PYTHON-INTERMEDIATE-4-GLOBAL-LOCAL-CLOSURE.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true,
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## 변수 범위\n",
    "*Python이라고 다르지 않습니다. 해 아래 새로운 것은 없습니다.*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "변수의 유효범위는 함수입니다. 함수가 끝나면 변수도 사라져요. (*다른 언어와 마찬가지로 stack기반이기 때문에 그렇습니다.*)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'v1' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-1-a46200acc9cf>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      3\u001b[0m     \u001b[0mv2\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m20\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mf1\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 5\u001b[0;31m \u001b[0mv1\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'v1' is not defined"
     ]
    }
   ],
   "source": [
    "def f1():\n",
    "    v1 = 10\n",
    "    v2 = 20\n",
    "f1()\n",
    "v1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "naming을 탐색할때 **지역변수 -> 전역변수** 순서로 찾아갑니다. (정확히는 LEGB rule. Local Enclosing Global Builtin순서로 찾습니다.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "slideshow": {
     "slide_type": "-"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10\n",
      "20\n",
      "10\n"
     ]
    }
   ],
   "source": [
    "# shadowing\n",
    "a = 10 # global\n",
    "def f1():\n",
    "    a = 20 # local이 global을 shadowing\n",
    "    print(a)\n",
    "print(a)\n",
    "f1()\n",
    "print(a)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(<module 'builtins' (built-in)>, OSError)"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 참고로 builtin은 여기에서 찾습니다.\n",
    "# builtin은 shadowing하거나 overwrite하지 않는 것을 권장합니다.\n",
    "__builtins__, OSError"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [],
   "source": [
    "# 왜 에러가 나지 않을까요?\n",
    "def f1(a):\n",
    "    print(a)\n",
    "    print(b)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "11\n"
     ]
    },
    {
     "ename": "NameError",
     "evalue": "name 'b' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-4-9824713f33d0>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;31m# 실행시점에 에러가 납니다.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mf1\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m11\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m<ipython-input-3-b4fcb6e89a2b>\u001b[0m in \u001b[0;36mf1\u001b[0;34m(a)\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mf1\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 4\u001b[0;31m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mb\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'b' is not defined"
     ]
    }
   ],
   "source": [
    "# 실행시점에 에러가 납니다.\n",
    "f1(11)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "11\n",
      "10\n"
     ]
    }
   ],
   "source": [
    "# !!실행시점에 b를 찾겠네요.\n",
    "# 전역변수 사용이 안좋은 까닭. 컴파일언어는 없으면 에러라도 내주겠지만,\n",
    "b = 10\n",
    "f1(11)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "Python은 코드 실행시점에 변수를 찾습니다. 즉 변수에 대한 잘못된 참조가 Runtime에 발견되기 때문에 주의가 필요합니다. (도구의 도움을 필수적으로 받는 것이 좋습니다.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10\n"
     ]
    },
    {
     "ename": "UnboundLocalError",
     "evalue": "local variable 'b' referenced before assignment",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mUnboundLocalError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-9-6bec29a49f88>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      7\u001b[0m     \u001b[0mb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m9\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      8\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 9\u001b[0;31m \u001b[0mf2\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m10\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m<ipython-input-9-6bec29a49f88>\u001b[0m in \u001b[0;36mf2\u001b[0;34m(a)\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mf2\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 6\u001b[0;31m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mb\u001b[0m\u001b[0;34m)\u001b[0m  \u001b[0;31m# 여기서 에러발생!