Recitation_01_.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# 10-701 Recitation 1\n",
    "\n",
    "\n",
    "## Goals:\n",
    "- HW1 linear algebra review\n",
    "- HW1 probability review\n",
    "- Lecture 1 and Lecture 2 review\n",
    "\n",
    "## Sources:\n",
    "- 10-701 [Lecture 1](http://www.cs.cmu.edu/~tom/10701-S20/Intro-DTreesAndOverfitting-1-13-2020.pdf) \n",
    "and [Lecture 2](http://www.cs.cmu.edu/~tom/10701-S20/Overfitting_ProbReview-1-15-2020.pdf)\n",
    "- [Andrew Moore tutorial](http://www.cs.cmu.edu/afs/cs/Web/People/awm/tutorials/infogain11.pdf)\n",
    "- 10-601 Intro to ML recitation 1 (Sida, Rawal, Rongye, George)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib\n",
    "font = {'family' : 'normal',\n",
    "        'weight' : 'normal',\n",
    "        'size'   : 16}\n",
    "\n",
    "matplotlib.rc('font', **font)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Linear Algebra review\n",
    "\n",
    "Consider the matrix X and the vectors y and z:\n",
    "\n",
    "$$ X = \\begin{bmatrix}\n",
    "1 & 5 \\\\\n",
    "2 & 3\n",
    "\\end{bmatrix}, ~ ~ ~ \n",
    "y = \\begin{bmatrix}\n",
    "1 \\\\\n",
    "2\n",
    "\\end{bmatrix},  ~ ~ ~\n",
    "z = \\begin{bmatrix}\n",
    "2 \\\\\n",
    "3\n",
    "\\end{bmatrix}.\n",
    "$$\n",
    "\n",
    "#### Question 1:\n",
    "- What is the transpose of matrix X?\n",
    "\n",
    "#### Question 2:\n",
    "- What is the inner product (dot product) of the vectors y and z?\n",
    "\n",
    "#### Question 3:\n",
    "- What is the product of Xy?\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "## Linear Algebra review\n",
    "\n",
    "Let\n",
    "$$ X = \\begin{bmatrix}\n",
    "1 & 3 & 2 \\\\\n",
    "1 & 4 & 10 \\\\\n",
    "2 & 3 & 1 \n",
    "\\end{bmatrix}, ~ ~ ~ \n",
    "Y = \\begin{bmatrix}\n",
    "2 & 4 & 5 \\\\\n",
    "1 & 2 & 3 \\\\\n",
    "5 & 6 & 1 \n",
    "\\end{bmatrix}.\n",
    "$$\n",
    "\n",
    "#### Question 4:\n",
    "- What is the top right value of the matrix XY\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "## Linear Algebra review\n",
    "\n",
    "#### Question 5:\n",
    "If X is a 2x3 matrix and Y is a 3x4 matrix, which is the following is true:\n",
    "* a - XY is a 3x3 matrix\n",
    "* b - YX does not exist\n",
    "* c - YX is a 2x4 matrix\n",
    "* d - XY is a 2x4 matrix\n",
    "\n",
    "#### Question 6:\n",
    "If $ A = \\begin{bmatrix}\n",
    "1 & 5 \\\\\n",
    "2 & 3\n",
    "\\end{bmatrix}$  then which of the following are true? \n",
    "* a - A is a 2x3 matrix\n",
    "* b - A is invertible\n",
    "* c - A has determinant as 0\n",
    "* d - A has trace 5\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Calculus\n",
    "\n",
    "#### Question 1:\n",
    "Evaluate the derivative of y with respect to x, where\n",
    "$$ y = x^4+x-3, ~ ~ ~ ~ \\text{for} ~ ~ x=3$$\n",
    "\n",
    "\n",
    "#### Question 2:\n",
    "What is the derivative of y with respect to x, where\n",
    "$$ y = sin(3x)$$\n",
    "\n",
    "\n",
    "#### Question 3:\n",
    "Evaluate the derivative of y with respect to x, where\n",
    "$$ y = 4x^3e^{5x}$$\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "## Calculus\n",
    "\n",
    "#### Question 4:\n",
    "What are the coordinates (x,f(x)) of the maximum point of the function:\n",
    "$$ f(x) = -x^2 +6x-3$$\n",
    "\n",
    "\n",
    "#### Question 5:\n",
    "Find the partial derivatives $\\frac{\\partial f}{\\partial x}$ and $\\frac{\\partial f}{\\partial y}$ of the function:\n",
    "\n",
    "$$ f(x,y) = x^2y^5+\\cos(x)\\sin(y)-4x-7$$\n",
    "\n",
    "\n",
    "#### Question 6:\n",
    "What is the integral of the function below with respect to x?\n",
    "\n",
    "$$ 6x^2-\\sin(x)+\\frac{1}{x}$$\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Probability and Statistics\n",
    "\n",
    "#### Expectation and variance:\n",
    "\n",
    "The variance of a zero-mean random variable X is $\\sigma^2$\n",
    "\n",
    "1- What is the variance of 3X?\n",
    "\n",
    "2- What is the variance of X+4?\n",
    "\n",
    "3- What is the mean of 3X?\n",
    "\n",
    "4- What is the mean of X+4?\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "# Probability and Statistics\n",
    "\n",
    "#### Coin Toss\n",
    "\n",
    "Consider five coin toss outcomes S = [1, 0, 0, 1, 0], where 1 denotes head and 0 denotes tail.\n",
    "\n",
    "1- What is the sample mean?\n",
    "\n",
    "2- What is the (uncorrected) sample variance?\n",
    "\n",
    "3- What is the probability of observing this outcome sequence? Assume the coin is fair, i.e. P(X=1) = P(X=0) = 0.5\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Entropy and Information Gain\n",
    "\n",
    "We saw in class that information gain can be used to determine which variable to split on while constructing a decision tree.\n",
    "\n",
    "### Entropy\n",
    "\n",
    "The entropy of a (discrete) random variable X taking values $V_1...V_m$ is defined as:\n",
    "\n",
    "$$H(X)=-\\sum_{i=1}^m P(X=V_i)\\log_2(P(X=V_i))$$"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "Entropy is a function of the probability distribution of X. If X is uniform, the entropy will be large (we have little information about the value of X):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "H(X) = 2.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "Values_X = np.array(['a','b','c','d'])\n",
    "P_X = np.array([1/4,1/4,1/4,1/4])\n",
    "\n",
    "plt.bar(np.arange(4), P_X, tick_label = Values_X)\n",
    "plt.ylabel('P(X)'); plt.xlabel('X'); plt.ylim([0,1])\n",
    "\n",
    "H_X = -np.sum(P_X*np.log2(P_X)) # python * is element wise, use np.dot for matrix multiplication\n",
    "\n",
    "print('H(X) = {}'.format(H_X))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "If X is not uniform, and instead has some very likely and very unlikely values, entropy will be low (we have more information about X):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "H(X) = 1.75\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "Values_X = np.array(['a','b','c','d'])\n",
    "P_X = np.array([1/8,1/2,1/8,1/4])\n",
    "\n",
    "plt.bar(np.arange(4), P_X, tick_label = Values_X)\n",
    "plt.ylabel('P(X)'); plt.xlabel('X'); plt.ylim([0,1])\n",
    "\n",
    "H_X = -np.sum(P_X*np.log2(P_X)) # python * is element wise, use np.dot for matrix multiplication\n",
    "\n",
    "print('H(X) = {}'.format(H_X))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "Think of encoding the values of X in the second case using the following code on a serial channel:\n",
    "\n",
    "| X    | a   | b   | c   | d   |\n",
    "|------|-----|-----|-----|-----|\n",
    "| P(X) | 1/8 | 1/2 | 1/8 | 1/4 |\n",
    "| code(X) | 00 | 01 | 10 | 11 |\n",
    "| N(X) (bits) | 2 | 2 | 2 | 2 |\n",
    "\n",
    "The expected value of the number of bits to use $N$ is:\n",
    "$$E[N] = \\sum_{i=1}^4 N_i\\times P(X=V_i) = 2\\times1/8 + 2\\times1/2 + 2\\times1/8 + 2\\times1/4 = 2$$\n",
    "\n",
    "Consider the code:\n",
    "\n",
    "| X    | a   | b   | c   | d   |\n",
    "|------|-----|-----|-----|-----|\n",
    "| P(X) | 1/8 | 1/2 | 1/8 | 1/4 |\n",
    "| code(X) | 110 | 0 | 111 | 10 |\n",
    "| N(X) (bits) | 3 | 1 | 3 | 2 |\n",
    "\n",
    "The expected value for the number of bits:\n",
    "\n",
    "$$E[N] = \\sum_{i=1}^4 N_i\\times P(X=V_i) = 3\\times1/8 + 1\\times1/2 + 3\\times1/8 + 2\\times1/4 = 1.75$$\n",
    "\n",
    "Entropy corresponds to the smallest number of bits on average per symbol needed to transmit symbols on a serial channel. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "### Entropy of coin flip example\n",
    "\n",
    "As an example, here is how the entropy of a coin flip varies as the probability of heads vary. When the probability is 0.5, we have the least information about the outcome (highest entropy)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'entropy')"
      ]
     },
     "execution_count": 130,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = np.arange(0.001,0.999,0.001)\n",
    "H_x = -(x*np.log2(x)+ (1-x)*np.log2(1-x))\n",
    "plt.plot(x,H_x)\n",
    "plt.xlabel('probability of heads'); plt.ylabel('entropy')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Conditional Entropy\n",
    "\n",
    "Conditional entropy of Y on X corresponds to how uncertain Y is given the value of X. Consider the example where X is the neighborhood in which a 10-701 student lives in, and Y corresponds to whether the student walks to school:\n",
    "\n",
    "| Neighborhood | Walks to school |\n",
    "|------|-----|\n",
    "| Squirrel-Hill | No |\n",
    "| Shadyside | Yes |\n",
    "| Oakland | No |\n",
    "| Squirrel-Hill | Yes |\n",
    "| Squirrel-Hill | Yes |\n",
    "| Oakland | No |\n",
    "| Shadyside | Yes |\n",
    "| Squirrel-Hill | No |\n",
    "\n",
    "The conditional entropy of Y given X is computed as:\n",
    "\n",
    "$$H(Y|X)=\\sum_{i=1}^m P(X=V_i)H(Y|X=V_i)$$\n",
    "\n",
    "It corresponds to how much information is given about Y once we know the value of X. In this case what is $H(Y|X)$?:\n",
    "\n",
    "| $V_i$ | $$H(Y|X=V_i)$$ |\n",
    "|------|-----|\n",
    "| Squirrel-Hill | |\n",
    "| Shadyside |  |\n",
    "| Oakland |  |\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "Conditional entropy of Y on X corresponds to how uncertain Y is given the value of X. Consider the example where X is the neighborhood in which a 10-701 student lives in, and Y corresponds to whether the student walks to school:\n",
    "\n",
    "| Neighborhood | Walks to school |\n",
    "|------|-----|\n",
    "| Squirrel-Hill | No |\n",
    "| Shadyside | Yes |\n",
    "| Oakland | No |\n",
    "| Squirrel-Hill | Yes |\n",
    "| Squirrel-Hill | Yes |\n",
    "| Oakland | No |\n",
    "| Shadyside | Yes |\n",
    "| Squirrel-Hill | No |\n",
    "\n",
    "The conditional entropy of Y given X is computed as:\n",
    "\n",
    "$$H(Y|X)=\\sum_{i=1}^m P(X=V_i)H(Y|X=V_i)$$\n",
    "\n",
    "It corresponds to how much information is given about Y once we know the value of X. In this case:\n",
    "\n",
    "| $V_i$ | $H(Y|X=V_i)$ |\n",
    "|------|-----|\n",
    "| Squirrel-Hill | $$0.5\\times\\log_2(0.5) + 0.5\\times\\log_2(0.5) = 1$$ |\n",
    "| Shadyside | $$ 0\\times\\log_2(0) + 1\\times\\log_2(1) = 0$$ |\n",
    "| Oakland | $$1\\times\\log_2(0) + 0\\times\\log_2(1) = 0$$ |\n",
    "\n",
    "$$H(Y|X)= 0.5\\times1 + 0.25\\times0 + 0.25\\times0 = 0.5$$\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "$$H(Y|X)=\\sum_{i=1}^m P(X=V_i)H(Y|X=V_i)$$\n",
    "\n",
    "| $V_i$ | $H(Y|X=V_i)$ |\n",
    "|------|-----|\n",
    "| Squirrel-Hill | $$0.5\\times\\log_2(0.5) + 0.5\\times\\log_2(0.5) = 1$$ |\n",
    "| Shadyside | $$ 0\\times\\log_2(0) + 1\\times\\log_2(1) = 0$$ |\n",
    "| Oakland | $$1\\times\\log_2(0) + 0\\times\\log_2(1) = 0$$ |\n",
    "\n",
    "$$H(Y|X)= 0.5\\times1 + 0.25\\times0 + 0.25\\times0 = 0.5$$\n",
    "\n",
    "#### Question 1\n",
    "\n",
    "- What happens to H(Y|X) if all the Squirrel-Hill student do not walk to school? What would that mean about how useful X is?\n",
    "\n",
    "#### Question 2\n",
    "\n",
    "- What happens to H(Y|X) if the following students are changed? What is it telling us about feature X?\n",
    "\n",
    "| Neighborhood | Walks to school |\n",
    "|------|-----|\n",
    "| Squirrel-Hill | No |\n",
    "| Shadyside | Yes |\n",
    "| Oakland | **Yes** |\n",
    "| Squirrel-Hill | Yes |\n",
    "| Squirrel-Hill | Yes |\n",
    "| Oakland | No |\n",
    "| Shadyside | **No** |\n",
    "| Squirrel-Hill | No |\n",
    "\n",
    "\n",
    "#### Question 3\n",
    "- What does conditional entropy represent? "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Information Gain\n",
    "\n",
    "How much would knowing about X help me? (how many bits would I save?)\n",
    "\n",
    "$$\\text{IG}(Y|X) = H(Y) - H(Y|X)$$\n",
    "\n",
    "In the previous example, 4 students walked to school and 4 did not:\n",
    "\n",
    "$$ H(Y) = -(0.5\\times\\log_2(5)+0.5\\times\\log_2(5)) = 1$$\n",
    "\n",
    "$$\\text{IG}(Y|X) = H(Y) - H(Y|X) = 1- 0.5 =0.5$$\n",
    "\n",
    "We use Information Gain to find features to construct a decision tree."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## Information Gain example\n",
    "\n",
    "| X1    | X2   | X3  | Y   |\n",
    "|------|-----|-----|-----|\n",
    "| 1 | 1 | 1 | + |\n",
    "| 1 | 1 | 0 | + |\n",
    "| 0 | 0 | 1 | - |\n",
    "| 1 | 0 | 0 | - |\n",
    "\n",
    "H(Y) = 1\n",
    "\n",
    "H(Y|X1) = ?\n",
    "\n",
    "H(Y|X2) = ?\n",
    "\n",
    "H(Y|X3) = ?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "## Information Gain example\n",
    "\n",
    "| X1    | X2   | X3  | Y   |\n",
    "|------|-----|-----|-----|\n",
    "| 1 | 1 | 1 | + |\n",
    "| 1 | 1 | 0 | + |\n",
    "| 0 | 0 | 1 | - |\n",
    "| 1 | 0 | 0 | - |\n",
    "\n",
    "H(Y) = 1\n",
    "\n",
    "- $H(Y|X1=1) = -1/3 \\times\\log_2(1/3) - 2/3 \\times\\log_2(2/3) =\n",
    "0.92$\n",
    "- $H(Y|X1=0) = -1\\times\\log_2(1) = 0$\n",
    "- $H(Y|X1) = 3/4 \\times H(Y|X1=1) +1/4 \\times H(Y|X1=0) \\sim 0.92$\n",
    "- $\\text{IG}(Y|X1) \\sim 0.31$"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "## Information Gain example\n",
    "\n",
    "| X1    | X2   | X3  | Y   |\n",
    "|------|-----|-----|-----|\n",
    "| 1 | 1 | 1 | + |\n",
    "| 1 | 1 | 0 | + |\n",
    "| 0 | 0 | 1 | - |\n",
    "| 1 | 0 | 0 | - |\n",
    "\n",
    "H(Y) = 1\n",
    "\n",
    "- $H(Y|X2=1) = -1\\times\\log_2(1) = 0$\n",
    "- $H(Y|X2=0) = -1\\times\\log_2(1) = 0$\n",
    "- $H(Y|X2) = 1/2 * H(Y|X2=1) +1/2* H(Y|X2=0)=0$\n",
    "- $\\text{IG}(Y|X2) =1$\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "## Information Gain example\n",
    "\n",
    "| X1    | X2   | X3  | Y   |\n",
    "|------|-----|-----|-----|\n",
    "| 1 | 1 | 1 | + |\n",
    "| 1 | 1 | 0 | + |\n",
    "| 0 | 0 | 1 | - |\n",
    "| 1 | 0 | 0 | - |\n",
    "\n",
    "H(Y) = 1\n",
    "\n",
    "- $H(Y|X3=1) = -1/2\\times\\log_2(1/2) -1/2\\times\\log_2(1/2) = 1$\n",
    "- $H(Y|X3=0) = -1/2\\times\\log_2(1/2) -1/2\\times\\log_2(1/2) = 1$\n",
    "- $H(Y|X3) = 1/2 * H(Y|X3=1) +1/2* H(Y|X3=0)=1$\n",
    "- $\\text{IG}(Y|X3) = 0$\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "## Information Gain example\n",
    "\n",
    "| X1    | X2   | X3  | Y   |\n",
    "|------|-----|-----|-----|\n",
    "| 1 | 1 | 1 | + |\n",
    "| 1 | 1 | 0 | + |\n",
    "| 0 | 0 | 1 | - |\n",
    "| 1 | 0 | 0 | - |\n",
    "\n",
    "\n",
    "$\\text{IG}(Y|X2) = 1$, X2 perfectly predicts Y (we didn't even need to compute this, after all).\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## In class example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9402859586706311\n"
     ]
    }
   ],
   "source": [
    "X_names = np.array(['Outlook','Temperature','Humidity','Wind'])\n",
    "\n",
    "Weather = np.array([['Sunny','Hot','High','Weak'],\n",
    "                    ['Sunny','Hot','High','Strong'],\n",
    "                    ['Overcast','Hot','High','Weak'],\n",
    "                    ['Rain','Mild','High','Weak'],\n",
    "                    ['Rain','Cool','Normal','Weak'],\n",
    "                    ['Rain','Cool','Normal','Strong'],\n",
    "                    ['Overcast','Cool','Normal','Strong'],\n",
    "                    ['Sunny','Mild','High','Weak'],\n",
    "                    ['Sunny','Cool','Normal','Weak'],\n",
    "                    ['Rain','Mild','Normal','Weak'],\n",
    "                    ['Sunny','Mild','Normal','Strong'],\n",
    "                    ['Overcast','Mild','High','Strong'],\n",
    "                    ['Overcast','Hot','Normal','Weak'],\n",
    "                    ['Rain','Mild','High','Strong']])\n",
    "\n",
    "PlayTennis = np.array([['No'],['No'],['Yes'],['Yes'],['Yes'],\n",
    "                       ['No'],['Yes'],['No'],['Yes'],['Yes'],\n",
    "                       ['Yes'],['Yes'],['Yes'],['No']])\n",
    "\n",
    "def entropy(vec):\n",
    "    # Computes entropy of a vector of discrete variables\n",
    "    values = np.unique(vec)\n",
    "    p_X = np.array([np.mean(vec==v) for v in values])\n",
    "    return -np.sum(p_X*np.log2(p_X))\n",
    "    \n",
    "H_Y = entropy(PlayTennis)\n",
    "print(H_Y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "Compute $\\text{IG}(Y|X)$ for every X:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x36 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x36 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAloAAABUCAYAAABazpmzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4wLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvqOYd8AAAD2VJREFUeJzt3Xu4XFV5x/HvT9NSkpIAisg9hgJRy00iYLkkiHhDgXKxhRYIFwO1RayKUASJxEIocpEgyk1AxGChQCqoQW4RaPOQpA33WgikUq5CQiBcA7z+sdbAnMnMnDnnzJ7b+X2eZ545Z6+197znPHvvefdaa6+tiMDMzMzMmu9d7Q7AzMzMrFc50TIzMzMriBMtMzMzs4I40TIzMzMriBMtMzMzs4I40TIzMzMriBOtDiPpNklDmnND0mRJIWlyk8IyszLNOE7NiiZpsaTFA6i/0n7t75Ohc6JVEElj8855fZ06pR34662MzWyoyvbvkHRdjTr75vKpLQ7PrOl8Tn9H2f/i0nbH0g1GtDsAW8lBwMh2B2E2AHtK2j4i5rY7EDPrY9cmbONaYC7wZBO2NSw50eowEfG7dsdgNgCPAmOB6cCktkZiZn1ExKImbGMZsKwJ4Qxb7jrsMLXGfkjaWNK/SVom6QVJN0raUtKluQl3bI3tfVLSf0h6SdJzki6T9J6i/w4bNh4ArgQmSvpMIytImihptqSlkl6RdJ+kb0j6o4p6k0pdj5J2lHRT3v+X1iifI2m5pCclTZf07lzvEEn35s96RNJhVWLaVNLpkhbmuF6VdL+kEyrjMhuIvI/eVqNspTFUZef0cZKOlfRw3ncXSvp0rjNG0g8lPZXLbpE0vpHt5+Vb5WNwed7fr5a0YY0Y+4zRyu+P5uKDy4YQRD4mv51/3q/G9o7P5QfU+Jf1HCdaXUDSBsCdwN7ArcC5wJvAb4AP1Fl1D+DnwBPAD4BFpK7JWUXGa8POicAK4FRJqldR0l8DtwAfA64GZgACTgOuqbH+DnmdFcD5QOWYsO2AG4Hf5/KlwLHAKXmszBnAfOBiYDRwkaRdKraxN3AI8L+53oWkY2wacFX9P9+sEGcBRwG/Bn4CbArMkvRR0vGwHTAT+BWwC3BD6eKiHklbALeTuhV/DvwQWBe4A1ijgbgWAt/LP98NfLvstZh0/LwFHFrls5WXLwWuaeCzeoK7Dou3aZ3BwFs1uI3pwNrAlyNiRmmhpG+Rdu5aPg9Miog7c/13AzcBkzymxpolIhZJuhD4ErA/8NNq9SSNISVCrwLbRcSDefk3gV8CnyNdCFxWseongIMj4sc1Qvg0sHtE/CJv71vAw6QvqSXA1hHxf7nsElLS9TXSRUvJj4EzI+L1sngFXAAcLmnHiLijgX+H9b5mnNMb+hxgy4h4DkDSr0gXJzeSkqu/jYg3c9kM4B+Av8x16jkX+FNgn4h4O9mRdBnp+KsrIhZKOhs4GlgYEVMr60iaDXxK0voR8f9lRROBjYEZEfFqf5/VK5xoFW8T4KTBrixpFWAf4HFSq1S575J29jVrrP7TUpIFEBFv5oNpEvBR0gBHs2aYBhwMTJN0VUSsqFJnL1KL0jmlJAsgIlZIOg6Yl7dRmWgtqJNkAdxSSrLy9l6SdANwGHB+KcnKZQskLQI2L99ARDxRudGICEk/AA4nJXtOtAyGeE4fgFNKSVZ2LfA6sDrwjVKSlf2MlGhtTp1ES9JGwE6kY6qyRelE4G+AflvFGnAB8BlgMvCdsuWlbvuLm/AZXcNdh8W7ISJU7UXqqujPeGAVYH5EvFFeEBEvk5pxa1lQZVnp6mL1RoI3a0REPAWcDYwDptSotmV+n1Nl/fnA8rI65eb38/F3V1n2ZJ2yp0hdJW+T9C5JX5R0p6TnJb2lNFaydAyt008MNnwM9ZzeqD77bkS8BTwDLI2IxyrqPpXf16W+LfL7ShcN+UasZt2MdT1pyMohpeEAuUV7H9J3WbXjsmc50ep8q+X339cof6bOui9UWVZK1ppx1WJW7nTgOeBESaOqlI/O70/XWP/psjqVy+uptp+/WafsDVZuzZ9BugpfjzR2ZDqpW740FmWVfmIwa7Za+3W983p/N26Mye+1vk/6O9YakhsFLiFdeJXGQx4ArMowa80CJ1rd4MX8vlaN8ve1KhCzevJt4KeSxhP+Y5UqpS+ItWtsYm2qf4kUOgO7pLWBvyO1IHwwIg6NiOPz2JOZRX62DQtB7WE6Y2osL0ppmoZa3ye1js3BuIi+g+IPA16mxhjOXuZEq/P9D/AaMEFSn4NV0qo0d/Cl2VB9H3gMOAaonEak1M29c+VKkj5CGqBbryu8KB8g3fl4U0S8UlG2Qxvisd6ylNRS2ofSlDytHsJxT37fsbIgT+9QdYqHKkotxjV7RiJiMemOyX0kTQS2Aa6KiGoXUz3NiVaHi4jXSIMg1wOOqCj+KrUHwpu1XL6TaCqpC/CYiuJZpBarwyVtWlqYLyCm51/rDXovSmlcysfKp5fIMf5TG+Kx3jIPGJuTDQAk/TFwZqsDyTeG3A5sI2nviuJpND6kZCmppW6DfupdAPwJ77RiDbtuQ/Bdh93iONKcJzMkfQJ4kNSStQPpoNmJ1ERr1gkuA74OfLB8YUQsk3QkaU6geZJ+Rjph7w58mDSAtuWJVkQ8Iela0q3x8yTdShpUvAcwmzSA12ywzgQ+CfxC0kxS99luwPO057E2R5EGw/+rpKtIc19NBNYntXhtUXvVJCKWS5oH7CzpcuAh0nfQ5eV3+ZLm6XqadDz9NiJub+Yf0i3cotUF8o77F6QWgV1Jt/GOIHXBlPrcX6y+tllr5dvOj69RNpO0D88FvkCankSki4m9I6LQ8Vh1HEwa+L4W6YtoK+CbrNwqZzYgEXEjaV9fBBwI7EfqUtuNNF1Dq+O5m/TdcQvpYuJI0h2CO5IufBp1IO/MfzeV1CLWZwLtPM1LqTXrR0OJu5upv/OapEn0ndivZFlEeIqANsoTkC4CRkaEB8WbmVlHkXQzKbFbPyKacldjtxlI1+GXSX3NJW/UqmjNJ2mdiKhsZj4G2Ih0d4eZmVnHkLQV8HHSIPhhmWTBwBKtB/3IlrZ6UNJdpPFZANsC25P6v6e2KygzM7Ny+YHRm5G65N8E/rm9EbWXB8N3j/NJg4a3A0aSZgK+GDg5Ih5vZ2BmZmZlppBu0noEOHC4zQRfaSBjtJ4B3ku6U2I2cFyest/MzMzMqmgk0dqa9KDJOaQ5cLYm3VG0Atg6Iuo9AsbMzMxs2Oo30aq6UprF+S5gekScUKV8CvnBsqNGjdpm/PjxQ43TrCkWLFjwbETUevxEYQZyTNz7+LKaZf3ZfL1WP9HDul27jgmz4WJQiRaApAeAxyLiU/XqTZgwIebPnz+ozzBrNkkLImJCO2Po75gYe9wNg9724um7D3pdG5464Zgw62VDnbC0XZMLmpmZmXW8QSVakiaQbt28q7nhmJmZmfWOfqd3kHQF8CjwX6Q7DrcmPWj1ceCcQqMzMzMz62KNzKN1H7A/6flfpfmbrgFOiohnhxqAx6OYmZlZr+o30YqIU4FTWxCLmZmZWU8Z6mB4MzMzM6vBiZaZmZlZQZxomZmZmRXED5W2juSbJMzMrBe4RcvMzMysIE60zMzMzAriRMvMzMysIE60zMzMzAriRMvMzMysIE60zMzMzAriRMvMzMysIE60zMzMzAriRMvMzMysIE60zMzMzArS0CN4JG0AnAXsBgi4CfhKRPyuwNjMzCzzY6nMulO/LVqSRgK3AOOBg4EDgU2AWyWNKjY8MzMzs+7VSIvWF4FxwGYR8TCApHuAh4AjgDOLC8/MzMysezUyRmsPYG4pyQKIiEeBO4E9iwrMzMzMrNs1kmh9GLivyvL7gQ81NxwzMzOz3tFIorUmsLTK8iXAGs0Nx8zMzKx3NHTX4UBJmgJMyb8ul/TbOtXfCzw7qM85bTBrWa/TaXX3qY1aGUuJjwlrp048JsyGC0VE/QrS08B1EXFExfLzgP0iYq0hBSDNj4gJQ9mGWblu36e6PX7rPN6nzNqnka7D+0njtCp9CHigueGYmZmZ9Y5GEq1/B7aXNK60QNJYYIdcZmZmZmZVNJJoXQgsBmZJ2lPSHsAs4DHg/CbEcEETtmFWrtv3qW6P3zqP9ymzNul3jBaApA3p+wiem0mP4FlcaHRmZmZmXayhRMvMzMzMBq6RrsNBkTRZUkh6XtIaFWUjctnUoj7fup+kqyUtkbR2lbJJkt6SdHQ7YmtEt8dv3cHnWrPOVliiVWYMcGwLPsd6z98DAZxbvlDSqqSxg/8JzGhDXI3q9vitu/hca9aBWpFo3QgcVe2q3qyeiHgaOBrYV9JeZUVTgfWBQyPirXbE1ohuj9+6js+1Zh2oFYnWd/L7CfUqSdpW0k2Slkt6SdLNkrZtQXzWwSLiJ8ANwPcljZH0EeBrwNSI6DO7uqS/knSXpJclLZV0paT1KupMlnR33seW5Z8PdfzWA3yuNetArUi0niR1nUyRVPVRD5K2AOaQnp04GTgIGA3MkbRlC2K0znYEMAo4G7gY+G/gu+UVJH0FmJnL9gG+BGwD3CppZK6zK3AJ8GtgD+ALwKXA6o7feoDPtWadKCIKeZEO4gD+jPRg6ueBH+WyEblsav796ly+etn6o0kPrr6mqBj96p4XcHjeZ14HNq8oWx14CTivYvmmwBvAkfn3E4AnHL9fvfTyudYvvzr71YoWLSJiCXAGcJCkzapU2Rm4PiKeL1vnBdLM8xNbEaN1toi4iHTFfl1E3FtRvBMwErgi32U1QtII4JH82jnXmwesI+lSSZ+VNNrxWy/xudas87Qk0crOIl01nVylbE3Sl1Clp0hN3GaQWoNer7L8ffn9DmBFxWsT4D0AETEb2B/YmPR0g+ckzZZU7VmeRej2+K07+Fxr1kFGtOqDImK5pFNJV1unVxQvAd5fZbX3A0uLjs263nP5/QDgoSrlL5R+iIgrgSslrQZ8HPgX0mD1sQXHWE+3x28dxOdas87SskQrOw/4Ku/cHVMyB/ispNUi4kWA/EXyeeC2lkZo3eg3wCvAuIiY2cgKeT+blbtXTpM0OnehtEO3x2+dx+dasw7R0kQrIl6TdDIrP+B0GvA54GZJp5EGbx5LGrdSrfnb7G0RsUTSccAZktYFZgMvAusBuwC/jIirJU0n32FF6j7ZkHR339x2JindHr91Hp9rzTpHK8dolVxCRfdIRNwDTCJ1kVwGXA4sByZGxN2tDtC6T0ScA+wL/DlwBak77STSF0lp8Plc0pin75GmSDiFNMnjXpXba7Vuj986ks+1Zh3AD5U2MzMzK0g7WrTMzMzMhgUnWmZmZmYFcaJlZmZmVhAnWmZmZmYFcaJlZmZmVhAnWmZmZmYFcaJlZmZmVhAnWmZmZmYFcaJlZmZmVpA/AGK4lRAiSQf+AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 720x36 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x36 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "IG = np.zeros(len(X_names))\n",
    "for i_var, X_var in enumerate(X_names):\n",
    "    values = np.unique(Weather[:,i_var])\n",
    "    n_values = len(values)\n",
    "    H_YX = np.zeros(n_values)\n",
    "    f, axs = plt.subplots(1, n_values+1, sharey=True, figsize = (10,0.5))\n",
    "    for i_ax,ax in enumerate(axs[:-1]):\n",
    "        value = values[i_ax]\n",
    "        # plot histogram\n",
    "        PT = PlayTennis[Weather[:,i_var]==value]\n",
    "        ax.hist(PT); ax.set_title(value); ax.set_ylim([0,5])\n",
    "        # compute H(Y|X=V_i)xP(X=V_i)\n",
    "        P_V_i = np.mean(Weather[:,i_var]==value)\n",
    "        H_YX[i_ax] = entropy(PT)*P_V_i        \n",
    "    axs[-1].set_title(X_var); axs[-1].axis('off')\n",
    "    IG[i_var] = H_Y - np.sum(H_YX)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "## Information Gain, first level:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.bar(np.arange(len(X_names)),IG)\n",
    "plt.xticks(ticks=np.arange(len(X_names)),labels=X_names,rotation=45);\n",
    "plt.ylabel('IG');"
   ]
  },
  {
   "attachments": {
    "tree_1.png": {
