZrVq1cXXx8aG8/9735CrUZNWDX5a+a/9ya5V24+8F6SpIohhCAh3Iv/DGzIlle78MmIJsQEuPHZqsN0fD+NIV/9yowtsptRkmoimWwV0TiKmQqbGUVrQ7E4vzq0W4cOGOLjufTN1yhW659u37BhQxISElizZg3Hjx8vvt7o7kH/f/yLjvc/yonfdjD5hSc5uGl9BUYuSdKtGLRq+jYO4YeHkvj15c681COWzNwCXpqzm+Zvr+C56TtYd+gSNtnNKEk1gky2ihS2bGE1gdaOUuD8h0YIge8TY7CcOMnVJUtua5+ePXvi4+PD3Llzycu7Vu9HCEGz3v0YNf5jPP0D+OnD/7D40/cx5eRUVPiSJN2GIE8DT3SMYcXYDsz7W2sGNwsj9cBFRk7cRLt3U3lv6QGOpsv3qSRVZ87PKKqKwjFbWE2odCCsauyK8ytCu3fujL5OHS599TWK/c/j0ev1DB48mJycHBYuXPiHMVq+YRGMePN9Wg2+lwO/rmHS2DEc+HWNHMslSU4mhKBJhDdv9W/I5le78Nm9TagX5M6XaUe454PVDPryV6ZtOklWvuxmlKTqRiZbRYpbtsyodKCz6cm35js3JkCoVPg9MYaCI0fIXrbstvYJCQmhS5cuHDhwgK1bt/7hdrVGQ+sh93Lf2x/i7uvHzx//l7njXyfr4vnyDl+SpL/AoFWT3CiESQ8mseGVzrzSM5ar+Rb+Oc/RzfjE1G0s3Xses1UWTZWk6kAmW0W0jjFbWE2o9QKtzUCupWrMEHLv3h1dVBSXvvzqtlq3AFq2bElMTAxLly7lwoULpW4TGF2be9/+gE4PPMqZA/uY/Pcn2bxgNrbbGB8mSVLlCPQw8HiHGJY9354FT7bhvhYRbDmeyeNTtpH09kom7zWz5XimLCMhSVWYTLaKFLVsWUxo9Wq0dj05lqoxTkKo1fiNeRzz77+TvWLFbe2jUqkYMGAAer2e2bNnY7lJ+QiVSk3TXv0Y/cEX1GrUhLXTJjPlpWfkcj+SVMUIIWgc7sW/+9Rn4yudmfxgczrV8+fXs1aGfLWB9u+t4r2lBzh8MdvZoUqSdAOZbBUpMWZLZ9CgtenJs1SdBWU9evdGFx1N+sefoNhur+vAzc2NAQMGkJ6ezrI/6YL08POn/4vj6PfCOKyWAua88y/mvft/ZJ49XR7hS5JUjjRqFR3rBfDR8CZ80smF/w1rTLS/G1+mHaHLh2tI/nQtE9Ye5eJVk7NDlSQJmWxdU5xsmYuTrarSjQggNBr8n3uWgiNHyFqw8Lb3q127Nq1atWLLli3s37//z7dv3pLRH3xJ+/se5PT+PXz/wpOs+v5bOWtRkqoog0YwoEkYPzyUxMZ/dua15HhUQvDWz/tp+Z+VjJq4ibnbT5NrlsMDJMlZNM4OoMoo0bJlMOiqXLIF4N61K4aGDUn/7FM8knuj0ulua7/OnTtz/PhxFi5cSEhICJ6enrfcXqPV0rzvIOLb38P6mVPZvmQh+1avpHm/wTTpkYxWbyiPuyNJUjkLcDfwcNsoHm4bxeGLOSzYeYZ5O84wduYujNo9dIkPpE+jYDrU80evUTs7XEm6a8iWrSIlki2jix6NoiXHVLWSLSEEAWOfx3r2HFemT7/t/TQaDYMGDcJqtTJ37lzstznI3tXLm26PPc2o8R8TXDeWtdMmM/HZx9i5bLEcRC9JVVztADf+3q0ea//RidljWjGgaSjrDqXz2JRtJL61ghdm7WL1wXQsNueXuJGkmk4mW0VKFDV1MToSr5y8qjNmq4hrq1a4tGrJpS+/wpZz+8mgn58fvXr14sSJE6xbt65M5wyIjGbgy68z7PXxeAUGsXLiF0waO4b9a1fd9uxISZKcQwhBYqQP7wxw1O+a/GBzusUHsXTPeR74bjMt3lnJq/N2s/FohpzRKEkVRHYjFikxZsvVxXE5L8/5dbZKEzB2LMeHDCXz+8n4P/nkbe+XkJDAkSNHWLVqFZGRkURERJTpvGFxDRj2+rsc27mVdSk/sPizD9i8YDZJ/YdQr1U7VGrZLSFJVZm2cGB9x3oBmCwNWH0wnZ92nWXO9tP8uOkkgR56ejcMoU/jYBLCvRDC+cuWSVJNUC4tW0KIHkKI34UQh4UQL5dy+1ghxD4hxG9CiJVCiFrlcd5ypdaASuNo2XJx1NzKyzM7OajSGRs2xL1bNzK/m4Q1M/O29xNCkJycjKenJ3PmzCE/v+zJpBCC6CbNGTX+Y3o/8yKKorD40/f57vnH2bV8MdYCJ6/gLUnSbTFo1XSvH8Rn9zZl27iufDKiCY3CvJi68QQDvviV9u+t4t1fDrDv7FW5woQk3aE7TraEEGrgc6AnEA+MEELE37DZDiBRUZRGwGzgv3d63gqhMYDFUfoBwJRfdRMH/+eexW4ycenzL8q0n8FgYNCgQVy9epVFixb95Q9RoVIR26YDD7z3Gf1eGIfR3YMVE75gwtMPs2XhHAryq14XrCRJpXPVa+jbOIRv709ky7guvDe4EVF+bnyz5ii9PllLlw9X87/lBzl4IVsmXpL0F5RHN2IScFhRlKMAQojpQD9gX9EGiqKsKrH9RmBkOZy3/Gn0YHUUNQUwmapusqWPjsZ72DAuT5+O9/Bh6OvUue19w8PD6dSpE6mpqcTExNC0adO/HIdQqajdvCUxiS04tfc3Ns2fxZofJ7F5/iya9OxL0559Mbi5/eXjS5JUuTyNWoYkhjMkMZyMHDNL9pznp11n+ST1EB+vPESMvyu9GgbTs0EwccHusqtRkm5DeSRbocCpEn+fBlrcYvuHgSXlcN7ypzGC1YzW4Ei2LOaqve6Y39NPkbVoERfGv0v4hG/L9KHXtm1bjh49ypIlS4iIiMDPz++OYhFCENGgMRENGnP+8EE2zpvJhtnT2PbzfJr06EPTXn1x8bh1yQlJkqoWXzc9I1vWYmTLWlzMNrF07wWW7D7H56sO82nqYSJ9XejZMJheDYJpEOohEy9Juglxp03CQojBQA9FUR4p/HsU0EJRlKdK2XYk8BTQQVGUPwyIEkI8BjwGEBgY2Gx6Gcob/FU5OTm4Fba8JG16gmz3GHaF/51DPyvsj1vJ4MZdKzyGO+GyciXus2Zz+cknKWjYoEz7ms1mtmzZgsFgoGnTpqhU5Ts5Ne/SRc5v38jlIwdRabT4N0ggsHEiWhfXW+5X8jmRqg75vFRNznherhYobL9gZct5K/sz7dgV8DMKEgM1NA9SE+2puqsTL/leqZoq+nnp1KnTNkVREku7rTxats4A4SX+Diu87jpCiC7Aq9wk0QJQFOUb4BuAxMREpWPHjuUQ3q2lpaVRfJ593rj4eNK2QxsO/bwOvcZIZcRwJ5TWrTm6ZSuBixcTPeZxhFZbpv0jIyNJSUmhoKCAHj16lH+Ag4eScfokm+bN5MD6NWTs+41GnbuT2Hcg7j6lt6Zd95xIVYZ8XqomZz0vfQv/fzm3gOX7LrB4zzlWHr7EL8cthHga6NEgmF4Ng2ga4Y1KdXclXvK9UjU583kpj2RrC1BHCBGFI8kaDtxbcgMhRBPgaxwtYBfL4ZwVo2jMVmE3os1c9QeCCp2OgJf+wekn/kbm1B/xfXB0mfavV68eSUlJbNy4kZiYGOqUYezX7fINi6DX0y/QavAINs2fxc5lP7Nr+WIadOpGUr/BePgHlPs5JUmqHN6uOoY2D2do83Cy8iys2H+BJXvOMXXjCb5bf4xADz096gfRvX4QzaN80KpleUfp7nPHyZaiKFYhxFPAUkANfKcoyl4hxBvAVkVRFgLvAW7ArMKm5ZOKovS96UGdResYs6XRqlCEgt3i7IBuj1vHjrh2aM+lTz/Fo1dPtIGBZdq/a9euHD9+nHnz5vHEE0/g7u5eIXF6B4fS44nnaDVoBJsXzGJ36jJ2py6jYefutBw4DDdvnwo5ryRJlcPTRcugZmEMahZGtslC6oGLLN59julbTvH9hhN4GrV0jg2gW/1A2tf1x0UnSz1Kd4dyeaUrirIYWHzDdf8qcblLeZynwmn0YLrqGGugsaEUVI+mbyEEQePGcTS5Dxf+M56wj/5Xpv21Wi2DBw/mm2++Yd68eYwcObLcx2+V5BkQSNdHn6LlwOFsmjeD3St/YW/aCpr07EPzvoMq7LySJFUed4OWfgmh9EsIJa/AypqDl1i27zwr919k7o4z6DUq2tXxo1t8EJ3jAvB10zs7ZEmqMPJnRUkaA1jTHZd1CsKiRlGUajHQUxcejt+Yx0n/+BNy1g7CrV3bMu0fEBBAjx49WLRoERs2bKBNmzYVFOk17r5+dHnkSRKTB/Lr7GlsWTiHXcsW49ewCZaWLdEa5ILXklQTuOg09GgQRI8GQVhtdjYfz2TZ3gss33eBFfsvohKQGOlDt/hAusUHEeHr4uyQJalcyc7zkgrHbAGodApqmxazrWpWkS+Nz8MPo6tVi/NvvYndXPa4mzVrRlxcHCtXruTMmT/McagwXkHB9Hrq79z/308Jr9+Is5vXM/G5x9idugy7vWqX35AkqWw0ahWtY/x4vW991r3UiUVPt+Wpe+pwNd/CWz/vp/17q+jx0Ro+XH6QPWeyZBFVqUaQyVZJhXW2AFQ6gc5mINdy+4s9O5tKpyPo3//CcuIkl776qsz7CyHo06cPbm5uzJ49G/NfSNjuhH9EJP1fHEe9ASPw8PNn2defMOWlZzm+c1ulxiFJUuUQQtAg1JOxXevyy3PtWfNiJ8b1jsPDqOWz1EMkf7qOtu+u4t8L9rD6YDomi/zxJVVPMtkqSaMHa37hRYHWpifPUr2WnXFt3RrPfv3I+HYCpv37y7y/i4sLAwcO5MqVK3e0nM+dcAsKZcSb79Pn+ZexmE3M+c+/mf32a6SfOFbpsUiSVHkifF14pF00Mx9vxZZXu/DfwY2IC/ZgxtZTPPDdZpq+uZzHftjKjC0nuXjV5OxwJem2yTFbJWkMxS1bWp0arU1PrrX6tGwVCXzlZXLWr+fsq68SNWPGX6q91bFjR1atWkV4eDhJSUkVFOnNCSGo27It0c1asGvZYjbOSeGHl56hfofOtBk28qY1uiRJqhl83fQMTQxnaGI4JouNDUczSN1/kZX7L7Bs3wUAGod5ck9sIJ3jAqgfIivYS1WXTLZK0hqKx2zpjBq0Nj05BTlODqrs1F5eBP3rNc488ywZEyfiN2ZMmY/Rrl07zpw5wy+//EJQUBAREREVEOmf02i1NOvdj/odOrNp/kx2LFnI77+uJTG5P837DkJnlANpJammM2jVdKoXQKd6AbzRrz4HzmeTesCReH208iD/W3GQQA8998QGcE9sIG1q+8qyElKVIl+NJWkMYCsAux29QYvWbiDPWr26EYt4dOvG1R49uPT5F7h37lymhaoBVCoVAwYM4JtvvmHmzJk8/vjjFVZ/63YY3NzoMPIhErr1Yt30KWycO4PfVi6l9ZD7aHhPN1RqtdNikySp8gghiAv2IC7Ygyc71SYjx0za7+msPHCBn3adI2XzKXQaFa1jfOkcG0Cn2ADCvOWPMsm55JitkjSFdV6sJgwGHVqbrloNkL9R0GvjULm6cuall1AKCsq8v9FoZNiwYZhMJmbNmoXN5vzBqZ4BQfR+5kXuffsDvINDWTHhc75/8SmObNssZy1J0l3I103PoGZhfHFfM7a/1pUfH2nByBa1OHYpl9cW7KXtu6vo/EEaby7axxo5yF5yEplslaQprOtkNWEw6tHa9eSYq2+ypfH1JfitNzHv20/6J5/8pWMEBQXRt29fTp48yfLly8s5wr8uuHY9hr0+nn4vjEOx25n/3zeY9earXDh62NmhSZLkJDqNija1/fhXn3jSXujIirEdGNc7jhAvI1M2nuD+7zaT8MYyRk/azOT1xzh2KVf+SJMqhexGLKk42TLj6uK4nJtXPbsRi7h36YLXkCFkTPwO13btcW1R9sHujRo14syZM2zcuJHQ0FAaNmxYAZGWnRCC2s1bEtUkkd9W/sKGWdOY+spzxLXrRNvho/Dwk2suStLdSghB7QA3age48Ui7aPILbGw8msHqg+msPpjO6z/tg5/2EeHjQoe6/nSs50+rGDnWS6oY8lVVUnGylY+rixGA3LzqP7048JWXyduyhbMvvUT0gvmoPT3LfIxu3bpx7tw5FixYgK+vLyEhIRUQ6V+j1mho0j2Z+Had2LxgNtt/XsDBjetI6J5MUr/BuHiU/f5KklSzGHVqOhWO4QI4kZHrSLx+T2f2ttNM2XgCnVpF8yjvwuQrgDoBbnKGo1QuZDdiScVjtszojI48ND+/+idbKhcXQt57D+ulS5wbN+4vNZur1WqGDh2Kq6srKSkpZGdnV0Ckd0bv4kq7EQ/w4EdfE9u6Pdt/XsCEpx9h/YwpmHKr36xSSZIqTi1fV+5vFcnE0c3Z+W/HWK8HWtciPdvMO4sP0O1/a2g9PpUXZu1iwc4zpGdXn9VEpKpHtmyVVGLMlk7veGhM+WUfWF4VGRs2IODvf+fiu++SOWkyvg89WOZjuLm5MWLECCZOnMj06dMZPXo02jLW8KoMHn7+9Pjb8zTvO5hfZ09j49wZ7Fi6iOZ9BtGkZx90BqOzQ5QkqQrRa9S0qe1Hm9p+vNobzl7JZ83BdNYcSmf5vgvM3nYagNggd9rW9qNtHT+Sonxkl6N02+QrpSTttTFbWr2jlIDZZHFiQOXLZ/QD5O/YwcUPPsDYsAEuzZuX+RhBQUEMHDiQGTNmsHDhQgYOHFhlm9l9w8Lp89xLXOw/hPUzp7Ju+g9sW7yAFv2H0KhrT7Q6vbNDlCSpCgrxMjI8KYLhSRHY7Ap7z2ax9tAl1h++xA8bTjBh3TF0ahVNa3nRtjBJaxTmhVpVNT8LJeeTyVZJRS1blny0BkeyZcqvOcmWEILgd97GPOh3To8dS/TcuWj8/ct8nLi4OO655x5SU1MJCAigXbt2FRBt+QmIjGbAP/7F2YMHWD9zKmk/TGDronkk9RtMg3u6yaRLkqSbUqsEjcK8aBTmxZOdapNfYGPL8UzWHb7EukOXeH/ZQd5fdhAPg4ZWMb60reOPNteOoihV9oeoVPlkslVSyTFbbkVjtmpWP73azY3QTz7h+LBhnH7mWSImT0KlL3uy0a5dO9LT01m5ciV+fn7ExcVVQLTlK6RuLEPGvcWpvb+xbsZUUid9zca5M2jWuz8J3XrJavSSJP0po05N+7r+tK/r+KF6KcfMr0cyWH/oEusOX2LpXsdSQp/uXkWrGF9aRfvSKsaXEC85fOFuJpOtkkqM2SrqRjTVsGQLwFCvLiHj/8OZ557n3GuvEfLuu2X+BSaEoG/fvmRmZjJ37lzuv/9+wsPDKyji8hVevxHD/+9dTu/bzab5s1g7bTJbFsymSc8+NOnRB6O7h7NDlCSpmvBz09O3cQh9G4egKArHM/L4bvGvXBQerNh/bbxXLV8XWkY5Eq9WMb4EehicHLlUmWSyVVKJOltF3YgFJht2xY5K1KyJmx49elDw7DHSP/4EfXT0X1o/UavVFg+YnzZtGg899BD+f6Fb0hmEEITXb0R4/UacO/w7m+bNYsPsFLb+NI/G3XrRrHd/3Lx9nB2mJEnViBCCKD9XOkdo6dgxEbtd4cD5bDYczWDDkQwW7znHjK2nAIj2c6VFYatXy2gfAtxl8lWTyWSrpBJ1trQ6R7Kltmm4ar6Kl8HLiYFVDN8xYzAfPUb6Rx+jq1ULj549y3wMNzc3Ro0axcSJE5kyZQoPP/wwnn+hjpczBdeuR/8Xx5F+8jib589i26L5bF+8kNg27WnWuz8BkdHODlGSpGpIpRLEh3gQH+LBw22jsNkV9p+7yoYjGWw4msFPu86SsvkkALUD3GgV7UvLaEfy5esmx5LWJDLZKqlEy5ZQCYRWQWszkGHKqJHJlhCC4LfexHLmDGf/8RJqT09cW7cu83F8fHwYOXIkkyZNYurUqTz44IO4uFS/8U/+EZH0fuZFWg+9j+2LF7I3bQX71qQSXr8RzXr3I7pJc4SqZrVwSpJUedQqQYNQTxqEevJo+2isNjt7z14tbvmas91RXBWgXqA7SVE+NI/yISnShyBP2fJVnclkqyTttTFbABq9Cq1NT0Z+BjFeMU4MrOKo9HrCv/yCE6Pu59RTT1Nr0ncYGzcu83GCg4MZMWIEU6dOJSUlhVGjRqHT6Sog4ornHRRC54fG0GboSH5b+Qs7li5i/n/fxDs4hITufajf4R70Lq7ODlOSpGpOo1bRONyLxuFejOkQg8VmZ/eZLDYcyWDj0euTr3AfI80jHYlX8ygfov1c5WzHakQmWyWpr81GBNAa1I5ky5ThxKAqntrTk/AJ33LivpGcfOxxIqdOQV+nTpmPExUVxaBBg5g5cyazZs1i2LBhaDTV9yVmcHMjqd9gmvXuz6FN69m2eAGrJn/N2pTJxLZuT+MuPQmMqSM/8CRJKhdatYqmEd40jfDmyU61sdrs7Dt3lc3HMtlyPJO039OZu/0MAH5uOhJrXWv5igt2R6OWLe9VVfX9JqwIag2oNGDJB8Bg0KLLMXAp/5KTA6t42oAAIr6byIl77+PE6AeJmPQdhrp1y3yc+Ph4kpOTWbRoEXPmzGHw4MGo1eoKiLjyqDUaYtt0ILZNB84fOcRvK5awf/1q9qxaTkBkDI279iS2TXtZOkKSpHKlUauKa3w90i4aRVE4kp7LluOZbDmWyebjmfyy9zwArjo1TWt5F7d8JYR7YdBW78/emkQmWzfSGIpbtlzcDBiyXMnIr9ktW0V04eFEfD+Zkw+M5uT9DzgSrr9QPysxMRGLxcLSpUuZN28eAwYMqPYJV5GgmDoExdShw6iH2b82jV0rlrD8289I+2ECdVq0Jr79PYTXb4hKVTPuryRJVYcQgtoBbtQOcGNEUgQA57Lyi1u+th6/zAfLDwKgVTvGhxW1lDWt5UWwp6z15Swy2bqRxlA8ZsvgqsXF5k6G6ZiTg6o8+uhoak2dwonRD3LigdFETJyAsWHDMh+nVatW2O12li9fjkqlon///qhq0OByvYsrCd1707hbL84ePMDetOX8vmEd+9ak4ubrR3zbjsS374xvWPWoPSZJUvUU7GmkX0Io/RJCAcjKs7D1hKPVa/uJy0zdeIKJ6xzfYUEeBprW8qJphDdNIrypH+IhW78qiUy2bnRDsqW3unAs76KTg6pculq1qDVlCicfeIATD4wm7H8f4tahQ5mP06ZNG2w2G6mpqahUKvr27UZaRZcAACAASURBVFujEi5w/NIMrRdHaL04Oj34OEe2bmL/2lVs+WkumxfMJjC6NnVbtqVeq7Z4BgQ5O1xJkmo4TxctneMC6RwXCECB1c7+c1fZfvIy209eYfuJyyze7eh61KlVxId4FLd8NYnwJsTTIMehVgCZbN1Io7+WbLlp0VoM7Lu0/65b50oXFkqtadM49cQYTv3tSYL+9S+8hw0t83Hat2+PzWZj9erV2Gw2+vfvX2O6FG+k1emJbd2e2Nbtyb1ymQPr13BgfRprp01m7bTJxYlX3ZZt8QqUiZckSRVPp7k24/HBNo7rLl41sf3kFXacvMz2k5f5cdMJvlvvaP0K9NDTJNybRuGeNA7zokGoJ55GrRPvQc0gk60blRizpXfRIBRBXp6JU9mniPCIcHJwlUsbGEDklCmcfv55zv/73xScPEHA2LGIMiZLnTp1QqPRsHLlSgoKChg8eDBabc1+87p6edOsdz+a9e5H1sULHNy0noMb1xUnXgGRMcQkJhHdNInAqBhZv0uSpEoT4GGgR4MgejRw/Oiz2Apbv044Wr92nrpSPPAeIMrPlUZhnjQK86JxmCfxIR646GT6UBby0bqR1nBdyxaAwerCrvRdd12yBaBydSX8iy84//bbZE78DvP+/YS8/z4an7ItZdOuXTv0ej2LFy8mJSWF4cOHV9s6XGXlGRBI8z4Dad5nIFfTL3Jw4zoObd7AhjnT2TA7BVdvH6KbJBLdNIlaDRPQGmTxQkmSKo+2xKzH0YWtX5dzC9h9JovfTl9h1+ksNh3NZMHOswCoBNQNdC9OwBqFeRIb5IFOI3803oxMtm6kMYClMNlycSRbnoovv6X/Rp+YPs6MzGmERkPwv/+NsUEDzv/fGxwbNJiwjz/C2KhRmY6TlJSETqdjwYIFTJkyhREjRlTLSvN3wsM/gMQ+A0nsM5C8q1kc37mNI9s28/uGdexOXYZaoyGkXjy1GiYQ0aAxgdG1UdXQbldJkqoub1cd7ev6077utfVuL1w18dtpRwL22+kslu+7wMytjoW2dWoVccHuxId4Uj/Eg/ohHsQGeWDUyc8vKKdkSwjRA/gYUAMTFEUZf8PteuAHoBmQAQxTFOV4eZy73Gn0YLoKXGvZinNtwMZzG50ZVZXgNWgQ+thYzjzzLMfvvQ//J/+G76OPIspQuDQhIQGdTsecOXOYOHEiI0eOxNvbuwKjrrpcPDyJb38P8e3vwWa1cObAPo5u38LJPbtYN/0HwDHrMSy+IRENGhMWVx+/iFqyrIQkSU4R6GGga7yBrvGOwfeKonD6cj67Tl9h9+ksfjudxc+/XVvvUSUgxt+N+MLkq35hIublcnf0apR0x8mWEEINfA50BU4DW4QQCxVF2Vdis4eBy4qi1BZCDAfeBYbd6bkrhMYA1nTAMWYLoKF7Asuy5nI06yjRnnf3osTG+vWJmjuH82+8SfrHn5CTtpqQ/76Lrlat2z5GfHw8rq6upKSkMGHCBO69915CQ0MrMOqqT63REtGgMRENHEsl5WVd4eTe3zi5eycn9+ziyFZHsq8zGgmuE0tI3VhC6sUTUqeeo5iq3eYYa2gruP6folw7SfEED1F4WZS4/s8v68yXIefiDce48XIpxxCq6895q+vuokkoklTdCSEI93Eh3MeF5EYhwLUEbO/Zq+w7m8Xes1ev64IECPUyXpeAxYd41PhZkOXRspUEHFYU5SiAEGI60A8omWz1A14vvDwb+EwIIRSl5DdBFaH545it2oZ6kAWpJ1OJbnh3J1vgWN4n9IP3cbunE+f/7w2O9u2H3xNj8HnoIVQ6HYqiYLcrKHbF8V2vON6AOP4DRSHQL4RRIx5g5pzpTJ40mT69+xMTXRsUsOQpZGc6ngPF7niJKCWO4fhbuel1xecDFDsohSd23KYUx6DceN0N294Yc/Gx7XYUuw2Uwv/bbI7rFLsj6bE7/sZuB8UGNovjn72g8HIBwm4pTIhK3mZ1XFe4nbAVEGkrIDLcgzxfNZk5eWTmmLj8+2+c+G1H8fPhrgVPnYKXXsFT57hccujE9Z9f195yosTlG11/m+NyEHB01Re32L/ksUu/3nHjn28noETSJoquKMzLbry+xLYlt6NEIle0jRA3HPv6fUXRMYSq9OMWb1d0vBu2uy6B5Nrf4vpzOf5UlbJdyeOVeIRKS2wLjxVw5TLnjyz+wz7iJvfx2nFKHu+Pfxff5aL79Ydzc/3lkonzdc/BDY93aecseS5R2rlKT+jFjdvwx21uetzCbcQtHoM/3NfbIcDl6BEyxOU/PESlbnw7hyz5ON3WYf647a3PVPjhVnRZKf5QLeX2ouuV624XJbbxUBRaAa00CkQoEAF5Zgvp2WYuZptIv/r/7d17fFTVufDx39pzzZUQICES7ncUBA0oFRTUUrVVatVWwcqxCq2Xl761Wj222vPq8WPbc+qxtZ4ewQuiVVS8i1bRFvEgiAjKnYBcEyQJkAC5TWZmr/eP2TNMJjO5kJnMJHm+fPZn7732mr2fzGJmnll7zd4eyvfWUbW1gU+BT4F0h0GfTCe9Q5OLnhl2HEqdPI7SJ48dDCs8DqWbllnLtsyeQPI+v+ORbPUDDoStlwDnxKqjtfYppY4BvYDUuw9OWLLlSgs8PU6vmzN6ncF7e97jpjNuSlr2rbWmvsZLdaWHhlofnjofDfU+vPV+/D4zkFiEJTqhhMfU+P0a7Q+UmaG5GSgLrge3+a1tjepG20825rcexl/nQX8Bev0/0cpAq9af5rIbo9E9N/PaG6/gqu+Du64AR0MPit/6NIHPZLIowGlN7ZAOrjQPpu8Q2neQGn8ZJ2rLKKmpCavkCBwv2IPUyjd1ESniQyZlHU92ACLCJ6xtVT0NoBTaMMAwTs6VgTas17BhoJU6OQ/bFmve8ku+7e8JTV4JrdpFM5XCfoVd64d9xxrYd6yhTbtobSVbQwVTvtu5k624UUrNA+YB5Ofns2LFioQfs7q6utFxRpQfoXftcT61ygwHfF28l9NHnM5LlS/xzPvPMMQdvcEa97IECyPWafql1fSC6bM6OKyODl8teOu0NQdvbWDS/jb+geFfmIOvwSjrhJXHWjcMExfHcJlHcelKXP5KbIYXleHHV6PwHlfgB+3T6AYTvBrMwIeVOtlFZC1rlB2UXWEeggMFGZT20njSKnD6/OTXeyjw1pOrGzBsGsOuMewmytAYNn+jb7SNvoGriOVgHQU6+E07rFyH9XQEerd0oIfLbz3Xfo32gfb5wQfaa4LPj/aaKK+J9vrB5wevHxr84PVBgw/V4EN7fShv0wbTkW8K1rG1y4V2ucHpRLucmC53oMzpgFjjtGygbQY4C2gwfdSYHmpNDz5tWseyeu0i3iabxNAKja81F+/krXX7a33aE7/4Yj9XsaJp6djxTd60qVFGM8dMpVwxPBYVvp6IIHXgRd/mfUevrwHTSna0YYSWTStBMo1AuTYMzPA6xsk6seq3/U/TKNNEaY1hapQ2rblG+TVK+1DxOnEU2THdumqNH9NkY/OxhTqn2hBXtMLwWG3+pp/3HSkeyVYpEH5PkkKrLFqdEqWUHehBYKB8I1rrBcACgKKiIj1t2rQ4hNe8FStW0Og49e/D0dWhsgMfraZXTjYzL7mDZS8vY5f3ABPdM6g4cIKqslpqqjw01PtpqPdh+uL7hqEMRUYPJ5k5LjL7ucnMdZHV001mTxeuDAeuNDsOtw2n247NrlBGYDKseeDsRTs/dOoqYeubsP1d2PsJUAsuBwweD/3PgdPOhZyBkNMfMvIafVMBMOvq8FdV4a+sxF9Vha+yEn9lVaMyf2Ulw745hGfHcQ643Ozr25fS0wo4kJlOWm0t/fcfYMD+/eQePRp48RgGRkYGyuVCORwouz0wD1sGAqf1NNbpPG2d5gsketrrRXs8mB4POjg1RPlG1Vz7pKVhZGRgZKRjZGRg65ltrceYMgNzW5Rtyp364xWavFZESpB2aco0TbxeLx6Ph4aGhlbPGxoa8Hq9oeXwdb+/bd90nU4nDocDp9MZmsLXm9sWXA9Odru90XJXvTB0oiXztRKPZOtzYLhSajCBpOpaYFZEnbeAOcBq4GrgHyk5XgsCv0b01oVWs3JdlGw/yraPMrhu42+wHUtjJcWkZTnIyU+n75AeONPsON127E7j5Ln1YI9SqAfm5PgFzcnxTEqBw2XHmWbD6bLjSLPhSrOT0cNFWrYTo7lvrIli+qH47/DlC7Dzg8A4op6DYML1MHwGDDwPnK27ZIORloaRloajoKBV9UcDKz74gOtGjWL71q1s272bXRkZFI8aSbbDwTC3m6FATl09eL2BpMlnza0Jr5dQAxiGtWiErSuUw4HhcgUSNpcLw+1COa1llxPlcmOkuWMnTunpbb64qxAiNp/Ph8fjaXOCFGveWsHkxuVyhZKc9PR0cnJyWpUQRVtftWoV06dPT+CzJTqbdidb1his24H3CVz64Wmt9Ral1APAOq31W8BTwHNKqV3AUQIJWWqyu63zen4wbFxw3Uhef2QDa97YTW5hD97OfZ7pkydx49QbU74nos3qKmH9c/D5QqjaD5n5MPFmGHsNnDah7QNFT5XTSeaAARQNGEARUFdXx/bt29m8eTMbdu9mvdbk5uYyatQoRowYQf/+/eWbnhBJorXG6/VSX18fmjweT6P1aGWRCZJpmq06ns1mCyVGwXkwOXK5XE22NTd3OBwJuV9rl/tsEO0WlzFbWut3gXcjyu4PW64HronHsRLO7grMfR5wptOzbwbXP3AunlofWbluVn6wmFcO/Y2bzB/jsHWRW87UVcHqx2HNX6HhBAycAjP+HUZ+F2zJH9aXlpbGhAkTmDBhAjU1NWzbto2tW7eyZs0aPv30U9xuN0OHDmXEiBEMGzaMjIyMZIcsRKfh9/tDiVBrkqRodVo6UWEYBm63OzS5XC4yMzNjJkHNJUjyxapr0jowwlRrjRn4abg15tQqCx+HatUNlelQzaZl1v9NW5KvT5j8T9JUY08LzH31oVNlTnfgNCHADaffwC0f3sJ7e9/jiqFXJCvK+PDWw5rHYdWfoP4YjJkJU++EgrZdGb4jZWRkUFRURFFREfX19ezevZvi4mJ27tzJli1bUEpRUFDA4MGDGTJkCP379+82twVKVaY28Wt/6A0zOAXL/NqPRuM3/aE3y0bbrHnosZixt4VNwX2aNC4LvokH14NvxuHrjebB5Rhv6KH9Btet+IJlpjab7ivGB0ijD5mwfQXrNNqXFdPBioN8vOJj8AZ+nIKXwOQ7OWmvbrQevl35W+6F0TYNdtB2jbZby06NTteBH2lY5U0mmzU3dKPRyuF/U2jdZ8Vfe7K80dUGiPKYKOUnZ43rhZ7nWPVi7TcikYxWL3Kfx44fY8GyBTFjbMs+m9Rtxz6D66F4W7nP8L8v/LUTvr/I11esxCdaIhUZTyIM6TGEX+T8IuHHiUWSrUihnq36qJvPO+08huUMY9GWRVw+5PLO211c/D68dzdU7oERl8L0e1M6yYrG7XYzZswYxowZg2maHDx4kJ07d7J7925Wr17NqlWrsNlsFBYWMnjwYAYOHMhpp52Gy+VKdugJobWm3l9PjbeGEw0nGs3rfHV4TS8ev4cGf0NgMhtOLvsbAtvMBnymD5/pw6/9gbnpx6d9HK06yhPvPIFf+/GaXvzaH9hm+vDpk/X8pv9kHSuBEk0p65/TdOIwHTi1M7TsMB04tAOn34ld23H4A2V2047dDKzbTTt2v52+9G32OTYx8dv8+Gw+fIYvsGz48KcFyvw2P37DH6pj2sxQebB+8H0uNCf6euhvMxU00LhuxGNCdSP32cK+2/SY0OzkPlTY9bmaxB1WHqqnmh4ncr+R+/QaXrJd2a2KsbX7DH9s+N/Spn1G/H1t2qdqHFf4c9TsvJltAIYyGtVFgYHR5v2F9kXjsmCdLGcW7CVpJNmKZLduAhwj2VJKMef0Ody36j5WH1zNt/p9qwODi4MTZbDsDtj+DvQaDj9+A4Z2/oGchmFQWFhIYWEh06dPx+PxsH//fvbs2cOePXtCP/dVSpGfnx+q279/f3Jzc1M2afb4PVTUVnC0/ihH649SWV/JkfojVNZXRi3zmt427d9hOHDZXDhtzsBkOHEYDuyGHZthw67soWWncpLjzsGurG2GHZsKzIPLkevBuaGMZiebsqGUwiD2tvA6NsOGQjXZFjxWZJlCnXwjVk3fuA2M0IdqeD0g9joKv99Pg6chMNUHfr3mqbcGeXtOLns8Hjz1Huo99YF52LilFtvI4Qicfks7eQou/JRcaWkpY8aMaXKaLrjscDhS9v93VyW/EE1NK/auSNqxJdmKFD5mK4bLBl/Gn9f/mUVbFnWuZGvrm/D2/w1csOvif4NzbwN71zzF5nK5GD58OMOHDwegtraW0tJSDhw4QElJCRs3bmTdunVAoIcsPz+f/Px8+vbtS35+Pnl5eTgciR2TV++r52D1QUqqSyirLaOitoLy2nLKassory2nvLacKk9V1Mem2dPIdeeS684lLz2Pkbkj6enuSbYzmyxHFhnOjMDckUGmM5M0expOwxlKqlw2Fw6jbR/CXe0DxDRNGhoaTmmMUrCspcsBKKWaJEC5ublRk6ZoZa0Zo7RixQomTpwYz6dGCBFnkmxFcoSN2YrBaXMye/RsHl3/KDuO7mBk7sgOCu4UNdTAsl/CVy9CwXj4wQLok+Ixx1l6enqj5Ms0TSoqKigpKeGbb76hrKyML7/8MtTToJQiNzeXXr16NZmysrJanaT4TB8HThxgV9Uuvq76mpITJRw4cYCS6hLKa8sb1VUoeqX1ok9aH07LOI3xfcaTl55HXnoevdJ6hZKrnu6epAXHFnZT4b+AC/YcRSZGLQ329nhif6EKcjgcjRKg9PR0cnNzoyZF0RInp9MpvUpCCEm2mgj2bHljJ1sA14y8hgUbF7BoyyIenvpwBwR2io58DS9dDxXb4YK74fy7oKv8irIdDMMI9WYFmaZJVVUVhw4doqysjIqKCo4cOcLu3bvx+Xyhena7nR49ejSZ7Ol2DjYcZF/9PvbU7WHXsV3srtpNg3nyVFFeeh6FmYVMLphMYVYh/bP60y+zH30z+tIrrRcOo+u3TbBHKTJBipY0xdrWmksFKKWaJEE5OTmtSpKCy/LLNyFEPEiyFamFMVtB2c5srhpxFS9se4Hbxt9GYVZhBwTXRsXvw6tzAxfyvP5VGHphsiNKaYZhkJubS25uLmPGjAmVm6bJiRMnOHz4MEeOHKGyspJjx45RdqSMfQf34a9veirJiZPRjtGMdY8lIyODnKwcemX1IiMtI/BB7nLjNty4fC7c9W4MDE54ToSuFG2327Hb7R3eK6K1xufzRZ2OHTvGnj17QuuRV9pu7RSeuMYSPP0WTHxcLhfZ2dnk5eU1KgtfjkyUpFdJCJEqJNmK1IoxW0FzxsxhyfYlPLP5Ge6bfF+CA2ujz54I/NqwYBz88DnoOTDZEXVahmGQlZ3FIf8hij3FfF79Oev86ziefhzSobezN+OyxzEsbRiFzkJyjVz89X5qampOTkdqOFJ6hPr6+lZfvBEIJV12ux3DMEKDu8OXw8tClyjQGtM0m12PLPP7/S2OQdqwYUOzsYZfTTs4Ba+nFDlFS5SCyzKoWwjRlUiyFcne8pitoPyMfGYOm8nru17np2f+lLz0vAQH1wqmCR/+Fj79M4z6HvxgYatvrSMaO1J3hP8t/V9Wlqzks0OfccxzDIDCzEIuGnARE/tO5Oz8synIKGh1YhDsOYo2fsjr9YZ6jMJ7j4LLLSVPpmlGTcRaWg+WGYbRKLmL7GXbunUrZ511VqPt4clTIq7ELYQQXYEkW5FauM5WpJ+c8RNe3/k6z255lrsm3pXAwFrB74U3boVNLwdus3PpHyDJV83tTExtsu3INlaWruSTkk/YfHgzGk2ftD5MK5zGpIJJTMyfSEFm6+7zGI1SKnRD2aysrDhGn3iHDh1i0KBByQ5DCCE6HUm2IrVyzFZQ/6z+XDb4Ml4pfoWbx95MT3fPBAbXDL8Xlt4I296GC++Dqb/suHsZdmKmNtlQvoHl+5azfN9yymvLUSjG9h7LreNv5fzC8xmVOwpDSa+NEEKIUyPJVqTQpR9aHrMVdPPYm3ln9zs8t/U55p81P0GBNcPXEEi0tr8Dl/wOzr2l42PoRPymn/Xl6/lg7wd8tP8jKuoqcBpOpvSbwkVnXcSUflPIdecmO0whhBBdhCRbkUKXfqhr9UOG5Azh4oEXs2T7Em4848bAbQE6SrBHa/s7cMnv4dyfddyxO5kdR3fw9tdv8+6ed6moq8BlczG131RmDJrB+YXnk+GQG1gLIYSIP0m2Itla/2vEcHPHzmX5vuUs2b6EuePmJiCwKEwT3rwtkGhd+gc456cdc9xOpLy2nGW7l/H27rfZWbkTu2Fnar+pXDb4Ms4vPJ90h/x4QAghRGJJshXJZgfD3uoxW0Gje41mSr8pLN66mFmjZyW+l0Rr+OA3sPEluPA3kmiF8ZpeVhxYwavFr7L6m9WY2mRc73Hce869XDLokuSNqxNCCNEtSbIVjT2tzT1bALeeeSuz3p3F37b9jXnj5iUgsDCrHoU1j8M5P4Opdyb2WJ3EweqDLC1eyuu7Xudw3WH6ZvRl7ti5fG/I9xjUY1CywxNCCNFNSbIVjd0FvtaP2Qoa22cs0/pPY9GWRVw76lqyndkJCA7Y+Ap8+G9wxtXwnYe79a8O/aaflSUrebn4ZVaVrkIpxdR+U7lmxDVM6TcFm1z6QgghRJJJshWN3X1KPVsAt4+/navfvprFWxZz+4Tb4xwYULIuME5r4Hnw/b8GbsXTDdV6a3lj1xs8v+15Dpw4QJ+0PswbN4+rhl/VrutgCSGEEPEmyVY0dlebx2wFjcwdyYyBM3hu63PMHj07vuODjpXAi9dBdkHgFjx2Z/z23UkcqjnEC9tfYGnxUk40nGBc73HMP2s+Fw24qFvcxFkIIUTnI8lWNI5TG7MVdOv4W1m+bznPbHmGO86+Iz4xNdTAi9cGksA5b0NGr/jst5PYVbmLhZsW8sHeDzAxuWjARdww5gbG541PdmhCCCFEsyTZisbuatN1tiINzRnKd4d8lxe3vcjsUbPJz8hvf0zL7oRDm2HWy5A3qv376yR2HN3BExufYPm+5aTb07lu9HXMGjWLwqzCZIcmhBBCtIokW9G0Y8xW0G3jb+P9ve/zly//woPnPdi+eDY8D1+9ABfcAyNmtG9fncT2o9v5n6/+h4/2f0SmI5N54+bx49E/Jsedk+zQhBBCiDaRZCsauwvqj7drF4VZhcwaNYvFWxdz/ejrGZk78tR2VLYl0Ks1+AK44FftiqkzKK0u5ZmKZ1j/9nqyHFnccuYtzB49mx6uHskOTQghhDgl3fOnbC05xetsRZo7bi5Zziwe+eKRU9uBpxpengPubLjqSejClzGo9dbyyBePcMXrV7CpbhNzx87l71f/nVvH3yqJlhBCiE5NeraiOcXrbEXq4erBT8f9lP9Y9x+sKl3Fef3Oa9sOlt8HR3YFBsRn5rU7nlS19pu13P/p/RysPsjlQy+nqL6IK8+6MtlhCSGEEHEhPVvRxGHMVtC1o66lMLOQP37xR/ymv/UP/PofsO5pmHwbDJ4al1hSTa23lofWPMRNH9yETdlYdMkiHpryED3tcjsdIYQQXYckW9E43Kd8na1ITpuTn5/9c3ZW7uStr99q3YPqj8Ob/wd6jwjc97AL2nJ4Cz9650e8tOMlrh99PUuvWMpZ+WclOywhhBAi7iTZisbuBm98ki2A7wz8DuN6j+OxDY9R661t+QH/fAiOlwauEO9Ii1scqcBv+nlq01Nc/+711PpqeXLGk9w96W7S7F3r7xRCCCGCJNmKph1XkI9GKcVdE++ioq6ChZsWNl/5m69g7QKYeBMUFsUthlRwqOYQc5fP5dH1jzJ9wHReu+I1JhVMSnZYQgghREJJshWN3Q2mF9oyxqoF4/PGc8XQK3h2y7PsP74/eiXThGW/hLTcLnf6cPm+5Vz11lVsPryZB771AH+84I/yK0MhhBDdgiRb0djdgXmcBskH/eLsX+C0Ofn957+PXuHL56Hkc5jxIKR1jUHitd5a7l91P3esuIMBWQN45fJXuHL4lSilkh2aEEII0SHalWwppXKVUsuVUjuteZMMQSk1Xim1Wim1RSm1USn1o/Ycs0OEkq34nUoE6J3Wm1vOvIWVJStZWbKy8cbao7D8tzBgMpx5XVyPmyzbj27nR+/8iDd2vcHcsXNZfNliBmYPTHZYQgghRIdqb8/WPcBHWuvhwEfWeqRa4Aat9enAJcCjSqnUvueK3RWYxznZApg1ahaDsgfx+7W/x+MP6zn7+A9QXwXf/SN08l4frTVLti9h9rLZ1HhreHLGk8w/az4Ow5Hs0IQQQogO195kaybwrLX8LPD9yApa62Kt9U5r+SBQDvRp53ETK0E9WwAOm4N7z7mX/Sf2s2DjgkBh5T5Y9xSMnw35p8f9mB3pmOcYd6y4g4c+e4hJBZNYesVSGQQvhBCiW1Na61N/sFJVWusca1kBlcH1GPUnEUjKTtdam1G2zwPmAeTn55+9ZMmSU46ttaqrq8nMzGxU1qd8Fadv/QNrJz5GbcaAhBz3ucPPsa5mHXcX3M2FX79En4pPWTvpr3jcvRNyvI6wx7OHRRWLqPJXcUXOFUzPno6h2p7PR2sTkXzSLqlJ2iX1SJukpkS3y/Tp07/QWke9jECLt+tRSn0I9I2y6dfhK1prrZSKmbkppQqA54A50RItax8LgAUARUVFetq0aS2F124rVqygyXF21MFWmDT+DOiXmAttnll/JjPfmMmy2lf4UdnH2M6bz+RvX52QYyWaqU2e2fwMj214jL4ZfXns/McY12fcKe8vapuIpJN2SU3SpNWUMgAADzRJREFULqlH2iQ1JbNdWky2tNYXx9qmlCpTShVorb+xkqnyGPWygWXAr7XWa0452o4SGrMV318jhuvp7sk9k+7h7k/u5qnefZg35RcJO1YiHao5xH2r7mPNN2uYMXAGv/3Wb8l2Zic7LCGEECJltHfM1lvAHGt5DvBmZAWllBN4HVistV7azuN1jODVzBMwZivcpUYPLq2u4a9ZaWyqLknosRLhvT3v8YO3fsBXFV9x/+T7+c8L/lMSLSGEECJCe5Ot3wHfVkrtBC621lFKFSmlnrTq/BA4H/gXpdSX1jS+ncdNrA7o2QJQK//Ab+ps9EnP555P7qG6oTqhx4uXitoKfrnil/xq5a8Y3GMwSy9fyjUjrpFrZwkhhBBRtHgasTla6yPARVHK1wE3W8vPA8+35zgdLvRrxLrEHaPkC9i9guxvP8DDw6Zw0/s3cc8n9/DnC/98SoPKO4KpTZYWL+XRLx7F4/cwf8J8bjzjRuxGu/4bCSGEEF1aan6qJ1tH9Gx98p/gzoGin3B2/tncPeluPi75mL9s+EvijtkOxZXF3PDeDTy45kHG9BrDazNfY+64uZJoCSGEEC2QT8poHAkes1W2BXa8C9P+FVxZAFw78lp2HN3Bwk0L6ZvRlx+O/GFijt1GR+uP8viGx1m6cynZzmwemvIQlw+5XE4ZCiGEEK0kyVY0wdOI3gQlW5/8EZyZMGleqEgpxa/P+TUVdRU8uOZBXDYXM4fNTMzxW6HB38AL217giY1PUOer47pR1/GzcT8jx53aF/8XQgghUo0kW9Ek8HY9HN4Fm1+D8+ZDem6jTQ6bg0emPcLtH93O/Z/eT423hlmjZ8U/hmZ4TS9v7XqLhZsWUlpdytR+U7mz6E6G5Azp0DiEEEKIrkKSrWhCA+QTMGZr9WNgc8Lk26Nudtlc/Gn6n7j7k7t5eO3D7D2+l7sm3pXw+wrW+ep4++u3eXrz05RWl3JGrzO4/9z7+Va/byX0uEIIIURXJ8lWNIYNDEf8e7ZqjsBXS+DMayEzL2a1dEc6j057lEe+eITFWxezvmw9D5z3AGN6jYlvPEBpdSkv7XiJV4tf5XjDcc7odQb3nnMvU/tNlXFZQgghRBxIshWL3R3/ZGvd04F9nntri1Vtho27Jt7F2fln8+9r/p1Zy2Zx6eBLuXnszQzNGdquMA7XHWb5vuW8t+c9NpRvwKZsXDTgImaPns2EvAmSZAkhhBBxJMlWLHZXfJMtnwfWLoBhF0PeqFY/7MIBF1LUt4gnvnqCV4pf4Z3d7zAhbwLfGfQdzi04l8E9Bjd7XS5Tm5RWl7Kzcifrytbx2TefUVxZDMCwnGHMnzCfy4deTt+MaLe/FEIIIUR7SbIVi90d3zFbm5ZCTTlMvq3ND812ZnPXxLu4eezNLC1eynt73+N3a38HQJYzi4FZAynILCDdno7dsFPjraHKU0VlfSX7T+ynzro4q9NwMiF/AvMnzOeC/hcwoueI+P19QgghhIhKkq1YHHE8jag1rPlvyBsDQ6af8m56unsyd9xc5o6by77j+1hftp5NhzdRWl3Krqpd1Pvq8ZpeMh2Z9HD1oG9GXyb2ncjwnsMZmjOUUbmjcNlc8fmbhBBCCNEqkmzFYneDN06369nzMZRthpmPQ5zGQw3MHsjA7IFcOfzKuOxPCCGEEIkht+uJxZEWv2Rr9eOQkQdjr4nP/oQQQgjRaUiyFUu8kq2KYtj5AUy8+eTFUoUQQgjRbUiyFYsjHby17d/Pmv8Gmwsm3tT+fQkhhBCi05FkKxZHWvsHyNccga9eDFzENKN3fOISQgghRKciyVYsjvT2n0b8ovUXMRVCCCFE1yTJVix2d/tOI/q98PnTgUs9tOEipkIIIYToWiTZiqW9A+S3L4MTB+Gcn8YvJiGEEEJ0OpJsxRIcIK/1qT1+7QLIGQjDZ8Q3LiGEEEJ0KpJsxeJIC8xP5ZY9hzbDvlWByz0YtvjGJYQQQohORZKtWBzpgfmpjNv6fCHY02DC9fGNSQghhBCdjiRbsQR7tto6bquuEja+DOOugfTc+MclhBBCiE5Fkq1YQj1bbUy2Nvwt0Bs2cW78YxJCCCFEpyPJViwOd2DeltOIphk4hThgMhSMS0xcQgghhOhUJNmK5VROI+5aDpV7YdK8hIQkhBBCiM5Hkq1YgqcRfW1IttYugMy+MPryxMQkhBBCiE5Hkq1Y2tqzdeRr2PUhFP0EbI7ExSWEEEKITkWSrVjaeumHtQvBcMDZ/5KwkIQQQgjR+UiyFUtberYaauDLF2DMTMjKT2xcQgghhOhUJNmKxd6GZGvTK+A5BpPkcg9CCCGEaKxdyZZSKlcptVwptdOa92ymbrZSqkQp9Zf2HLPDhHq2WjiNqDWsfRLyz4D+5yQ+LiGEEEJ0Ku3t2boH+EhrPRz4yFqP5UFgZTuP13FCyVZ98/UOfAZlmwL3QVQq8XEJIYQQolNpb7I1E3jWWn4W+H60Skqps4F84IN2Hq/jGDawuVru2Vq7EFzZMO6HHROXEEIIIToVpbU+9QcrVaW1zrGWFVAZXA+rYwD/AK4HLgaKtNa3x9jfPGAeQH5+/tlLliw55dhaq7q6mszMzKjbzvvf2ZTlX8Cu4dEvUupoqGLy6ps4eNol7Bou47Xipbk2Eckj7ZKapF1Sj7RJakp0u0yfPv0LrXVRtG32lh6slPoQ6Btl06/DV7TWWikVLXO7FXhXa12iWjjNprVeACwAKCoq0tOmTWspvHZbsWIFMY/zRTaFebkUxtq+8j9A+yi88t8o7D08USF2O822iUgaaZfUJO2SeqRNUlMy26XFZEtrfXGsbUqpMqVUgdb6G6VUAVAepdpkYKpS6lYgE3Aqpaq11s2N70oNjrTYv0b0+2DdMzBkGkiiJYQQQogY2jtm6y1gjrU8B3gzsoLWerbWeoDWehBwJ7C4UyRaELiwaaxkq/jvcLwUJsrpQyGEEELE1t5k63fAt5VSOwmMx/odgFKqSCn1ZHuDS7q0HKg7Gn3b5wshuxBGXNKxMQkhhBCiU2nxNGJztNZHgIuilK8Dbo5SvghY1J5jdqj0XlC2pWn54Z2wewVc+BuwtespFEIIIUQXJ1eQb05Gb6g93LT886cC90E8a07TbUIIIYQQYSTZak56b6irDAyGDwq/D2JmXvJiE0IIIUSnIMlWczJ6B+bh47Y2viz3QRRCCCFEq0my1ZxgslVjnUrUGj5/EvLHyn0QhRBCCNEqkmw1Jz2YbFUE5gc+g7LNMEnugyiEEEKI1pFkqznBnq3gIPm1C8HVA8Zek7yYhBBCCNGpSLLVnFDP1hGoLoetb8L4WeDMSG5cQgghhOg0JNlqTnouoAI9W+ufBdMLE29KdlRCCCGE6ETkipzNMWyBhKtqP+z6EIZeKPdBFEIIIUSbSLLVkvTegcs9aD9M+9dkRyOEEEKITkZOI7Zk/HWQ1hNO/wH0n5TsaIQQQgjRyUjPVkum/CIwCSGEEEKcAunZEkIIIYRIIEm2hBBCCCESSJItIYQQQogEkmRLCCGEECKBJNkSQgghhEggSbaEEEIIIRJIki0hhBBCiASSZEsIIYQQIoEk2RJCCCGESCBJtoQQQgghEkiSLSGEEEKIBJJkSwghhBAigSTZEkIIIYRIIKW1TnYMUSmlKoB9HXCo3sDhDjiOaD1pk9Qk7ZKapF1Sj7RJakp0uwzUWveJtiFlk62OopRap7UuSnYc4iRpk9Qk7ZKapF1Sj7RJakpmu8hpRCGEEEKIBJJkSwghhBAigSTZggXJDkA0IW2SmqRdUpO0S+qRNklNSWuXbj9mSwghhBAikaRnSwghhBAigbptsqWUukQptUMptUspdU+y4+lOlFJPK6XKlVKbw8pylVLLlVI7rXlPq1wppf5stdNGpdRZyYu861JK9VdK/VMptVUptUUp9XOrXNoliZRSbqXUWqXUV1a7/D+rfLBS6jPr+X9JKeW0yl3W+i5r+6Bkxt/VKaVsSqkNSql3rHVplyRSSu1VSm1SSn2plFpnlaXEe1i3TLaUUjbgceBSYAxwnVJqTHKj6lYWAZdElN0DfKS1Hg58ZK1DoI2GW9M84K8dFGN34wN+qbUeA5wL3Ga9JqRdkssDXKi1PhMYD1yilDoX+D3wX1rrYUAlcJNV/yag0ir/L6ueSJyfA9vC1qVdkm+61np82CUeUuI9rFsmW8AkYJfWerfWugFYAsxMckzdhtZ6JXA0ongm8Ky1/Czw/bDyxTpgDZCjlCromEi7D631N1rr9dbyCQIfIP2Qdkkq6/mttlYd1qSBC4GlVnlkuwTbaylwkVJKdVC43YpSqhD4LvCkta6QdklFKfEe1l2TrX7AgbD1EqtMJE++1voba/kQkG8tS1t1MOsUxwTgM6Rdks46VfUlUA4sB74GqrTWPqtK+HMfahdr+zGgV8dG3G08CvwKMK31Xki7JJsGPlBKfaGUmmeVpcR7mD1ROxbiVGmttVJKfiabBEqpTOBV4P9qrY+Hf/mWdkkOrbUfGK+UygFeB0YlOaRuTyn1PaBca/2FUmpasuMRIVO01qVKqTxguVJqe/jGZL6HddeerVKgf9h6oVUmkqcs2IVrzcutcmmrDqKUchBItP6mtX7NKpZ2SRFa6yrgn8BkAqc8gl+Ww5/7ULtY23sARzo41O7gPOAKpdReAsNQLgT+hLRLUmmtS615OYEvJpNIkfew7ppsfQ4Mt3454gSuBd5Kckzd3VvAHGt5DvBmWPkN1i9HzgWOhXUJizixxo88BWzTWj8StknaJYmUUn2sHi2UUmnAtwmMp/sncLVVLbJdgu11NfAPLRdTjDut9b9qrQu11oMIfH78Q2s9G2mXpFFKZSilsoLLwAxgMynyHtZtL2qqlLqMwDl3G/C01vqhJIfUbSilXgSmEbgDexnwW+AN4GVgALAP+KHW+qiVBPyFwK8Xa4EbtdbrkhF3V6aUmgJ8Amzi5BiUewmM25J2SRKl1DgCg3ptBL4cv6y1fkApNYRAj0ousAG4XmvtUUq5gecIjLk7Clyrtd6dnOi7B+s04p1a6+9JuySP9dy/bq3agRe01g8ppXqRAu9h3TbZEkIIIYToCN31NKIQQgghRIeQZEsIIYQQIoEk2RJCCCGESCBJtoQQQgghEkiSLSGEEEKIBJJkSwghhBAigSTZEkIIIYRIIEm2hBBCCCES6P8Dk6cLOSd1KvUAAAAASUVORK5CYII=\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "labels = list(diabetes.columns)\n",
        "labels.pop()\n",
        "#labels = [\"Theta0\"] + labels\n",
        "fig, (ax1, ax2, ax3) = plt.subplots(3,1)\n",
        "fig.suptitle(\"Convergence des thetas\")\n",
        "fig.set_figwidth(10)\n",
        "fig.set_figheight(20)\n",
        "for i in range(len(labels)): \n",
        "    ax1.plot(range(MAX_IT), coefs[:, i], label=labels[i])\n",
        "    ax3.plot(range(MAX_IT), coefs_lasso[:, i], label=labels[i])\n",
        "    ax2.plot(range(MAX_IT), coefs_ridge[:, i], label=labels[i])\n",
        "ax1.plot(range(MAX_IT), coefs0, label=\"Theta0\")\n",
        "ax2.plot(range(MAX_IT), coefs0_ridge, label=\"Theta0\")\n",
        "ax3.plot(range(MAX_IT), coefs0_lasso, label=\"Theta0\")\n",
        "ax1.legend()\n",
        "ax1.set_title(\"Regression\")\n",
        "ax1.grid()\n",
        "ax3.legend()\n",
        "ax3.set_title(\"Regression+L1\")\n",
        "ax3.grid()\n",
        "ax2.legend()\n",
        "ax2.set_title(\"Regression+L2\")\n",
        "ax2.grid()\n",
        "\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "K6IDKxoHUUAC"
      },
      "source": [
        "**TODO: Analyser les résultats**\n",
        "\n",
        "- Que remarquez-vous ?\n",
        "- Analyser la convergence des valeurs des paramètres entre les trois modèles (pourquoi dans les deux modèles avec régularisation, les valeurs se rapprochent plus de 0, pourquoi il y a des paramètres qui convergent plus rappidement que d'autres)\n",
        "- Pouvons-nous déduire les attributs jugés importants par le modèle ? Si oui, lesquels ? Si non, pourquoi ?\n",
        "- Est-ce que la régularisation peut être utilisée comme sélection d'attribut ? Préciser (avec argumentation)\n",
        "\n",
        "**Réponse**\n",
        "\n",
        "- - On remarque qu'en utilisant la régularisation, certains paramètres se rapprochent au 0 plus rapidement que ceux de la regression normale. L2 pousse les coefficients vers zéro (mais pas 0 ), mais la régression L1 élimine les coefficients (égalent 0). Certains parametres n'avont pas pu s'annuler vu leur importance.\n",
        "\n",
        "\n",
        "- - Les valeurs se rapprochent plus de 0 car le terme de régularisation ( λ ∑ |w| ou λ ∑ w^2 ) augmente avec l'augmentation de la magnitude des coefficients du modèle,donc la minimisation de la fonction objective a pour effet de réduire la magnitude des coefficients. Si la valeur (λ) est suffisamment élevée, certains des coefficients peuvent devenir nuls.\n",
        "\n",
        "  -  La régularisation L1, ajoute un terme de pénalité à la fonction de coût qui est proportionnel à **la valeur absolue** des coefficients du modèle. Cela a pour effet de réduire la magnitude des coefficients du modèle. Cependant la régularisation L2, ajoute un terme de pénalité mais qui est proportionnel au **carré** des coefficients du modèle. Cela a pour effet de réduire la magnitude des coefficients du modèle, mais de manière plus **progressive** que la régularisation L1.\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "- Oui, DPF, Theta 0, Pregnancies sont des attributs trés importantes. \n",