\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      7\u001b[0m     \u001b[0mb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m9\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      8\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mUnboundLocalError\u001b[0m: local variable 'b' referenced before assignment"
     ]
    }
   ],
   "source": [
    "# 그렇다면 이것은 왜 에러가 날까요?\n",
    "# 변수검색하면 b가 안에도 바깥에도 있는데?\n",
    "b = 6\n",
    "def f2(a):\n",
    "    print(a)\n",
    "    print(b)  # 여기서 에러발생!\n",
    "    b = 9\n",
    "\n",
    "f2(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "- python은 line단위로 실행하는 것이 아니었나요?\n",
    "- **실제로는 함수의 body를 통채로 compile합니다.** compile후에 한 바이트코드씩 수행.\n",
    "- 컴파일 시점에서 f2함수의 b 변수는 **할당연산이 존재하기 때문에 local변수로 결정**되었으며 실제 f2 수행 시점에서 local에서 b를 찾으려는 시도를 합니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "11\n",
      "6\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(None, 9)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "b = 6\n",
    "def f2(a):\n",
    "    global b  # 이게 필요합니다.\n",
    "    print(a)\n",
    "    print(b)\n",
    "    b = 9\n",
    "f2(11), b"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "전역변수를 참조하는 것은 상관없지만 <span class=\"mark\">값을 할당하려고 한다면</span> global을 챙겨주세요."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "ename": "UnboundLocalError",
     "evalue": "local variable 'x' referenced before assignment",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mUnboundLocalError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-1-689637822814>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      3\u001b[0m     \u001b[0mx\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;36m1\u001b[0m \u001b[0;31m# x = x + 1\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 5\u001b[0;31m \u001b[0mfoo\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m<ipython-input-1-689637822814>\u001b[0m in \u001b[0;36mfoo\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0mx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m10\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mfoo\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 3\u001b[0;31m     \u001b[0mx\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;36m1\u001b[0m \u001b[0;31m# x = x + 1\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      4\u001b[0m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0mfoo\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mUnboundLocalError\u001b[0m: local variable 'x' referenced before assignment"
     ]
    }
   ],
   "source": [
    "x = 10\n",
    "def foo():\n",
    "    x += 1 # x = x + 1\n",
    "    print(x)\n",
    "foo()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "ename": "UnboundLocalError",
     "evalue": "local variable 'lst' referenced before assignment",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mUnboundLocalError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-2-a7232171f6eb>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mfoo2\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m     \u001b[0mlst\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;36m5\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;31m# lst = lst + [5]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 4\u001b[0;31m \u001b[0mfoo2\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m<ipython-input-2-a7232171f6eb>\u001b[0m in \u001b[0;36mfoo2\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0mlst\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m3\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mfoo2\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 3\u001b[0;31m     \u001b[0mlst\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;36m5\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;31m# lst = lst + [5]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      4\u001b[0m \u001b[0mfoo2\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mUnboundLocalError\u001b[0m: local variable 'lst' referenced before assignment"