     "image/png": "iVBORw0KGgoAAAANSUhEUgAABVgAAAMWCAYAAAAEeO33AAAMKGlDQ1BJQ0MgUHJvZmlsZQAASImVVwdYU8kWnltSSWiBCEgJvQlSpEsNLYCAVMFGSAIJJcSEoGJHFxVcCyqiWNFVEUXXAojYsJdFsfeHIirKuqiLDZQ3SQBd/d573zvfN/f+98yZc/5z7sx8MwBoxHIlkhxUE4Bccb40LiyINS4llUVqBwggAyYAwJHLk0kCY2OjIAaD73/K+1vQGsp1B4Wvn/v/q2jxBTIeAEgsxOl8GS8X4oMA4O48iTQfAEIX1JtPzZdATIQsgY4UEoTYQoEzVdhTgdNVOEppkxDHhjgNADKNy5VmAqCu4MUq4GVCP+pLIHYS80ViiJsg9uMJuXyI+yAekZubB7GGDcQ26d/5yfyHz/Qhn1xu5hBW5aIUcrBIJsnhTv8/y/G/JTdHPhjDHDaaUBoep8hZUbfsvEgFpkF8XpweHQOxNsQ3RHylvQI/E8rDEwfsP/JkbFgzxX9GaXxucCTEhhCbiXOiowb0fhmiUA7EsPZogiifk6Aai/KleXED/tFpAllI/CDmSpWxFDYl8uzEwAGfG4UCzqDPxkJhQrKKJ3q1QJQUDbE6xA9k2fGRAzYvC4Xs6EEbqTxOwRn+cwxkSEPjVDaYRa5sMC/MWyjiRA/gqHxhQrhqLDaJx1Vy04M4SyAbFzXIky8IDlHlhRUJxIkD/LEySX5Q3ID9NklO7IA91iTICVPozSBukRXED47tzoeTTZUvDiT5sQkqbrhOFjciVsUBtwNRgA2CAQvIYUsHeSALiFq66rvgl6onFHCBFGQCAXAY0AyOSFb2iOEzHhSCPyESANnQuCBlrwAUQP2XIa3q6QAylL0FyhHZ4BnEuSAS5MBvuXKUeChaEngKNaKfovMg1xzYFH0/6VgagzpiCDGYGE4MJdriBrgf7oNHwWcAbC64J+41yOubPeEZoZXwhHCT0Ea4O1lUJP2BOQuMAW2QY+hAdunfZ4dbQa9ueBDuC/1D3zgTNwAO+CgYKRD3h7HdoPZ7rvKhjL/VcsAXxYmCUoZRAig2PzJQt1N3G/KiqNT3tVDxSh+qFnuo58c82N/Vjw/fkT9aYouwA9g57CR2AWvC6gELO441YJexowo8NDeeKufGYLQ4JZ9s6Ef0UzzuQExF1WRONU6dTn0DfSBfMC1fsVjYeZLpUlGmMJ8VCHdrAYsj5jmOYLk4OXsBoNj7VVvLO6ZyT0eYF7/p5m8EwPdgf3//kW+6yEYADpQCQL39TWc9Cy7nkwCcr+TJpQUqHa54EAAVaMCVog+M4d5lAzNyAe7ABwSAEBABYkACSAGTYJ2FcJ5KwVQwE8wDxaAULAerwTqwCWwFO8EesB/UgyZwEpwFl8BVcBPch3OlA7wC3eA96EUQhITQEQaij5gglog94oJ4In5ICBKFxCEpSBqSiYgROTITmY+UImXIOmQLUo38jhxGTiIXkFbkLvIY6UTeIp9RDKWhOqgRaoWORD3RQDQSTUAnopnoFLQQXYAuRSvQKnQ3WoeeRC+hN9E29BXagwFMDWNippgD5omxsRgsFcvApNhsrAQrx6qwWqwR/unrWBvWhX3CiTgDZ+EOcL6G44k4D5+Cz8aX4OvwnXgdfhq/jj/Gu/GvBDrBkGBP8CZwCOMImYSphGJCOWE74RDhDFw7HYT3RCKRSbQmesC1l0LMIs4gLiFuIO4lniC2EtuJPSQSSZ9kT/IlxZC4pHxSMWktaTfpOOkaqYP0kaxGNiG7kEPJqWQxuYhcTt5FPka+Rn5O7qVoUiwp3pQYCp8ynbKMso3SSLlC6aD0UrWo1lRfagI1izqPWkGtpZ6hPqC+U1NTM1PzUhurJlKbq1ahtk/tvNpjtU80bZodjU2bQJPTltJ20E7Q7tLe0el0K3oAPZWeT19Kr6afoj+if1RnqDuqc9T56nPUK9Xr1K+pv9agaFhqBGpM0ijUKNc4oHFFo0uTommlydbkas7WrNQ8rHlbs0eLoeWsFaOVq7VEa5fWBa0X2iRtK+0Qbb72Au2t2qe02xkYw5zBZvAY8xnbGGcYHTpEHWsdjk6WTqnOHp0WnW5dbd1Rukm603QrdY/qtjExphWTw8xhLmPuZ95ifh5mNCxwmGDY4mG1w64N+6A3XC9AT6BXordX76beZ32Wfoh+tv4K/Xr9hwa4gZ3BWIOpBhsNzhh0DdcZ7jOcN7xk+P7h9wxRQzvDOMMZhlsNLxv2GBkbhRlJjNYanTLqMmYaBxhnGa8yPmbcacIw8TMRmawyOW7ykqXLCmTlsCpYp1ndpoam4aZy0y2mLaa9ZtZmiWZFZnvNHppTzT3NM8xXmTebd1uYWIyxmGlRY3HPkmLpaSm0XGN5zvKDlbVVstVCq3qrF9Z61hzrQusa6wc2dBt/myk2VTY3bIm2nrbZthtsr9qhdm52QrtKuyv2qL27vch+g33rCMIIrxHiEVUjbjvQHAIdChxqHB47Mh2jHIsc6x1fj7QYmTpyxchzI786uTnlOG1zuu+s7RzhXOTc6PzWxc6F51LpcsOV7hrqOse1wfXNKPtRglEbR91xY7iNcVvo1uz2xd3DXepe697pYeGR5rHe47anjmes5xLP814EryCvOV5NXp+83b3zvfd7/+Xj4JPts8vnxWjr0YLR20a3+5r5cn23+Lb5sfzS/Db7tfmb+nP9q/yfBJgH8AO2BzwPtA3MCtwd+DrIKUgadCjoA9ubPYt9IhgLDgsuCW4J0Q5JDFkX8ijULDQztCa0O8wtbEbYiXBCeGT4ivDbHCMOj1PN6Y7wiJgVcTqSFhkfuS7ySZRdlDSqcQw6JmLMyjEPoi2jxdH1MSCGE7My5mGsdeyU2CNjiWNjx1aOfRbnHDcz7lw8I35y/K749wlBCcsS7ifaJMoTm5M0kiYkVSd9SA5OLktuGzdy3Kxxl1IMUkQpDamk1KTU7ak940PGrx7fMcFtQvGEWxOtJ06beGGSwaScSUcna0zmTj6QRkhLTtuV1seN4VZxe9I56evTu3ls3hreK34AfxW/U+ArKBM8z/DNKMt4kembuTKzU+gvLBd2idiidaI3WeFZm7I+ZMdk78juz0nO2ZtLzk3LPSzWFmeLT+cZ503La5XYS4olbVO8p6ye0i2NlG6XIbKJsoZ8HXjIviy3kf8if1zgV1BZ8HFq0tQD07Smiaddnm43ffH054Whhb/NwGfwZjTPNJ05b+bjWYGztsxGZqfPbp5jPmfBnI65YXN3zqPOy573R5FTUVnR3/OT5zcuMFowd0H7L2G/1BSrF0uLby/0WbhpEb5ItKhlsevitYu/lvBLLpY6lZaX9i3hLbn4q/OvFb/2L81Y2rLMfdnG5cTl4uW3Vviv2FmmVVZY1r5yzMq6VaxVJav+Xj159YXyUeWb1lDXyNe0VURVNKy1WLt8bd864bqblUGVe9cbrl+8/sMG/oZrGwM21m4y2lS66fNm0eY7W8K21FVZVZVvJW4t2PpsW9K2c795/la93WB76fYvO8Q72nbG7Txd7VFdvctw17IatEZe07l7wu6re4L3NNQ61G7Zy9xbug/sk+97+Xva77f2R+5vPuB5oPag5cH1hxiHSuqQuul13fXC+raGlIbWwxGHmxt9Gg8dcTyyo8m0qfKo7tFlx6jHFhzrP154vOeE5ETXycyT7c2Tm++fGnfqxumxp1vORJ45fzb07KlzgeeOn/c933TB+8Lhi54X6y+5X6q77Hb50B9ufxxqcW+pu+JxpeGq19XG1tGtx675Xzt5Pfj62RucG5duRt9svZV4687tCbfb7vDvvLibc/fNvYJ7vffnPiA8KHmo+bD8keGjqn/Z/mtvm3vb0cfBjy8/iX9yv53X/uqp7Glfx4Jn9Gflz02eV79wedHUGdp59eX4lx2vJK96u4r/1Ppz/Wub1wf/Cvjrcve47o430jf9b5e803+34+9Rfzf3xPY8ep/7vvdDyUf9jzs/eX469zn58/PeqX2kvoovtl8av0Z+fdCf298v4Uq5yqMABhuakQHA2x0A0FMAYFyF54fxqruZUhDVfVKJwH/CqvubUtwBqIUvxTGcfQKAfbBZBUDf8K04gicEANTVdagNiCzD1UXliwZvLISP/f3vjAAgwfPMF2l/f++G/v4v2yDZuwCcmKK6EypEcQfdHKBAN/X4c8EP8m+JMHJ8DI4PwAAAAAlwSFlzAAAWJQAAFiUBSVIk8AAAAgVpVFh0WE1MOmNvbS5hZG9iZS54bXAAAAAAADx4OnhtcG1ldGEgeG1sbnM6eD0iYWRvYmU6bnM6bWV0YS8iIHg6eG1wdGs9IlhNUCBDb3JlIDUuNC4wIj4KICAgPHJkZjpSREYgeG1sbnM6cmRmPSJodHRwOi8vd3d3LnczLm9yZy8xOTk5LzAyLzIyLXJkZi1zeW50YXgtbnMjIj4KICAgICAgPHJkZjpEZXNjcmlwdGlvbiByZGY6YWJvdXQ9IiIKICAgICAgICAgICAgeG1sbnM6ZXhpZj0iaHR0cDovL25zLmFkb2JlLmNvbS9leGlmLzEuMC8iCiAgICAgICAgICAgIHhtbG5zOnRpZmY9Imh0dHA6Ly9ucy5hZG9iZS5jb20vdGlmZi8xLjAvIj4KICAgICAgICAgPGV4aWY6UGl4ZWxZRGltZW5zaW9uPjc5MDwvZXhpZjpQaXhlbFlEaW1lbnNpb24+CiAgICAgICAgIDxleGlmOlBpeGVsWERpbWVuc2lvbj4xMzY4PC9leGlmOlBpeGVsWERpbWVuc2lvbj4KICAgICAgICAgPHRpZmY6T3JpZW50YXRpb24+MTwvdGlmZjpPcmllbnRhdGlvbj4KICAgICAgPC9yZGY6RGVzY3JpcHRpb24+CiAgIDwvcmRmOlJERj4KPC94OnhtcG1ldGE+ChwxtLEAAEAASURBVHgB7N1nrFVXev/xRe+9915MN82mgw2DOy7jZGbSRxlNNFImUjIvoigvRlGivI0ySiLF/8kknkwvLhiDMRiMDZhieje9mt6bwfD395ns6+vre+FyuO2c+13SmXs5Ze+1Pvswsn48+1l1bn82kkMBBRRQQAEFFFBAAQUUUEABBRRQQAEFFFDgngXq3vMn/IACCiiggAIKKKCAAgoooIACCiiggAIKKKBACBiw+kVQQAEFFFBAAQUUUEABBRRQQAEFFFBAAQVyFDBgzRHOjymggAIKKKCAAgoooIACCiiggAIKKKCAAgasfgcUUEABBRRQQAEFFFBAAQUUUEABBRRQQIEcBQxYc4TzYwoooIACCiiggAIKKKCAAgoooIACCiiggAGr3wEFFFBAAQUUUEABBRRQQAEFFFBAAQUUUCBHAQPWHOH8mAIKKKCAAgoooIACCiiggAIKKKCAAgooYMDqd0ABBRRQQAEFFFBAAQUUUEABBRRQQAEFFMhRwIA1Rzg/poACCiiggAIKKKCAAgoooIACCiiggAIKGLD6HVBAAQUUUEABBRRQQAEFFFBAAQUUUEABBXIUMGDNEc6PKaCAAgoooIACCiiggAIKKKCAAgoooIAC9SVQQAEFFFBAAQWqUuDWrVuJx9mzZ9OFCxfStWvXUpMmTVKrVq1S8+bNU6NGjYqmw2vHjx9PFy9ejPcVvXCHX+rWrRvHaNiwYfxs3Lhxatq0aTx4LZdx48aNmAPzvXr1aurWrVtq2bJlLocq92euX7+ePv7442pZe3aNzp8/H+e/fPlywuDmzZsp88U0u2b169eP58u9uFrwxlOnTiUefF8+/fTTcq24QYMGRd9dvrf8vcC5+N+Jch2o2JsuXboU15Br2aJFi/juFns551/5PnDMc+fOxc/evXundu3alXk8/g7z4DN8h1hbmzZtiv5e1qlTp8zP+oICCiiggAIKKFDTBQxYa/oVcn4KKKCAAgoUmAAhHcHpsmXL0gcffJD27NmTBg4cmCZNmpTGjh2bunTpUrTikydPppdffjmtWbMm7d+/v+j5O/1CcNO1a9d4EIT269cvDRo0KA0dOjTnEJBQaO3atWnFihVp9+7d6a/+6q/SQw89dKdp3PdrrP3HP/5xWr16dc5r79u3bxo8ePA9rz27RqtWrYrzb926NcJeAmaCP1zx5JqNGjUqgrL7CQHvG6sGHoDvym9+85v4fuNWntGxY8ew5fvbq1ev+O7izJ9zHXv37k1cx/fffz/+fv3lX/5lrof6wucIblnjkiVL0rvvvpv+4R/+IT311FNfeE/xP+zatSu+S++9916Eq/ydnzVrVnw/+U4ZsBbX8ncFFFBAAQUUyDcBA9Z8u2LOVwEFFFBAgTwXOHLkSISVCxcuTPv27YtAicCuc+fOEd4VXx5hKQEhg+D18OHDETZ+9NFHUVFJFen06dNTjx49ElWUBINUX/I7VYMcn/cS0BKyDh8+PA0YMCCqT6kWLG3weY5DZR7VszwIgQk6t2zZEs/zWmUPQifWfvv27Vg7boTMrOeTTz6JasQZM2aUe+0jRowo99oJkVeuXBnB3KZNm9KhQ4di3VeuXElUBh88eDCuxdGjR8Nm6tSpEQhWdlVvZZtX5PH5vj744IOJ0JTvLQ+uHd8nxsiRI9PEiRPju8qf+c5R2UnQSHX3iRMnwp8gkpCV69e+ffuo8ub9pQ2OQeVz9r3l54cffpjWrVuXtm/ffscK09KOV9ZzzI/AlL/D/EPJjh074vtR1vt5vnXr1vF3vXv37vGPFAsWLIi/q3ynxo0bd19Vunc6r68poIACCiiggAJVIWDAWhXKnkMBBRRQQAEFigQIPX/6059G6MMty3/yJ38SlZBUW5YcBEovvvhiVLxye/Hy5cvT/PnzI/DjFuW2bdumb33rW4mgkVupCWsIfwgACUMJl6iSJaziXN/4xjfSc889l/r37x8ha7169UqeMgJNQqpjx47FHDds2BA/169fn7hVvkOHDhHufumDFfwEa//qV78aa6dakApEQinCTgJW1v7nf/7n6ZFHHilz7VQu0mYgW/vzzz9/x7UTSlOtSyD30ksvJQJUwmRCP0JrjkOId/r06Tju5s2bIyxv1qxZzMGA9fMvAaHh6NGj4ztJGMn1+/nPf14UsE6ePDn94z/+Y7jxKb67hOiE23zXCLj5vmPOcb75zW9GtTDBO0Esj5KDa8N3hfNxDTkOD/6BgMH3934GYT+D7xR/LxYtWpR27tyZyvrHiuLnopKc7zT/kPLf//3fae7cubFm/kGD8NgK6OJa/q6AAgoooIAC+SZgwJpvV8z5KqCAAgookOcC9KXkdvszZ86kTp06ReBHld+dBgEOQRMtBPj8K6+8Ev0cCf4InLJwhp8EjzzHsQluuJWfsIpwct68eRG+fu1rX4uwimq6sgafnzZtWlQaUv1K9ej9BlRlnetOzxdfO8Ema+eW8/Ks/eGHH47buEuuncrK0tZOiP3aa6+ljRs3pp49e6aZM2cWVaYSrhHe0S6A8A5Tgl5aGXDbN+6EaI7PBQhB+S726dMnvr/cUk/gz+C68o8CBNcMfueaUOmJI6Hq9M+qs996662oPv2Xf/mXuAX/mWeeiffx96G0wfGoeOX68f39wQ9+UBSwlvb+e3mOAJdrThBMaEsofC+DNbI2/m4R5hMm8+C4DgUUUEABBRRQIJ8FDFjz+eo5dwUUUEABBfJQIKvUI7DjdnM2xmFzqzsNKk15EEDxyEIpQkaCquzP/ORBkEO1HH0saR9AZSXBLNV8ixcvjp6hvI9gl58lqwE5F0EXQRDnYKMijlkdo/jaWQvrZZRn7YRsBHEEs4Si77zzTqydY5RcO8EZ1apUJLJeKiwJtAmX8eN6EcDSYoHNiQh5Dxw4EJ+hUnjIkCHVwVOjz5ldI8JnHnynssF15fufDX7nwcZhtMzgWtMegL8vBOT8owT/gMD1nD17dvQx5RicIxt8j3kP14vPMwjkK2rwvaCqmSpnKk/vdfC9y/6+833iH1n4buZyrHs9t+9XQAEFFFBAAQUqU8CAtTJ1PbYCCiiggAIKVKsAARQhKb0uCXao5nvjjTfSr3/96wioqPAkOC0edGVVh8WDq2pdRI4nZ+2EqGxERdjM2rktm42XWHPJtVOdS8DK+glL//iP/ziMigfQBIRUBBPg8T6OR3/bbdu2Rdia41T9WCkCBKlUv/7e7/1eeNP2gpCVKmKqQLnVvmTon/1jQ2V9d/l+8I8UnIc2G/RIdiiggAIKKKCAAgqk9OXmTaoooIACCiiggAIFIkDQREBI9SChIbfGs3HUtWvX4tZkNq6igq744DOEh/ysrKCq+Pkq6/ds7VSbPvDAA3HLebZ2Nlti7VT1ZoNAlgB1+me3pU+YMCFCPaohsypJjsfvhH5UEXP7ORWyVCLSD9fbvDPJivnJd5Dgn0pUWl3Q05XqVm7Pp4UDt9aXrPzkGmXf3YqZxe+OwvWlryvVq1SFE9gTsnIuhwIKKKCAAgoooIABq98BBRRQQAEFFKgFAllP1uHDh0fvVYIrNotiE6jiIWMhUmRrHzZs2BfWTi/Q4mvHhCpfqn3Z4f5O4TJBH++hQpaQjUpKwlhHxQpwDQi96eFKxTGtA2jNwOZihOSE25U9aB1BS4C9e/dGhTOVqwSsDgUUUEABBRRQQIHPBfxn588t/E0BBRRQQAEFClyAiks2AGLXe/o/bt++PQKrAl92LI+1E7Ky9rNnz6YdO3Z8Ye1UJPIaj7uFpVQFE6ryGQJcKmSpbHRUjgAtGahi7dq1a1QL0/uWStaqCFiPHTsWFbNUy1INzfeD6+9QQAEFFFBAAQUU+FzAgPVzC39TQAEFFFBAgQIXIKBikyZ6R2YhI9V5tWF06dIl1k4wytoJl4uvndv/Cc54ZG0BynJhs6PDhw/HplcEblSz0hfUUTkC9GMlxKafMLfrE7DiX5kBKy0f6PdKwEr1KhvR8R0ifLc1QOVcZ4+qgAIKKKCAAvkr4D8/5++1c+YKKKCAAgoocI8CbAxEEEjVJSHj9evXv1DFeY+Hy6u3Z2vPwmVCUm43z2VQ/ctmR8ePH4/gbfTo0RHe5nIsP3N3AUJs/mGgQ4cO0euWDa/oh1qZAStBLv2JCVh50DqCKmiHAgoooIACCiigwJcFDFi/bOIzCiiggAIKKKCAAncQuHjxYrQYoAcrVcFsHkZ1o6NwBAjRFy5cGBWrM2fOjOrZwlmdK1FAAQUUUEABBSpWwBYBFevp0RRQQAEFFFCgDIFPP/00bim/fPlyVN61bds2Nkmyn2MZYDXwafpwcts4lY0ff/xx6t27dyJ8Y+MjNr5yFIbAyZMn0/79+6NKlpYAbA5XkdeXlgds2EU1NZtonTt3Ll27dq0w8FyFAgoooIACCtRKAQPWWnnZXbQCCiiggAJVL0CQsm/fvtich7MT2rAzOj1BHfkhQF9Obk8/dOhQIjCn9+pXvvKVRFjuKByBdevWpbVr16aBAwemwYMHx9/RivyHkD59+qRZs2ZFVSzV0Dt37kynTp0qHEBXooACCiiggAK1TsAWAbXukrtgBRRQQAEFql6AXp+EcvPmzUt79uxJEyZMSI899liaNm1aopqtqgabOlGdR39Jztu9e/fYFb2qzl+d58nWTtCdVRDS27O8g3CV6mMqG/F79NFH09ChQ1Pr1q3Lewjfl6MAvYK5foSRbEBGL1YeFT1oC8AGWjy4xlQmU2nKOSty0E/22WefTXXq1Ikq2VdffTUqWKmWpVK2YcOGFXk6j6WAAgoooIACClS6gAFrpRN7AgUUUEABBRQgmGPX8/feey82Vho1alQaN25cGjZsWGw4VVVC3NpO0EtgRdUl1XkVeetzVa0jl/MUX3ubNm3KvXY2UuJx5cqVaA1AQE0wS0DOLd6OyhfAntCTjdkaNGiQevXqFcFn3boVezMam5a9//77sakVleVUrfJ35ciRI19YJO9jLoSwDL4f/JnNsPgM3487VabTr5eN5vj/BILjpUuXxnr4/wM+Z8D6BW7/oIACCiiggAJ5IGDAmgcXySkqoIACCiiQ7wIEetxuTCi3bNmyNH/+/KiKYyf7yZMnV1kVJLe3b9u2LSoxqQAkYG3ZsmW+85Zr/gTLW7dujbW3b98+PfDAA+VaO31X6Y+5e/fu9NFHH4VZjx49DMHKpV4xb6JydePGjRF0Fg9YqQCtyEEF64cffhi37NMXdfHixaVeZwJfAlX68DJoF/HSSy9FUEr4+/jjj0cLgLLmtmvXrvj/gblz58Y/uPD/C7QLoVqWKlaHAgoooIACCiiQbwIGrPl2xZyvAgoooIACeShAtVqnTp3SxIkTo4L0N7/5TVq5cmVUj7IDfWXfZk4AxIOQcNOmTYnQsGfPnmns2LEF3z+0tLUTkLL2du3a3fHblLUFOHr0aGxuRRhNgNa5c+cvfY7WAzwIyAgBHfcvQGUo14/+pPREpXKUf6wYMmRI6tevX9xif/9n+fwIXN9BgwZFZTIBalmDOTRv3jwqVqlOZ1B5yt9jnr9bBSptJhYuXJg2b94cfw9nzJgRgT+fdSiggAIKKKCAAvkoYMCaj1fNOSuggAIKKJCHAtxOTmsAKvEYbGxDUEcVa2UPQiqCoB07dsT5CYCoqJ0yZUqV9oCt7HWWdnzWTsVhLmvn9nDCPUJpglV655bVj5PrSChHdawBa2lXIrfnuAZUivIPElyL7B8GqECu6ApWglv6oxLs8o8Qdxp8n/7iL/4iUfXKd+Jb3/pW+vrXv55oW3C3DbGopqYVwaVLlxL/wMI/vHTs2PFOp/M1BRRQQAEFFFCgRgsYsNboy+PkFFBAAQUUKBwBwiCCF8IYfue2c0LPuwU59yPAsamqJAxasGBBWrVqVVTXPfPMM2n69OkRrhZqGMja6ZG5ffv29NZbb6UPPvgg1v7000/fde1cG8Ivrg9VrPTGpNq1NCt8Cdn27duX9u7dG2FZZVQiEuByHgI8qmTLqpRkQzX6ejJXqip5X/FepdxuTxBMyE7oz+tlhcb38927389y7Vgz146WGvQ35R8ouH6EoMy9ogNWHMprwfmz9zIPrsmd+q4W9+A7w/eL71YWyBa/RsXf6+8KKKCAAgoooEA+CBiw5sNVco4KKKCAAgoUoADhSrZJzt2WR1hIJSahDCEbvzOyW6j5MwENQQ/HJSDMAlw2ZaL6j53K+TxVevR8HD16dLk22CKsY4Mojpudk4pQHncKuag85LP0siTUY35s/sPt1S1atCjXuTnf/aydisds7cwhW/uYMWNKPX8WSLOpEaE0ASshGpsS4VlysyPmxzppIXDixIlYJ5+509oJNHEr7+Aac27OsX79+vgYG5RloW923fkOEEjSBoJQmbCve/fu8T7Cv+w4BMFU5PJ6165d43V+zwI+jpFdN64xzzPn7LZ4jlXekVWCMn/8OV42sM4CxuzcvI8Hc+B7Q+UqASvtAagMpoKY7y7Xo7SwOzs2PzkO6+B6ZINgkwAa/7t9PvtMZfzkWjEXhwIKKKCAAgooUCgCBqyFciVdhwIKKKCAAgUsQBBLUETQtGbNmgjwWG4WmhHW0OeVQYhFJSUPNmWiapUd2AlXZ8+eHQ9uS75b/9E42Gf/Q59IQsosqCIUY7MsdkDv27dvmbdDE26+9957sVEQ82Z+w4cPj81/6H/ap0+f7BR3/Jmtnc2HVq9eHcEZHyjP2nn//s/6XRLusXbCuTutnbUR6r3zzjvphz/8YVGVIWFcFgKWnCxBIXMcP358euGFF6KvLmvnFvBFixbFNWPtBKKzZs2K3q+4lXdwfKzffffd9P/+3/+L69u7d+/0N3/zN3HO4td9z5496ZVXXkm//OUvI0Cd/lmV8ve+970IJwn1+E6wsdL//M//xOtsrMTrBLHZcQiRmfvbb78dQS1hJNfr0UcfTQTTbMRU3sHcCTo5L9eP7002MMtC+swW++y7SysNvjdcO8757LPPpoceeijmercep5yDfq2sg+Nlg8pdwmf82eTNoYACCiiggAIKKFAxAgasFePoURRQQAEFFFCgEgSo4lu3bl3cfs4t0ll1IqEVg2o8ArUtW7YUbaxDSEhwxYNQlHBw6NChESoRqHGbNdWIZYVUBF9UybLTOefetm1bHJ+AlkF1JgEkFZv0kGWDHkLLkoMqRKouOQ5hGfNgPTxPuHW3gJXqQ0I5Ki5ZO+Eh4RifZ7B2qnK3bt1atJaSa6clA2vnXFQ/3mntmBLK0UqBqkmCZdZKMHmnQThICEkfTfqCtmrVKlE1XHLthHvZ2u8lYOXcVCYTonNMrisVpcyXADMbvIe5cJ0IMpkH3lxPwmgGr/OdyF7nOnPc7HXew+v0CKWCd8OGDWHLcxyHILa8ASvXiutOaMr5+O4SdmeDY//bv/1bHD+71R6f7LvLOvieEFwTzE+aNCkqbu9U/ct3gu8l31vOzRw4bzb4B4eXX345Npzje0HgzcZWDgUUUEABBRRQQIH7EzBgvT8/P62AAgoooIAClShAAEbF4bJly6IaNTsV4SgPQq9f/OIX2dPxk9vw2QioR48esTETwR8B1eTJk4uCyC98oMQfOCZVjOxy/q//+q9F4RvhHLf2M6gM5MGg8rG0gJXXCAAJVgkEGQSE8+bNi2ArnrjD/7D2N954Iyo3Ccayka2dMPVuaydUzdZ+t1vbCfZYN/0+qfol9CtPT03Wx+3r7GrPxmEMAlbWzlyztRMSv/nmm+Vae7ZWfhKcEgISNrIREteHW/tLBo2EybQO4D2dOnWKOfGZLLzkdYL17HUqmHkf8y8+CFuZO9c1mzshN5Wk9D8t7yBA/dGPfhSbuREuZyP7DhGA8ig+WBffXR6E4tM/q8Dle4tteQaBMv8g8NJLL8U1zD6TnZPg+Mc//nE8PW3atDRu3LicA1b+PvD94NgYlnTMzu1PBRRQQAEFFFCgNggYsNaGq+waFVBAAQUUyFMBwrA/+qM/ilvbCbjKMwjSCH54ELbys3Xr1mXeyl/ymISChFsvvvhiVHwWr24s+V7+PHLkyNKejvDuqaeeitvKCRyplOSW79/+9rflmgvhIGvn1v6qWDthIgEpt94T7hLglmcQtBHe9v7stv1scN2efPLJ6HNbfO1UG3N97mVkASu3x//TP/1ThJ+Eesy1eGhMwEfAOmfOnKja5TUC1GxzLo5DwPr4449HNTOvEwxThcw1z0avXr2i1QEVozhQKUvAn4XO2fvu9pNgn16pVJVy7cszCI2z72z2/WWO5R2slTYGf/d3fxd9g+/0Oa5RedtklHYcwuC///u/DyO+A5zXoYACCiiggAIK1FaBe/sv3Nqq5LoVUEABBRRQoFoECP3YjKoqR1bpSBg3YsSInE9NCDh48OB4cOs3vTAJ23iOwPdug7WXVRl7t8/m8johI2He1KlTc/n4Fz5TfO0ElLQ3ICQu79qLH4xglGrSrLKz+GvFf88qbtnIi0dpgwBz0KBB8SjtdZ7juvOgZyxz5hZ7brunCjWraC3rs8WfJ6jlUZWDcJZHedsY3M/cqCp+5JFH7ucQflYBBRRQQAEFFCgYAQPWgrmULkQBBRRQQAEFaqoAPTzZpIkWAWw0RfuC2jKKr51q3Hxa++nTp2OzK9Ywc+bMaC9QW66b61RAAQUUUEABBRQov4ABa/mtfKcCCiiggAIKVIAA1Y30lGRDJTZRYhMgbuVmU6JCGmwOxa3h2SZZ9A6lepLKSG6vLuRR1trZWKmmr53WCFw3eqju3Lkzeosybzbwupfb9Qv5+ua6NlypBGYDN9oK0KaAFg387lBAAQUUUEABBfJZwP+ayeer59wVUEABBRTIQwH6Po4dOzZuYyaI27VrVzpw4EDs/l7evp/5sGw2SiJUZYOkrVu3Rr9L+rWOGjXqvnpf5tPa9+/fn7Zs2RLrpd0Ca6/pIWV23dhYjNYG3bt3j1YRbBZWntYO+XB9qmOOtIo4depU/IMDLRdot8A/OAwYMOCe+/JWx/w9pwIKKKCAAgoocCeBet//bNzpDb6mgAIKKKCAAgpUtAA9Lglc6HHJzvX0KM12nGdzn0IY9A6lnyu9KqnYpS8omx6xTl4r5JGtnark4mtnY6mavnbmx4ZZhKn9+/ePvrFUWhbfCKuQr11lrY1etmwUNm/evPhHBzYkY8OxyZMnR3/de938rLLm6XEVUEABBRRQQIFcBGwRkIuan1FAAQUUUECBnAWync45ABVs3DZM8EioVUi3CrMWHtwCTdUmAV1tCZGytdP6geudT2vP5s5GUVSzEgrznOP+BPjuE7gPHDgwNv9iMy4q2Qnga8vfi/sT9NMKKKCAAgooUJMF6tz+bNTkCTo3BRRQQAEFFCh8ASpY6cdK6ELQ6lBAgcISYKMwWoLwd7xevXqFtThXo4ACCiiggAK1XsAK1lr/FRBAAQUUUECB6hfIqlcNXqr/WjgDBSpDgCpm6jpqeouIyli7x1RAAQUUUECBwhewgrXwr7ErVEABBRRQQAEFFFBAAQUUUEABBRRQQIFKErChVCXBelgFFFBAAQUUUEABBRRQQAEFFFBAAQUUKHwBA9bCv8auUAEFFFBAAQUUUEABBRRQQAEFFFBAAQUqScCAtZJgPawCCiiggAIKKKCAAgoooIACCiiggAIKFL6AAWvhX2NXqIACCiiggAIKKKCAAgoooIACCiiggAKVJGDAWkmwHlYBBRRQQAEFFFBAAQUUUEABBRRQQAEFCl/AgLXwr7ErVEABBRRQQAEFFFBAAQUUUEABBRRQQIFKEjBgrSRYD6uAAgoooIACCiiggAIKKKCAAgoooIAChS9gwFr419gVKqCAAgoooIACCiiggAIKKKCAAgoooEAlCRiwVhKsh1VAAQUUUEABBRRQQAEFFFBAAQUUUECBwhcwYC38a+wKFVBAAQUUUEABBRRQQAEFFFBAAQUUUKCSBAxYKwnWwyqggAIKKKCAAgoooIACCiiggAIKKKBA4QsYsBb+NXaFCiiggAIKKKCAAgoooIACCiiggAIKKFBJAgaslQTrYRVQQAEFFFBAAQUUUEABBRRQQAEFFFCg8AUMWAv/GrtCBRRQQAEFFFBAAQUUUEABBRRQQAEFFKgkAQPWSoL1sAoooIACCiiggAIKKKCAAgoooIACCihQ+AIGrIV/jV2hAgoooIACCiiggAIKKKCAAgoooIACClSSgAFrJcF6WAUUUEABBRRQQAEFFFBAAQUUUEABBRQofAED1sK/xq5QAQUUUEABBRRQQAEFFFBAAQUUUEABBSpJwIC1kmA9rAIKKKCAAgoooIACCiiggAIKKKCAAgoUvoABa+FfY1eogAIKKKCAAgoooIACCiiggAIKKKCAApUkYMBaSbAeVgEFFFBAAQUUUEABBRRQQAEFFFBAAQUKX8CAtfCvsStUQAEFFFBAAQUUUEABBRRQQAEFFFBAgUoSMGCtJFgPq4ACCiiggAIKKKCAAgoooIACCiiggAKFL2DAWvjX2BUqoIACCiiggAIKKKCAAgoooIACCiigQCUJGLBWEqyHVUABBRRQQAEFFFBAAQUUUEABBRRQQIHCFzBgLfxr7AoVUEABBRRQQIGCF7h+/Xo6duxYOnfuXMGv1QUqoIACCiiggAIK1CwBA9aadT2cjQIKKKCAAgoooEAOApcuXUpbtmxJhw4dyuHTfkQBBRRQQAEFFFBAgdwF6uf+UT+pgAIKKKCAAgooUJUCO3bsSD/5yU/S+fPnv3TaTp06pcGDB6cZM2aktm3bFr1+/PjxtH379vTRRx+lgwcPJoLIOnXqpEaNGqWePXumgQMHplGjRqUWLVqkhg0bFn2upvxy69at9PHHH6dt27ald955J+Zf2tzq1q2bGjdunKZOnZqGDx9e2lsq5LmLFy/GfN577720YcOGMo/52GOPpSeeeKLM131BAQUUUEABBRRQoHAEDFgL51q6EgUUUEABBRQocAEC0h/96EfpyJEjqUGDBql79+4RjBIu9unTJ56bOHFiKHz66afpzJkzEUwuXrw4AlY+d/Xq1URoSchKKDts2LD4naC1S5cuiWPxWk0ZzHXv3r1p0aJF6aWXXoo1lTa3Dh06pGnTpqUxY8aU9nKFPXfjxo0IuAmtCVkZN2/eTLQoIAgmgGUQchuwBoX/o4ACCiiggAIKFLyAAWvBX2IXqIACCiiggAKFKNC+ffv03e9+N40fPz4qN5s0aRJha7t27WK5165dSwSry5YtSxs3bkzjxo1Lzz77bFStUsVKaDlv3rz09ttvp927d6cXX3wxPf3003EswtuaMghYqRRdvXp1hMNlzat58+YRFvfo0aOst1TI81T6Dho0KH3nO99J3/jGN+KYZ8+ejdYE//mf/5lWrFhRIefxIAoooIACCiiggAL5I2DAmj/XypkqoIACCiiggAJFAtzOT0sAKjb5vXjVKZWr+/fvj3B1586dqXPnzmnEiBHp4Ycfjt8JX7t27RqVsFRcEsBSvZq9j/C2JgyCYFocEAZTWfuHf/iHEQCXNjcqWCdPnpwqO2AlfOZB0JoNDGm38Morr2RP+VMBBRRQQAEFFFCgFgkYsNaii+1SFVBAAQUUUKBwBLKeo/RSLTm4VX3NmjVp5cqVEbzOmTMnKlhpI8DIql1nzpwZVaGbNm2K9zZr1ixCVm5vJ7AtHtqWPAd/5tZ4WhEw6tWrl+rXr9j/tCQo3rp1azp27FgEmN///vcjGI4T1qD/yfrX4udQQAEFFFBAAQUUqH0CFftfwbXPzxUroIACCiiggAI1ToCAldvq2QyLik4qXYtvfMWECWgfeOCBdODAgdjwip8Eso888kjq1q1batmy5V3XReUrG28xBgwYEO0K7vqhe3jD0aNH0/LlyyNU7d+/fwTD9/Bx36qAAgoooIACCiigQJUIGLBWCbMnUUABBRRQQAEFqk6Ais/169enc+fORbjaq1ev1KpVqy9MgICVTa64tZ3b6/fs2RO34tNagA2vqMq8WwXrli1boo8rB/7KV75SYQErfVc/+eSTxKZeH3zwQcyR+b///vuJHrNUitJztXXr1hEc322eX1i4f1BAAQUUUEABBRRQoIIFDFgrGNTDKaCAAgoooIAC1S1AwEoFKzve05+VULJp06alTougkkCVatcrV65E71YqR7N2AqV+6P+e3L59e3rzzTfjT/R0rahB6wE2jqL3KptbrVu3LqpXCVSZF/Pl8eCDD6aJEyfeNQiuqHl5HAUUUEABBRRQQAEFShMwYC1NxecUUEABBRRQQIE8FqAv6vXr12MFVKrSG5WfpQ2eJ4TlPXyOvqdUvpZnUGV6+fLleCu/V9TIAlZC4XHjxkXYSuB65MiRxIZSBMBUzxK80tZg/PjxaciQIVHRygZUDgUUUEABBRRQQAEFqlLAgLUqtT2XAgoooIACCihQBQLcMk9wyq32WdhKaHm3Tahu376dLl26FJWsVTDNMk/BPKi+pRfsk08+mQ4fPhy9Ynft2hXzo0L3o48+io282rRpk77xjW9EFeuwYcOiFQIbbjkUUEABBRRQQAEFFKgqAQPWqpL2PAoooIACCiigQBUJUMXZpEmTdO3atQhL6WXKcyU3uqqi6dzzaRo3bpz69u0bPWJpA0A1LpWrVNfSK3bz5s1pyZIlEbyePHkyvf766+nUqVPp29/+dho0aNCX+s3e8wT8gAIKKKCAAgoooIAC9yBgwHoPWL5VAQUUUEABBRTIBwGqOulVSrB64cKFtG3btggd8yVgpQKVTbZ4ZIMKXNoQELyyaVf79u3TqlWrokfroUOH0tq1a9OYMWOi12zJDb2yY/hTAQUUUEABBRRQQIHKEDBgrQxVj6mAAgoooIACClSjQOfOndOoUaNi4yr6qa5fvz6CSTaGytdB6EpV7oABA2ItU6dOjcrV/fv3p48//jgeS5cuTR07dky0CnAooIACCiiggAIKKFBVAgasVSXteRRQQAEFFFBAgSoSoMpz1qxZae/evdGrdNGiRRFOcvrhw4enZs2aRfuAnTt3puXLl8fmUWxWRRsBqkO7du0aM2UzKapfFyxYELfnl5z+6tWri57ilv1vfvObRX8u+QubUM2ePTv6quZSSUtf2WzQX5awlXU+/vjjafHixYmglY2vJkyYkL0tqndZwzvvvBMVrkUv5PALbQsee+yxRMuCLl263LWfbQ6n8CMKKKCAAgoooIACeSpgwJqnF85pK6CAAgoooIACZQkQAI4bNy59+OGHieCUTaGyMPT06dNR5UlIeeDAgQhXT5w4EYErt9b37t07QlACTapfCVh/+ctfJm7Dv9PYsWNH4lHWmDlzZlSW0r4gl4C1+HGZO4/u3bun6dOnp+3bt8e5WQ+9WLNBD1qqWwlYf/Ob32RP5/STdgVUBhNA89OhgAIKKKCAAgoooEAmYMCaSfhTAQUUUEABBRQoEIGmTZumHj16pK997WsRmHLr/O7du9N//dd/pV/84hfxHJWshLCEsfRpvXHjRmrZsmVUhXbr1i0vJAhqx44dG4HxrVu3IiRmHQ4FFFBAAQUUUEABBapSwIC1KrU9lwIKKKCAAgooUAUCWb9Seq5yKz2BKRWoWXVnu3btog0AoSQVnlR6durUKY0YMSJ16NAhNWrUKGbJ+wgwv/Od70Q/15JT59b8NWvWxNPcOk8LgLJG//79I7wtvnFVWe8t7/Pctp/Nl4rb+vXrJ9aeDYLmnj17pueeey56t2bP5/ITk0mTJhVV/+ZyDD+jgAIKKKCAAgooUJgCBqyFeV1dlQIKKKCAAgrUcgGCRgJSHmx4VXzcvn07ffLJJ+lnP/tZog/rzZs3I/x8+OGHU+vWrYveSujKY+LEiUXPFf/lr//6r4sCVj77z//8z8VfrvTfaRNAqMpP1ksFLoFyNpo3b5540KvVoYACCiiggAIKKKBAZQkYsFaWrMdVQAEFFFBAAQVqqAC30V+8eDFt3rw5bd26NQLKoUOHRj9TAtl8GZcuXUrHjx+PtRCkDhs2LPqy5sv8nacCCiiggAIKKKBAYQgYsBbGdXQVCiiggAIKKKDAXQVoCfDpp5/G5lYbNmyIcJU/U31KK4B+/foVtQe468FqwBvOnDmTNm7cGJtx0Y91/PjxsYYaMDWnoIACCiiggAIKKFCLBAxYa9HFdqkKKKCAAgooULsFCFivXr2aVq9enf793/89HTlyJPXp0yd9+9vfTvRQpWdpPg2qV5ctW5ZOnz6dOnfuHBW4gwcPzqclOFcFFFBAAQUUUECBAhAwYC2Ai+gSFFBAAQUUUECBuwlcvnw5HT58OC1dujS9++67sbkVmzZNnTo1Nrdq3759YqOoqhjnzp1L27ZtS5s2bUrbt29PDRs2jFv76RXLplX0h33nnXfShQsXoq8qlbUEwWxYxXsJifk84SphMe0NqF5lUy/6sDoUUEABBRRQQAEFFKhKAQPWqtT2XAoooIACCiigQBUJZBtZXbt2LV25ciUdO3YsQsklS5ZE5Wr37t3TrFmz0owZM1LXrl1TgwYNqmhmKdE7lXD11VdfTW+99VZq3LhxeuCBByL0HTduXMxl/vz5af/+/VFVO3r06AiBCVJbtWqV6CFLsLp79+7Y3GrChAlp9uzZqUuXLnGsKluIJ1JAAQUUUEABBRRQ4DMBA1a/BgoooIACCiigQAEK3Lx5M504cSLt2bOnqFqUlgBUiD7//PNR8dmrV69E5Wq9evWqVIC+r1ShUqnK4OfevXsjcO3WrVuaOHFiGjlyZLyHEHXhwoVRrUoLgx49ekQlK78TvP7pn/5p6t27d+rUqVNUt1bpQjyZAgoooIACCiiggAKfCRiw+jVQQAEFFFBAAQUKTIBb6+lLSqDKRlCNGjVK/fv3L7rNnlvpqRjldvv69XP/z8HJkydHNSl8/F7e0aJFi6hIrVu3bhoyZEgc4+jRo4mNt6i8paL20UcfTQTAhw4dStevX0/0j2W0a9cuwlQ2taJtAJ9nfVVZgVvedfo+BRRQQAEFFFBAgdohkPt/UdcOH1epgAIKKKCAAgrkncDZs2ejcvXkyZMRoo4YMSKqPqnyrMhBJSyPex2EowSoPKhmJRBetGhRtDFo3rx5bFj1+OOP3+thfb8CCiiggAIKKKCAAtUiYMBaLeyeVAEFFFBAAQUUqDwBbvunOpXKT6pEmzVrFo/KO2PuR6Y9wIoVK6JFwLRp06IqNfej+UkFFFBAAQUUUEABBapewIC16s09owIKKKCAAgoocN8C3DJ/+fLl2MCqSZMmqU6dOkXHrMmBKpNk7lSunjp1Kh08eDDCVfrAErD27NmzaB358gsh8fnz54t6yubLvJ2nAgoooIACCiigQMUI1K2Yw3gUBRRQQAEFFFBAgaoUuHHjRtxST0iZb4OAlU2uVq1alV5++eUIXB988ME0ZcqU2MQq39Zz8eLFCIoJvB0KKKCAAgoooIACtU/ACtbad81dsQIKKKCAAgoUgAB9S3/729/GxlCtWrVKHTt2TP369YvNplq3bl2jV0jbAloYsEkVoSqbWXXv3j21bNkyUclak8elS5fSiRMnIhzesmVLTPXKlSuJvrc7d+6syVN3bgoooIACCiiggAKVJGDAWkmwHlYBBRRQQAEFFKhoAVoBEEY2atQoDk2gl4V6/fv3T5MmTUqjR49O+RCwNm7cOA0ePDhCVn4ncM2HQV9bAtbly5en+fPnf2nKffv2jefatGnzpdd8QgEFFFBAAQUUUKAwBerc/mwU5tJclQIKKKCAAgooUFgCZ86cSbt27YrNq0qurGnTpolQj0pQAst8GLQK4EFFK498GPRbpWL16NGj6eTJk2VOmSC8d+/eZb7uCwoooIACCiiggAKFI2DAWjjX0pUooIACCiiggAIKKKCAAgoooIACCiigQBUL5EepQBWjeDoFFFBAAQUUUEABBRRQQAEFFFBAAQUUUKA8Agas5VHyPQoooIACCiiggAIKKKCAAgoooIACCiigQCkCBqyloPiUAgoooIACCiiggAIKKKCAAgoooIACCihQHgED1vIo+R4FFFBAAQUUUEABBRRQQAEFFFBAAQUUUKAUAQPWUlB8SgEFFFBAAQUUUEABBRRQQAEFFFBAAQUUKI+AAWt5lHyPAgoooIACCiiggAIKKKCAAgoooIACCihQioABaykoPqWAAgoooIACCiiggAIKKKCAAgoooIACCpRHwIC1PEq+RwEFFFBAAQUUUEABBRRQQAEFFFBAAQUUKEXAgLUUFJ9SQAEFFFBAAQUUUEABBRRQQAEFFFBAAQXKI2DAWh4l36OAAgoooIACCiiggAIKKKCAAgoooIACCpQiYMBaCopPKaCAAgoooIACCiiggAIKKKCAAgoooIAC5REwYC2Pku9RQAEFFFBAAQUUUEABBRRQQAEFFFBAAQVKETBgLQXFpxRQQAEFFFBAAQUUUEABBRRQQAEFFFBAgfIIGLCWR8n3KKCAAgoooIACCiiggAIKKKCAAgoooIACpQgYsJaC4lMKKKCAAgoooIACCiiggAIKKKCAAgoooEB5BAxYy6PkexRQQAEFFFBAAQUUUEABBRRQQAEFFFBAgVIEDFhLQfEpBRRQQAEFFFBAAQUUUEABBRRQQAEFFFCgPAIGrOVR8j0KKKCAAgoooIACCiiggAIKKKCAAgoooEApAgaspaD4lAIKKKCAAgoooIACCiiggAIKKKCAAgooUB4BA9byKPkeBRRQQAEFFFBAAQUUUEABBRRQQAEFFFCgFAED1lJQfEoBBRRQQAEFFFBAAQUUUEABBRRQQAEFFCiPgAFreZR8jwIKKKCAAgoooIACCiiggAIKKKCAAgooUIqAAWspKD6lgAIKKKCAAgoooIACCiiggAIKKKCAAgqUR8CAtTxKvkcBBRRQQAEFFFBAAQUUUEABBRRQQAEFFChFwIC1FBSfUkABBRRQQAEFFFBAAQUUUEABBRRQQAEFyiNQvzxv8j0KKKCAAgoooEAhC5w5cybt3Lkz7dmzJx05cqSQl+raFAiBli1bpr59+6aBAwemPn36qKKAAgoooIACCihwHwIGrPeB50cVUEABBRRQoDAETp06lZYsWZLeeOONtHLlysJYlKtQ4A4CPXv2TE888UR6/vnnDVjv4ORLCiiggAIKKKBAeQRsEVAeJd+jgAIKKKCAAgoooIACCiiggAIKKKCAAgqUImAFaykoPqWAAgoooIACtVegfv36qXPnzqlr166pW7dutRfClRecwNmzZ6MFxscff1xwa3NBCiiggAIKKKBAdQoYsFanvudWQAEFFFBAgRon0LRp0/TQQw+lJ598Ms2ZMyfVresNPzXuIjmhnATWrl2bXn/99bRw4cJ09erVnI7hhxRQQAEFFFBAAQW+LGDA+mUTn1FAAQUUUECBWixQp06d1LBhw9SsWbPUunVrA9Za/F0otKW3aNEiNWrUKNWrV6/QluZ6FFBAAQUUUECBahWwJKNa+T25AgoooIACCiiggAIKKKCAAgoooIACCuSzgAFrPl89566AAgoooIACCiiggAIKKKCAAgoooIAC1SpgwFqt/J5cAQUUUEABBRRQQAEFFFBAAQUUUEABBfJZwIA1n6+ec1dAAQUUUEABBRRQQAEFFFBAAQUUUECBahUwYK1Wfk+ugAIKKKCAAgoooIACCiiggAIKKKCAAvksYMCaz1fPuSuggAIKKKCAAgoooIACCiiggAIKKKBAtQoYsFYrvydXQAEFFFBAAQUUUEABBRRQQAEFFFBAgXwWMGDN56vn3BVQQAEFFFBAAQUUUEABBRRQQAEFFFCgWgUMWKuV35MroIACCiiggAIKKKCAAgoooIACCiigQD4LGLDm89Vz7goooIACCiiggAIKKKCAAgoooIACCihQrQIGrNXK78kVUEABBRRQQAEFFFBAAQUUUEABBRRQIJ8FDFjz+eo5dwUUUEABBRRQQAEFFFBAAQUUUEABBRSoVgED1mrl9+QKKKCAAgoooIACCiiggAIKKKCAAgookM8CBqz5fPWcuwIKKKCAAgoooIACCiiggAIKKKCAAgpUq4ABa7Xye3IFFFBAAQUUUEABBRRQQAEFFFBAAQUUyGcBA9Z8vnrOXQEFFFBAAQUUUEABBRRQQAEFFFBAAQWqVcCAtVr5PbkCCiiggAIKKKCAAgoooIACCiiggAIK5LOAAWs+Xz3nroACCiiggAIKKKCAAgoooIACCiiggALVKmDAWq38nlwBBRRQQAEFFFBAAQUUUEABBRRQQAEF8lnAgDWfr55zV0ABBRRQQAEFFFBAAQUUUEABBRRQQIFqFTBgrVZ+T66AAgoooIACCiiggAIKKKCAAgoooIAC+SxgwJrPV8+5K6CAAgoooIACCiiggAIKKKCAAgoooEC1ChiwViu/J1dAAQUUUEABBRRQQAEFFFBAAQUUUECBfBYwYM3nq+fcFVBAAQUUUEABBRRQQAEFFFBAAQUUUKBaBQxYq5XfkyuggAIKKKCAAgoooIACCiiggAIKKKBAPgsYsObz1XPuCiiggAIKKKCAAgoooIACCiiggAIKKFCtAgas1crvyRVQQAEFFFBAAQUUUEABBRRQQAEFFFAgnwUMWPP56jl3BRRQQAEFFFBAAQUUUEABBRRQQAEFFKhWgfrVenZProACCiiggAIKKFArBC5evJhOnjyZzp49m27dupU6dOiQ2rRpk1q1alXj13/jxo30ySefpHPnzqXTp0/HOvr165d69+5d4+fuBBVQQAEFFFBAAQUqX8CAtfKNPYMCCiiggAIKKJCzwO3btyOQvHbtWrp+/Xoi7ONx8+bNOGbdunVT/fr1U+PGjYsePFcTxqeffpp4MPdDhw6lLVu2pIMHDybWNG7cuJh3TQ5YCYIJVi9duhTB8EcffZR2796d9u7dm55++mkD1prwJXMOCiiggAIKKKBADRAwYK0BF8EpKKCAAgoooIACZQkQpl65ciVt3Lgx7dixIx04cCAex44di4+0aNEide7cOY0YMSKNHDkyjRo1KjVv3rysw1Xp85cvX45qzw0bNiQe69atS8y7U6dOadCgQVU6l1xOhjve+/btS7t27UrvvvtuBMUNGjSIgDiXY/oZBRRQQAEFFFBAgcITMGAtvGvqihRQQAEFFFCgAASonOR29D179qSdO3dGMMkt6lSuNmvWLHXr1i1WSZXl+fPn09q1a9PHH3+cCDUHDx6cevXqVe0KVNISRhL4UslKUHnixImoXGUNTZs2rfY53mkCzL9hw4apUaNGqV69ehG2UoHbv3//VFOqhO80f19TQAEFFFBAAQUUqBoBA9aqcfYsCiiggAIKKKDAPQlQPblt27b06quvpl/96lcRmj744INp2rRpaejQoalPnz5xPG5bX7ZsWXrllVfSihUr4vb1F154oUYErISoWRhMYLxo0aK43Z62AD169IgerPeEUsVvbtKkSaLXart27VLr1q3T66+/Hq0CqL7lOYcCCiiggAIKKKCAAggYsPo9UEABBRRQQAEFapgA7QBWrVqVli5dmtgcaubMmWnKlClp+PDhqWvXrqlt27ZRGcq0+TOhK31OCVi3b98et7J36dIlDRw4sMqCwAsXLsQmUASPhKqMOnXqROUqa6BylepPXmdzKCpDeb0mD+bHg4256L1KL1aqbnv27JlatmxZk6fu3BRQQAEFFFBAAQWqUMCAtQqxPZUCCiiggAIKKHAngatXr0ZI+cEHH6QFCxZEWDpkyJD0xBNPpAkTJhRVrRY/BpWVPKh4pT3A8uXL04cffhghIGFmZVdaUpl68uTJeBBEEpxmASvzpKUB4erRo0fjPX379o258b58GWfOnCkKWGl3MGDAgAi582X+zlMBBRRQQAEFFFCgcgUMWCvX16MroIACCiiggALlFqCH6nvvvZfefPPNtHnz5kRLgMcffzzNnj37rhtX0XP1gQceSI0bN46NmKhmnThxYrnPnesbCVXfeOONqPKkqpOKWjbdygabdLFR1JEjR+IpAl+qa/MpYKU699ChQ4kAnN63XBc26nIooIACCiiggAIKKIBAXRkUUEABBRRQQAEFqleADaCo8tywYUN67bXXYsd6bkUnIB0zZkxUobLR0p0G4SaVrGzGRMUlvVm5pb0yBvOlJcHWrVvTwoUL01tvvRUhKuFjyY2rCFipXmV9t2/fjmCSMDifAlYqdOmHy8ZWBKsdOnT40jorw9ljKqCAAgoooIACCuSHgBWs+XGdnKUCCiiggAIKFLAAt9Hv2bMnZa0BWrRokcaPHx8BKxsqlWfUr1+/qK8pweqxY8ciBCUMJeS8detWBJz8mUFvUapdGzRoUHR4AlAevP/69esR1nJc3kO4mL1GKwKqOtlca/78+dH7dcaMGbFxFe/jdTaI4nfWRmXuqVOn4pxUr94tYM3mwVxpQcAx+J3nGRw3Wy9BbVm9XLPjsB4eHAcHHnyGMLr4cYog/u8XPs95mT+9bZk7D1og8FmOx3F5Dw/mxfN48ZM/OxRQQAEFFFBAAQUKX8CAtfCvsStUQAEFFFBAgRouwK3n77zzTlr62aZWBIrc6s/GVdxqX95KTzaS4nZ9gr7ig2Bzy5Yt8RqVrVRjEjBSaUr7AXq8ZoPPnj9/PtoTLFmyJG6Hp98oVbRUyDI3qlbXr18fm3AROhIMc+6VK1emv/3bv039+vVLDz30UHruuedSq1atIoCkRcDx48fjnKzpbi0CmB8BLxtL0U+WalxaDBDqElxy3KFDh6Zx48ZFCM1zpQ0CUI6zadOmxMZhHIdKWuZLRTDVqMOGDYv5jho16kuHoEoXLz7DcWh/0KNHj5gDITZrwoP2AbwPI6p42YyMNVZ2/9svTdgnFFBAAQUUUEABBapFwIC1Wtg9qQIKKKCAAgoo8DuBc+fOpb1796Z169ZFoEjwR3jIplZt27aNCsvyWHEcwlRCRaoyCVCz4JFqTYJTwtDVq1dHZSWbTZXs0Uo1JkHm2rVr0+uvv57Gjh0blagjRoz4whSo7OQ8nI9H+/btIwxmAyiqN7OKUgJZwkxaBFDVSujIe6nQLWuwWdfp06cjFCa85MFzHIvzcj4CW47JsQlbu3fv/oVNpwho+Qxr4fMEzKyd55g3g88fPnw4jsWx8SJAZQ3ZIGDlXASsHJNAljXgyp95nZCZIJjwl2CcuTAyg+xY/lRAAQUUUEABBRQoXAED1sK9tq5MAQUUUEABBfJAgOrH5cuXR8hKYEfVIxWQhJsEpeUdhJK0BSAk5fZ8jkP1K4HmpEmTItQknCR0JODkeVoEFB9Uae7cuTN6wRJMUo1afHA85tazZ8+o+uRcnJNNn+bMmZO++tWvRvhIyMqxCXWp7OTB7fIEkG3atCl+yC/9TpUtYfMPfvCDmCvreOKJJ2INAwcOjErZefPmRbUvx6Xq9plnnvlSwEoouvSzimCOQ9jZu3fvOM7o0aNTnz590ty5c9OCBQsSlboEr9lxigesVBYTnlKpyvw7duwYD37nfVSsZn1uCawfe+yxsOC6ZeH2lxboEwoooIACCiiggAIFJ1D+/2ovuKW7IAUUUEABBRRQoPoFCAmpgiSMpBqTatG73UJf2qwJT+kVSuhJKMmt/YSABH08qCDlXISGVGJyq3vJDakIWHfs2BHVnhyHTbMIFbOgl6CS8JQqUEJHwkVCV6o2S1aRMkfmREDJ++hbSshZVvUqISctDtgwa9GiRVHFSiUvoSW38ROKsi4CUs5PmwLCaQJSwuhsUElK6EsF7ooVK+K9U6ZMSfSI5Tjcwk/IS+jMe999991oHUCLA95XfGBF5StrxYDAmQe/48TGV8yJcJVeubxGuO1QQAEFFFBAAQUUqF0CBqyKdoilAABAAElEQVS163q7WgUUUEABBRSoYQKEkFSNEkJSVUrASgB6r4NwldvZud2dTaSo9iweZhJecls8r1N5SShY/HWep83Arl270sGDB+P0zIfb5glViw/mSj9TjklLAypaCWJLDgJdAlbCXUJN5lW8QpT3c9s/t9tTgUtoSlXp+++/H/OfPHlyevHFFyMIzuZASMsg/N28eXP0VsUwOw4VuqtWrYoKVVoAEHw++uij6cknn4zwMwuLBw8eHC0CqLTFjopd1lV8ZAErVbVYESJzXqpjWT9OzIdzEAY7FFBAAQUUUEABBWqngAFr7bzurloBBRRQQAEFaogAwSKhJuEeoSfBH71X73VwDAJKKkGpqiRgLR5mEhISsFKlyvG5XZ+gNRuEhlTS0g+WwJJBuErlKCFq8cE5qB6lApSKTY5FeFpyZEEkASvhZFkBK3OiTyq381MVSujJBlwErFS+ckt+aYPb+rPb+/mdFgtUrf7Hf/xH2r9/f1Tx/v7v/35iAyvmWdZt+1TrEqZyjOIjC1ipmMWTal2CVTYk69ChQ1TOErDi7VBAAQUUUEABBRSovQIGrLX32rtyBRRQQAEFFKgBAgSEBJpUkBImEmjSKqC8g6pTglOCUcLarKKUStjixyHspKKTkJEqTKpTiwenBKYEtLyPVgC8TtVp8ZA2mxPBKhWfBJBUrnbu3PkL58reRwUrgSRrJMwlpOTcxQfhJVWu9F3duHFjVMsSMo8ZMyYssorT7DO0UqAFAGEqgSeVpayDEJeAmONw+z7zHzJkSISgBMXFj0O1K9W3zI/zE75ilVXJci5e5zz8JOBl/mwQxnv37dsXgXH//v0jXC3umM3TnwoooIACCiiggAK1R8CAtfZca1eqgAIKKKCAAjVQgICPsJBBT1R6oxavLC1ryoSEDG5vp48o7QGowKSykv6rI0eOjNezW+fpI0qISiBKBSthLoPXefD5tWvXRi9YXiur2pTPUA27YcOGCCSpTCWcJOzMRnbMrEUA6yNYJTgtHmLyfoLlDz/8MK1ZsybOTY/VCRMmxBpKqwwtXhXLPKkgJQQmaH7vvfci+MWUc40bNy5C2myt2fxoSUDYzJqzXrOso/imX4TWBKlU1xKqErK+9tprMX9aArBmeuU6FFBAAQUUUEABBRQo/X4rXRRQQAEFFFBAAQWqXIDqUkLWkiFkaRMhxCS4pHJ1/vz5ReHps88+GxtBZZ8hDKUylYCVSksqZKk4zQZVqISNVJlmQSKBZWkBKwEuxyHUpaUBAWhpGztlxyRg5TOEl2wuVdrgdYJfHgSfzK20Y2afJfjctGlTunjxYpyfSldC5ex2fuaHI0E15y2tLQB2uNH7loCXClcqfosH21SvZgErz+PBg+peNsRi8yvOyZwdCiiggAIKKKCAArVbwIC1dl9/V6+AAgoooIAC1SxA1STVnYSqhHc8yjMIBqnAZKMnKkAJDemF+sgjj8SmS9kxuJ2fW+dpQ0CVKUEnt8/zfio9CVYJEwlMs3MT8hI6lqz8JAzlvWwkRbhLsEnf2OKVn5yXY9IDlYCV83BOgtPs+Nnc+MkxOR7HJazkvLQSIOwtPngfgWa2ZipLCVHp00olKfPJzkmoSpDc+7Pq1pIBK2umnQCtC2hNQGUqt/qPHz/+C+0LmA/H47wc/8EHH4yqYOyYAwEtwTTzcCiggAIKKKCAAgrUbgED1tp9/V29AgoooIACClSzAIFdtiFVFnhy2/rdBiEmrQG4LZ4gk7YAhKu0BujUqVPRx3kfu90TtNKHlDA366tKOMjzhK/cYp9VZFLpSnhaMjglbORYWbUp5ymt2jQLMJkXwTHvY52lDUJV3seD3wlNhw0b9qVwl0CZHrMEo/R/JRilDQCbYfXs2TOCTqpbWQfnxINH8d6rnJ81ci4206JClR6umM2YMeMLm1VxLILU4ufBd+jQoXF8KlhpqYCfQwEFFFBAAQUUUKB2Cxiw1u7r7+oVUEABBRRQoJoFCB8J+OjpyYZKbNLEbe53GlRevv3222nRokVR+Tl8+PA0c+bMCFgJMou3GCgesBKccps75yTMJCilpykBJRWjnJ9wl1viCX0JNaku5SeDADi7nZ+wlapUjkf1K8FmNrJKU47HXDg+7QR4D1WfxY+ZHZdjM6iyZS4lK08JgZcuXRrVuISe06dPT1OnTo3AmM9kFbnMi0pZ1sqjZNUsAfFbb72V9n9WnUrY/Mwzz8RGWASthLHMg/lzDWivQPUqVbVs+JWFtqyFNRCw0qqAgSfrZZ7ZWuIF/0cBBRRQQAEFFFCg4AUMWAv+ErtABRRQQAEFFKjJAgR49BHlJ2Hd+++/H7euUwVKWMggPMyqTQlECfbefPPNaA9AgEm4yoNNnUre1k/ASiUmlZYEkYSdBIvnz59PVGkSjnKbP6Em4SBhKmEjzxG2cqt8VlHLfAhY+RzHImhkwyyCRR7ZIKAkwOQ5QsvsnISWfLZkCEmlbFYty3yZKyEloSXnZq5UjBIq83lC3SeeeCI2wyJsZfCT1gbMi8+xLh78zrzxIzCl+pWetbxGS4UXXnghEVBnI3sfc+UaZIE0gS+BMi48mCMbffE+jsXnqIzl+mTXLTumPxVQQAEFFFBAAQUKW6B+YS/P1SmggAIKKKCAAjVbgPBx2rRpEUiuWrUqLVmyJCovCSapbCUMJSQkYOS29o0bN8YmTwSNDz30UHr44Yfj81RZljYIAnfs2BG311OZyvG41X7NmjURfNIflVCS6lKqMgkKs0EFKMFnVgVKcEgFKIEnfUt5nj8TMmbv4bOck42omCMBLMEkQTC9YAcNGhShJcdm8DwtAVgfn2F9CxcujOpUwmJC0WXLlkUrBKp7CaO/8pWvRNUpbQyywdrGjh0boSgh8PLly6MtAVWurInKWY5DgL1r164IaLnln5AVk2xgw6ZghKUEtrRAoOI3GzhT9Tp37txYO8ek/yutDdj8Kwt4s/f7UwEFFFBAAQUUUKDwBQxYC/8au0IFFFBAAQUUqMECVItSJUkPUMI5QlbCQG7/Z5d7Qj4CVoJGqiMJLWknQKA3ceLENGrUqOhFmlVylrbULPwkaKSalcG5aCdASEl1Jg/CTkJUHpyPYJSAlLCX5/g8FaYEilTV8h7CVYLQrK8rx86qRvlJJSrzJpjkWDwIQ7PBPAhGCWV5P8HmvHnzonIWG85F+Mq5CVdpDTBlypSYe/GNsAhJCaqZJ4b0an399dfDkvVnVbCcj9AVbyp+MWB92WB9+PNc9+7dIxAuHrDixfnp38pcV65cGRWsHIs58LjTtcjO408FFFBAAQUUUECBwhH4/L8mC2dNrkQBBRRQQAEFFMgrAapEH3300dj5nuCUjateffXVuB2fwI/Ajg2deFABSrBK2EiwWVqf0eKLJ6DlmNzuTsi4ePHi9OSTT6ann346QtGskpRgkECR2/e5LZ8glupaWhcwCDgJOwkeCUIJIXkPbQCoQO3YsWPRaVkPlbEci3YBBJ6ElM8991yEuEVv/OwXAs9JkyZFaMznsrX//Oc/j3CZ444ePTrWPHv27DgO685C4+xYtCvgdcJbjkOlKud96aWXon8qa+M4WTsFwtviAW12HAJazAlSCYWHDBnyhQpWerFyHPrgUvFLpSyhN9fo+eefj+tTcm7Zsf2pgAIKKKCAAgooUJgCBqyFeV1dlQIKKKCAAgrkkQCBHBWkhJBUR3JbOj1G6clKBSivE+xlDypLCQgJO+9WLUko+73vfS+CUKpOCUypmM0+n4WBVHVSzUkgynnoM8p7s9c5F69/+9vfTk899VTochweVNQWnwef/e53v1vUa5Wgtnfv3jHf7HjZ5eFzhMC0HJgzZ05UlXKbPuvmnISghJ3MBR/+XPxc2XHwI3iln2pWzUqgzHEIkamw5TgEv7zOsUvOhWNxnlmzZkVVLedhY6vi1bl8hgcVq1Tj0i6A87JGHqXNLZujPxVQQAEFFFBAAQUKU8CAtTCvq6tSQAEFFFBAgTwUoPqSB1WqFTUIFHncbQwdOjTxKGsQSBKk0v/0boMglIC4PIOwktYEfIZHroNgkyA1CzpzPU7WxuBunycE5+FQQAEFFFBAAQUUUOB3267qoIACCiiggAIKKKCAAgoooIACCiiggAIKKHDPAgas90zmBxRQQAEFFFBAAQUUUEABBRRQQAEFFFBAgd8JGLD6TVBAAQUUUEABBRRQQAEFFFBAAQUUUEABBXIUMGDNEc6PKaCAAgoooIACCiiggAIKKKCAAgoooIACBqx+BxRQQAEFFFBAAQUUUEABBRRQQAEFFFBAgRwFDFhzhPNjCiiggAIKKKCAAgoooIACCiiggAIKKKCAAavfAQUUUEABBRRQQAEFFFBAAQUUUEABBRRQIEcBA9Yc4fyYAgoooIACCiiggAIKKKCAAgoooIACCihgwOp3QAEFFFBAAQUUUEABBRRQQAEFFFBAAQUUyFHAgDVHOD+mgAIKKKCAAgoooIACCiiggAIKKKCAAgoYsPodUEABBRRQQAEFFFBAAQUUUEABBRRQQAEFchQwYM0Rzo8poIACCiiggAIKKKCAAgoooIACCiiggAIGrH4HFFBAAQUUUEABBRRQQAEFFFBAAQUUUECBHAUMWHOE82MKKKCAAgoooIACCiiggAIKKKCAAgoooIABq98BBRRQQAEFFFBAAQUUUEABBRRQQAEFFFAgRwED1hzh/JgCCiiggAIKKKCAAgoooIACCiiggAIKKGDA6ndAAQUUUEABBRRQQAEFFFBAAQUUUEABBRTIUcCANUc4P6aAAgoooIACCiiggAIKKKCAAgoooIACChiw+h1QQAEFFFBAAQUUUEABBRRQQAEFFFBAAQVyFDBgzRHOjymggAIKKKCAAgoooIACCiiggAIKKKCAAgasfgcUUEABBRRQQAEFFFBAAQUUUEABBRRQQIEcBQxYc4TzYwoooIACCiiggAIKKKCAAgoooIACCiiggAGr3wEFFFBAAQUUUEABBRRQQAEFFFBAAQUUUCBHAQPWHOH8mAIKKKCAAgoooIACCiiggAIKKKCAAgooYMDqd0ABBRRQQAEFFFBAAQUUUEABBRRQQAEFFMhRwIA1Rzg/poACCiiggAIKKKCAAgoooIACCiiggAIKGLD6HVBAAQUUUEABBRRQQAEFFFBAAQUUUEABBXIUMGDNEc6PKaCAAgoooIACCiiggAIKKKCAAgoooIACBqx+BxRQQAEFFFBAAQUUUEABBRRQQAEFFFBAgRwFDFhzhPNjCiiggAIKKKCAAgoooIACCiiggAIKKKBAfQkUUEABBRRQQAEFPhf45JNP0u7du9OiRYvS2bNnU506dT5/0d8UyGOBffv2pY0bN8b3ukGDBnm8EqeugAIKKKCAAgrULAED1pp1PZyNAgoooIACClSzwNWrV9OaNWvi8dJLL1XzbDy9ApUj0LNnz8o5sEdVQAEFFFBAAQVqoYAtAmrhRXfJCiiggAIKKKCAAgoooIACCiiggAIKKFAxAlawVoyjR1FAAQUUUECBPBZo0qRJ6tWrVxo9enRq3LhxHq+kZk390qVL6fz58+nEiRPp2rVrqV69eqlp06apZcuWqX379qlRo0bxXM2ade2YTceOHdPAgQNTmzZtaseCXaUCCiiggAIKKFCJAnVufzYq8fgeWgEFFFBAAQUUqPECN27cSJcvX060ByAIdFSMwK5du6LVwhtvvJGOHDmSmjVrlvr165dGjhyZZs6cmbp06WKgXTHU93yU+vXrJ/5hgWvCT4cCCiiggAIKKKBA7gJWsOZu5ycVUEABBRRQoEAE2PCndevW8SiQJVXLMvh3e8LqkydPpm3btkW4umrVqvTxxx8nqlnZQOzAgQOJ99WtWzeNHTs2jRo1KnXo0CGCvmqZtCdVQAEFFFBAAQUUUOA+BQxY7xPQjyuggAIKKKCAAgqkdPPmzQhQT506lbZs2ZLmzp2b3n///fidMJVx69atCF/37NmTNmzYEGErnxs2bFjq1q1btA/gvXXq1JFUAQUUUEABBRRQQIG8EbBFQN5cKieqgAIKKKCAAgrUXAH6rO7fvz8tWLAgrVixItEegLCV1gv0XP3000/TxYsXI0SlYphq1nbt2qXu3buniRMnpsmTJ6epU6emFi1apIYNG9bchTozBRRQQAEFFFBAAQVKCFjBWgLEPyqggAIKKKCAAgqUT4Dq0ytXrqSjR49GpeqaNWvSsmXLIlwlQO3Ro0f0WT137lw6fvx4BKz9+/dPbLBEIEvrgLVr18Yxzpw5E4HsAw88EH1a27ZtG5tglW8mvksBBRRQQAEFFFBAgeoTqPf9z0b1nd4zK6CAAgoooIACCuSrABuCHTt2LC1cuDD96le/Sj/96U/T4cOHo8cq1anPPfdc+vrXv54IWOnLymPOnDnpsccei5CV6laqXnl+x44d0VKASlc+2759ezdfytcvhvNWQAEFFFBAAQVqmYAVrLXsgrtcBRRQQAEFFFDgfgW41Z9gdeXKlWn16tVp69atib6qWT/VkSNHpjFjxqTRo0dHWNq0adPoq0pv1UaNGkVbgIEDB6a+ffumIUOGpPXr10fQev78+bR06dKodh0xYkQaN25cGj9+fGrevHlq3Ljx/U7bzyuggAIKKKCAAgooUCkCBqyVwupBFVBAAQUUUECBwhJggyoCVELQffv2xSZV8+fPj5CVwLVJkyZp8ODBadq0aWn69OnRV7VNmzZRvVq8p2q9evUSzxOsErD269cvqllXrVqVtm3blg4dOpQOHjwYFa20ELh9+3bq3bt36ty5c2rdunXi8w4FFFBAAQUUUEABBWqSgAFrTboazkUBBRRQQAEFFKihAjdu3Iiw9L333osq00WLFkUfVdoEsDEVG1U99dRTadSoURGI8lz9+nf+T03aAFCdStBKpeqbb74ZFbFskHXgwIE0d+7cqG6dMWNGbIA1ZcqUeH8NJXJaCiiggAIKKKCAArVU4M7/1VtLUVy2AgoooIACCiigwO8E2KzqwoULaefOnRF2Umm6cePGtHv37tS1a9dEO4Bhw4ZFQPrwww+nTp06pZYtW5aLj8pWHryfMJYHYeumTZuiQpYK1i1btiT6srIpFv1dOdeAAQMi1KXdgEMBBRRQQAEFFFBAgeoWMGCt7ivg+RVQQAEFFFBAgRoqQLjKBlV79+5NCxYsSK+88kqEnFeuXImeqPRJnT17dpo1a1bq2bNnotdqrqNLly6JB5WsbHj1s5/9LK1YsSJt3749wlae++CDD9LTTz+dnnnmmQhi27ZtG6EsvV15OBRQQAEFFFBAAQUUqA4BA9bqUPecCiiggAIKKKBAHgh8+OGHEXLSFuCjjz5KR48eTQ0aNIjNq7htn02ohg8fHlWrxfus3s/SmjVrltgA68/+7M/Sgw8+GD1e2UyLlgFHjhxJr732WmyqNXny5PTQQw/FXOj/yrwcCiiggAIKKKCAAgpUh4ABa3Woe04FFFBAAQUUUKAGCnAr/tWrV9PJkyejUpU+q8uWLUtr166NzabatWuXhg4dmiZMmBBVq9zO37FjxwpdCbf986A/Kw+Oz6ZY69ati7YEtAkgbGWzrVOnTqWzZ89Gy4Bu3bpF24C79X2t0Ml6MAUUUEABBRRQQAEFPhMwYPVroIACCiiggAIKKBACN2/eTPQ9Xbp0afrlL38ZVavHjh1LtArINqL62te+lsaOHRuhZ0VVrZbFT49XQlYqWenLStsAesBSTbt69er4uXjx4vTiiy9Gq4JBgwbddWOtss7l8woooIACCiiggAIK5CpgwJqrnJ9TQAEFFFBAAQUKSGDXrl3R65TgcsOGDXEbPr1WCTjHjBkToSo/2WSK57glv7L7nlKNWq9evQhN6fdat27dOD/zo6L1+PHj0R927ty56eDBg9G/ddSoUWnw4MExP6tZC+gL6lIUUEABBRRQQIEaLGDAWoMvjlNTQAEFFFBAAQUqS+D27dspawnAbfZsKDV//vxEv1VaBNDXtFOnThFoPvfcc9HvlArRqh6EuFTK0gKAxwMPPBBtCthQix6xBKuErTt37kz79u1Lp0+fjnYGnTt3Tq1bt451VHYQXNUmnk8BBRRQQAEFFFCgZgkYsNas6+FsFFBAAQUUUECBKhEgYL1w4ULavHlz+u1vfxshJbfeE7ZSoUoV6JNPPpkmTZqUevToES0BqmRidzkJPVlpUUCAun79+ugRyyZY+/fvj16xhw4dihYHTz31VJo4cWIEsm6AdRdUX1ZAAQUUUEABBRS4LwED1vvi88MKKKCAAgoooED+CbA5FBtFbd++Pa1ZsyaxmRW9V+nBOnz48DRy5MgIMdnMiqpVNp3i9vyaMBo3bhwVra1atUotW7ZMbdu2jUpbwlbaHLAJ1tGjR6OKlf6x9G9lDYTEfNa2ATXhKjoHBRRQQAEFFFCgsAQMWAvreroaBRRQQAEFFFCgVAEqVhm3bt2KW+lpB7BgwYKoYL18+XKEj4SQtAOYNWtWBKw1JVQtuSDmRVg6YMCAeNAblqD45Zdfjv6xhKxvvvlmbIQ1ZMiQxMZcjz/+eKL6lZ6utgwoKeqfFVBAAQUUUEABBe5HwID1fvT8rAIKKKCAAgookCcCly5dSkeOHEnLly9PH3zwQdxOz58JHAkqp06dmqZPn54IJAla8ymEbNeuXQTCbdq0SbQLoI/sjh070rlz56I3689+9rO0ZcuW6CNLRStVug4FFFBAAQUUUEABBSpKwIC1oiQ9jgIKKKCAAgooUAMFrly5Ehs/0aN048aNsZEVYSO3z3N7Pb1WR48enZ544okIWPPxNvrmzZsnHr17906ErbQPYIOubdu2RSsE1r137950/PjxsPjkk0+ihyuBLOutqZW6NfDr5JQUUEABBRRQQAEFShEwYC0FxacUUEABBRRQQIFCEaC36tKlS9PixYvjlvmTJ09Gr9UWLVrEJlCPPvpotARg06imTZvmVeVqadeIfqtdu3ZNU6ZMiUrdX//611HNSrj69ttvR+hKhSutENjAq3v37tHTtbRj+ZwCCiiggAIKKKCAAuURMGAtj5LvUUABBRRQQAEF8kjg2rVr6fz581GxSm9SbpvnlnlCRqo7qVodO3ZsbGY1bNiw1LNnz6jkzKMlljnVJk2axKZc/GzQoEEiSF67dm1iEywqWtkA6+LFi+nGjRtp9+7d4UAoiwHvp2WCQwEFFFBAAQUUUECBexEwYL0XLd+rgAIKKKCAAgrUYIFPP/00gkOqVgkPqd58//3343eCQ26fp//oY489FhWc/LlZs2Y1eEW5TY1b/glWhw4dGg8CZQJUDHbu3JlOnToVrRIIXfft25eo4uUzeNBqoH79+nlfyZubnJ9SQAEFFFBAAQUUyEXAgDUXNT+jgAIKKKCAAgrUQAE2stq1a1dasmRJWrhwYYSHZ86ciVvg2dxpwoQJ0WeVwLF9+/a15tb4fv36pZYtW0a4vGLFirDB6fTp02FFf1qqfAmeR40aFWFsPm3yVQO/ik5JAQUUUEABBRSoVQIGrLXqcrtYBRRQQAEFFCg0AapWeRASbt++PW6Hp2qV1gBs4NStW7fELfDjx4+PBwEimzvVptG6desIWLGgRQJVqlnLACpYN2/enI4cOZKuX78eP0eMGJH69OkTG2VR9eomWLXp2+JaFVBAAQUUUECBexcwYL13Mz+hgAIKKKCAAgrUGIGbN2+mK1euxAZOc+fOjYD13LlzEQr27ds3bn//gz/4g6jKbNu2ba3tMUpFKn1ZR48eneg7Sz9WNv/6yU9+EpW+BK2HDh2Klgrjxo1LX//619O0adMKsoVCjfnyOhEFFFBAAQUUUKBABAxYC+RCugwFFFBAAQUUqF0CVFseOHAgbdiwIULBTZs2Ra/VCxcupB49esTt8PQWJSykGjPrLVq7lD5fbXbLPxWpVPb27t07PfLII9EqYd26dYnH1q1bE/1rV61aFaH1hx9+GH5DhgwJw8+P5m8KKKCAAgoooIACCnwuYMD6uYW/KaCAAgoooIACNVrg1q1biQcVqtzSvnbt2vTOO++kN954I1HJ2qhRo9SrV68IBQkPCVgJVx1fFCBsZUMr2gX079+/qB0AfvRmJWQ9ceJEUeCKN20Y3ATri47+SQEFFFBAAQUUUOB3AgasfhMUUEABBRRQQIE8ESDku3r1alq+fHmaP39+BKxUsV6+fDl17Ngxbn1/5plnEhtaERwSIDrKFqCatWnTprGxFVW/tA9YtmxZotXC0aNH0+HDh9Orr74aQevq1asTtiNHjowetnzWoYACCiiggAIKKKAAAgasfg8UUEABBRRQQIEaLkDV6pkzZ9LevXsjVGXHe6pX6RlK1SUbV2WbWNESgM2c2NjJcWcBKlkJSlu2bBlBa7NmzaKVQocOHRLtAWgZQG/WHTt2RNXwtWvXosJ18ODBacCAAal79+53PoGvKqCAAgoooIACCtQKAQPWWnGZXaQCCiiggAIK5JvA7du3Y8pUrRLs7dmzJzay+uEPf5iOHz+ebty4kRo0aJD69euXHnvssTRnzpw0ZsyYfFtmjZlv/fr1owqYSuDJkyenRYsWhfdbb72VDh48GCErYSvh9qRJk8K7TZs2qWHDhhHS1q1bt8asxYkooIACCiiggAIKVK2AAWvVens2BRRQQAEFFFCgXAJUrbKR1e7du6Oa8t13342NmE6ePBk721OlOn369Oi3Onz4cKspy6Va/jcNGzYsWiyMGDEiWjLgf+zYsQhbr1y5EiE3G2PNmDEj2jFQ9epQQAEFFFBAAQUUqJ0CBqy187q7agUUUEABBRSowQKXLl1Kp06diqpVblXPWgKcPn06qizZ1Z5qVTax4nb1Tp06JW53d1ScQOfOnVPbtm1T7969o+dq8+bN0+bNmyPwpjXDhQsXokfrxYsX41oQyFL9SrsB2g54PSruWngkBRRQQAEFFFCgpgsYsNb0K+T8FFBAAQUUUKDWCbCL/Zo1a9IvfvGLtGnTptjV/pNPPkldu3aNikk2W6JysnHjxtEmoNYBVdGCacFAyIo1gTZh98KFC+O6nDt3Lm3cuDHt3LkzeuDOnj07Au+hQ4dGP1cD1iq6SJ5GAQUUUEABBRSoAQIGrDXgIjgFBRRQQAEFFKjdArQDoKcqvVUJ7FasWJHYtZ6KSSok2Xxp6tSp0Q5gwoQJiQpWN7Gq/O9MtglWkyZNIsimXQChdo8ePaJdA+F3tgkWAfiBAwfS6NGj08MPPxwtG9q1a1f5k/QMCiiggAIKKKCAAtUuYMBa7ZfACSiggAIKKKBAbRYgWM16ehKoLlmyJNHvc/v27RGscvs/t6k/++yzacqUKdHvs1GjRrWZrFrWziZY3bt3jzYAY8eOTUuXLo3qVp5nEyyqWz/66KPYDOvq1asRtA4cODD6uHK9CGutaq2WS+dJFVBAAQUUUECBShcwYK10Yk+ggAIKKKCAAgqULUCv1V27dqU33ngj2gJQEUmv1RYtWkQlJDva8+jZs2diIyVuW3dUnwCBatOmTePaUMlK6wDCVloHXL58uailAz1ZH3roofTEE0+kfv36ReVr9c3aMyuggAIKKKCAAgpUpoABa2XqemwFFFBAAQUUUKAMAW79J0xdv359tAOgLQC3mN+8eTMCOQI6Kla55Zy+nlRBEu45qlegbt26iUe2oRWtGgjD+UnVMVWsXFeuL0E5G5Y9+OCDadCgQalLly6xYVb1rsCzK6CAAgoooIACClS0gP+VXtGiHk8BBRRQQAEFFChD4Pbt2/EKP6lcXbBgQZo3b15atmxZPE+vVapUqXqcM2dO9FplV3pHzRSgH2ufPn1i8zF65M6fPz8eBKtsVEZPXVoH0FLghRdeSNOnT4+WAVmrgOxnzVyds1JAAQUUUEABBRQor4ABa3mlfJ8CCiiggAIKKHCfAtxCfvLkybR8+fL0wQcfRPhGtWPDhg0TGyhRrTpu3LhE9aq3ld8ndhV+nMpiqlgnTZoUla2jRo1KK1eujOt77ty56Mv6v//7v2nr1q1xfWn50LVr19S8efMqnKWnUkABBRRQQAEFFKgsAQPWypL1uAoooIACCiigwP8JsJEVwSphKiHbW2+9ldauXZtOnDgRwdzw4cPTo48+mqiCpG8nwZsbWeXP16devXqJB6F4586d04ABAyJoJXTdtm1bOnz4cIStVLUeOnQoerUSolP9SsUylcsOBRRQQAEFFFBAgfwVMGDN32vnzBVQQAEFFFAgTwTow5lthLR48eJ04cKFdP369ahcJVB95plnIljt3bt3bKBEj09Hfgo0adIkde/ePVo8cG0XLVqUlixZEtefkJ12EPTdZXOsmTNnpscffzwC2fxcrbNWQAEFFFBAAQUUQMCA1e+BAgr8f/beq0mOI7vfThoQjvAgvPfee28IECBoQHKJVZC7ywvFSpfSnT6APoAUoTtpdbG7XENyF6AF4b0HCO+99957vHzO+y+qORyAADmDme5+MqKme7qrsjKfqsyu/OXJcyQgAQlIQALlQODmzZvp0qVLsTx8/fr1iSBWmzdvDmtGrByxWkVkGzRoULgGaN68edLfajlciKecJeI41sdsNWrUCL+rDRs2TB06dAgr1p07d4Y1M/fE9evXIxAWriG6du0aVq8EyzJJQAISkIAEJCABCeQXAQXW/LpellYCEpCABCQggUpOgABWiKtYKxJR/ssvv4zAR8eOHUu4CqhXr174WB07dmx69913w9oRq0dT4REgCBa+ddu1axf+WadPnx5Wy7iJOH/+fFq8eHHau3dvuI2YMmVK6tu373e+d3E5oCVz4d0T1kgCEpCABCQggcIkoMBamNfVWklAAhKQgAQkUEEE7t27F0vAly1blubOnZv2798fYivFwWp1woQJEVUeH5xNmzYNwa2CiuppnxIBhFZ8reIKonv37gnrVcTVpUuXhtCKdTP3CcGvRo4cGYGwsHLmOJMEJCABCUhAAhKQQOUnoMBa+a+RJZSABCQgAQlIoJITQFTFpypBjPbt2xfi2fLly9Pq1atTlSpVIvBRp06d0pAhQ9K4cePCSrFx48aVvFYWr6wIPP/884mNIFiNGjVKXPs6derExv1C4KtNmzbFPXThwoUIfoYQy/7169c3CFZZXQjzkYAEJCABCUhAAuVEQIG1nMCarQQkIAEJSEACxUPg7t27CWGMAEYfffRRLPk+fvx4QnglcNXw4cPT+++/H8vFiSzP8m9TcRJ48cUXw0VE27Ztw5oZtwFz5swJa1b8sx48eDAhzo8aNSpNnTo1/PPWrFmzOGFZawlIQAISkIAEJJAnBBRY8+RCWUwJSEACEpCABCongV27dkXwqnXr1qWNGzeGuHrt2rXwrdqnT580dOjQRDT5Ll26JMRVLBmfeeaZylkZS1XuBLj2COwEwMINwMSJExMBzvDVumbNmrRjx47v3AacPn063EkQCI17CKvXF154odzL6AkkIAEJSEACEpCABJ6MgALrk/FybwlIQAISkIAEipwAQaywWL169WpEgMfX6sKFC8Pq8OzZsxGYqEWLFiGYTZo0KSGOESHeJIFcAgjtWLMiwrdq1Sp17tw5NWzYMBHwDH+suJsgANbhw4fjPfcdwbJwMVC7du1UtWrV3Ox8LwEJSEACEpCABCRQgQQUWCsQvqeWgAQkIAEJSCD/CNy/fz9dunQpAll9/vnnEbBo9+7d6cqVK6levXrhN5Ol3cOGDUssA8fq0CSBRxHAshkLZyxaCXI1a9asEO1XrlwZYivuJwiMhQ9fXAfgcgKrV5MEJCABCUhAAhKQQOUgoMBaOa6DpZCABCQgAQlIIA8InDhxIgSvbdu2JVwC4CuTZdxYtGKlijUiItjAgQNDaMVCUX+reXBhK7iIBEJj437JBPkGDRqEZeuePXsS/nzxz3rjxo2wmuae69evX9xzuBrQmrWCL6Cnl4AEJCABCUig6AkosBb9LSAACUhAAhKQgAQeRYCl2Vit3rlzJ0QurAvnzp2bsFq9efNm+FUlkNUrr7wSQYvGjx//qOz8TgKPJIA1K1asCPa8fvbZZ2nRokVp69atcc/hPgA/rS+//HJ67733UrNmzVL9+vXDt++zzz77yLz9UgISkIAEJCABCUigfAgosJYPV3OVgAQkIAEJSKBACNy6dSsdPXo0LV68OCK9r169OnxiYjWIsMqS7REjRqROnTqlli1bFkitrUZFE0Bobd++fZo2bVrq3r173H9YTRNUjfvxyy+/TFu2bIngVwRSY6tbt25FF9vzS0ACEpCABCQggaIkoMBalJfdSktAAhKQgAQk8GMErl+/Hr5WCTS0cePGsFrFipDl2gQa6tixY+rfv38aO3ZsuARgaTfLvE0SKAsCCPhsWKc2btw4AlsRBIv/Dx48GCI/YuuZM2fS+fPn0+3bt+OeZF9cDeg2oCyugnlIQAISkIAEJCCBxyOgwPp4nNxLAhKQgAQkIIEiI3D27NkQVj/66KOwXEXEwl0AgaywWp0wYUK4Bcgiurs8u8hukKdYXYRVhHwsWceMGZMIrrZs2bK4P/EHfOzYsQi6Nm7cuDRp0qSwpn7ppZeeYgk9lQQkIAEJSEACEihuAgqsxX39rb0EJCABCUhAAjkE8LN68eLFiNjOcuy1a9eG70vEVawCe/ToEUuysVxF7EL4ev55H6dyEPq2HAgQKI1gVk2aNEnVqlWLew73AdyD33zzTTp8+HD4Z8WKlffDhw9Pffv2TZ07d07Vq1fXsrocrolZSkACEpCABCQggVwCjghyafheAhKQgAQkIIGiJHD37t0IWMVya1wCfPrpp2n58uXh4xJhC2vALl26pIkTJ6bJkyen5s2bx5LtooRlpSuMAOIqIisbLiq4J2vWrBkW1lhcY826efPmdOLEicT/9+7di3sVtwIci1BrkoAEJCABCUhAAhIoewIKrGXP1BwlIAEJSEACEsgzApcuXYrI7PPmzUsLFiwIK8Bz585FLXr27BkuAVh6jdUgVqv6t8yzC1yAxW3QoEH4AEb8x2UFEwIrV64MtwHr168PX8GrVq1Ko0ePjiBsiLH4CTZJQAISkIAEJCABCZQ9AQXWsmdqjhKQgAQkIAEJ5AEBrFZxCXDgwIFwA4AYtWLFirBaRUBt2bJlWAkOGTIksQ0YMCAR2d0kgcpAAItUNu5JAlthpZoFwcIK++jRoxEI69q1a+n06dNx/3br1i116NAhXAxozVoZrqJlkIAEJCABCUigUAgosBbKlbQeEpCABCQgAQk8FgECVZEyf6tYrH711Vdp8eLF6caNG+mFF15ILVq0SAQM+uCDD1Lbtm1DuHqszN1JAk+ZAJMBuKxg69q1a+rXr1/6/e9/H5MFuLzg/saXMBMEU6dODfcCuBXIgrI988wzT7nEnk4CEpCABCQgAQkUHgEF1sK7ptZIAhKQgAQkIIFHELh582YsnyY40JIlS8Jidd++fYkAQZ06dYrgQERs7927d4irBLcySSAfCOA2gOBWWLWOGDEifLNu2rQprFl37NiRbt26FdbaiLB9+vRJvXr1iuBZ+VA3yygBCUhAAhKQgAQqMwEF1sp8dSybBCQgAQlIQAJlQuD+/fsR8OfixYshNm3YsCEtXLgw4XOVJdRYrbJ8eujQoeGzcsyYMalZs2Zlcm4zkcDTIsBkAFurVq1Su3btUr169cLv6rp16yLo1datW0NgPXToULgNIAgWrjCyIFjPP+/Q4GldK88jAQlIQAISkEBhEfApqrCup7WRgAQkIAEJSKAUAvhbZfk/S6UXLVqU5syZkxCZEFcJ/IPV3/vvvx8WfW3atAmRqpRs/EgCeUOACYI333wz9ejRI2HFOmvWrAiAhc9hfA3zumXLljR+/PjEhAIuBvQxnDeX14JKQAISkIAEJFDJCCiwVrILYnEkIAEJSEACEig7Aljo4YeSoD9Y8W3cuDFEJf5HTOrevXu4AsA/5bBhw8I/pSJT2fE3p4ojgFU2lqkEs6pdu3ZYs3bs2DGtXLky4RLj1KlTMeFw+fLltGfPnjRw4MDUs2fP8OOa+WetuNJ7ZglIQAISkIAEJJBfBBRY8+t6WVoJSEACEpCABH6EAEGs2PCpeuXKlbRt27ZwB/Dxxx+HqIQP1urVq4dLgIkTJ0YwK4ID1ahR47vAPz9yCr+WQN4QwEKbrXPnzrE1bdo0zZ07N6xaz549m44dO5aWL18ekxC40MCtAG4GqlWrlnAZoNiaN5fagkpAAhKQgAQkUIEEFFgrEL6nloAEJCABCUig7AngbxURdfv27Wn16tUhruJ78vjx4+FrFSu+CRMmpCFDhoRrgJdeeinEJKOpl/21MMfKRQDfrK+++moEc8NtwOLFi2MCApF1zZo1MQGBheu4ceOifbA/kxEmCUhAAhKQgAQkIIFHE1BgfTQfv5WABCQgAQlIII8I4FMVlwA7d+4MwQhfkwS0un79egSt6tSpU/hZxe9kly5dwu9kHlXPokrgZxHAVQAuMJhUwJK1bt260S5wnXHw4MEIgLV///509erVdO7cuXAb0LZt29SoUaNwNYC7AZMEJCABCUhAAhKQwA8JKLD+kImfSEACEpCABCSQpwRY8oyv1T/96U+xBPrIkSMJP6zt27dPuAN45ZVX0uDBg0Nkqlq1ap7W0mJL4OcRQGjF/zCTDPgfxkUAbQZLb9rQF198kdavX58mTZoUbYYgWLjQUGD9edw9WgISkIAEJCCBwiWgwFq419aaSUACEpCABIqGwOHDh8NSFZcACENYsGLN2rhx4zRo0KAQVRFW27VrF1Z7+pYsmlvDipYgkLnCoA2wtWnTJlxnEBALtwGbN28OtwGnT58OFwK4D8ASfPjw4REEq2XLlqlKlSolcvVfCUhAAhKQgAQkUNwEFFiL+/pbewlIQAISkEDeEiCIFb5WWcqM1epXX30VQhBLnLG2Qwjq1q1beuONN0JgRVxFUDJJQAL/R6BBgwYR2Ir2gVVr69ato/0wUYG4yuTFqlWrop0RNA6LV1wM4GrAiYr/4+g7CUhAAhKQgASKm4CjjOK+/tZeAhKQgAQkkLcEiHi+Z8+eWM6MALRjx450+fLliICOUIQ7gMmTJ4efVSKju7w5by+1BS9nAli1VqtWLREArkmTJuFSg6BX8+bNS3v37g1x9csvv4w2hsD68ssvh2U47UpXG+V8ccxeAhKQgAQkIIG8IKDAmheXyUJKQAISkIAEJJARuHDhQgirLGfGyg4hCCs7xNWuXbvGMuaBAwdGgB6EVgQgxdWMnq8S+CEBBFbayIsvvhgb/yOeIrauXbs2LMQJgrVt27aEFeulS5fSgQMHwuK1Q4cOqc23bgZMEpCABCQgAQlIoJgJKLAW89W37hKQgAQkIIE8IfDgwYN0//79dOfOnYTQgzXd7NmzQ1xFGKpevXosW54wYUJ67bXXUr9+/WIJc55Uz2JKoFIRQFhlGzJkSOrRo0fCPyvtbffu3WHFihuOZcuWfRc4rlGjRuEuAJcBiLNsJglIQAISkIAEJFBMBBRYi+lqW1cJSEACEpBAnhK4fv16On/+fJozZ05aunRp2rhxYzpy5EhYp7Zv3z6WK48fPz517tw5fEiy3NkkAQn8fAKdOnVKv/zlL1Pfvn3DYnzRokXR9k6dOhUuBI4fP562bt0aYiw+j/HPahCsn8/dHCQgAQlIQAISyC8CCqz5db0srQQkIAEJSKCoCNy6dSudOXMmrFZ37dqVPv/881iuTGCrhg0bhksA3AGMGjUqIbBiyaq4U1S3iJUtZwIIpliwtm3bNiYvCG6F2wDcBZw8eTJcczDZcfr06fgftxwEmKN9miQgAQlIQAISkECxEFBgLZYrbT0lIAEJSEACeUgAf48sRSbYzsKFC9PZs2fTzZs3IyDPiBEj0htvvBG+Vps1axbiqkuT8/AiW+RKT+DZZ59NderUCdcb+DlevHhxWJPPnz8/HTt2LMTWffv2pSVLloTbgClTpqTRo0dX+npZQAlIQAISkIAEJFBWBBRYy4qk+UhAAhKQgAQkUCYEsFq9du1a2rx5c1irrly5Mm3fvj2xFJnAO3369EmIq0Qz79mzZ0JcrVGjRpmc20wkIIEfEmDiAl/HbLjfwF0AAbEQWzds2BC+kI8ePRqBr5gMwep8y5YtIci2a9cu4aMVkdYkAQlIQAISkIAECpWAAmuhXlnrJQEJSEACEsgzAgSxQlhFnEGsIZAVYg0Bde7evRtBq1h+jCuA9957L0Qbfa3m2UW2uAVBoHXr1qlFixbhdxWRlQkO3AZgxcpkyIEDB+L/t956K9x39O7dOwTZqlWrhtCqpXlB3AZWQgISkIAEJCCBHAIKrDkwfCsBCUhAAhKQQMURwHJ13bp14RIAtwCINPh1JBE8h4jm48aNCwtWA+lU3HXyzBKAABapmTVr06ZNw6Ica3OCYNFuDx06lP785z+H4Dp27Ng0dOjQ1KVLlzgGS1iTBCQgAQlIQAISKCQCCqyFdDWtiwQkIAEJSCDPCGCZik9VguTs3LkzfDhiCcfy4ueffz41btw4xNX+/fuHgNOrV69wCZBn1bS4Eig4Alih0kYbNGgQrjsIMMd7gltt3bo17d69O9o1oiuW6fhqxZIVK/TmzZtH4KyCg2KFJCABCUhAAhIoWgIKrEV76a24BCQgAQlIoGIJPHjwIN25cyedP38+ESzno48+Cms3/id16tQpDRs2LP3mN78JX6tGJa/Y6+XZJfAwAliz4mu1ZcuW4RKAYFdff/11mjlzZliynjhxIi1dujTaNO0Zi1b8Kecm3Qbk0vC9BCQgAQlIQAL5RkCBNd+umOWVgAQkIAEJFAiBgwcPRiCrBQsWRKCcPXv2hK9VhBqCV40cOTINGjQodejQIfw3Fki1rYYECpYAS/+xZMVStU6dOqlt27bhixXXHxcuXEiHDx9Of/3rX8NCnWB1TKB07Ngx/CsrsBbsbWHFJCABCUhAAkVBQIG1KC6zlZSABCQgAQlUPAEsVu/du5cuX76cTp06FcLL8uXLw9Lt4sWLiQA4iC24AUBczXw2VnzJLYEEJPA4BLBkfeGFF1KrVq3CvQcuPnAHULdu3RBVCVi3fv36xOQKLgSuXr2aLl26FEIsFuq1a9d+nNO4jwQkIAEJSEACEqh0BBRYK90lsUASkIAEJCCBwiRw//798Le6adOmNGPGjBBYsVpFcEWIIRr5O++8E1arLDV+8cUXCxOEtZJAERCoUqVKat26dWrUqFFMmOAuYNasWWnx4sXhFmTjxo3p+PHjafXq1eEyYMyYMRHArgjQWEUJSEACEpCABAqQgAJrAV5UqyQBCUhAAhKobASyqOJYrxHEClHl6NGj4RKAoDeDBw9OI0aMSP369Qs/jjVr1owo5ZWtHpZHAhJ4PAJYs1arVi02/K3SvrFQxbfy5s2bv/O3/M0336QrV66E2IpVK+5BmjVrFi4GHu9M7iUBCUhAAhKQgAQqnoACa8VfA0sgAQlIQAISKDgCuANgu3v3brpx40bauXNnWrZsWfrkk09iafDt27fDV2ObNm3SuHHj0pQpU+K14EBYIQlIIAgwkdK+ffsIgkVQO6zYt2zZEsIqVq30EUzATJs2LazY2Re3Ic8//3zCP6s+Wr2RJCABCUhAAhKozAQUWCvz1bFsEpCABCQggTwmcOvWrbRv376EmLJq1aqwWjt27Fj4aGzatGkaP358Gj58eOrRo0dYreZxVS26BCTwGARwG1C/fv2EOwACYC1cuDDhh5kgWFixYtl67ty56C9GjRqVBgwYkAh6h9BKAC2TBCQgAQlIQAISqKwEFFgr65WxXBKQgAQkIIE8JXD9+vVE0Kpt27aFO4AFCxakHTt2pDNnzqQmTZqkLl26hK/F0aNHx+tLL70UomueVtdiS0ACj0kAkbR69eoRBIu+gIBYDRo0CJcA9BEEv6LfuHDhQmwEw+vbt2+4FSAIln6ZHxO0u0lAAhKQgAQk8NQJKLA+deSeUAISkIAEJFDYBM6fPx9Lf3/3u9+lJUuWRJRwXAXgf3HIkCHp9ddfT2+//XbCmi1b/lvYRKydBCRQkgDtv3///hHcburUqWn69Onps88+CzcBR44cCdcB+GfFivW9994L/8wKrCUp+r8EJCABCUhAApWFgAJrZbkSlkMCEpCABCSQxwTwqYrV2fbt2yOAFct+8a+ImwAsVPv06ZMGDhwY1mjdunVLNWrU0KdiHl9viy6Bn0sAn6pMsBDQDhcAY8eOTY0bN06bNm1KCKsEw7t8+XIIrljFI7TSh9CXZNav+mX9uVfB4yUgAQlIQAISKCsCCqxlRdJ8JCABCUhAAkVI4M6dO+nmzZvpxIkTEbyKYDVLly4NkRWL1RYtWoSF2iuvvBJBrPC9iqBikoAEJAAB3Aaw4YuZwFZMwDRr1izgHDp0KJ08eTLNnTs38f7o0aPp2rVrqWfPnql58+bfibOSlIAEJCABCUhAAhVNQIG1oq+A55eABCQgAQnkMQEszPbv35+++uqrtGjRohBBCFKDZRq+E8eNG5deffXVEFrr1Kmjr9U8vtYWXQLlTQCfrAS/qlWrVliqMmGDm5Fly5ZF34KVPBaugwcPTm+++aYB8sr7gpi/BCQgAQlIQAKPTUCB9bFRuaMEJCABCUhAAhDAnypWqwSkwQ3A6tWr05o1a9KuXbviO6xWsUIbMWJECCHdu3ePwDbSk4AEJPAoAliyYuGeBb/CTyvBrbBoJfjVvn370t69e6P/uXHjRkziYM1K4Lz69evbzzwKrt9JQAISkIAEJFCuBBRYyxWvmUtAAhKQgAQKh8CDBw/S/fv309WrV9Pp06fTwoUL0+zZs9OcOXPic6KD40Nx9OjR6R/+4R/CNUCjRo0KB4A1kYAEngoBhFW2QYMGhXg6fPjwCID16aefhsX8nj170s6dO9OqVasicN7777+fmMjB3zPW888+++xTKacnkYAEJCABCUhAAhkBBdaMhK8SkIAEJCABCTySAIFmiO69cuXKWLaLxSp+Ee/duxcWq0QEHzJkSCzb7dixY8IHq0kCEpDAzyHAxA3+VnEJgKXqggUL0rp16yKg3pkzZxIB9c6ePZvof3AdgGuSzIfrzzmvx0pAAhKQgAQkIIEnIaDA+iS03FcCEpCABCRQpAQINIOvVYQNfK0iarBEt0aNGuErceTIkWnUqFER6Rur1WrVqhUpKastAQmUJQEsWevWrRtb69ato8/BUpV+BncB9E0E1uMVy/pLly4ZBKssL4B5SUACEpCABCTwWAQUWB8LkztJQAISkIAEipsAwurMmTPDJcCpU6fS7du3U7169UJQnTZtWois7dq1C2EVP4omCUhAAmVNgAkdJnJ69OgRwfM++eSTNHfu3ITLgK1bt343CcSED/1S+/btQ4gt63KYnwQkIAEJSEACEihJQIG1JBH/l4AEJCABCUgg3blzJ128eDEsxNavXx9uATZu3JiOHz8efg47d+4cS3FZjjtgwIAQMV588UXJSUACEig3AvhWRWTFzyr9zRtvvJFatmyZ1q5dmzZv3px2796dDhw4EP0X7gPom/r165for5gQIniWSQISkIAEJCABCZQHAQXW8qBqnhKQgAQkIIE8JoCv1XPnzoVVGO4Apk+fHsIqn9eqVSusVSdMmJDGjBmTsFpF8HjmmWfyuMYWXQISyCcCCKVsI0aMiGB6WKo2bNgw/EHjJgChFcEVFwInTpxIN2/eTPiFJggfxyHQmiQgAQlIQAISkEBZEvDpoixpmpcEJCABCUigAAiw1HbhwoVp8eLFIVTgEoDUpk2bNGnSpMTyWyzD6tevr6/VArjeVkEC+UyASZ+BAwemFi1apNGjR4crkxUrNrVWTwAAQABJREFUVqSdO3fGxmQR1veIsa+++mr0Yw0aNMjnKlt2CUhAAhKQgAQqIQEF1kp4USySBCQgAQlI4GkSuH//frp27VpCSMXia9myZbFhAYarAMSIbt26xVJbBIyuXbuGmPE0y+i5JCABCZRGAItUgl4RCAt3AfiAxlJ1w4YNYYWPW5Pz58+nq1evpitXrkQ/1r1795QFzNKatTSqfiYBCUhAAhKQwJMSUGB9UmLuLwEJSEACEigwAg8ePEj4K1y+fHn6/e9/n3bs2BERuakmIgTWqh988EEEl6lZs2YIGAWGwOpIQAJ5TgChFJF1ypQpadCgQWn79u3pr3/9a/ryyy/T2bNnE1atK1euDCtXLPF/8YtfpGbNmukuIM+vu8WXgAQkIAEJVBYCCqyV5UpYDglIQAISkEAFEDh48GDatGlTWvStr9VvvvkmLL7wtYoF2LBhw9LgwYNDYMV/YfXq1RNBZvS3WgEXylNKQAKPJJDbL+E2gMBWv/zlL8P6fs2aNeEmYNu2beE2gD6O90we4V6AfXF5YpKABCQgAQlIQAI/lYAC608l53ESkIAEJCCBPCXAsn8EBlwCrFu3Ls2fPz8E1iNHjkRk7latWoUbgMwSrEuXLoqqeXqtLbYEipEAk0Fs9erVi+BWTZs2DQEVK1f6PcRVLFwPHz4c7gNwHdCpU6fUqFEjg2AV4w1jnSUgAQlIQAJlQECBtQwgmoUEJCABCUggnwggJuBrdcaMGbFslqBW+CesU6dO6tGjR3r99dfThAkTQmzgM5MEJCCBfCSAf1ZE05dffjkxUTRq1Kjo93AVgG9WrPYPHDgQE01jx45Nb731VvhzxQLWJAEJSEACEpCABJ6EgALrk9ByXwlIQAISkEAeEyCa9r59+9KWLVsiAMzq1asTVqs3b94MYbVPnz5p6NChqX///mH1VaVKFf2t5vH1tugSKHYCuDRBZGWjP8NHK69t27ZNq1ativ7w9OnT4T7gxo0biT6SIH4E9cNtwIsvvljsCK2/BCQgAQlIQAKPSUCB9TFBuZsEJCABCUggHwncv38/3bt3LyEe7NmzJ82ePTtcAhBhG1cBBK0ikBUWq1h54Xe1Ro0a+VhVyywBCUjgoQQQS9latmyZevfuHX6mFy9eHFasCKsEwcJlCn6nsWbFxQD7VqtWLSaaEGtNEpCABCQgAQlI4GEEFFgfRsbPJSABCUhAAgVA4Nq1awkLrXnz5oWAwJLYkydPRs3atGkTFqsIq1hr4Xu1atWqBVBrqyABCUjg4QReeuml9Oqrr4al6s6dO9OcOXPS5s2bw20AQf/Onj2bdu/enUaOHJlGjx4dbgZ0G/Bwnn4jAQlIQAISkEBKCqzeBRKQgAQkIIECJEAQK4RUfK0SzAWBFV+rfNawYcPwR0j0bASE4cOHp9q1a4fFVgGisEoSkIAEvkcAy/327duHFSsTTVipNm7cOFyn4DZl165d6cyZM+nChQvp4sWLqWfPnrE/wbJwN/Dcc899Lz//kYAEJCABCUhAAgqs3gMSkIAEJCCBAiRw/vz5xPJXXAIsWbIkRALcBSAkDBgwIL3yyitp0qRJqVmzZuGT8JlnnilAClZJAhKQwMMJILQisL7//vuJCSf6ys8//zwRBIvJqK+//jo+o6+cOHFibPXr11dgfThSv5GABCQgAQkULQEF1qK99FZcAhKQgAQKjcDdu3fT7du305o1a0IgWLp0aVhiYYHVoEGDCN6Cj1WssYiojcWWLgEK7S6wPhKQwOMSYGIJa1Q2XKTge5V+ET+sa9euTbgPOHr0aPSpWLNu27YtDRkyJPXr1y/201/145J2PwlIQAISkEDhE1BgLfxrbA0lIAEJSKDACRDECpcA+A08ceJEmjlzZlq4cGGIAwgICAYEdRkzZkx6/fXX9SdY4PeD1ZOABJ6cQN26dRNbu3btor/EHQD/r169OvrWU6dOJfyz4tOa/rZHjx6pRYsWqV69eokAWAbBenLmHiEBCUhAAhIoJAIKrIV0Na2LBCQgAQkUJYGbN2+GmIqo+sUXX4TFFdZWiKtYq+Jndfz48RHQheWt+BA0SUACEpDADwlg1Y9wOnXq1OgzsWalX92yZUvC9cqsWbPiPVas9KtTpkwJ1yv2qz9k6ScSkIAEJCCBYiKgwFpMV9u6SkACEpBAwRDAahWXAPv27Ytlq+vWrYtlrN98800s+2/SpEmIq4MGDUps3bp1S0TONklAAhKQwMMJYInK0n82fFYzKcVr27Zto689cOBABA68du1aunLlSli00r926NAhNW/e3Amsh6P1GwlIQAISkEBBE1BgLejLa+UkIAEJSKDQCDx48CARrAqr1cuXL0cgq+nTp0f0a6yr8CXYsmXL8CH43nvvhchKICuTBCQgAQk8GQHE1cxtABNVBA3EghU/10xuHTp0KM2dOze99dZbafLkyYmgWbgMoB9mBQGbSQISkIAEJCCB4iCgwFoc19laSkACEpBAgRDAcnX//v0hqM6