        "Il est possible de déduire l'importance des attributs en utilisant un modèle avec régularisation. Et si la force de régularisation est suffisamment élevée, certains coefficients deviennt nuls ou presque nuls, et les autres coefficients restants >0  peuvent être utilisés pour déduire l'importance des caractéristiques correspondantes.\n",
        "\n",
        "- Oui, la régularisation peut être utilisée comme méthode de sélection d'attributs dans certains modeles, et cela en eleminant les variables dont le coefficient est presque ou carrement nul. Mais la regularisation marche seulement avec les modeles ou il existe la notion des poids (coefficient), donc pour d'autre modeles comme les arberes de décision on ne peut pas utliliser la regularisation. "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "tags": [],
        "id": "f_46hma1UUAC"
      },
      "source": [
        "### II.2. Sélection d'attributs\n",
        "\n",
        "Ici, on va utiliser le même dataset précédent sur plusieurs types de sélection d'attributs.\n",
        "\n",
        "#### II.2.1. Sélection d'attributs par filtrage (Choix ANOVA)\n",
        "\n",
        "Nous avons extrait les F values ANOVA pour les différents attributs. \n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "IKsEH6_fUUAC",
        "outputId": "ff10c32b-b826-4b27-ddd4-b65376a655e1"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "from sklearn.feature_selection import f_classif\n",
        "f, a = f_classif(X_train, Y_train)\n",
        "\n",
        "labels_diabetes = list(diabetes.columns[:-1])\n",
        "x_pos = range(len(f))\n",
        "plt.figure(figsize=(15,5))\n",
        "plt.bar(labels_diabetes, f, color=\"green\")\n",
        "plt.xlabel(\"Les attributs\")\n",
        "plt.ylabel(\"Valeur F de ANOVA\")\n",
        "plt.title(\"Comparison entre les attributs de diabetes2 en terme de ANOVA\")\n",
        "\n",
        "plt.xticks(x_pos, labels_diabetes)\n",
        "\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KVia9RFDUUAD"
      },
      "source": [
        "**TODO: Analyser les résultats** \n",
        "\n",
        "- Quels sont les attributs les plus importants d'après F values ?\n",
        "- Pourquoi un attribut est considéré comme important dans ANOVA (aspect mathématique et logique : les valeurs de l'attribut et de la prédiction Y) ?\n",
        "- Pourquoi nous avons utilisé ANOVA et pas Pearson ?\n",
        "\n",
        "**Réponse**\n",
        "\n",
        "- Glucose, BMI\n",
        "- - Mathématiquement : un attribut est considéré comme important dans ANOVA si la variance entre ses classes (MSB) est grand et la variance dans chacun de ses classes (MSW) est petit.\n",
        "  - Logiquement : les individus de la meme classes sont tres proches et trés loin des individus des autres classes  \n",
        "- On a utilisé ANOVA parce que c'est probleme de classification, si c'était un probleme de regression, on aurait du utilisé pearson"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "U9sew7MAUUAD"
      },
      "source": [
        "#### II.2.2. Sélection d'attributs par filtrage (Convergence et performance)\n",
        "\n",
        "Ici, on va utiliser ANOVA pour pré-sélectionner les caractéristiques qui donnent 50% de variance. On veut comparer la convergence du modèle normal contre celui avec pré-sélection ANOVA.\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 663
        },
        "id": "d19HbcCvUUAD",
        "outputId": "bf86890f-1276-4c12-fa87-193c245779d4"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1440x720 with 2 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABIcAAAKGCAYAAADQ/fsEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeXxU5dn/8e+VEAgEJOzKHhSrsggFFWtRqo9b3a1o0brU/bG26lPrU5dWW6t2ccGF1seCuwX3BfVnXbFuUMWCImFT2QUFhATCluT+/XHPyUyWSSbJZJacz/v1mteZ5Zwz90wCOfOd676OOecEAAAAAACAcMpJ9wAAAAAAAACQPoRDAAAAAAAAIUY4BAAAAAAAEGKEQwAAAAAAACFGOAQAAAAAABBihEMAAAAAAAAhRjgEAABCycyWmtl/pfD5nJntkeR9bjazQcncZyOff6yZLUzX8wMAgOQgHAIAoAVEgoetkQ/vwaV35LH7zGyhmVWa2TlpHiqymHOuo3PuC0kyswfN7A8t+Xw1Ay7n3DvOue+05HMCAICWRzgEAEDLOS7y4T24rI7cP1fSJZI+TuPYJElm1iaMz91Y2TTWpgrDawQAAHUjHAIAIMWcc5Occ29I2tbQumb2QzObb2alZrbKzK6MeewEM5tjZiVm9rmZHRW5v7eZvWBmG8xsiZldELPNDWb2lJk9amYlks4xs85mNsXMvoo8xx/MLDfOeILtH4+M6WMz2zfm8V9HxlIaGfdJMY+dY2bvmdkdZrZe0g1x9v+EmT0c2cdnZjY65vG9zWyGmW2MPHZ8zGMPmtlfzez/RSq13jOzXc1sopl9a2YLzGxkjafcLzLOb83sATPLj+xrnJmtNLP/NbM1kh4ws5yY17c+Ms6u9fzsfhV5T1eb2bk1HmtnZrea2XIzW2tm95pZ+8hj3c3sxchr3GBm75hZncdsQSWPmV0o6QxJV0Ve+/TI473N7Gkz+8bMvjSzX9Txs4z9XdjfzD6IPPdXZnaPmbWNrP+vyKZzI89xWvA+NeLnM8nMXor8bGeZ2e6Rxyzye/F15Pf5UzMbGu+9BQAAyUU4BABAZpsi6SLnXCdJQyW9KUlmtr+khyX9SlKhpIMlLY1sM03SSkm9JZ0i6WYzOzRmnydIeiqy3WOSHpRULmkPSSMlHSHp/HrGdIKkJyV1lfQPSc+ZWV7ksc8ljZXUWdLvJD1qZrvFbHuApC8k9ZJ0U5z9Hx95DYWSXpB0T+Q150maLulVST0l/VzSY2YWO63pVEnXSeouabukD+QrtLpHXvPtNZ7rDElHStpd0p6RbQO7Rl7jAEkXRp7vREmHyL+330qaVNcLMB/UXSnpcEmDJdXsbfTHyPONkH/f+0j6beSxX8r//HrIv0/XSHJ1PU/AOXef/M/yz5EqteMigdJ0+Uq1PpIOk3S5mR0Zs2nN34UKSVfIv18HRra5JPIcB0e22TfyHI/XeM2J/Hx+LP970UXSEkV/B46Q/x3eU/5351RJ6+t7zQAAIHkIhwAAaDnPRSooNprZc03cx05J+5jZLs65b51zwVS08yTd75x7zTlX6Zxb5ZxbYGb9JB0k6X+dc9ucc3MkTZZ0Vsw+P3DOPeecq5S0i6QfSrrcObfFOfe1pDvkP8THM9s595Rzbqd82JIvaYwkOeeedM6tjozpcUmLJe0fs+1q59zdzrly59zWOPt/1zn3snOuQtIjkoLKpDGSOkr6o3Nuh3PuTUkvSpoQs+2zzrnZzrltkp6VtM0593BkX4/Lh1+x7nHOrXDObZAPKmL3VSnpeufc9shYL5Z0rXNupXNuu3zl0ylW93SsUyU94Jyb55zbopgqKTMz+bDpCufcBudcqaSbFX3Pd0raTdIA59zOSF+fesOhOPaT1MM59/vI+/WFpL+r+s+26nfBObc18t7NjPx8lkr6P/kwLBGJ/nz+7Zwrlw+jRsS85k6S9pJkzrli59xXTXjNAACgCQiHAABoOSc65wojlxObuI8fyYc3y8zsbTM7MHJ/P/kqnZp6SwoCh8Ay+cqRwIqY6wMk5Un6Kgiy5AOBnvWMqWr7SMAUVCnJzM4yP9Ut2NdQ+SqUup47njUx18sk5UcCmN6SVkSeM95rWxtzfWsdtzvGey2RffWOuf1NJGQKDJD0bMxrK5avtOlVx2voXce+Az0kdZA0O2Zfr0Tul6S/yFfVvGpmX5jZr+vYfyIGSOodE1BulK9Cih1vtZ+Hme0ZmdK2JjLV7GZV//nVJ5GfT82fbUdJigRJ98hXYn1tvmn7Lgk+LwAAaCbCIQAAMphz7kPn3AnyYc1zkp6IPLRCfipUTasldTWzTjH39Ze0Kna3MddXyE+/6h4TZO3inBtSz7D6BVciU5f6SlptZgPkK1MuldTNOVcoaZ4ki/PcjbVaUr8a/XdqvrbG6hdzvX/kOQI1x7pC0tEx71Ohcy7fOVfX839Vx74D6+SDqiEx++nsnAuCklLn3C+dc4Pkp9j9j5kdlsBrqWu8X9YYbyfn3A/r2eZvkhZIGuyc20U+TDIlplk/H+fcXc65UZL2kZ9e9qsEnxcAADQT4RAAAClmZm0jjY9NUp6Z5dfVcDiy3hlm1jkyhatEfqqT5HsR/dTMDos0Su5jZns551ZIel/SLZH9DpefgvZoXWOJTN15VdJtZrZLZF+7m1l9U4lGmdnJkWqey+XDpZmSCuTDhm8i4/+pfOVQssySrza5yszyzGycpOPk+xM11c/MrK/5xtLXyk89i+deSTdFQjCZWQ8zOyHOuk/IN3jex8w6SLo+eCBSWfN3SXeYWc/IvvoEvYDM7FjzTaZN0ib56qTKWs9Q21pJg2Ju/1tSqfmm2u3NLNfMhprZfvXso5P879lmM9tL0n838ByxmvzzMbP9zOyASN+iLfLN2hN5zQAAIAkIhwAASL1X5StHvifpvsj1g+Ose6akpZEpPhfLN1CWc+7fkn4q3x9ok6S35acRSb7Hy0D5So5n5fvmvF7PeM6S1FbSfPkmy0/J97yJ53lJp0XWPVPSyZHeOPMl3SbfBHqtpGGS3qtnP43inNshHzYcLV9981dJZznnFjRjt/+Q/3l8IT9N7w/1rHunfIPsV82sVD4QOyDOWP+fpInyDcSXRJax/jdy/8zIz/Z1SUHj5sGR25vl38u/OufeSuC1TJHvT7XRzJ6L9Fk6Vr6vz5fy79lk+YbP8Vwp6XRJpfIBVs2w7AZJD0We49Qar7k5P59dIs/3rfxUtPXy0+sAAEAKWNP6GwIAgDAysxsk7eGc+0m6xwIAAIDkoHIIAAAAAAAgxAiHAAAAAAAAQoxpZQAAAAAAACFG5RAAAAAAAECIEQ4BAAAAAACEGOEQAAAAAABAiBEOAQAAAAAAhBjhEAAAAAAAQIgRDgEAAAAAAIQY4RAAAAAAAECIEQ4BAAAAAACEGOEQAAAAAABAiBEOAQAAAAAAhBjhEAAAAAAAQIgRDgEAAAAAAIQY4RAAAAAAAECIEQ4BAAAAAACEGOEQAAAAAABAiBEOAQAAAAAAhBjhEAAAAAAAQIgRDgEAAAAAAIQY4RAAAAAAAECIEQ4BAAAAAACEGOEQAAAAAABAiBEOAQAAAAAAhBjhEAAAAAAAQIgRDgEAAAAAAIQY4RAAAAAAAECIEQ4BAAAAAACEGOEQAAAAAABAiBEOAQAAAAAAhBjhEAAAAAAAQIgRDgEAAAAAAIQY4RAAAAAAAECIEQ4BAAAAAACEGOEQAAAAAABAiBEOAQAAAAAAhBjhEAAAAAAAQIgRDgEAAAAAAIQY4RAAAAAAAECIEQ4BAAAAAACEGOEQAAAAAABAiBEOAQAAAAAAhBjhEAAAAAAAQIgRDgEAAAAAAIQY4RCAFmVm15jZ5HSPAwAAAABQN8IhIATMbKmZbTWzzTGXexLYboaZnd+c53bO3eyca9Y+Us3MzjGzd9M9DgAAgMaocaxXWeP474wm7K/Zx4IAskObdA8AQMoc55x7PZk7NLM2zrnyZO4TAAAATeOc6xhcN7Olks5P9vEfgNaJyiEgxIIKGTO71cy+NbMvzezoyGM3SRor6Z7YSiMzc2b2MzNbLGlx5L47zWyFmZWY2WwzGxvzHDeY2aOR6wMj259tZsvNbJ2ZXRuzbo6Z/drMPjez9Wb2hJl1rbHtTyPP9a2ZXWxm+5nZJ2a2sWY1lJmda2bFkXX/aWYDYh5zke0XR7adZN7eku6VdGDkdW9sobcfAAAgJRo4xso3s0cj9280sw/NrFe8Y0EArRPhEIADJC2U1F3SnyVNMTNzzl0r6R1JlzrnOjrnLo3Z5sTIdvtEbn8oaYSkrpL+IelJM8uv5zm/L+k7kg6T9NtIICNJP4/s+xBJvSV9K2lSHeMdLOk0SRMlXSvpvyQNkXSqmR0iSWZ2gqRrJJ0sqUfktUytsa9jJe0nabikUyUd6ZwrlnSxpA8ir7uwntcBAACQDeo7xjpbUmdJ/SR1kz8O2trAsSCAVoZwCAiP5yLfBgWXCyL3L3PO/d05VyHpIUm7SerVwL5ucc5tcM5tlSTn3KPOufXOuXLn3G2S2smHP/H8zjm31Tk3V9JcSftG7r9Y0rXOuZXOue2SbpB0ipnFToG90Tm3zTn3qqQtkqY65752zq2SP4AZGbOvW5xzxZGpbzdLGhFbPSTpj865jc655ZLekg+4AAAAWpv6jrF2yodCezjnKpxzs51zJWkcK4A0IBwCwuNE51xhzOXvkfvXBCs458oiVzvW3ryaFbE3zOzKyPStTZFpWJ3lK5HiWRNzvSzm+QZIejYIsCQVS6pQ9bBqbcz1rXXcjt3XnTH72iDJJPVJYBwAAACtSX3HWI9I+qekaWa22sz+bGZ5aRwrgDQgHAJQH9fQ/ZH+QlfJT8vqEpmGtUk+iGmsFZKOrhFi5Ueqgpqyr4tq7Ku9c+79BLaN97oBAACyUdxjLOfcTufc75xz+0j6nvy0+7Mi23FMBIQE4RCA+qyVNKiBdTpJKpf0jaQ2ZvZbSbs08fnulXRTMPXLzHpEegc1dV9Xm9mQyL46m9n4BLddK6mvmbVt4nMDAABkkrjHWGb2AzMbZma5kkrkp5lVRrZL5FgQQCtAOASEx/TImSaCy7MJbHOn/Hz0b83srjjr/FPSK5IWSVomaZtqTDtrhDslvSDpVTMrlTRTvgF1oznnnpX0J/kS6RJJ8yQdneDmb0r6TNIaM1vXlOcHAADIIPUdY+0q6Sn5YKhY0tvyU82C7Ro6FgTQCphzVAoCAAAAAACEFZVDAAAAAAAAIUY4BAAAAAAAEGKEQwAAAAAAACFGOAQAAAAAABBibdI9gJq6d+/uBg4c2GL737JliwoKClps/6iO9zu1eL9Tj/c8tXi/U6+l3vPZs2evc871SPqO0WQteQzGv93U4z1PLd7v1OL9Tj3e89RKx/FXxoVDAwcO1EcffdRi+58xY4bGjRvXYvtHdbzfqcX7nXq856nF+516LfWem9mypO8UzdKSx2D820093vPU4v1OLd7v1OM9T610HH8xrQwAAAAAACDECIcAAAAAAABCjHAIAAAAAAAgxDKu5xAAAC1l586dWrlypbZt25buoWSNzp07q7i4uMnb5+fnq2/fvsrLy0viqAAAQDbhGKxx0nH8RTgEAAiNlStXqlOnTho4cKDMLN3DyQqlpaXq1KlTk7Z1zmn9+vVauXKlioqKkjwyAACQLTgGa5x0HH8xrQwAEBrbtm1Tt27dOChJETNTt27d+JYQAICQ4xgsdZp6/EU4BAAIFQ5KUov3GwAASBwTpFJT3mvCIQAAAAAAgBAjHAIAIIVyc3M1YsQIDR06VMcdd5w2btyY7iEBAAC0ehyD1Y9wCACAFGrfvr3mzJmjefPmqWvXrpo0aVK6hwQAANDqcQxWP8IhAADS5MADD9SqVaskSZ9//rmOOuoojRo1SmPHjtWCBQuq7h8zZoyGDRum6667Th07dkznkAEAALIex2C1cSp7AEAoXX65NGdOcvc5YoQ0cWJi61ZUVOiNN97QeeedJ0m68MILde+992rw4MGaNWuWLrnkEr355pu67LLLdNlll2nChAm69957kztgAACAFOMYLDMRDgEAkEJbt27ViBEjtGrVKu299946/PDDtXnzZr3//vsaP3581Xrbt2+XJH3wwQd67rnnJEmnn366rrzyyrSMGwAAIJtxDFY/wiEAQCgl+u1SsgXz3cvKynTkkUdq0qRJOuecc1RYWKg5yf4aDQAAIMNwDJaZ6DkEAEAadOjQQXfddZduu+02dejQQUVFRXryySclSc45zZ07V5I0ZswYPf3005KkadOmpW28AAAArQHHYHUjHAIAIE1Gjhyp4cOHa+rUqXrsscc0ZcoU7bvvvhoyZIief/55SdLEiRN1++23a/jw4VqyZIk6d+6c5lEDAABkN47BamNaGQAAKbR58+Zqt6dPn151/ZVXXqm1fp8+fTRz5kyZmaZNm6aFCxe2+BgBAABaG47B6kc4BABABps9e7YuvfRSOedUWFio+++/P91DAgAAaPXCdgxGOAQAQAYbO3Zs1dx3AAAApEbYjsHoOQQAAAAAABBihEMAAAAZysyOMrOFZrbEzH5dx+PnmNk3ZjYncjk/5rGzzWxx5HJ2akcOAACyCdPKAAAAMpCZ5UqaJOlwSSslfWhmLzjn5tdY9XHn3KU1tu0q6XpJoyU5SbMj236bgqEDAIAsE5pwaN066b33pE8/7aZNm6L3V1ZKTz0lfe970s9+lr7xAQAA1LC/pCXOuS8kycymSTpBUs1wqC5HSnrNObchsu1rko6SNLWFxhrXnXdKt9++nzp0qH5/To70pz9Jxx6b6hEBAICaQhMOzZ8vnXiiJA2r8/HiYsIhAEDLy83N1bBhw1ReXq6ioiI98sgjKiwsTMq+161bp91220133323Lr744qr7Bw4cqFGjRunpp5+WJD311FN68cUX9eCDD0qSnnvuOf32t7/Vzp071aZNG91444068cQT9bvf/U4lJSW67bbbqvY1Z84cTZgwQcXFxZKkESNGaK+99tK0adOS8hpQTR9JK2Jur5R0QB3r/cjMDpa0SNIVzrkVcbbtU3NDM7tQ0oWS1KtXL82YMSM5I4+xfn1P9e9fqDZttlS7/513uuuhh1apY8fPk/6c8KdsbomfJ+rG+51avN+p19z3vHPnziotLU3egJqgsLBQQ4YMUXl5uQYMGKD77rsvacdg69ev1+DBg/WXv/xF5513XtX9Q4cO1YgRI/Too49K8sdcr7zyiu69915J0osvvqibbrqp6hjsuuuu07HHHqubb75ZO3bs0A033FC1r08++UTnnnuuPvroI0nSQQcdpMGDB1cdz9W0bdu2Rv3MQhMOffe70scfSx999JFGjx5d7bGf/ERK0u8EAAD1at++vebMmSNJOvvsszVp0iRde+21Sdn3k08+qTFjxmjq1KnVwiHJn451/vz52meffardP3fuXF155ZV67bXXVFRUpC+//FKHH364Bg0apAkTJujII4+sFg5NmzZNEyZMkCQVFxeroqJC77zzjrZs2aKCgoKkvA40ynRJU51z283sIkkPSTo00Y2dc/dJuk+SRo8e7caNG5f0AY4bJ82YMUM1973rrlKXLv00bly/pD8n6n7P0XJ4v1OL9zv1mvueFxcXq1OnTskbUBO0b99en3zyiSR/DPbwww8n7Rjs0Ucf1ZgxY/Tss8/q8ssvr7rfzDR37lytWLFC++yzj9q3b6+8vDx16tRJc+fO1W9+85tax2BDhgzR+PHjdcopp1Q7Bps+fbrOOOMMderUScXFxXLOaebMmcrJyanzGCw/P18jR45M+DWEpiF1x47SyJHS4MGbNXKkql169PDTywAASKUDDzxQq1atkiR9/vnnOuqoozRq1CiNHTtWCxYsqLp/zJgxGjZsmK677jp17Ngx7v6mTp2q2267TatWrdLKlSurPfbLX/5SN910U61tbr31Vl1zzTUqKiqSJBUVFenqq6/WX/7yF+25554qLCzUrFmzqtZ/4oknqsKhqVOn6swzz9QRRxyh559/vnlvBuqySlJsctI3cl8V59x659z2yM3JkkYlum26FRRIW7Y0vB4AAMmW6cdggwcPVpcuXVJ6DBaayqH65ORIO3emexQAgJS6/HIpUsGTNCNGSBMnJrRqRUWF3njjjarS4wsvvFD33nuvBg8erFmzZumSSy7Rm2++qcsuu0yXXXaZJkyYUFWCXJcVK1boq6++0v77769TTz1Vjz/+uH75y19WPX7qqafqr3/9q5YsWVJtu88++0xXXnlltftGjx6tSZMmSZJOOeUUTZs2TQcccIBmzpyprl27avDgwZKkxx9/XK+99poWLFigu+++W6effnpCrx0J+1DSYDMrkg92fiyp2ptsZrs5576K3DxeUnHk+j8l3WxmXSK3j5B0dcsPOXGEQwAQUhyDSWr4GGzChAkpPQYLTeVQfXJzqRwCAKTG1q1bNWLECO26665au3atDj/8cG3evFnvv/++xo8frxEjRuiiiy7SV1/5z/sffPCBxo8fL0n1/uF//PHHdeqpp0qSfvzjH2vq1Op9h3Nzc/WrX/1Kt9xyS6PGe/LJJ+upp55SZWVltSllH330kbp3767+/fvrsMMO03/+8x9t2LChUftG/Zxz5ZIulQ96iiU94Zz7zMx+b2bHR1b7hZl9ZmZzJf1C0jmRbTdIulE+YPpQ0u+D5tSZgnAIAJBK2XYMdtppp6X0GIzKIfnKoYqKdI8CAJBSCX67lGxBz6GysjIdeeSRmjRpks455xwVFhZW9SJKxLXXXquXXnpJkm8SPXXqVK1Zs0aPPfaYJGn16tVavHhx1TdMknTmmWfqlltu0dChQ6vu22effTR79mztu+++VffNnj1bQ4YMkST17dtXRUVFevvtt/X000/rgw8+kOTLmRcsWKCBAwdKkkpKSvT000/rggsuaNobgzo5516W9HKN+34bc/1qxakIcs7dL+n+Fh1gMxAOAUBIcQwmqeFjsH79+qX0GIzKIfnKIcIhAEAqdejQQXfddZduu+02dejQQUVFRXryySclSc45zZ07V5I0ZsyYqrOMxZ4R7KabbtKcOXM0Z84cLVq0SJs3b9aqVau0dOlSLV26VFdffXWtb67y8vJ0xRVX6I477qi678orr9Qtt9yipUuXSpKWLl2qm2++uVo59IQJE3TFFVdo0KBB6tu3ryorK/XEE0/o008/rXq+559/vtbzAfXp2JFwCACQehyD1Y1wSL5yiGllAIBUGzlypIYPH66pU6fqscce05QpU7TvvvtqyJAhVc0FJ06cqNtvv13Dhw/XkiVL1Llz51r7mTp1qk466aRq9/3oRz+q80DhvPPOU3l5edXtESNG6E9/+pOOO+447bXXXjruuOP05z//WSNGjKhaZ/z48frss8+qypnfeecd9enTR717965a5+CDD9b8+fOrSrGBhlA5BABIF47BamNamagcAgCkzubNm6vdnj59etX1V155pdb6ffr00cyZM2VmmjZtmhYuXFhrneuvv77WfcOHD1dxse9NHHwjJUnt2rXT6tWrq6178skn6+STT4475u7du2tnzJkbDjnkEM2cObPaOrm5uVqzZk3cfQA1EQ4BAFKJY7D6EQ6JhtQAgMw1e/ZsXXrppXLOqbCwUPffn7EtZIBGIRwCAGSysB2DEQ6JhtQAgMw1duzYqrnvQGsShEPOSWbpHg0AANWF7RiMnkOicggAwsQ5l+4hhArvN+IpKPDHX9u3p3skAIBU4JggdZryXhMOicohAAiL/Px8rV+/noOTFHHOaf369crPz0/3UJCBCgr8kqllAND6cQyWOk09/mJamagcAoCw6Nu3r1auXKlvvvkm3UPJGtu2bWtWuJOfn6++ffsmcURoLWLDoW7d0jsWAEDL4hiscdJx/EU4JCqHACAs8vLyVFRUlO5hZJUZM2Zo5MiR6R4GWiEqhwAgPDgGa5x0HH8RDonKIQBItyuvlP7f/2t4vS1b9qv6QBnIzZXuuksaN65FhgaghRAOAQCQOQiHROUQAKTTN99IEydKw4dLu+/e0Lpb1KNH9XToueekl18mHAKyDeEQAACZg3BIVA4BQKI2bZI2bEjuPp980gf0Dz7oA6L6zJgxX+PG9ax23557SsuXJ3dMAFoe4RAAAJmDcEhUDgHIHu+/L112mVRentj6Xbr4yppddmn+c2/ZIhUVSd9+2/x91bT33tKwYU3btn9/admy5I4HQMsLwqHNm9M7DgAAQDgkyVcOEQ4ByAYTJ0qLFiU2hWr9eumtt6Q5c6SDD27+c0+f7oOhG2+U+vVr/v5iHXCAZNa0bfv3l/75z+SOB0DLo3IIAIDMQTgkXznEtDIAqfLVV1JJSfX7uneXOnaUli6Nv9327T6gOfdcadKkhp9n4UJpr72aV1Uzb570+uv++tSpUp8+0jXX+P83M8WAAf493bFDats23aMBkCjCIQAAMgfhkKgcApA6S5b4KVQ1p4UVFkpjx/rwpyGnn57Yc/Xv75fN6cdz1lnSf/4Tvf2b32RWMCT51+mctHKlNGhQukcDIFHt2/vltm3pHQcAACAckkRDagCp849/+DB6ypToB6Nly6Srr/bB0EknSePHx9++SxfpoIMSe6727aUePZpeObRwoQ+G/vhH6aKL/LSvZPQuSrYBA/xyyRJf2VSfdu1afjwAEpOb65d8QQcArdDvfuebZaJJ2p5/fsqfk3BINKQG0Hxbtkh33SVt3Vr/eg895Pv/nHtu9L7KSj9NbOVK6aqrpDFjkjeuAQOkRx6RFiyQRo6U7rwz+tjll0sffRS93aGDdMUV0p//LO3cKX39tQ+EzjzTVzZlqoED/fLIIxte9ze/kX7/+xYdDoAEBVWIfEEHAK3QxIlSfn70QA2N0sRWnM1COCQqhwA038sv+148Uv2NlXNzffgSKydH+vnPfeXQAQckd1wDBvgA6J13/OXWW6W8PB/+3HWXn4Y1cKCflvXaa9J77/nr3/uen6512mlS797JHVOyFRVJ998vrVlT/3r/93/SBx+kZkwAGhZUDnEMBgCtUFmZdPHF0i23pHskWWn7jOpu88sAACAASURBVBkpf07CIVE5BKD5gkbSmzY1berVVVf5S7IFDV+7d5fWrZNWrfJh0KpVPgS6+mrpvPP8OmPHSu++K519tvTgg8kfS0sxk37604bX+/hj6ZNP4j8+d640ebJ/X+rStq30619LPXs2bZwAqgsqhzgGA4BWprzcnymkQ4d0jwSNQDgkKocANN+yZX7qVab15OnVyy8vv1y67jo/zoEDo32IgqbVkm90/e67iTe8zjYDBkgvvujDn7qqu665Rnr1Valz57q3X7/e/3xvuKFFhwmERrVpZatWST/5ifTkkz7NBgBkr6DPAuFQVsmwc86kR3Aq+3jfFgNAQ5Ytqx60ZIrrr5fefDPa5Do4c1mwDJo5S9L55/vpcYcfntoxpkr//v6sSOvW1X5swwYfDF1xhX+8rsu4cdLjj/O3AkgWM3+prJQ0a5Y0Y4b06afpHhYAoLnKyvyScCirEA4pOuedA34ATbVsWfWgJVMUFEg/+IHUr5+/HVQMBcvgfsn3Ijr66Pp7JmWzILwLgrGlS6WhQ33PoiFDfAX0aafF3/6003xj74EDfdPw4LgHQNNVTe3fsMHfsXlzWscDAEgCwqGsxLQyVZ/znkNcBqAJli2TDjkk3aOIr3173ysntnKoZ09/f1jEhkOjRkkPPCAVF0tnnOEDsYEDpe9+N/72Z5zhixoWLpTeeMMHRfWtD6BhVVP7g3Boy5a0jgcAkASEQ1mJcEicLQNA82zaJJWUZGblUKwBA3yI5VzmToNrScHrnTVL2n13aepUP1Xs4YcT275TJ2nSJGn2bGn0aB8yEQ4BzRNM7adyCABakSDoJxzKKtTJiLNlQFqyxPfBvOqq6PTCqVOl559P77iQuSorpRNPlM45Zz+NHu3vy/SwZcAA31cnJ8cvBw5M94hSq1s33zT8T3+S9t1XWrxYmjCh8fsJfs4rViR3fEAYVU0rW7/e30HlEABkPyqHshKVQ4pWDhEOhdftt0uPPeavn3iirwYIzthELyrU5Z13fHg4YsQO7bFHgb73Pemww9I9qvpdf73vsRM4+eT0jSUdzKS33/ZhsCS1aycdeWTj99O9u5SfH52iB6Dpak0ro3IIALIf4VBWIhxSjVOpInR27PBnIDrmGH9WpwcekD7/PN2jQqps2OArSMykYcOq9+CpqJDmzpV27qy93d/+5ps933zzpzr66INTN+BmGDq0ejgURsOH+0tzmPlG3qkIh95+W7r33rpD6rZtpd//PnwVYGhdak0ro3IIALIf4VBWIhwSlUNh969/+WPSiy6SOnaUJk/2F8lXB6B1O+UU6a23/PX//m/pr3+NPvb3v/v74jnrLKl9e1LlMOrfPzXh0DXX+ICyb9/q9zsnLVokjRwpXXFFy48DaCm1zlZGOAQA2S8IhwoK0jsONArhkGhIHXZr1vjlXntJt90mff/7/oPXW29Jzz7rj1P5f631Ki6Onr596lTpjjv8dCNJeughae+9/e9FTWbSgQdK//lPaseLzNC/v/TPf7bscyxfLr3/vnTTTT4kiuWcD7Ppe4Rsx7QyAGiFqBzKSoRDoiF12JWU+GWnTtKuu0qXXupvd+zow6G1a6VBg2pvt3ChDw1if2+OPVY66aSWHzOSY9s2Hw4eeKA/+9TLL0vjx0s9evif68yZvnnx0Uene6TINAMGSKtXV5+GmGzB/y2nnVb7MbPUVS8BLYlpZQDQChEOZSXCIVE5FHalpX65yy7V7991V7+MFw7ddJOvNAnW27RJeukl6bjjpDb8y8oKQdVF//7S4YdLBx9cvRJon32kM89Mz9iQ2c491/crKy9v2efZYw9p993rfqxfPyqHkP1yciTbvi36QYLKIQDIfsH/6S35LRqSjo+wonIo7EpL/e9Azf+7evXyy7VrfXAY25R461bpmWeks8+O9id65hnpRz+SXnlFOuII3ywWmW3ZMr8cMMAHem+/nd7xIHv06+cD4nTq398H0sk0b57UpYvUp09y9wvEk5srtd+6IXoHlUMAkP22bJHy8vwFWSMn3QPIBFQOhVtJiZ9SZlb9/iAcWr3af3ufnx+9dOni/88744zo+j/8oVRY6CuH2rWTpkxJ3WtA08SGQ0C26d/fT4vcvj05+9uxw5+xb+zY5OwPSEROjtRhW0w4ROUQAGS/sjKmlGUhKodE5VDYlZb6cKimnj398r33pC+/lCZM8B+cAt27S+PGRW/n50tPPy3NmiX94Q/SnDktOmwkwfLl/t9/zTNBAdmgXz+/XLky/tSzxnj9db/88svm7wtIVE6OlLtjq79RUEDlEAC0BoRDWYlwSFQOhV1pae1+Q5KvguzaVXr1VX/7V7/yp42uz6GH+suUKdL69ckfK+q2cKH02GP+DE6HHir94Afx112xwv98Kip8A+reval4RXbq398vzz3XVzNKPiS69dbqlZC/+Y306acN72/+fL/s0SO54wTqk5sr5VRE5m137kzlEAC0BoRDWYlwSFQOhV28yiHJN6L+6CN/8Lr33onvs1u36IlX0PJuucWfdl6Spk+vv2rrvvt8ZVfw7/7001t+fEBLGDXKN1HftMlfNm+Wnn/e/06PGuXXWbTI/74PHOg/d9enQwd/lsZ163yPNUJTpEJOjqKd3QsL/VxJAEB2IxzKSoRDonIo7IKeQ3UZPz4aDuXnJ77Prl2pHEqlTz6RjjxSGjxYeuSR+tddutT3GFq6NBUjA1pOYWH1JuobNvizJ/7jH9KIEf6+qVN9FdG77ybWZHryZOmCC3yvtbw8X1kHtKScHMkqYsKhJUvSOyAAQPMRDmWlhBpSm9lRZrbQzJaY2a/rePwcM/vGzOZELufHPHa2mS2OXM5O5uCThcqhcKuvciioKon3eDzduhEOpUp5ufTZZ9Lw4X6aTVBFEc/y5dHpOEBr0rWrD0lvv92ffa9NG+mGG3yD6UTPPhb0Mfr4Yx+iTpzYYsMFJPkvX6rCoc6dfWf02NODAgCyD+FQVmqwcsjMciVNknS4pJWSPjSzF5xz82us+rhz7tIa23aVdL2k0ZKcpNmRbb9NyuiThMqhcIvXc0jyjYonT26411BNXbsyrSxVFi3ynyWGD5fatvX3LV9evXl4rGXLOBsTWq/bbpMOOMD33wqcdFLi2wfB6cSJPnj9wQ+kbzPqLzZam5wcycojYVBhoV9u2RK9DgDIPM75s/aUltb9+Jo10ne+k9oxodkSmVa2v6QlzrkvJMnMpkk6QVLNcKguR0p6zTm3IbLta5KOkjS1acNtGVQOhVt908ok6bzzGr/Prl2ljRv9h6s2TN5MuooK6YUXpK1b/bQ/yYdDZWX+erxwqKLCn9mJyiG0VnvuKV13XdO3DyqH/vUvaa+9/L+r2KlrQLLl5Eg5lTHTyiTCIQDIdP/+d8Pfth58cGrGgqRJ5GNrH0krYm6vlHRAHev9yMwOlrRI0hXOuRVxtq1V3G5mF0q6UJJ69eqlGTNmJDT4pti8eXOt/c+f303SMM2a9ZE2buQsGclU1/udSZyTSkoO0YYNyzVjRvLO37xhQx9Jg/XSS++pc+fUlcdn+vudLG++2VM33rhP1e2CgnKtXfueSkryJH1Pr722SAUFq2tt9/XX7VRRcaC2bVuoGTO+SspYwvKeZwre75bXq9cYrV2br+9//wu9/fZy3nO0qGrTymLDIQBA5lq40C+ffjr+3PWhQ1M3HiRFsmoapkua6pzbbmYXSXpI0qGJbuycu0/SfZI0evRoN27cuCQNq7YZM2ao5v6DY5Dvfne09tuvxZ46lOp6vzPJ1q1+OuGwYQM0btyApO131Srp7rulvfY6KKUVlZn+fifLrbf6KX+vv+6b7Xbv3kZdux6iyko/tSw/f0+NG7dnre3ee88vDz/8Oxo3Ljk/mLC855mC97vlLVwoffONNGjQIOXkDOI9R4uq1pA6OKUep7MHgMwWnNnlmGOkdu3SOhQkTyLh0CpJ/WJu943cV8U5F9t6d7KkP8dsO67GtjMaO8iWxrSy8AqmyTa24XRDunXzS/oONc6HH0pXXRU9q3E8H3wgXXll7anMOTl+WsyDD0aDoFhBk/ABycsBgVanc+eGT3sPJEturpSzvY6eQwCAzLV0qT+lKcFQq5LI2co+lDTYzIrMrK2kH0t6IXYFM9st5ubxkooj1/8p6Qgz62JmXSQdEbkvo9CQOrxKSvwy2eFQ165+yRnLGucvf/E9hNq2rf9yxBHSxRfXvY+f/1waMqTu7XbbTTrjDGmPPVL7ugAAdauz5xCVQwCQ2ZYulQYOTPcokGQNVg4558rN7FL5UCdX0v3Ouc/M7PeSPnLOvSDpF2Z2vKRySRsknRPZdoOZ3SgfMEnS74Pm1JmEyqHwWblS6tkzWl3SUpVDL70kHX10NIDMRmvWSL16+elbLeXzz6V166Tp06ULLpDuuqvp+7rsMn8BAGS+OqeVUTkEAJlt6VLpwAPTPQokWUI9h5xzL0t6ucZ9v425frWkq+Nse7+k+5sxxhZH5VC4lJf7qUejRkmzZ/v7dt01uc/Rs6df3nuvb+R/+unJ3X+qzJsnjRghPfqo9OMft8xzLFgg7b139PaZZ7bM8wAAMk9urpRTUWNaGZVDAJC5ysulFSukCRPSPRIkWSLTylo9KofCJZhKFgRDd94pjRmT3Ofo1MkHK1L0VOvZ6KGH/L+LKVNa7jk+/9wvb73Vnz6bpvAAEB7VppVROQQA6ffPf/pv0nv3rnY58JRT/PW+fX1AxLSyVidZZyvLalQOhUvQhDpwyiktM2VqyBAfdHzySfL3nQwrV0r33y9dc43Upo7/Cf7wB/94bq70xhu+oqeu9+n44/17WFkp3XKLr5IqKkp8HGvX+uXJJzduOwBA9svJkYyeQwCQOWbO9B8ULrig2t3rV69W7969/Y127aQTTkjD4NCSCIdE5VDYxIZDeXnJn1IWa9gw6YUXJOdatmdPUzz+uHT99dLo0dIPf1j9sa+/ln7zG3/9gQek226r++xfGzb4LxdOPFF66y3puuukOXOkJ59MfBxBONSrV9NeBwAge/lpZZFwaJdd/B9LKocAIH3KyqT8fOm++6rdvWjGDPUeNy49Y0JKMK1MVA6FTWw41K9fNBxsCcOH+0bL8+dn3hehy5f75WOP1X7s00/98vXXpXPO8be/+KL25aGHfJD0zDPSww/7bRp7Rsu1a6WOHaUOHZr8UgAAWSonR8qtjPQcysuTCgoy7w8mAIRJWRkH5iFF5ZCoHAqb2HBowICWfa4RI/xy6FAfmhQXZ87UqWXL/PK55/xxeMeO0ceCcGjYsPr3cfTR/sxsp50Wve+bbxo3jrVrqRoCgLCq1nOoTRsfDlE5BADpQzgUWlQOicqhsEllODR2rPTII9Ltt0vbt/vrmWL5cn9WtbIyHxDF+uQT/1hw1rV42raVXnpJmjRJ+tvf/OsNQqdEEQ4BQHjl5kbCITOfFFE5BADpRTgUWlQOicqhsAnOVia1fDiUkyP95Cf++vTpfuruxo3Rx484QjrqKN+T6N57pWOOkfr3b9kxBZYtk8aPl155xU8tC8Y5a5bvM3TYYYnt54AD/EWSFi/2IVFjeiytXSt95zuNHz8AIPtVVQ7l5fk7OnakcggA0olwKLSoHFK0cohwKByCyqGiIunQQ1P3vJdd5v+vnTzZX+65x/fzKS/3p7u/5BLpd79LzVg2b/bNpAcO9GcXe+21aGPo00/3y4MPbvx+BwyQtm71fZYSReUQAIRXVc+h4LSZIZpWtn27tG1bukcBADUQDoUW4ZCYVhY2QThUXNy0AKSpTjjBBzIlJf7yxBM+GHnpJV+pI/mzfK1f7/9PLivzwVFLCJpR9+8vnXGGD0Yff9wvly6VLrwwerayxgiqnhKdWrZzp3+9hEMAEE5V08qCcKhjx1BMK/vyS//Zq0MHacmSdI8GAGJs2UI4FFJMKxPTysKmtNT3ymnsWbWS7Yc/9M2cTzzR395jD3+A2L17dJ3u3f0BZGyz6Hjef18688z9q46v6xN8UzlggDRkiG+cPXWqdNJJPiQdNSrxaWGxgml6xx4rtW8v9eghvfmm9PTTvirKuerrB4Es4RAAhFNOjpTjyqtXDq1Zk95BpcDs2dG/gR995I8BACAjlJX5g3iEDuGQqBwKm9JSqVOndI/CB1TPPCPNnOmDmFNPld54Izola8kS6e9/982hv/e9hvf3+uvSypUddOaZiQU7hYXS6NH++rhx/rmCip+m9mIaNkz6n//xr2H1aj+m+fOlF16QNm3yoVFN7dpJxx3XtOcDAGQ3P62sXGoX03MoBJVDixfXfR0A0o5pZaFFOCQqh8ImU8IhyU9ri53adu650evLljUuHFq+XOradbsefrjxJVH9+/sK0jlzorebok0b6bbb/PW5c31F0vLl/rXsv7/00ENN2y8AoHXKzQ1nz6HFi6Vdd/Uvm2llADIK4VBoEQ6JyqGwKSnJnHCoPv37S7vsIn36aWLr+1PTb5fUtHBIkt55p/rt5gj2sXy5v4wa1fx9AgBal5wcKdeFr+fQkiXS4MH+GJTKIQAZhXAotGhILSqHwqa01Icumc7MT9NKNBxatkzq1atppz0JppG9+67vg1RQ0KTdVFNY6EO4BQukb75JTuAEAGhdqk5lH1s5VFbW6r+xW7zY9xkaPJhwCECGIRwKLcIhUTkUNpk0rawhw4b5nkTDhklXXRV/Ped8dU6vXtub9DxBcLN6dfJCHDO/r3ffrf4cAAAEcnOlNm6nlBfpORR8O3HTTekbVAsrLfU9twcP9pd166QLLpAuukiaPDndowMQahUV0vbthEMhRTgkKofCJpvCofPPl044wVfYT5sWf7116/wZyHr2bFrlUI8e0X8HQ4Y0aRd16t9fKi7215va5BoA0HrVOlvZAQf45W9/K5WXp29gLejrr/2yTx/p0EP938cXX5QefbT+L4IAoMVt3eqXhEOhRM8hUTkUJsXFfppTIg2eM8GoUdKTT0qXXy498ED89ZYv98umTiszi/7+n3FGk3ZRp9hAiMohAEBNtXoOHXKIP7PBL3/pG1N37hx/4//8x58Gs0cPX2abn5+aQTfTzp1+2bat/zu/dKm//atfSffck7ZhAcg0jzziT/ubiOOPlw48sPnPWVbml4RDoUQ4JCqHwmLHDn/MKUl9+6Z3LI3VubNvpF1REQ0zYy1a5JdNnVYm+X8HlZXSf/1Xk3dRS1FR9HqfPsnbLwCgdcjNrREOSdGpZZs31x8Ovf22nw+9erW0dm3WlKju2OGXbdtWv799e18F7Jz/0gaNt2aNrxCP1bNn/b9GQEbasUM65xx/vU0DH9l37JD+/W/pjTea/7yEQ6FGOCQqh8Li5Zd9Y+TJk6Wzz073aBonOKgpLfWNnmt64gl/StxBg5p++t/58/3fg4b+/jTG+ef7Y/yiomg7CQAAAjk5UpvKndX/SHTs6JcNndI+tpNzQ+tmkPrCIcm3+8iSIqiM8sUXvsm3c9Xv79cvWmENZI3ly/2H0wceiIZE8UyY4MOhZCAcCjXCIVE5FBaPPCL16uWDoWQGIKkQhEObNlUPh8rLpVtukV56SfrFL6TcXFf3DhLwne80c5B16NpV+tnPkr9fAEDrUGtamRStHGoo8FmyJHo9i8KhYFpZzS9NgkBo61bCoab45BMfDN18c3Qq+2uvSQ89JG3Y4I9JgKwRzDeNLcOPp6hIeuqp+FMMGoNwKNSy7CNyywj+DREOtW7vvScdfXT2BUNS9XAo1ocf+p6dkq/SWbMmteMCAKA5cnOlXDUjHOrZ03d4Dj7QZIGGKoe2Na19YOgFWeHFF0tduvjrnTr5cGjJEmn//dM3NqDRvvzSLwcObHjdoiL/jfGqVc1v8kk4FGqcrUxMKwuDNWt8O4IRI9I9kqaJFw4tW+aX8+ZJe+2V2jEBANBc9VYObd4cf8MdO/w368Ef9lZQORSEQ8HJgtA4n3/uq4OCYEiSdt89+hiQVZYu9f8vJtK0MwiQgkCpOQiHQo1wSEwrC4M5c/wyW8OhYCrZxo3V7w/m0HMmMABANsrNldq4RvYcck669Vb/rd6++9a/bgZqqHKIcKhpliyJhkGBQYOijwFZ5csvfcOsRKY8BFPPgqlozUE4FGpZOMEm+agcav2CcCg4hsw28SqHli/335B16pT6MQEA0Fw5OVKbxvYcmj1buvZaf33sWOkvf2kV4VDQZ4hpZU3z+efSmDHV72vf3hdeUDmEjOOc7w8Rb0rsvHmJ9RuS/LfEZtK//tX8szZ+/LFfEg6FEuGQ/IHJQXpXndf2lLRnuoeDFjB7tv+/sq4zfWWD+sIhqoYAANkqJ6cJPYcWLPDL997zp6eqb90MxLSy5Hv0UV9occYZtR/bfXcqh5CB3nlHOuSQ+te55JLE9tW2rf+/8P77/aW5cnLo4B5ShEOScr5dr3c1VqUP7ird/VW6h4MkKy31p7Gv64AhW9TXc6i5XxAAAJAuflpZjXAomFYWr+fQokX+w8vo0dGkJYvCIaaVJdfWrdKZZ/rrBx5Y+/EBA3xBBZBRvv7aL++/P36F0OjRie/vjTeSVyLXo4fUrVty9oWsQjgkSXffLUnqtJlTPbU2y5dLN97oKzZ/+tN0j6bp8vP9QWRdlUNjx6ZnTAAANFdOjtRGNXoOBdMZ4gU+ixf7T/xt20ZDpSwKhxo6lT3TyhonyBBvv1364Q9rP96xY1adzA5hEfxSHnxw7WZZTdGvn78AzUA4JPmykkBpKQ1cWpGbbpImT5a++93a89CzTWGhb0j9+99LH3zgpypv3EjlEAAge/lwqEblUE6OL6OpLxwaPDi6bocOWRUOUTmUXMGPPl7rgCz79UBYBP/Q6e2DDMLZyiRp0yaVKfIX+bPP0jsWJNXcudJ++0n//rfv05bNOnf2veluuMFX1H/7rfT970tHHZXukQEA0DS5uT4ccrk1vq8sKKj7E71z1cOh+tbNUA2FQ1QONU7wo4/3GbugwBdpcOIZZBTOCoYMRDgkSSUlmplzkL8+b156x4Kkqaz0P84xY6JnpMtmnTtL77/vj4tfe02aNcv3shs2LN0jAwCgaYLKIVfzdM0dO9Yd+HzzjVRSIu0ZcwKRLAuHGppWRuVQ4wQ/+qCPeU3B/byvyChBOBSkwkAGIBySpJISzc8dpu15BVQOtSLLlvkDhtYSnuy2m1+OGycNGpTWoQAAkBQ5OVKedsrl1khKCgrqbki9eLFftuLKIUKMxgk+Y8cLh4LCDPoOIaOUlflvr2umxEAaEQ6Vl0tlZSrJKVRJh12ltWvTPSI0wrffSp9+Ks2fHy0XXrfOX3/ySX+7tYRDU6ZIr74qPfFEukcCAEByNHpa2aJFflkzHHrmGWngQOn881tsrMnS0KnsmVbWOIlWDmVRfogwKCvzyWW2971Aq0I4VFoqSdqcs4s253eX1q9P84DQGAcdJA0fLg0ZIk2aJL37rj/74jHHSP/7v76/5ZAh6R5lcvToIR1+uF8CANAaVE0rSzQcWrzY/3EfOLD6upIvGf7HP/z86wy2Y4f/PFhzynu7dn5J5VDjEA4hKwXhEJBBCIdKSiRJW3Ij4dC6dWkeEBJVUiIVF0tnneX78cyf78/iJUmvvCINHSq99x4nnwMAIFMF08oqa4ZD8XoOLV4sFRVVP7tZbCqwdav01VctM9gk2bHDVw3VLBgw832Hsikc+vZb6ec/T2+1UyINqWPXAzIC4RAyEOFQbDjUrhvhUBYJ2kOdcorvwbN8ebQVgST9939L+++fnrEBAICGRaeVNaLnUGwz6mBdKXr/kiXJH2gS7dxZu99QoH377JpW9tZb0j33SHPmpG8MDVUOBZ+/CYeQUbZuJRxCxiEcigmHSttSOZRNghPLDR0q9e/vq8nnzZP69ZMmTJB+8pP0jg8AANQvR5XKkat7WtnixdK0adH76jqNvRT91H/00X6Z4eHQjh31h0PZVDkUNHlOZ6DVUEPq4H4aUiOjUDmEDEQ4FAmHNud2Vmm77v4fajb9VQ6xefP8H/wBA/wlCIdOOMG3HNhll3SPEAAA1KeNyiVJlTk1wqGxY/3ymmui961e7Y/TaoZDq1f75aGH+ulmsWXEGSiYVlaXbJtWFuRymTCtLN4ZwZlWhoxEOIQMRDgUCYfK2uyikrbd/X00pc54f/6zdNddvmooJ8dXDm3e7PuLDx2a7tEBAIBE5MmfuqtW5dDZZ0sXXli93KOu09hL0R5De+/t+xHNmCE99JC0cqX00UdVx3qZojVNK8uEyqEtW/z7lhPnUw3hEDIS4RAyEOFQTDhU2rabv4+pZRkvOJ371Vf75YAB0cd+8IPUjwcAADReVeVQzZ5DUu0zlsULh376U78cMEAaNUqaOVM65xzpkkuk/fbzZ67IIPVVDmXbtLJMqRyKN6VMoucQMhThEDIQ4VBMOLQpL1I5RDiUkTZs8I0PJT+F7KKL/BQyyVcOBWr2qQQAAJkp18WZViZFw6Hg1PSLF/uSm379qq/3hz/4dKJtW+nhh6UvvvDfFL3+un88OINFhqivcqgx08q2bZMqKpI3rqbIhMqhsrL6wyF6DiEjlZXFnwsJpAnhUEmJZKZtuQXRcIhpZRlpv/18O4ENG3x+V1cgdPfd6RkbAABovGjlUJxwyLloWrJ4sbTHHv4UZ7HMpHbt/PW8PD+1bOjQ6Ha9e7fQ6JumoYbUiQQtDzzg1033WVmzoXIo+PxN5RAyCmcrQwaq4y9xyJSUSJ06STk50Z5Dr73mz49e8+ADabN5s/8iUJLee88vY6eSFRZK5eX8yAAAyCZtnO85VGlxwiFJmj/f9xWaO1caPjyxHe+xR/T6brs1c5TJ1dC0sm++aXgf777rl598krxxNUUmVA5t2VL/Z+ycHP844RAyCtPKkIGoHCopkXbZRWZSaV5XX887ZYo0cWK6R4aIDz6QuneP3n7n4l53igAAIABJREFUHb+MrRySCIYAAMg2VeFQvJ5DknT88f7y5ZfSvvsmtuOafYkySEMNqROZVrZkiV+Wl0uVlckbW2NlQ+WQRDiEDEQ4hAxEOFRSInXuLDOpwtpI//63dMAB0l//mt6/tqjy8cfS9u2+t6QUPxwCAADZJa9yuySpMq9d7QeDT/xffSUdc4w/ILj22sR2HBsOZVgqUF/lUKIhRhAOSentpROMNZ1NtBvqOST5x+k5hIyxc6dPdgmHkGEIh7ZulfLzZRbpdzhsmHT55X4OU79+1f/6Ii3WrvUlwX/7m28rMHOmv92nT7pHBgAAmqNNuS85qcjLr/1gx47R64MGSSNHxi+5qWngQKlNZKpahqUC9VUOdezYcDi0ZYu0enW0lVI6X16mTCtLJBzKsIwQYRb8wyEcQoYhHIo0qqkKhyTpRz+STjrJ/+WdOzetw4MPh7p39zP+ggOhPn2ix3wAACA7ta2sJxyK/cRfWNi4HbdpIz32mA+JMiwVqK9yKJEQI+jBGLRfSufLy5ZpZYRDyCiEQ8hQhEMVFbXDobw86a67/HXOXJZ2a9ZIu+7qr3fp4pdHHpm+8QAAgOTIrfDTyiraJjkckqRTT/XVRhmWCtR3trKCAv/4zp3xtw+K2oP2S2GuHHrjDWnFioY/Y3foIL3yivSf/6RmXEC9CIeQoai9qKiQ2rSpHg5JUrdufkk4lHZr10q9evnrCxb45bnnpm88AAAgOfKCaWVtWiAcCvaRYeFQfdPKgpe8ZUv8l7x4sV9SORQ9Htxnn/rXGzpUevtt6X/+R3rrrZYfFzLMrFnSxo3pHkXUsmV+STiEDEM4VFflkORPF9G+PeFQBli7NtpX8qGHpGeekcaMSe+YAABA89Xbcyg2HOrcuWlPkIGdiBuaViY1HA51707PIcmfV+aCC6Sf/az+9e6+W1q3LnpSE4RIcXHmfnDo2TPdIwCqIRyqqJDy8mqHQ5KvHiIcSivnqlcOnX66vwAAgOxXFQ61qedsZVIoK4fiWbzYf2mWyLotLd2VQ2VlUteuDa9n5s858/jjifUoQisSTDt4+GFpjz3SO5ZYBQX+lxLIIIRD8SqHJMKhDFBa6k8oF4RDAACg9QjCofKWmlYWnBveOZ8QZIBEK4fiWbxYOvzw6IyUdIZD6awcKi/372WiQU+QC3z+eXRKHkIg6OB+3HFN/38ECAkaUhMOZbS1a/0yaEgNAABajzaRhtR1hkP5+VJO5FC1OZVDzqX3dFo11NeQumNHv9y8ue7Hg9PYx1YOpWta2c6d0cbZW7em/vmD151oOBS0KAgaeiMkvvjC//9BMAQ0iHCIcCijBeEQlUMAALQ+uTvr6TlkFv3k35xwSMqoqWXNmVYWBBuDB6e/cig2lEpH9ha87kR7+u6+u18GDb0REl9+KQ0alO5RAFmBcIhwKKNt2OCXwcnjAABA65EbmVa2M7eOcEiKpiXNaUgtZVRT6uZMK1u0yC8zoXIodozpCIcaWznUubPv/3vNNVK7dr5K6403Wm58yBBffCEVFaV7FEBWoOdQQ+HQhg1SZWW0rBkptWmTXzb1mBAAAGSuNpHKofLcOhpSS/6Tf0GB1KaJh6zpLq+pQ3Mqh4Kqlz328CfVrW/dlpZtlUOSdN990syZPqC7/Xbpk0+kww5rmfEhjb7+2lcMOSctXer7DQFoEOFQQ+FQZaVPKLp0Scvwwm7jRr8kHAIAoPXJ3blNO9VGlZZb9woFBc3rFZJh08oqKvyhZVMrhxYv9n0YO3Xyt9u3j1855Jw0b57fV48eic2scU769NPEeggFU9w6d86OyiFJOuEEf9m504dDGfJrgWT7r//yv8iBvfZK31iALEI41FA4JEl33OFPazB6tLRqlXTQQSkfZlhROQQAQOuVu3ObtilfFRVxVigoUPwHE5Bh4dCOHX7ZUOXQ3LnSM8/467m5/rNuQUH0NPaB4GRsdSku3kWHHhrdx9dfN3za9zff9M/VGH37+n2nWlMqhwJ5eb4YLYNmGyKZ1q6VjjlG+tnP/D+2sWPTPSIgKxAO1RcO7b23/2t6443V76+1IgLO+be0qdXfNW3a5E9WEu8gCgCA1szMjpJ0p6RcSZOdc3+Ms96PJD0laT/n3EdmNlBSsaSFkVVmOucubvkRN05u+XZtU74qK+OsUFTUvGAnw8KhoMKmXZxZdEHQ8X//5y+BO+6QLr/ch0OxM2QKCuK/tC++8K/9ggukv/9d+uabhsOhNWv8cvJkqXfvBl5M5PmfeUa6//6G10224HU3pnIoVn3vHbJcWZn0ne9IRx+d7pEAWYVwqL5waL/9fDqxfr3/Dyb4i755c/Rco6jmrLOkRx+tOz+rrJSmT/dBfqLh0aZNVA0BAMLJzHIlTZJ0uKSVkj40sxecc/NrrNdJ0mWSZtXYxefOuREpGWwTBZVDccOhKVOa96Vcurs21xCcaCNeSJOb678Q27FDOu886Re/8IXra9dKJSW+Qqdm5VC8l7ZqVXu1bSsde6wPh4Jq7PoE+zrqqP/P3p2HR1We7wO/T9gCIWFfAyEsYQcRgqggxKUulbpvtdbS2tr6rYpeP7VWra173Vpt64Z2caF1bZUqVq01igoKCrKHhCUQ9i0hyWTP+f3x5OXMTCaTWc465/5cV64zme28GULmzH2e93mB7OzYfqb//Mc708qCMRyKbtUqoLQ08m0jRgCTJhnfb9hgNEvv0kX6OLU1ddJyui6/HImUlBH5HMOhaOEQYDRCvPxy47RIaamcWundGzj2WFuH63YvvSTbw4dbt2l6/XXg0kuBt94Czjmn9WMffRT44APp/f3LX0oFKMMhIiLyseMAlOi6vgUANE17GcC5ANaH3e8eAA8CuNne4SUvrb4WdejS9syx9DZWMYuVyxpSq0Vwo63CqqaenXCCdDXo0UOOh1Qz6uBwKFrAsXNnV4wcaRyPHTnS/vgSmaqVni49fFoOqW2TzLQy9TiXZIau09Qkv39thX49e8qxvnL66UBZmfH9668DF15o7RjbVF8vZ6QZDhHFjeFQY2P0cEh57DHpNXTVVdL9XtX0NjcDmmbLUL1k3Tpg1qzQ6xYskO3mza3vv3+/BELZ2cCuXcDQoQyHiIjI97IB7Aj6vgzAjOA7aJo2FcBQXdff0TQtPBwarmnaSgBHANyh6/qS8B1omnY1gKsBYMCAASgsLDRx+IaqqqqIz52zfzdqkY5vvlmDbt0Omr7fTocOYSaATStXYtewYaY/f7yWLesNYDJKS79GYWFbaU0BAKC6eiUKCyvQpctxKCqqxFtvHQQwHuXly1FYKMlIQ8MU7Nqlo7Dwm1bPsmPHVAwefACbNm0FMB2ffbYWHTseiDq+tWtzAIzAihUfo1On2Cq2du4cCmAk3n//E3Tt2lYJmPm++SYbQB5WrvwUW7Y0xv14Xc9HaWktCgvXmjKetn7HvSgQ6IDa2pNw0UU7cPrpe0Nue/vtQVi0KBsfflh4NAzcv/8knH76fpx22l7ccssx+PzzIvTps9vSMbb1enesrMQsAMU7d2Jnivx7uEUq/Y57gROvN8Oh9iqHlMxM4NvflsuffWZcv2QJMHu2pUP0kuxs6dm9Zk1oOFRSAnz4oVzevr314154Qc46vfOOTE1T92E4REREFJmmaWkAfgdgXoSbdwPI0XX9oKZp0wC8qWnaBF3XQxIJXdcXAFgAAPn5+XpBQYElYy0sLESk567o9jAOIR0TJkyCJbuuqgIAjM7OxmiLfrZ47GiJ+k4/fWpIBVAkl156LAYMkNXJ0tO7oVOnAQCAyy6bfrQoYvBgqUYKf211Hdi9uwnnn5+F007rCwAYOnRiu6/xf/+rGmDPifnc5+rVsj3uuNlRK6LMtmyZbE8/fRa6do3/8QMGAF27do/4e5mItn7HvWhvSx50yilD8ZOfDA25rbwcWLQImDGjAN27y3nyujrghBMG4ic/GYhbbgGGDh2DgoIxlo6xzdd71y4AQN7kychLkX8Pt0il33EvcOL1ZjgUazgEyLtIejrw2mvGdQsXMhwKopZWXRt2EubZZ+Vgo1+/yPOXFy0Cpk0Dxo8Hhg2TucuAhEOxznknIiJKMTsBBH8yG9JynZIJYCKAQk0+yQ8EsEjTtHN0XV8BoA4AdF3/StO0zQBGA1hhx8BjldZQF321smR5cFqZ0r+/bHv2lA/kmzbJymDBs2UyMoClS4HLLgt9bEMDUFfXAaNGAVlZcl34tLJnngE++kiOz269VXrIVFfL88dTFK9m/s2fD7z4on0F9dXVsq9EZx5GW+nN79R0u0gzs9R1gYC0YFVTz7p1C73NMdEGT0RRMRyKJxzSNEkuioqko/IJJwBfhPd+9DfV7HDhQnlpKiunITNTwp65c2UacKRwaNs2YM4cuZyTA7z3nvx7sHKIiIh8bDmAPE3ThkNCocsAXK5u1HW9AkBf9b2maYUAbmpZrawfgEO6rjdpmjYCQB6ALXYOvl1//CO6rfsStTiu7YbUyUpLk/TAJc1lDh6UIfXs2f59VcjSo4e0ugxfxh6QxZjWrpXmweFGjqzCKad0PxoOhTekvu8+CYwqKoCRIyUcCgTib/A8o2Wi48KFwP33y3GcHdRYEw2jMjKkwTe1Fms4FH7fTp0kbGQ4RORNaU4PwHHxhEOALKkKABMnSlOctWuBmhpLh+glFRVAbq5MKRs4EOjdux4DBwKnnQb8+teSrYWHQ42NMhVNHUzk5MiZnCVL5CCK4RAREfmRruuNAK4F8B5kWfpXdV1fp2na3ZqmRVjaIcRsAKs1TVsFWeL+Z7quH7J2xHG6/np0qK6MvlqZGVy0LNXBg9IgOi3KEfjWraH9GXv0kMqh4mJg9OjQ+151FbBxY+Sv555bgbFj5cN6RkbryqGKCmDevNAKmkQWeTrmGKN1QElJfI9NhqpySpSLfi1cJ9FwSNNc0Oib4RBRwlg5FG84dO+9wHHHAWecIacbmprkdM0JJ1g+VLdraJC/x1ddBdxxh1xXWLgmZK5kTo4cGFVXG2emdu2Sl1H1iVRbVUkUy9k1IiKiVKTr+mIAi8Ouu7ON+xYEXX4DwBuWDs4kUVcrM4OLUoADB9qfUpabG/p9z55SOdTU1LpyKFZqxTOluRmorJQpZxkZxufp4OOzeIwaJdviYuCUUxIbY7wSHavieIjhYomGQ2rrinAokUZURD7HcKipCejYMfZwaNo0+QKMNRtnzQK+/NK43qfUQUe0Sh8V/JSWSn8hdRkIrRwKxsohIiKiFBNUKlSLdGg+qhzq27f9+wXr0QNHw7NEw6GsrNDKoaoqOe7t0SP0w3wilUOA9ELq0sXeyqFEx6q46NfCdTwdDqkZHawcIoobp5XFWzkULDsb+NnP5ADnlVcsGZ6XlJfLNlqlz5Ahsm1ZSACAsTJZeOWQ0pERJhERUWoJKhWytCE14KoU4ODB2JpRBws+SWZW5ZC6rCqH1MuT6FSttDTpW1RcnNj4EpFs5ZD6ueM+/veBaOGQes2jhUOO/nfjtDKihDEcSiYc0jTgqadksvWaNZYMz0tiqRwaIKuwYs8e4zpVOTS0ZT2Wfv1kdbOHH5bv+bediIgoxQSlQY3oaG3PIYdLGXTd+EokHFIn3dLSgBEjEhtDeOWQuhypcijRwGXUKOk9dCios9XBg9Ke08zwr7YWWL0a2L8/+WllTU3SFoFCebpyiOEQUcL8HQ6pd+pEwyFl8mSGQzDCoWiVQwMHynbvXuO67dulxDr4Df7HPwZuugn45BPgyivNHysRERE5qLHx6MWeKE/phtTXXCPBTlqadCTo1y++x6uTbjk5MnUrEW1VDvXoEfryJDNVa/x4ma6mVi8DpE3npEnAgw8m9pyRXHednJf9+mugd+/En0cdd7qkqMxVUiIcYs8horj5OxxSpzGSDYcmTZLltg6FLQLyzDPARRclNUQvUdPKolUOZWXJirLBlUMbNkgpciQnnSSrbBAREVEKCSol6YOD1k7tcTgcWrtWFrv9zW+Au+8Gfv7z+B6vjqsSnVIGtF05lJUV+mE+mRXAfvlL4MQTpe9QdbXsY8sWuW3VqsTHHm7lSmnz+cYbwB/+kPjzMBxqm/p9iFSZ1VY4pO4b3ODcEawcIkqYv7u5mBkOAVI9pJbYAoB335UvXZcpaCkulsohTZOpZSocam6WN/nvf9/68REREZFLBIVDfXEApSkcDtXXA2PGAL/+dWKPV8dV4cvYxyNa5ZBZ08qysoAbbgA+/1wCouCpZGb1ItJ1ea7vfx+44ILknis85CCD+u8SqfhGvW7BfaqCr3e8ckg1pGblEFHcWDkEJB8OqWW3Nm4Mvb60VI4IVElNioulcgiQqWVqWllxsSylmp9v7diIiIjIRYKmlR1Cb2srhxz+tNrQAHTqlPjj1dSpZMKhnj1lyteJJ8r3wZVDZk0rA4zqpuJiIxA6+WS5HMu/8ZlnSsX4Cy9Evn3/fhl7MlVUCiuH2hYIyO9spN9bT0wra2vwRBQVwyEg+XAoO1uW1FKdlRX1ffAcqhTQ0AAUFMhUsOCve+6R2zMzoz8+uHLoq69kO22aZcMlIiIit2k5Bqs7fg6+i3+k9LSy+nqgc+fEHz9sGPDnPwM/+EHizzFvnpycW7pUwpVIlUPNzVJ0kUyT51GjZBscDp15prz87R0O19cD770n4/j668j3Uc/JcMha0ULCTp0kwHN1OMSqIaKEcFoZkHw41KGDLLW1bZtxXWUlcPiwXN6zBxg3LpmRusrWrcDHH0s/oPBl5ydObL9HkDo4AWSbnm4UXxEREZEPtFQOBS64AvuWDbA+HKqpkdQhzf7zoslWDmka8KMfJTeGnBzgiSeACy+UKV8VFfK8GRlGdqZm4yRTOdS9OzBokIQ4TU1y/nTKFLmtuFhua8vmzcbltgIbFQ4lU0WlcFpZ26KFQ+r3JjgcSkszAlBXLGXPfkNECWE4BCQfDgFAbm5o5VDw5RSrHFI/2r33ArNnx//4gQOBAwfkjePvfwfmzpXCKyIiIvIJdQzWcgBgeTgEyIfG7t1b3/7ww8DrrwPz5wOXX2767pOtHDKLquopKZHqoaws+VCvKj3Ce8ckKi8P+Oc/5d906lSjymfTpujHjZs2GZfbChc2bZJfmdzc5MYIGL8W114L9Oolx6evv87ZSED7+Up4n6pu3Yz2quo2x1quMhzyhHXrJLC2dKVKjzvjDPv/GPn7I7ma725GODRsGPDBB8b3PgiHcnISe/yAAfJaz5kjC7z99KfmjY2IiIg8oOUYTOso5ca2hEPV1ZHDoaeekrLoN96wJBxKtnLILGpl2OJiqRxSPSLVy3PwYOj3ibr2WuAvf5HLV10lx4udO7fflFqFQ7m50SuHhg8356Ti2LHyz33okCw6vGgRUFYmz+93iYRDwbfpOlBXJ7MDbFdTw3DIA55+Wr7693d6JO41Z479S3b7Oxwys3Jo2DBg1y7j9FCKh0NpaVIqnIhTTpEzR3V1wPe+J98TERGRj4RVDlkq2vyhujrjmK2y0pLdu6VyKCMDGDw4tHIIMF6e/ftDv0/UxRfLV7CRI2MLh/r3lzG2NdWruNicfkOABBcLF8rl114DLrmEU8yUZMMhdb0j4RB7DnlCSYlMOW2rvxgBhYW1tu+T4RBg3rQyXQd27DDeAdPTgX79Ui4c2r5dgqFEz4KNHSs9i4iIiMingqq3ARsrh8Jt2WLMa6iqsmT3bqkcAmRq2WefyUm+vn3lOvXy/O53od+bKS+v/XCoqEh6CaWnR/6nUsvYFxSYP77w5dn9zqxwSK20Z4rmZvnP1EKrr5dwN1xVFSuHPGDzZqMfGbkHVysDzAmHVA3qqFHy275iBXDssXL6Izgcqq+3+AjIHAcOALt3G80Jg5WWtm5ETUREFvn1r2X+R48ewG9+4/RoiMzhRM+hSJ/8VWIxeHDKVw4BwAknyI9cVCQn6wDjc/Rbb8nWimO8vDypFIjWX2T7djnX2tbicrt2SeBgRjPqcMFtqci8cMhUxxwjyWHL15wzzgj5/ujXp5+2v3QyOaqxUWbyqqmu5B6sHALMCYdmzgRuuQV46CHglVekRu7HP5Z3utWr5clrauTd8ZprgDvuMOVHsMI//ymrWQAyT3zbttCGcqWlwIknOjI0IiJ/qakBHn9c3jsOHJDedmYGRM8+C9x2W9tvgA89BIwYYd7+iBS3VA6pcGjqVGDNGkt276bKoQceAH7xC7ncs6dsgyuFli6VlWfNlpcH1NZKT59IPSt1Xc6lDhokAVK0fyqzppUFY+VQqFjCoX37It/XknBI16WD8amnyheALVu2YERb709nnWXizslsO3bIWwDDIfdhOASYEw517Ag8+CDw/vuy6kUgAEyfLin3m28CS5bIu9quXcAf/gDcfDPQpYspP4bZPvxQ+jXOnQu8/DJw+LBRFqrr8sY+dKizYyQiSim6Drz0ktH0Q1GdY3/7W+D554FPPjF3nw8/LMv0nH565PuMGsWlRMgaLcdgtjSkVp9Wn38e+OKL0Nvefhvo00cSi6VLTd+1rsuHILdUDmma/JcPFvzBfvJka/arqn2KiyOHQ+XlMkNo0CC5HB7SHDgA3HijXLYiHGLlUKjq6ujTC4OXq6+uNqYoAtGz2ITV1cl/ptNOA269FQCwvbAQI6yYY0iWKymRrVpBkdyD4RBgTjiknH02cN99cnn6dElR/t//A849V04dZWXJwf+bbwKXXmrCDs23fDmQnw9cdJGEQ6WlRjhUWysHOeEHFkRElITiYuDKKyPfNmYMcPLJktzv2iVhTVoCs8LffNM41QsAe/fKfv/8Z+BHP2r7cYWF8e+LqD2qcsiOaWXDhsk0kxdeiHz7BRfI7RZMK1MtUtwSDkUSHAJY1apFTVXbsSPy7bt3y3bQICm6Dw8W/vIXYNUqOay24gQlK4eA++8H/vhHubx3b/TfhYwMmZo4aJB8rDnvPOM29bjvfCf0PPhJJwGvvtr6ufbskRWMjxxpfdvZZwPPPQcjtWMvIVd7+mngrrvav59qW8LKIfdhOASYGw799KfyV27AADlNomnAM8/ImSkAmDcPOPNMYOVKV4ZD9fXAN98A8+cbb+SlpdI+CTB6NUZaCZaIiBK0fbts33kHmDUr9LZu3WTqzZAh8oF63z5g4MD4nn/5cuD881tf36uXnAkgspudlUP9+8t65eq4L1znznJir77e9AZB9fWydcu0skjs+LytjhvbCl9UODRwoNHPRteNtgZFRVLgtXXr0ZmIpmLlEPDRR/Kan3uuvO4//nHb9/2//5P2Pur/7RVXGLfl5wM33RQa9qxcKW0rIk2x/PJLWanuootCG1gvWwa8/rrMftZUmsBVyFxt0SLjd6g9OTlyWEPuwnAIADp2NC8cGjq0JeIOcskl8hV8n+Cl7l2goUHKdbdvlwOZ6dNDwyGF4RARkQXKymQ7erSxvnS47GzjvvGGQy+8IKdw16wJLRPIyuIfdHKGnQ2p1X46RjnsVQ1sq6pMXWLJS5VDVi473t5UI7V2y6BBct+mJjkeVZUnmzYBEyZYEwwBrBwCJBibOFHOabfnuOPkK5L0dJmxHOxvfwN++EP5TBE+lUhNMXrmmdD/eo89Jp9NDh4E+rJyyBM2bwZmz47td4jcKaZwSNO0MwE8DqADgOd0Xf9tG/e7EMDrAKbrur5C07RcABsAFLXcZZmu6z9LdtCmsaJyKBbDhrkuHFq7FnjiCVmsIz9fZjD06SMBPcMhIiKLqXBIBUCRqFNsmzcD119vVBvFYt8+mTpjRbMOokS0TCuzpXIoFiocqqw0NRzyQuWQCmCOP966faiCj7Yqc4KnlQUHScHh0DnnWDc+y1bY8pBAQI79raCmD5WURA6HevVq/d9O3a+kBOjbleGQ26kVyC64wOmRUDLaDYc0TesA4AkA3wJQBmC5pmmLdF1fH3a/TADzAYR1+sNmXdenmDReczkZDr3/vk07i82BA7J95ZXQGQ25uZHDoWhN6oiIKE47dxqJfFtUOPT009I494ILYm8A16GDzBcmcgu7K4fao856qQMdk3ihcmjIECkuPPts6/aRliZ/3qJNK+vWTTK64HCod29pUL1vnzVL2AePLz2dlUNWZS/BQU+4SIFR+GOOH8VpZW63Y4f8vWOTaW+LpXLoOAAluq5vAQBN014GcC6A9WH3uwfAgwBuNnWEVnIyHNq92/R57clQ4VDwagOADPWf/wQ++wyYOZOVQ0REligra3/yff/+8gmmsFDKPF991bo5FkRWc3PlkIm8UDkEAN//vvX7yMiIXJlz5Ij0Khk4UHrdqIBixQpg2zZjCXsrwyE1PoZD1jy36iW1eXPr2zZvBmbMaH398OHy+7B5M4DBrBxyO65AlhpiCYeyAQSvLVAGIOS/sKZpUwEM1XX9HU3TwsOh4ZqmrQRwBMAduq4vCd+BpmlXA7gaAAYMGIBCC1dGqaqqOvr8WWvWYCqAb9auxf79U1FdnYHCwuWW7VsZWFODsbqOZa+/jtrBgy3fXyyWLs0GkIeNGz/Dnj0NR68fOHAYgOG47LIAXnzxSyxb1hfARGzcuAKNje2fXQt+vcl6fL3tx9fcXqn6ek/buBH1ffpgTTs/27SRI5FZXIxtp52GbUtavZ1aIlVfc3JYWOWQ44J7DpnIC5VDdgle/jzY/PnywfK00+R7VTkUPj1l4kTrx+fFaWXNzXLOQL22OTnAMcfE/zxWhkOaJqHBsmXAv/9tXK/rMkPhe99r/ZguXeRnWbIE+KIpgBkAPlnRDRUVcvuaNX2sWGAwbh07SjsOK3t22aWxURqT19ZGvj3aa/7BB7JlOORtSb8ja5qWBuB3AOZFuHk3gBxd1w9qmjYNwJuapk3QdT1ksUJd1xcAWAAA+fn5ekFBQbLDalNhYSGOPn/LUsDHTJ2KAcv7Y88ewMp9H9XcDDz0EI4fNAiwY38x+OibUUDaAAAgAElEQVQj+cM9d+7MkOO0ggKgXz/gkUe6YcaMgqNtMQoK8mP6zx/yepPl+Hrbj6+5vVL29S4vB045pf2fbfVq4PBh5A4ciFy1jI/FUvY1J2eppew7uKRySJVE+7RyyA5tVebs3Svbl14y7qf84Q/A+PEyvczqZa+9Wjn08cfAqaca33fpIhlnvLmrleEQAEyeLP/GkXpHTZrU9mP+/W+g14c1eA3ANf+va9DUlTYe5IA//hG49lqnR5G8t9+OvLCpIfpr3ru39A0j74rlz8ZOAEODvh/Scp2SCWAigEJNDlQHAlikado5uq6vAFAHALquf6Vp2mYAowGsMGHsyXNyWhngqqbUBw9K64pIbyTTp8trs2EDp5UREZlG14G5c4EvvpA/wrGs6ZqeziMvSg1qKftOLuk5ZNG0MlYOGdqqzAkEZIWjAQPk++Bw6JJLjOudGp/brVsn2w8+AN55R1b5qqkxfqVj0dws1SJWhkPPPAPccEPr67t0kQAwkoULpRl577cDwG+Al9/qhvqWdRtWrFiB/Px8y8Ybq9NOA9aHN1vxKPW79Pnnkf9mtfeaDx58tPaCPCqWcGg5gDxN04ZDQqHLAFyubtR1vQLA0U41mqYVAripZbWyfgAO6brepGnaCAB5ALaYOP7kBJ21sjUcGjpUynRcFA4dOND2CgWqjHfdOoZDRESm+eorYPFi6QI7apQ9TT+I3MJtlUMWTStj5ZChrcqc6mqpUg++n9K/v/XjCt6vGt/770tbtwsvBM46y74xJKKkRMZ+6qlGf6ZAIL5wqKal37OV4VC3bsC0afE9JjOz5TFfSmo36biuUoYAoLKyKu7ns0JeXuRG215UUiIBzwknRL7dLa85WafdcEjX9UZN064F8B5kKfu/6Lq+TtO0uwGs0HV9UZSHzwZwt6ZpDQCaAfxM1/VDZgzcFE5VDnXuLGd+XRYOhTejVkaNkoOadeuM+bTsB0dEvlVeDvztb0ZJQKI++ED+qL70EtCzpylDI/IMt1UOWTStjJVDhowMYwGUYNXVRlE9EHqMadPs2aP7PXxYLt97r/S62bLFG+HQqFGhzbzjrYAKuL3fsx3pVYJUL6VU0NbKceQfMc1G1XV9MYDFYdfd2cZ9C4IuvwHgjSTGZy2nwiFA3gVdFA4dPCgFTZF06gSMGSPh0Jgx8ubOkkEi8q1HHgHuu8+c5/rJTxgMkT+5bbWyrl3l4IY9hyzTVkPqQCC0Wij4sp0yMnC0t+amTbL1wjSzkhKjyj9lwyE1QBcuZT9qFPDKK65ahDphJSXAt7/t9CjISS5ZIsIhTodDy61fGS1WBw4AU6a0ffuoUVKqmp3NKWVE5CO6Hvrm0NwMvPACcMYZwBsmnPtw7ZE4kcXcVjmkaTKHhZVDlok2rSz4T6GaDjV7tj3jUlTPoSNHjCbZbg+HGhuluum88+T7lA6HOnZ0Zco6apQcGmzdKifRvaqyEtizh5VDfsdwCHAmHMrJAf75T/lr4nAZjq5Hn1YGyHD/+19ZGpPhEBH5xgknSMPocI884tzpbaJU4LaeQwCQlSXJgIlYOWSI1pA6+M9p9+7A//4Xf3+aZKnwSvXtaavSyS4rVgDz5xv/VSJpbJQAMi9Pvk/ZcKimxrWDU6/9eefJn5B49e0LvPZaYj/eiy8Cf/pT/I+LRC1fr34e8ieGQ4BzlUP19RLRDh5s445bCwTkD0K0cGjYMOnRWFbGcIiIfKKpSSo8Tz4ZCF7KPSsLuOACx4ZFlBJU5ZBbppUBloZDrByKXDnU3Nw6HALkz67dMjJkLGpK2ZQpUpXjlDfflF42p58e/X7nny/FrEAKh0OBgGsHd+yxwPe+Jy064nXokKxL8c03bTeBjuavf5WpYMcdF/9jI8nLCz3cIf9hOAQ4Fw4BwLx5wKJFRqdnB/znP7KdMKHt++TkyHb9emD0aOvHRETkuEOH5JPL+ecD113n9GiIUovqOeSWaWUA0KMHUFFh6lOqaWWsHDIqh3TdaDTtpj7DqlLonXdkfJMmAWvWODee4mJg+HDg3Xdjf0xKh0Mu7DcEGOtKJKKoCBg7Vv6tEwmHVI+gF19MbP9E4fzdVtjJcGjaNKBLF1mt5r33bNxxa3/4A5CbG301BpVlHTjAyiEi8ol9+2Rr51rKRH7hxsohC8IhVg4ZVHWQCoQAI5RwwyzdwYPl93DhQvnA3qePEWY5oaQk/ik+KRsOuXhaWTKGD5fuImoqYzxqaoAdOzgNjMzFcAhwJhwaOFDWy+zYEfjySxt3HGr1auCTT4Cf//zotP+IVOUQwHCIiHxi/37Z9uvn7DiIUpHbGlIDMq2MlUOWUQFQ8NQyddkNn/t/+lNg7Vpg1Srg009lTE1Nxr+hnXRdAgOGQy1cXDmUjM6d5QR9SUn8j928WbYMh8hMnFYGOBMOAfJHbvLkyM1ObfLHP8owfvSj6Pfr318KnerqGA4RkU+wcojIOmFL2btCjx7sOWShSMGFCofcUDmUlhbaYiE4zLL732/fPlk9Kt6Vo1I6HHLt4JKjVoSOl3oMVxcjMzEcApwLhwBgxgypX3Vg1TJdB159FbjkEqB37+j31TQpsf3mG2DAAHvGR0TkKIZDRNZRx2AdXVQ5xJ5DlopUOeSmaWXhgoOWXr3ksq4Dn30GzJxp9E0yW2Mj8PDDcpmVQy1qatr/sOJReXnA0qXycTAeqmcsK4fITAyHAGfDoeOOA556SjqSjRtn664PH5YTZMccE9v9Fy+Wssf8fGvHRUTkCvv3y9F/nz5Oj4Qo9YRVDrkiHMrKkuVb6+tNKxVh5ZAhWuWQG0OJSOP9xz9kZaoXXwSuuMKa/S5ZAjz6qFyeNCm+x3bqJHlryoVDgQCQne30KCwxbRrwxBOJ/T4NHw707Gn+mMi/GA4BQMeOzoZDgPQdsjkcKi2VrWo23Z7Bg+WLiMgX9u2TYChaQzYiSozqOdRBqqZdEQ716CHbI0eAvn1NeUpWDhlUddDOncaJRjdXDkWqdPr0U9kmsmx5rA4flu0HHwBDhsT/eLUqXDxqauRcSJcu8e/PMitXAm++KZd37oz9bLbHzJsny8e35OVx4WwOMhvDIcDZyqGxY4HMTAmHfvADW3e9bZtsYw2HiIh8Zf9+NqMmskpjoxx/pcncHFeFQxUVpoVDqnKI4ZBR4XDeedIvZdQo71UO7dwpWytnOKn95eYm9vhEwiHV0seqqXIJufNO4O235bKmAVOmODsei2iaVAARuQHDIcDZcCgtDZg+3ZGm1PFWDhER+UZzM7BlC/sNEVmlqelovyHAJeFQVpZsTew71NAgxYc2t5V0pSlTgGuukW4KZWWh4ZAbK4eihUO1tdbtN9lqqvBw6NAhWfI8mh07XBjQFRcDF1wAvPGG0yMh8g2GQ4Cz4RAgU8seeUTeadLTbdttaam8EbCdBhFRmLlzpaT90kudHglRamqpHALg7DFYsODKIZOY2L7I89LSZArNU08ZoZDXppWpcCjeypx4JNv/JzwcmjMHWLu2/ceNHp3Y/izR1ARs3Qqcc47TIyHyFX+HQ2pyp9Ph0LHHylg2brS1ZLK0VKqGXFVCSkTktE2bgHfflaUc777b6dEQpaagyiHXhUMmLmff0MBwKJgKXFR44ZVpZbou/5Z79hjXWSXZ1yQ4HKqvB9avB777XeDii6M/zubWp9Ht3CmDHznS6ZEQ+Yq/wyG3VA5NmCDbdescCYeIiKjF++8Dd9whFQ2PPQYMGuT0iIhSkxsrhyyYVlZfz35DwcKnark5HFJBVmWlVNWUlBi3BVcTmS0QkNw00d+b4HBo61aZJX3WWcD555s3Rstt3ixbhkNEtvL3DGi3hEN5efIusG6dbbtsaJAzCWPH2rZLIiJ303XghhukivP66xkMEVmpqcl94ZAF08pYORQqfKpWICArZLlxUUgVWBUVSTB0wQXAgw/KdVZPK0smLAsOh1SglZeX/LhspcKhUaOcHQeRz7ByCHA+HOrcWU5J2BgOrVkjLY6OP962XRIRifp64PPPjb/Bcei5alVCj4vJ1q3Ahg3As88CP/6xNfsgItHY6N5pZawcsowKPd54A7j5ZjkW7dXL2TG1RY111SrZXn+99O959FHvhEPFxbJ1VcZyzz3A4sXR71NWJv9xhg61Z0xEBIDhkGzT0pw/MJkwAfj6a9t2t2yZbBkOEZHtHnxQlqhNgOUTbzMz2YSayA5hlUOu0LmzBFYmzhli5VCorl1l++mnEpzdcoss2utG6enyu6nCIdWwOZGl4uORbDiUkRFaOdSzp8sWn3niCSkXizZ9Yfx46V7uxpIyohTGcMgtJc3HHgu8/rqcuR4+3PLdLVsGDBgA5ORYvisiIkNTE/Dcc8Ds2cC998b98JUrV+LYY4+1YGAthgyRgIiIrOXGpeyB0E/WJmDlUKgOHSR0qa2VfEBN03IjTZOQ5tAhoHt3YOBAuT7ZX5FPPpHFMOfPj3y7GZVDu3bJQl/Ll0vVkGsC2MpKYO9e4IEHgFtvdXo0RBSG4ZBbwqHvfx/41a+k1PKBByS5sdDq1cC0aS56syCi1HH77cALL0S+rbFRlnt55BHgpJPifuqKpqaEHkdELuPGhtSAfPI3sXKork5CEDJ06ybhkJrF52bdusmvQ16eccycbOXQnDmyve46IC1C99dkw6GzzgK++EJmZg0aJAU4rsFeQkSuxnDILQcmQ4bIGpN//aucTli50tLdbdsGzJpl6S6IyK/+9S+ZR1FQEPn2Pn2Ac8+1dUhE5DJuXMoeMH3OEMOh1jIypBpHLQ7nZmoanJpSBpj3K7JjR+RVg6urkwuHvvMd+XIlhkNErsZwyC3hEAA884wsIVZSIutOApFPKSSpokK+cnNNf2oiImDnTjlV+fjjTo+EiNyKlUO+pYIPL4RDV18NfPABcOWVxnXdugEHDiT/3MXFkcOhQMDyCQTOUcuncYl6IldiOOSmA5OsLOCyy4DbbgNOO01OUzz9tOm7KS2VbaQ3JCKipBw5Il9Dhjg9EiJyMx9VDnlh+pSd1HL2XgiHbr9dvoIl+yui8sfiYjncDxcIGK+RazU1SZ/UeK1aBfTvz95+RC7FcMhN4RBgdLsrLDR1KdVgDIeIPErXgaIi6XAabMQI6ZbpBjt3ypbhEBFFw8oh31KVQ14NzVQfokR16WKEQ5Ek23PIFrfdBjz0UGKPnT3b3LEQkWkYDrntwESFQ7pupDgmYzhEvnfkCPD55/L/bOJEYOhQp0cUm4ULpXl9uG99C3j/ffvHE0lZmWyzs50dBxG5m5srh/bvN+3pGA615qVpZZEkUznU3AyUl8vlZcuAd94xbuvYUZpVeyIc2rFDPrM8/HD8jz3+ePPHQ0SmYDjk1nAIAA4eBKqqgPvvl+lmkyebsovSUllGtH9/U56OyHvmzwf+9je5fPzxwNKljg4nZk8+KUum/Pa3xnVPPAFs3Bj/c91/v1QoxkPT5LX79rfbvg8rh4goFj6pHFJLtpPBS9PKIklmKfuKCqOt6NKlwNy5obc/+qhHwqGaGvkgccUVTo+EiEzEcMhtBybB4RAALFggS9uvWgUsXmzKLkpLgZwcLmNPPnX4MPDyy8Dll8sp3U8+cXpE7duzBzjnHGD5clkC/oILjNu++gr4+OOQv2ftamwE7r0X6Ns3vhBn0ybg1ltlndy2/oCoyqHBg2N/XiLyHzdXDnG1MkulwrSyQEB+Z+M9lj54ULZPPgnk54fedvbZwLp1HgmHPDFIIooXwyG3hUP9+oUO5p57ZPvuu1IdMHZs0rsoKeEiAZRinnhCKlpi+U+s6/J1883A228Db7whp3bT060fZ6KefVaCoR/+ELjqqtDbhgyRv2X79gGDBsX2fEVFctbv/vvjO+v3zDPAz34mAVVbS6l8/LH8HXPz60lEzvNJ5VBdHf8chvN65VC3blL9U18ff/B36JBsc3KA6dNDbxszBli9Wv4vuD53qakBunZ1ehREZDJ/h0PNze47MOnYUco09+6V78vLgWuuAZ56SnqKJBkO6bqc/GcvOEopS5YAvXpJcBGL3FxgyhSpyANkKpSbEtPFi4FPPzW+/9vfpK/QX/7S+r6q8qesLPZw6KuvZDt1anzj+u53pXLollui36+gIL7nJSL/aWoCOncG4LJKZlYOWS4Veg4B8msS77+tqhzq3bv1baNHAy+8ELoP1woEWs92ICLPYzjUckTimnAIkD+2nTsDu3dLWHTXXfLh0IQG1Xv2yAmxvLzkh0nkGmVl0lhaVdrFKjhYcUs4VF0tIUxVlTFNrFMn4MYbI99fNX4uKws9DdnYKB++Ilm+XI48x4yJb2xZWbJ0rTr12ZZYQyoi8q/GxpDKA9ccg2VkAA0N8tWpU9JPx3CotVSYVgZIPtKrV3yPVW+fffq0vi0vT/5bBO/DtTitjCgl+Tsc0nUgLQ2Ay8Kh/Hz5gLhtGzBjhkzRyMkxJRzatEm2o0cn/VRE7lFWllg5XHA45KS6Oqklb24GPvhAVlNbsgSYNav9x0b6GUpKgGOOiX72+8QTY+9RFKxnT/kiIkqGG6f2A6Gf/JNML5qbJWNiOBQqFaaVAYkVmLVXORS+D9fitDKilOTvcKi5+Wg45CrPPSdb1RsFkHXnt29P+qkZDlHKaWqSaWGJrI7llnDo9ttliRJl4kRg5szYHtu3r1QaBv8MTz8tzRDuuaftAOiMMxIfLxFRshob3dmQWiUX1dVJh0P19bJlOBTK69PKuneX7fjxoVMidX12u1Mkm5rko0ekiqPgY3O1D9di5RBRSmI45MbKIUXTjHednBzpj3LwIPD448Cddx49qIrVypXA1VdLlfTQoRaMl8gJ+/bJh4xEfqm7d5cqGCfDoZoa6SV0xhnSVBsAJk2KvQlHWppMLfvHP4ANG+S6jz8Gzj0XuOMOa8ZMRJQsL1QOJamuTrYMh0J5vXLo1FPl3Ev4r0hp6Q4MGzas3cePGxf5vM2ECcDvfie55CmnmDRYqwQCrBwiSkEMh9wcDgUbNkw+BL/8srwjnX22TDmLw5//LNvvftedBVNECVHBTiKVQ+pxCxZIuJKAmcn2pWhsBCoqpMlzokeD8+YBb75pvBZjxrTfNJqIyEluXco+uHIoSQyHIpswQdprxpCjuFL37pHPvRQWbkVBQeI/lKa13V7QVXRdTmyxcogo5TAc8ko4lJMj26VLZbtjR9zhUHk5MHw48PzzJo+NyEnJhkP33w+8917Cu9+3cyeyVVPoRA0eDJx8cuKPv/NO+SIi8gq3LmVvYuVQba1sGQ6FOvFEWXOFPKqhQcJdVg4RpRyGQ25crSwSdXrls89km0D/oYoK764MQRTRpk3A974nlxOdK/md78hXgooLC5HNpduJiOLjo8qh9PSkn4rIPWpqZMvKIaKU4+/JRWGrlbnaqFGy3bZNtgmGQ1xkiFLKPffI2eezzoq8LiwREbmTDyqHOK2MUpL6v8FwiCjl+DscCptWBrjo4CTc4MFAZqbx/Y4dcT9FeTkrhyiFHDgAvPqqdFlfvNgDCS8RER3lo8ohhkOUUlQ4xGllRCmH4ZBXwiFNkzUzFVYOkd+9956sE/zDHzo9EiIiihcrh4i8idPKiFIWwyGvhENA0uEQK4copaj/A2PHOjsOIiKKX3MzwyEiL2LlEFHKYjjkxXBozBhZ1l4l9zFobgYqK1k5RCmkrAzo1cuYAkBERN7R1HT0GMxVOnWSbUND0k/FcIhSEnsOEaUsF74r2yhstTLA5eHQMcfIVq2MpJbwjsGRI/KzMRyilFFWlvjy9URE5Cy3Vg4xHCKKjtPKiFIWwyEvVQ6ddhpQWAhcdpl8H8fUsooK2XJaGaWMsrLEl68nIiJnBVUOMRwi8hBOKyNKWf4OhyIsZe+ag5NINA2YMwcYNky+jyMcKi+XLSuHKGWwcoiIyLvcWjnUsoKaGeFQba1sGQ5RSmHlEFHK8nc45LXKISU7WwbMyiHyq7o66bvFcIiIyJvCjsFcc/ylaVI9VF+f9FOpyqH09KSfisg9WDlElLIYDnkxHOrcGRg4kJVD5F+7dsmW08qIiLzJrdPKAAmHOK2MKDI2pCZKWQyHvBgOAUBODiuHyB90HTjhBPlPqr5GjJDbGA4REXmTW6eVAQyHiKLhtDKilNXR6QE4yuvh0DffxHTX6mrg6aflMiuHyHM++ghYtgz4/veNUAgAuncHTjrJuXEREVHiWDlE5A11dcCjjwKVlfL9kiXyn7ZzZ2fHRUSmYzjkpaXsg+XkAP/+N9DYaDRPbMO99wKffy7BECuHyBP+9S/g97+X/5DbtgG9ewMLFrBxAxFRqvBR5RA/Q5OnffklcPvt8nmjJdDFjBnGhyciShn+nlbmtdXKgs2eLctgvPlm1LtVVgJPPQWcey6wdauxQiuRazU3A7/4BVBcLEfUo0cDjzzCYIiIKFXoeqtjMFcdf5kYDnXuzM/Q5HHV1bJdskR+qevqgKVLnR0TEVmClUNeDYfOPlum2Dz2GHDRRW3e7b//lX5DN97IKWVkk9pa4OGHgaqqxB5/8KAEQwsXApdfbu7YiIjIec3NsvVB5RCnlJHnsQE1kW8wHPJqONShA3D99cANNwDLlwPTp0e828GDsg1u1UJkqRdeAO68U46IEz1dOnEicOGF5o6LiIjcoalJtmHHYK5h4lL2DIfI81QDai5dT5TyGA55NRwCgB/+EPjVr4Crr5btBRe0uotawp69hshSq1dLtQ8APP44MGUK8PXXLjziJyIix4VVDgEuO/7q3NmUyqH6evYbohTAyiEi32A45OVwKCsLuOMO6c9yzz0Rw6GKCvkRu3d3YHzkDxUVwMyZodPI/vpXBkNERBRZhMohVx1/mTStrKGBvR4pBbByiMg3GA55dbUy5ZZbgLVrgU8+iXhzRYVkSGn+bj1OsaqtBfbsie8xCxdKMPSvfwEjR8qR8Jgx1oyPiIi8zyc9hxgOUUpQlUMMh4hSnr/DIQ+vVvb117JQ2V13AVqvXsDhwxHvV17ORtQUh299C/j00/gfN2MGcN555o+HiIhSjwqHvFQ5VFIiPfWmTZMlYGPAcIhSgqoc4qqxRCnP3+GQh6eVTZsm25tuArJ69waOHJEy7aD5+4BUDrHfEMXkyy8lGPrZzyTsiUdBgSVDIiKiFOSFaWV1daHX/f73wJNPAn37xhwONTYCHf19pE2pIBCQqiG2CyBKef5+y/JoOLRsmXF5zx4gq1cv+aa8HOjTJ+S+rBzysRtvlN4/saqrAzIzgYceki0REZEVvDCtLLiPHgBs2CDbw4dDjh+jYeUQpYSaGjajJvIJhkMeDIfef9+4vGcPMFqFQ4cPtwqHKiqAYcNsHBy5w+7dwJ/+BJxwAjB1auyPmzOHwRAREVnLC5VD4dPKNm6UbVOTHFypY68oGA5RSlCVQ0SU8hgOeTAcKioyLu/eDeMApaXvUFWV9AV+6CFOK/OFnTuBH/0ICAQwRf2D798v9ex//jOQl+f0CImIiAxurxzq3FnWoVcqKuSAa+JEWQRk/36GQ+QfrBwi8g1/r2Hl0dXKioqMnkN79sA4QDl0CACwaROwbx8wb57kRRGnlZ14InDNNXYMl6z2+98DH34IdO4MvWNHOajNzgZ+9SsGQ0RE5D5eqxxSZ+VmzZLtgQMxPQ17DlFKYOUQkW/4+y3Lg5VDui7HKPPmAatXh4VDLZVDO3YY949YObRyJbB0qXw99ZQdw05dq1YBTz/t7C/OK68AF14IvPIKviksRAGbQxMRkZu5vXIoPBxSU8pmzZL3/BjDoYYGLvBEKYCVQ0S+4e9wyINL2e/aJdPGxo0DBg6MMK2spgbHXz8b+7ANAFCF7vhv80cAco0nefJJ2XLVgeTdeCPw+edA797OjaFnT+CWW5zbPxERUTwiVA65Sng4tHOnbFXZdhzhEKeVkeexcojIN/wdDnmwckidvBozRsKhVpVDCxdiwPYV+HvaFejSqRkX1v0dIw4uR0g49M03stV1oLoayMiw8SfwoP/8B1i/vvX1VVVAYaE0d7r5ZtuHRURE5ElhlUOAy46/wsOh6mo5XhwyRL5nOER+UlPDpY+JfILhkMfCoc2bZTtqlIRD27dDapbT04HbbweysrCt5zG4q98LGDuwHBcu+Tt6Ve0IfZIdO+T+tbVyeexY238Oz9i1CzjnnNarlii9ewNXXWXvmIiIiLzMaz2HAgE5kZaRIcdPDIfITzitjMg3GA55LBzavl1OtGVnA4MGSdsgABL0AMDEiXjw8P3IydZw7Ik9UbUkA+n7txtPUFcn5UannAL8738MhwA5SP3gA3ltwr31lnSUXL0aGDas9e3p6dIAmoiIiGLjtZ5DgYB8ONY0oG9fNqQmf+G0MiLf8PdblgfDodJSCYY6dpQVVRcskMAoR91h8WK8Na4Hvn0icNvtGqqeycHYbkGVQ2re/MyZEg6dfjpwzz3AHXfY/aO4x7PPRl+57bzzgEmT7BsPERFRKnN75VDnzpHDISCucIiVQ5QSWDlE5BsMhzy2lH1pqVHAolZUHT0aeP6kP+HSH6SjLr0Hdu8GcnLk2Kb3MUOBsqDKIbWU2fHHG9c98ghwww1A9+72/BCRVFcbc+bs9vjjwNSpwHPPRb59zBh7x0NERJTKVOWQW8OhTp2A+nrj++rq0HDok0+AggIZ+C23AGedFfFpGA5RSmDlEJFv+Dsc8uBqZaWlwOzZclkVs9TVAZct+Tku/QTYuUWuy1GlREOHGg2oASMcGjnSuK6iAhg8uPURzPDhwLJl9tREf+97MoXLKS+9BBx7rHP7JyIi8gsvTCtrbjYqzIMrh37wA6OqaOlS4PXXGQ5R6tJ1Vg4R+Yi/wyGPTCtrbASuvhqYP19mhanKoY4d5W91ICDfNzW1NKhGWDi0d68kSF26GHcYOhT45z+lq/WyZcCWLaE7LSqSPjzBO7TShg1SCnXjjY80A/cAACAASURBVNbvK1y3bjK9joiIiKzn9mllKtFpaJBjJ9WQGgCuuEK+AGD8eODIkTafhj2HyPMaGuT/KyuHiHzB329ZHgmHtm0D/vpX+QJCs5rPPgPuvBP497+B3bsjhEPqwsSJcoSyd6+ssNWtG3D++XLbCSe03ul770k4tGOH9eGQrst+5s4FLrjA2n0RERGRs7xQOQSEhkP9+7e+X2YmUFnZ5tOwcog8r6ZGtqwcIvIFhkMuDYdefRXo1w84+WSgqir0tuCsZsoU4P/+T8Kh0lIjHBoypOUOZ50FXHmlsZoZAJx0UvsDGDpUtjt2RL+fGQ4fljcftU8iIiJKXV6qHAJCp5UFy8qKWjnEcIg8T01PYOUQkS8wHHJhOLR7N3DppXJZ141wKDsb+MlPgDlzQu+vioNUODRggKywDkCmjT3/fPyDSCQc+tOfgH/8A/j0U+MFjYXax9FEi4iIiFKWlyqHgNCG1MGysuSgrQ0Mh8jVdu0CHn00dGW+cCr8ZOUQkS8wHHLhamW//71se/eWrapYfu21yDPAVDi0fbt8mTILLDMT6NEj9nCooQG47jq5vHevhFKxUvtg5RAREVHqi1A55CqRKodUz6FgUaaV6br8mOw5RK71978Dv/sd0KtX9PsNGmSsgkNEKc3fb1kuXa1s2TLZqr/VqnIoMzPy/bt3lyCptFS+Jk40aSBDhxrBzU03SQNrQF6zBx4ALr4YWLsWuOSS0LLqTZsSC4dYOURERJT6wiqHAHccfx3VubNsk5hW1tgoW1YOkWtt2iQ9LPbtc3okROQSaU4PwFEunVamjjMOHZKtCoe6d2/7McOGGdPKjjajTpYKh7ZulXKmAQNkRbHGRuD22+X1e+ABuc8ppwCXXSaPKyqKbz9lZXJqLZ5AiYiIiLzJKz2H6utlYO2FQxEGr3IlhkPkWps2AaNHOz0KInIRhkMuDofKy+X4SVUsRwuHcnKAlSvl+MXUcGjtWgl+0tJkXtsLL0ggVFwMHHss8Mor0gjphReAl16SVT02bWr/ud96S85W9O4NPPIIMHhwyBlEIiIiAjRNO1PTtCJN00o0Tbs1yv0u1DRN1zQtP+i6X7Y8rkjTtDPsGXEMvNRzqLZWBtdWONTcbKzoFIThELlecTGQl+f0KIjIRfw9rcxl4dCTTwL/+pcRDuk6UFHR/rQyQCqH3npLLpsWDl1xBbBzp7xO111nTPu66CLgww+lkd3IkTLlDJCDvLy89iuHdB246y6Zv3/OOXLd7NkmDZqIiCg1aJrWAcATAL4FoAzAck3TFum6vj7sfpkA5gP4Iui68QAuAzABwGAA/9U0bbSu6012jb9NXqkcamgwVmtqq+cQIAduYeGRmlbGnkPkSlVVchzPyiEiCuLvtyyXhUNffAF8/LFczs6WXObQIakc6tjRmAIfSXAgZFo4dNJJkZe979QJeO65yI8ZMwZ45x1gwgQAwPTq6tYHVE1NEiAtWCBVR0RERBTJcQBKdF3fAgCapr0M4FwA68Pudw+ABwHcHHTduQBe1nW9DsBWTdNKWp5vqeWjbo+XKodUONRW5RAg4VDY1HhWDlFSGhuBzz6TyjUrlJbKluEQEQXxbzik665rSF1ebhxM5OYa4VBVlZyciraaR/AKZaaFQ4n4+c/lNW15Iav370dGv36t7zdnjlQmERERUVuyAQQvG1oGYEbwHTRNmwpgqK7r72iadnPYY5eFPTY7fAeapl0N4GoAGDBgAAoLC80ZeZiqqqqjz9171SpMBvDVypWorK1FRcUUaBpQWLjKkn3Hq9f69TgGwNdffIGGrCzMALB+2zbsC3tt+pSWYhKAFR99hKpdu0Ju27evC4ATsGVLEQoL217u3krBrzlZz8zXu++SJZh4552mPFc0XwQCqPHo7wh/v+3H19xeTrze/g6HAFcsZd/UBGzeLOGQkpsrJwxUOBSt3xBgBEJduwJ9+1o21PadfLJ8tVhfWIj+BQXOjYeIiChFaZqWBuB3AOYl+hy6ri8AsAAA8vPz9QKL3rMLCwtx9LlbmilOmz4dyM9Hr15y/GXVvuPWUtk0ddKko9VB4/PzMb6N8eWPHg2E3bZli2wnThyDgoIxVo00qpDXnCxn6utdXCxb1aPTCj17Ysa4cdY8tw34+20/vub2cuL1Zjjkgsqhiy4C3nxTAiFFXVbTytoLh1TlUE5O9AojIiIi8oydAIYGfT+k5TolE8BEAIWavPkPBLBI07RzYniscyJMK1NXuULwUvbV1XK5vWllYVQlOHsOUUJUk/OZM4E+fZwdCxH5hn9XK1NHIQ6HQ4cPSzAEANu2GdercOjwYWNaWTT9+slCYY5OKSMiIiIzLQeQp2nacE3TOkMaTC9SN+q6XqHrel9d13N1Xc+FTCM7R9f1FS33u0zTtC6apg0HkAfgS/t/hAi80pC6vj56Q2oVDqllZYOohtTsOUQJUeFQpFCSiMgi/j2f4ZJw6N13I1+vKoFirRxKS5MFv2bONHd8RERE5Axd1xs1TbsWwHsAOgD4i67r6zRNuxvACl3XF0V57DpN016FNK9uBPBzV6xUBnirIXV9vVyO9CE9eLWyMGxITUlRoWR6urPjICJfiSkc0jTtTACPQw5MntN1/bdt3O9CAK8DmN5y1gqapv0SwFUAmgBcr+v6e2YMPGkuCYcOHox8fe/eckJK9RyKZbrx+++bOzYiIiJylq7riwEsDrsuYqdaXdcLwr6/D8B9lg0uUV6pHHrtNWOs0aaV3X8/cPHFIQdrDIcoKTU1EgyxVwQR2ajdaWWapnUA8ASAswCMB/BdTdPGR7hfJoD5AL4Ium48pAR6AoAzATzZ8nzOc0k4VFER+fqsLGksvXdvbA2piYiIiDwhQuWQqwwaJAdiCxcCL78M9OrVaql6APLhfcQIWV722WdDbmI4REkJBDiljIhsF0vPoeMAlOi6vkXX9XoALwM4N8L97gHwIIDaoOvOBfCyrut1uq5vBVDS8nzOUwcmDq9WVl5+9NgoRFaW9A/avl2mlbXXc4iIiIjIE8IqhwCXVQ717Sul3TU18rVvn1ElFEzTgE2bJDhavz7kJtVziA2pKSE1NbIEMRGRjWJ5y8oGsCPo+zIAM4LvoGnaVABDdV1/R9O0m8MeuyzssdnhO9A07WoAVwPAgAEDUFhYGNPgE1FVVYXCwkJ0qK7GSQBKtmxBWWEh1q7tA2ASli9fgSNHqizbf7gNG8agZ8/eqKnpgEDA+OdYteoTdOkyGmvX9kJlZUccPrwThYVbbBuXWdTrTfbg620/vub24uttP77mZDq39xwCJNWJJdnp0AGYPBnYsCHkalYOUVJqalg5RES2S/p8hqZpaQB+B2Beos+h6/oCAAsAID8/Xy8oKEh2WG0qLCxEQUGBlOwAGDV6NEYVFBxdaGLatHzk51u2+1aeeALo3x+oqwO2tGQ/HTsCp58+G59/DrzX0qFp3LgcFBR4bymyo6832YKvt/34mtuLr7f9+JqT6dzecyhe48cDCxZI6NXyMzEcoqQEAqwcIiLbxTKtbCeAoUHfD2m5TskEMBFAoaZp2wAcD2CRpmn5MTzWOS7pOVReDvTsKRXMgLwPZGXJeNSKZQCQ3areioiIiMiDvFA5FI9x4+TD/A6j0J7hECWF08qIyAGxhEPLAeRpmjZc07TOkAbTR5dO1XW9Qtf1vrqu5+q6nguZRnZOy2pliwBcpmlaF03ThgPIA/Cl6T9FIlwSDlVUAD16GAtc5OYa09pzc437zZ5t77iIiIiILJFqlUPjxsl21ixgzx4A7DlESeK0MiJyQLtvWbquN2qadi2A9yBL2f9F1/V1mqbdDWCFruuLojx2naZprwJYD6ARwM91XW8yaezJcUk4VF4OjBwpC14AwC9+Ie8HQGjlUF6eveMiIiIiskSEYzBPh0PTpgGjR0tz6i++AM49l5VDlJxAABgwwOlREJHPxHQ+Q9f1xQAWh113Zxv3LQj7/j4A9yU4Puu4aLWyHj2MaqHvfhfo3FkuDw2akOe6ZV6JiIiIEpFq08q6dQPeflsCopYmlgyHKCmsHCIiB/i32FUdhThYOaTrMq2sZ0/gyiuBQYOMYAiQy489JlXKRERERCkh1aaVAcZZviNHADAcoiSx5xAROcC/4ZALppXV1gL19RIOjR8vX+Hmz7dvPERERESWS7XKIQDIzJRtS+UQew5RUgIBVg4Rke1iaUidmlwQDpWXy7ZHD/v2SUREROSoVKwc6tpVfh5OKyMzsHKIiBzAcMjBcKiiQrY9e9q3TyIiIiJHRagc8jxNk6llnFZGZggEGA4Rke0YDrFyiIiIiMg+YZVDQApUDgEytYyVQ5Ssxkb54rQyIrIZwyEHw6FDh2Tbq5d9+yQiIiJyVCr2HAIkHAqrHGLPIYpbTY1sWTlERDZjOOTgUvbbt8s2eMl6IiIiopSWij2HAJlWFtaQmpVDFLdAQLasHCIim/n3fIYLlrIvLZUzSoMG2bdPIiIiIkelaOWQnpmJuv2VOFDGyiFKAiuHiMghrBxyIBx6/nlgzx4Jh4YOPXpsRERERJT6VOVQUPV2KoRD++uysHnVEQwdCpSUAJ07p0izbbIXwyEicoh/z2c4FA6VlQHz5gE33yzh0LBh1u6PiIiIyFWam+XAK8XCoT3VmZiM9ViMs9D15Xpc1xXAqQDOOgu46Sanh0dewWllROQQhkM2h0Nr18r200+l59C3vmXt/oiIiIhcpbk5pGw6VcKh3ZWZmAzgLPwHnzbPRM+uGrCmSA74GA7R/v1AfX3799uxQ7asHCIimzEcsjkcWrdOtkuXypaVQ0REROQrTU0hy9inSjhUejgLALAvbSBOav4U98wH7tjxU+CttxweGTnunXeAuXPje0yPHtaMhYioDQyHbF6tTIVDCsMhIiIi8pUUrBxqaAC2HcwEABzMygXKgYkTAZRnHl3BjHzsiy8kEH3qqdgaUWVmAtOmWT8uIqIg/g2HHFqtbN06YPp02W1VFTBrlrX7IyIiInKVFKwcKioCKpq7AwAaBgwByoFJkwB8kyk9ZJqauAKJnxUVAbm5wNVXOz0SIqI2cbUym8OhDRuAGTOAZcuk/1BenrX7IyIiInKVCJVDXrdmDZAJqRAaMmMIbrsNGD4cUgECyBlB8q+iImDMGKdHQUQUFcMhG8OhhgapLO7f37p9EBEREblaWOUQ4P3KoTVrgLo0WV2q9ylTcN99LT+iCoc4tcy/mpuBTZsYDhGR6/l3WpkD4VB1tWzVcQIRERGR76Rgz6E1a4AdY68Bfj0QuPhi4waGQ95SXw907NgqvIyoocH4PAFAq68H6upa36+sDKipYThERK7HyiEbwyF1XNC9u3X7ICIiInK1FOw5tGYNMG5yJ+CSS0LnyamDPoZD7vfKK0CXLtJJvL1fyPffB9LTQ77mnHFGq+uQng6MGiWPGTvW+p+BiCgJrByycbUyNd2c4RARERH5VopVDh05ApSWttFrmJVD3rFhg7E9cADo16/t+y5ZItv77jv6IWLLli0YMWJE5PtnZQEnnWTiYImIzOffcMiB1cpUOMRpZURERORbKVY5tHatbCdNinAjwyHvCASMyxs3Rg+HioqAESOA2247etX2wkKMKCiwbnxERBbjtDJOKyMiIiKyT4pVDq1ZI1uGQx5XU2NcLiqKft+NGzlNjIhSDsMhByqHGA4RERGRbzU3p1zlUGYmMGxYhBsZDnlHTQ0wcKD0Cdq4se37NTcDxcVsME1EKce/08ocrBzitDIiIiLyrRSbVrZmjfQwDu5DfRTDIe8IBOQMbr9+wPr1oZVEwbZvB2prWTlERCmHlUOsHCIiIiKyj4unla1dKwtRRVqRHAD27QMWLwZ27ZLvdV3CoYhTygCgWzc51mQ45H41NUDXrsC4ccC778q/XaQvFQoxHCKiFMPKIYZDRERERPZxaeVQUxMwbRpQXw/88Y/Atde2vs/VVwNvvQWccgrw4YfA7t3AoUNRwiFNkwM/hkPup8Khu+8Gpk6Nft8ePYATT7RnXERENmE4ZONS9uq4ICPDun0QERERuZpLK4cCAQmGAODrryPf56uvjNtV1RAQJRwCZGoZwyH3CwSkMmjMGOAXv3B6NEREtvPvtDKHlrLv1i3keIiIiIjIX8Iqh9yiutq4vHp169sPHQLKyoDhw4HycrnMcCiFqMohIiKfYuWQzeEQm1ETERGRr7m0ckiFQ337AuvWSYYVfEJPBUZXXAHcc498v2YNMHgw0Lt3lCfOzJSSo5tvNq7r0gW44QbZGbkDwyEi8jn3nbaxi0OrlbHfEBEREfmaS3sOqXDo+ONlMaolS4CDB43bVTh0+eWy/fRTYNWqdqqGAGDmTGDvXuDJJ+XrT38C7rsP+Pe/Tf8ZKAlqWhkRkU8xHLK5cojhEBEREfmaSyuHAgHZzpol25NPBvr3B7Ztk+9Xr5ZCnzFjgBEjgN/+Vq475ph2nvj3v5fkSX3t2CHXB89jI+excoiIfI7TyjitjIiIiMg+HqgcevNN4PPPgYcekiwnN1eCoMmTZbxvvAGsXCk/xty5ce5IBRA1NWYOn5LFcIiIfI7hkM2rlfXpY93zExEREbmeSyuHVDiUkQHMmSO9hB56CDhyRPKstWuBn/5U7jNlinwlRAUQqlSJ3IHTyojI5/w7rcyh1co4rYyIiIh8zaWVQyqryciQbVaWbI8cATZvlsKSyZNN2FFaGpCeznDITZqagIYGVg4Rka/5NxxiQ2oiIiIi+zU3uzIcCq4cAkLDIdWM2pRwCJAKFYZD7qGm+DEcIiIf47Qym8IhXQf27wf69bPm+YmIiIg8wQPTygAjHKqoAHbvlnGOH2/Szrp2Zc8hN1FBHaeVEZGPMRyyKRwqLwfq6oBBg6x5fiIiIiJPaGoCOnU6+q3bwiGVD3TrJoeJR47IV1aWiYUlrBxyF1YOERFxWpld4dCuXbJlOERERES+5tLKoUBAxpKeLt9rmgRCR45IcKQqikzBcMhdGA4REbFyyK7Vynbvli3DISIiIvK1CA2p3UAFQMHjUeFQba3J4VDXrgyH3ESFQ5xWRkQ+xnDIpsohhkNEREREaFU5BLijcihSdZAKhxoaLKgcYs8h91BBHSuHiMjH/BsO2byUPcMhIiIiIrh6KfvwwhEVDjU3WxAOlZeb+ISUFE4rIyJizyE7w6GMDCAz05rnJyIiIvIEl/YcilY5FAiw51BK47QyIiKGQ3aGQ6waIiIiIt9rbg5p7OOFcMj0htTsOeQunFZGROTjaWUMh4iIiIjs58HKofR09hzyvL//HXj33ci3bdkiW4ZDRORjDIdsCId0Hdi4ETjzTPOfm4iIiMhTXFo5FAgAvXqFXqfCoaYmk2cccVqZ/e68E9i3D+jXL/LtJ57IM7lE5GsMh2xYyn7tWnkvKigw/7mJiIiIPEXXXdmQuq3KoepqCYcsmVam6yFBGVmkthbYuhW44w7grrucHg0RkSv5t+eQjauV/fe/sj31VPOfm4iIiMhTXFo51FY4BEi2YPq0Ml0H6upMfFJqU3Gx/N6NHev0SIiIXMu/4ZCN08o++gjIywNycsx/biIiIiJPCauWcUM4pOuysnz4qrI9ehiXTQ+HAPYdssvGjbIdN87ZcRARuRjDIRvCoTVrgPx885+XiIiIyHNcOK1s+3apHAovLOnf37hsSTjEvkP22LBBftFGj3Z6JERErsWeQxaHQ4EAUFoK/PCH5j4vERERkSdFmFbmtG++ke0xx4ReP2CAcdn0nkOA/8KhsjLgBz+w/+fesgUYNszkruJERKmF4ZDF4VBxsTwnpzgTERERoVXlkLrKSSocmjgx9HrLwiG/Vg69/z7wv//JKi2dO9u33ylTgO98x779ERF5EMMhi1crU1OcGQ4RERERwXUNqQMB4K23gJEjW/ccCg6HTF/KHvBfz6ENG4AuXWS1lg4dnB4NEREF8W84ZNNqZRs3ynPn5Zn7vERERESe5LKeQ+eeC3z1FXDxxa1v69LFuGzJtLLPPgutHhoxAsjNNXFHLrNhg/T9YTBEROQ6/g2HbJpWtmmTrFKmjgGIiIiIfM1FlUPNzcDSpXIS77HHot/X1HBIdbq+6abQ64cNA7ZtM3FHLrNhAzB9utOjICKiCBgOWRwOHT4M9Otn7nMSEREReZaLKoc2b5ZVym69FRg8OPp9TQ2Hxo0DVq0CKiqM6556Cli0yMSduExNDbB1K3DllU6PhIiIImA4ZHE4VFnZev46ERERkW+5qHKorVXKIjE1HIq00w8+kABF192xhJvZiorkZxs3zumREBFRBGnt3yVFhYVDitkHJ1VVQPfu5j4nERERkWe5pHKosRF4/XVpfzNhQtv369hyKtX0cChct27yQtTVWbwjh2zYIFuGQ0RErsRwyOLVylg5RERERBQkQuWQE265BXjlFWD8eCA9ve37ffvbsrUlHAJSdwWzDRskFBw92umREBFRBP4Oh2w4MGHlEBEREVGQsGlTVp2ga8+uXbJ97bXo9/vHP4Cvv7YhHFKrlwSvXpZKNmwARo4MXQKOiIhcw7/hUISSZnW1mVg5RERERBTEpmOw9tTVAZMmAWPGRL9ft27AscfaMCBVOZTK4RCnlBERuZa/G1JbfGDS1CTv7wyHiIiIiFq4ZFpZbW306WS28/q0sspKYMkSo3VDMF0HNm0C5s61f1xERBQThkMtrAiHqqtly2llRERERC3CKoeCr7ZTXZ3LZjh5fVrZAw/IVzRTp9ozFiIiihvDoRZWhEOVlbJl5RARERFRizYqh5wIh1Sxjit4vXJo1Spg7FjgxRcj396lS/Rl4YiIyFH+DocsPjCpqpItK4eIiIiIWrik51BtLdCrl737jMrrlUPr1wMnngjk5zs9EiIiSoB/G1I3NwMdOhz9lpVDRERERDZwUeWQK3sOeTEcqqoCSkuB8eOdHgkRESXIv+FQU5PlZ61YOUREREQUxkWVQ67qOeTlaWUbN8qW08aIiDzLv9PKmppYOURERERkNxdVDrkqHPLatLLKSuC662S7a5dcx8ohIiLPYjjUgpVDRET/v717D9KrrvM8/vmmL6HTJMglBEjCvQETcIMTLqvCtihLXChwphwBd2oZh5kMFkzhqLWDu4guJVWuO8PolkEnpThoLZt10XEzVrbUUtrLKhCQCCaZaAiQ0EnIlXQ63aRv3/3jnGOf7n6e7qe7n/M7p/u8X1Vd5/I8zzm//Drp/OrT39/vAEAABaocKuS0splSOfTkk9Jjj0ltbVHKduON0gUX5N0qAMAUlTccYs0hAACA8KgcqmymrTm0eXO0ffZZacGCfNsCAJg21hyKUTkEAAAQgHthwqFCVQ41NUW/uJwp4dCWLdLSpQRDADBLlDscClQ5RDgEAAAQK8C0sqEhqa+vYJVDUlQ9NFOmlW3ezBpDADCLlHdaWaA1h1paRtwGAACg3AowrayvL9oWqnJIisKhri5pYEBqTA3TDx+Wjhyp7RoNDdKSJWPPDw1Ju3bVp6Pdpa1bpXe/e/rXAgAUAuFQLIuByZEjVNoCAACMUKVyKKQ334y2hascammRvvY16Uc/kl5+OTp3+LB05pnRPLharVkztqrns5+VPv3p+rVVkt72tvpeDwCQm/KGQ0NDmZc0v/aatHhx/a4HAAAw442qHEqErBxKcpZCVg5J0iuvDJ/bvz9q8F/+pXT11RNf42/+RnrqqbHh0C9/KZ1/vvSpT9WnrXPnSn/4h/W5FgAgd+UNhwJUDr36qnTxxfW7HgAAwIyWDLRyXnOosJVDlaqDkgWqb7ihtjBm3brhJ4mlbd4sXXut9Kd/Oq0mAgBmJxakHqVeAxN3aedO6eyz63M9AACAGS8ZaOW85lBhK4d27Bh7LlmguqWltmssXx49SWxwcPjckSPRekPLl0+/jQCAWYlwKMWsfgOTw4ejBanPOac+1wMAAJjxClI5lIRDhascqtQJSeVQMuVsIsuXS2++qZa9e4fPbdkSbS+9dHrtAwDMWuWdVjY0lGk4tHNntKVyCAAAIDY0FG1zrhwq7LSyxIknDu9PpXJI0kV/93fS974XnUsGplQOAQCqKHfl0JyRf/x6hkOvvhptqRwCAACIFaxyqHDTyh59NNomgZA0+cqht71NuvZazT1wQPrVr6KvAwek971POvfcujYXADB7lLdyKONpZckvaAiHAAAAYlQOje/DH5Y6O6MnivX3S01Nk68cammRfvITPdPRofb29syaCgCYXcpdOZRhOLRrVzTgOO20+lwPAABgxmNB6oklIVASCiXbWiuHAACYgvKGQxmvObRnj3TmmSPGPgAAAOVWkGllha0ckoZDoGQ6WbKttXIIAIApKG84lPGaQ0k4BAAAgFhBppXNyMohwiEAQIbKHQ5lWDm0ezfhEAAAwAjjVA6FNOMqhxoaovWHAADICOFQCpVDAAAAGapQOZSgcihWqXKopYW1CgAAmSIcSqlXONTbK73xBuEQAADACAVZcygJh2ZM5RCLUQMAMlZTOGRmq8xsm5ltN7P7Krx+l5m9aGabzOznZrYsPn+umfXG5zeZ2Vfq/QeYsqGhzNYc2rs32hIOAQAApBRkzaFkWtmMqhwCACBDjRO9wcwaJK2RdL2k1yRtNLP17r4l9bbH3f0r8ftvlvSwpFXxay+5+4r6NrsOMqwc2rMn2hIOAQAApBSscqi5Odw9a0blEAAgB7VUDl0pabu773D3PknrJN2SfoO7d6UOWyUF/O99igiHAAAAwipQ5VBT05gi8mKgcggAkIMJK4ckLZa0K3X8mqSrRr/JzO6W9DFJzZKuS710npk9L6lL0v3u/rMKn10tabUkLVq0SB0dHbW2f9K6u7vV0dGhK7q61HPokDan7jU09C7t2rVXHR3bp3WPn/70LEkXaceO/6c33uifXoNnuKS/EQb9HR59Hhb9HR59jroqsOcwaAAAIABJREFUSOXQsWMFLsZJGpYOhwrbWADAbFFLOFQTd18jaY2ZfUjS/ZLukLRH0tnuftDM/kDSd81s+ahKI7n7WklrJWnlypXe3t5er2aN0dHRofb2dqmlRa1nnKH0vRobpcWLl6i9fcm07vHjH0cDnVtueefo4qTS+X1/Iwj6Ozz6PCz6Ozz6HHVVkMqhQj9VNqkSSk8rmz8/v/YAAEqhlmLaTklLU8dL4nPVrJP0fkly9+PufjDef07SS5IumlpT62xwMLMFqffvl049dcysNQAAgHJLBlqEQ9UxrQwAkINawqGNktrM7Dwza5Z0m6T16TeYWVvq8EZJv4vPL4wXtJaZnS+pTdKOejR82iqsOTRnzvAvtKZj/37p9NOnfx0AAIBZpSDTymZEOMSC1ACAgCacVubuA2Z2j6TvS2qQ9Ki7bzazByU96+7rJd1jZu+V1C/psKIpZZJ0raQHzaxf0pCku9z9UBZ/kEmrEA41NUn9dVgiaN8+aeHC6V8HAABgVinAtDL3godDc+ZIc+dSOQQACKqmNYfcfYOkDaPOPZDav7fK574t6dvTaWBmhobGhEPNzfUJh/bvly69dPrXAQAAmFXGqRwKpasrylsKGw5JUaUQlUMAgICK+ADPMCqsOdTUJPX1Tf/S+/YxrQwAAGCMCpVDiVCVQ3v2RNtCh0MtLVQOAQCCKnc4lMG0soEB6dAhppUBAACMUYA1h3bvjrZnnRXmflMyb5505Ij0xhs8yh4AEAThUEo9wqGDB6Mt4RAAAMAoBVhzaEZUDs2fLz3xhHTyycPHAABkqKY1h2alCuFQPdYc2rcv2hIOAQCA6TKzVZK+qOihIF9198+Nev0uSXdLGpTULWm1u28xs3MlbZW0LX7rU+5+V6h2V1WAyqG9e6PtGWeEud+UfPnL0i9+Ee03Nkq3355vewAAs155w6GhoUzWHNq/P9qy5hAAAJgOM2uQtEbS9ZJek7TRzNa7+5bU2x5396/E779Z0sOSVsWvveTuK0K2eUIFqBw6cCAa8y1YEOZ+U3LVVdEXAACBMK0sZbrTyn77W+k974n2qRwCAADTdKWk7e6+w937JK2TdEv6De7elTpslRQoYpmiAlQOHT4czdYK/ZQ0AACKrLyVQxmEQ0n17623SpdcMo22AQAASIsl7UodvyZpTDmJmd0t6WOSmiVdl3rpPDN7XlKXpPvd/WcVPrta0mpJWrRokTo6OurW+LTu7m51dHRo3s6dulLSlq1btS++19atp0tapqeffkZ79vRkcv+0bduWae7cE9XR8Uzm98pT0ucIg/4Oi/4Ojz4PK4/+JhxKaW6WeqYxJtm5M9o+9tiYSwMAAGTC3ddIWmNmH5J0v6Q7JO2RdLa7HzSzP5D0XTNbPqrSSO6+VtJaSVq5cqW3t7dn0saOjg61t7dLW7dKkpYtX65l8b2SNYCuuOJKvfWtmdx+hIcekhYvlrL6sxbF7/scQdDfYdHf4dHnYeXR3+WdVjY0VLFyaDprDu3cGS1uOHfuNNsGAAAgdUpamjpeEp+rZp2k90uSux9394Px/nOSXpJ0UUbtrF1BppWdckqYewEAMFOUNxwaHKy4IPVkp5Ul6ypKUTh09tl1aBsAAIC0UVKbmZ1nZs2SbpO0Pv0GM2tLHd4o6Xfx+YXxgtYys/MltUnaEaTV4ynAgtTJmkMAAGBYOcMh96qVQ5MJhw4ejC6xdm10vHOndM45dWwnAAAoLXcfkHSPpO8reiz9t9x9s5k9GD+ZTJLuMbPNZrZJ0bpDd8Tnr5X0Qnz+CUl3ufuhwH+EsQpSOUQ4BADASOVccyj5rdUkw6EdO6TOTumaa6LjT3862j7xhPQXfxGFQzfdlEF7AQBAKbn7BkkbRp17ILV/b5XPfVvSt7Nt3RRUqBwKffs33iAcAgBgNMKhlObm8dccuuiiaDaae/T1zW9G5xcvlg4ckHp7mVYGAABQVc6VQ0eORPchHAIAYKRyhkODg9F2kmsOJR/r65OOHZO64ud9HDkivfJKtE84BAAAUEXOaw4dPhxtWZAaAICRyh0OTXHNob17hwcXUhQOvfBCtL98eZ3aCAAAMNvkXDl0KF51icohAABGIhxKqTUc2rNHev31aH/hwmju+q9+Jc2fL11wQZ3bCgAAMFsUpHKIcAgAgJHK+bSyKuHQRGsOJfbsiRaflqTLLosqh55/Xrr88jEz1QAAAJDIuXKIcAgAgMrKGWUkv7Wa5JpDc+dG2927pVdfjY7b2qIS5V//OgqHAAAAUEXOlUMHD0Zb1hwCAGAkppWlNDVFY5ahocoVQPPmScePD1cOLV0qveUtw7+FWrEi43YDAADMZEkClFM4tHt3NMY7/fTs7wUAwExSzsqhccIhqXr1UE9PtN2zR3r5Zemcc6JwKHH++XVuJwAAwGyS87Syzk5p0SKpsZy/HgUAoCrCoZTm5mhbad2h/v6oakiSvv51aeNG6ZprpJNOGn4Pj7EHAAAYR87TynbvlhYvzv4+AADMNOUMh5KBySQqh7q7R358yRLpE58YGQ4x2AAAABhHASqHGK8BADBWOYtqk8qhCgtSS5XDoaNHo+2nPiVdeKH07ndLra0jw6Hk8wAAAKhgnMqhEDo7o8pvAAAwUrnDoUlUDiXh0KWXSh/84PD5dDgEAACAcVSoHBr9UlZ6e6OHiFA5BADAWOWcVjaFNYeSaWUnnjjyfBIOjboUAAAARstxzaHdu6Mt4RAAAGOVMxyawppDSeXQ/Pkjzy9YEG2pIAIAAJhAjmsOdXZG27POyvY+AADMROUMh6ax5tDoyqGFC6Ptgw/WsX0AAACzUY6VQ/v2RdtFi7K9DwAAMxFrDqXU8rSy0ZVD8+aFeboGAADAjJdj5VBPT7Rtbc32PgAAzETlrhyqsubQ6HDo2DGpoyPaHx0OAQAAoEY5Vg719kbblpZs7wMAwExUznBogjWHRi9I/ZnPSI8+Gu2PnlYGAACAGiUJUI7h0Lx52d4HAICZqJzh0CTXHHr++eF9BhQAAABTVIBpZVQOAQAwFmsOpVQLh7ZulZYulT784RG/6AIAAMBk5DytzEyaOzfb+wAAMBMRDqVUWnNo715p927p4Yelv/7rQO0DAACYjXKsHOrtlU44gV/0AQBQSbmnldWw5tBzz0XblSsDtAsAAGA2G6dyKGu9vSwPAABANeUMhyZYkDqpHNq9W/rhD6Py47e/PWD7AAAAZqMKlUOjX8pKby/rDQEAUE25p5WNsyD1pk3S5ZdHxzfeKLW2BmwfAADAbJTjmkM9PYRDAABUU87KoQmmlX3849KGDcPnb7opULsAAABms5zXHCIcAgCgMsKhlGRB6tdfl774xWj/iiukP/7jgG0DAACYrXJ+WhnhEAAAlZVzWtkEaw5J0r590fFTT1WcFg8AAIDJyqFyyF36znekI0dYJgAAgGrKGXtMsOZQ4owzCIYAAADqJofKoSeflD7wAemZZ6gcAgCgmnJGHxOsOZQ466xA7QEAACiDJAEKGA51dQ3vEw4BAFAZ4VBKsuZQgnAIAACgjpLKoYDTyubOHd4nHAIAoLJyhkM1rDkkSWeeGag9AAAAZZBD5dDx48P7hEMAAFRWznCoyppDZtLLL0uPPx4dUzkEAABQRzksSH3s2PD+vHnZ3AMAgJmunE8rqzKtTJLOPVc6dCjaJxwCAACooxwWpE6HQ1QOAQBQWbkrhyqEQ5K0YoX0hS9If/RHAdsEAAAw21WoHMpad/fwPuEQAACVUTlUwZw50r33BmwPAABAGVA5BABAIZWzcqjCkzIAAACQsQqVQ8nv6pLf3dUb4RAAABMrZzoyQeUQAAAAMlChcqgxrmMfGMjmlulw6IQTsrkHAAAzHeEQAAAAwqhQOdTUFG37+7O5ZToc6uvL5h4AAMx0hEMAAAAIo0LlUBIOZVU5lF6Quqcnm3sAADDTlTMcSgYmhEMAAADhJJVDFaaVhagc6u3N5h4AAMx05X5aGQtSAwAAhFPhoSAhppVdeKF02mnSn/95NvcAAGCmK3c4ROUQAABAOBUqh7KeVnbsmNTWJm3YkM31AQCYDcpZOkM4BAAAEF6FBalDTCtrbc3m2gAAzBblDIeS0UfyqyoAAABkb5wFqQmHAADIT3nDoTlzWHMIAAAgpHEeZZ/l08oIhwAAGF8505H+fqqGAAAAQqtQOcS0MgAA8kc4BAAAgDAqVA4lS0BmEQ4NDkrHj0snnlj/awMAMJsQDgEAACCMCpVDZtGwLItw6NixaEvlEAAA4yMcAgAAQBgVKoekaGpZFmsO9fRE23nz6n9tAABmE8IhAAAAhFGhckjKrnKory/azp1b/2sDADCbEA4BAAAgjKRyKHA4xLAPAIDxEQ4BAAAgjKRyaNS0sqambKaVJYFTc3P9rw0AwGxCOAQAAIAwqlQONTZSOQQAQJ4IhwAAABBGlQWps5pWRuUQAAC1IRwCAABAGOMsSJ3FtDIqhwAAqA3hEAAAAMIIPK2MyiEAAGpDOAQAAIAwcnqUPcM+AADGRzgEAACAMNzHrDck8bQyAADyRjgEAACAMIaGxlQNSTytDACAvBEOAQAAIIxxKodYcwgAgPwQDgEAACCMKpVDPK0MAIB8EQ4BAAAgDPeg08qoHAIAoDaEQwAAAAhjaCjotDIqhwAAqA3hEAAAAMKoUjnEmkMAAOSLcAgAAABhVFmQurGRNYcAAMgT4RAAAADCGGdB6nTl0De+IV1//fRvR+UQAAC1acy7AbkgHAIAAAivxkfZ33FHtN27VzrjjKnfjsohAABqQ+UQAAAAwqjxUfZLlkTbX/5yerdLAqfGcv46FACAmhEOAQAAIIxx1hxKVw5deGG0/cUvpne7ZMhXIY8CAAAp5QuH3KXBQcIhAACA0MapHNq3T7r5ZqmnZ3g62N/+rfSZz0z9dn19rDcEAEAtShcOWVKzTDgEAAAQ1jhrDknSP/+ztH27dPSoNHdudO5LX4o+NhUUiwMAUBvCIQAAAIRRpXIovSZQV1f0deut0le/Kh08KG3bNrXbUTkEAEBtagqHzGyVmW0zs+1mdl+F1+8ysxfNbJOZ/dzMlqVe+2T8uW1mdkM9Gz8VcwYHox3CIQAAgLDcq04rSxw9Gn3Nny+94x3RuamuPUTlEAAAtZkwHDKzBklrJL1P0jJJt6fDn9jj7n6Zu6+Q9HlJD8efXSbpNknLJa2S9Eh8vdxQOQQAAJCToaFxp5VJw5VD8+dLF18snXKKdOed0rp1k78dlUMAANSmlsqhKyVtd/cd7t4naZ2kW9JvcPeu1GGrpGRm+C2S1rn7cXd/WdL2+Hq5IRwCAADISZXKofS0sgMHosfaL1gQ5Uif/Wx0fv36yd+OyiEAAGrTOPFbtFjSrtTxa5KuGv0mM7tb0sckNUu6LvXZp0Z9dnGFz66WtFqSFi1apI6OjhqaNTWDXVGO9S8vvaS9Gd4Hke7u7ky/nxiJ/g6PPg+L/g6PPkddTbAgtSR1dkbb+fOj7Uc+Eq09dPTo5G9H5RAAALWpJRyqibuvkbTGzD4k6X5Jd0zis2slrZWklStXent7e72aNcbT8Yjjkssu0yUZ3geRjo4OZfn9xEj0d3j0eVj0d3j0OepqnEfZJ0aHQ8l+V5cmjcohAABqU8u0sk5JS1PHS+Jz1ayT9P4pfjZzTCsDAADISZXKofS0st27o+2CBcPn5s+ncggAgCzVEg5tlNRmZueZWbOiBaZHzPo2s7bU4Y2Sfhfvr5d0m5nNNbPzJLVJemb6zZ46wiEAAICcTLFyaMECKocAAMjShNPK3H3AzO6R9H1JDZIedffNZvagpGfdfb2ke8zsvZL6JR1WPKUsft+3JG2RNCDpbncfzOjPUhPCIQAAgJxMYc2hZJ/KIQAAslPTmkPuvkHShlHnHkjt3zvOZx+S9NBUG1hvcwbjbIpwCAAAIKwqlUPpaWVJhdDoyqGphEP9/VJr6+Q/BwBA2dQyrWxWoXIIAAAgJ1Uqh4aGxr519JpDvb3RI+4ng8ohAABqQzgEAACAMKpUDvX0jH3r6MohaWT1UC2VRKw5BABAbQiHAAAAEIZ7xXDo2LGRx3PmSCeeOHycBEXJlLNNm6STT5Z+85vxb0flEAAAtSldOMSaQwAAADkZGqo4rWx05dBll0kNDcPHSTiUVAv9+tfS4KD04ovj347KIQAAalO6cIjKIQAAgJxUqRw66aRom4RA73znyNdHTyvbuXPkthoqhwAAqA3hEAAAAMKosiD1X/2V9LWvSe95T3R81VUjXx89rWzXrpHbaqgcAgCgNoRDAAAACKPKgtRNTdKf/Zm0dGl0PDocmmrlUH8/lUMAANSidOEQaw4BAADkpErlUOJzn5Oeekq6+OKR56daOdTXx5APAIBalC4conIIAAAgJ1UqhxLz5o2tGpJGVg65UzkEAEC9NebdgNAaenujnfTzUQEAAJC9CSqHqklXDh05InV3SwsXSvv3S88/H01HO+006c03pZdfHv4clUMAANSmdJVDDcmzUgmHAAAAwpqgcqiaxkaptVU6cEDasSM6d9110fbtb5cuuCCqErrjDmnZsuEvaThYAgAA1ZWzcqilJRplAAAAIJwpVg5J0ooV0jPPSBddFB1/9rPS7bdLP/iB9MgjUmentHmzdPXV0kc/Gr2noUG64YY6tR0AgFmsdAlJ47Fj/AoJAAAgD1OsHJKkd71Levhh6fTTpSVLomqhCy+M1il65BHp1VejrzvvlG69tc7tBgBglivntDLCIQAAgPDcpxwOXXNNNHVs/fpoP7nMOedE2xdeiNYiOvvsOrUVAIASKV841Ns7/MgLAAAAhDM0NOVpZe94x/BH29uHzy9dGm1/9rNoSzgEAMDklW9aWU+PdNJJeTcDAACgfKZROXTyyVF10IEDUVCUaGmJnlz2859Hx0klEQAAqF3pwqGGnp5oojoAAADCmsaC1JK0fHnl82efLT333PA+AACYnHJOK2PNIQAAgPD6+6NHiNVZUi3U3BxVEQEAgMkpXTjUeOwYaw4BAADk4eBB6dRT637ZSy8d3k6jMAkAgNIq3X+fVA4BAADkZP/+TEp7HnhA+u1vpZ/8pO6XBgCgFMq15tDgoBrefJNwCAAAILShoahy6LTT6n7phgapra3ulwUAoDTKVTnU3R1tCYcAAADCOnw4CohYFAgAgMIpVzh09Gi0Zc0hAACAsA4ciLYZVA4BAIDpKWc4ROUQAACYAcxslZltM7PtZnZfhdfvMrMXzWyTmf3czJalXvtk/LltZnZD2JZXsH9/tKVyCACAwilXONTVFW0JhwAAQMGZWYOkNZLeJ2mZpNvT4U/scXe/zN1XSPq8pIfjzy6TdJuk5ZJWSXokvl5+qBwCAKCwyhUOUTkEAABmjislbXf3He7eJ2mdpFvSb3D3rtRhqySP92+RtM7dj7v7y5K2x9fLTxIOUTkEAEDhlOtpZSxIDQAAZo7Fknaljl+TdNXoN5nZ3ZI+JqlZ0nWpzz416rOLK3x2taTVkrRo0SJ1dHTUo91jdHd3a8fTT+t8ST/dskVDL72UyX0wrLu7O7PvJ8aiv8Oiv8Ojz8PKo7/LFQ4dPx5tTzgh33YAAADUibuvkbTGzD4k6X5Jd0zis2slrZWklStXent7eyZt7Ojo0PkLFkitrbr2hvyXPyqDjo4OZfX9xFj0d1j0d3j0eVh59He5ppX19UXbpqZ82wEAADCxTklLU8dL4nPVrJP0/il+Nnv79zOlDACAgipnONTcnG87AAAAJrZRUpuZnWdmzYoWmF6ffoOZtaUOb5T0u3h/vaTbzGyumZ0nqU3SMwHaXF1Xl7RgQa5NAAAAlZVrWhnhEAAAmCHcfcDM7pH0fUkNkh51981m9qCkZ919vaR7zOy9kvolHVY8pSx+37ckbZE0IOludx/M5Q+S6OmR5s3LtQkAAKCycoVD/f3RlnAIAADMAO6+QdKGUeceSO3fO85nH5L0UHatm6TeXqmlJe9WAACACphWBgAAgOz19lI5BABAQZUzHGJBagAAgLB6eqgcAgCgoAiHAAAAkD2mlQEAUFilC4eGGhsls7xbAgAAUC4sSA0AQGGVLhzyxnKtwQ0AAFAIVA4BAFBY5QqH+vs1xJQyAACA8FiQGgCAwipXOETlEAAAQHA2MCANDFA5BABAQZUuHBoiHAIAAAhqzvHj0Q7hEAAAhVS6cMiZVgYAABBUQxIOMa0MAIBCKl04ROUQAABAWFQOAQBQbKULh1hzCAAAICzCIQAAiq1c4RBPKwMAAAiOaWUAABRbucIhKocAAACCo3IIAIBiIxwCAABApua8+Wa0Q+UQAACFVLpwiGllAAAAYTX09UU7VA4BAFBIpQuHqBwCAAAIi2llAAAUW+nCISqHAAAAwmJaGQAAxVa6cIjKIQAAgLCYVgYAQLGVKxzq79cQ4RAAAEBQv68cIhwCAKCQyhUOUTkEAAAQ3BwqhwAAKLTShUOsOQQAABBWw/HjUnOz1NCQd1MAAEAFpQuHqBwCAAAIa86bb7IYNQAABVa6cIjKIQAAgLAaenul1ta8mwEAAKooXTjklDMDAAAE1dTVJZ16at7NAAAAVZQnHHKXBgbkVA4BAAAE1Xj0qHTKKXk3AwAAVFGecKi/X5J4lD0AAEBgVA4BAFBs5QmH4keoUjkEAAAQVlNXF5VDAAAUWOnCISqHAAAAAnJnWhkAAAVXunCIyiEAAICAjh3TnIEBppUBAFBgpQuHhnhaGQAAQDgHD0ZbKocAACis8oRD8YLUVA4BAAAEdOhQtCUcAgCgsMoTDrHmEAAAQHhJOMS0MgAACqt04RCVQwAAAAExrQwAgMIrXThE5RAAAEBATCsDAKDwShcOOeEQAABAOIRDAAAUXnnCoeZm6ZJLNNjamndLAAAAymPePB0791zphBPybgkAAKiiPOHQFVdIW7eqa9myvFsCAABQHh/9qDZ+/et5twIAAIyjPOEQAAAAAAAAxiAcAgAAAAAAKDHCIQAAAAAAgBIjHAIAAAAAACgxwiEAAAAAAIASIxwCAAAAAAAoMcIhAAAAAACAEiMcAgAAAAAAKDHCIQAAAAAAgBIjHAIAAAAAACgxwiEAAAAAAIASIxwCAAAAAAAoMcIhAAAAAACAEiMcAgAAAAAAKDHCIQAAAAAAgBKrKRwys1Vmts3MtpvZfRVe/5iZbTGzF8zsR2Z2Tuq1QTPbFH+tr2fjAQAAAAAAMD2NE73BzBokrZF0vaTXJG00s/XuviX1tuclrXT3HjP7iKTPS7o1fq3X3VfUud0AAAAAAACog1oqh66UtN3dd7h7n6R1km5Jv8Hdn3T3nvjwKUlL6ttMAAAAAAAAZGHCyiFJiyXtSh2/Jumqcd5/p6T/mzo+wcyelTQg6XPu/t3RHzCz1ZJWS9KiRYvU0dFRQ7Ompru7O9PrYyT6Oyz6Ozz6PCz6Ozz6HAAAYParJRyqmZn9iaSVkv5N6vQ57t5pZudL+rGZvejuL6U/5+5rJa2VpJUrV3p7e3s9mzVCR0eHsrw+RqK/w6K/w6PPw6K/w6PPAQAAZr9appV1SlqaOl4SnxvBzN4r6T9Lutndjyfn3b0z3u6Q1CHp8mm0FwAAAAAAAHVUSzi0UVKbmZ1nZs2SbpM04qljZna5pH9QFAztS50/2czmxvunSXqnpPRC1gAAAAAAAMjRhNPK3H3AzO6R9H1JDZIedffNZvagpGfdfb2k/ybpREn/28wkaae73yzprZL+wcyGFAVRnxv1lDMAAAAAAADkyNw97zaMYGb7Jb2a4S1Ok3Qgw+tjJPo7LPo7PPo8LPo7vKz6/Bx3X5jBdTFFGY/B+LcbHn0eFv0dFv0dHn0eVvDxV+HCoayZ2bPuvjLvdpQF/R0W/R0efR4W/R0efY564O9RePR5WPR3WPR3ePR5WHn0dy1rDgEAAAAAAGCWIhwCAAAAAAAosTKGQ2vzbkDJ0N9h0d/h0edh0d/h0eeoB/4ehUefh0V/h0V/h0efhxW8v0u35hAAAAAAAACGlbFyCAAAAAAAADHCIQAAAAAAgBIrTThkZqvMbJuZbTez+/Juz2xhZo+a2T4z+03q3Clm9kMz+128PTk+b2b23+PvwQtm9vb8Wj4zmdlSM3vSzLaY2WYzuzc+T59nwMxOMLNnzOzXcX//l/j8eWb2dNyv/8vMmuPzc+Pj7fHr5+bZ/pnKzBrM7Hkz+158TH9nyMxeMbMXzWyTmT0bn+NnCuqGMVj9Mf4Ki/FXeIzB8sEYLKyijcFKEQ6ZWYOkNZLeJ2mZpNvNbFm+rZo1/lHSqlHn7pP0I3dvk/Sj+FiK+r8t/lot6cuB2jibDEj6uLsvk3S1pLvjv8v0eTaOS7rO3f+VpBWSVpnZ1ZL+q6S/d/cLJR2WdGf8/jslHY7P/338PkzevZK2po7p7+y9291XuPvK+JifKagLxmCZ+Ucx/gqJ8Vd4jMHywRgsvMKMwUoRDkm6UtJ2d9/h7n2S1km6Jec2zQru/lNJh0advkXSY/H+Y5Lenzr/DY88JektZnZmmJbODu6+x91/Fe8fVfTDe7Ho80zE/dYdHzbFXy7pOklPxOdH93fyfXhC0nvMzAI1d1YwsyWSbpT01fjYRH/ngZ8pqBfGYBlg/BUW46/wGIOFxxisMHL7uVKWcGixpF2p49fic8jGInffE+/vlbQo3uf7UEdx+eblkp4WfZ6ZuLx2k6R9kn4o6SVJb7j7QPyWdJ/+vr/j149IOjVsi2e8L0j6j5KG4uNTRX9nzSX9wMyeM7PV8Tl+pqBe+DsTDv9uA2D8FQ5jsOAYg4VXqDFYYz0vBozm7m5mnnc7ZhszO1HStyV91N270kE9fV5f7j4oaYWZvUXSP0m6JOcmzVpmdpOkfe7+nJm1592eEnmXu3ea2emSfmhm/5J+kZ8pwMzDv9tsMP4KizFYOIzBclOoMVhZKoebQwUtAAACGElEQVQ6JS1NHS+JzyEbryclbvF2X3ye70MdmFmTooHJ/3D378Sn6fOMufsbkp6U9K8VlXEm4Xq6T3/f3/HrJ0k6GLipM9k7Jd1sZq8omnpynaQviv7OlLt3xtt9igbfV4qfKagf/s6Ew7/bDDH+yg9jsCAYg+WgaGOwsoRDGyW1xautN0u6TdL6nNs0m62XdEe8f4ek/5M6/x/ildavlnQkVTKHGsRzeb8maau7P5x6iT7PgJktjH9bJTNrkXS9onUGnpT0gfhto/s7+T58QNKP3Z3fItbI3T/p7kvc/VxFP6d/7O7/XvR3Zsys1czmJ/uS/q2k34ifKagfxmDh8O82I4y/wmMMFhZjsPCKOAazsnwPzezfKZpH2SDpUXd/KOcmzQpm9j8ltUs6TdLrkj4t6buSviXpbEmvSvqgux+K/2P9kqKna/RI+rC7P5tHu2cqM3uXpJ9JelHD84H/k6J57/R5nZnZ2xQtBNegKEz/lrs/aGbnK/qtyimSnpf0J+5+3MxOkPRNRWsRHJJ0m7vvyKf1M1tc0vwJd7+J/s5O3Lf/FB82Snrc3R8ys1PFzxTUCWOw+mP8FRbjr/AYg+WHMVgYRRyDlSYcAgAAAAAAwFhlmVYGAAAAAACACgiHAAAAAAAASoxwCAAAAAAAoMQIhwAAAAAAAEqMcAgAAAAAAKDECIcAAAAAAABKjHAIAAAAAACgxP4/NiyrkuF4bN4AAAAASUVORK5CYII=\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "from sklearn.feature_selection import SelectPercentile\n",
        "\n",
        "selecteur = SelectPercentile(score_func=f_classif, percentile=50)\n",
        "X_train_new = selecteur.fit_transform(X_train, Y_train)\n",
        "X_test_new = selecteur.transform(X_test)\n",
        "\n",
        "reglog_anova = LogisticRegression(penalty='none', solver=\"saga\", max_iter=1, warm_start=True, random_state=0)\n",
        "reglog_anova.coef_      = np.array([[1.] * X_train_new.shape[1]])\n",
        "reglog_anova.intercept_ = INTERCEPT_\n",
        "\n",
        "coefs_anova   = []\n",
        "coefs0_anova  = []\n",
        "f1train_anova = []\n",
        "f1test_anova  = []\n",
        "\n",
        "for it in range(MAX_IT):\n",
        "    reglog_anova.fit(X_train_new, Y_train)\n",
        "    coefs_anova.append(reglog_anova.coef_.flatten().tolist())\n",
        "    coefs0_anova.append(reglog_anova.intercept_[0])\n",
        "    f1train_anova.append(f1_score(Y_train, reglog_anova.predict(X_train_new)))\n",
        "    f1test_anova.append(f1_score(Y_test, reglog_anova.predict(X_test_new)))\n",
        "\n",
        "coefs_anova = np.array(coefs_anova)\n",
        "\n",
        "fig, (ax1, ax2) = plt.subplots(1,2)\n",
        "fig.suptitle(\"F1 score par nombre des iterations\")\n",
        "fig.set_figwidth(20)\n",
        "fig.set_figheight(10)\n",
        "\n",
        "ax1.plot(range(MAX_IT), f1train, color=\"blue\", label=\"Reg\")\n",
        "ax1.plot(range(MAX_IT), f1train_anova, color=\"red\", label=\"Reg-ANOVA\")\n",
        "ax1.legend()\n",
        "ax1.set_title(\"Entrainement\")\n",