     ]
    }
   ],
   "source": [
    "lst = [1,2,3]\n",
    "def foo2():\n",
    "    lst += [5] # lst = lst + [5]\n",
    "foo2()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1, 2, 3, 4, 5]\n",
      "[1, 2, 3, 4, 5]\n"
     ]
    }
   ],
   "source": [
    "# 잠깐 복습, 다음의 차이는?\n",
    "lst = [1,2,3]\n",
    "lst.extend([4,5])\n",
    "print(lst)\n",
    "\n",
    "lst2 = [1,2,3]\n",
    "lst2 = lst2 + [4,5]\n",
    "print(lst2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "![](img/lst-visualize.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1, 2, 3, 5]"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lst = [1,2,3]\n",
    "def foo2():\n",
    "    #lst += [5]\n",
    "    lst.extend([5])\n",
    "foo2()\n",
    "lst"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "(TIP!) 내가 원하는 실행context에서의 local, global 변수를 찾아볼 수 있습니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1\n",
      "6\n",
      "{'a': 1}\n",
      "9\n",
      "9\n"
     ]
    }
   ],
   "source": [
    "b = 6\n",
    "def f2(a):\n",
    "    global b\n",
    "    print(a)\n",
    "    print(b)\n",
    "    b = 9\n",
    "    print(locals())\n",
    "    print(globals()['b'])\n",
    "f2(1)\n",
    "print(globals()['b'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "당연한 것이지만 종종 헷갈릴 수 있는 부분. 코드의 실행 순서를 감안하여 위치를 조심하세요. 인터프리터는 소스를 한번에 읽고 실행하기 때문에 아직 지나가지 않은 곳의 정보는 미리 알 수 있는 방법이  없습니다. (컴파일언어는 보통 two-pass로 소스를 탐색합니다.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10\n"
     ]
    },
    {
     "ename": "NameError",
     "evalue": "name 'z' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-3-d272fd034e06>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      6\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0mz\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;31m# 왜 에러가 발생할까요?\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m \u001b[0my\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;31m# 이것이 문제입니다.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      9\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mz\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     10\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0mb\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m<ipython-input-3-d272fd034e06>\u001b[0m in \u001b[0;36my\u001b[0;34m()\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      6\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 7\u001b[0;31m     \u001b[0;32mreturn\u001b[0m \u001b[0mz\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;31m# 왜 에러가 발생할까요?\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      8\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;31m# 이것이 문제입니다.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      9\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mz\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mNameError\u001b[0m: name 'z' is not defined"
     ]
    }
   ],
   "source": [
    "def x():\n",
    "    return b\n",
    "    \n",
    "b = 10\n",
    "print(x())\n",
    "def y():\n",
    "    return z() # 왜 에러가 발생할까요?\n",
    "y() # 이것이 문제입니다.\n",
    "def z():\n",
    "    return b"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## CLOSURE"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "함수내 함수를 정의하는 것을 단순히 closure라고 하지는 않습니다. 그렇지만 함수내 함수를 쓰면서만 closure가 발생하는 것은 맞습니다."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "**확장된 변수 범위**를 갖는 **함수**를 closure라고 합니다.\n",