fPz8sqvgfSyqWtfbp0yeWrRIRu8230bERB0wSkIAEJPDTCCCS4jYAC9Y33ngjtW/fPg0dOjQtWbIkRNZLly6lRYsWpWPHjiVWEOCShf63du3aBhH8acg9SgISkIAEJJCXBBRY8/KyWWgJSEACEigmAlissmGhevz48Qi6wuB+3rx56caNGzGI79KlS4iro0aNigF+p06digmRdZWABCRQLgQQWKtUqRKuArBoJWggk1cIqGvXrg13AWfOnEmrVq0K/9cIrlevXo19WE3A/gbAKpdLY6YSkIAEJCCBSkVAgbVSXQ4LIwEJSEACEvghAcRVhFQiWM+cOTMsp3bv3h2f4VcVcXXatGnhaxU/gNWrV/9hJn4iAQlIQAI/mwAia61atVLHjh3DkhUXAawm2LFjR9q7d2/4ZF2xYkUaPnx4mjBhQgTBUmD92djNQAISkIAEJFDpCSiwVvpLZAElIAEJSKCYCWAZdfDgwfD5RyCr9evXp8OHD4dFVffu3cPX6pAhQ1KvXr0iwApWVfr9K+Y7xrpLQALlSQD/qmwEviK4Fe4DsFSlb8Y368mTJ8ONC0EIWXVw5MiR1Ldv38SqAly2vPDCC+VZPPOWgAQkIAEJSKCCCCiwVhB4TysBCUhAAhIojQBBrNhu376drl+/HlZRWEPNmDEjrKPwtYqI2rlz5/Tyyy+nSZMmhRVVaXn5mQQkIAEJlB+BJk2ahAsA+mPE1ho1aoQfVlYYYM2KX9bNmzeH79Zbt24lVhg0aNAgVhlg1epkWPldG3OWgAQkIAEJPG0CCqxPm7jnk4AEJCABCTyCAOLqzZs3wwJq9erVacGCBQnL1VOnTsVRzZo1S2+++WbC1yoD+qZNmz4iN7+SgAQkIIHyJsCkF6sIGjVqlLZu3RqWrLNmzUqHDh2KFQcff/xxWLiOHj06DRs2LPXv3z8sYPHtapKABCQgAQlIoDAIKLAWxnW0FhKQgAQkUAAEsE49e/ZsBE3ZuHFjDNLxu4q42rx587B+6tGjR5o4cWIM5hs2bJief96f8gK49FZBAhLIUwJYobLsH9+sbHXq1En4xq5Zs2basGFDwpoVtwHnzp2LVQm4fSFYIS4DcC1AP47LAZMEJCABCUhAAvlNwFFZfl8/Sy8BCUhAAgVEAHEVP35/+MMfEtarp0+fDncBDNZHjBgR7gAImvLiiy/qx6+ArrtVkYAECocAqwxwHdCvX79YffD3v/89LVq0KNy9LFu2LCxcCYz1zjvvRBAs/LIqsBbO9bcmEpCABCRQvAQUWIv32ltzCUhAAhKoYAKZr1WE1C1btqSVK1cm/K3u2bMn3ARgDTVo0KDwsTp48OCwYEVcxWpV330VfPE8vQQkIIFSCNA341+1evXqqWvXrulXv/pVuHNZu3Zt9PFYsOKb9fPPP0/79u0Ln61Dhw5NrE4gYJZiaylQ/UgCEpCABCSQBwQUWPPgIllECUhAAhIoPAJ37twJEfXw4cNp27ZtYeGEuIrQWq9evVg62q5du7BaHTlyZGrbtm0sOS08EtZIAhKQQGERyNwGYM3KhouXVq1ahd9Vgl7hm3Xnzp3pyJEj8f7y5cvpxo0bsS8uA/gNMElAAhKQgAQkkF8EFFjz63pZWglIQAISKBACV69eTUePHk0zZsyIQFZYrV64cCEsmAYOHJgIhjJu3LgYcBN1GssmkwQkIAEJ5B8BXAaw+qBDhw7hW3vhwoVpyZIlIbDip5XfAj4bO3Zs9P0EMTRJQAISkIAEJJBfBBRY8+t6WVoJSEACEshzAtevX0+7du1KBK9at25d+Fzdv39/un37durSpUvq3bt3Gj58eOrTp0/q3Lmzkabz/HpbfAlIQAJMkOHahcBXvGarFPC5vXXr1phcw4r15s2b4XubIFg9e/YMq9caNWroNsBbSAISkIAEJJAHBBRY8+AiWUQJSEACEshvAvfv30/37t1Lt27dSidOnEjz5s1Ls2bNCstVBtu1atVKLVq0iIAnU6dOjejSWK2aJCABCUigMAjgW5WtU6dOCfcvTKbhd/ujjz5K27dvD7+sa9asSQcOHAhXMdOmTUtYsuJiAGHWVQyFcR9YCwlIQAISKFwCCqyFe22tmQQkIAEJVBIC+NY7depUIoI0y0KxXsUHH6lNmzapf//+afLkyRHkhP8RXE0SkIAEJFCYBAiC9dJLL6URI0bEbwBC6/Lly2NVw8WLF8Mv9//8z//E7wUrGtj69etXmDCslQQkIAEJSKBACCiwFsiFtBoSkIAEJFD5CGC1SrTovXv3hkXS0qVLw/8evlbxxzd48OA0YMCARATpMWPGJIKbaKVU+a6jJZKABCRQlgQQWLFKZWvZsmVMqtWvXz9EV4Ie4jZm9+7d4aP1/PnzCfcB+O0mWBa/E7Vr104E0jJJQAISkIAEJFB5CCiwVp5rYUkkIAEJSKDACNy9ezetWrUqffHFF2nOnDkJyyTcBVSrVi316tUrvffeeyGwtm/fPlWpUiUx6DZJQAISkEBxEcD/dqtWrdKwYcPCkvXTTz9NWfCrRYsWxSQdVq5vvvlm7MP+uBswSUACEpCABCRQeQgosFaea2FJJCABCUigAAggqmJptGXLlhgor169OoKYYIWE5VHHjh3DYhXLVXzwNW7cOATXAqi6VZCABCQggZ9AIAuAhb9VRFZ+KwiCuHbt2tj4/cC1DEESd+7cGa4FunfvHqIsk3OKrT8BuodIQAISkIAEypiAAmsZAzU7CUhAAhIoXgJXrlxJ586dC/+qWKz+/e9/TydPnkx37txJTZo0CatVfOm9/vrrCatVokObJCABCUhAAoikuI7p3LlzTMThjxvB9cGDB+EygN8W/HgfPHgwnTlzJuFqBv/dTNLVqVMnVa9eXYgSkIAEJCABCVQgAQXWCoTvqSUgAQlIoLAI4DuPQCUs6cTKCP+rDJqxWn3ttdfSkCFDQmRt0KCBvlYL69JbGwlIQAJlRgD/qu3atUt169aNlQ6shEBc5fflxIkTafbs2WnHjh2pZ8+eacKECREAC2HWJAEJSEACEpBAxRFQYK049p5ZAhKQgATynAD+VG/cuJGOHz+edu3a9Z1LAIRWrFYbNWoUFkaDBg2KJZ1YrWLJapKABCQgAQk8jAACKwGw8NeNuwDev/TSS/F+48aNMYFH8ESsWFk5cejQofDnjciKRStuAwyC9TC6fi4BCUhAAhIoHwIKrOXD1VwlIAEJSKDACdy7dy/dvn07lmpiXfThhx8mhNXMapWAJX369Em//e1vE24BGCAbxKrAbwqrJwEJSKAMCbACgqX/WKqyEqJr165p3rx56U9/+lP81uAuYN++fREQC7/f77//fuxfq1athF9XfbOW4cUwKwlIQAISkMCPEFBg/RFAfi0BCUhAAhIojQCD2vXr16dVq1alzZs3h0URAUiwUO3bt28EKhk5cmTq0KFDWCFpTVQaRT+TgAQkIIHHIYBVauvWrdOUKVNSt27d0tKlS9OKFSsSFq2nTp2K9/hp7dWrV6yc6NevX4iyj5O3+0hAAhKQgAQk8PMJKLD+fIbmIAEJSEACRUCAQCNYrLIcE5cACKv4WiXK8+nTp8NSCJ95WBiNGDEiDR48OPziFQEaqygBCUhAAuVMAGvU+vXrx4bAWq9evfDRSmCs/fv3x2oKBFfes5Li8uXL4cKmefPmCYvWF154oZxLaPYSkIAEJCCB4iagwFrc19/aS0ACEpDAExBAXN26dWv65JNP0po1a8Lv6s2bNyOCM8s333777fTyyy9H5OfatWs/Qc7uKgEJSEACEnh8AqyUwKJ16NChMdk3a9astGfPnoTbgJMnT8Z7Vle8++67qXv37ongiiYJSEACEpCABMqPgAJr+bE1ZwlIQAISKBACBBI5evRoBLHCcvWbb76JJZks++/du3f4Wh02bFi8YsVao0aN8H9XINW3GhKQgAQkUMkI4JuVwFdYphIMq2XLlt+5DNi9e3cIrQRhPH/+/He/UwitWLSaJCABCUhAAhIoewIKrGXP1BwlIAEJSKAACNy/fz/dvXs3Xb16NTFYxRXA9OnT07p16xIBrurWrZvatGmTxo8fn8aOHZvGjBnjEswCuO5WQQISkEC+EEBcRWRlI6gi4imWqkz+4SYAX+Hbt2+PSUF+p/hNI/gVQRerVq2a8OtqkoAEJCABCUigbAgosJYNR3ORgAQkIIECI3Dnzp1YZjl79uywXEVYZdklfvAQVkePHh3uALBYbdasmRarBXb9rY4EJCCBfCKAWEpgKwItDhkyJC1ZsiR+u3bt2hUrML766qt0+PDhNGDAgPj96ty5s9as+XSBLasEJCABCVR6Agqslf4SWUAJSEACEniaBK5fv56IxMygdNOmTeHbDr+r+LVDSGVQijsAAlkxUCV4CJZAJglIQAISkEBFEXj22WfDkhX/3/xW8duEZSuTgzt27Ahr1lu3bkVQxjNnzqRBgwaFIIvVK8dg2Yrlq0kCEpCABCQggZ9GQIH1p3HzKAlIQAISKFACly5dShs2bEgfffRRmjlzZrgIYFklqVevXun1119P77zzTgxcHYwW6E1gtSQgAQnkKQHcBjRs2DCNGzcuJgERWD/77LN04sSJdPHixRBY8SPOBOKrr76aJk+enDp16uQqjDy93hZbAhKQgAQqDwEF1spzLSyJBCQgAQlUEAHcAVy7di2tXLkyEcQKf6tYsGLtw0C1W7duEakZH3cECalTp07CWsgkAQlIQAISqEwEsok/Xgm4yO8XomvXrl3Txo0bQ1jdsmVL2rt3b/r0008T73EtMGrUqITLm0aNGlWm6lgWCUhAAhKQQN4QUGDNm0tlQSUgAQlIoKwJYJlKlOXjx4/H8skvv/wyfNaxnJIIzSyd7NGjRww8sfLBtx3BrUwSkIAEJCCByk4AYRV3AUwU9u7dOwRU3AYgvuKPFXGV37v9+/fHJCNuAxBkCZTFb6BBsCr7FbZ8EpCABCRQmQgosFamq2FZJCABCUjgqRK4efNmCKu4Avjb3/4WQaxYQonwypJJoi7jEoD3WK3io84kAQlIQAISyCcC/Ha9+OKLsRKjdevWqWfPnunrr79Oc+fODbcB27Zti0BYuA4YPnx4euWVV1Lbtm1TvXr18qmallUCEpCABCRQoQQcKVYofk8uAQlIQAJPm8C9e/cSLgEIWrV9+/bv3ALs3LkzLHbat28fbgAIYDVw4MB4X79+/addTM8nAQlIQAISKBMCmUsbJgqxSsU6tWbNmiGirl+//rsgWPgfv3z5cjp16lRYvLKCA7cBWLSaJCABCUhAAhJ4NAEF1kfz8VsJSEACEigQAg8ePAjLVHytnj9/PlwBYL0zf/78sOCpVq1aatWqVRoxYkSaNm1aWK02bty4QGpvNSQgAQlIQAIp/LK2adMmsY0ePTp98cUXYc3K6g22NWvWRKDH/v37p/Hjx4c1K64GqlatGqs4MrFWlhKQgAQkIAEJfJ+AAuv3efifBCQgAQkUKAGW/eNzjiWQc+bMiSBWWLFircOSSaxVJ0yYEME+WrRoEcspCxSF1ZKABCQgAQmEJevIkSNT06ZNE/5Xly9fHoGw8EuO24Bz586Ff9bBgweH6wB+G3Ub4I0jAQlIQAISKJ2AAmvpXPxUAhKQQEESYHk8guKFCxfS2bNn4xWLThJWKfhpI+owlirPPfdcvLKMkGBPvOZ7IrAHdWR5JKIqwT5IuAUgivKwYcPi83yvp+Wv/AQQ/K9fvx7t8NKlS+nKlSvRNm/fvh2F5x6tVatWBKchQA1bvib6E5Ya45qjc+fO4f+RYHH0MSYJSKDiCOAugJUbiKZYtNLP8HrkyJGEj3KeC2ir/G4+ynKV5woCZfFcwftHJX6H6ds4J+fDBQ//P83E5CrlJTVq1Ch17drV/uhpXgDPJQEJSKBACTzz7ZLJBwVaN6slAQlIQAIlCDBg2rVrV1hxYqmC77UDBw7EXoggiKgMthAeEXjwu4b/tTfeeCPvhUd+7u7fvx8iDyIW/2c/gQg9DDTZFH1K3DT+Wy4EmNg4evRoWIxt3bo14QOYtonFGInI34iRBJxhGzp0aLmU42lkSru7detWuOhgkoe+JmtriC0mCUigYgnQRtmY+GEihPe5v48IrLRZRNbShFb6r9/97ndp2bJl0Zc9qjb8xtK39e3bN1zy4IqAQJJPM/3xj39M//u//xunHDNmTPq3f/u3eOZ5mmXwXBKQgAQkUHgEtGAtvGtqjSQgAQmUSuDkyZNp9+7d4W+NwRDWc1htIp4SXZjBFJ8huGK9cuLEifi+e/fuBSE6IuQwsGPD36pJAhVBAP+/uKpYuXJltEcEDSY2WKY7bty479oalmBnzpyJ/Wi7LNnFjQUTIPmWEGSYsDFJQAKVk0AmnCKk/pTfRyZlX3311Wjn5LFjx47wdU5tx44dG+4HsFhlooVJpH379iUmlnjeOHToUPiCxYoUa9nyTEyuMrlFH0wZSR07dgxBuTzPa94SkIAEJFAcBBRYi+M6W0sJSEACIZpiXfLll1/GcmTEVQJYIOqwRI+Bz+nTpxOBn1iuvHbt2tSnT59EoCcCXJgkIIGfTgAhFVcATG6sWrUqLV68ONpb7969Y8k87RFrcdoiaePGjWENNmPGjBBZjx07FkIsAgQTIqVZkf300nlkeRDAEhCL3cxatzzOYZ4SqAwEWO3CswT9HBNCWOfT33Hvjxo1Kv3mN78JP69M4iKoTp8+PS1YsCCCTPK8QWrSpEm5C6xY5rKSh9fMIpdXkwQkIAEJSKAsCCiwlgVF85CABCSQBwT27t0bgxks47BK/e1vfxuvCDqZfzWWJWPRyhJegkER0KJDhw7hlzUPqmgRJVBpCVy9ejVEVSYw2Hr27JneeuutECXwcYwPwlzLMdodExtYli9cuDCOxTci4ipLarUIrbSX+ruCEZEdkQnhiOtmkkChE0BAxUofS1H8uRM8C3/nvNKf8axB3/b2229Hf7d9+/ZwjUL7IMhk27ZtyxURZeCZZ8iQISG00o/iH1qRtVyxm7kEJCCBoiGgwFo0l9qKSkACxU4A61Ss5xB6sILDco5lfZl1KhZxDH4YcGBJh5jTsmXLEAcQXE0SkMBPI5BF5J49e3YsS61du3aIpCydZVlsaeIbn9H+iN69Z8+esHbFqpXPEAQUWH/atXgaRyEw4U8XcZV+letd2jV+GmXxHBJ4mgQIoHnw2wBSCK1MGrH8niBSWX+VuelBZOU5A2GTiQj6SFbRlHeiPfI8gw9YysR7rG8pl0kCEpCABCTwcwkosP5cgh4vAQlIIE8IENkXv6pYxCGqIrI+TDhlYITwg8iK3zSTBCTw5ARoayxFZWLj66+/TnPmzAmxDV+Fr7zySvhUfVSuCABdunQJUZWl5vgMpE0+DSHiUeXyux8S4FqTuE4sj/773/8eog0iEmKOSQKFTID7nw2/0ayWwc8pVqs8R2RuT7L64w+dZwwmHniPKw36tKwNZfuV5yu+rPPRn3V5MjFvCUhAAhL4+QSe/flZmIMEJCABCeQDAQLpMNDBYgShddasWeELrbSyMzB67bXXFAZKg+NnEnhMAliLY326aNGiEFcREhBMJ0+eHKLpj2WDtRUia2ZljnuPU6dOhYj3Y8f6/dMlgO9JfEnizuHjjz8Odyxc7169eoWQ9HRL49kk8HQJ4NeUiQUE1hs3bsTEEs8b+JbGgjU3IaRi3c3GeyZxcZPysAnf3GN9LwEJSEACEqjMBLRgrcxXx7JJQAISKEMCuAPAkgorOALmfPXVV+ESAD9pDHAyEYdTYt06YMCA7y1rxSKF5X/4VsPahEEU1icMivDVijVKlrDaYz8EIVwTIO7WqVMnXBIgOjGwQnxgUMaG6Mvx+IBlwMW5EJI4H+85tmHDhvF9rp9KysA+lIeNvEjUhbxYlstnWUR2zs25qC8iMlY0WPZyDvbje/KkbFi3sBybpYMIXZSLfYiAvH///siHpYXZebK6+yqBjABLxZcvXx5BrYiajTUXy/sR3Wh3j5OwiKQ9kbhPuUf5P2s/3L8Ie3zHvZtZvdLmssS+fEdwGdoC9z/B62gDD0vkR9vifGycBxGR9kBbpD8hD8qXtcOsDJyPctB+eKXtHDlyJHwzsg9tK/NLmrs0l7zoN2hj2blp1+RHu6W9cs4s0FfJsrMvx8GDMtEPZPnj/gRu9Ed8ljEo2W9xHP0TnKgzS51J9DsZU95zPP0S5+CcTFpRx/nz56cVK1aE1TLWxge/XS5NnagzvijpS0wSKDQCtBvudQRW2ittDIG1W7duP7BgpZ2zL30i+9Iu6BMfx40G7THrG2hXtL+sTfGe32ieHWir/D7TX/CcQr9DP8bx7ENfQaKN0pY5FlcF9B3km+3D57Rz+krKTR7kz77kybH0hTwHsK9JAhKQgASKm4ACa3Fff2svAQkUEQF8nhHlF+HgwIED6dNPP40BAQIIAR9yBweIpogo2cAFTAgIRD9H/ECkQEzgewYwRAgmcFaWGDQxCFqyZEmaOXNm+GEjMA/nR5zElyTiUxZtmIEQwtO7774bIg4DLwTg1atXp927d4cVDMf+4he/iIFMdh7yoEzkQ5mwoKFMDOzIC2tBPsOqbN68eZEXggr1feedd2Lwt3Xr1igT+2FtSL2oI0HAPvjggxDCMg6wW7p0afqP//iPOMfIkSOjTJzHJIGSBPAr+Le//S1t2bIl7iP8qdIOEBIy4a/kMbn/Z+2IQX3JxGfc95s2bUoEiuHe5R5GyPv3f//38G+YHYMQwKTKf//3f6dF31rTcv/jpoAgWw9LiAi0LXyJ4uKA85AP4iK+Y3FxQB4IEYcPH459aauUgXZJOegXEDAQZv/yl7+kZcuWRRv8x3/8xzR16tTUqVOn7wnNiB60PdrYypUrY1/6HfoS+gj6mClTpgTD0pbdI4iuXbs2bdu2LcrEMfRvJERZ8qc/QnwZOnRo5EW/kyV4U+/169dHhHPqTN1ICCgs9ycxoQTDPn36RD9AQMDFixcHW+pPHTgXfc6GDRuiH3r99dfTP/3TP32vT43M/COBAiBA34BrAH6HSTxD8DuMeMrESG6indG+mXyinRDwb9KkSfHMkbtfae/JP+uXaGecg2cXRFTaO/0Fzw74fuX3mf6CyRX6RzaeP2ijCMEc88///M/RF3As5aevpv+i76Z/5fO+ffvG7zx1pD8k/3Xr1sX+9AM8m/C8wTOTSQISkIAEipuAAmtxX39rLwEJFBEBhI7Ro0eHFRsDHCxIGOSQsMzAggRRhISgmImK8cG3fxAoEBIZpCDWbN68OQYfJZf/sT/WLAxEEB4QPLDWw5IFSzasPSgLeTCgQcQYPnx4DHYY9DCIYRCF+ML/iLTkgyiCqJObEKoYSDEgYtCEcMuAiOWGWKJwDgQSLFAYCCEUIUZxjoEDB8bACbEEEQaLOkQrrPzYGGhRR0QdrFRIGRfqwUCNV8QmkwRyCWQWWtzb3OOIkFhAcc/RDhBXGdz/WKKdMshHYCTRBrjnuQ8ROhEXEBoRGhAdmBzg3Nz3tAPuexICA8dyXqwz2TcTC2OH//eH8/H9mjVroq3QDij3iBEj4jwIpIifbORP1G+EFNo0/QfH0gY5P22K72h/CK98RjvllXwRMOgH6Buy/ghhlP4isz5FIEGc5TjOTbkQZKh7JrDyPcfDmskS+ifqTZ/CayawMmlD/lwP2jwMS14D+oW5c+eGWEp+iEOIyNQPC1rOjSjE9SR/uFIWIp9nZWbiCqGXfgnhiP6D47nuJfvUXPa+l0A+E8AylLbFbzbtArGRtkd7p41gFZpZhiOC0r4QXn/961/Hbyjt42EWrLRvfpuZrGAFDm2cNkwfQBtHYKUfoJ+hDLR1nkuyNk7fxzMB52dymd92LFBpl/SJlJO+gGcAPqfP4hw8w9DO+Z3nmYR+mI02zf+Lvp2sok9i0hYh1yQBCUhAAhJQYPUekIAEJFAkBBBKEDMQYbJldgxIGDBkAwyERAZDiAklEwMmBB1EBoQKBBIGSAxScq1fOQ6hgn0YyCCmMLhi0EViYET+CCkIJwi948ePDxET6zGORSRCEGVg9eGHH4ZVCoMmhI3cxOAKYYfvGIAxmGOQhnhEOfmfQRUCCPsRaAiRlUEaohT1bt++fWzsj0iCoIWgy0ANKxWWOWcCK4IRAzfywrqWvBByTRLIJcBSdNoWYiMDcdoUYh2CG6Li4ybaDWIE7QQBgEkOhE2EA+5dNizI+R6fyrQbxD0mG7hvaR8k2idtFyt22gLCAW0tSwgIlJl8EBa+/PLLaDu071GjRqVBgwaFWEoZsAZncoGN8yEu0K9g4UmeWHeRH4lX2iCTGjBA5KB9IWrQlvmedkcbpT8gbyZUsPIdM2ZMCC/kT7nIB/ETcRjBl2PZaH/kydJ82ittdOLEiQmLUc6Z9U1Yn7Mf/QR9Edz4PjfR91AGzsM5sar71a9+FZM99JNMuFBOyky/R725Bkw80S/Alr6DY/E9OW3atBBoc8/hewkUIgHaDr+Z9He0b9oX7YznAPoW2in9AxsiLPsgqiJM8ntNeyqZsjZO26O/wfocgZQ2zcQmlqO0P87DMw37cH7e07Z5ZuFZgz6Evowy8ttPn0rfRV9MH5kJu/SZnJM+hucDhFb6EvbheYc60AfS/9Bn/eEPf4h+iTLRV5skIAEJSEACCqzeAxKQgASKiACCwMsvvxyDEgYYWIQgOnz++ecxKEG46d279/eW4ZfEw8CDAQriJwOPzAI0dz8GOAysMjcCiA9sDKpKSwgvCMCUBaGGMlA+hAwS4hDfZ2JJaXlknyFCMXAjL/JgeR+iC/8z0IMBgz8GXbmJQRr7MpBCYEKkQiDGSi9LHIO4+q//+q8RyAYxhvxMEsglwH3G/cMAHcsq7m/uGwb5T5JoY7Q12hwDe/LByrzkBAgWWgi4CBy0GcRIRAkEBBL3NkIE9zKiA+2d/bNEO0HwxCKLSQiECJa/v/nmm9HuyB9xgfLQhhEi2Eq2ZyYoEDI5BwLHnDlzYvICsZMyUw+EZwRU/NEibDABgzD6ySefhPCClT0TLizhR0xGuMjqTplgQNlhTF3pw3A/gLhCfrj+oB1T39zyUXbKx3FcB/KgXrmJ7zI/jHxOPbBqo+y8x50AEz/kiyiUCdjsC2/EGwRuzk2/SF9kkkAxEKBtMTmTCaz83pOw+OZzRFXaC1b8WIVnk0XsR3spLWX9EhbzWIYf/HYylkmibNKD43KtwukP2BBJmRSlr+C3P0v0YbRv2jnPEzxnlHwOyPbllbzp5xB/eX7gPf0v5coEVfoF8irZJ+fm43sJSEACEigeAgqsxXOtrakEJCCBGAQwSMAijQECgxAssrAAxV0AQgwb4giDoVyBIsOHgIIVCsIRAghCQ67IiGUH4hKDKvZjkMIAhMEOeZMQO8iHMiBgZJ+RH/vyinVoJooyCGNgkztYioP+3x9EDazGEE4oC//zyiAOwZZBH4MzhF/2YZBWUvzIyolFK+IL+SFYIQohPsEEgTfjgohLHR41QMsto++LhwAD/Exg5V7nHsfKMbOUelwSiBa0g2yiAtGOezNrR1k+3IMsSUeopM0gZiAmZIl7m417l/aPwEm7zRL3On5H8SPKMntEToREBAjECNoPVl0It0xO8B0WmpmogOCAsIF4kQmsiB9Yi1I2ltTSdqkP56VNUhbaFaLk7Nmz45W2j7hKgL1cdyXsi9UYbRNBg7rSd9A+WaK7YMGCEJMRQBFXqWNWtqyO8KCeJPq90gRWuFI2+hvqSr25jvQZXDs+hw2JSZ/cCR/6PfpRzsO+LF9mf5MECpkAfR2/kbQTXkn0EfQxtGH6IlZ7YP1N26V9YslP+8l9biiNEf0HzxE8m5AHv9tMVJEHbSy3H6Qc9D+0Qz7nGYJnDn63s8QkCP1T9uxCfiX7ZPprysw+1IN86d8QWemf6Dvos+gbSOyTrcrJzuOrBCQgAQkULwEF1uK99tZcAhIoYgIsZUVgYGOAwFI3Bg1Y3CEGMZhAQHyYwIqIw6CDARO+UnMtwbC4QxTCoowBDeJEJgxlyBFHEDwRXLBkYz/yxMcqAy8GVli6IVgg8FAmhJmHCawMdhChqAvCCkITy3sRXLB+wyoVn28MvhB8ELs4T8lE/ogiDOCwvIEH5UI0yQZqiEKIL9QLK7vS8imZr/8XF4FMYMU6k4E69xSi4I8JCiUpYUmK8Elb4D7O2lGusMAxCJC0Iwb/tAOERI4tmbiXmfRgX/qALHHMH//4x2gjCLEsvUXQ5JxZW/3zn/8c7YF6vPbaa+E6IBMx6QtowwgTCBiIrbQXyotgST9Bos3Qt5Av+2TC7hdffBFWbPQlLMnNnfwgH9o/lqlZ4njaMqIL7ZS+hDJhKcu+ucJndgznov4ci/hC31cyIQoTiAxG9BmwJz+EGoQY+gfYkMgnN9FHcL3hjhCDEETdTRIodAL81uN2hDZNX5S1Y35/6ROYtKGtZsvs6QdLe7YoyYk2i1U9k8CItwSdpP3xG17yePpczsV+tH/6XNptbqJvog+gT6acTMSUtJ7l9536ILRSD54tyJPnCfoA+jmeb3ALQKKNt/l2pUDWF+aez/cSkIAEJFB8BBRYi++aW2MJSEACIQ4wQMHKCpETgQJLMIQWrEUYnCBolhQqGMQwSGJjsIEQwmApV/DBqgxhApEUQYVBDAJsJkgwQOE8iBi85zuEGAYzDHbYjwEQ3zMgQmDFEgbBt6RAhUUqgx/Kw76IMQy+sK5D1EEsIiHGcE7251xY11DH3JSVjzLDhQEjAi+DKcQohCESQgquAygLS5YZqJkkkEuAdoLYwGCd97STktZUufuX9p57D2GBdkBbRchHGKWdlJxoYHBPe+QVYQBhgvZXMvEZbYA8EAERRhFvafNYmHEvM7lA22GiBKEB/6n0D7QLRAasz7jvqVPWvsiT/RAYSdSd9oGlPEJm1rb4LnuPtRlLf2lLiLiZVSx9TrZP7v65ggoiJhZzBL6ibNQHK9lsEib3ePaFB/Wh7dKOEVhz9+E8pExgpS9hwol+BTcgTNxQPiZmSjuOYxF36Pf4HtGFfqhk/8l+JgkUEgH6N9oL7Yu+h3ufSRj6C/qATOhkooLfaiYn6W/oZx42AUGetFX6H6zTOY6+jzbO8wT9Tm47ZMUKv++Ugb6HfOnLcvchT/oKfs/pl+njKCf9ZpY4T/ZsQv9EHjx7YOma9UvUkfOwLwlxlX4hm4DN8vJVAhKQgASKk8D/P/IszrpbawlIQAIFTYABBcIi4gKCZWmJQQ+WVgSMQFRkP6xNEFsYSOQmBiUMKrAAYUCDUIHog9jDgIdjETMQJbD8ZKDDIIsBSq4wSh4MdHhFWGFgwqAJIYZBDIIqx2YDHT7jewZDJa1NEHbIizLxHhGGcjEYY3BEXpQL8QPBijoxsMKyBpGltMSAC8GG/SgHohF1gifHM/CjfiQEnVzr3dLy87PiJMB9x0binuK+fBzBLbtnEQKyYCvcy7TTTBjNndAgf9pg1hYRTWn3TJxkifuWz8ibsiAYIpDyP0IiQi73Oe2LZfgcyz3OZAmWWgiziBss3588eXJMONCmM4GVdkL7op2QaNO0L9os9S4t0Ydgac4SYNpWJqCUrFtpx9KeKTfiLPnQnmmLiDklLcmoN30agiz9DXV/mNU514cyY8U6cuTIyAtf0viSzYJXkUdugiF9D5Zu8IIhkzjweZy65OblewnkIwHufYRT+jvaffa7SFtkooHf9mzygz6Cfg1R9mGJ/gARkzbOswhtjDaOmEnbzfqd7Hj6A/oCngNItD/6zCzRZukHMhGWctEvsV/uMwVl4nmCiV/qQhum7Dx/8GxBOXgOoq/knPS7Wd0ep2/PyuOrBCQgAQkULgEtWAv32lozCUigyAkwSEGkYfBPYIlcS41cNCwrZlkuQgoBbhg4YDWHUJObskEMgxBERaw6cq1AEWEYPDEAQWBhEMTgCoE1V4Qkb4QIysW5sUxlAIPgwyAGsYbBEIMrRBEGVFj/YXFXUjzhnFiaItYywGEAxkCOwRN5UYdsQMQAjMEU+TAoomylJTghNjNQ5HgsWRBVSYg5/M/nnC9XZCotLz+TAAS4b7n3SgoDpdFhYM89i7CAhSdtAVcUBHbJXdafeyzti3uaeze753OFQO5bBFDaG+2DiQ8SggGWp5yLdscxtKUVK1bE/U+eWG7SFig/93tp9zzHZVZdWI1xDtrQo0QHRFzKRH+AEEk/geBBu/2xhCi76Nulw5SViRB8tj5MyCX/uXPnfhd9nPZPfUpLXB/6Lix0ye+//uu/ok+cP39+XEMmpBB6EGmzlNWd83Dd+A5uJfuqbH9fJVBoBBBXER1pC/yO045pR1niWYFVJVik0k6w9CbY5qMSfRJuBZi0xUIUf9C5omnusZybSRAEVvopnidy90Uw5ZmDc5PoA3h2KdmXcTxlpB3TRzDpQ11yn03Ii0kX9s1EWvoF23vuFfG9BCQggeIloMBavNfemktAAkVAACs0BEGs3x4msCKCMEBAfHlUQuhB2EBs5BiEBISaXMtOBjB8j8jD5wxgGMjkDrawqMWShEEMQgzWswxisgEKAi7LlrGK4TwIsJSv5LJoyopIy+CKQQ/nYhDH/pmwgwDLgAmhlvNhAYgIyyDsYdZllJXzMXgiT0QcykyiXJSdZdKU/XHEoDjQP0VHAOGPe5/7ByET4ZR77sfuGSYyNm7cGBv3HRaViJwstyfP0hL5Zhvfcz7aH6+0EcRMzotQQB5ZW2MfJkW4pykfIuewYcPiFLRt2jhbrlhR2vlpW7QNRAfaDG0QIbK0NpsdjxhDf4H4S5tjIgVR43FEaMRcrFLpaxBySxNY6YPoS5iAwWKOetIvIK7kWrCyX2atjzgMRzhRD6xYuQa4CUDwQcShX4BJljIrd8oEbyaD4JgxzvbzVQKFRoB7HwEUgZX2RR9CG0ZQzZ1U5TeX38yZM2dGm88mVvh9pp2V1ub5XadP4Ry02dIEVr6jf0XwxMUR73EPUFJg5RmIiSSeE+iT+O2mby7ZF9NfUBfy5Zlk1KhRcW76BBLPJJSJ/eizqBP5PKqfK7Rrbn0kIAEJSODRBBRYH83HbyUgAQnkJYFMNEBgZTCAwMJnJQcUVA5RIdv4nwEPgyM+y00Mhhj08IpQgVCDaIuowGcIsLwy4ELU4XPyYQDCAIXBF+dHsEDwpDwMhAhageCZJYQKxAyEF0TfzJqM/BBlGIxlAmomsHJeRCCEFqxdssTnCDEM/hA8WP6MFSDlIC/yJK/cAR7/syGkYOXKoI2BFYMuxBoGae+//37UKzuPrxLIJcD9xL2NJSPtgvbHK/doaW2QYxE7ucewJP/888/Dgpt7FsFz+PDhIVrknuNh78kHwZCNNocIyb2O8MH9nDvZwb7cz9zjJMRc/Kzm9gd8zn65G/XL2gztmHMgPJAP7QbLL7ZHJcpGG4cN7Z++pLRJoNzzZuWiT0Okof3TRzCBlGuRnvUV9H8IK1jDcy7ETyZPuA6wzvol+gnqQZ3IB8GE77Ac5hy4B+B8s2fP/oHlXXbN6LcoXyawktfD+txHcfE7CeQLAdoGkyr8vtIm6dt4LkDA5DkiS4iVTITQ7miT/P4jZNJf0B9lfUm2P6+0J/alH2NCAxcluXnSxnnWoF3Sxul/6FPoS2iD7Eu/Rzumj2MfhFY+RwCmnyIPNtotZaA8lIs+p823ovDo0aO/d06+43mC/oL+lAldXmnn2bmcWMm9ir6XgAQkUHwE9MFafNfcGktAAkVAILPYwBJu06ZNMQjhs9IS4gi+EFlCxwAJERKhhYFJbmJQkQmsDFLYB/GUgQWDG/JHDOU7BjQMdrAUw2oUcYlj+YzBFZawCFAMdNg/91xY8JEf5+M7rMEQPciDJdMMlLLEOakjgyMsXBg05VqTMADkXAik1I2BGvswoMLqhQjk5FtawoqN/akLwiqMyJuBVybwlnacn0mA+5F2xL3PvXLwW79+WFE+rA1CDCtQ7kfE1UXfLn9HeMDn6dSpUyMgzONQ5VyIA1htc04EA9onnyME0J4elRAzssmY3P0QIfictom1K+0qSwgsiByUH+sxrL5yLTyz/Uq+wog6Zm2J49kof26iz0BIYWIGfpSF8/M/5aXvKNmH8D3WsZSLY+ir6IvoA2jXnAfRFD68p0+iP6CvRIjNEpZztHfaPfVHREJQzU2ZwAqHTGCl/vz/qOudm4fvJZCPBLi/M+GStoywye8r7ZG2liXaOJ/jqoh+kTbNcVmAu2y/3FeeS7L2zvEl8+T5ABGW32b6OARcnhOyvo7fbfJnv9xykg/9Af0A50Acpl+jX+EZhX6TSVjKWbJfoU/J3BsxGZS5ZaEvWrx4cZQltw6+l4AEJCCB4iOgwFp819waS0ACRUCAgQODBYQDxMp58+aFQMiABEEhSwgMWHfhYxAxAhEGK1Ai/GbCR7YvwkPmIiAbTDEoweIOQQPLEQQJRA8GLAyi+Iw8ec+Ah88z4QExBssWzpM7GEMMReygDixTZqCDuEE9EDgoP+dFGGXAQ50YCCHWsn+uWMvgivpxTj7PLFvwTYs1Ct+TV2kJyzis3agPljQIrNl5ckXc0o71s+ImwP2MuNq/f/+4f7lvGYBzL9IGMhERcY57mPt92bJlsQQdCykG77RDrElpi7Srx0nc/xxLW+U8tG9EPyzJuXdLs65C+EAQpMy0JcqCIEGbpZyUnb6BsvM97SErP2WibSHo0kbJC6uu3CX4Dys3fQjiJRZvnIt+JPN/mFmEIX7Q5yDGlJwIoV4k2jCiCP0FZaPO9H0ch4iL1VzJBHfyQ5ih7+J4An0tWbIk2jqCLnUkT/otygM/6ley7WcCK/vxPazZn+vK8SYJFCqBTLikLfFbT3tm0pV+hraXJX57aYe4DWHClPaBMMrvMG32xxLnYWKHPobff/o2+gp+l2lv9Hu5fRL5sR/tj76F/hBhlFfKiTBK/0EetH/aMP0cfRn9CZO/PJ/wbJJbD/o/ykz/QH0QajkHeWeTNT9WF7+XgAQkIIHCJvB/v36FXU9rJwEJSKCoCGSDEF4ZmPzud79Ln3zySVh75goViJaIq59++mkIHIiZWM2xJBkBJDchICCyMNhhAIVwyoADq1IEFY5lMII1B58jXmSJvBjocD4GIhyPeEkeJROiSmZJhkUKS40RTLDuY3+OQxSlXmzkixiKeFqyzAyWsGJBaGaQx+COhJUggyQsVUoek5WHPBGA+Z7BFwIM4hUuCB52THasr8VNgHbAfcsSUzYG8LNmzUoLFiyIAXomBmT35/Tp09N//ud/xn3JZ++++2764IMP0htvvFFqG3kY3ax90L4QI2iL3KuIAblCQXZ8Vk6W2CMcIq5+9dVX0a4oM4k2S7kpP4IG58idfEFIRIDlfPQD1Ls0UTM7Z/ZK28aijeXEnIvAWllQL0QX2ifBcOif8H3K/1idUWbETs5FPZkooWz0BYgliB1MhrBh1YafRMRb+h2EHfox+gyEXdiQL3VEAKdfoL+gvXONsHKl7ggziC64ayjZZ9HHkj+v9A30e5SFiSLKY5JAoRLIfSagTSI4MslQWkLY5BmByRzaIe2UdvcwgRXRNGvv7Es/gLDK8wvPIQTAQiDFjQdtE8GWdkr7o62Tsj6CPpVnCNo6bZ7fdfbFbyt9CH0fx9Nf0rchrnJsycRzBP1D1hfR11E28mESDHHZJAEJSEACxU3g++s/i5uFtZeABCRQMAQQEBgEDB48OAY9iAUMIrCSY2OAw0AEsRQLDAYUiKpY3LHUnsFNSUGGARSDIwYhbIu+XcaM1SgCBqIn32MVgrjCQINBDcIJYisDK86HdR7iBYMcxA8sSUomLFIQPxi8Ieh8/PHHIYxi0cdACoGD8jPY4XvqSvnx0YYAmpsYPJEXZUAAIlgNTLCyIy8GRbliUe6xDO4YiCEEUZ+JEydGOcjTJIFHEaAdcJ8g6L/99ttxfzIw53798MMPE4JqZoFJPtzDiP0EsuKVe5ljESWeJNGuOJ77e8SIEfGeNlGyLWd5YrWKyEnbJCFMfP311yEa0LbJi36DdkJ7pT3QPnOtxLH6QmAk0T88rsBK+yKIFAIL+dEvzJgxI23YsCHEDcoNI85NfXjN6oFV77Rp00J0oS/iOCxW23xrQUf5KMeYMWOivAgrXAv6NEQTJk5oz0zIsC/9DP0J7OBBXvSR7M/1QihFOIITFsWUI0two4+jnLyHBeIq/Rr8FFwyUr4WCgGeK2gzTKow+ZC52qBt4js9m4Bh0ib3/uf5gOB3TKjQ1nju4Hngr3/9a/R3tF2eI7I+j74BERRhlXb40UcfxQQQv/W0W15pZ/y+00fRxvmcNk7fQl60Vb7j3PxPuXnmQZxlhQH9LeWl70G0pS/h/PRh9HUlE30S+2f1/8tf/hLnyJ5N6FtMEpCABCRQ3AQUWIv7+lt7CUigQAkw2EAYxB8iAwysqxikMCjC6oMBBYIjYihCAwMNIpUjPjIY4fiSicESFlxYjCFKIiZk0cIZlGTHIMqQF4MQBjiIR4gVCCm8Z0DEUkFEJM5dMiF8YpWCJSsDMZbksT8iC/8jxiBqkB//YzWGyMuAibLnJhhwHOVlYIWAhFUZAyK+g83DEnlnIiviEsG4SrNqedjxfi4B7nXaIPcN7QcraCyeuLdJDNiZfEB04P7lvkRY5d7LxMQnocg5uN8RA7lnaUuPSgiKtGHaAW0T4QGhAaEEazEEDsqUlbFk+yJv8qB9cz6EicxK/FHn5TuOYTKHRPuEDeIkbGib9B20e+rA5E9uou+gH0DUpKz0NZlrA74jX/oQEn0c/QP9BYIMG2ILIiyJc3OdKD8CDf0jG4nrgEUeeXIdKUcmAPE9wiz9HteOxDWjXORZMtBP7OAfCeQ5AdoR/QITFzxH0HfQpugHuP/5jAkXJhhyE20JEZX+Bl/tPH/QTvh9px9EnES8pH0xYUEfRp7sR370R/hb53ecfoHnDPIh0UaZdOFY2jrPFbRx2jUpe3bJnil4DqCtcwwJK3z6EvoEfvPJP3ciJXb69g/15NmGfDiePod9KQf1o7wmCUhAAhIobgLPfGvV9P1oAsXNw9pLQAISKAgCiIgspWOwz6CHJfWIrAxWeGWQhDjA9wyQGCwgarCxP1vJhIjBwAJxk+8RSBiMIGDkHoO1KAMYzkG+DJzYl58bLEQoF98hnPB5Jsxm52MgxYAnOw+DJgZQnIvzMPiiPuTFYIuNgRGDKPLKLTvWNpSZVwQkBkHkw74Zm+y8JV9ZsszSPwaTDApfe+21yN9BVElS/v9jBLDmxmKLjTbIvU3iXuL+zjbaA/dodp//WL4lv8fakvzJBzGUvGh00ZoAAEAASURBVB6VaJO0JY6hfXOv01ZIWb9AXrRjyljavc9EC6IKbRqB5HEFVs5NHwQb+gTYwIXycB7EF85Ley1pGYYYQh9HfSkv56bt5x7DcST6B8QgzkedaPsIOrwn0R/Sl/DK+WFBX0HiOmR50s9lfVDWx1B+9kcY5pW+ibKyH9xK4xUZ+0cCeUqAtkE7ob3SbtlI3Pu0i+xZgkml0vof2iyrT7LnD/anrdAm6WP4Xaat0sY5B/vTNrP9yZNjsucKzp3lyfnJhzbOfrRdUslnl6xPoU2T6EuyZxP6H55NOEfJZxOeSzgXZcv6bp5NsmcgGJgkIAEJSKC4CSiwFvf1t/YSkIAEJFCCQDZ4XLlyZVjXYJWHhVpmpVZid/+VgAQkIAEJSEACEpCABCQggSInoIuAIr8BrL4EJCCBYiaApQxbruUJ1iy4U8AKFysVlgC2+XZpo0kCEpCABCQgAQlIQAISkIAEJFAaAQXW0qj4mQQkIAEJFAUBlvrhkoCliSxpxHqVJYAE2sHtAL5pMz9uRQHESkpAAhKQgAQkIAEJSEACEpDAExNQYH1iZB4gAQlIQAKFQAALVXzBESQHH2oE5cCXI59j1UqQC3yvmiQgAQlIQAISkIAEJCABCUhAAo8ioMD6KDp+JwEJSEACBUuACMBz585Nn3/+eVivDh48OAJjECl4xIgREYW4YCtvxSQgAQlIQAISkIAEJCABCUigzAgosJYZSjOSgAQkUNgEiB7Mkvo9e/bEtnv37lhSj/9SIu4iTPbp0yeW1mfReSszEaINs/yfLYuU3rFjx8TWrFmzcBtQmctv2SSQLwSIBr5ixYroN3DJMWTIkNSrV69EBO4s0ne+1MVySkAClZ8Azyk7d+5MBw8eTJcuXYrVKVWqVInJ1Ox3Xt/qlf86WkIJSEAC+UZAgTXfrpjllYAEJPCUCbBcHnGVQQr+SRcvXpzmzZuXFi1aFIIr4mqLFi3S8OHDQ6xEuMwHgRUfqz179gyXAIg8nTp1CuGnVatWT5mwp5NAYRO4du1aWIvPmjUrbd++Pf3Lv/xL9BV169ZVYC3sS2/tJFAhBI4dO5ZWrlwZEzv79u1LZ8+ejaCVPKuMHTs2vfzyy7FiheeVqlWrpueff/57wS4rpNCeVAISkIAE8p6AAmveX0IrIAEJSKB8CeCX9OTJk2n16tXp66+/DoHkwIED6fr166lz584hSiKuIlZiEYI1aD6kBg0ahBVd8+bN07PPPpsYaCH4mCQgAQlIQAISyF8CPI+wOmXYsGEhtPLsgs91tpkzZ4Y1PVb1EydOjGcXJlyfe+65/K2wJZeABCQggUpBQIG1UlwGCyEBCUigchLACgTrDwJBrVq1Ki1btixdvXo1rM5Y5suGv9LevXtHUCgsQXAZkA+pWrVqYcHCIMwkAQlIQAISkEBhEGAClZU0uPvhlQnU9evXxwTx4cOH04YNG9KJEyfieebIkSMx2dqyZcvEcVizMulqkoAEJCABCTwpAQXWJyXm/hKQgASKiMCWLVvSjBkzEkt7sfy4f/9+WIX06NEj/frXv04DBw4Mn6WIqvkirBbR5bOqEpCABCQggaIkgEVqvXr1wn0Rk8G4N5o9e3aaPn16On78eNqxY0fatWtXWr58eZo0aVKaPHlyPNPUqFFDgbUo7xgrLQEJSODnE1Bg/fkMzUECEpBAwRBAQD137lwigBUWq2vWrAnrVfyXMVDBBQD+y4YOHZoQWV1WVzCX3opIQAISkIAECoZANumbTQB369YtIZ527do1nm94xiEIFit15syZk7Bk5Zln5MiRqUOHDqlJkyYFw8KKSEACEpDA0yGgwPp0OHsWCUhAApWewOXLlyMQBFYdBIfAZ9mhQ4cSAWqI9s2gZMCAAWnKlCnhEoAl9vosq/SX1QJKQAISkIAEipoAIivuAho1ahTuAHiP24Bvvvkm4VN+z549sUqHV555WJ3TvXv3ePZhP1wGZIJtUYO08hKQgAQk8EgCCqyPxOOXEpCABIqHAAMLrDjmz58ffsouXboUA4qmTZumN998M40ePToEVvyZIa7qo6x47g1rKgEJSEACEsh3AkwKV69ePVbhYKW6bdu2tGjRovTZZ5/F6h1cBhDUkwBYiKxTp04NQZZjFFjz/epbfglIQALlT0CBtfwZewYJSEAClZYAVqsMJhhksFyOQcXOnTvTjRs3ImgVkXj79+8fwaw6d+6cWrRoUWnrYsEkIAEJSEACEpDAwwggkiKyMlFcs2bNEFtxG1C/fv1wh8QKHixaCezJJDNBPXkGwiVS69atwy3Sw/L2cwlIQAISkIACq/eABCQggSIkcPfu3XTnzp1YErdu3br0l7/8JQYURNWtWrVqIpou7gDeeeedNGHChPBbRmRdkwQkIAEJSEACEsh3AjzTNG/ePNwAYK2KJevcuXPTrVu3IgjWpk2bIgjW6tWr0xtvvJHGjBkTz0I8I3Gsq3jy/Q6w/BKQgATKnoCj5bJnao4SkIAEKj0BhFR8jy1dujQhsO7duzdduXIl1a1bN40bNy6i7iKwtm3bNiw8HEhU+ktqASUgAQlIQAISeEICiKVYsSKyErizT58+admyZfF8RNDP/fv3p48++ihcJ7EPz0jt2rULK9gnPJW7S0ACEpBAgRNQYC3wC2z1JCABCWQEbt++na5fv54OHz6cNmzYENYaiKwEssK/GC4A2LBYZRCBf7IXXnghO9xXCUhAAhKQgAQkUFAEmEDmWQdr1oYNG4YrpAYNGsSE85YtW0JgRWS9ePFiYnIaF0q9e/cOkZVgWfXq1SsoHlZGAhKQgAR+OgEF1p/OziMlIAEJ5BUBxNWDBw+GJcbChQvT1q1bYykcEXK7du2aXn/99fT222/HYAFrDl0C5NXltbASkIAEJCABCfwMAgitiKaTJ09OI0eOjMCfBP/EdcCFCxfSmjVr0u7du8Mn64gRI8J1gALrzwDuoRKQgAQKjIACa4FdUKsjAQlIoCQBBgUEbsCf2Pr168N69ciRI+n+/fupe/fusRwO32J9+/ZNTZo0SVWqVIkgECXz8X8JSEACEpCABCRQqASyIFis6uFZaNCgQWHJyiQ0vlg3btyYzp8/H4FBCYJ19uzZ2AeXSgizBM8ySUACEpBA8RJQYC3ea2/NJSCBAiZw7969RCArlrTt2bMnzZs3Ly1evDj8riKs1qpVK3Xp0iWNGjUqAjdgqcGSOJMEJCABCUhAAhIodgKs4unYsWO4DMgmoBFecRtw/PjxeJ46duxY+LC/du1auA1o3759PF9hCYvrAQRbkwQkIAEJFA8BBdbiudbWVAISKCIC+AjDcvWrr75Ki76NjIvVxalTp9KdO3dS69at05AhQ8IlAIMB/I7Vrl27iOhYVQlIQAISkIAEJPDjBBBLmYB+5ZVXYtUPK4Fws8TzFVasuA3ARyvWrqwGGjZsWDxnIcaaJCABCUiguAgosBbX9ba2EpBAgRMgkNWZM2fCJQABrIiEi7UFy9gYIBAdd/DgwbENHTo0lrM5CCjwm8LqSUACEpCABCTwkwg899xz4TapadOm4aOeFUBMStevXz/t3LkzxNUDBw7EBPbly5fDupVnrZ49e8ZzF/ubJCABCUigOAgosBbHdbaWEpBAARN48OBB1I6l/1euXIngVV988UX68MMPI9otXxLIigf+SZMmxda2bdsCJmLVJCABCUhAAhKQQNkSqFatWrhX4hkKi9bp06enmTNnxrPXoUOH0uHDh8NX6/Dhw9Ovf/3rEFmZxM7cBegyoGyvh7lJQAISqGwEFFgr2xWxPBKQgASekAD+Vm/dupXWrl0bFqvr1q0Lq4rr169H0Cp8rbJsDWuKTp06pZdeeukJz+DuEpCABCQgAQlIQAIQwD9rzZo10+jRoxOWrf369UvLly9PPH/hoolXgmH17t07DRw4MOHnvnHjxqlq1aoClIAEJCCBAiagwFrAF9eqSUAChU0Ai1WWo+FbleVp8+fPD3+ru3fvTnxHRNvswX7ChAn6Wi3s28HaSUACEpCABCTwFAhkbgMIgtWkSZPwuVq3bt2Eteq+ffvCTQCBRY8cOZIIhMUkePfu3VPLli3DzYCumZ7CRfIUEpCABCqAgAJrBUD3lBKQgATKgsDdu3cjei3BFj755JN4kL948WLCDyuWqmPHjk1vvvlmiKxYWmBxYZKABCQgAQlIQAISKBsCNWrUSAit+LnHxz2T3QsWLIhAWLgMOHfuXAQaxW3Aq6++GkFGW7VqVTYnNxcJSEACEqhUBBxtV6rLYWEkIAEJPJoA7gDu3LkTFhIbNmyIZWhEtN21a1cEYcA6giVpLFfDerVbt24RiOHRufqtBCQgAQlIQAISkMCTEsisWRs2bJgQW/GzijuADh06pE2bNsVE+PHjx9OKFSvSpUuX0p49e+IZDb/4mdXrk57T/SUgAQlIoHISUGCtnNfFUklAAhL4AQEsU69evZrOnj2blixZkmbMmJE2b94c1hEEXmjTpk3q27dv+sUvfhEiK2KrSQISkIAEJCABCUigfAlUqVIl1alTJw0YMCDEVSa5P/vsszR37twQVXEVgDunnTt3xiQ5AUrbt28fPlwRZl1lVL7Xx9wlIAEJPA0CCqxPg7LnkIAEJFAGBIhQi7Xq119/nbZu3Zr4/8qVKxG0atCgQQk/qyxBI+BC7dq1y+CMZiEBCUhAAhKQgAQk8CQEcMvUunXrNG3atLBWxR/rypUr4xnuxIkTad68efEch0uBESNGpFGjRsWz25Ocw30lIAEJSKDyEVBgrXzXxBJJQAISCAJYN+Bn9cKFCxEwgeVly5YtS0uXLo1lZlitssSsZ8+eadiwYeH7C5cAJglIQAISkIAEJCCBiiGANStbrVq1UqNGjcJ1AC4ECIi1Y8eOdPTo0fTNN99EoFKe8Vid1KNHj4RvVqxgsWg1SUACEpBA/hFQYM2/a2aJJSCBIiGAwHrz5s3w3zV79uw0c+bMhN/V+/fvJx7UCarwq1/9Ko0ZMybcA1StWrVIyFhNCUhAAhKQgAQkUPkJsKIIK1UEVFYa/fGP/x979x1ta3Hf93+QAIkOohfBpVfRu+gdBKbJtrAlIifOcklz1nKSlaxkrfyTrLgkjuM4TmzHkvWTLRQVJIREE3BBdC69iCaKKKJX0QX87musQY+29zl773PP2ad9Zq3n7nP3fsrMe2a+bcrz/9VtAwRV77vvvvqCUrbd4YcfXs4888yy8847J8A696s1OQyBEAiBvgQSYO2LJV+GQAiEwOwSeOaZZ+oWAGar2hbg7rvvrv+37GzXXXctBx10UDn00EPLjjvuWDbbbLOy6qqrlg984AOzm+k8PQRCIARCIARCIARC4H0CXnrlRVhrrrlmsTe+ffIFUQVVr7/++jqj1bYBS5cuLT/84Q+rbWfbJwFZs1nZd0khEAIhEALzg0ACrPOjnpLLEAiBBU7AbFUzU81Y9ZbZu+66q9x4443l29/+drn33nvr8rH111+/2ALgqKOOKkcffXQ1whc4lhQvBEIgBEIgBEIgBOY9AauMHAceeGB9CZZVSAKothJ44okn6rYBtg94/vnn68tLX3311foSLNsKGFxPoHXeN4EUIARCYBEQSIB1EVRyihgCITA/CLz55pt1qZhZq16AYM/VF198sWZecPXkk08uxxxzTNljjz3yMoT5UaXJZQiEQAiEQAiEQAj8HAGBVXvoW4FkNZJtoOyxv2zZsnLHHXfUgOt1111Xf2P3mc264YYb/tw98p8QCIEQCIG5RyAB1rlXJ8lRCITAIiPwyiuvlGeffbbceuutdcmYLQHMYH3qqafK1ltvXXbaaadqXJv1sPvuu5eNN964eMFVUgiEQAiEQAiEQAiEwPwi0F6Ctdpqq70/i3XTTTctW221Vd0SyozWl19+ua5qsmXUvvvuWwOyVjGx/1ZeOS78/Krx5DYEQmCxEIh0Xiw1nXKGQAjMOQK2BHjrrbeKvbfMWPjSl75UZzBYHmY/VbNWDz744HLcccfVmaveRpvA6pyrxmQoBEIgBEIgBEIgBEYmINBqZqptnwRP99tvv3LOOeeUyy+/vDz55JN10P32228vjuOPP76wA53v07XZe39k5LkgBEIgBGaUQAKsM4o3Nw+BEAiBiQmYnWAbAFsCXHHFFfXlBq+99lpZd911q5HtjbJmLWy33XbFW2gzY2FilvklBEIgBEIgBEIgBOYrAbafIOvZZ59d9tprr3LttdfWLQPuueeecv/99xf24Z133lntQ4HYffbZp9qL87W8yXcIhEAILEQCCbAuxFpNmUIgBOYsgbfffrsayY8//nh9eZW9VhnR3iZr6f8OO+xQvPjgsMMOq4c3zgquJoVACIRACIRACIRACCxMArYLsErJ/qybb7552WCDDepKJoHX9hKsBx98sM5sffrpp4uXYNlCyrlenpVB+IXZLlKqEAiB+UUgAdb5VV/JbQiEwDwnYAbCo48+Ws4999zyne98p/zgBz94/0VW9lf1MgNbAmyxxRbVyP7gBz84z0uc7IdACIRACIRACIRACAxDQLCUDWjQnV1oNdN5551XvPTq3nvvrbNaH3jggbptwEknnVTOPPPMGohdc801h7l9zgmBEAiBEJhBAgmwziDc3DoEQiAEGgEzV+2hddNNN5Ubbrih7rlqJsIbb7xRZ6xa6mXWqk8vtlp99dXr/lrt+nyGQAiEQAiEQAiEQAgsXAIrrbRSLZzZqI7NNtuszmpdY401yo477liuvvrq+hIs+7N+//vfL2+++Wa57777yt57711fgiUgm736F277SMlCIATmPoEEWOd+HSWHIRAC85SAl1gJrL7yyivFcq5LL7202BLgyiuvrCUy22DJkiXlkEMOKaeddlrZbbfd6lKveVrcZDsEQiAEQiAEQiAEQmCaCLATHVtuuWXZZpttykYbbVT3XV22bFm1K2+88cY6s9Uerc8++2zdKmDTTTet15gJm1VQ01QRuU0IhEAIDEkgAdYhQeW0EAiBEBiVgJkFZhl4idVFF11U7rrrrvoiq7feeqsaymYcnHzyyXUJmL1Ws7xrVMI5PwRCIARCIARCIAQWPgGB0+OPP77OZGVPXnLJJXX/ftsF2MffQP4dd9xRV0MdfPDBdTVU9vBf+O0iJQyBEJhbBBJgnVv1kdyEQAgsEAIPPfRQ3Svr1ltvrVsCmGXw0ksvFcu8bAUguLr//vuXAw88sDCazTRICoEQCIEQCIEQCIEQCIFeAl6C5VhrrbXKhhtuWD/NbLX1lG0CHnnkkfLCCy/Uff29SPWAAw6oL8EygL/KKqtkNmsv0Pw/BEIgBGaAQAKsMwA1twyBEFh8BN57771aaJ+2BmDwfvOb3ywXXnhhXbb1gQ98oM5Q3XnnncvZZ59dA6ve/poUAiEQAiEQAiEQAiEQAsMQsNppu+22q4fB+uuvv7588YtfrJ9WTT3xxBPl2muvrS9RPeGEE4pDUNberPZ4bfu8DvOsnBMCIRACITAagQRYR+OVs0MgBEKgLwFBVXuteunAVVddVa655ppy22231e+8CXb77bcvRxxxRNlvv/2KIKt9tJJCIARCIARCIARCIARCYCoE2JcHHXRQ3Zf18MMPr4P7Vk798Ic/rHuz2jbAi1VtGbDXXnvVbQPMgk0KgRAIgRCYGQIJsM4M19w1BEJgERF48cUX695XDz74YJ01YNaqpVqvvvpqsTTLy6ss1bJ3llmrtgMwozUpBEIgBEIgBEIgBEIgBKZCwB6rZqduttlmZYcddihbbLFFWWeddWpw1WxWL8O6++67y3PPPVeef/75su+++5atttqqbjHgBVixRadCPdeEQAiEwMQEEmCdmE1+CYEQCIGhCHjBwGWXXVYPM1jNGGC4MmLPOOOMuueqmQP2XxVczfKsobDmpBAIgRAIgRAIgRAIgQEE2JZsTqujrJjafffdy9e+9rUaXH322WfLBRdcUFdVCbB6uaptA1ZfffUEWAdwzc8hEAIhMCqBBFhHJZbzQyAEFj0B+6y+9tprdZ8r2wBcd911dQnW/fffX954442yySabvP8SK8uy7JW1wQYbLHpuARACIRACIRACIRACITB9BNq+qoKsDgFWwdP11luv3HzzzXUWK/vUtgE/+clPyuuvv16suNpll13KjjvuWM83KSApBEIgBEJgxQkkwLriDHOHEAiBRUTgrbfeqsapN7R6kdVXvvKVcvvtt5fHHnusvkDAlgBmDrSZqx/5yEeqwbuIEKWoIRACIRACIRACIRACs0Bg/fXXLw626B577FHs03rJJZdUW9ULsH70ox+Vq6++uhx55JHlmGOOqS9gta2Al2Bl24BZqLA8MgRCYEERSIB1QVVnChMCITDTBMwA8MIAL7LyIoGHHnqozmZlzHqB1aGHHlpfZmUfLLMHVllllZnOUu4fAiEQAiEQAiEQAiEQAj9HYOutty6f/OQn68tVzWa98sori22tbBtwxRVXFJMF7rzzzmq77r333jUYK9CaFAIhEAIhMDUCCbBOjVuuCoEQWEQELKnywipLqq6//vpy+eWX1+CqWQDexrrzzjuXXXfdtb7Iap999qkzBlZeOeJ1ETWRFDUEQiAEQiAEQiAE5hQBL8FyrLvuumXTTTetL7cySeCWW26pM1kFV5955pkacH300Uer/brtttvWl2XNqYIkMyEQAiEwTwgkAjBPKirZDIEQmD0CtgWwrOqrX/1qufTSS+t+Vu+++259UytD1HYAp556ah359yKrvJV19uoqTw6BEAiBEAiBEAiBEPgZAe8BsGWVWaq2DfDSq4suuqjcd9995Z577in2aPWy1hNPPLHas1ZhJYVACIRACIxOIAHW0ZnlihAIgUVC4MUXXyyPPPJIufbaa+uLrO64444aaPUSgd12262YrXrIIYfUv73Yqu1ftUjwpJghEAIhEAIhEAIhEAJznIAXYdlf1bHDDjuUVVddtb6A1UxWM1offvjh8vzzz5fvfe97dTbrsmXLykEHHVRfhMW+zXZXc7yCk70QCIE5QyAB1jlTFclICITAXCBgZqotAexPZZ+qG2+8sVx44YU1wOo3MwC8OOCoo44qhx12WDn44IPLWmutNReynjyEQAiEQAiEQAiEQAiEwIQEBEzNaN1uu+2KVVjsWkHWu+++u3jPgH1ZBV5tHfDSSy+Vvfbaq24tsOaaa9YVWlmlNSHa/BACIRACJQHWNIIQCIEQ6BB4++23i5mr559/ft1rVYCVkfnOO+/ULQDMWD3ttNPKjjvuWPeoWn311TtX588QCIEQCIEQCIEQCIEQmLsEzGRdZ5116pYBgqxWZHl563nnnVfsxeodA1/72tfqC7COPfbYOqFAoNVKrQRY5269JmchEAKzTyAB1tmvg+QgBEJgDhB4/fXXy1NPPVW+//3vF29atS3AXXfdVQ1NLwYQUN1///3rwRA14m+/1aQQCIEQCIEQCIEQCIEQmC8EbBlg2b/DS7D83+d6661XbIdlb9aHHnqoBlhNMHjsscfq33vuuWfZZptt6oQD1ySFQAiEQAj8PIEEWH+eR/4XAiGwiAi89957xeElVmap3nTTTeVb3/pWOffcc8sbb7xRSay22mr1hQAnnHBC3fj/ox/96CIilKKGQAiEQAiEQAiEQAgsZAJeamXrABMITDK45JJL6vZYJh1cf/319bvNNtusnHXWWXWLLBMMvI9g5ZVXrsHZBFsXcutI2UIgBEYhkADrKLRybgiEwIIiYK/VN998s1x55ZV1Y3/bAfzgBz+owVUzVHfZZZdyxBFHFCP2O+20U1l//fUXVPlTmBAIgRAIgRAIgRAIgRCw9N+kArbvuuuuW7beeuu6N+s111xTtwx48skn67YB9957b9l3333rOwh23nnnutWAQGtSCIRACIRAyR6saQQhEAKLj4DlTq+88kp54okn6ousLr744hpgvf/+++ubVS1/2mOPParxeOSRR9a9Vi2bSgqBEAiBEAiBEAiBEAiBhUZAgNWx0UYb1QkF9mi1RdaGG25YZ7B6CZZtA5577rn6Iqznn3++bq21/fbbF7NbvTgrKQRCIAQWO4EMNy32FpDyh8AiJOBFVo888ki59NJLyznnnFP3WWUwmtG6ww47lOOOO66+yOpjH/tY3dDfywCSQiAEQiAEQiAEQiAEQmChExBotSWWbQO83NWWAbbQ8ul9BV4G+8ADD9TJCWzmY445JgHWhd4oUr4QCIGhCCTAOhSmnBQCITDfCZi1aq9Vo+9eXnX11VfXPVcZiDb59xIre0859t5772JEfq211prvxU7+QyAEQiAEQiAEQiAEQmBoAvZUtezfYdsAtrF9V20fcOutt5Zly5YVM1i9DMvkBEFX9rSJCVtttVV9WZYgbVIIhEAILDYCCbAuthpPeUNgkRHwEiv7rL788svVALz22mvL5ZdfXvddffbZZ8uaa65ZlixZUvbbb79yxhlnFPtJWRKVFAIhEAIhEAIhEAIhEAKLnYD9WL0IS6D1uuuuq0HX2267rU5auOWWW+qWW4KtZrOyp01asI/r6quvvtjRpfwhEAKLjEACrIuswlPcEFhsBARYH3/88bp/1Pnnn1/sIfXoo4/W5U0CqQcffHA5/PDDq0G45ZZbZtbqYmsgKW8IhEAIhEAIhEAIhMCkBGyXJWh60EEHle22265us+UlsV6C9dJLL1U72/ZbVogdsfwFsUcffXTZfffdJ71nfgyBEAiBhUYgAdaFVqMpTwiEQHn33XfrkiUzVBl7N9xwQ7n++uvfNwItebINgJH4Qw89tC5pMov1Qx/6UN3gPwhDIARCIARCIARCIARCIAT+joAl/6uuump96ZUXX9l66yMf+UjdEuCOO+4o99xzT53Q4CWyr776annhhReqDW7LLXu5Cs4mhUAIhMBCJ5AA60Kv4ZQvBBYZATNWGX2MO7NVL7zwwnLBBRfUv+0ptd5665VtttmmnHXWWXWE3X5SeYnVImskKW4IhEAIhEAIhEAIhMCUCdhv1bZaJ598cvnOd75TvvzlL5cbb7yxPPnkk3V2q60D2Nhnn312XS3mvQbdfVnZ5EkhEAIhsNAIJMC60Go05QmBRU7AiPnDDz9cli5dWg2922+/vY6or7/++nWpkr2h9t1332oUbr755iUG3iJvMCl+CIRACIRACIRACITAyAQETL0E68ADDyzsbAFWh5Vjr732Wn0J1uc+97ly55131hVje+yxR93L1Yqx2N8j484FIRAC84BAAqzzoJKSxRAIgcEEzFh9+umn6xIlo+YXXXRRuf/+++teq5YmGWU/8sgj695RDLwPf/jD9e2og++cM0IgBEIgBEIgBEIgBEIgBLoE2rYBXoJl0gJ727HGGmtUG9w7D5YtW1Zs2fX8889XO92+rLblsqLMeUkhEAIhsJAIJMC6kGozZQmBRUzgiSeeqEuULrnkkjp6/uMf/7isssoqZaONNiqnn376+5vt2y8qe60u4oaSoodACIRACIRACIRACEwrAe832GGHHWqA1YxWL8C69NJLyxVXXFH3YjUJwgux9tprr3LmmWfWl8t6WVZSCIRACCwkAgmwLqTaTFlCYJEREER97rnnim0ALEny5tJ77723zlrdbLPN6surDj744GrEmcFqU35B16QQCIEQCIEQCIEQCIEQCIHpIWA2a1sdZr9V70QwS9XM1ttuu62uMHvmmWeqvf7222+XBx54oL5wln2+8cYbl9VXXz3bBkxPVeQuIRACs0ggAdZZhJ9Hh0AITI3AT37yk/L6668XS4++//3vl6985Svluuuuq3utmp266aab1n1WTzzxxHLGGWfU/aF8nxQCIRACIRACIRACIRACITAzBMxkdey22251K4Add9yxzmL1wllbd3kJ1re+9a06IcKkCC/Jsm2AQKz9XFddddWZyVjuGgIhEAJjIJAA6xgg5xEhEALTS8A+Tt/73vfq8iPLjWwP8Morr9Sl/4ceemjda9XLrLbddtu6v9MHP/jB6c1A7hYCIRACIRACIRACIRACITAhATNat9lmm7L22muXffbZp1x88cXVdvcSrMcee6y+CMuLaffcc89yyCGHlP3337+eP+EN80MIhEAIzHECCbDO8QpK9kIgBP6OwJtvvlm8yIoh5m2kl112WV1mZAarN5cKpu60007lsMMOK7YFYNBZopQUAiEQAiEQAiEQAiEQAiEwXgJmsq6zzjr12GqrreoWAOuuu27dEsAWAYKstvnyEiyHLQT23nvv4qVZzrNtQFIIhEAIzCcCCbDOp9pKXkNgERMQXH3kkUfKOeecUy666KLy4IMP1pFvs1Ntkn/ccceVz3zmM+/v42QvqKQQCIEQCIEQCIEQCIEQCIHZJcBeP+igg+rWAaeccko599xzyze+8Y261dd9991XfvCDH5SbbrqpHHDAAeXTn/50fY9CAqyzW2d5egiEwOgEEmAdnVmuCIEQGCMBe63edddd5eabby7XXnttHek24m3z/D322KMaYrYDsH/TJptsUjfYz5YAY6ygPCoEQiAEQiAEQiAEQiAEJiGw0korFfb5GmusUd+VcPzxx9d9V235dcstt9TVaWaxeqfCCy+8UF9Qa8uAXXfdtU6e8C4F90gKgRAIgblMIAHWuVw7yVsILFIC7777bvGG0RdffLG+yOrSSy+tezZdffXVlQjjzKzVww8/vJx00klll112KRtttNEipZVih0AIhEAIhEAIhEAIhMDcJ7DKKqsUh31Xbedlm68NN9ywvuDKBIqnn366PPTQQ+WHP/xhfSHWSy+9VGe9brHFFnXLgLy0du7XcXIYAouZQAKsi7n2U/YQmKME3nrrrTp6bZ9Vbx297bbb6j5NP/7xj4u3kRrRPu200+qeq2atZgnRHK3IZCsEQiAEQiAEQiAEQiAE+hAwYcK2ATvssEM5+eST65YBbH/vWrBlwHPPPVduvfXWcuCBB1a7nw+w6aab9rlTvgqBEAiBuUEgAda5UQ/JRQiEwHICZq56iZW9mBhX119/fd2P6fnnn69BVDNVDz300PoSK9sCGPVeddVVwy4EQiAEQiAEQiAEQiAEQmAeEbBlgJdZrbnmmjVwaoKF2axmtnqJbfMJbBdmksU+++xTZ75uv/329cVZZsImhUAIhMBcIpAA61yqjeQlBBYhAXupOn7yk58UBtSyZcvqS6wuueSS+kbRd955pxpf9mCyX5OXWdlvNSkEQiAEQiAEQiAEQiAEQmB+E1h55ZWL47DDDqt7rrLzrWDzUtsf/ehH5e67767vYLj99tvLUUcdVU499dS6Vdg666xTr8uLbed3/Sf3IbCQCCTAupBqM2UJgXlIQGD15ZdfrtsALF26tAZY77nnnhpcNaq99dZbl2OOOabsvffedQlRlgbNw0pOlkMgBEIgBEIgBEIgBEJgAAGzWT/2sY+Vtddeu85Y9YLbG264ob4Iy7YBr776an35ra0FBFu9k2GDDTYYcNf8HAIhEALjIZAA63g45ykhEAJ9CNi4/qmnnqqG0lVXXVW++93vlscff7y+4GrzzTevBpb9Vo8++uj3R6ozSt0HZL4KgRAIgRAIgRAIgRAIgXlOwEusvF9hvfXWK9tuu23dDsyWYAKvXnzlRVg+n3322bptgAkYO+20U/noRz9aX5SVrcPmeQNI9kNgnhNIgHWeV2CyHwLzmcCjjz5arrnmmvKlL32pLv954YUX6p6qS5YsqbNWjz322LrnKmPLPksrrbTSfC5u8h4CIRACIRACIRACIRACITCAgECpIOvhhx9evIPB59e//vVy8cUX18kYN998c92n1RZifvvUpz5Vttxyy7ybYQDX/BwCITCzBBJgnVm+uXsIhECHgJdYvf3223X0+aabbio33nhjXfJjI3t7rQqselOoTeztv2TZjyVCSSEQAiEQAiEQAiEQAiEQAouDgEkVXoK12mqrlY022qiYbGEVm1mtVr3Zl/WRRx4ptg3gWzz33HN1O7E99tij7LjjjjU4uzhIpZQhEAJziUACrHOpNpKXEFjABN58883y2muv1S0BBFbPPffcGly1zMeyHwbTXnvtVU4//fRiWwCj1nk76AJuEClaCIRACIRACIRACIRACAwg8OEPf7g4Nt544zqb1Z6rH/nIR+ps1SeffLJuNeb9Dffff3/1M7zfYfvtt6+TNPgSXqCVFAIhEALjIBBpMw7KeUYIhEBhAN15553l29/+dt2s3vYAXm4luGqjetsBnHjiiWXDDTcs3gpq1DopBEIgBEIgBEIgBEIgBEIgBBDwAlwvv7XvKt/h/PPPL9dff33dl/XWW2+t2wfccsstdYuxE044oWy22Wb1mtALgRAIgXEQSIB1HJTzjBBYpAQs2fEiKyPKjB0zV70N1IusLPXZbbfdir2TbAtg1qo9lpJCIARCIARCIARCIARCIARCoJeAvVk33XTTOhljiy22qBMy/N/WY7YLaBM4+B+2Ddhzzz2rf+ElWKuvvnpms/YCzf9DIASmlUACrNOKMzcLgRBA4L333qvH66+/Xh566KH6EqulS5eWO+64o76oygxVG9GfeeaZ5ZRTTilbbbVVnckaeiEQAiEQAiEQAiEQAiEQAiEwGQHBUgdfwiSNK6+8snzjG98o3/3ud2tg9fLLLy+O448/vq6Qs0rObNY11lij3jYvzp2Mbn4LgRCYKoEEWKdKLteFQAhMSMCIscDqFVdcUa677rr6ls8nnniijhrbfP6AAw4oRx55ZN2EfvPNN6+zWSe8WX4IgRAIgRAIgRAIgRAIgRAIgT4E7MdquzHbjB188MGlTergi5jc8corr5Tbb7+9/vbxj3+88D3WWmutPnfKVyEQAiGwYgQSYF0xfrk6BELgpwTeeeed99/i6c2etgK4+OKLy2233VbP8NKqnXfeuRx++OHvH94MmhQCIRACIRACIRACIRACIRACUyFgVuo222xTA6d8DcHT9ddfv34+9dRT1Rd54IEHyrPPPlteffXV+lLdrbfeugZkvQAr732YCvVcEwIh0I9AAqz9qOS7EAiBkQnYb9XMVctxLM8RXLX/kWRJztFHH123A/BWT3sl2YM1KQRCIARCIARCIARCIARCIARWlMAqq6xSNtlkk3LGGWeUvffeu9x1113l61//ep30wSfho5jJeuihh1a/xLYBebHuilLP9SEQAl0CCbB2aeTvEAiBkQnYZ/WZZ56pS3BsMO8NngyaJ598shgd9uIqy3X22Wef8rGPfay+yTMzV0fGnAtCIARCIARCIARCIARCIAQmIPCBD3ygOGwZYGaq4KnPbbfdts5ive+++8pjjz1WA64vvvhifSnW7rvvXn0V74No+7NOcPt8HQIhEAIDCSTAOhBRTgiBEOgl4CVW7777bnnjjTeKpTdGg88///xy4YUXlhdeeKG+4Ipxs++++5YTTjihfOITnygbbbRR723y/xAIgRAIgRAIgRAIgRAIgRCYVgJrr712cVg1t+eee5bLLrusrrB77bXXivdCCLR6V8Sxxx5bjjvuuPoSXnuzenFWC9ROa4ZysxAIgUVBIAHWRVHNKWQITC8B2wGYuXrDDTeUa665pi65sZG84KrZqbvttls1WMxaNYPVCHJSCIRACIRACIRACIRACIRACIyLwKqrrlq22GKLctJJJ5XtttuuvgzrkksuKd4XYQUeP0aw9eabby6HHXZYOeqoo6rfktV246qhPCcEFhaBBFgXVn2mNCEwowQEVl9++eXyyCOPlHvuuadcd911ZdmyZXVbAMtq7K+64447lgMOOKAceeSRZcstt6ybzM9opnLzEAiBEAiBEAiBEAiBEAiBEOghYIsAL71yeOGuYKu/vR9CUNVKPFubeY+EfVpfeeWVGohdsmRJ+ehHP1oEaJNCIARCYFgCCbAOSyrnhUAIlDfffLM8+OCD5YILLihf+tKXqlEi4PrOO++UPfbYoxx//PF1SwBBVsaLJTZJIRACIRACIRACIRACIRACITCbBKyo45+YELLffvvVF2DZJsBWZw8//HD50Y9+VK666qr6EizbBpx++ukJsM5mheXZITAPCSTAOg8rLVkOgdkgcOedd9bZqt/73veqIfL444/XjePbjFWzVm0JYJP4Nddcs3zwgx+s+xnNRl7zzBAIgRAIgRAIgRAIgRAIgRBoBFZaaaXqn/BRvIj3tNNOKzvvvHN9Ue/VV19dbHf27LPP1i3QvATL//k3JpGsv/765cMf/nC7VT5DIARCoC+BBFj7YsmXIRACjYBtAbzM6qabbirf+MY36siuDeKNAls+I6h6yimnlF133bUuu2nX5TMEQiAEQiAEQiAEQiAEQiAE5hoBAVPHDjvsUH0Yk0NsfWZCiW0Dnn766XL//fcXgdZVVlml7LXXXgmwzrVKTH5CYA4SSIB1DlZKshQCc4mAvYhsC2AT+CuvvLLuTbTNNtvU5TNHHHFEfTOnN3QyTJJCIARCIARCIARCIARCIARCYD4Q4L8Isv7ar/1anTRiiwAvwbr33nvry6/MbLWPq71bN9hgg/lQpOQxBEJgFgkkwDqL8PPoEJgPBGzuvu6665ZNNtmkGiBrr7122X333Yvg6i677FI3iXdO9ludD7WZPIZACITA+AmsvvrqdVBuww03LE8++WR9U/PGG29cndbx5yZPDIEQCIEQCIG/IyB4KsjqaC/E2mijjeoLfL38aqeddqovvaLHkkIgBEJgEIGV3lueBp2U30MgBELgoosuqm/b9EZN2wFYKpMUAiEQAiEQAiEQAiEQAiEQAguBgNDIq6++Wu65557yzW9+s+y2227l4IMPrrNXV1tttYVQxJQhBEJgBgkkwDqDcHPrEFhIBLxZ84UXXihGcO2/ut566y2k4qUsIRACIRACIRACIRACIRACi5iAAOs777xTXn755eKFvnweWwN86EMfqi/IWsRoUvQQCIEhCCTAOgSknBICIRACIRACIRACIRACIRACIRACIRACIRACIRAC/Qh8oN+X+S4EQiAEQiAEQiAEQiAEQiAEQiAEQiAEQiAEQiAEQmAwgQRYBzPKGSEQAiEQAiEQAiEQAiEQAiEQAiEQAiEQAiEQAiHQl0ACrH2x5MsQCIEQCIEQCIEQCIEQCIEQCIEQCIEQCIEQCIEQGEwgAdbBjHJGCIRACIRACIRACIRACIRACIRACIRACIRACIRACPQlkABrXyz5MgRCIARCIARCIARCIARCIARCIARCIARCIARCIAQGE0iAdTCjnBECIRACIRACIRACIRACIRACIRACIRACIRACIRACfQkkwNoXS74MgRAIgRAIgRAIgRAIgRAIgRAIgRAIgRAIgRAIgcEEEmAdzChnhEAIhEAIhEAIhEAIhEAIhEAIhEAIhEAIhEAIhEBfAgmw9sWSL0MgBEIgBEIgBEIgBEIgBEIgBEIgBEIgBEIgBEJgMIEEWAczyhkhEAIhEAIhEAIhEAIhEAIhEAIhEAIhEAIhEAIh0JdAAqx9seTLEAiBEAiBEAiBEAiBEAiBEAiBEAiBEAiBEAiBEBhMIAHWwYxyRgiEQAiEQAiEQAiEQAiEQAiEQAiEQAiEQAiEQAj0JZAAa18s+TIEQiAEQiAEQiAEQiAEQiAEQiAEQiAEQiAEQiAEBhNIgHUwo5wRAiEQAiEQAiEQAiEQAiEQAiEQAiEQAiEQAiEQAn0JJMDaF0u+DIEQCIEQCIEQCIEQCIEQCIEQCIEQCIEQCIEQCIHBBBJgHcwoZ4RACIRACIRACIRACIRACIRACIRACIRACIRACIRAXwIJsPbFki9DIARCIARCIARCIARCIARCIARCIARCIARCIARCYDCBBFgHM8oZIRACIRACIRACIRACIRACIRACIRACIRACIRACIdCXQAKsfbHkyxAIgRAIgRAIgRAIgRAIgRAIgRAIgRAIgRAIgRAYTCAB1sGMckYIhEAIhEAIhEAIhEAIhEAIhEAIhEAIhEAIhEAI9CWQAGtfLPkyBEIgBEIgBEIgBEIgBEIgBEIgBEIgBEIgBEIgBAYTSIB1MKOcEQIhEAIhEAIhEAIhEAIhEAIhEAIhEAIhEAIhEAJ9CSTA2hdLvgyBEAiBEAiBEAiBEAiBEAiBEAiBEAiBEAiBEAiBwQRWHnxKzgiBEAiBEAiBEAiBEAiBEAiBEAiBEAiB2Sfw1ltvlTfffLN8+MMfLqusssrsZ2gFc/CTn/ykvPPOO+Xtt9+ux7vvvltWW221suqqq5YPfvCDZaWVVlrBJ+TyEAiBcRBIgHUclPOMEAiBEAiBEAiBEAiBEAiBEAiBEAiBFSbw9NNPlx/84Adlp512KhtvvPEK32+2b/DKK6+U5557rjz55JP1eOmll8pee+1Vttpqq7LuuuvWIOts5zHPD4EQGEwgAdbBjOb0GUa5Xn755fLwww+XZ555prz22mtliy22KJtvvnnZZJNNfk4YP/LII+Wuu+4qL774Ynn99deHKpeRszXXXLMe66yzThXwG2ywQfH3VBPl8aMf/ag89thjZa211ir77bdfHZ2byv1eeOGFcuedd1ZFhMMwySin566xxhpl7bXXrmX6yEc+UjbccMORRgeNMho9VQ7lefbZZ8urr75aRx1XXnnl+gz3VB8+jbAagRw2uZ+yMSAo3WGSUc5WX61s66+/flFnoyQjwtqSsj311FPvtxnllfBzT+Xy6Rhl9FiZ7rjjjsrsxz/+8VBZ+9CHPtS3bMo3amr19sQTT9S6e+ONN2q9KYO2vdFGG5UlS5YU91aXo4wau6d6e/7552t7GCZv2kZvvWG63nrrDXP5hOe89957tXza56OPPlq22267sttuu014/kQ/aA8Mvscff7we++67b9l6660nOv3nvseh9Xf3YSRuueWWZbPNNqtsP/CB7FTzc8Dm8X/MviCr6KMmt7r6qCsjyBYygD4ia4ZJ+gnZo6905duo/aSbT+2arPWdtugZ9AFnjQzwnFHTbOhajpiy3H///VVmD5tnZf7Yxz5Wy6pvkneTpQcffLB8//vfr/VGbg6TVl999fflW9eOmArb9jz2TpMr5DQ7YhhZYkaQulbn2qBPbVBZWv2TvZtuumnV3W32UHvuoE/3164feuihel92wqBEv7R2rW2rB21aPrTHFUn6o3ahrOQvToN0pjxjpB3TGxjh43v5UW/ypn+wM4dJrX3KhzwpI9nw0Y9+tLa5UWyjYZ6Xc0JgKgSabtB/yRj2qXbKXiETuvJxunSYvqQ/TGdqNh9drI8eeuihI9mxK6LDyLKmn5VtKv4iWUUu33ffffWYiA255Fw2+7AB1qbD+I+uHSbNhA7zXO1NPvhZ2hO7iT3U9A5+nj2sH9J4YTds2fjCzf9o+plv57upJuXgCymTfrPPPvuMdCs6gu/QmCgP3crHxaTpZ22Nbzhs4vvRzz/84Q9r+3LPQYldgEXT0foqG1E+5GdFUtOL9KxEP0+lv3TzoD25n74v6fuDbGTXsBNcJz6kjHSzWBK7fRjbqj4s/7xPYHJL+v3T8sdcJUCAEhRf//rXy7Jly2rw46STTirHHHNM7fxdo/W2224r//2///dy9913T+iA9QaSWqDJ6JngzC677FIEV7qOUe81XVYCPL1JB7788svLBRdcUIM0u+6665SFFCPi85//fLnyyivrKGa/53l+N48EJcdAmQSJlGnPPfesTkdXiHSv6ZahPUNwmxK45ppranluvPHGyl+gl0Lcdtttq7A88cQTywEHHFCDke4/0X27z/C3ev2Lv/iLct1111VnrT2397zu/dRLK9s222xTy2b0s9eh6l7TvV97BgUvmHbhhReWa6+9trYZwrcFsXGjMCkD7cHnsAaA5xlx/l//63+Vm266qTpx3Tx0/+7mk8BXNke3bBRdN3Wv6X7f/lZGCuSGG24o3/3ud2vbYdwos3ozEq6+PvnJT9aRY9+NUm/33ntv+eM//uNy++23V+OgPbf3s5tP9aM9Kpt2o03uvffef8/o7l7Tez//b/XXfmM83HPPPeXSSy8t3/nOd8qnP/3pgQHW3nu4Fza33HJLufjii8tFF11Ufu/3fm/oAKt2tHTp0vp8gyvK9gu/8Au1bAy7bp9r+c7n/CRAH6nvb37zm7V/cWTIv2OPPbYapN0Aq8G+//bf/luVLQzxfqm3vTNo9ZGuPhqln7S2LZ+eed5559X+r20bHKMvORa77757OeSQQ6oMYFRLvXnpl9/23a233lplwDh1Lfl8xRVXlC984Qvl6quvblkZ+MlB+N3f/d1y5plnVieiG0DodzE9R3Yrm8Bbv9TLioOv3hzNjqAzGlv36L2me99Wb93vOA/sCHJNG3AMkiXuQyaS/4LEbBB2ExkpmKJ9tmDtEUccUdhSnPbWbifLY8ubIKQ8ffWrX62MJgpCd++l3WnTTQfQQYLedGvXeexe057X/ezHSR1pD8rKiSK7e+2Bdo92PduGzL/qqqtqWTDCR1n0D3Uob6effvr7QY1BeeNwf+9736v5IBd23nnncsIJJ5RTTjml6t2urdryk88QGDcBukFQp+kwcqbpMP2mKx8NpP/X//pfaz/ni/RLvf2i6TCysNl6bOluUKX3mn73nei71ofZoeeff36VQ0ceeWQNskx0Tb/v6UR2LDmv7/ZLvfkkK5sMa3J+FH+x+wzy54EHHiif+9zn6tH9rfs3hgcddFA58MADu19P+jfb/8/+7M/qhCM2ab/UWzYBvVa27bffvtqxyjasDuv3DHWlvWljZCN/C2t+nDZ31FFHVV+k37UTfccf/cu//Muq3wQo+6XesgmiaY/Kt8MOO9Sy0c/s85Z6r2nf+2xtrvudurvsssuq/jj66KOHDrC2e9FbfCh6iw9KXwtG4y1v2jT9Id8tyDlZHlve6P5vfOMb1fZzT7quX+rey/1NCmmM6C42onw028A9utf0u2crW/c3vqf4BZtBop+7sqB77qC/2/2b7/61r32tXvJXf/VXAwOs5Bf7UT78rYwnn3xyvU47aIwH5SG//4zASssr5O9HwH72e/6a4wQuueSSGiQSwKAAKJn999+/7Ljjjn9vBquAHaeWAPc3BXzzzTdXwaWYrvvVX/3VOnLGUSH4HRQd49p1nDjCWFCUgKNEJzLW3ZPwMmJkNJgS8UloCvYSAocddlh11gaNrrhXv2RUi5HDIJI3ZXJvDoFnc+w+85nPVEOGgJAXToKD8lEmAs5ooaChMnHUlGkigcJBw4Wj+e1vf7s+T8CQQsDKMxhhFAFjSjDQPU877bTKqzcg2K9cvqP4lU0gwMEJxM5zdVvKUMDMKBMh77nypWzyos4wkQ9GCGVNKShbVyl0n+9612pPlL3nyK96xogCEVRWXqOA/o/bqaeeWvbYY4+hg264KxtBrh2oN0EJ7VFiwCib+sNSvrpla/Wm3ThXvWmTeA9y1DyHEcLIorBbG/Yc95UXxgEHV5DlE5/4RFUyEzHr8vO3MilbG5FWb4xVdSeZQaqf4drK1tok9vKg3jiy+rGycbiVbVAAASP10/qa/tb6uD7yO7/zO+U//If/UPMx0T/q3H20P9c7zIpzH1yUT4BFGYZJnHp1jPf1119fBwy0FeU67rjjqjwZ5j45Z+4TEPAyaEF+aN9NHzHWuoEqJdEmuv2k6SPfSQa9yADyS5/uym59pMk393V/8o1zQL5NZOjq79o2/aBNC165N5munZJrZAM5TrYYCDj88MNrMGlQ36uZ/uk/w+haZTMbcDJda+aUAQllI8Mn07WM9D//8z+v/Us5hklNT/2X//JfaoCVYe+7yRJ5wDEh38xyUm8GyvCUzJb45V/+5fcDg3g7mh3hOvKNs0LGkgP+nkwv0uVkpGc32UbPk6naEUeATBqL1ds7AABAAElEQVQko+WDLhV44LS5npwzi8T91YU2QdayczDXBuiByWyCLi9tSRtWB8pKZuJFb2uz0qc+9alq+3hWa5PyJh90gOs4gnQru0ywV54GzSgS3HUfMrrJbs9WP9qke+DEMeyXmm0j7wZ28dQemt6Xf/2jzbYxkH/88cfX+03WNj3L4Kx6x6HpATZM0wP0cFIIzDYBOoxfRYdp04J3+iAbrNmjLY9sKn2dPtInmg7jZ0kmN5Dz+m1Xh+nbrZ/r6+5Lh+kLdJh+OtXE/iPjvvSlL9XgKhltsoAgy0R6sd+zXNcGCJWNHcsG9CmZhKBsZFjTYWSospETzY6lw9jmdBhfaJCcaHkxCHPOOedUH4vMmSipn1/8xV+sAUl1NExyb+Ugj8hFf5ORZLXEL/2lX/qlv6fDDMJ2y9Z0mACiANxkOqxfvrQdspDuJh9dzz9g9+PGp+cHjJLoYe0QfzrA3+wdf0tkNl9UvUn0hXprZdMetU26h52u3ujCyXx0+pl+9wxsH17u6/OzlE8f4Sv80R/9UX3eoH/MGHcfkzn0Q31H/+Dn0UWe4/50JV9F3bMz9ZlBtotnsyPlyzMwMuNXf8WIDaifatfad1c/Y+SZrsGIrsVI+2MfeL7JOJMlut7zMfF85ZAX/qE+pn/Qz2TBVJJ76/u4CbKrC/fU9/XByZJYCruh+WnsoyaT6HjtO2k0ApNb0qPdK2fPAgFCTGeiIAjlNqLTb7mJDiIYRwHqSBwUQpUBLumIgpFtVhmBQtA61zmexfCmEPyfYOZMCR4SLP0Ckgx2nZ5Q1IEJJnllZEuesSJJOQk3eSGY5ZHTqmyEvt/NQiEsfd+EJMVKERGs/m55cg/nGJmk2PoJTPwoYgKZE6J8jAZKgMD3XIJYANczKHDPoKTcb1gl7J6ce2XDTtk4PAwBwt3vnD9BU+zl23meRWk05eYTb3Ug70bLXStg2psYZ+5PuRk5pIw5wZSs2cwYeobRVuUjhCkK32s3FM4wyb3cU57dR9mkFmDF8owzzqhKVZmdR/m1uhLcpZDUs2uUzaHe1Ln89Cb11MpnlA4LZaM8ONLqhsL71re+VduDsrknI4qikadhEqPIIc+MMWWjWFuAleHEeNMeGASNgX7GGHKefCiXv+VTP+UUMzb6la3lS7twP+1Y4ICydY/WxzEYJsmv5+JFBmjD8iOPjGm/D5sYZoJk2hsj8itf+UrNn76ivnBIWhgEyBqygzwld+kjMqGfzFPvDu2MXtCXtV33kARfBKIYroxeclCb1gbd3/lkDznsmtbutSmyqDkQXbL6BbnhHvoCZ0b+yHpyzT19T74xeskEQcdRl7f16loyoKtryRu61v31J3nXz/RXZXN+07Wc3Ml0rT5PTjH8OcV4OcjNiRxq9/MsfVD/tPyVThgm4aVu5JmjIK/YtQCrAMHZZ59d5ZTnq19lEwglh+hH53KayTt5+fjHP16dOTwmsiOch2HTA7jQVZK8DJOUWb3S22Raq191gKP24f5YOpeewYe+Eizpl7fe57oXnSwwruzqU1nJdPpA4hA2Rk3WapPqz/nsBzYGPvIhb9pg01P6Q7/U7A/tgaxVjtaO8FNeNtlESV48Tx/R15SD7qPP9Av91LOtYqAf8aCP2iDlRPf1vVlZDtd4jsFpZVZ39HACrJPRy2/jIqDPNx1m4Kb5VP10GFuODtOG9Q2y0N8twEpOnnXWWVX/dXWYPqZvO58OY583HUbG0WH6aj8dNhEHfV/fJjuaHdnyMdE1k30v7/QYGaafkpvkI/kkkUVNhk2kw8gwsrb5i3y1yfzFlh8MPBMfOoSfOVES/HbPfvUz0TV8XXa9uiLrmg5rAVa+dCtb02FkNx2m/FjwsawM8EmnHXzwwVVeTqTDunlRPu3F9dqa52JNxguyTzXA5hkCkcpH3tM5yibv2pkk0NbK5v8Y0BM4KJvr6Gf/pzuUjb5iKykbu6I3df2Opp+1PZwk9xw2sSVwMftVsI8feMTyAUaH/kAvmWmJm3PoZvYe/sMEWOkf7YXOUXb9zv2UVd71U+30V37lV6puU37tXn21wLzzsFJG7Jyjz+ozGOkP/ZI+So/Sz/qpe6p79oHf+JiT6ed+92zfqQN9VT2TYS1OoC0Mk/jN9DPG8mdVtHLS1Xhpn0mjEUiAdTRec+7sZpATFJwqimGQE8AgZhhwqBjJLRFOXQEqIMLxajPpjNSYcfY3f/M3dZbS//k//6c6BM5jHE820sbpIviN9BEklp5NZyL0OcKULeOlKQGCTnmboeJTuZ1LwQq8CmKarWkpiqUDjJPf/M3frPfqFzAkdEy9J8gYGQQ7Z8p9CVoBQ4rBSLj7EqgCg0uXj4ibbUVRjZKUhfHgOepbmQhTigB79c0IUE51QFASqpTiEcuVEkfGslFLFhkDytY2Te/NB2Xx5S9/uQp8zzCSy8B0X/nwbA4nbspL2bXlDUaPLQUeJcm7e1NouLfUyuaZrWwUqWcrm7bIELGM6//9v/9XR+spKtdRnNp3bxIsFGykzBgdZqYqH4MDR2XTjhlzFJV7O0+b8NuwAdb2XGVT357huS3JY+tnyubZzsNTH1Fvgs/qyxJmdUexq1+MObuTJfdWvzhQlC2wMdk1/X6TH3XDcGD4MSYYvVNJyqxOKGlyps38YLgkLRwCglbaCYNaX+UoDNJHZDI94/wWMEOkXz/RB7v9xAAQeUXGGaUnk/Un/ahffyWHDUjKFxnimc7XDxmXZLNBDP2GU6j/Dxr5n6z2pqJr9fGma7/4xS/W7T3MTCX75bVX19I5fqN39FdysTmcTQ/25pFM+Nu//dvqHNCFOIySmkzmUJDLS5frtpY8s+tk+FsQwnm402PsCHKNHfCnf/qn1ZnDStnUb79EBruejeN6bcxg7yiJ7DdwqG0ZAMCZbm1OBcfEOfS2tqwdcBLJXG1KGxk2tfar/bh/W67nen2itTu8yETlx4msxUjQhU3CRtMOf+u3fqvqLIzce7JE1hpMF7jWt+joYZxceo/twskzy4lNaXBM/hzyZ7WKusCFXnK+AcNhk/ywqzDhxLsPmyopBOYCAbK0DWqRDeTxKDrMgF9L+kyvrdd0GJ+IrUeHmalJ5tBhDy8PirlGP2cXjpL4gWxytnQbzBnl+n7nTqTDmpxvsog8a/5i02H8ATpMIIy/SKY6jywlWyZKfDhBPvr6iOU+zGc/+9mJTq2sWmBowpP6/NDVYTjz11pqZSPnpK4O4+s1Hfb55dvT0Rd/8id/UmUZmTaZDmv3p3/5FYKEfDI+Ap1mWwDydkWTfKs3vhLdpW211K9s9AUGfAY6Ufuhe9hj6qGtSFE2/mW/5L7sJzqarhEobRNm+p0/0XfaLTa40Df8T32EXabOMLatDN2MPduDvpJ3gdZhU+ubBhHFJPiRLXkGfp6nXGwk98ZJjIFepZcxEpw1ME8/q0OMWrtp9+v9ZDctWR5/EHtge1n1JZC5IokdaFCVDax9KdOoSXkxlz/l1g/5ru6bNDqBBFhHZzanriDIGMWSDkF5TZZ0oCakCJHu+b6nLH1K/nYwLggX5xKuFLjO20aYCD4CmJJzbru+3YOgkjcCTSKspjsRaA7KSX6agJOX9ptntv83Vs6lPHEksO1Vybg699xz67mUk3ObEeEczgDlwVASGGQoEUoYEfRGJ+UDX/d1P+dzkgQmR00t//LhGY1vK0tvWZ2nTM71qX04GBAcyOboUajdssmX8whoI2vyr04pAPXanqMeOcGY+L3NbJmKEO4tW2OjbO7fntnK2i2b8mGNsbJRcpSktsl46+bZfeWPslA+1yxZruAoQ4xa2/RM3/mdU6rOfDKqBC2757a8TvQp7/LgcF1L/crWPc+52o/8CnJrkwwVS744sox+HBqbdl+fzWho9aq/dp/dPXeyv13fDs8RvJHHqSZl1p4c/m6juFMdrZ1qPnLdzBJQr119NEzb6/aTfvqotXPtpvWTJtvoErJVP6GPOLcGAM18Jbd6+4n8MB4Fn3rln77jfvo/+UaOkPec1qkk+e22+35l87tE7jiUT9marjXzUz7IN4Es+krZurrW7/qRclkuJ6Dnb/dpcq2bf4a4OtKnPd/5EzlN3eu6f/eWjY5vyW/K0tJEZeNIOVe9cVaaHUH2qTe/teQemPi+lcl5w6Ymb7QV96W3LX/k6DdOzmn6DqM280oQkd5s7XrYZzYO2px2Je8tadPddu17ZWwcWztna3C+zKg2MOAe5H8b8Gz3a5+eSca2PuV757dntfMm+sRWv/B8No38NN6ucW99h6PpnvQkHaz/DZvc231dj7M+Nsr1wz4n54XAVAiMqsP0OUfrs/3kfOt/3fOaDiNzyCUynJw3sEK28anIZX2+XT9ZeVxv4NogON3n/iuaWtma7dbkk/v6jczyKfWT8+ReV4cZEJMvOkwZMWvX15v89B+TJdjpdLHAmc/pTq1sytSO9oxWtvb/VjYc6Kmmn01I8RsbXbBP2dRbPx3mXuScgVsDi4KrAlj8SJM6BO7YHcPUdcvXRJ+9ZZPvltxfnlvqLZtz1Rv9rO3Rz/JLjymbT9936809evUzTqMmAXgzSh9ePsjgfm1WaNc+8Sw6TfDYNlJmkmLq02904zCp1bE60webnvM9Rq0e2v/9rp04V/k9yyxUepl/JojtewPJ+LX7dfPinvj6zbmSvtpl2T1/lL+1RbaKPuX5U0nyIX8OeY1+ngrFn12TAOvPWOSvAQQIPIfRJArEyJZRJjOIOGkCjYRFV7D4u/v/AY8Y+8/yJhhHuRl9FTw2c9EslxZA7QapOFkO33HOzOboJkLXb4JxWBFYzqd8GQ1GZMeVmrA32qpsBDClrnycJMFe9SXPLRlFM6tQUIGg5vw4lKWbXEPZtGt9ToeS6D5jsr/ljSPIMKHQGCqcUMsazI4xg0vZKIqWWoDV6J56x0AAvSnSdp56d537ch4FWbV1BofzZ7o9e7Y8HLF85J7B9fByY0PgSD/T5oy4amMY9KZ+3/WeM+j/TckOOi+/h8BsEiB/9BMzNvUHxqUBFAMRBkQcZEC3f+tPttpwvut7k77ddSD0xenoU73PGfT/pmvNAGq6lgwiA8wSIQearlU+eTZzhEwbJIfJc7rA/QTRcJqqQT6oHP1+l1eH2To+OUd0oyBD28db/TTd4h7qZUXkLkeBM6QuOWa40hPdhDlHntynB11jJg/dom0JJo4rqQ8zdTl/nHLBdcEJTASHfd+vzvze5TZqfrV37QFrz+htS/6v3zjP32wM/+/2sVGfmfNDYLES0HfYlFYs6c9tMMfqAvJcXyQLB/WvFghh25JzBkAEoWYzNR1mVl/TYeQ8HdaCpvLYlVeCxA5y17n0nGtN+Gg6wCc53gJUs1HGXh1m1Q59ij0dtmTJkqqfu/JTuegQctwWZQYVrejiK/LRzPqdC0lgVPlsj4Ezm4qO5gtZXcFm0Ca7ZVMX01EfGLYVDdo+W42d0y+x/fiwDy/3j+hoE2ycO2yAtd89h/3Oc8Q89Fm+smcLQmvzZqXipt32Jm3dOdOZ2AcGhdh0+r+88Y3JDG0uafYI/N3899l7fp48DwlQioKplhJQJJZeG3UiiMfphEwnOkrFrB+BOQqS8rAni71gBF1bIjgZRJYoWN4zWeKoWfrh/NlMBL16ohwpBTOeBMbN0OldmkdJql/OFaHNCMCgt16NkpvZSagLWggGahPjTkYTOcwUMYO0OcVmAfTOqG1bBFDGlA/l6PquodDyTxEKVjj8zcDTHkadxdTuN5VP+bMFBSOCoUmRM3Qs22HsJIVACPzdDACBMfJYfxFA5PAYGOntrwJDBtQmMnINLjHyzbrh/JohIdA1W2kyXUs+S2Q2Z2OyJZfd/FuWbVUF45sjTi+Q4eNOnokt55KjyTlVZ2yJFV0u11sWMpzTyGaxb/lk5dVG2AD0ST/d0Hvvmfp/szUE1NkRnFoDtHSyZcwzkehFz+nnHHoeZ077cfi76d6Jzp+JPOaeIbDQCAiKdHUY+24iHdav7AIr5IJ+yFacjmBXv+dM5Ts6TJ7IXj4WO9yKAEevX8HGNZHBbET+hWCst6p7Oevv//7vv7/dgJm6cyHRI/weS8b5Ck2HqbteHcYW4Ufwu/gnyu5avtNk+mg2yknvCVQK+rZtZqx44Xuom5lK/Eozl9k2fDNM6e1+SVtvK3X0Fz4tu21ciZ1g4pIJL9q2/OiH/GX6cVyJD69NYSYv0cXjIj/4OQmwDmaUM3oIcE4FDzmfBJyOLfjEMW1OX88lc/6/nBkBLU4Vh4aBIhDH4TOzpiXnMYacR/hPlgg6DghF4TqjSp4x7tTKZkTQwdlkoAiI9wYh5U9ArwUVlN+s17YJuPMZQJxgSyIEawWl7Wc6GyOwlJz6UGctwNuWmfTu7yn4SgH7niPZnYnTWyd+V3/qzt+UmPbQaxD2Xjed/5c/hlcb7RcQeXj5aK0gqyBSUgiEQKn9WHBRUEyAlWNjJhBHgMPWTWa/MNh7HVByjHwwYMRAJufIQLNj24Bb9z7j+rufrjWzp6tryac262SYgCA5Ro6Qm3SY4OxEAeeZLKdnejb9IR/qSrnYEoJ305lavRtopLsNIE6UtA2OExsHTwFHf7vHOBM9rV7lmU4WqODk0tvqcCbSZHpRndA7ZuuoIzOZDGq0feRmIj+5ZwgsBgJsPQEtuobP0NVhk9mc/C2DLg7BPfYqWUV2zJVEzvN9zNTkLypP02F0bje1ACs5xLZXJoFYKwrts+4FtPbn9Ll06dIayJrNyQZsCTrMICUdRj4K9PXTYSbqtJmufCjlw4Xt4vvzzz+/bnFmC7c2w1WwUfl6g7VdZjPxN72nbPSgySsmCckHm0qe2EszkdhffFPtmi+K6aAAKx2JO5uGfhxXYg80Hx8jfY9/aQauQPpMJ/6gNsUW8MnPV19kSdLcIDBei3FulDm5mAYClKZgHWFC2BKKZlWMWxFMQ1F+7hYMAOUSPCS0LFUx4toSIeYYJhHA7eDQmT1KYcxGojAZboIG8k/Bc2Z7jRPGmX121KPfBfMoCwEL+/tZkk+JCELYp7Zt0n/mmWfO2igsY9Lov1FyvGditilOlHevQTjTdcnZV2cCR+rQCO+4l6vOdBlz/xBYUQKcFQM8gk9SCyJO5pz2PtO5dBjnjYPnhYy2gHFffW+2UtO1BrvoWrMkDCJNVRZxxskRS+Tpu9lM9KLBMTNj2uxItsR0B1jpiIkctd7yq+s2+OY3jARl1cNsJIEXe9GZ0WVmDL1Mf487aW8CHrbiYRdxKg2sWgJM7yaFQAhMnQCZQ84YwJCaDpts0gq71GoEnwJ1gjz64mzqq34EyF5+lb0quzqs119sAVa2PB1AHwtYCp7xRcgduvmiiy6qs/V+/dd/vcqhyQbM+uVnOr8TZKPDzErt6rDe5dnqk+wURGTXGzDDxeQN27bxpdtLwAQ0DTzaOsIAlrqdDRmrPQn6ty1q1Af7Y6YSn5ttQtd4Nn95Ir2Lu5Uv6t7sYPmin8adBNgNHtg/mW+mjSrHTCf9yPO0K0FmAWms9I+kuUEgVtHcqId5lwsOraAPRUCREG6E71SdvrkCwAgQx1P5OKK9AdZRDBeKExPCljLgkAhyzkYidC0H5SxS7oKk/QKsjDzn2KPONUuXjxIrgxFWhp6ZrOqbIP/0pz9djQAzLAn32TAAsOQ8U8QOfzNgjCD3Lg9WHuWTT4ad83Bg6AyjlCi0cSd1pT7aS0W0yQRYx10Led5cJ2DQaMnymX5m1ZHR3aVmE+WdM0SmkQMcOA6Sw4wZszsZ+GSfWeSM6NlKZFZX17bBzFF1LZmnvGagkI9m58729jUcTM60mTzNOW3Lz6eT9yh6GyOOrron862MsFpn2ADtdObbvegngRczac3UMftrXAFWgw5mc+Mh8KN/2EOwDTqrK7af/jGRIzzdPHK/EFiIBLo6TPkGBVitxKK7mo4yEM8OJyNGkXfjYDmsDqNzyVsBRUErckYZ2b3KyR8TiBRIs4ruC1/4QvWvzjjjjBpk4reNOw2rw8jRNiNVUJDeo3/pvk9+8pM1mMofEVBWVgOq7BiDxia9WP7NHxjGV5kuBgJ32pVBPvaGfGmXM+ULqW8rRtkC+oPnT+RXYtF+F5hnF3VXm04Xg0H3YR8ahNd25X8cAVa+rfakzOSA2EJbcTPX+v4gfgv59wRYF3LtzmDZCD+KgfEvmd3H6CcY53NSHuUivAkuyoTwnkrChJLksBlps4xEsGw2EqFLEbSljoQz5d07w0u5nUdgM3IEU43MGSXHQuJIcc4FYc2sEdgYp9Lv5SeoygBw+FtAm0GGezdps87xm2AD5YSBYMpcTYwLdaZOMGZAMDhHDa7M1fIlXyEwHQTIra4+Mriln/TOkOk+i2wj253HqbGHnRk29BjZwSn0nT7IiSL3yJBxJ88UwGpyqula+R8lMcqVVbCMjDQgJSA9m4lDzMkUPFQeDqZjphy4Ycqq7g0+WglB9lp5YhBxtgKInEh1xSmnr+ks+nscCQt