        "ax1.grid()\n",
        "\n",
        "ax2.plot(range(MAX_IT), f1test, color=\"blue\", label=\"Reg\")\n",
        "ax2.plot(range(MAX_IT), f1test_anova, color=\"red\", label=\"Reg-ANOVA\")\n",
        "ax2.legend()\n",
        "ax2.set_title(\"Test\")\n",
        "ax2.grid()\n",
        "\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        },
        "id": "yckUJDgxUUAE",
        "outputId": "2802821c-6549-47db-a1c1-c7c06a80c051"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "labels_new = selecteur.transform(np.array(diabetes.columns[:-1]).reshape(1, len(diabetes.columns)-1)).flatten()\n",
        "\n",
        "for i in range(len(labels_new)): \n",
        "    plt.plot(range(MAX_IT), coefs_anova[:, i], label=labels_new[i])\n",
        "plt.plot(range(MAX_IT), coefs0_anova, label=\"Theta0\")\n",
        "plt.legend()\n",
        "plt.title(\"Regression+ANOVA\")\n",
        "plt.grid()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "x5ne1I2uUUAE"
      },
      "source": [
        "**TODO: Analyser les résultats** \n",
        "\n",
        "- Pourquoi la performance de la régression sans filtrage est meilleure dans l'entrainement et pas dans le test; et l'inverse pour la régression avec filtrage ? (P.S. nous ne parlons pas ici du sur-apprentissage; mais des attributs : nombre et qualité)\n",
        "- Est-ce que le filtrage PEUT prévenir le sur-apprentissage ? Si oui, comment ? Si non, pourquoi ?\n",
        "\n",
        "**Réponse**\n",
        "\n",
        "- Le modèle de régression sans filtrage a des performances élevées sur les données d'apprentissage, cela peut être dû au fait que le modèle est capable d'apprendre et de capturer efficacement les relations entre les données. Cependant, ce modèle peut aussi entrainé sur des variables considéré comme **bruit** donc il peut qu'il ne généralise pas bien aux nouvelles données et peut avoir de mauvaises performances sur les données de test. D'autre part, si un modèle avec filtrage a des performances inférieures sur les données d'apprentissage, cela peut être dû au fait que le modèle est incapable d'apprendre toutes les relations entre les données en raison du nombre limité des variables. Cependant, si les variables sélectionnées sont de **haute qualité** et sont capables de capturer les pattrens et les relations importants dans les données, alors le modèle peut avoir de meilleures performances sur les données de test car il est capable de bien généraliser à de nouvelles données. \n",
        "\n",
        "- Oui le filtrage peut prévenir le sur-apprentissage. Car en sélectionnant un sous-ensemble de caractéristiques, le modèle devient moins complexe et est moins capable d'apprendre le bruit et les variations aléatoires des données. En conséquence, le modèle a plus de chances de bien se généraliser et d'avoir de bonnes performances sur les données de test."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GtpH5qYTUUAE"
      },
      "source": [
        "#### II.2.3. Comparaison entre les différentes approches de sélections d'attributs\n",
        "\n",
        "Ici, nous voulons comparer des modèles appartenant à des différentes approches de sélection d'attributs : \n",
        "- **ANOVA** : c'est un modèle qui utilise la sélection d'attributs par filtrage. Ici, nous prenons les attributs qui participent à 50% de la variance (4 attributs dans notre cas)\n",
        "- **L1** : c'est un modèle qui utilise la sélection d'attributs intégrée (embdded)\n",
        "- **FFS (Forward Features Selection)** : c'est un modèle qui utilise la sélection d'attributs enveloppante (wrapper). Il commence par tester un seul attribut et ajouter d'autres en validant la solution à chaque fois. Ici, nous avons choisi 4 attributs. \n",
        "- **BFE (Backward Features Elimination)** : c'est un modèle qui utilise la sélection d'attributs enveloppante (wrapper). Il commence par tester tous les attributs et éliminer d'autres en validant la solution à chaque fois. Ici, nous avons choisi 4 attributs. \n",
        "\n",
        "Nous voulons tester les modèles par rapport à celui sans sélection d'attributs en se basant sur les critères suivants :\n",
        "- **F1 entrainement** : ici on veut répondre à la question \"combien un modèle peut-il converger vers sa solution ?\"\n",
        "- **F1 test** : ici on veut répondre à la question \"combien un modèle peut-il généraliser ?\"\n",
        "- **Nombre des itérations** : ici on veut répondre à la question \"combien faut-il d'itérations pour qu'un modèle converge ?\"\n",
        "- **Temps pré-entrainement** : ici on veut répondre à la question \"combien de temps faut-il pour préparer les attributs avant l'entrainement ?\"\n",
        "- **Temps entrainement** : ici on veut répondre à la question \"combien de temps faut-il pour l'entrainement ?\"\n",
        "\n",
        "*P.S. L'exécution prend un peu de temps*\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "7nVTQiUWUUAE",
        "outputId": "062bcdf1-9223-4d2a-a123-c690adf905f2"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                 Criteres       Reg    Reg+L1 ANOVA+Reg        FFS        BFE\n",
              "0         F1 Entrainement  0.465574  0.465574  0.379928   0.567568   0.585526\n",
              "1                 F1 Test  0.380165  0.380165   0.37931   0.564516   0.539683\n",
              "2   Nombre des itérations    [1407]    [1404]     [148]     [5000]     [5000]\n",
              "3  temps pré-entrainement         0         0  0.001377  17.371314  27.896622\n",
              "4      temps entrainement  0.146683  0.146661  0.010435   0.343466    0.33066"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-2488b39a-0713-46bd-8c02-535e260b1e6f\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Criteres</th>\n",
              "      <th>Reg</th>\n",
              "      <th>Reg+L1</th>\n",
              "      <th>ANOVA+Reg</th>\n",
              "      <th>FFS</th>\n",
              "      <th>BFE</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>F1 Entrainement</td>\n",
              "      <td>0.465574</td>\n",
              "      <td>0.465574</td>\n",
              "      <td>0.379928</td>\n",
              "      <td>0.567568</td>\n",
              "      <td>0.585526</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>F1 Test</td>\n",
              "      <td>0.380165</td>\n",
              "      <td>0.380165</td>\n",
              "      <td>0.37931</td>\n",
              "      <td>0.564516</td>\n",
              "      <td>0.539683</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Nombre des itérations</td>\n",
              "      <td>[1407]</td>\n",
              "      <td>[1404]</td>\n",
              "      <td>[148]</td>\n",
              "      <td>[5000]</td>\n",
              "      <td>[5000]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>temps pré-entrainement</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0.001377</td>\n",
              "      <td>17.371314</td>\n",
              "      <td>27.896622</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>temps entrainement</td>\n",
              "      <td>0.146683</td>\n",
              "      <td>0.146661</td>\n",
              "      <td>0.010435</td>\n",
              "      <td>0.343466</td>\n",
              "      <td>0.33066</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-2488b39a-0713-46bd-8c02-535e260b1e6f')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-2488b39a-0713-46bd-8c02-535e260b1e6f button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-2488b39a-0713-46bd-8c02-535e260b1e6f');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ],
      "source": [
        "from sklearn.feature_selection import SequentialFeatureSelector\n",
        "import timeit\n",
        "\n",
        "CMAX_ITER     = 5000\n",
        "creglog       = LogisticRegression(penalty='none', solver=\"saga\", max_iter=CMAX_ITER)\n",
        "creglog_l1    = LogisticRegression(penalty='l1', solver=\"saga\", max_iter=CMAX_ITER)\n",
        "creglog_anova = LogisticRegression(penalty='none', solver=\"saga\", max_iter=CMAX_ITER)\n",
        "\n",
        "temps   = {}\n",
        "temps_p = {}\n",
        "\n",
        "temps_debut = timeit.default_timer()\n",
        "creglog.fit(X_train, Y_train)\n",
        "temps[\"reg\"] = timeit.default_timer() - temps_debut\n",
        "\n",
        "\n",
        "temps_debut = timeit.default_timer()\n",
        "creglog_l1.fit(X_train, Y_train)\n",
        "temps[\"l1\"] = timeit.default_timer() - temps_debut\n",
        "\n",
        "temps_debut = timeit.default_timer()\n",
        "selecteur_anova = SelectPercentile(score_func=f_classif, percentile=50)\n",
        "X_train_anova= selecteur_anova.fit_transform(X_train, Y_train)\n",
        "temps_p[\"anova\"] = timeit.default_timer() - temps_debut\n",
        "X_test_anova = selecteur_anova.transform(X_test)\n",
        "temps_debut = timeit.default_timer()\n",
        "creglog_anova.fit(X_train_anova, Y_train)\n",
        "temps[\"anova\"] = timeit.default_timer() - temps_debut\n",
        "\n",
        "creglog_ffs = LogisticRegression(penalty='none', solver=\"saga\", max_iter=CMAX_ITER)\n",
        "temps_debut = timeit.default_timer()\n",
        "selecteur_ffs = SequentialFeatureSelector(creglog_ffs, n_features_to_select=4, direction=\"forward\")\n",
        "X_train_ffs = selecteur_ffs.fit_transform(X_train, Y_train)\n",
        "temps_p[\"ffs\"] = timeit.default_timer() - temps_debut\n",
        "X_test_ffs = selecteur_ffs.transform(X_test)\n",
        "temps_debut = timeit.default_timer()\n",
        "creglog_ffs.fit(X_train_ffs, Y_train)\n",
        "temps[\"ffs\"] = timeit.default_timer() - temps_debut\n",
        "\n",
        "creglog_bfe = LogisticRegression(penalty='none', solver=\"saga\", max_iter=CMAX_ITER)\n",
        "temps_debut = timeit.default_timer()\n",
        "selecteur_bfe = SequentialFeatureSelector(creglog_bfe, n_features_to_select=4, direction=\"backward\")\n",
        "X_train_bfe = selecteur_bfe.fit_transform(X_train, Y_train)\n",
        "temps_p[\"bfe\"] = timeit.default_timer() - temps_debut\n",
        "X_test_bfe = selecteur_bfe.transform(X_test)\n",
        "temps_debut = timeit.default_timer()\n",
        "creglog_bfe.fit(X_train_bfe, Y_train)\n",
        "temps[\"bfe\"] = timeit.default_timer() - temps_debut\n",
        "\n",
        "\n",
        "pd.DataFrame({\n",
        "    \"Criteres\" : [\"F1 Entrainement\", \"F1 Test\", \"Nombre des itérations\", \"temps pré-entrainement\", \"temps entrainement\"],\n",
        "    \"Reg\" : [f1_score(Y_train, creglog.predict(X_train)), \n",
        "             f1_score(Y_test, creglog.predict(X_test)),\n",
        "             creglog.n_iter_, 0, temps[\"reg\"]],\n",
        "    \"Reg+L1\" : [f1_score(Y_train, creglog_l1.predict(X_train)), \n",
        "                f1_score(Y_test, creglog_l1.predict(X_test)), \n",
        "                creglog_l1.n_iter_, 0, temps[\"l1\"]],\n",
        "    \"ANOVA+Reg\" : [f1_score(Y_train, creglog_anova.predict(X_train_anova)), \n",
        "                f1_score(Y_test, creglog_anova.predict(X_test_anova)), \n",
        "                creglog_anova.n_iter_, temps_p[\"anova\"], temps[\"anova\"]],\n",
        "    \"FFS\" : [f1_score(Y_train, creglog_ffs.predict(X_train_ffs)), \n",
        "                f1_score(Y_test, creglog_ffs.predict(X_test_ffs)), \n",
        "                creglog_ffs.n_iter_, temps_p[\"ffs\"], temps[\"ffs\"]],\n",
        "    \"BFE\" : [f1_score(Y_train, creglog_bfe.predict(X_train_bfe)), \n",
        "                f1_score(Y_test, creglog_bfe.predict(X_test_bfe)), \n",
        "                creglog_bfe.n_iter_, temps_p[\"bfe\"], temps[\"bfe\"]],\n",
        "})"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "quQIlK52UUAF"
      },
      "source": [
        "**TODO: Analyser les résultats** \n",
        "\n",
        "- Pourquoi la sélection d'attributs envelopante prend plus de temps d'entrainement pourtant elle utilise moins d'attributs ? (en se basant sur les résultats)\n",
        "- Pourquoi la régression avec L1 prend plus de temps d'entrainement que la régression normale pourtant elle prend moins d'itérations ?\n",
        "- Pourquoi la sélection d'attributs enveloppante donne des performances mieux que celle par filtrage pourtant cette dernière sélectionne les meilleurs attributs en terme de corrélation ? (dans notre cas, c'est le même nombre d'attributs = 4)\n",
        "- Si vous avez un dataset avec plusieurs attributs et une quantité énorme d'échantillons, proposer une solution avec moins de temps de pré-entrainement+entrainement et plus de performance (par rapport la solution sans sélection d'attributs) en se basant sur les résultats de cette expérimentation.\n",
        "\n",
        "**Réponse**\n",
        "\n",
        "- la sélection d'attributs envelopante prend plus de temps d'entrainement car elle implique l'entrainement et l'évaluation de plusieurs modèles, chaqu'un sur un sous-ensemble de variables .\n",
        "- L1 prends plus de temps d'entrainement que la régression normal à cause des calcules supplémentaire dû a l'ajout du terme de régularisation (probleme d'optimisation plus complex)\n",
        "- La sélection d'attributs enveloppante donne des performances mieux que celle par filtrage car elle choisit les attributs qui offrent la meilleure performance. Ces attributs peuvent ne pas etre les meilleurs en terme de corrélation et la selection des attributs en terme de corrélation ne garantie pas la qualité des données obtenu.\n",
        "- Pour minimiser le temps de pré-entrainement+entrainement et avoir plus de performance, on propose d'effectuer un filtrage au debut pour minimiser le nombre des variables, puis on fait une sélection d'attributs enveloppante pour trouver le meilleur sous-ensemble des variables."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "FibS_xq6UUAF"
      },
      "outputs": [],
      "source": []
    }
  ],
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3 (ipykernel)",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.9.12"
    },
    "colab": {
      "provenance": []
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}