    "- 함수 내에서 참조할 수 있지만 함수 내 선언되어 있지는 않습니다.\n",
    "- 함수몸체 밖에 있는 non-global변수를 참조할 수 있습니다.\n",
    "- 이 변수들을 free-variable이라고 합니다. (*global이 아니면서 local scope에 bound되지 않는 변수*)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "10.0"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 아직 class는 모르실 수도 있지만;\n",
    "class Averager():\n",
    "    def __init__(self):\n",
    "        self.series = []\n",
    "    def calculate(self, new_value):\n",
    "        self.series.append(new_value)\n",
    "        total = sum(self.series)\n",
    "        return total / len(self.series)\n",
    "\n",
    "avg = Averager()\n",
    "avg.calculate(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "10.0"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 함수로 간단히(직관적으로) 표현하고자 할 때\n",
    "def make_averager():\n",
    "    series = []  # free-variable이라고 합니다.\n",
    "    def averager(new_value):\n",
    "        series.append(new_value)\n",
    "        total = sum(series)\n",
    "        return total/len(series)\n",
    "    return averager\n",
    "\n",
    "avg = make_averager()\n",
    "avg(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "- encapsulated된 변수범위를 포함한 함수를 표현할 때 사용합니다.\n",
    "- closure 함수는 물론 class로도 표현할 수 있어요.  private 영역을 포함한 객체인스턴스를 생성하는 것과 유사하다고 볼 수 있어요.\n",
    "- 함수로 표현하는게 간단할때가 있고, class에서의 self 표기법을 탈출할 수 있습니다."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "함수의 free-variable을 검사할 수 있어요.\n",
    "\n",
    "> In computer programming, the term free variable refers to variables used in a function that are neither local variables nor parameters of that function. The term non-local variable is often a synonym in this context.\n",
    "\n",
    "https://en.wikipedia.org/wiki/Free_variables_and_bound_variables\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('new_value', 'total')\n",
      "('series',)\n",
      "[10]\n"
     ]
    }
   ],
   "source": [
    "print(avg.__code__.co_varnames)\n",
    "print(avg.__code__.co_freevars)\n",
    "print(avg.__closure__[0].cell_contents)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "300"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# inner는 closure일까요?\n",
    "\n",
    "a = 1\n",
    "def outer(b):\n",
    "    c = 2\n",
    "    def inner(d):\n",
    "        e = 10\n",
    "        return d*e\n",
    "    return inner\n",
    "\n",
    "oh = outer(10)\n",
    "oh(30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('d', 'e')\n",
      "()\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "print(oh.__code__.co_varnames)\n",
    "print(oh.__code__.co_freevars)\n",
    "print(oh.__closure__)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(600, 1)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 나는 어느 c를 보아야만 하는가\n",
    "a = 1\n",
    "c = 1\n",
    "def outer(b):\n",
    "    c = 2\n",
    "    def inner(d):\n",
    "        nonlocal c  # or nonlocal c로 테스트해보세요\n",
    "        # c = 30\n",
    "        e = c*10\n",
    "        return d*e\n",
    "    return inner\n",
    "\n",
    "oh = outer(10)\n",
    "oh(30), c"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8\n",
      "8\n",
      "8\n",
      "8\n",
      "8\n"
     ]
    }
   ],
   "source": [
    "# closure가 사용될때 나오는 흔한 예\n",
    "def create_multipliers():\n",
    "    functions = []\n",
    "    for i in range(5):  # local, free-variable\n",
    "        def _(n):\n",
    "            return i*n\n",
    "        functions.append(_)\n",
    "    return functions\n",
    "#     return [lambda x: i * x for i in range(5)]\n",
    "for multiplier in create_multipliers(): ## 0&n, 1*n, 2*n, 3*n, 4*n\n",
    "    print(multiplier(2))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "http://www.pythontutor.com/visualize.html 의 도움을 받아서 살펴보면,\n",
    "<span class=\"mark\">i는 각 함수의 로컬변수가 아닌 외부변수(자유변수)이기 때문에 실행시점에서 해당 값을 참조하게 됩니다.</span>\n",