9SWDALGO2gDakfugj/UVQx0qWLE8cR43kGQuVwKg6jB1rYJBMMIHDILwkwDrXUq8Oo3/I+V4dRtY1edfKQEcLXPFB6GM2vkEmcui73/1uPV/w0SDkbARY++kw5evVYYLC5Ci/ki4XQJZncpMfpf6V1erASy+9tK6kMXPUzN02COnccfpadJ6DjOfbs43YHzOVlL+tqmR3TVaf2gEbzSd/CPN27Uzlr999BVbZUTjJPz9ZHc9k4geyWdkBbFM2lDrqnVQ0k3nIvQcTSIB1MKOcEQJTIkChUo6Er+CqpRaCAPMhURqU+llnnVWD6JxNypUio0QYdpYsKg/nvym6uVw2RihDjCIWXDXSaOTRQUknhUAILB4CnDsGqb5PjrUXCVnGx/kza9/sPLKC7GhvmJ+vhJSJ80aWG0i0rHG2BvzmMsM2U0cejz/++Lr/aVvVMpfzPRN5E7zhWAuACwjoB/qHALR+YTsNelRf8rKy2VqhMxNlzz1DYC4TMHmDf3HE8pcktfcPzOX8TjVvgrPkj1UEJqrQ05/73OfKeeedV2eykkW2DDj99NPrpJCpPmemrzNQ1faF53PYXsXsULpYGU2CESwzWEXOChbS183XwsALGiea0TnT+c/95w4BqxjtRWyik34xW4O/c4fI3MxJAqxzs14G5soIiZEkn2Y2mEU427NRBmZ6kZxg1pS64cgaeTUDhhBsgcj5gEHgQUCV8cLBYrxQ9EZgfRo99+ZCbY8hwCgwgjaXlb9RY06goImljZY8+s4sJSPJZuxyrpWbAnOesnEelU39MYSSpk4AR0Fuy7Cx1kcEvM3QTZq/BMyaIfPIO8aewJ1Zd3M5CaxyaPR/n2aHyLfjqquuqjP2tE9t1dYoZJ1z52tSP5Z4k2eCxWbzK3fSzwiYGULuc4bxsQ8vZ1jdL8Zk4JSONLuKfmwzrmwTYICV7BY0uOaaa+rvwwZYtTt7MuJNZgj+07uCDRnsXIwtbfbLTIdpz02H8anarNDZz93PcqCvmMFp9po+KY/66EJNZmw62I7ksMHQ/fffv/ohy5Ytq/LazFbvh5jLib3bZjaykdhHdEzXp1BOdWlAz6xWS/LJV4dtYub7CtG5XD/zIW9kVBvc5DOxWdmk0+13C/BbtaPNimeYLU420s/ZgmD4lpIA6/Cs5sSZlh04NHZ7s5glSUjbCJsjMK4kDxw1h9SU4LieP1PPaWVSPoKEsT+KQHGdWR+CkJbyUZ42K//4xz8+U1ke6r7y1eqslU2d9Sa/UeICqJb/ffWrX617y/zLf/kvq5IXaPACMLNX7IXE2CN8GQuWc1jeM+40bNkoB0YYY4VBZvRfWTl7XtShrszObcuF7VPIiHV/Rh3nctwjhcOWbdzMp/o8suqYY46pfcRSZY4CY9LovPY4Sl+bah5y3fQRaO1T8I4+8snw088EUMaVWj5G0UeMUu3O0ZJVBkcsnxHkOzqWMSvg1l6q4fdxJ2VresmzR9W1jY2goVlP6kfdzIXg6oqWbbrqQj4kukydWxovWHjyySdPu/Myap7VPT3V8jhOOdmW7LY8y4vtg6xu0T8MPgiSCrjSjyeccEI7ddJP/ctMNPrVtXSBQU3PY0f0s00mvWF+DIEpEmjysekwclLgcq7pMPkkB/SVb3/721U+WRknr90g3RQxzOhl8r4iOqxljs52CP7Qy2x1vorJEuptNtJUytYGdyeqN4FkE3PIRgEuE0L4XOp/nGkqZVuR/LH/HZ47KDmnHc4dp17s5q2rn+V9pvKh7skoA5uec8ghh9TBFYOg053YiGbHCujyl9vEKitbtd3o5+GIJ8A6HKc5c1YbTTBj4Itf/GJt7JyA0047bawOLQfEjANGts5u9qyZP2Y6zOekPAKjRm4IGUs4jOYMm3DhNHAWOAq/+qu/WgNIw14/U+dRROqLMaINCXRZVtQbEDXCKrAooOBQ/oMOOuj9ffKMljNuLrvssvKd73ynOleW0l5yySVVIAsmjztRcMrGOfa3YKgZWr0BBMaMOjUryd9t9NtSCwFXbXjJ8lkLgsXY2N8JDwpMufvdc6bLqq6UzaEO7bXD8Z8JpTrTZXF/+T/llFOq3LLUVMDegATFrU36TJo/BBh92qbZ7H/zN39TZYulb/QROTGu1PSiAR9JOxO4n8iBmShfzfExQ8ZMpr/+67+uzo3BmNly4CbStcPOWiAT6SR6zYwY8m02Bov6MTegpc6wxZ6clb9xG/DaMTlk/zScOC8YjTsf/Rhph/LUVithRFbORmLr0T0CrJ/61KeqvmULcMDkcdhksNOgpjbMTjJYi797Wwa8kGfkDcso542HAN1BBrEB6TDycjZ0GJ+DLuWDSGansUebDqMH7EPqIK/oODJqPtiC8s4+FyAiQ8gw/XxYHdbbElwrAMnGV38CQT5nI9Fh6q2rw5RvRXSHcvE3+DLsfvrb4e9xpraS0UQuvr1yaZczlfhn2jWeg5L6tsLQpz5gJmd3sHzQ9dP1OzYGPbRBsQKMlGO6E79c32cPiLWY0TxT/pJ2Rwbqn8ok3kQ/ixdYyWk2edJgAgmwDmY0584gZAl1Al1DJ/go4akqq6kUUGdjVAvYLaQAKyfh4eWzT5WPEzBsgJVRxuhhqHHSOIsUhRFmf8920mbkiyKgkDhonKTepY+UhWCJl1lYTioYqQwUl2sIXg6wMrmfEWT3NKpG6M9GgBV3o7vt7dMUbb8AgjwzXOw9qDzqx5tIlQEfxoR7qUuHPuYaCtNhpsC4k7rC12i9PFJu83k/QEYnWeUgNwQNHHgrX9L8I6DuyEv6iKE7G/qIA9d9w21zTgVvRkmtfXJuGO/f/OY364xGssAxG8lzV0TXkmn0GhnJOCfH6LXeAajZKBuZy3HgfJO1dEvTL+PID5mDDy70mfzQc2b4zqQjOUrZ9CuzSOhm+prens0AKzuTHjLDzyAZfvKI4bCJ7CcnHO4neOB694oeGJZizpsuAs3e046tUmo6jEwaV+rqMM9sOqw3wGplGXuQfyLfvUlfYjOSZy2R/RdccEENkJD/4x6kX1Ed1srRPpXdhIim32dTZjQdJig4mQ5jG8k3e1cwXT0JxPULkKtz9dQmwLCvHONObCArOpRNn+Dj4U5+z0TCg26xKkLA0mBDW9XQ+zzc+du46LPyNRsBVoMGfEj6SzCSbaWepzuJS5jRrO8re79BSDJE32dTNdmAJT3NxhIv4j/2u7ab36aftUNt2j2an9Y9L39PTiAB1sn5zLlfm3Er6MUYMPvu/PPPr0apmawCYuNIOrhZPZxanZFwI3zHGeSdiXISYsrFIDBKOmyAlWNAIVBGhJsZMK5fkVHM6SwfA8SsWi+koJAmCrBqUxdffHFd7qsuOXOOVq8ErsCkWa3K/PnPf77OPhGU5XDNRpIPS82xF5AUBO4XYG15o4QpbUammXaMHakpFbNpMMCCwre9A6d/NhKFSXmrOwoTe/luhuVs5GlFninQ48UolripL/wtyV6yfKbGOJ2ZFSlDrv0ZAXWmv3kBzhlnnFEDkuqWvNF2h10y/LM7Tu0vMo1sE4iUOKdTmcHans4A7Tpw7fvZ+KRrzVxgOJNRBnvoFg7PMIlM1NewIb/JD7N/hr1+mGdM9RxGu2XmAgDNOVW+ccoC7ZRcMoPNygV6rJ/TO9Uyruh1nFurLTibsx1gbWURnNcGpzqThVNKx5IVZseceuqp5dhjj62ztnoHfdsz8xkCM0GAbUuHmRzQT4d50d04Eh3WluN6XtNhzdZjp7IFvfiIj2KFVT8ZTveS+fyRllxDzvJnbCf3W7/1W9WWbL/P9Ke8NB3GJ2o6rPkVoz4fEzLavdyDPJrqvUZ9du/5AmtNh8kDuah8vb4fm8I2fuqQf0nnCcTNJV3TWzZBYNsvCNazHdjpMznwyN/i4/Ch6Ts2HZ3c+kA3f7iziQSBMcR8qvqoe99R/zbJzNZLgpAtwDoTk3FMIOJnkxHaz1/8xV/0zaq+7/cWYGX3/ft//+9r3Xn/xW/+5m/WPYz7XvzTLwVjrVCln239xU876qijVsimnux5C/W3BFjnWc1ysAhxzp8Gb2RC52akmjI+jgArh4SBfNNNN1UlQShaDspp6ycI5wNiQokBQ3hRlgwXgcUDDjhg0tEoxozr8CDI1I1gg1G4fvt1cpYJP04EhTWO1IwtS/kZZpQRQ8tear0jfow8zmYbyZfHbj677c8Lowhg5adclG3cybObgSOIIHCqHxiAmGj5BOfd0dtWBWrdz4ttONvuKxihTwnWjDvJSzNwBFgZD2ZWeWHafHVCGU1eIKSutCV9Za4Ee8ZdvwvheU0eMIo5bgzAG264oX4yjMcRYNVPyCsDIwb8ODLaFJ3UTwYPw51xSh6Q72Qk433QqP8w9x31nF5dy3BvurYrlye7L/mvv5ntwcFTV72yb7LrZ+o3ZZMnKyXoHIN+ZBv5PQ5nGRf6jt7WfshWA2lt1lC33NoYRw7zcc381QYx4kQKkAhSyB+bhBM/m0ne8POJif4xiuNNr5mNI+iijE2vYTvO4PpsMsyz5waBpsO0X4O9bD8yiS6jw8YRYG06zOzUXh3WAnB0mX0RyQByU9+bLBnAsAWTRKaeeeaZdVYkm7afjJvsXivym/7N7jNIJKhIhwn00NHD6rDe55vNjwFu/M+2Kq33vJn+/yg6TDBVPfAv+EoCZuS44H5var6I8gnUWpXHp+wN2vZeN13/bwPkgof6gUG+NpDOd9RnZiIpo4labEg+Jd1Ht/RL/DN9pRtg7fVn+103Xd+1OmI/YMSWwMYLn8VnpjsZAGIXafeeNVkSFP/a175WbVj1pu+LZ+DL/huU2ENsDgFuckZ/1cf013G1wUF5nA+/J8A6H2qpTx4pFQdlRcAInlFeM5kIXZ1NBzcKJ7BLyFEaHCMdeKoKcybzPdm9lclBYXPyBFcJbQKSI6tczcDpvQ8B6+D4UAaUEQYEUW/CjbIUiGVsUKwzzarVl7xR5spmeYFnE5j9AhDySXhznlzPEHBwyLuCVfDSLFEjdi05f1yplY0iMHuVQmG4CIQ2422U4IpyK6eRWkFo/UlAheNnti7FNK7UyqadCIwomz6nbSmb/jZRmxxXHqf6HP1MefQVdcWBme1gwVTLkut+RkBfcQiSkYcGqsicmUytn+gbrZ+QB4KIXkblc6r9hF7jVHCg6Fh7so5zWXYrmzysiK4lx9kHZCQ9QDf1c+hmsp56703WKh953cqmzTQ7gl6ZySCbZ9PbBjpbPcsjB6bfwBW9baYt3WCQaxwBVozkr+kjgwcCQIIS8jlKMLOX/3T8X970O22LnpWvUQYh1XezWfEU9HEkhcBsEejVYezA2dBh5DRZSIcJ2LTBMP1MANgxTCKzWoCV3fhv/+2/nbHAWL/89NNh8tTKZhuefjNw+92r9zsz7Php9DNZaNCJPzyu1NVhglDkM5nIX1BvZFmvDmuDo+xfA0x0MvsXB75VN2hJb/MVyVdBNUFHdnLXB5upsiob/chGZ8cpm7zxF02IWLJ8Fms3r9OZD343HgYA+AoYabv9kn7CftCmnM9uGFcbwKj5iwYJbTUn7+wrk3tmwlbE3jFMspe01c3qUf/4x//4H1cdPcy1zmEXiRewP9iL6kS9J41GIAHW0XjNybObMc4RmMlE6FMiXmh07rnnFsvpKZPPfvazVbBwZmdK8M5UubATUBQsJpSMsnKifvmXf7nOyGL8T6TUsCDkOQvKzQmfaNakZzSFhRtHbqZH21q7MBr/t3/7t7WMlLwXb5l92q9sjDkCWbCeAqHgHAy9bkBY2Tn/PtW7AORMl6fbBpqDbLbBOeecU501edAW1QNjtNfA6V7f+zdDTb1oA2aF4/ALv/AL5cQTT6wzmKdqCPY+Z5j/t7LZ1/arX/1qHUVkXP2jf/SPqnON9yhlG+aZ4zpH2bQrsiRpYRJockefmsnU2pL9oo3Wk1MGIf7hP/yHVS/pJxPJ7kH54jQtXbq0Ojnkitm5w4z8D7rvsL/36lrOpME+8k0wa1hdy0lxrYOs5tDNxP5gw5bLecpGH9oT0B63Bv3oo8985jNVzwxbtlGe2T3X8+ltjhndLTBIb3tuv2Sgi27gcNCDE82o6XftVL/TdwR36KNLL7201p3Za5bqcd7GqY/6lUGAw8utrKxgRwj6alvDpqbjZtpmHTY/OS8EugSaDpvp9tl0mL5Eh1kaTcaw9fhWK6LDuuWZjb+bDjOTlr9IBwk+0WECkSsi5+l6s309g873wrxx62c6zAo+W14JRNla5tOf/vT7OqyXOb/KOYLeZjyyWwSt2iSe7qoNAS5BOzrA98o40wOPLb9tQJFuVj6y3izuX/mVX6mB4xbwb+dP5yceAqpY0btW5k40oIC51XB0OZvIRBiB6HEkdU/30c/qkbzQBm3TaKBmtvXzijJocmmm5d+K5nOuX58A61yvoSHzp0Po5IOSUUVKiQAX8DOLpCX3IDgEC1sAx3kEmMP5puxzPAl+isHyaXtUGjnqKoh2z+6n0TiBOzNFWxJs8X8CabJlK57tWtPxOUXyRxhbLsMB9n9OifMocmWUMCEkHE3otcCoMhHiyiQIKbhqtIaApwiNQvYrk3tzzoxcWeYmPxx5wcvGrZWvfeLqWX7n4HYD0RQFnpw4eZPPbtmc6/vGv9WzT2WWn5ZPPFt9mWGChTwKIGPFETrssMPq/ds1LY8+OZq2DpBXI5eupTAELHH2KS9mjPlN3v1uWYQ8t2QmmfpSJvlQJtdTgD7bzFd5b23A+S31KxvOrWyCu8omwKoeGB8Cq8rGAe5Xtnbv7qd2oTycZ8pccFViUNiCQ1tQ5m6gBptWNuX0LAZBK1ubBaVs2rx6w6klZdMeuvXmvFa2NjrPeDBCy+C2p68Du2HLZgQcI3lsqfV7wfCJAgrt3Papfwo4yaNEhrinenWfYfPTriVn3CNp4RIYVh8hoA+2ftLVDf36Cbnb+gndRXYbiGDsMrIFQvUTQah+7VLfMitD/vyu/Vp+12Z16h/6NgeIA0eW7LvvvlVuku8tuY/z6CPXuJdBkKaPyIxhdC1npemMQbr2mGOOqYHIYXRtyyc5ST7qr2SuWSjKO1kir5o+6qdr5bmVTQDX+Z7TUrMjMGllM2On1Zs6Vm+CCuqQ3FY2OmQyHdru3+qoa7vIJxnr+iZ/2/ndT+1HfukN57sOz8lsD/KejnIOXUxuq3tloN8kzxSopQfIaEn71bbJSbPhPLsl96C7MaJbnKuOGiP8XUMnYW2gT4BV/XlW49ru1++z5VMe/S3Jj7rynH4DonSh5/pdMtiMjzKp93ZPtpL6k3fBcYGTUVcijCIjambyTwiMkcCw7bPJwiZbyJeWmixka+nrUq8OY6PRYXSJGXBNh40i59vzRvmkw5qc7+ow/Vh/l+dWNr8rl2taamVj2zd51KvDlE0AyiAa34OcFwjrLRu5wgYgqwQWyRq+iHzQV3wMib4hey0hd6578tWsVmk+hfPktZVtWB1GTrfUyjaZDmtlo4fMoOUL+5xIh8kfH0ngHA/Xk6P8TINTbBbMyV4BZNzkSdn4ZD7lh95qZSOv8ad72Cp4+T99ogzuRd/yBVrCmv/RLRuuznXIG7b8fDqJPaVs8sB3bXXd7tf72eqoaxOoW/mmcwzITZQaI0FVOlqfYId5LnaeTRfLH/+U3afsbDS2l+vpVW0EX+e5RltyPf3c7B98sNCm+VnaiaTN05MOjJrOb3x8urcVGPJo4hH9LM9t4KDrL05U1pZPdemZku/kR13SvbOV5Aeflq/Zysd8f24CrPO9BkfIv87iIJQJICNpAmUtNSVHIDUBQcBz0ByUgWv83mYKUZaTGdbdDkpYu56D2xKF7P+EbnOMusHHdh4BaDTtK1/5Ss27821If8opp1SDhENL8An62RuWcJQoGgK5KRTfMXDcT5kE1uRJOd3TKJ1ZIoJl7tmblIcQpCwuuuii8sd//MdVMLfn9Z7f+/9PfvKT5Z/8k3/yc8YAh8aLysxWpKwpA+d94hOfqGUj5Dl5nC2BUmWSCMBWNnUiNZ7KZnNqSpLCc08zMk866aRaNgZMv0SRGYlTPixdjx3FJB9GidWjAMSf//mfV0Wj/uVVgKMlSvrCCy+sZdJu1C3DUZ35bA6te1MogonqoiVtkcLDupVNObCi2HBQb85hmNhjxksyKFH5nCx126S+wBAwCu3wN07qSBtohl33fhxrs6+0RXnWbjz79NNPr2Xzf0nZGHryqT5aamXTZlrZMHWOg0HhGr9bdqJNUt7K1vplu1fvZ7dsLZ/6ekvNmGAItDbQ29+693Cd/ik/8ij5Xf0yftxHGXrvUU/MPyEwCYHWzvQT8kYbE8hvqfUTsqL1E45C6yfkvGvIQ33DDEjOln47UXtkkP/f//t/q9xwX+1XYK8Z3fqHGYNeICIv5LDBFrKte08zjegjMtvf5BtdRHaQb805JV8Y0PLJaWqp6Vr9ufXpiXQt58uMJvmcTNe2e3c/OVdkpXy4j0DYIMOdbmz6iOwgz8hXszOUjUOBuXtyNJSNLGiJ3OJQYdLKhmurN86xa+hXAQUvXZAvTsVEqbUVv+MkuNe1Xeh+93e/Jn+79dWu1344Zn/6p39a9SP9OUzi3P7u7/5uDcTTSXSaGT7qX9LmtBW6swXrW/ulr5qz3J6lzXs2RvLpXPK6MbL/mUCD3+1j/Bu/8Rv1GQITk6VWTue0fOon/pbUmT7AqW/toMuJbUQPNl2s3JYltnp3H/czK41tYBBSmxckb+WuD8o/IbAICOhvjkE6jExqOowt3/p502HuIVB09tln14AWfTbTiT5qOoxMIGv05abD2Ify1XSY/t7VYf3kPPu5la35i+xxkwR+/dd/fcL9o8k/trL8/OEf/mENUlqWTO7QewaWJM9ne9MfbP9/8A/+QR3g4QN0E93ddBhbFX/yuVeHkcGCcMPoML5ZK5sJOXi4r7z99m//dpWTk+kwgU9+mIAlPUTP8unIX3nD28EWwo4cFjBk+zvoNkke/Eb30IXu+Yu/+IvVt+Ov4N38XraJstEtLak37OS96Wfs6Sn35vfRP9qDwKUXonk2H3CipJ205F70Q9cmUF6+G3ur+dVdvdOu9UmPnHXWWdX/+eu//uu6Ytb9MXKt+5933nmVP5/kU5/6VNW78ispG//UEnmzwjHVljAS4G9tpelnfrIBDjaE5Fld/YwRvffw8nbS6t/57A/PFC/47PJZ2fpsv0HLetOf/tPl1PLZ9em1R3WGtXqVJuL001vmYw4TmDwSMYcznqyNTkBHJvgIW0qHoOoKXkrjd37nd+oIWDMGCBd/E9qMe8qXo0ZQMr4F5CZLgmItGEnotZGldg2Fydlxb0r4137t16rx335vn4Qh4UNxtMAiR629BIhDx5FVJs+gRCT/p7AJVQquJeVRLocZuIQ6J9r+QG0UsJ3b/fQcAtCSewpWnuRnUPI8ypKDrJz+bkleGTGtbJwcipfy4Axi2C1bE9Lq7j//5/9c68D9JcK41ZfvKGbCupVNOdu57fndTwrMbFBKQ51QKBh/4QtfqAqdsmrK270E2M3ypIib4nI/TCgQ9Sa/ykeBe7bgvHsIwGKpvnrbIuX1H//jf6xKpuVX2fytfPJJWRq1pPy1RUGSZjB0y9T7N0XKEGQAeA7Frx4FZ7RtI8zuywDpl1rZfCqHslHS2Gg/6k/ZKOR+ZcPh3/27f1eVs7JI3X6mbQjqMLjVm7JhO4yibQFUwX8Ouud3+zjjSfkZAsprObV8d5My6bPkhfuoI+2PY9CS8jFi3UdbEThPCoFRCOgH2pH22a+faH//6l/9q9rX+/UTbc+AEXnV+smgAJTgoHPIGwMH5Bs5TuaTk/q8/swJous4T/phb2r6iAxwkB+cIte3GSqT6Vq661/8i38xlK5tZRuka3vz6P/KiTP5MewSQ/qoV9eSG5wJM4bobEFo9UZOeAZZ15LBsn/2z/7Z+7La9+Rbk91YkxdsiGZH9BvIavfzSRdqI9oLO8XfgrstkeF/9Ed/VAfE1Bk7wuBUSxwkg49tWwL5bUHHds5EnwLx2gbdQOc1B4wTJuHF+RaUN3vJjE5yUx3LoyC3g85p6ctf/nIdSMOk6Sxt3OG7Fthsepvd0BzIdo9+n+pNnXCQ1VGT/80e4ug220U/IP/p8JbkReBVGcl3OpItwlE1uErH6h8CygLO9J28Nmew3SefIbAYCLCxmg5rPlVXLtFh+okBt6bDmh2rn9NhJicIADVZOEiHTRfXQTqMzWc7OHK+X9nohH/+z//5+3Jdvrp2LFl46qmnVrk5SIdhg4UgFV+CvGSbCyAKENJfZDBblSwzmEkmGpjrpxcn0mFkuWsENw3Sqaumw7r1ZiVbPx3W5DMdZtBRuVq9DdJhjQ+ZKSkvrmSyySrag3bSdAI9wrcRYO4GbuktMpiMVof0GhlNBptBq821oGorW1c/01OeqSytTXb1M9/nl37pl36ubPI6WeIb0BX0Hp+V3uk+076zf/AHf1DLrH7pnX715hnqWF2btMKe0hbofPpUUmY+FiYYGeTjU7ct7Jq/2fS7Txz0U/qbfeZvE1m0bff3e5tAwp7j77Jj1EXzu1rd++TztvpRn5i1wHHN5AT/tJm86sfR9LM8S2ypz3/+89VOwQCnYfddneCR+XoWCSTAOovwx/3o1rl1agJWwJHT5WiJUu0mQpAyoSx9OjgwDIJhEmEo0NaEGEHOYe06rQSg3wkoiqNfasFJwplgJQQFkziPFAvHgrKSOALd/XiU29ESod3K0j45T7Y8aAqnndv76bnyS8B7rpkbwyTCn5LyjG7eXOs+DAtLMBgUmLXRUr8JcOIjcRq7jqPgq6MlBoQyMdJa2ThAlMowST45URS15wiKMyIpS3mSN0aAe1MsnDNGBsXVFJznqEvl9Fx/C9pRYGbGKA8WlIm26FNiMDlaorS7QT1twD27bVEe2shuu26iT0aXtuh52rkyaTOeoT1TmsrDmFPGiZL2Q/mpS/WjbO6hLSobhay+lE0b1ebVr6MlLLuJU4upZ7d6a85197xBf2ubDC7Gojxh3wy6dq3fHTg0xd5+8+k7wQMyQnBVW2c89hqQ6tPRraPuffJ3CExGgEzW//RFco/s7dUNTe61+5AbZAC95JPc0f8Z2MMk/Uy/0k8k+kaf1cbJNoZyuyfHyfN6ByBcR87SQ5w159MLyiAoRca4H6N9RXUtmdTVz549Smq6juxxLw7DoETOCqD16lqBVXKB7Gj1RveQ+/SnoyWOSzfRR+qLbmnyjW7oysTu+b1/q5vWXsgt5ei1XdgADnXW6rfdh4xu8t/f2ouBvmGSNkPWy6u/PUP7cz0nSxtq9of6lshOf2vb8uK6Xh3cdejpiNae26fneW6v3J0sz8omP/ROm0Hkfo5uEuwlu7X9bmITmt3qd/fRluVffap7usG9sJc3ulh7SQqBxUhA/2iykB1Eh+k/jpbogW5qOqxro9NhXfnZPX+qf5PhBgqlfrKODqPv6DAymQ7T7+kw8qzJWzKMrCAb2Mld369XzvfzF4fRYeQfOSKfAmsGP8lPMoi+8Yktn4SPwFein+nmfjqtq8OUsfmLvTrMM+gw9xmkw5rf0dVhZCDOwybl5DsYLHM/EzMcyttmUKoLdpB7O5Szm+gSuqHrL6o3tjodR6a3NokdX4YP2E1d3eP7bpk8n4z37G5gt3t979+ey+bBk+7pZxNoTw73dv5ESTuj8wQW+UN8UEFjjNgB7u0eBnvNcMaiOyihPWCrPrV/1+in/D19kY5s+ln71Ta0W0c3Ob8leWLneW77bHU0iv7zbH2L/9fqoNen90wMtXlMp5JwM3EMZ/4fOzBp/ARWWm4w/Wxu9/ifnyeuIAFvh/vLv/zLepd//a//dfm93/u9Ce+osxLiDGZCZZhEqREuhFb7FLTp53T2ux/h5nmeS7hOlghOwqtfcKs5re7hb0Lqf//v/10DWWYDElKTCe3ucym5bnmUS3mUq41Wdc/v/o1bCz4RgMN2H/fFknLsFeTKpG6Uy0EpmtXrb2WjPEYpW6un9omnso2S1JtnyptyqkPKQXkbP/XFUHB/PH3fUlMOrUw+Lam02b2RTHulahPO86xhkvsrUyuXz1Zvw1zveQwsQRtGFa4UsU98lEN5esvSe+9Wtm5bNCPJiOTv//7v18CEZ2knw5ZN2+htk/LjGCWpJ/ljEPt7ssTA1t+69eZ8dexa5XMf9T5Z0p67Bs5k51oCZlSWAcEA+qu/+qtqCEx2TX6bPwTMyvwf/+N/1Az/03/6T8uf/MmfTJh57Us/0dZG6Sfd/q/PjNJP9Emylv7wd7dtk2eO7j31y97+oUDkWVdm+//nPve56uSQARzsVrZx6dp+oMk5+VAuMs6AUL/ydK9VL+1QrqZrOSXKRma4r3obtmyY9so3+ZGvYRJd5HnkkfxMlrQHDmPXRlHP8up6+rVb75Pdy2/agLyScfLrWnXfZcS5pbMNuP6n//Sf6nOcI8/D2gi97brVGW7DJmX0TGVUR5MlNglO2kRLrm/tGmdlbfmXD3mSz27e8Bk1mY1EDxiU5EzbtsMWO0khMNsEzFz8n//zf9ZsmKFpG7CJ0mzosIny0vu9AT/yTtLH9fWuf9NPh+mHAktsdIGrcekwMsbRbGufZH6T0/LdZA/5I2BEvtNl3TI1Bk02+2w67M/+7M+qTOQjYzFuHdby5rPJaTLWwRbxnUTONr+m+SX1h5/+06t7BFTpHizoZz5MK1vj172+39+9utn/R9XP6kx78TlZUnf4q8fJUtcHbYy0EfXtWvnjw2gX8ttS02GtDbjWJCGMzBb/N//m39R8+n1Y/eyZvYyaDhxF/7V6Vz909GTJfXFSxlGTNsHHwqufPTTofla/0M8C0AYo+GlWjCaNRuBnrXK063L2PCSgo06ls65IUQklx4qOoLi+3cPIjyUeRtoEVpcsNwR8jiMRet28TMczKUSHZC8kyx2Vx8iY0VqjUeNOrd4osamkbtCDY26vKfcyM1ZAzt9TvfdU8uMaClFAlfFCsZpJwLnzOSjo0H1mK5t7UWKWmjBgjSi713S3j+6zB/2tjI6u0zzomt7fGRMYOXoHA3rPzf9DYKoEunJvqvcY9brWrke9rvf8bh83W8TADXlAZjc9O25d25tH/zeg5xgldeul6VpBVTqJzFbGUWZtjPLsic7lUDmanpzovIm+J98dHBbHiiT3IV+bjDUDiWNplpCZneQnuTkbspN90nTrVMrp+qm0mRXhmWtDYL4S6MrKuVYG+sgxUerqMDMPLcnna7CJlavpsYmun87vyUxHN08rcv9uvTy8PGBsVi5/kT81WzqsW56unO5+lwRvoQAAQABJREFUP8zf3XoxsGebIzNN+R7sm65uGuZ+03FO08/TaRdM1Qft1WH8M8FUK2+XLI8V+N1g6bCTQqaDT7tHt961x5lKs9EGZqos8/m+CbDO59pbnncdliAy2mN0xKif70YJGM0HBEZilM9olKVvS5fvm0dZWj4xVadrrpRb2dSfslGYNtC2/MCyoXEHIaeLiTI5lEkAwjIPBtT/z96bAFtVZff/23lgdmAQRURBFFABQQRUlFkRUNt2iqaHJJ3ETnUnqSRVSaXqn3TlV5WpqpOudHWbsjttm6BJq00ryiggMyiTTIIMMsgkCCgiDvjns5JDv36+93jvce99d/jsqsN9795zz9n7cx5nnfXda6/FLBjOelM0Ht4wsKfaqo6NB1Py3XHsIcfz3db1QHuq5y3l78OM/7/8nfMz9ygeyvibsJUPgey6VoI9IjqE+xsOHPds7mssq8+Et1K9qtn/VcbGhF852dpcXJPs/s+zFoyIwsTJJfUEm61uAlXtAM+p2IFye16tm4CfFjOBSrJhmU+V2TCWQJdDTuWabBjP/uXmLzKxS65QVszgLxL8Uemtun0mLR3PMaQUqG8qqUplmP2/0U/LzV+AAmtuODbZUcjnghhHKDdh5+TLaWxYeZMNoh4nJuKQJS+ErhM1yI2S3D7k7Cl1gZXlIeRkIfk4kTA4aVneolI1mNygWSpC9WQicpkVJ3cTDwFNEdlTjz+xeu/C2Fh2xN8iDzjMRLLcnYc3xmn7MgHEKPJE8X+Xh3ruUfw9lOoEwpdH6DsQ4P8CD/vYI5ZA8cq1zmVkQzGQRlxj+SX3ACaQGDP3be4DpT7WzNZij/g/W062Nhd/Ozgh2GxW0cAIcRB7TY7V+uasy0U/SvUYPMeRpoe/M6KI+Psq9f8zpXot7PeXCWDDMp+qnG0Y///wN7iH8f8RG0aUX7nYMGwX9plX8lByj+aZk2CIUm7YHq4XY8Pn57kD28P1Y/VapTd8DRgx8QkjfOjMPjOBYKudAM822Gfy3OPnEjBUDhpL7SPO7ycKrPnlm/ej83DKsmv+Y7AUmzyQ3ExIvI2QVU6RAUS7MR7Gxbgxlk0VDZnLC8u4srFxQ2NsLDVsitQAuR4X14slLdn1KhcHNLteRKvxQMrfYqHSVOTyGhXiWDgpGO0smT9LUDHa3Lcauny5EP31HI0nwP8Dqt7SePgn+hGnDYeV+zbRQeXSsnsAkyrZPbtcHuCzsSGAZfe3crC1ufzbyxhx34cRf+MKhbUTZpkmkxJEFFFVGseXAAEKkbjyo3ZuflJYAjybZDYMARIbhk9FBGS52TDIch/DhmX3+XwuXS7klWRcbNkkDj5VOdiwbFy84iPy7IG4Wur+Yi7/NjJG+J+sdOX/dOfj/39LXVzPJaPqxyK3L4X6sM+syMFvJ/8qhZ8NHKpOq36/W+SqfpyKdi8i6F5//fX0/PPPx6wDM1ijRo2Kh1ZC4stlRgsBmRkVkjdz8+Q/P6kRykFAzsbGwxwCBE4ar6U8NsZEQ2Qh2osxsRSwHAQWxsbG2IjG5G+xXMaWjxsdub3IwTt58uSI+OX/7ciRI2NDkCv1JdX5YFaqxyR9C9eaYnaspuAexrWmqB32qFwilvn/z/997BE/c38rN3tUjrY2F/+vuN40VmhgA3Da2LBtPJvYaiZANPuyZcvSlClTYrKN/y+Iq3fffXdMKCuy1szNdwtLgOXEmQ0jBRT/r7Fht9xyS+rbt2/ZLMPObBj3MJo2rLB/Z409G9eN6GPsM/dQ/I9S9xcby6Km72X2mVWGMIIPz53a55po/fo97nvkq8VPW7duXfAipd/o0aMjgIgcv7aGEVBgbRivotubGwjLbhFamYHg5kJUAFEVRAtyAy6XRug/IitODOMqJ2eGsSFEIkgg1pVLY0yMrdzyrPH/jL9FXvlbLGUxPN9/ayzRQmxjWRP3KP4WOh+fTSbKnof6cpkEyjfHUjg+0cpcb+wRtgkRkqhO7BGz4eV0rfm/z/2Nxt90udmjcrW1ufp/xN82jHDcyuk5K1d8qh+HFVYIVu+++24I0/yf4b5A9BV2oFTTIVUfp7+XNoHqNozn16o2jL/bcmnasNK8kuXqL+byamCb2bDN2ueTk2WVYWafuQdynyP6F18N+0wBNVvDCCiwNoxX0e/NjZfIChoPrAo/RX/J7KAEypoA+ZDYmEUuJ+ekrC9ajgaHA0chIF65/tqjHIH1MBIoMQJEXfFsynNpOU20lNhlsLsNJKANayAwd5eABEqOgPY595dMgTX3TJv0iDwMILLScGbLKaqmScF6cglIoFEEuB+xcT9SYGsUwpL9kvaoZC+dHZdATgloB3KK04MViIA2rECgPY0EJNBkBLTPuUevwJp7ph5RAhKQgAQkIAEJSEACEpCABCQgAQlIQAISqBACp1fIOB2mBCQgAQlIQAISkIAEJCABCUhAAhKQgAQkIIGcE1BgzTlSDygBCUhAAhKQgAQkIAEJSEACEpCABCQgAQlUCgEF1kq50o5TAhKQgAQkIAEJSEACEpCABCQgAQlIQAISyDkBBdacI/WAEpCABCQgAQlIQAISkIAEJCABCUhAAhKQQKUQUGCtlCvtOCUgAQlIQAISkIAEJCABCUhAAhKQgAQkIIGcE1BgzTlSDygBCUhAAhKQgAQkIAEJSEACEpCABCQgAQlUCgEF1kq50o5TAhKQgAQkIAEJSEACEpCABCQgAQlIQAISyDkBBdacI/WAEpCABCQgAQlIQAISkIAEJCABCUhAAhKQQKUQUGCtlCvtOCUgAQlIQAISkIAEJCABCUhAAhKQgAQkIIGcE1BgzTlSDygBCUhAAhKQgAQkIAEJSEACEpCABCQgAQlUCgEF1kq50o5TAhKQgAQkIAEJSEACEpCABCQgAQlIQAISyDkBBdacI/WAEpCABCQgAQlIQAISkIAEJCABCUhAAhKQQKUQUGCtlCvtOCUgAQlIQAISkIAEJCABCUhAAhKQgAQkIIGcE1BgzTlSDygBCUhAAhKQgAQkIAEJSEACEpCABCQgAQlUCgEF1kq50o5TAhKQgAQkIAEJSEACEpCABCQgAQlIQAISyDkBBdacI/WAEpCABCQgAQlIQAISkIAEJCABCUhAAhKQQKUQUGCtlCvtOCUgAQlIQAISkIAEJCABCUhAAhKQgAQkIIGcE1BgzTlSDygBCUhAAhKQgAQkIAEJSEACEpCABCQgAQlUCgEF1kq50o5TAhKQgAQkIAEJSEACEpCABCQgAQlIQAISyDkBBdacI/WAEpCABCQgAQlIQAISkIAEJCABCUhAAhKQQKUQUGCtlCvtOCUgAQlIQAISkIAEJCABCUhAAhKQgAQkIIGcE1BgzTlSDygBCUhAAhKQgAQkIAEJSEACEpCABCQgAQlUCgEF1kq50o5TAhKQgAQkIAEJSEACEpCABCQgAQlIQAISyDkBBdacI/WAEpCABCQgAQlIQAISkIAEJCABCUhAAhKQQKUQUGCtlCvtOCUgAQlIQAISkIAEJCABCUhAAhKQgAQkIIGcE1BgzTlSDygBCUhAAhKQgAQkIAEJSEACEpCABCQgAQlUCoEzK2WgjlMChSLw6aefpr1796YjR4586ZRnnnlmOuecc1Lr1q3TueeeG58fO3YsffbZZ+mjjz5KH3/8cTp69GjiPbbTTz89nX322bHv+eefHz/z3mmnnfalYzf1G/T3888/T4cPH46xZ+OgX2eccUb0nbE3b948fs5nf+nLJ598kj788MN08ODBWk9FXy6++OLgXOtOfiABCUigCQi89957Nd6/MrvQsmXL1KJFi3r1jPshNob7M/dH7nuZDarXAQq8E/3FhtBnXrGrNGwJdgR7SP/POuusvN6/6Qe2/MCBA2Gna8KAjS52njX12/ckIAEJlAsBbMWePXvCD6k+Ju7R5513Xvhe+GFZ++KLL8Ie8l02fDD8GN7H1mBfsDfNmjWL72ffK5ZX+smGfcz6j81iDDR8RfqPrcRmMh7Gla/GeTk/fhfPGrW1Nm3apAsuuKC2j31fAiVP4Nd3mZIfigOQQHEQ2LdvX/rHf/zHtHTp0i91qF27dqlbt27pwQcfTD179ozPMUY40gsXLkxvvvlmeuutt8KZ432MYpcuXVKPHj3STTfdlK644op4QPjSgYvgDQz8/v3706xZs2Ls69atOyEyIyjTd8Zx2223xc/57DIO+TvvvJOmT5+e/ud//qfWUw0dOjT9xV/8RTyA1LqTH0hAAhJoAgLPPPNMjfcvhNXLL788jR07No0YMaJePdu1a1datGhRmj17djg+3Pe6d+9er+82xU67d+8OW0ifsSVbt26NbuCY0e9+/fql3r17J2wqzm++2s6dO9Prr7+enn322USfamrYtj//8z9P1157bU0f+54EJCABCeSZAP7T3//93yd8sOoNe4nNeOCBB1Lbtm1PfExwC6Lq/Pnz04IFC8LW4Mfgf+G3XHLJJWFvBg8enPr06XPie8X0A/3fvn17+JBr1qxJGzduDB8S4RVx9eqrr07XXXddGjBgQOrQoUNq1apV3rp/6NChtGXLlvTCCy/Es0ZtJ/r617+evva1r9X2se9LoOQJKLCW/CV0AMVGAHFv5cqV6bXXXouu3Xjjjalr164xc3jhhReGccfoMdPHtmrVqvTGG2+kefPmpbVr14Zx+uCDD2JGklnXDRs2pE2bNoUBRWTFSOJgF0v0EdFQPKQgaL799tsJA4vD26lTp5jF5GGF93kAYBzMHjNuHniYTc1HyyK8mCHF+aUhABPRirO+fv36eK9jx44xex2/+I8EJCCBIiKAo5TZkauuuiphS7h/cv/HUSIC/2SNey3RJKtXr04vvvhiOGHYDmxMMTaiXnCQ58yZkxYvXhyiJvYFRzdr27Zti6jWd99998SEHSzysbIDGwXvSy+9NCKAsHfYEfqwYsWKsCv0t1h5Zsx8lYAEJFDOBN5///2wb0yK4YPgL2EnidjEfuAPZNGbiI/YFSJeCWphAhJ/q3379uFbYTOxL0zsYYfZF1vQuXPneq8ayTdrbDv+JsE8S5YsCR8SXwvbhD3CVuFrbjkueLLhg8Gkb9++eVtJCF+eLxCxM9+LFSD0B3sJUxrBLTYJlDMBBdZyvrqOrSgIEK36zW9+M5yzbGkKjmC2pIMoy6effjoMYPUlFcxMYjARJufOnRvCIIbryiuvLBqBFSOPAcV4MgN86623pmHDhoWAiuO5bNmy9KMf/ShE5OXLl8c14Ts8+ORLYOVBCePOwxAz1jScYh4yfvjDH54QWOMD/5GABCRQ5AS4r/7DP/xDOI7c32j1ERS51+IkEglKVAn3wV69ehXtaFmKj8NI9C5RRYx7zJgxafz48dFnJhyfe+65sDVTpkwJu0qahPqIzY0ZNHaKDYcwc8qZTJw4cWI45thxmwQkIAEJFA+Biy66KH3nO99Jt99+eyztJ+iCltlM7uX4V4in3MtZDcGKEAI/EFMRAvFbJk+eHEEw7Mv7X/nKV4pGYMX2ENBC/3/1q1+Ff0PkbdVG2gCiWhGRX3311fTQQw+F+Mk4s+eIqvuf6s/YYlaYEDULYxop8/Bjv/e9750QWE/1PH5fAsVOQIG12K+Q/St5AszmkfsGcTUz8gxq8+bN6eWXXw7n9/rrrw8HkghXDBRGEqNEpCUCJZGtiJgIrRj8u+++O3K+FQMcRFTSGzA+HGFETfLRMZNJ5A+G9uGHHw5j/vzzz8fDCmkPxo0bl7fuZw9R2SsnIv/SZZddVrQpFvIGwwNLQAIlT4DJKETE6nakroHhgBE5gq0hqgUHsdgbYjDiKiImtgSnd+DAgScEVGzH/fffHzYS+zlp0qQY0m//9m8Hm1yPL7MhVV9JS8BW1Z7n+rweTwISkIAEGkeAezPP/Phf+CLZ/Ts7GpGfpH7Bfxk9enT4BkS5sj/CIJNq+DN89z/+4z/C92KSkuCRYmmsSsEfJHIXG/nII4+Ef8MY8CGJ5EVYZYUk9hSxFV+SoB58SFIg5LplnLNXjk9KAgTdfKbyyfU4PJ4ETpWAAuupEvT7EjgJARziqjOFGG8cX4wfaQFYknLLLbekm2++OZbVY/SYLeVzxFWWeDBLyYPAjh070qzjOU7JBcQyj/o0HGy+S8PwMrNLn3LVMOTkkGUJK33ioSY7Pj/z0ELeVR4GWLLCGNiIrCpk4xqwREgjX0jqnksCEsgFARw9bEFDGvd+7AjLHbEpWURJQ46R7cv3sSPYLvrCZCATh7lq2Ab6SYQq6QG4T5M3jiWNTNJljRUcbNmKCV6xaXfeeWciPys2Lp8Nxx2HkclDBdZ8kvbYEpCABBpHAIGPZ/7aVslhC1ktgS0jNyuTl1XtK34Y7+PbTJgwIV4JcCHQpb6N9Gj4bjQCZ7CZuWjYSuwwk5Gs8iBilCAdfMgsJzn2esvxFXusGmSs/M5YmGxlNSSCbKEafiA2Ot+2uVDj8TwSqA+B3Kks9Tmb+0hAAiEy4kiSDgDDg1GkUAkOZSbEYujJD4rDyAMAG8shcZaZjSS6tb6Nh4Jp06bF7uRFHTlyZE6TnPMgQl5Y+orDXd3p5AFHYbO+V8v9JCABCeSGAHnlKPyBc0f0PnaksQ1xFRuE04jdGjVqVKSqaezxqn8Ph5G+kr8cp/Saa66JnLP0vaZGdCsOMBE8OJJEI91www0xzpr29z0JSEACEpAABPBLSJWDEIuNyYJCMjq8zyQa/k1j7SbiZlbsGNuUq+hXVqIQtYrYS99Jg4AfiQ/JuLLoXVZ7EImb5WLFVmaBO3zfJgEJ5I+AAmv+2HpkCdRKAANIZOfw4cMjSocl9VUbxh2Rla1Hjx7hcJLEnGWe5NDDGa1vY2nISy+9FLuTZ4icdrmsIomzTQEQxOGaHkSIcMU5R1Rm3J2PO8bkR61p3/qOyf0kIAEJSKBmAkzeEd2CUImDxWoBokewK41tOGSzjq+e4LjYD2wJucBz1XAaEUqJsMG+YROxFbVFvRARgx3B7hCZwwoJRGQ2mwQkIAEJSKA2AvgfrITAJiKuVg8MIeqTyUR8F+wnKw2Z9GvICjhsUuZ7cb5cCayMif5i/5jopF/Vo2MZF74ZG+IuQTmsDMGO83xQ6BWEtV0H35dAuRJQYC3XK+u4ipYAhg+nEePIVpsDmQ0A8RJnFqcWo8oMZUPESWYsZ8+eHYfje+QeymWjP8zy1tR4SMGYr1u3Lqp1MtuKsWfpJ9+zSUACEpBA7gggTuIYEg2KuMoEHcLlqTbSDTDJx70ccRPHM5cNgZUJRCod4/yxIoJJyGxVR/VzYXNY5YH9ZPKOfrGSwiYBCUhAAhKoiwC+FBGqNTX8FmxQNuGH30SaGpbV1+br1HQcCmhlvhd2OFeNviOcIqyynSxVD5GsPXv2jO8wFmwqx7BJQAL5I6DAmj+2HlkCNRJAYCUyNRMYTyaWYgwxphhEZinJdVo94rXGExXBm4i5VONkaQoOOUtZKFrCchYY2CQgAQlIIHcEcAoRKXG6cAZ5PZmNyd3ZG38kljHu2bMnIm34ORNYa7MTjI3JRwRWJhFZqYEIbJOABCQgAQk0lgCTlNgSamQQhUpQCKnVKIZVDL4XviB2DyGYVj29QfVxY0PZH9+TSFzytTZEKK5+PH+XgAROTkCB9eSM3EMCOSWAkTuZQax6QiJAWQLJUnsihxAnMZLF2DD4RCLRZ5aikP+VHEQkVacxC4xhJ5LVGdRivIL2SQISKEUCOIVMaLGEf/fu3YmoFSbkmMgrhXstoiqRt2zYEYqOILJmE5HVrwnCMbm9+Rx7g8BK+hybBCQgAQlIoCEE8Fuwn/gt2FCKLc76v5Q4l19+efgtrCKszR415Fynui/2vCE2HbEYu8oYWfXRv3//sK2n2g+/LwEJ1E5AgbV2Nn4igaIgsGPHjij8gZEkZxA5VFn6WYwNx5hqlTygLFu2LE2cODGqXPLQgjNMNO769etDIKbyZW3LP4txbPZJAhKQQLESyKoEYy8QHHEKsRelkmsN20HOWDZ+xjbUtfQx+5zoXMb70UcfKbAW6x+n/ZKABCRQxAQIYCFPKQEhU6dOTZMmTQqxFUEVe/T222/HRpBIqfkt5DVfu3Zt+GbkKCfVQfWcrUV8aeyaBEqSgAJrSV42O11JBFg2iWDZrVu3iAAllw5iZTE2opCYBeaVCCTyDjFrSmVooot4SPnFL34RSzrvu+++1Pl4EZNiHUsx8rVPEpCABGoiQP7S6dOnx+QVTiCrBBAfS0VgZUw4smw0Vnqw1dayz3nNvpd9t7bv+L4EJCABCUigOgEEVjbSsVG8kUCWFStWpF27dkWasylTppzwW0gZwCrCuuxT9eM35e+kaSOCtU+fPpGnnLytPB/YJCCB/BFQYM0fW48sgVMikC1ZYVaVCFAKXWEgiUyqmmKAh4IjR47E7CuRPNUb+fiyhsjJTOaBAweyt37jlTw9pCFAHK0t991vfKGGX3ByeUhhKQoRSBQq4ZWiK6QMoJIlUUqdOnWKvECkDCD6ij4x1lNtzMyytJSxlELuwVMdr9+XgAQqlwB2giJPTMJxb8d5YmKLe3BDGsfhPowjhr2p3pgc43Ma+xIVgwNaU8PxJFcdSyq5DzdkOWNNx6vve1kUL3YEu3gqDQeUMWAHSy1i6VTG7XclIAEJVBoB/Bb8KlZ9IKAS6UkRrDfeeCNW3bEijzQ0CKv4LO3atYtgElYWZincqjOr6s+wT232ku9l9gZ7mas0BKQNwiYSpIM9HDx4cLrpppsidznnzGw+feMZ4lQavhb2EmaMoVTE51MZs9+VQF0EFFjrouNnEmhCAoilOLEYaQTKW265JSpBVhcNKR6FY/30009HpGj1Llc18uz3Z3/2Z7WKp4i3jz76aIi5/NzQRt9wqps1axZpDDDgRNzeeOON6aWXXkoTJkyIWWAMPtFWiLBXXHFFCK/MEPP5qbbx48enu+++O3IQOkt7qjT9vgQkUMwEsBNUKsZJIuqG+3ZWFLEh/WbyDXvz3HPPRVHC6t/lc6J5aPz8r//6r+mpp56qvlv8jqP6W7/1W2nYsGFxf2/sZF2NB6/jTYpdYVewhVlf69i9zo/69esXtpBoJpxpmwQkIAEJlCcBgkoQBhEJ8VsQJK+99tq0cOHC9NOf/jRyspKGZtGiRSG+IlQSGcrvP//5z0PErE6GVSVZmzlzZtq4cWP265deOR6+F7nTEXlz0QhaIaiFCVMCTyjUhb+VNQJd+PzZZ59N9O9UGvyw+TyDMIaqQUCncly/K4FSJaDAWqpXzn6XLQFmUjHwGEfylSJWMvNIioCaltPzIMC+b731Vsy21gWG2dZVq1bVuguOM446y/wb05i1JNqnasQPgitRsfSRY5PfCEd4+fLladSoUZFOgAcXIm2ZLT7VRpQvokMpLY091TH7fQlIoPIIUMwKUZTIGhwcIm9wzqpPwtWHDPdL7v9btmw56X0Y+1SXs0gEzh133BF5Ueu7bB/bQfQLG3bqZI0+YKfoNwJuNpnG5B2rJYjoPZVGlBIOKFFANglIQAISKF8C2KyqkaPYLWwAr/grpD0jihU/C5vL+1l0KPam6krBmigR6FI12KX6Pvh2ubI39BX7iL+1Zs2aeCZgkhDhE/uaNfbB5iOynqrvhZ83dOjQWE1ZX5uf9cNXCZQjAQXWcryqjqlkCWCY2Fjyj3HEkBNJg7NaW8MxxaHGyaxpWShGNHMSWaqJ+Fnb8g1mcDlWbZ/X1oe63ud49Ou6666LBxgMfpYwnuWoNPrFw01N/a/r2DV9xnEKtSS1pvP7ngQkIIFCEEAMJYIGW8EkFssaufcxYZU1JuBwBLEBmeODA8Z72Bn2z6JN+Bn7UNN9mO9wrOwY7FebkMs9mGM2xI5wLMRhIoiYION8iKf0qabj0BfGyT5MQhK5i/PIebFjNY0hY1KfV8ZX27nr8333kYAEJCCB0iSAzcGn6ny8TsS9994bE3aLFy9ORKUiuNLYpy57g83F/6Jh37AptbVc2RvsMzYRG4qgi3g6bty4WEVY07nrsvk17V/be9jchtr82o7l+xIoBwIKrOVwFR1D2RDAGOP0MiNKxCczgpdeemmd42PpB1Gbf/mXfxmOdvWdWTL55JNPxtvk5/vWt74VTmz1/fidKKDu3bvX+nlN36nve8xwIgBUrwzNQwzLVh588MEYR32PV9t+LOnkeNXPU9v+vi8BCUigFAns2LEjBFaiSWfNmhWVj6uPA6GSyBgibLKVCfz8ve99L3Xt2jVyYY8ZMyZdddVVsUrid3/3d9OIESOqHyZWU/z4xz+OZZHYCexI3759v7Qfb+B4Yke431eNCqpx5/97E0cvE1iJyCU6l3Fh/2q6l7PskXEgFGcCK6s8evXqFd+pKjLXdd7aPiPihzFgt2wSkIAEJFB5BJiow44RzVq9UV+CNDjYKPy26u1nP/vZiXQ7BMl84xvfqL7Lid/J7UpxylO1N4i6+I74kPiTFBOm/zU1JiQ55+OPPx4ibE371Pc97DzHIu1bbROv9T2W+0mgHAgosJbDVXQMJU8gm3Uk0ThLHFm2gdOIIMpr9Ub0Do4zs558zlabEeWYmcCK0zh27NiI9ql+zHz/jgFmGSuznPSbBwkcZ35HJM6E4nz3w+NLQAISKAcC3E+5pyOcYjNqKk6VjbNqdAkRLoiuOGIUBMQpY6KL6FG2m2++OfvaidcFCxZEjmzyzrEvudbIdZ2rhlPWoUOH2PiZ6BuEViJzaxJY6XsmsMKBCFYcXkRWNpsEJCABCUjgVAhgN/Gv8FmwS4iS2TJ7fBg2hMWaGvlbJ0+eHB8xmUkQSb5atuKDVYFMTiL4YjtZOcjkZdWGv4kPyUQok5o8Q9gkIIHcElBgzS1PjyaBRhHAOGIQcRjJkUpEaufjS1NYdlFTI0cdBhJREsNfCg0nnjEyq8ps8PXXXx8VOUuh7/ZRAhKQQLERGDRoUBQk5N6KDamrvfLKK+lP//RPI78pETd//dd/HXaGia8sf2ld38/3ZzivTBKy8TOOImIuEak1NQRWIniJYEUkRmAthnHU1Fffk4AEJCCB0iPAZCSrIbCx+FqsaiCXabG1rJ9bjqcNwi727t07JisRUas3nhUIvKHhQ2JvbRKQQG4JKLDmlqdHk0CDCWC8yZfDhgHv0aNHzCjWFLXDvjiWGzZsCINPftZSEVgRj8lhtG/fvijWRZTUydIfNBimX5CABCRQIQSy1Qv1GS6OVBbJgqhK5AoRo8XSiBQiTQGrNujrrl27ohAiInJNjVx4S5cuDYGVZYnYE5ZZ2iQgAQlIQAK5IMCqEIpbUTSRdAFM+LFCoibhMhfna+gxEEur+pBMMhJRy0qUmvKQE5yDD4YPyef4kAqsDaXu/hI4OQEF1pMzcg8J5IUAM45EczLbyIbAytITnEVmF7MZxqonz6J2WDrJMk2Wf+SrZVG1mfjLebLlMhjmzHgTSctyU8bDQweCL30j+pbfGSMPAOvWrUuvvvpqRLESnXvTTTfFWPPVf48rAQlIQAJNT4CVC2ykMcjsBOIwk4jYCRw8bEunTp0iQggHkejVlStXhm1kCWM24UgkETYJB5HVHtgahFkidhCObRKQgAQkIIG6CODf4LdkecmxHfgu+DXYI5bR47dsOR4ROnPmzChsxYQkdobl/vkUWDOfCTuH3eRc2DlsJv3LgmqwpfQ/y1nO/girTJ6SYoeteuM9/E2+Q0qd2vKoV/+ev0tAAg0joMDaMF7uLYGcEcB4E6Xzi1/8Ik2dOjWcRiKMaksLwIlxLjGoFCEhaTpGN1+Nc73xxhtp3rx5afbs2fHAgeG+7bbbYtaTJf40IlMnTpwYgjAPKYinRCHxOQ8qCMXkImKMr7/+eswA33LLLfE5kUo2CUhAAhIoXwKIoeRwxY7g4GG3yOFK1OkNN9zwG0v7KVD4zW9+M+zimjVr0osvvhi2EZtBw95wHI7H5B62cPDgwSeic8uXoiOTgAQkIIFcEMD/mjRpUkzUcTxW0zFRhz0ihRniJZGriKtsiKvYmkIEhpAeh9V+r732WlqyZEnYS0RdfC/Sx2UpCrLilU8//XT4WAisiK+ZAFsTJ2wmftmdd94ZY+ZnmwQkkHsCCqy5Z+oRJVAnAWZOEUk3bdoUEZ3Tp08P48h7zJrW1YjyQYAdN25cwhHNp8DKA8bOnTujGiWGnllVkqbzYEIfMoGVHHhEG61fvz7EVKKNqGBJ9BEPKoyXsWLImS3FUeaVmda6HgTq4uBnEpCABCRQGgRYksgKBkRRlvZjt7Aj2Aac2qq5U7ExCK8Isdg7ij7OmjUrIm4YLcdau3ZtCKqItEOHDo1Ju3xGFJUGZXspAQlIQAL1IYDtwbYwiUegC8EerB5kVQSpZiicuHHjxlh1kQWMDBgwIAJIiCTNZ8OnIsIUgRfbh5+05XgkLStA8KnIN84+9J1VgWz4XPiQJ2sEweCjXXDBBZHvPEsbdLLv+bkEJNAwAgqsDePl3hI4ZQKIqER1Eh36gx/8IERMjGV9Go4pDigzmNksZn2+19h9MNiZ8EtEK9UpEYRZyvlbv/VbcVgMNsYah3n58uXxmi0DZaknM688EJDrB8eZKFgeXmwSkIAEJFD+BLAdLHXEqcU2YCuI0MF23Hvvvb+RC5ZUANiXsWPHpiuvvDJRnAsHEmeThhjL56NHj44oHHKWuxKi/P+GHKEEJCCBXBFgQq5Vq1ZxuLfeeivSBRDdSSOa9dprrw0BkpyrDz30UPxMMcVCTOThI2IvsZs0+kWQCv5X//79I0iF3LBz585N3//+96NwZbZvfKGOf0gxgMCKbaWgpE0CEsgPAQXW/HD1qBKolQAGmlyrFO/4u7/7uzCkOJ31aTikRI/eeOON9Tb0RPk88cQTcXhy7iCG1qcxa4ogytKYu+66K/rJTO/kyZN/I/IUwRRnl4hWoo6IdMWRpjHbykMMgir7sdH/Qjyk1GeM7iMBCUigEgiQO+5f/uVfwlnDBnQ+nsqlIY0Jvb/927+NIovcwxuS/xtnFeGUJY7kEcdhnDJlSqxqqG4Lst8RTXFusSGkw3nvvfeiu9hARFacQ4RWbEv2nYaMx30lIAEJSKAyCWDD8MEI/mAyD5uU+WFEqCKmYluwlfhAvNfQaM8HHnggihZDGBtY38Y5SUfQvXv39PDDD0ffWLXBZCOrOtiwiyNHjoy+Ia5mPtfJzoFfh09GVK5282S0/FwCjSegwNp4dn5TAo0igJHG2SS6ky3fDUPK1tCGESf6lI1GYnSW/bNUhSjarGGsyVvEZpOABCQggeIjQEoZtsY2ol7uv//+Rn2diCA2Gkv8SRNAWhnEVgTTmhrOLZtRNjXR8T0JSEACEmgsAexOvn0wAlTYGtqYQGQCkw3xFHuJLSSXOcE5pFvjd/2uhpJ1fwkUjoACa+FYeyYJlDQBxNX58+fHQwnLMm0SkIAEJCCBhhDYunVrev755yMiiKgeHEWbBCQgAQlIQAK/SYAUbaQC2Lx5c6wAOZVJ0t88sr9JQAL5JKDAmk+6HlsCxwmQb5WoHWYea4vWKUZQ5AFi9pS0ACR+ZxaVZZs4xRQmKbXGgwpVNz/44INS67r9lYAEKpwAeUu5F7NigDxqpdSwI9x3sSM4ivSf5Y9EELGao5QaSzGx6eTAq++yzFIan32VgAQkUOoESFW2f//+uFfje5XScnjsCjaTIlukBmClBytAWIlYiis6Dh8+HOnjyCtrk0ClEFBgrZQr7TibjABLIamgjGEkjw9L73Eqi11sRWAlLxERRy+//HIsL2VJCrn86pvHtamg03cesBBVM6NOBU7GQuVQmwQkIIFSIoCzuHr16rAj3H9ZJohQSS65Ym/cixEk58yZE/fkPn36xPLGhuaBbYpx0necXQRuHF9+Z7IRW5Ll7GuKfnlOCUhAAhKomQC+C5N5b7/9dthM/C7yj2IzsZ3F3LAz+C6IqxQVxu+ixgWvjKGYW2Yj8bu4BjRqc3AtmJi0SaBSCCiwVsqVdpxNRmDSpElp+fLlIaq2b98+oj/vueeeRuVFLeQgmPHFeSdalSqaJFW/+OKLS2JJJw8nOMCzZ89Ov/rVrwIbjjFOMjlkbRKQgARKicCSJUvS3/zN34SDiMBKgadRo0al22+/veiHgVOI7bjzzjsTk1/kc8WelELbuXNn2O8XXnghHEX6n0UTf/zxx6UwBPsoAQlIoKIIEPX5s5/9LOF/IaoS4MLE3vjx48MWFTMM6nTQZ/qLraQ4MHUvEImLvbFKkAnIl156KVLK0V/8MaJYySFrk0ClECj+/62VciUcZ9kQwDAOHjw4lnJWHxQGktnTUliuQh/pL2kBSLpOxG2xz55W5U3/YZ1FCvPKtWE8ffv2jV379evX4MqgVc/hzxKQgATyRaBXr16JybjqjXsZ9+aGVjWufpxC/c59mNUbpAXgvnzOOeeUhA2ED/2FM7wzW4IdRCBmPDScd6pO2yQgAQlIoGkIIEaOHj06VktU7wH37lLxvbA3bB07dgxxFXuT2Z7q4yq237GXbFXtJX0n1zrXJ2uZ7cx+91UC5UbgtOOz8V+U26AcjwQkIAEJSEACEpCABCQgAQlIQAISkIAEJCCBQhA4vRAn8RwSkIAEJCABCUhAAhKQgAQkIAEJSEACEpCABMqRgAJrOV5VxyQBCUhAAhKQgAQkIAEJSEACEpCABCQgAQkUhIACa0EwexIJSEACEpCABCQgAQlIQAISkIAEJCABCUigHAkosJbjVXVMEpCABCQgAQlIQAISkIAEJCABCUhAAhKQQEEIKLAWBLMnkYAEJCABCUhAAhKQgAQkIAEJSEACEpCABMqRgAJrOV5VxyQBCUhAAhKQgAQkIAEJSEACEpCABCQgAQkUhIACa0EwexIJSEACEpCABCQgAQlIQAISkIAEJCABCUigHAkosJbjVXVMEpCABCQgAQlIQAISkIAEJCABCUhAAhKQQEEIKLAWBLMnkYAEJCABCUhAAhKQgAQkIAEJSEACEpCABMqRgAJrOV5VxyQBCUhAAhKQgAQkIAEJSEACEpCABCQgAQkUhIACa0EwexIJSEACEpCABCQgAQlIQAISkIAEJCABCUigHAkosJbjVXVMEpCABCQgAQlIQAISkIAEJCABCUhAAhKQQEEIKLAWBLMnkYAEJCABCUhAAhKQgAQkIAEJSEACEpCABMqRgAJrOV5VxyQBCUhAAhKQgAQkIAEJSEACEpCABCQgAQkUhIACa0EwexIJSEACEpCABCQgAQlIQAISkIAEJCABCUigHAkosJbjVXVMEpCABCQgAQlIQAISkIAEJCABCUhAAhKQQEEIKLAWBLMnkYAEJCABCUhAAhKQgAQkIAEJSEACEpCABMqRgAJrOV5VxyQBCUhAAhKQgAQkIAEJSEACEpCABCQgAQkUhIACa0EwexIJSEACEpCABCQgAQlIQAISkIAEJCABCUigHAkosJbjVXVMEpCABCQgAQlIQAISkIAEJCABCUhAAhKQQEEIKLAWBLMnkYAEJCABCUhAAhKQgAQkIAEJSEACEpCABMqRgAJrOV5VxyQBCUhAAhKQgAQkIAEJSEACEpCABCQgAQkUhIACa0EwexIJSEACEpCABCQgAQlIQAISkIAEJCABCUigHAkosJbjVXVMEpCABCQgAQlIQAISkIAEJCABCUhAAhKQQEEIKLAWBLMnkYAEJCABCUhAAhKQgAQkIAEJSEACEpCABMqRgAJrOV5VxyQBCUhAAhKQgAQkIAEJSEACEpCABCQgAQkUhIACa0EwexIJSEACEpCABCQgAQlIQAISkIAEJCABCUigHAkosJbjVXVMEpCABCQgAQlIQAISkIAEJCABCUhAAhKQQEEIKLAWBLMnkYAEJCABCUhAAhKQgAQkIAEJSEACEpCABMqRwJnlOCjHVJ4EPv3007R37960evXqNG/evPTFF1+U50AdlQT+j8DZZ5+dunbtmrp375569OiRzjjjDNlIQAISkIAEJCABCUigYASmT5+e5syZU7DzeSIJNDWBIUOGpNtvv72pu+H5S5CAAmsJXrRK7fJnn32Wdu7cmWbMmJH+6Z/+KX3++eeVisJxVwiB5s2bpzvvvDONGzcuXXPNNQqsFXLdHaYEJCABCUhAAhIoFgIzZ85M/+///b9i6Y79kEDeCZx++ukKrHmnXJ4nMEVAeV5XRyUBCUhAAhKQgAQkIAEJSEACEpCABCQgAQkUgIARrAWA7CnyQ6Bly5apQ4cOEdnXrFmz/JzEo0qgwAQ+/PDDtHXr1rR9+/Z05MiRAp/d00lAAhLILQFWn3Av4752+PDhdOaZZ6YLL7wwtW7dOmG7+d0mAQlIQAKlQaBLly7p5ptvLo3O2ksJ1JPAsWPH0oIFC9KWLVvq+Q13k0DNBHyqrZmL75YAAcTVoUOHpj/+4z9OnTp1KoEe20UJnJzApk2b0vPPP58mTZqUVq5cefIvuIcEJCCBIiZA/vQ9e/akl19+Oe3YsSOdd955qW/fvqlnz57p8ssvV2At4mtn1yQgAQlUJzB48OD0xBNPpNNOO636R/4ugZIl8PHHH6dvfetbCqwlewWLp+MKrMVzLexJAwlg2Cn6c9ZZZyWKAdkkUA4E+Hvm75rcPzYJSEACpUqAPOlHjx6NwihTp05NS5YsSfv37w9BlckjivfddtttIbQSEWWTgAQkIIHiJ8DzKX6XAmvxXyt7WH8CRLDqe9Wfl3vWTkCBtXY2fiIBCUhAAhKQgAQk0AACX3zxRcJROXDgQCIin8KUEydODHEV0fWcc85J77zzTlq/fn0iJQoibIsWLRJpf/jMJgEJSEACEpCABCQggVIkoMBailfNPktAAhKQgAQkIIEiJcBSu7feeis9+eSTafHixWn37t3p4osvTu3bt08dO3ZMy5YtS9u2bUsvvPBC5GUlcr9///7pkksuKdIR2S0JSEACEpCABCQgAQnUTUCBtW4+fioBCUhAAhKQgAQkUA8CFLRCXJ0/f35ErvJ68ODBEFZHjhwZRSnbtm0bxa3mzp0bxfxIHcBSU3K19uvXL1166aXmZa0Ha3eRgAQkIAEJSEACEiguAgqsxXU97I0EJCABCUhAAhIoSQIIrIcPH07Tpk2LtABbjlfjveiii1K3bt3SfffdF1Gq559/foiwpBDYt29fpAqg+BUpAlq1apUoYHnmmT6eluQfgJ2WgAQkIAEJSEACFUzAJ9gKvvgOXQISkIAEJCABCZwqAfKo7tmzJ82ePTuW/a9evTq9//77kQ5gzJgxafz48RG9muVYHTJkSIip5F4lhcDGjRvTyy+/HHlbP/roo3TDDTekzp07n2q3/L4EJCABCUhAAhKQgAQKRkCBtWCoPZEEJCABCUhAAhIoLwIIouRYJR0AIumUKVOiEi+5VgcOHJiGDRuWBgwYEAWsssjUyy67LJF3lWjXM844Ix05ciSOsWDBgnTuuedGugA+J/o1E2XLi5qjkYAEJCABCUhAAhIoNwIKrOV2RR2PBCQgAQlIQAISKBABIlVXrlyZfvKTn0TxKsRSBFSKVj3++OPpiiuuSOedd17kWa3aJYpeEdnKZ+RtJSfr1q1b04QJE9Inn3wSAuygQYOiOFbV7/mzBCQgAQlIQAISkIAEipGAAmsxXhX7JAEJSEACEpCABIqYAJGriKuTJ0+Obd26denYsWMhro4dOzYNHz48XX755al58+YR0Vp9KESuIq5ef/318RGRqkSwbtq0KS1atCiORU7XPn36pC5dulT/ur9LQAISkIAEJCABCUigqAgosBbV5bAzEpCABCQgAQlIoHgJfPHFFxFhSloAIlczgfXzzz8PcZX8qXfddVcaPHhwCKinn356nYMhwrVdu3YJwRaB9tChQxHJShEsCmLR2rRpEz+bLqBOlH4oAQlIQAISkIAEJNCEBBRYmxC+p5aABCQgAQlIQAKlRAAhlYJW5FwlLQCRq7yHSDp06ND02GOPRVoAcqmedtpp9RoawinfbdasWSJqdcmSJSGyTpo0KYTXs88+O/Xu3TsE3Hod0J0kIAEJSEACEpCABCRQYAIKrAUG7ukkIAEJSEACEpBAKRIgynTv3r1p5syZadq0aZFzNUsLcMcdd6QRI0akXr16RWEqUgDUt7EvAi0i6tGjR+P7c+bMSe+++25avHhxojgW73OuSy65JPKz1vfY7icBCUhAAhKQgAQkIIFCEFBgLQRlzyEBCUhAAhKQgARKlABpAdjIubp27dr0zDPPRFGqrKAVaQG+9rWvhbiaLetvzFA7deqU2BBcKXw1a9astHHjxrR58+bI40oka+vWrVOLFi1qzOvamHP6HQlIQAISkIAEJCABCeSCgAJrLih6DAlIQAISkIAEJFCmBIgefe+999LUqVPTCy+8kNasWRMi6KWXXpooaEXO1c6dOycE0Fy0G2+8MaJWOR4FrzZs2JCIaKUfpCMg0pXcrTYJSEACEpCABCQgAQkUCwEF1mK5EvZDAhKQgAQkIAEJFBkBolR37twZeVFJCzDreFQpDXG1X79+afjw4WnQoEFR0KohaQHiILX8c/nll0c+1sOHD0ce1w8++CDSBcydOzeKXSGynnfeeRHNSq5XmwQkIAEJSEACEpCABJqagAJrU18Bzy8BCUhAAhKQgASKlMC+ffvS0qVL049//OO0atWqiCIlD+rAgQPTH/7hHybEUNIC1LegVX2HSSoAImMpfEW6gIULF4bI+txzz0UfzjrrrDRgwIDUvn37+h7S/SQgAQlIQAISkIAEJJA3AgqseUPrgSUgAQlIQAISkEBpEqCg1YEDB9KUKVOioBVpAT799NPIkTp69Og0bNiwdOWVV6bmzZvnJR8qha1atWqVrr/++ihuRcTq/Pnz05YtW6LwFTlh6U/fvn0jPcHpp59emqDttQQkIAEJSEACEpBAWRBQYC2Ly+ggJCABCUhAAhKQwKkTQLj87LPP0p49e6Kg1UsvvZSmT5+ePvnkkxBX+/Tpk+6555500003xTL9fAub5Hbt2LFjRLEiqB46dCiKXtE/0gOQluDCCy+MKFqiWm0SkIAEJCABCUhAAhJoCgIKrE1B3XNKQAISkIAEJCCBIiSAiLl79+702muvpaeeeiq99dZbCdG1Q4cOaeTIkemrX/1q6tatWzrnnHNynhagNhyIqLfeemukC2CfxYsXRyQrRbfIEUu0K5GupCuwSUACEpCABCQgAQlIoCkIKLA2BXXPKQEJSEACEpCABIqMAGkBiAxFXCU1wKJFi2L5PwWtEDhJC8CSfMRVRM1CNaJkiWKl0UciVRGC6St9PPvssyMvK/u1bds2+leovnkeCUhAAhKQgAQkIAEJQKBwT8fyloAEJCABCUhAAhIoWgLkXF29enVErhIl+uGHH6bLLrssRNXf+73fS1dffXUsxW+qARBF+5WvfCUEVSJX58yZkzZt2hTRrMeOHYv3EYIRgG0SkIAEJCABCUhAAhIoJAEF1kLS9lwSkIAEJCABCUigyAggVu7bty+x5P7FF1+MtACnnXZaiKvjx4+P1AAsvyfnKe83VSNClY08sLzSn4ULF6b169dHASxyx7L17t07denSpam66XklIAEJSEACEpCABCqQgAJrBV50hywBCUhAAhKQgATIrXr06NG0a9eutHTp0kgLQEEr3ictAELmqFGjIj3AeeedFwWlioEaha9at24d6QLoK5G327dvT4cPH44IW95r0aJFatmypdGsxXDB7IMEJCABCUhAAhKoAAIKrBVwkR2iBCQgAQlIQAISqE4AIXLv3r2Rx/SJJ55Ia9eujdymLMW/7bbb0u/+7u+mTp06JcRVIkaLqTVv3jyNGDEiCl8hEi9ZsiRE1okTJ4bwSo7Yfv36ncjdWkx9ty8SkIAEJCABCUhAAuVHQIG1/K6pI5JAowlQPGTbtm3hcL/33nsR1UREUNZwWC+88MLUrl272MjN16ZNm+xjXyUgAQlIoEQIcL/fv39/mjZtWiJq9c033wxxFUF1+PDhsXXv3j2W4Z9xxhlFNyrs0UUXXZRuuOGGSAuACDx37twQWRFbEYQphIXISjRuIYtyFR0sOyQBCUhAAkVNAP9r586dka6HAo74YVUb/hY2Dx+Moo/YNZsEJFB8BBRYi++a2CMJNAkBIoB2796dFixYkFauXJnWrVuXli1bFiJr1qHzzz8/ipz06tUrlo4S4YRTSwXnYotuyvrsqwQkIAEJ/JoAUasUhMJ5I2L1+eefT7Nnz04ff/xx5Fwlf+mDDz4Y93ju+U2Zc/XXva79J3LDsn3++eeR7oDCXBS+2rFjR9ims846KyYCmzVrpshaO0Y/kYAEJCCBJiBA3nB8MCY5KS751ltvpRUrVoR9rtqdK6+8Ml177bUxqTho0KCwaxR0dPKwKiV/lkDTE1BgbfprYA8kUBQEKBQyc+bMcLTffffddOjQochrV7VzPABs3rw5op54ENiyZUsaMmRIuv322yPXXdV9/VkCEpCABIqPwCeffJKIjnn11VfTs88+m1atWhWdJCLmrrvuShS16tq1a8nlLr355pujz0Tb4qQismLTEI5pRLqSu9UmAQlIQAISKBYCb7/9dtiqWbNmhT3+4IMP0sGDB7/UPXKls/Jk48aNie+888474X9hr20SkEDxEFBgLZ5rYU8k0CQEMNYUCJk/f36aPHlyGHciftq2bRtVmCkUQoQqEU84qlSaJmff6tWrI2IIZ520Ad26dYvvNMkgPKkEJCABCZyUwJEjR2IJIhNqpAaYM2dOrD5gqSEC5bBhw9KAAQNCqCy1qBhS1mC7SGuDyMpYcUhZlXHuuedGugA+Z4klUT82CUhAAhKQQFMRIHIV/4uAFXKHM9mJj4X/hWhKSoBshSArNEjpw8qTrVu3hp0jEIZVJuxzySWXhP1rqrF4XglI4NcEFFh/zcKfJFCRBDDmy5cvD4GVKtIYfCJ9hg4dmliCgnCKQ0r0KlFPCLFEBc2YMSNmUBFdcVbHjh2rwFqRf0EOWgISKBUCOGikgHnyyScjBQwiJOJq//790+OPPx6TaqR9Kfa0ALXxvvjiiyMCF6cTmzVv3rxwRp955pkQWHFEBw4cmNjPJgEJSEACEmgqAvhPpGNjEpA0PQSskF/11ltvTbfcckvkDyeAhYlBbPWiRYtiUpSc6aw0nHU84pXUNzR8MPazSUACTU9AgbXpr4E9kECTEkBgff3110/kq6OoCeIqy0RxvBFXcbiZPc2iWXFeEVWJgsJhX7NmTTjoTToQTy4BCUhAAjUSYKXC+++/H6sUWKmAU8eqBKI+ccyIXGX5fPPmzUs6nzaRq9gnJglp2CmcV9IF4JwyZgpf9enTJ8TkGmH5pgQkIAEJSCDPBBBUybfKcn/E1iuuuCJE1XHjxqUePXpEVCr+F3YNu8VEaMuWLcOuUdCRoBjStpEygM9tEpBAcRBQYC2O62AvJNBkBFieQkQTQivONY7p4MGDI3q1aqcoYtW6devYcGAx+tu3bw8Dj3GvXu2y6nf9WQISkIAECk+AglY4XhQwZBniK6+8kqZMmRITZoir3O/HjBkT93vu6eVSrBCxmChVhGUmB1lKSb46RGbGSYQuyy+zycLCXxnPKAEJSEAClUyAVRaIqyz5p1HEihUWRK8SyVq1IbLyORGtrMTAphEcQyTrtm3bYvVh1f39WQISaDoCCqxNx94zS6AoCFBxGceTpOqtWrUKA17dsFfvKLl+WMJC9WlmXanWjFBrk4AEJCCB4iGAuIi4SmqXn/zkJxG5ynvkeLvjjjvSY489FpGc5Cgt1bQAtdFmTIyRJZSMecmSJeHIvvzyyyG8kmOWSFaEZpsEJCABCUigkARIyUae8CxABd+KKB6+B8QAADRsSURBVFYE1Noa9qxv377pjTfeiF34LunbOJZNAhIoDgIKrMVxHeyFBJqMANE8CKoY9CyxOmkB6mrsixhLvh+WXJIbyOUpdRHzMwlIQAKFJUD0Js7Xq6++GgWtli1bFvfrTp06pSFDhqQRI0ak6667LpYbEh1Tbo0xtW/fPvXu3TvysZIu4LXXXosiX4itCKxEEGHDLBBSblff8UhAAhIobgLYKFZSIKrid5GirWPHjnUKrHyH1YZMINKwYQS6sFrFJgEJFAcBBdbiuA72QgJNRoDlJjfeeGNE+LCk8uqrrw6DX1eHEFOp1JzNmJZb5FNdY/czCUhAAsVMIHO0WA6/du3aNGHChCj2hOCapQX4+te/nnr16hVL5It5LLnoG4IyG+kPmAykmAjLMsnLiu3CUSX9DTnGyyVFQi64eQwJSEACEsgfASb9rrnmmhNiKSl7mBSsq1gVqzGw5QirNGyYPlj+rpFHlkBjCCiwNoaa35FAGRHA8bzvvvsiigdHk2hWZkfranv37k0rVqyIvK04pV26dIlZ17q+42cSkIAEJJB/Ajhe3KOpNPzCCy+EyErUC+IqxTPuuuuu1Pl4jtK6liHmv5eFP0O/fv3CccWppeDVhg0bEoVCmDBkspBIVyKJbBKQgAQkIIF8E8DXuu2222LJP+fC/yKiFXtdW2OSkMLC5G1FWMX/6tatW52ibG3H8n0JSCA/BBRY88PVo0qgZAhccMEFia0+LavAvGXLljRz5szI7UdFS/PY1Yee+0hAAhLILwGcr507d0a+0WnTpsV9mjNeeumlUZ14+PDhUcSQybS6nLj89rJpjn755ZdHxC6rL3BMyTtOgZA5c+bE+9g3Uubg4CLC2iQgAQlIQAL5IoCdQRytb/vkk08isGXx4sVp/fr1seKia9eu6dprr624CdP6MnM/CTQFAQXWpqDuOSVQogSI8mHZ6cqVK6PAFfn9iIS66aabEs6rTQISkIAEmo7Avn370tKlS9MTTzyRVq1aFcsIyS9KZeI/+IM/iPs1ImKlLilEPCWC9/zzz4+8dUSyUqTxF7/4RfzO0kzsGcs0bRKQgAQkIIFiIUBRYgJcpkyZEnaelDakGCDdj5OCxXKV7IcEUlJg9a9AAhKoFwGWUWbLTmfNmhXRq4iqAwYMSFneoHodyJ0kIAEJSCCnBMjJduDAgTR16tTYWEJIqgDSAowePToNGzYsXXXVVYkKxJWcZ5TCVhRoxGZlEavz5s1L77zzTiIqiEaUEHnJsW+VzCqnf6AeTAISkIAEGkWAvKv4YBRnRFwlepVJQopU9u3bN9IE1JW3tVEn9UsSkECjCSiwNhqdX5RAZRCgYArb/v3707p169LEiRNj5hRHldQAt99+exTGokiITQISkIAECkeAezMrC5j8QlR98cUXI/cqIiH5tckreu+990ZUJpGrCob/e21YeUG1ZjjhuJIuYPPmzTFxmKVPIHUOTqyOa+H+nj2TBCQgAQn8L4HM/yJydc+ePYnglkmTJsVkas+ePdOdd94ZNp5VKjYJSKB4CCiwFs+1sCcSKEoCOO9EQs2YMSP98pe/TMuWLYtcP0OGDEljxoxJt956a0RFFWXn7ZQEJCCBMiaAOLh79+40e/bs9POf/zwmwXDKcLhGjBiRvvrVr8YEGMsHKzUtQG2Xnxy0t9xyS+RdZR8iWFl+SRQwEcF8TqSr6W9qI+j7EpCABCSQLwLYcvwv0v08++yzaf78+enQoUMJcXXUqFHpnnvuicJY+Tq/x5WABBpHQIG1cdz8lgQqhgDLTolcpRAIDigRPxj38ePHxzJKooBsEpCABCRQWAKIgES1cG9m2eDChQsjQpWCVkx8UdCKpe6Iq6w4sP0mAaJ5M/sFSxghWMOU3Kxnn3122Dv2o7ozv9skIAEJSEAChSBAgMuGDRtCWCXI5eDBg2GL7rjjjrDxpP2xSUACxUfAJ+7iuyb2SAJFRYAqyy+88EIYeIpaUa1y6NCh6Rvf+EZR9dPOSEACEqgkAhQcJLLlySefjNxsiISkBejfv396/PHHE9WFWeJuq5sAIutDDz0UKzEQWIkG3rhxY2z8ToqA2267TYG1box+KgEJSEACOSRA9Oprr72Wpk2blt5+++0oUol9x15169Yth2fyUBKQQC4JKLDmkqbHkkAZEiCCdcWKFWnXrl2pZcuWsey0X79+ZThShyQBCUig+AkcOXIk7du3L5ayv/TSSyEEEn1JQatx48alkSNHpg4dOpg7tIGXkoIhWVQrha+ICM4KX1FkhHy2Xbp0aeBR3V0CEpCABCTQcAJM8CGsMuGHDaLuBXlXyQ9uk4AEipeAAmvxXht7JoGiIEBydQw80VIUBmH2lChWmwQkIAEJFI5Alo+NyS5yYZMrlMgW3ictAAIgedlID0BBK3KI2upPAPt28cUXp1atWkW0Krltd+7cmWbOnBk84cwkY4sWLSLtQv2P7J4SkIAEJCCBhhEgRcCOHTsSKwkRWIlavfnmm8MONexI7i0BCRSSgAJrIWl7LgmUIAGMOktPMfQ47G3atAkHswSHYpclIAEJlCwBBL69e/dGftB///d/T2vXro2coUSrIqr+3u/9XqQIQFwlEtPWcALnnntuuv7668PWkXP1V7/6VXrzzTfTxIkTE5HDpAsgr22Wu7XhZ/AbEpCABCQggfoRIE0AG6158+YRverkaf3YuZcEmoqAAmtTkfe8EigRAjicVKQmlx+vzZo1s2BKiVw7uykBCZQHASa59u/fn6ZPnx5RqytXroxJL9ICjBgxIg0bNix17949cb/W+Wr8NYcdUarkr0WkRlTlvTVr1kSeW96j0CNpcogatnhY41n7TQlIQAISqJ0A9ubCCy+M9D/sRYALk3w2CUiguAkosBb39bF3EmhyAiyXJDcdqQJwKBFYbRKQgAQkkH8CRK0eO3YsUWBw3bp16fnnn0+zZs0KkY+CVuRko+AF6QGYBDvttNPy36kKOAOpAthgf84550TO282bN6ft27dH+gCcXJxdJxwr4I/BIUpAAhJoAgLYmSuvvDJRC4PWtm3bJuiFp5SABBpKQIG1ocTcXwIVRuCqq65K3/72t2MpahbFWmEIHK4EJCCBJiFAtOSePXvSq6++mv77v/87lqtzH0ZYHT58eBo6dGjiHo0IaMs9AXLeIXKTKmfGjBkRxcq1ILKVdsMNN0Ru8tyf2SNKQAISkEAlEyDdz7333puGDBkSGK644opKxuHYJVAyBBRYS+ZS2VEJNA0Blqew2SQgAQlIoHAEEPEoaLVgwYJICzBnzpyImCQVwOjRo0NcZal6Ltu+ffsizysRM6xaIDUBjdQD7dq1iwJQ5IGrWuiJ1AUUQfzggw/SwYMHY2PpPKsfKBzVunXrkl35cNFFF0WKgExkPXToUNq2bVuaP39+MCE3OblasZGK3Ln8S/RYEpCABCqbALbFosKV/Tfg6EuTgAJraV43ey0BCUhAAhKQQBkTQLhcsWJFevLJJ9Py5cvTxx9/HKkARo4cmR555JEQPHM9/PXr16fZs2enZcuWpQ0bNqStW7fGKdq3bx8Rs6SLIarz6quvPiEobty4MSI7+S65YdlYOk+xqK9//evR51JOLYNQjJBNLlaih5955pn01ltvpWeffTZWdrCMc+DAgZFSINfXw+NJQAISkIAEJCABCZQOAQXW0rlW9lQCEpCABCQggTInQNQoEaRTpkxJkydPjtyrRJBec801aezYsbFckPygRLfkupFnu3///pHLlWhNhF2iN4leJc9rr169ElGdCI1Zo/gh77OMfsuWLSEEDx48ON1xxx1RLIoI1lJuFBohOpVceDSuz6zjeXDfeOONtGjRohBZifyFW8+ePUt5qPZdAhKQgAQkIAEJSOAUCCiwngI8vyqBSiDAss+dO3eG84xDjzNNXiCbBCQgAQnkjgBC5qeffpp2796dVq1alV555ZUQWRH4iBwl5yrRq4iZ+WqXXXZZLHcnDcDevXvT1KlTQ2Dld/rQo0ePL1Ux7tChQ7yHnUCIJS0A/SSNAfYiH0JwvsZf13EpMEI0K0Iy0axcJ4pekSMX0ZVrh/jMPtrIukj6mQQkIAEJnIwAKWiwq6TroWGDTNl2Mmp+LoGmJ6DA2vTXwB5IoKgJsFT07/7u7xLRTDjOf/VXf2WUTlFfMTsnAQmUIgGEO0Q78nv+9Kc/jchVhDyW2t95553p/vvvL8gydKI1SQHQpUuXExiPHj2aduzYEYJp9UrGFOKi39OmTQtH8L777kt9+/YNsZVcrOXUSAdAlOppp50WIuvLL78cUb5cM1I4YCe5VgjRNglIQAISkEBjCWBPvv/978czAcf4zne+kx588MHGHs7vSUACBSJQXk++BYLmaSRQSQQoXvL6668n8gFSwCSbSa0kBo5VAhKQQD4JEAH53nvvpZkzZ0bU6NKlSyMKkqhRBLvbbrstFaqCMKIuUZjkXUVkRTxFPGT5P5Ns1QXWzZs3R65YinJdfvnlEWnLfuUYxUk0MSkPEKARWRGejx07lrL8s4jNpHPgva5du8bP+fy78dgSkIAEJFCeBIhgJd/3woULY4Bf/epXy3OgjkoCZUZAgbXMLqjDkYAEJCABCUigtAgQqUJRKQoozZkzJ5acd+/ePQ0dOjQ98MADIVwWekQsRSQSdfHixQnxdNOmTVHRuHp05ptvvpnmzZsXguyAAQMS+VfLvZEKgA2RFUH1Zz/7WQjQpAvgPRzjjh07KrCW+x+C45OABCQgAQlIQAJVCCiwVoHhjxKQwJcJ9OnTJ/3whz+MSB2imrJCH1/e03ckIAEJSKAhBMhxTQTo7Nmz0/Tp09Pq1atTixYtoqDUuHHj0ogRI5os5xp5V4lIXbNmTeSBQwAmmjVr5Ghdt25d2rhxY+RZpahVpRV5uuqqqyJalfyrRB8vWLAgeBEFjNBK5DHFwWwSkIAEJCCBhhDgWeBP/uRPUha5euONNzbk6+4rAQk0EQEF1iYC72klUCoEKHpCBJVNAhKQgARyR4D0K5m4Sv7SuXPnRsTjNddcE0Wibr/99igslbszNuxILVu2TETRIhoSvYqQSsENlsEjKG7dujWibYluRWgkncGll15ar5McOHAg8pUePnw4ITLzSiOtAJGzF198cSzF5z3OR4qCffv2JZixP0vwWa5//vnnnyi6xQQgG/0uVO5X+knOWvpDkTLSPOzatSuRu5zf6SP9Id1Cs2bNGI5NAhKQgAQkcFIC2EMmLm0SkEBpEVBgLa3rZW8lIAEJSEACEigDAuTtJGp1woQJ6Z133oll5UQ7jh49Oj388MMFKWhVF0aWwBMx89JLL0XKAgTW7du3Rx5ucnKvXbs2BNa777478sQ2pLoxeeWWL18eaRFWrVoVeeboS+fjuVsRlkeNGpX69+8f3UNQRcydNWtW5KJbuXJlCK4Im+Q5RVSlsT8pDRB6ifwpVEM4ZaUHUas4xM8991xiTFnePMRhCn+5+qNQV8TzSEACEpCABCQggaYhoMDaNNw9qwQkIAEJSEACFUiASEwEVSrQz5gxI3J3shwfsXDMmDFpyJAhEfF49tlnNykdokPJI0qUJiIihbi2bdsWwiHiKiJrv379EjlZidA866yzau0vEZ5EvW45XiiLoolE7lIwkRUSCLNEwE6ePDll4mm3bt1OCKykJnj++edDkEVs7dWrV2rXrl1q06ZNnI9jEmVLQS76wNL8QjaEVfgwBiJW4URe1hUrVkS6ACJZEV7JTXv99dfHPoXsn+eSgAQkIAEJSEACEigMAQXWwnD2LBKQgAQkIAEJVDABREaWuyMIIqwisC5dujSWmCOuDh8+PARWUgQUQyNClA0xE5F1x44dEUlK3ylsRX7Wb37zmxGZiaBYW0NgZNzkbF2yZEl6+umnY2k/Yx44cGCIuIitb7/9diJVAkyq5npFYCXKF2H62muvTffff38Iu/QrE18nTpwYaQEuuOCCiCStrS/5fL9t27YhFn/++efRFwpekVIhi2Tl/U6dOiXEdLjaJCABCUhAAhKQgATKi4ACa3ldT0cjAQlIQAISkEAREjh69GjK0gI89dRTIViSn5Ml7USujh07Nl1yySVF13OW7SP6InCytB8xlGhMIm2JGmUMdTXEVQTTZ555Jop4Ec1JGoRBgwbFeBEeESM/++yzWNqPCIlQmjVyvJLblOMg5JLntXXr1lFYi76xPP/WW2+NAlsdOnSI97PvFvqVCNaqAjmCMRG7ROYSkczY6WvVfQrdR88nAQlIQAISkIAEJJAfAqfn57AeVQISkIAEJCABCUggI0AEayYWEq3JcncEOUQ3oh+vuOKKiOzM9i+WVwRUBE8KRyFyshFJirDJ+yeLxiRil/yps2fPDpF2wIABsfyfAlqkISCvK0v8EaARbIcNG5YQTrPG+fmdyE/4kY+VYlcszeczxMrx48fHK0v1eb+pGukJYEJqhaywFQIyAvG7774bEa2kELBJQAISkIAEJCABCZQfASNYy++aOiIJSEACEpCABIqMAMIfkZfkDkUIJBIUUZGl90RwIrwhODalQHgyZORLJecq4io5R+uT73Tx4sWJiF2KZBH1+uijjwaHTHDm85/97GcRwUpRrUceeSQE1awvRLSOHDkycrTCadKkSRHpiiBNo3gUAixidbE0hGCE4wMHDpzoEkI0hcPqSqdwYmd/kIAEJCABCUhAAhIoOQIKrCV3yeywBCQgAQlIQAKlRoACTOQNZXk9wuqUKVNiyTyRjdOnT0/kKh01alSiwBPLyYtFMCSPKFGoRK6Sd/Xqq69OrVq1Omn/iDIlRypL5CnqRcQq6RAY27p160JwXbNmTaQPIAqVqFiKQGV5SrPry+9wQYjmWORxRagkohWxF9G3WFgRrUrOVSJ2Z86cGdG2iOZE7Q4dOjTSA5DGwCYBCUhAAhKQgAQkUH4EFFjL75o6IglIQAISkIAEiowAkalEryIyUjQKkfXjjz8OEY6K8wiSLHlHgCQis6lFVsRC8qIiAG/bti0EVsRMBOAWLVqclC7jQWwk/yj5W2+77bYQRsnHunnz5sjnSk5XWPTu3TsKXhGVSoRnVcGU9An9+vVLa9euTfv3709vvPFGHBfhko3IUCKC6xNNe9JOn8IOpHxgbAirkydPTlxT+gavESNGpDvuuCME4WKOUD6F4ftVCUhAAhKQgAQkUPEEFFgr/k9AABKQgAQkIAEJFIoAkZcIqA899FDkXZ0wYUJEeiI2/ud//mcIrV/72tfqld80n31G/EXQJHKUKNZPP/00BFIiWBnDyRrfRWAlepXoXI6HsMoyf0TTm266KQp7IdYilPKKIFldKCX3K8IzRcCInEWs3XI8ovaXv/xl9AchumfPnl/63sn6l+vPuX6/+tWvIhp51apVIaD3798/3X333ZFXlpQKVYXjXJ/f40lAAhKQgAQkIAEJNC0BBdam5e/ZJSABCUhAAhKoIAIIhmzkDuWVSFFSBMybNy9t2rQpvfrqqyFgkkqAZfOkFmgKYe7QoUOxjB9Bkz4SJUrhJsRhfj5ZI6csxbwo8EQhL5bzE71LQ2DlWKRMQFSt3hBzOSeNiE84XXrppRHpSjQovDj2a6+9FkI0nJqCEf0jly6pDujT1KlTgxn9HThwYESu3n777cGsPqI0x7NJQAISkIAEJCABCZQmAQXW0rxu9loCEpCABCQggRImgOBGNCjCISIjUaKIhiyB37VrV0Rkkq+Twlg1RXbme+gHDx4M4ZCl/giGpAegPx07dqzXqUmBwJhYOn/BBRdEHlLERlrVKFWiW7ONzxBK+S6FonifsRO5yis5YMePHx9L8Uk9MH/+/HiPaOCmaOSlJS3Ayy+/HDl1Fy1aFKkdEHzvvffeSAtwww03NEXXPKcEJCABCUhAAhKQQIEJKLAWGLink4AEJCABCUhAAhBATGR5PNGqRHmSLmD58uWxLH/ixIkhTiLUETWKwFjIduDAgch7isBKH4m4JeK0MS1LEYBwWn0cRKuSPoCIWYRccrKSXoDcr4i8iLpZ5CuRs+Q0ZZ9jx44l+oiA21SNolszZsyIyFUKenE9uZbDhg2LolYU6LJJQAISkIAEJCABCVQGAQXWyrjOjlICEpCABCQggSIjgCDHRnGnLL8oXSR3KcvOESSJ/mS5OdGuCJDsn8+GGMryfERO+oDASp5TxFVyoda3UayKaFeEUyI9N27cGCItYjHjoIAW6QcQUVlmz/5E63J+0gpwblImdO/ePaJUEaARY+kP6QfgALOmWHpPn7dv3x4FraZMmRIFuxCOBwwYkEaOHJmGDBkSQjB9tklAAhKQgAQkIAEJVAYBBdbKuM6OUgISkIAEJCCBIiVAnlWW4N9///3xivCIIMkyeKI7EfQuu+yyiCRF7MxnIzKUc7L0fe3atREl2qZNm4iwbci5EUsRHBFDtxwvSkW+VMTQiy66KF4Z05IlS9LWrVtjnzvuuCMEU9IHcO6lS5dGXlqKYhHByvgRVydNmhTiK4Is7yPAVk05kE822bHp87PPPht5V1esWBF5colcfeyxx9L1118ffW0Iq+y4vkpAAhKQgAQkIAEJlC4BBdbSvXb2XAISkIAEJCCBMiCAQEhUJ9Gqffr0Sb/zO7+TXnrppUROT8Q8lqFT7Iml5wiKuS58RcQqUaabN29O69evjxQFFN1iCT7vI7jSDwRRzk3eWCJaEVHpV02NQlajRo2KSNO5c+dGfllSICCcdunSJQpdcSx+5jhEuyKaEplKfxB6SQnA0vt//ud/DvE16yORrDfffHO666670o033pj3qN5sfEQUkyN31qxZIa7ChGvGdUEg7tu3bwjI1dMgZN/3VQISkIAEJCABCUigfAkosJbvtXVkEpCABCQgAQmUCAFEVqIeERwvueSSSA9AlOeqVasiLyuRnAiObHyOiFebuNnQISNmIlpSXGv16tURPUsUaZZDlKXunI/I2nXr1oUYTDQqS/Rra1lqA47N2BYvXhwiLaJkjx49EsWf+vXrF4Jx5+NpA6o2xpgV0yL6FaGZhghNX7p27RpCNMImvAoRwXr48OEo2jV9+vTIuUrkKmMkYpU8uQi9WZ+rjsWfJSABCUhAAhKQgAQqg4ACa2VcZ0cpAQlIQAISkEAJEEBEpKjU8OHDE0vzf/rTn54QPVmWjtD6la98JaJIG5ITta6hI9RyTkRP8sGyrJ+oVSI2aUSVIrAibrIf4iqvdQm82TH79++frrrqqvTQQw9FNCzHQ0Bt0aJF5HTl5+oN0RLRlfMfOXIkNvZBSKUvnJuxk1ahUNGiCN2vvPJKiKtE+RJ9i8B7zz33pN69e0fkavVx+LsEJCABCUhAAhKQQOUQUGCtnGvtSCUgAQlIQAISKHICCIhs5BxFpCRqFEFxzpw5sXyf4lAUgiLnJ8IeYuep5vvMzklEJlsuWnZMcq6yNaTlsh8NOW9N+1KkizQFpGmYOnVqFN5CHCZiFRGcKFzGVyiht6Y++p4EJCABCUhAAhKQQNMTUGBt+mtgDyQgAQlIQAISkMBvEECwQ2R99NFHI9pz27ZtkSN14cKFUfxq//79EcFJxCnRlLRCLJX/jU6W+S+kNyBieOLEiWnKlCmRg5aIW0TVxx9/PF1zzTWRi7bMMTg8CUhAAhKQgAQkIIF6EDi9Hvu4iwQkIAEJSEACEpBAgQkQBUrhJwS97373uxG1SrQkxZ4otPSjH/0ovf766+m9994rcM8q43QLFixIzzzzTJo2bVrkjyWylvQMpDsgD2yuUjRUBk1HKQEJSEACEpCABMqbgBGs5X19HZ0EJCABCUhAAiVKAIE1K3zVtm3b9P777ydSBCxZsiQKUVEEizytpAxAhCVH6qmmCyhRVDnpNhw///zz9MEHH6S9e/eGiE3k6tq1a6OgV58+fdLo0aPTwIEDI3KVfLk2CUhAAhKQgAQkIAEJQMAnQ/8OJCABCUhAAhKQQBETIAUA0ZLjx4+PfJ8ffvhhiH67du1KEyZMSKQLILKVlAIUfrI1jgApASjutWbNmihoRd5VilshdA8aNCg99thjqUePHiGu1lXgq3Fn91sSkIAEJCABCUhAAqVMQIG1lK+efZeABCQgAQlIoOwJIPCRX5Uo1r59+6aHH344coLOmzcvimCRl5VUAkRXUnypdevWyejKhv1ZfPrppxEhTDExtvnz56ft27dHhPCtt96aRo4cma6//vpgm+W8bdgZ3FsCEpCABCQgAQlIoJwJKLCW89V1bBKQgAQkIAEJlAUBBFYKX1HUql27diG4knt13bp1af369WnLli3xeatWraL4EpXujbI8+aUnLQAb6RY2bNiQnnvuuTR37ty0Y8eOiAq+4YYbIudq7969U4cOHU5+QPeQgAQkIAEJSEACEqhIAgqsFXnZHbQEJCABCUhAAqVIANH0/PPPT0OGDEmIqD//+c/TsmXLotDVpEmT0oEDB9Kjjz6aunfvni6++OJSHGJB+0zOVdICUDTshRdeiPy2MCSfLRHBY8aMSb169UoUuLJJQAISkIAEJCABCUigNgIKrLWR8X0JSEACEpCABCRQZARIF8DWuXPnEAGJYqWwFcvat23bll577bVY1v7RRx+l/v37p2bNmln4qoZrmBW02r17d1qxYkWaNm1aMKSQGPlsSQcwfPjwdPPNN4dQbfGwGiD6lgQkIAEJSEACEpDACQIKrCdQ+IMEJCABCUhAAhIoDQLkAW3fvn3kY23Tpk3kCyVNwKZNm9KTTz6ZDh8+HEvaO3XqpMBayyU9evRopFf48Y9/HFHA7777bojWpAP47ne/m7p27XoiHUMth/BtCUhAAhKQgAQkIAEJBAEFVv8QJCABCUhAAhKQQIkRICcr6QJIE0Dhqz/6oz9Kv/zlL9OCBQvSoUOHIo/osWPH0j333JPII8oSd3Oy/u9FpqDVkSNHolDY1KlT05tvvhnMiFwdNWpURK4irlIsjGhhmwQkIAEJSEACEpCABE5GQIH1ZIT8XAISkIAEJCABCRQhAcQ/lq5feeWVqWPHjunDDz9MpAYgJ+vbb7+d9uzZkyh6RbQrIivpAvi5UhtpARCdSauwcePG9Morr6QZM2Yk0gSQr5Zcq3fffXcaMGBARK4qrlbqX4rjloAEJCABCUhAAg0noMDacGZ+QwISkIAEJCABCRQNgTPPPDMKX915550RqfrZZ5+l1atXp3379qX/+Z//iVfE1S5duqQLL7ywaPpd6I4grhK5unjx4vTUU09F7tW9e/eGSD3keNGwBx54IF133XUhthIhbJOABCQgAQlIQAISkEB9CSiw1peU+xUdgf3796fly5dHrjmr+xbd5bFDjSRAZBUFV3bs2NHII/g1CUig0ggQacl2ySWXpH79+qUDBw5E5CqFr4jOJG0AqQQo2sTnLVu2rLhIVtICIDjDYsqUKWnRokXpgw8+SG3bto1CViNGjEh9+vQJgdqCVpX2P8jxSkAC9SXA5N33v//95CRUfYm5XykQ4Blh3bp1pdBV+1jkBBRYi/wC2b3aCbD0kQ0H0iaBciTQvHnzchyWY5KABPJE4JxzzkmdO3eOwleIhDt37oxUAevXr09bt26Ns5Jn9KqrrgqRtRKWwJMWgEb6BNImTJgwIURWClqRY7VHjx7p61//eqQH6NChQ+zrPxKQgAQkUDOBJUuWJDabBCQgAQl8mYAC65eZ+I4EJCABCUhAAhIoSQKIpueff34aNGhQOvfcc9MzzzyTXn/99cSqj2nTpoXQ+OCDD6aePXvGUviSHGQDOk26hE8++STGPmnSpLR06dJ08ODBiOgdO3ZsGj16dLr22mtDbG3AYd1VAhKQgAQkIAEJSEACv0FAgfU3cPhLMRPAaaRYR/fu3RN55silZpNAORNAHOndu3cs+3UpVjlfaccmgdwRwFaykW+VCE2EVdrChQvT5s2bIwcpaXUQHvv375/OO++8yEGaux4Ux5GIXP38889jpcvatWtDYJ01a1YUuKKgFQLzyJEj0+DBgyNNQCUX/yqOK2YvJCCBYiVw9dVXp7vuuqtYu2e/JJBzAl27ds35MT1gZRA47fgD6P+unaqM8TrKEibAnyoO4dGjR6NKcgkPxa5LoF4EEFVZ5suyXzZF1nphcycJSOA4gcxmknf01VdfTT/4wQ9iiTx5Ryl0NW7cuPT7v//76bLLLktt2rQpO2ZZQat58+alf/u3f4vc1qQF4F5K1Cpj79atW2rfvn0644wzvL+W3V+AA5KABHJFgBQrH330Ua4O53EkUPQEKAzKZpNAQwkosDaUmPtLQAISkIAEJCCBEiCAyErhhk2bNiWERpbIz58/P9IEkIf1lltuSWPGjIniToisZ55ZHgubGDNC8owZMyJyderUqVH4i7zWd9xxRxT7ouAXq2KI4LVJQAISkIAEJCABCUjgVAmUx5P0qVLw+xKQgAQkIAEJSKDMCGRR8KQLuOSSS2IFyIEDB9KqVavShg0b0vbt2yNCg5ytvXr1SgiQpSyyIigTuUpaBApavfjii4m0ABT7Ii3Addddl+6777504403RuRqmV1uhyMBCUhAAhKQgAQk0IQEzvj/jrcmPL+nloAEJCABCUhAAhLIIwGEVoRTUgOwJH737t3p/fffj6jOPXv2xOull14axbFKeUkcOVc//vjjEFV/+tOfpsWLF6e9e/cm8qsSqfvII49EtC5iaykLyXn8U/HQEpCABCQgAQlIQAKNJKDA2khwfk0CEpCABCQgAQmUAgEEVgpfIZ6yLJ6fP/nkkxBaiWg9ePBgFISisB4FsMhJylZKjbQAiKmkQJg8eXLknSX/7EUXXRSpECjQkhW0Ypw2CUhAAhKQgAQkIAEJ5JKAKQJySdNjSUACEpCABCQggSIlQIEnilo9+uijkQ5g27ZtafPmzWn9+vWRLgDRtUOHDumKK66IqE+GUezF9UgLQKMA5saNG9NPfvKTiFwlLQCCcs+ePdO3v/3tdM0116R27drFvv4jAQlIQAISkIAEJCCBXBMwgjXXRD2eBCQgAQlIQAISKGICRLCSd/Xyyy+PglekCzh8+HAUhiJ9QOvWrVOLFi1in2IXWD/77LNIC/Dyyy+nX/ziF2nJkiUnClrdc889ady4cZEWgDGRKsAmAQlIQAISkIAEJCCBfBAwgjUfVD2mBCQgAQlIQAISKEICiKtnn312ovBV27ZtIxcry+tff/31tGnTpkgXgBhJoxjUeeedF/sX21CIXCXnKjlkKWg1derUNGPGjEgTQI5VIldHjx6dBg0aFJGr5lwttitofyQgAQlIQAISkEB5EVBgLa/r6WgkIAEJSEACEpDASQkQzdmyZcuI8CRPKRGs69atS7t27UoTJkwI4ZX3KX5Fcaxia8eOHYvI1eXLl6cnnngirVq1KnLKIh4jqj722GOpR48eISKXWj7ZYmNtfyQgAQlIQAISkIAETk7AFAEnZ+QeEpCABCQgAQlIoKwIZIWviFAlXQApAcjBSkTooUOHQnClABafZcvri0WoJOJ2//79UcjqlVdeSbNnz47IWwp0DR8+PCJXBw4cGAW7yDtb7GkOyuoPy8FIQAISkIAEJCCBCiVgBGuFXniHLQEJSEACEpBAZRNAeESApKhVVgDqvffeS2vXro3CV1u2bIn0AK1atUrdu3ePiNemFFlJC8B28ODBtGHDhvTcc8+luXPnpnfffTeRFqBXr17p4YcfTr17945iXZV9dR29BCQgAQlIQAISkEAhCRjBWkjanksCEpCABCQgAQkUGQGEVnKUIqR27NgxokMpfIWQyeu+fftCsGzWrFlia6pGztUjR46k6dOnp6effjotXrw4+kbfKWb1wAMPREErUhuYc7WprpLnlYAEJCABCUhAApVJwAjWyrzujloCEpCABCQgAQkEAQpfsRHJSkoABFUESqJDt27dmj777LNYbv/RRx+lm266KfYh12mhGlGr9IH0BStXrkzTpk1Lr7322ok8sdddd12kBrj55psjkrWQfSsUA88jAQlIQAISkIAEJFDcBBRYi/v62DsJSEACEpCABCRQEAIUvmrfvn0ss2/Tpk3avn17euedd9KmTZvSk08+GXlZO3TokDp16hSpAwrSqf87ydGjR9Nbb72VfvSjH6Vly5ZFWgDyx95www3pu9/9buratWv03XyrhbwqnksCEpCABCQgAQlIICNgioCMhK8SkIAEJCABCUigggkgTrIhtJ577rmRLoAl+US0Hj58OH344YcRRUpBrJYtW8Y+RL7ms1HQinNTzGrixIlp4cKFieJbnH/s2LFpzJgxqX///lGIi8hVBdZ8Xg2PLQEJSEACEpCABCRQGwEjWGsj4/sSkIAEJCABCUigwgggmCJUXnnllSGwIqoicK5YsSIKS7FMH3GTfYgeJScrgmyuG2kBjh07lii6RQQtAuurr76adu3aldq2bRsFre6+++40YMCAKNClsJrrK+DxJCABCUhAAhKQgAQaQsAI1obQcl8JSEACEpCABCRQAQQQLMnDipjJRroAil7t378/7dixI0TXzp07Rz5W8rbmuiGuEj07b968SAtAQSvEXcTckSNHpq997WtR0Iq+nXHGGUau5voCeDwJSEACEpCABCQggQYRUGBtEC53loAEJCABCUhAAuVPAIGVaFbE0+bNm6dzzjknffLJJ2n37t2xRJ/I1o8//jjSBJCvFeEToTMXjbQARK7Onj07TZkyJSJX33///XThhRem22+/PY0ePToNGjQoXXTRRXF+o1dzQd1jSEACEpCABCQgAQmcCgFTBJwKPb8rAQlIQAISkIAEypgAwiqRqo888kiIrDt37kxvv/12FJyiABZL+RE6r7rqqkgdcCo5WTkWjZQEGzduTP/1X/8VOVfffffdyLHao0eP9I1vfCPSA1BsyyYBCUhAAhKQgAQkIIFiIWAEa7FcCfshAQlIQAISkIAEipQA6QKIZL3ssssiVQARpR988EH8zNJ9RFaKX5GTtbHts88+i6hYolafeeaZRFoAUhIg8t5zzz3p3nvvTX379k0XXHBBXvK+Nrbffk8CEpCABCQgAQlIQAJGsPo3IAEJSEACEpCABCRQKwGiUtm6dOkSkaSInkSbLlq0KG3evDl99NFHIXoikPbv3z+dd955UQSr1gNW+4Bjff7555Fjde3atWnatGlp5syZkSbg4osvTj179kwjRoxIt9xyS+SDzUdRrWpd8lcJSEACEpCABCQgAQk0iIACa4NwubMEJCABCUhAAhKoTAIIm+RB/cpXvhKC6r59+yJdwLZt29JTTz0V0azt2rWLKNezzz673pAQWI8ePZpWrVqVfvjDH6aVK1cmUhEQuYpg+61v/f/t3ctLVWsYx/EHhTArb0hqahliXsEEiSKRkMiyHFSOEp2ekf+PE3EkQsOwUtG0m1AhBKWWGzMqonCgFl4gssPvge1AT51VpPtVvwvCy17u9e7PChbrWe/7e/6xsrIy03v/rZzXyINjRwQQQAABBBBAAAEEIggQERABiV0QCE1AyzE1gyglJcW7PIc2PsaDAAIIILD7BOKNr1T41CxVxQKo8ZUaUikuQDNZ9b1iAtLT071A+n+ZrGpo9eXLF5+12tfXZ2NjY95EKyMjw5qamryh1enTp02NtHRcGlrtvv9XfCIEEEAAAQQQQGA3CDCDdTecRT7DnhFQx2Z1blZjEd2U6gZUN5xsCCCAAAIIbIeACpyanaq4gCNHjvi1aHFx0SYmJiwWi9m7d++8wJqamurNqJTbqvzWjZtmra6trZmyXNU069atWzY6OuozVxULUF1dbdevX7fa2lrLzc3d+Of8jAACCCCAAAIIIIBAUALMYA3qdDAYBH4t8ObNG7t7967Nzc35DWt+fr7PIvr1X/EqAggggAACf1dAhVYVThUZoGuRiqz6p9gArbLQ9wUFBevF1o1HV+aqHhqqqNrd3e0NrXRtUwzBlStX7MaNG97QSsXW/yrQbnw/fkYAAQQQQAABBBBAIJECFFgTqc+xEYgooMYhigRQR+WbN296J+ejR4/6zatiAtgQQAABBBDYToF4XIDiALKysjyyRr/Tcn9dr1Ro1QxVrbLQ68pOjeenagWGiqmKA9BDw3v37vn+ihyoq6uzy5cv+9fDhw/7+27n5+JYCCCAAAIIIIAAAgj8icDmNVt/8i78DQIIbJmAllEq4+7t27f27NkzvxktLi72PDpm9WwZO2+MAAIIIBBBQAXUwsJCa29vNz34U5H1/v37Nj09bZ2dnT5LVVECRUVFPjtVb6mGVlqR0dXV5dc1NbRSpEBVVZV1dHRYeXm5N7SKcHh2QQABBBBAAAEEEEAgCAEKrEGcBgaBwM8FNNNHTUPu3LljDx488BlBmg2kTDotpWRDAAEEEEAgBIHS0lIvtGpWq65XMzMzHgGgh4QtLS1WUVFhhw4d8uvZwMCAvXz50nPFdU1TLMCFCxespKTE0tLSQvg4jAEBBBBAAAEEEEAAgcgCFFgjU7EjAokR+Pz5sz1//twePXpkk5OTphmtuvlU7h0zWBNzTjgqAggggMBmgby8PFNm6tevX/1h4PLysn348MELqpmZmR4doNcHBwc9FmBhYcEUC6DCbFNTk509e9ZnrnJt22zLbxBAAAEEEEAAAQTCFqDAGvb5YXQI2Pj4uPX09NjU1JTftIpEHZz379/vSzEhQgABBBBAIBQB5azW19f7g0Dlhz9+/NhisZj19vZ6UVUFVj0sVE5rTk6OnT9/3pqbmz0eQJmr8ZzWUD4P40AAAQQQQAABBBBAIIoABdYoSuyDQAIENLNHyyvVBERFVjUMUdadZgFpiWVSUlICRsUhEUAAAQQQ+LmAMli1yqKsrMyuXr3q1y2tvFDO6vz8vL1//95ntypL/OLFiz5r9eTJk5aRkUHszc9ZeQUBBBBAAAEEEEAgcAEKrIGfIIa39wR0I/r9+3dTNMDIyIg9efLEZmdnHUJLKdVERAVWNgQQQAABBEIV0GzUxsZGH56aWg0PD3tcgGIDTpw4YWfOnLHW1lY7fvy4N7gK9XMwLgQQQAABBBBAAAEEoghQYI2ixD4IbKOACqxqajUxMeE5derEHN80a1WNrZi9GhfhKwIIIIBAyAKanapIG+WqKk9cDw+vXbtmly5dsvz8fJ/hGvL4GRsCCCCAAAIIIIAAAlEEKLBGUWIfBLZJQDNX1W35xYsXnlunnDpFA8Q3RQSouVVKSkr8V3xFAAEEEEAgWAHlrKrAuri4aMeOHfOogIaGBquurvZrGQ8Mgz11DAwBBBBAAAEEEEDgNwQosP4GFrsisNUCagii7suKBujv7/eZrN++fVs/7MGDB305JREB6yR8gwACCCAQuICuWYoLUPOrpaUlS09P96Jr4MNmeAgggAACCCCAAAIIRBagwBqZih0R2HoBNf94+vSpL6P89OmTd1POzc31G1H9rCWWKrIqJoANAQQQQACBnSCgxlf79u3za5pWYugapt+xIYAAAggggAACCCCwWwQosO6WM8nn2NECigZQ44/Xr1/7zNVXr155VEBRUZF3VtaN6MLCgt+gqtOyblDZEEAAAQQQ2EkCycnJXmTdSWNmrAgggAACCCCAAAIIRBFIirIT+yCAwNYKqLiqvNWHDx96gfXjx49WUFBgbW1tVltbu37wAwcOeIYdEQHrJHyDAAIIIIAAAggggAACCCCAAAIIJFSAAmtC+Tk4AmYrKyumgurQ0JCNjY3Z3Nyc5eXl2alTp6yurs4yMzP99dXVVY8KUGRAamoqdAgggAACCCCAAAIIIIAAAggggAACAQhQYA3gJDCEvS2gzsqxWMxu375t4+PjnrNaU1Nj586ds8rKSvvx44fNzs56IVbRANnZ2d55eW+r8ekRQAABBBBAAAEEEEAAAQQQQACBMAT+BbwMuBMdpx2AAAAAAElFTkSuQmCC"
    }
   },
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "We pick \"Outlook\" as a first variable to split on:\n",
    "\n",
    "![tree_1.png](attachment:tree_1.png)\n",
    "\n",
    "We have to keep splitting for Sunny and Rain"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Overcast = Sunny\n",
      "Weather = [['Hot' 'High' 'Weak']\n",
      " ['Hot' 'High' 'Strong']\n",
      " ['Mild' 'High' 'Weak']\n",
      " ['Cool' 'Normal' 'Weak']\n",
      " ['Mild' 'Normal' 'Strong']]\n",
      "PlayTennis = [['No']\n",
      " ['No']\n",
      " ['No']\n",
      " ['Yes']\n",
      " ['Yes']]\n",
      "0.9709505944546686\n"
     ]
    }
   ],
   "source": [
    "print('Overcast = Sunny')\n",
    "Weather_subset = Weather[Weather[:,0]=='Sunny',1:]\n",
    "PlayTennis_subset = PlayTennis[Weather[:,0]=='Sunny']\n",
    "\n",
    "print('Weather = {}'.format(Weather_subset))\n",
    "print('PlayTennis = {}'.format(PlayTennis_subset))\n",
    "\n",
    "H_Y = entropy(PlayTennis_subset)\n",
    "print(H_Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x36 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAloAAABUCAYAAABazpmzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4wLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvqOYd8AAADwxJREFUeJzt3Xu4XFV5x/HvT9NSkpIAipF7hAeIWrlI5FIuCSLeUKAJ2EILhIvB0iJWQSgCiaAQRC4SVK4FRAwWCqQSNQECEWjzkKRNkEstBNJSrkJCIFxD8vaPtUbmTGbmTHJmz+Wc3+d5zjMne6295z159pp591prr62IwMzMzMya7z3tDsDMzMysv3KiZWZmZlYQJ1pmZmZmBXGiZWZmZlYQJ1pmZmZmBXGiZWZmZlYQJ1odRtI9kvq05oak8ZJC0vgmhWVmZZrRTs2KJmmxpMVrUH+189rfJ33nRKsgkkbkk/P2OnVKJ/BJrYzNrK/Kzu+QdFuNOgfn8kktDs+s6fyZ/q6y/4tr2x1LNxjU7gBsNUcAg9sdhNkaOFDSbhExp92BmFkP+zbhGLcCc4Bnm3CsAcmJVoeJiP9tdwxma+BJYAQwGRjT1kjMrIeIWNSEYywDljUhnAHLQ4cdptbcD0lbS/oXScskvSJppqQdJF2bu3BH1DjepyX9m6TXJL0k6TpJ7yv677AB4xHgRmC0pM81soOk0ZJmSFoq6Q1JD0n6pqQ/qqg3pjT0KGlPSXfm839pjfLZkpZLelbSZEnvzfWOkvTb/F5PSDqmSkzbSjpf0oIc15uSHpZ0emVcZmsin6P31ChbbQ5V2Wf6VpJOkfR4PncXSPpsrjNM0mWSnstlsySNbOT4efuOuQ0uz+f7zZK2qBFjjzla+fXJXHxk2RSCyG3y2/n3Q2oc77RcfliN/7J+x4lWF5C0OXA/MBa4G7gUWAn8BvhQnV0PAH4BPAP8GFhEGpqcVmS8NuCcAawAzpWkehUl/RUwC9gduBmYAgg4D7ilxv575H1WAJcDlXPCdgVmAr/P5UuBU4Bz8lyZC4B5wNXAUOAqSftUHGMscBTw37nelaQ2djZwU/0/36wQFwEnAHcAPwW2BaZJ+gSpPewKTAV+DewDTC9dXNQjaXvgXtKw4i+Ay4BNgPuADRqIawHwg/z7QuDbZT+LSe1nFXB0lfdW3r4UuKWB9+oXPHRYvG3rTAbescFjTAaGA1+NiCmljZLOJJ3ctXwRGBMR9+f67wXuBMZ4To01S0QsknQlcDxwKPCzavUkDSMlQm8Cu0bEo3n7t4BfAV8gXQhcV7Hrp4AjI+InNUL4LLB/RPwyH+9M4HHSl9QSYKeI+J9cdg0p6foG6aKl5CfAhRHxdlm8Aq4AjpW0Z0Tc18B/h/V/zfhMb+h9gB0i4iUASb8mXZzMJCVXfxMRK3PZFODvgb/Ideq5FPhTYFxE/CHZkXQdqf3VFRELJF0MnAgsiIhJlXUkzQA+I2mziPi/sqLRwNbAlIh4s7f36i+caBVvG2Di2u4saR1gHPA0qVeq3PdJJ/uGNXb/WSnJAoiIlbkxjQE+QZrgaNYMZwNHAmdLuikiVlSpcxCpR+mSUpIFEBErJJ0KzM3HqEy05tdJsgBmlZKsfLzXJE0HjgEuLyVZuWy+pEXAx8oPEBHPVB40IkLSj4FjScmeEy2DPn6mr4FzSklWdivwNrA+8M1SkpX9nJRofYw6iZakLYG9SG2qskfpDOCvgV57xRpwBfA5YDzwnbLtpWH7q5vwHl3DQ4fFmx4RqvZDGqrozUhgHWBeRLxTXhARr5O6cWuZX2Vb6epi/UaCN2tERDwHXAxsBUyoUW2H/Dq7yv7zgOVldcrN6+XtF1bZ9mydsudIQyV/IOk9kr4s6X5JL0tapTRXstSGNu4lBhs4+vqZ3qge525ErAJeAJZGxFMVdZ/Lr5tQ3/b5dbWLhnwjVrNuxrqdNGXlqNJ0gNyjPY70XVatXfZbTrQ633r59fc1yl+os+8rVbaVkrVmXLWYlTsfeAk4Q9KQKuVD8+vzNfZ/vqxO5fZ6qp3nK+uUvcPqvflTSFfhm5LmjkwmDcuX5qKs00sMZs1W67yu97ne240bw/Jrre+T3tpaQ3KnwDWkC6/SfMjDgHUZYL1Z4ESrG7yaXzeqUf6BVgViVk++Dfxc0nzCf6hSpfQFMbzGIYZT/Uuk0BXYJQ0H/pbUg/DhiDg6Ik7Lc0+mFvneNiAEtafpDKuxvSilZRpqfZ/Uaptr4yp6Too/BnidGnM4+zMnWp3vv4C3gFGSejRWSevS3MmXZn31Q+Ap4GSgchmR0jD33pU7Sfo4aYJuvaHwonyIdOfjnRHxRkXZHm2Ix/qXpaSe0h6UluRp9RSOB/PrnpUFeXmHqks8VFHqMa45MhIRi0l3TI6TNBrYGbgpIqpdTPVrTrQ6XES8RZoEuSlwXEXx16k9Ed6s5fKdRJNIQ4AnVxRPI/VYHStp29LGfAExOf+z3qT3opTmpexevrxEjvEf2xCP9S9zgRE52QBA0h8DF7Y6kHxjyL3AzpLGVhSfTeNTSpaSeuo276XeFcCf8G4v1oAbNgTfddgtTiWteTJF0qeAR0k9WXuQGs1epC5as05wHXAS8OHyjRGxTNJXSGsCzZX0c9IH9v7AR0kTaFueaEXEM5JuJd0aP1fS3aRJxQcAM0gTeM3W1oXAp4FfSppKGj7bD3iZ9jzW5gTSZPh/lnQTae2r0cBmpB6v7WvvmkTEcklzgb0lXQ88RvoOur78Ll/SOl3Pk9rT7yLi3mb+Id3CPVpdIJ+4f07qEdiXdBvvINIQTGnM/dXqe5u1Vr7t/LQaZVNJ5/Ac4Euk5UlEupgYGxGFzseq40jSxPeNSF9EOwLfYvVeObM1EhEzSef6IuBw4BDSkNp+pOUaWh3PQtJ3xyzSxcRXSHcI7km68GnU4by7/t0kUo9YjwW08zIvpd6sf+pL3N1MvX2uSRpDz4X9SpZFhJcIaKO8AOkiYHBEeFK8mZl1FEl3kRK7zSKiKXc1dps1GTr8KmmsueSdWhWt+SRtHBGV3cwnA1uS7u4wMzPrGJJ2BD5JmgQ/IJMsWLNE61E/sqWtHpX0AGl+FsAuwG6k8e9J7QrKzMysXH5g9HakIfmVwHfbG1F7eTJ897icNGl4V2AwaSXgq4GzIuLpdgZmZmZWZgLpJq0ngMMH2krwldZkjtYLwPtJd0rMAE7NS/abmZmZWRWNJFo7kR40OZu0Bs5OpDuKVgA7RUS9R8CYmZmZDVi9JlpVd0qrOD8ATI6I06uUTyA/WHbIkCE7jxw5sq9xmjXF/PnzX4yIWo+fKIzbhHWqdrUJs4FirRItAEmPAE9FxGfq1Rs1alTMmzdvrd7DrNkkzY+IUe2MwW3COkkntAmz/qyvC5a2a3FBMzMzs463VomWpFGkWzcfaG44ZmZmZv1Hr8s7SLoBeBL4D9IdhzuRHrT6NHBJodGZmZmZdbFG1tF6CDiU9Pyv0vpNtwATI+LFAmMzMzMz62q9JloRcS5wbgtiMTMzM+tX+joZ3szMzMxqcKJlZmZmVhAnWmZmZmYF8UOlu8SIU6ev9b6LJ+/fxEjMzMysUe7RMjMzMyuIEy0zMzOzgjjRMjMzMyuIEy0zMzOzgjjRMjMzMyuI7zo0s67kO3HNrBu4R8vMzMysIE60zMzMzAriRMvMzMysIE60zMzMzArSUKIlaXNJN0taJukVSbdI2qLo4MzMzMy6Wa+JlqTBwCxgJHAkcDiwDXC3pCHFhmdmZmbWvRpZ3uHLwFbAdhHxOICkB4HHgOOAC4sLz8zMzKx7NTJ0eAAwp5RkAUTEk8D9wIFFBWZmZmbW7RpJtD4KPFRl+8PAR5objpmZmVn/0UiitSGwtMr2JcAGzQ3HzMzMrP8o5BE8kiYAE/I/l0v6XRHvM8C8H3hxbXbUeU2OpLtt2Y43dZsohNtEc7SlTZgNFIqI+hWk54HbIuK4iu0/Ag6JiI0KjM8ySfMiYlS74zDrFG4TZtYNGhk6fJg0T6vSR4BHmhuOmZmZWf/RSKL1r8BukrYqbZA0Atgjl5mZmZlZFY0kWlcCi4Fpkg6UdAAwDXgKuLzA2KynK9odgFmHcZsws47X6xwtgPy4nYuA/QABdwFfi4jFhUZnZmZm1sUaSrTMzMzMbM019FBpK5ak8ZJC0suSNqgoG5TLJrUpPLNC5QfWL5E0vErZGEmrJJ3YjtjMzPrKiVZnGQac0u4gzFrs74AALi3fKGld0hzRfwemtCEuM7M+c6LVWWYCJ1S7sjfrryLieeBE4GBJB5UVTQI2A46OiFXtiM3MrK+caHWW7+TX0+tVkrSLpDslLZf0mqS7JO3SgvjMChERPwWmAz+UNEzSx4FvAJMioscq+pL+UtIDkl6XtFTSjZI2ragzXtLC3D6W5d+Pbt1fZGaWONHqLM+Shk8mSKr6WAxJ2wOzSc+ZHA8cAQwFZkvaoUVxmhXhOGAIcDFwNfCfwPfLK0j6GjA1l40Djgd2Bu6WNDjX2Re4BrgDOAD4EnAtsH4r/ggzs3KFPOvQ+uQ80hfORKDaFfiZwFvAvhHxMoCkO0hrnU0ExrYmTLPmioinJZ1Empe1Atg5IlaWyiWtD3wXuCwiji/bPp/0lIojgMuA3YFnI+KkssPPaMGfYGa2GvdodZiIWAJcABwhabsqVfYGbi8lWXmfV0ir9I9uTZRmxYiIq0g9u7dFxG8rivcCBgM35LtxB0kaBDyRf/bO9eYCG0u6VtLnJQ1tVfxmZpWcaHWmi4AlwFlVyjYkfRFVeo40nGjW7d7OP5U+kF/vI/V4lf9sA7wPICJmAIcCW5OeYvGSpBmSqj2z1cysUB467EARsVzSuaSerfMripcAH6yy2weBpUXHZtZGL+XXw4DHqpS/UvolIm4EbpS0HvBJ4HukyfYjCo7RzKwHJ1qd60fA13n3TsSS2cDnJa0XEa8C5C+TLwL3tDRCs9b6DfAGsFVETG1kh9xGpuVh+PMkDc1D7WZmLeFEq0NFxFuSzmL1B+eeDXwBuEvSeaSFHk8hzV2pNtRo1i9ExBJJpwIXSNqENMH9VWBTYB/gVxFxs6TJ5DtxScPsW5DuTpzjJMvMWs1ztDrbNVQMkUTEg8AY0jDJdcD1wHJgdEQsbHWAZq0UEZcABwN/BtxAGg6cSLrgKE2en0Oas/UD0hIP55AWAz6o8nhmZkXzQ6XNzMzMCuIeLTMzM7OCONEyMzMzK4gTLTMzM7OCONEyMzMzK4gTLTMzM7OCONEyMzMzK4gTLTMzM7OCONEyMzMzK4gTLTMzM7OC/D+46FNTEj1qGgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 720x36 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x36 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "IG = np.zeros(len(X_names)-1)\n",
    "for i_var, X_var in enumerate(X_names[1:]):\n",
    "    values = np.unique(Weather_subset[:,i_var])\n",
    "    n_values = len(values)\n",
    "    H_YX = np.zeros(n_values)\n",
    "    f, axs = plt.subplots(1, n_values+1, sharey=True, figsize = (10,0.5))\n",
    "    for i_ax,ax in enumerate(axs[:-1]):\n",
    "        value = values[i_ax]\n",
    "        # plot histogram\n",
    "        PT = PlayTennis_subset[Weather_subset[:,i_var]==value]\n",
    "        ax.hist(PT); ax.set_title(value); ax.set_ylim([0,5])\n",
    "        # compute H(Y|X=V_i)xP(X=V_i)\n",
    "        P_V_i = np.mean(Weather_subset[:,i_var]==value)\n",
    "        H_YX[i_ax] = entropy(PT)*P_V_i        \n",
    "    axs[-1].set_title(X_var); axs[-1].axis('off')\n",
    "    IG[i_var] = H_Y - np.sum(H_YX)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZQAAAFHCAYAAACGb7ohAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4wLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvqOYd8AAAIABJREFUeJzt3XmYXGWZ/vHvnYUtIIuEcWEJEQYIOjoYhRFEFgGVTdkEBBGBoA4MoCD7KiACsonIIooCjiyCgCCCbCLD8gsqSFAkmhhQlAAh7EuS5/fH8xY5Fp10dffpqnT3/bmuurr71DnVb6XSdde7KyIwMzPrq2GdLoCZmQ0ODhQzM6uFA8XMzGrhQDEzs1o4UMzMrBYOFDMzq4UDxczMauFAMTOzWrQ9UCQtL+lbku6W9JKkkDSmxWuHSTpU0lRJr0h6QNK2/VtiMzNrxYgO/M5VgB2A+4E7gU17cO3XgAOBw8v1OwJXSNoiIm7o7uJll102xowZ0+MCm5kNZffff/9TETG6u/PU7qVXJA2LiDnl+z2BC4CVI2JqN9ctBzwGnBQRR1eO3wKMjoj/6O53jx8/PiZOnNiX4puZDTmS7o+I8d2d1/Ymr0aY9MJmwELAJU3HLwHeI2nlPhXMzMz6ZCB1yq8JvApMbjo+qXwd197imJlZ1UAKlGWAZ+PNbXTPVO43M7MOGUiB0iuSJkiaKGni9OnTO10cM7NBayAFygxgKUlqOt6omTxDFyLi/IgYHxHjR4/udpCCmZn10kAKlEnAwsC7mo43+k4ebm9xzMysaiAFyo3A68Bnmo7vAjwUEVPaXyQzM2voxMRGJG1Xvn1/+fpxSdOB6RFxRzlnFvCDiNgDICKelHQacKik54HfAJ8GNgK2ausTMDOzN+lIoABXNP18Tvl6B7BB+X54uVUdDrwA7Ae8DXgE2CEiftY/xTQzs1Z1JFAiorljvaVzImI2cHy5mc3XmEOu73QRBq2pJ23e6SLYAmgg9aGYmdkCzIFiZma1cKCYmVktHChmZlYLB4qZmdXCgWJmZrVwoJiZWS0cKGZmVgsHipmZ1cKBYmZmtXCgmJlZLRwoZmZWCweKmZnVwoFiZma1cKCYmVktHChmZlYLB4qZmdXCgWJmZrVwoJiZWS0cKGZmVgsHipmZ1cKBYmZmtXCgmJlZLRwoZmZWCweKmZnVwoFiZma1cKCYmVktHChmZlYLB4qZmdXCgWJmZrVoe6BIWkHSlZJmSnpO0lWSVmzx2hUl/UDSNEkvS/qTpOMljervcpuZ2fyNaOcvk7QYcCvwKrAbEMDxwG2S/iMiXpzPtaOAXwIjgSOBacAHgGOBVYFP92/pzcxsftoaKMBewFhgtYiYDCDpQeBRYG/gtPlcuy4ZHJtFxE3l2G2SlgEOlLRYRLzUf0U3M7P5aXeT11bAPY0wAYiIKcBdwNbdXLtQ+fpc0/FnyeehugppZmY91+5AWRN4qIvjk4Bx3Vz7S7Im8w1J4yQtLmkjYD/g3Pk1l5mZWf9rd6AsA8zo4vgzwNLzuzAiXgHWI8s8CXgeuAX4GbDPvK6TNEHSREkTp0+f3ttym5lZNwbMsGFJiwCXAcsBuwIfAQ4iO+O/Pa/rIuL8iBgfEeNHjx7dlrKamQ1F7e6Un0HXNZF51Vyq9gA2AFaJiD+XY7+SNBM4X9K5EfFAbSU1M7MeaXcNZRLZj9JsHPBwN9e+B5hRCZOG+8rXNfpYNjMz64N2B8q1wDqSxjYOSBpDDgm+tptr/wEsLWmVpuNrl69/q6mMZmbWC+0OlAuAqcA1kraWtBVwDfAYcF7jJEkrSZol6ajKtReRHfE3SNpN0oaSDgJOBe4nhx6bmVmHtDVQytDejYA/ARcDlwJTgI0i4oXKqQKGV8sXEVOBdYDfkbPrbyAnSp4PbBIRc9rwFMzMbB7a3SlPREwDtu3mnKl0MVExIh4GduifkpmZWV8MmGHDZma2YHOgmJlZLRwoZmZWCweKmZnVwoFiZma1cKCYmVktHChmZlYLB4qZmdXCgWJmZrVwoJiZWS0cKGZmVgsHipmZ1cKBYmZmtXCgmJlZLRwoZmZWCweKmZnVwoFiZma1cKCYmVktHChmZlYLB4qZmdXCgWJmZrVwoJiZWS0cKGZmVgsHipmZ1cKBYmZmtXCgmJlZLRwoZmZWCweKmZnVwoFiZma1aHugSFpB0pWSZkp6TtJVklbswfVrSLpC0lOSXpb0iKT9+rPMZmbWvRHt/GWSFgNuBV4FdgMCOB64TdJ/RMSL3Vw/vlx/O7AnMBNYFVi8H4ttZmYtaGugAHsBY4HVImIygKQHgUeBvYHT5nWhpGHAD4FbIuJTlbtu67/implZq9rd5LUVcE8jTAAiYgpwF7B1N9duAKzBfELHzMw6p92BsibwUBfHJwHjurl2vfJ1EUn3SHpd0pOSzpK0aK2lNDOzHmt3oCwDzOji+DPA0t1c+47y9TLgJmAT4GSyL+VHdRXQzMx6p919KH3RCL9LIuKo8v3tkoYDJ0laIyL+0HyRpAnABIAVV2x5MJmZmfVQu2soM+i6JjKvmkvV0+XrzU3Hbypf/7OriyLi/IgYHxHjR48e3XJBzcysZ9odKJPIfpRm44CHW7h2fub0qkRmZlaLdgfKtcA6ksY2DkgaA6xb7pufn5PzVzZrOv6x8nViPUU0M7PeaHegXABMBa6RtLWkrYBrgMeA8xonSVpJ0ixJjb4SIuJp4OvAFySdKOmjkg4BjgJ+UB2KbGZm7dfWTvmIeFHSRsDpwMWAgFuA/SPihcqpAobz5sA7Dnge+BJwIPAEcArwtX4uupmZdaPto7wiYhqwbTfnTCVDpfl4kBMb2z65ccwh17f7Vw4ZU0/avNNFMLMaeLVhMzOrhQPFzMxq4UAxM7NaOFDMzKwWDhQzM6uFA8XMzGrhQDEzs1o4UMzMrBbdTmyUtAiwL/D7iLixHBtG7ute9RywTUS8VnchzcxswddKDWV34Fjgj5VjIndQXAyYXW4fAj5Xc/nMzGyAaCVQtgEuLsuhNJsQERtGxIbAWcB2dRbOzMwGjlYC5b3AjS2cd18518zMhqBWAmUp4MnqgYiYDewE/LlyeGY518zMhqBWVhueCSzXfDAiLms6tBzZMW9mZkNQKzWU3wJbtHDeFsDv+lYcMzMbqFoJlIuAXcvuil2StDWwK/D9msplZmYDTLdNXhHxI0nbAldJuha4HphW7l6RrJlsCVwbET/qt5KamdkCrdUdG3cAjgT2Bz4JRDkust/kBHJ7XjMzG6JaCpQyqusYSScC44Hly12PAxM9O97MzFpZemVs06F/lFvD8tLc7d8j4i/1FM3MzAaSVmook5nbxNWK4b0si5mZDWCtBMru/V4KMzMb8FoZ5fWDdhTEzMwGNu+HYmZmtXCgmJlZLRwoZmZWCweKmZnVwoFiZma1cKCYmVktHChmZlaLtgeKpBUkXSlppqTnJF0lacVePM4hkkLSr/ujnGZm1jNtDRRJiwG3AqsDu5F7qKwK3CZpVA8eZyxwBE1bE5uZWee0unx9XfYCxgKrRcRkAEkPAo8CewOntfg43wEuBVaj/c/BzMy60O4mr62AexphAhARU4C7gK1beQBJOwNrAYf2SwnNzKxX2h0oawIPdXF8EjCuu4slLQ2cDnw1Ip6puWxmZtYH7Q6UZYAZXRx/Bli6hetPAf5E7nPfEkkTJE2UNHH69OmtXmZmZj00YIYNS/ow8FngixHR8v4sEXF+RIyPiPGjR4/uvwKamQ1x7e7QnkHXNZF51VyqzgMuBB6XtFQ5NgIYXn5+OSJera2kZmbWI+0OlElkP0qzccDD3Vy7Rrl9oYv7ZgAHAGf0qXRmZtZr7Q6Ua4FTJY1t7D0vaQywLnBIN9du2MWxM8gth/cltyo2M7MOaXegXADsA1wj6Qhyr/qvAY+RTVoASFoJ+DNwXEQcBxARtzc/mKRngRFd3WdmZu3V1k75iHgR2IgcqXUxOTlxCrBRRLxQOVVkzWPADBowMxvq2j7LPCKmAdt2c85UMlS6e6wN6imVmZn1lWsAZmZWCweKmZnVwoFiZma1cKCYmVktHChmZlYLB4qZmdXCgWJmZrVwoJiZWS0cKGZmVgsHipmZ1cKBYmZmtXCgmJlZLRwoZmZWCweKmZnVwoFiZma1cKCYmVktHChmZlYLB4qZmdXCgWJmZrVwoJiZWS0cKGZmVgsHipmZ1cKBYmZmtXCgmJlZLRwoZmZWCweKmZnVwoFiZma1cKCYmVktHChmZlaLtgeKpBUkXSlppqTnJF0lacUWrhsv6XxJf5T0kqRpki6VtHI7ym1mZvPX1kCRtBhwK7A6sBuwK7AqcJukUd1cviOwJnAW8HHgEGAtYKKkFfqt0GZm1pIRbf59ewFjgdUiYjKApAeBR4G9gdPmc+03ImJ69YCku4Ap5XGP6pcSm5lZS9rd5LUVcE8jTAAiYgpwF7D1/C5sDpNy7K/AdOCdNZfTzMx6qN2BsibwUBfHJwHjevpgktYAlgP+0MdymZlZH7U7UJYBZnRx/Blg6Z48kKQRwLlkDeXCvhfNzMz6YiAPGz4b+BCwS0R0FVIASJogaaKkidOnv6nVzMzMatLuQJlB1zWRedVcuiTpJGAC8PmIuGl+50bE+RExPiLGjx49ukeFNTOz1rV7lNcksh+l2Tjg4VYeQNLhwMHAvhFxcY1lMzOzPmh3DeVaYB1JYxsHJI0B1i33zZek/wGOBw6PiLP7qYxmZtYL7Q6UC4CpwDWStpa0FXAN8BhwXuMkSStJmiXpqMqxHYEzgBuBWyWtU7n1eISYmZnVq61NXhHxoqSNgNOBiwEBtwD7R8QLlVMFDOdfA+9j5fjHyq3qDmCDfiq2mZm1oN19KETENGDbbs6ZSoZH9djngM/1V7nMzKxvBvKwYTMzW4A4UMzMrBYOFDMzq4UDxczMauFAMTOzWjhQzMysFg4UMzOrhQPFzMxq4UAxM7NaOFDMzKwWDhQzM6uFA8XMzGrhQDEzs1o4UMzMrBYOFDMzq4UDxczMauFAMTOzWjhQzMysFg4UMzOrhQPFzMxq4UAxM7NaOFDMzKwWDhQzM6uFA8XMzGrhQDEzs1o4UMzMrBYOFDMzq4UDxczMauFAMTOzWjhQzMysFm0PFEkrSLpS0kxJz0m6StKKLV67iKRTJD0h6WVJd0tav7/LbGZm3WtroEhaDLgVWB3YDdgVWBW4TdKoFh7iQmAv4ChgC+AJ4BeS3tc/JTYzs1aNaPPv2wsYC6wWEZMBJD0IPArsDZw2rwslvRfYGfh8RHy/HLsDmAQcB2zVv0U3s/405pDrO12EQWvqSZu35fe0u8lrK+CeRpgARMQU4C5g6xaufR24rHLtLODHwGaSFq6/uGZm1qp2B8qawENdHJ8EjGvh2ikR8VIX1y4ErNL34pmZWW+1O1CWAWZ0cfwZYOk+XNu438zMOqTdfShtJ2kCMKH8+IKkRzpZnjZZFniq04Volb7R6RIsEPyaDTwD5jWr4fVaqZWT2h0oM+i6JjKv2kfztV09qUbN5Jku7iMizgfOb7WAg4GkiRExvtPlsNb5NRt4/Jq9WbubvCaRfSHNxgEPt3DtymXocfO1rwGT33yJmZm1S7sD5VpgHUljGwckjQHWLffNz3XASGD7yrUjgE8DN0XEq3UX1szMWtfuQLkAmApcI2lrSVsB1wCPAec1TpK0kqRZko5qHIuI35JDhs+QtKekjckhwysDR7fxOQwEQ6qJb5Dwazbw+DVrooho7y/MZVZOBzYBBNwC7B8RUyvnjAGmAMdGxDGV44sCJ5ATHJcCHgAOjojb21J4MzObp7YHipmZDU5ebdjMzGrhQDEzs1o4UMzMaiRp4Va35BhsHChWG0nDO10Ge2M4PZLU6bIMNeVv4P+AcyQNufUFHShWC0kjImJ22QRtC0mfKcPCrc0iYlaZAHy+pCU6XZ6hJCJmA98FPgocLWnVDheprQb9Wl7W/yQNL29iSwB3AIsD7wSGS/oVcAxwb/ljs/Z4H7AH+Xpc0uGyDBnlb+E7kl4mg2W2pG9ExB86XbZ2cA3F+qxRMwF+CTxH7sb5fmDDcjsW+LfOlXBI+jNwL7Clkv/W+1mjll5+vA34DvBZYD9J/965krWP/5NZXT5KLvx5OFkb+SOwIjCcXBrn750s3GBWDYvG9xHxT+Bicqmi90TEnA4Vb0iQpLLhH5IuAa4E1iBXAZkAHDEUmr8cKNYrjQ74ypvZWHLl5z9GxBxJOwH/CxwWEadIequkvTpU3EFL0rDy7z1S0qjyfaMz/mfAI8DekkZ2sJiDXpQZ4pK+SX64OgzYDtgAOAj4DHDUYK+pOFCsx0o78WxJbyGbtCD3hVgSGCbp48ClwOERcVJ5g9sU2EvSGp0p9eBTXoc5Zfvrm4DbJG1G2eYhIqYBdwJbAIuWazzyq59IGgWsD1wfETdHxLMRMSUivgnsSy4ZdZCk1Tta0H7kQLEeKVX72WVdtcan3+WA35affwVcT67P9vVy2RrAF8gtCP7YgWIPGpIWlzS81Exml9FcbwXuAp4kV+2+RtKxkt4GnFwu/TLM/SRt9SpBPYwckNI4Nrxy32XkcOLdgGMkvasT5exvDhRrWflEHKWZaxPgfrJJ68kyiuUSche7PwE3SRpVVoX+PrAYsFfleuuhUhP5AnB6qZksDjxdjh0VEVuQzSy/BPYnF149H3gWeF/5EGA1aK7pRXqeXLB2U0nLlsAfWe57GngceIhsEnul/aXuf14c0nqkfCI+A1gBeD4idmi045f7DwN2BN5F7rI5E/gbsHlEvN5oLutQ8Qe00qSyI7kNxBXAB8iN5fYEHqvWPspM7c8CHwE2Loe3jIjr21roQaj5/3AZ3dXokH8vcAP5unwiIl4sx98JnEuutH53RLzc/pL3PweK9YikDwL3AK8Dl0bE58vxkRHxevl+dXIexGLk8NU7yyfqN/7wrHWleWRaI5DJ5qtvAE8A60XElHJeo4O++ev2ZMfwVDJkXnXTV+80fXg6jtwxdmHg5Ii4swx+2A34Otmv+G0gyL6VDwPrNl6vwciBYvPVVQhIWo8cFrkUsFtEXFaOvxEqXTzOG3+I1royKuhm4KCIuLwcO5PsaF+efB0+39WOpU1vfkeRQbRaGVJsfSDpx8B65AemfyNr5PtFxDmlaXIdcv7V+4A55PDhz0bEAx0qclu4Ldvmq8yAH1U+jTWO/Zqc3/AsOWply3L89Xn1jzhMeu1J4MiIuFzSQuXf91iyD+tAYBvg+2Vi6RvK4Ik5ldfjLrLdfsitL1UHVdapk7Qu8Hbyb2BT4GPARcCZkvYHZkXEHRGxAVkz+QCw8WAPE/DSK9aabciJWe+MiD0ASvV+Z+BH5Ph6IuK6xjwIN6n0TelwfzUingV+WD71Xgf8nRzc8IykH5XTTy7XfLZ8AHg7+XodFxH/LG+Ge5KTTCe3/9kMbI0RdeX7o8kh2M8Avyk1w6mSDgFmAacAcyRdGBEvRsSDHSt4B7iGYq34OfAVYBdJFzUORsSt5Nj65YHDS1u9h6b2UZmr8xtg/cpoooWAV8l5P6eV5sWnycmjXwU+CVwtaU/gJ+Ror6fKtW8hazofc3NXa8rw7MslvaXSbPhW4Evkv/eS1WbG8locQQ6YOAnYp7nWOBQ4UOxfqCx9XhURTwE/IJdV2XkeofJBYLM2FXMomAOcB3y49GM9D+wK3A58Cji9hMpTZKjsQ7bbf41s2lq+DFsdFhEzgK9ExP2deCIDTQnxnYCRwBujuUporA3cTYb9pyUt1HT/EWS/1oGUyaRDiTvl7U3K8NQNgZ83DY9cBtid/AT2fWDvypITawEPeEhw35V+j9WB75G1v88A90TEq5KWAs4CNgJ+ChxQGf01muwcvtuj6nqnDIs/CzikhDWSDga+FxHTy89jyObHUcD/AL+oDkaRtDSwcET8o72lXwBEhG9D/AYIGFb5+RTyE/Iu1ePlvreRiw7OAc6kfCip3D+8089nIN+AkY2vwHhyDs9vyc7dhcp9S5E1xseBbzWuaXqcYe0q82C6kR3sc4BvlZ83IvtGbgeWqZw3hlz54S/kiLs3vQZD8eYmryGsfBoj0pzKTOpvkGtxnU/2m7zx/yTyU9ePyfb5fYEjq48ZrqH0Wpkw97pyjbSryNFcrwHvJffWWKc0cz0L7EfOhN8K+J6adssMj6rrrfuAy4HtJH2U3E/mEDJAri79KETEVGBz4GXgm+Q2AUN+AU4HyhBVwmRfSV8rPy8MTJd0YmRV/1DgarIdf5fGH0tpXx5LrmS7NXBiJ8o/GMXcfWV+RTannEU2PU4g37guBv5L0kKVUHkAWIKcPGe9pDQsIp4BjiMn5X6pfEA6k9zbZGXgqqZQ+QT5738kOcFxaOt0Fcm3ztzIP4JjyOr9ucCjwC+AMZVzlidrKi+Tnb7jgDXJ6v/RlfNGdPr5DOQblWZD4OPkkNSPVo6NJDcse5BcgHN9so0ecjHCYeV7N3P1/N9+MeDdXRz/Yvnb2KPyGnwVmEbWWt5aOXdFYGynn8uCcOt4AXzr4IufbfHfLn84U5nbRj+8cs7byU9nc8iFCJ8AJjZChKY+FN9a/rf/MPCjLo7vRLbZjy8/N/6dR5KLQM4p//6bNL1ODpOevwYLlf/Pc8jm2zUr961CLrI5rfpakEvYTCObG5ft9HNY0G5u8hrCIptNliY7flcETi3HZzeGD0fEExHxRWAHcgLdicDakRPohkf5S7PWlX/b9ciAbjaJDJT14Y2VCkZEjiK6lQz+NYF9otJfFe4z6bGIeI2s9UHOszpS0oHlvsnAhcBywE5lPsoscnHHM8jhwxfNa2WIocrDhoe4stjjy8DnyTb5b0fEvuW+ea4M7FWD+0bSYhHxUhmifWREHFKOL0F28u4G7BwRP6lcsyXZn/IVYLJDpPcqC2duRc7veZxc8PSTwBTgCxExRdKp5GTGj0TE/yvXjiCbxG6MiEc78wwWTA6UIaSbgFgJOIAcV18NlXeQCxMe0L6SDh2S9iBnV/84InYux9YjO4bXIyeT3kU2Tx5J1ia3j4hwqPedpCXJgSdzyNFcy5JzrEaSE0ZvBU4gVxtYp9RqbB4cKENEY5JbGd11IDkB7iVyLa57IoerjiFrKfuRi91dTs78XRV4h9+86idpNLnO1nHA1RGxQzm+LvnJeTfyze05Sod8ea28XloPldrgF4CfRcQjleMrk6PlTo+Io8uxE8lFHVcB/kAOijgbOME1w3lzoAwBmrsH/BLkCJXh5M5x/0U2d10AnBMRr5VQmUB2Ur5CvoltWN7EvAR9H8yrRlFCZQI57+TqiNi+HB9Fzn9YjQyU2xr9W+EZ8D1WVgI+DfgdcElEnFa5r7GHyRcj4ppy7APkskL7ldNuAT4VES+0teADiANliCiTFn9Brk30mYj4u6SrgS2Bf5Cz479TQmUZ5u7x8HO/ifVdJdQXBbYl13m6Bfh7RLwiaTlgLzJUrmrUVOb1OG0r+CBS5lKtDZxD1rp/TdbWHyGHD58LvAAcERF/r1y3Ddln8j+RW13bPDhQhghJu5NNK7uUzsargbXI5VXOJYcHH0epqTRd6zexGpQl6e8hRw4tSw5Z/TbZZzWzEirHAFc0+lSsXqVf8HNkrXBxsg/lDLIm+FNgz4j4qf51F9JFImJQ7gNfJw95G6Qaw34rS2j/mmxOaYxceR+wQ0TcSS74OIJsrz+kecVhh0nfVIaWHkfOYdiSXG3gIbI55VBJS0bEk2Tz41HAjqpsamb1KbWPU8l1um4nd7K8j9we4BrgbEmjY+6imzhMWuNAGYRKh+2ssibUA2Vo5GRyH423AB8hayWN5cxfJPtSxpK1FgdIDSpvRo1+p5HkaK57I5ft2IIcwbUH/xoqF5GfoB0o/SQiXouIv0TEduTIxifIZenHkKtInCBpUX+Y6hkHyiDTmGxY3sxOIpfxeDTSHGAZYA1yVnyjT+RtwPVkoGxTrldXj2+tqfSZLCZpv9Lk+AHy9Wg0obxOThht1BIPlrR0mUz6w8akxs49i8GtUXOMiAvIzvdDyT1lliBHdXltrh7yf9ZBptLxuxG5mN2ZwJ8qpzxHjq3fSdITZIf8YcDMyA2C3GdSg/I6LE4uk/JWcrHHRch9yH9WOuIXjtzjZHvgMnIexGPkUjeNx/FAiH4Sle2qS43xFEl3kUOLTywrSVgPuFN+EJJ0Hbm09jRg44j4czUkJH2CnF/yQXL5j0eZOzTY8xv6oDLfZxhwNFkr+SqwJLkawe7AaRFxYDm/ESojyCauIx3mndH4v1/tjLeecQ1lcNoDuIHsD9lL0jHlE/HwiJgdETdImkw2dY0id5zzDn81KGEyCtgGWAm4LiIeApD0N+B54MuSiIgDS5g0RhAdVs7z69ABjQ9SDpPec6AMcF1NNoyIJyVtRk5i3J3smL8i5i7oODsi/kSlKawc95tYPfYmRxG9RPZNAbl/hqTGZLovS5oTEV9tHkHk18EGKnfKD2Dlk+wcSQtJWkPS+pKWLs0oT5P9KDPJVYK3L+fPVhcrpLqZpffUtFsiuSLtkeRkuZ0krdi4IyKmkYs/ng4cKGmfthXUrJ+5D2WAalpO5TqyA/7tZJ/IWcBPI+IPZVmPX5Mzsw8iZ2G7Sl+zskba+hFxY/l5GNknchi5pMfZEfFE5fyVyfko57hGYoOFaygDVGNIKhkWc8jJWRuQ28SeABwkadmImA6sSy4pcQk5B8VqVMLjdOAGSTtCjiCKiCPIJW0OJbdbfnvjmoiYEhFneWiwDSb+jzwAVUZi7UruJf5l4IEyQmVxcn2iuyPiqdLM9ZSkDcnlJW7rXMkHp9LseCVZSzyjvD7/W+47uEzpORiYLem8iHi86XrXUGxQcKAMQJVhvePIuQ2PlDDZBfghubjdBcq9HsZJ+m1E/JPcXtbzTPqgTPgcVhmCPTIiXo+ImyW9Ri6bckYZxVUNlTnk3iaPk2tHmQ06bvIagCqz2EcCsyLiZUk7k2FyWEScWJph9iEXG1y0er3DpOfKUGDKJLjZkhYtS3O8Lmmhct8dZL/JQ8CZkrZrXB8RhwL/TW4razYoOVAGgOZRWZUayk/IGsiPmVsSweYHAAAHf0lEQVQzOanctzqwCTnKyzN++0DS+4GfSKr2P10P/FXSqMgl/6uhchLZr/XdplD5jvtMbDBzoCzgKkODF5b0HkkfLt8vRM4z+R6wFXBzqZksKmmdcnwRcvter83VN8sAmwKHS/pQOXYpMAu4t4tQuZmsicwGLpe0afXB3Gdig5U/KS3AyqTFWWVo8E1krWNJYCrZDn8+c1ek/bykO8m9x18nVw/eoDqZse1PYBAoHew3S/oYufrAwpL2i4gLJb1ErpV2n6QPRsSLldrk28gNtKaUr2aDnuehLKAaM+DLp94bydrk98lltr8IbAj8jNyq9wVyOPDm5Ozs35MbNHmnxRpU1njanNwv41fAoRFxr6SdyFB5ihy2/Sw5H+gC4IyIuKE8hkPdBj0HygKoukgduYDjCeQb2N2Vc04hO9xPIN+43jRZ0W9i9SjNhSNLs9a+ZID8Ajg8In6jXC34ZLL2+BDwDjLk319C3Qtu2pDgQFlAlY7be8kRWi8BH2z0pUTEq+Wca4A1gdVL01YjiPwG1kuSxgObRsSJlWONVQnGk81X9wAbk5tjHVBC5T3kqLrlyFWev+LmRhtq3Cm/gCrNVD8gJ8utRXYKU1anbWz8811yU6zx5b6ofrWeKf0f2wLHSzqiHGuEyVrkIIjLImIzsnlxXeD00n/y+4jYG9g+IvZrjOZymNhQ4kBZQHS1YCNwNrly7RxgT0mrQYZKuf/fgBnk0GDro8hVm08il0s5TtKxlTC5k1y65oBSA/wFuYXvh4ATJW1cHmNW5fHcd2VDikd5LQA0d1OmkeT2o3Mi4tnSxHUxOfz3nDxVF5LNLmuTfSi/Bx7pVNkHm4iYKekEYDhwZFl/a0dKmETES/BGzeXG0lF/I/AHPJrLhjj3oXRYZTTXEsCPgX8H/kmuFnxq5by9yK1hhwG/I2slLwGfKp3Fb9oXxXqvvB5Hkv0ik8k+rFeazmn0Wa0N3O8aiQ11bvLqsEZHO/Bzcg7JdWQT1/GSzq6ceiG5hexscte/oyNi88aEOodJvSLieeBEckTXu8nFHZvPaQyAuNcz4M3c5NUxTTWKlchhpl+NiAclLU+uILx/aVr5YgmeS4GFyEmNu0v6a0T8NSJe68yzGNwi4llJXyfXTDu6LPh4bNM5UfneNRQb0hwoHdDUZ7IY8C5gWkQ8CBARj0v6Jrk0/QGlZeVLZcTQd5U7BJ4BvEXSVyJiaoeeyqAXEc9JaqxGcJRy296vdbRQZgsoB0qblSaSxnIq15Kzqt8BPCHp3RHxEEBE/E25/3iQNZWlImLnct95khYh2/hdO+lnlVCZAxwr6W8R8b1Ol8tsQeNO+TaqdMCPJCfFvQrcD4wm9yo5DTg1Iv5RueadwLHAGsCHq30lJWS8knCbKPeX2RU4181bZm/mGkqbNK3NNQ54lOxYnyxpUXIRwcOAWZLOjLL/eKmpHAo8VTqB35h57TBpr4iYSc4NeqPZssNFMlugOFDapBIm1wIrkkN+p5T7XgaOUO7q15ihfUajphK5L3wjlDzzegHgMDF7MwdKG5Uhvv8E/pNcSPCtwJON2ktEHKW5W8UuJenwiHi6cr2HBpvZAsvzUPpRWaW28f1wgIjYDbgYWBY4R9LoUnsZVu4/hmxWeTfwTNsLbWbWS+6U7yeVocHDAAELN5btKPd/C9iOXK7jgIiYXp2b4pWDzWygcQ2lH5SO81mSFidrI78GHpF0VmURwX2BK8hl0E+v1FRU7neYmNmA4hpKjUqz1pwSBouTQ4KfJgNlYWATcn/y/46In5RrziSXTH8Q2Nkjt8xsoHKnfA0krVSWQJldfhbwdXLNrV2AqaX2cRqwP7Bko/YREftJWppcx+u5Tj0HM7O+cpNXH0n6MPCopE9UDo8gO9Xvi4i/lDD5NBkmB5dZ1qMk/TtARHwW+GS1c97MbKBxDaUPSpj8khyVdVvlrpFk09bIct6ngf8FDouIU8rqwl8BpkuaFhGvNMLEQ4PNbKDyp+FekrQOOULrNOCoMjmx4VXgcWAtSV+mhAlwcrn/3cBmwPDqHhsOEzMbyNwp3wuSPkjWSP4CbBwRTzYP85X0PuAmcr7JyRFxSLl2HLkX/CvAJp75bmaDhQOlhyStTzZzzSA3u7oIODEiXmiaRzKMHL11NvA3MlyWIPcgnwOsExGvV9fmMjMbyNzk1QMlTG4ETiW36r0P2BM4TNLiTTPe5wBXkU1bTwAbkRtpXQ+sXcJkhMPEzAYL11BaVJYunwF8CzgiIp4vW75eSdY6vsvcmsq/1DpKyIyMiFcrx1wzMbNBxYHSA5JWAf4eES81AqHsbXIFbw6VxnL1nu1uZkOCA6UPWg2VzpbSzKw93IfSByVMhkfE68D2wP+RfSqHVvpUNP9HMTMbHFxDqUFTTeVyYB3gJ8BBTfNTzMwGLddQatBUU9mB3N53DDnXxMxsSHANpUaVmspwcgV6d8qb2ZDhQKlZ8+RGd8qb2VDhQDEzs1q4D8XMzGrhQDEzs1o4UMzMrBYOFDMzq4UDxczMauFAMTOzWjhQzMysFg4UMzOrhQPFzMxq8f8BZH6GHeCqFCIAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.bar(np.arange(len(X_names[1:])),IG)\n",
    "plt.xticks(ticks=np.arange(len(X_names[1:])),labels=X_names[1:],rotation=45);\n",
    "plt.ylabel('IG');"
   ]
  }
 ],
 "metadata": {
  "celltoolbar": "Slideshow",
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}