    "\n",
    "![](img/closure-visualize.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "2\n",
      "4\n",
      "6\n",
      "8\n"
     ]
    }
   ],
   "source": [
    "# closure가 사용될때 나오는 흔한 예(수정됨)\n",
    "def create_multipliers():\n",
    "    functions = []\n",
    "    for i in range(5):  # free-variable\n",
    "        def _(n, number=i):\n",
    "            return number*n\n",
    "        functions.append(_)\n",
    "    return functions\n",
    "#     return [lambda x, y=i: y * x for i in range(5)]\n",
    "for multiplier in create_multipliers():\n",
    "    print(multiplier(2))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "![](img/closure-visualize2.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 ()\n",
      "2 ()\n",
      "4 ()\n",
      "6 ()\n",
      "8 ()\n"
     ]
    }
   ],
   "source": [
    "# 사실은 좀 풀어쓰면 더 명확해 보입니다.\n",
    "# closure가 사용될때 나오는 흔한 예\n",
    "def create_multipliers():\n",
    "    ret = []\n",
    "    for i in range(5):\n",
    "        def _(x,y=i):\n",
    "            return y*x\n",
    "        ret.append(_)\n",
    "    return ret\n",
    "    # return [lambda x,y=i: y * x for i in range(5)]\n",
    "for multiplier in create_multipliers():\n",
    "    print(multiplier(2), multiplier.__code__.co_freevars) # 자유변수가 사라졌습니다."
   ]
  },
  {
   "attachments": {
    "hard.jpg": {
     "image/jpeg": "/9j/4AAQSkZJRgABAQEAeAB4AAD/2wBDAAgGBgcGBQgHBwcJCQgKDBQNDAsLDBkSEw8UHRofHh0aHBwgJC4nICIsIxwcKDcpLDAxNDQ0Hyc5PTgyPC4zNDL/2wBDAQkJCQwLDBgNDRgyIRwhMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjIyMjL/wAARCAFNAhQDASIAAhEBAxEB/8QAHAAAAQUBAQEAAAAAAAAAAAAAAwECBAUGAAcI/8QASRAAAgEDAwIFAgQEBAQDBwEJAQIDAAQRBRIhMUEGEyJRYTJxFCNCgQdSkaEVM7HBJGJy0RZD4SUmNFOC8PFEFzU2N3OSorLC/8QAGQEBAQEBAQEAAAAAAAAAAAAAAAECAwQF/8QAIhEBAQACAgMBAQEBAQEAAAAAAAECEQMhEjFBBFEiMmET/9oADAMBAAIRAxEAPwDJ1w60tcBX1HwDhXd6Qd6cvSiFA5pwHNIOtPFEKBRI03OifzNj/emCpFqdt5CW52kuf2FTK6xaxm6obtf+Puhx/mkZoOOTTY2MjXEjfqnY0TFeWXb6FmpHDjnvUq3wCpJA2nnJ4wfeo6jDAk8GrzRbAXMzyThPKhUu4DYOOxx357U34xJh5XQV7o2r6rbRWFlbbbZF3TTzHEZ74FWWmeH1hsDBczhEkQq6p1LCpN1M/kW1xG0l3byuARjaiZOCyD//ABNS5AVkkUbpnmIIkB5iU8EJ7HFcLnuvbhxYydpmnaHpr28fmO5RDlW/mI7VNt/8EaYQxWJbOBG0nQAH/vUK0aaW5uIokEo08Ksdui7AxPct3NTodOvbRHMhj/EHIL+ZvfaOSPv7Vztyv12nhGgP+DpDtutOt8rx75PzTz4b0OaMTJatbOBkNGfTx7VRWwQbJRDI0kwIjX9SqOu//b96tLNpIxZL5hMUwfaWPJpLlPpccEXVtBuHWWa0nE8OwHYD6hWfYOgCyq4bdt9XYVsZ4Hs7hJoCMkYIPYUC5tYtRi3FB5rHpjv711w5bPblycONm4zUZAiJHUcftUhCWTr9VBubeSyDRTIUcH+tFjGP/pP+1eiXbx+NnQgO5hgZ5xUi2dmnlYt1bAFRIs7Ec8+o1Ks8mFnAHJJFZdMUpWKiLjkZP2rlVmgtAf8A5ZINDnYx2uQTuKkYqRsZI41PHlwjb+9Ytb0hYzJn4xTtxJxx8Uw8ORSJjzOaqVKhAEcvPNSEOO36RUSDBhlJ65qUp4Jz2FAEuC756U0sCT196Ev6v+qiKfntUQe19Tbvin53I33pIBtjJpM4Q/Jo1DepHxUa5GbogewP9qkgZbHQUOVR+IJ/5aRmo8akKTnqcVLdvLsyo43imRxgWBYhSxmAH2206bBkSNgx2DnHSpfbU9I8q4YA9l/vQ5JCqjYOSOabdTEkY6FsCkfG3aPq65qsGDKxjd9WfTTtxZgp4Y/UKE7gITnO0cD5ptrMpkXzAS5G7A7UNptxIqyR5OFxjFQ3mIV8ng+mmSyebJuY7hmgsXKtkDZng0kS0xWyqq2cAnj/AEp7gmCAHgEEn75pgIER3ZZifqpzHzZYkd/T8VUdIdpLfSW9qa+5QhJHHvSdbl8jOBwTSXDYUMSScUQke1bVnc5BJwB3qRcsBbwqciTaCfkVHUbYYxuILc4p9y5fYM8DHXrxQda+iFTInBBOfanIN2xT+o5Px7UOWUSKE5CnnA6URCFZm9161QrOeWHbgUIn1cYximlm9CoeO9MYBsKDg/HegIXAAwATQ3lJGNxO3qKZI6EhY1xjhjXEgf5dB2XkGTwp5prOMYbPwKdI4CAqeO/3oTEsu4nntQF9SR+rDE9KE7Zxk4+KY2/apY0hB8r3GaApXdHngYpASHyemOlC7DcML2py7lG5zkdqIezALzzntQi7SME6Y7015MAkde1cHVVCn6jyaoeWCkjPI70wNgEHqOc0i4dCfY8VL0vTptVuhBFt2xnfJI30qPY0DIrWe4jEirwenFdVjda+1vcNDZ28bRJ6coOM11ZGEpRTsUgFe180o60oFIKcOTQKBTwKQcDNPA6UTL/muxg1M08Kb87hwtvIT98DFRsDqTQJb1rW8RF5aQY/Yis8l1i68eO84pYtrRl16NKcUReajQSgM644WQ4qSp4rzSvbnD1wiMxVnVQSVXqaFZX0l/bSCML+JQbl2nCgDqpPuadLvMXkxD82YelW44HU1H06ESPBZ2UoUSM091cSDGyNTkj756Vz5K78OPW2rivrye5g0chgZEWdFUf5HPKf74/er14I571nsZZUQOvnMnPl9iB7Y65rOreXNxpst1bOY3nzHDKUwzKTjf8Afb1HtVujvpdiNIiJW5uI1DhSS5TIIZvbJ4rj09OmphtWhzPHFCJ0kLbSeM46se+OtGsoY7dFuJJ2uWIGJP5ee69/+qs3pF5Heym3nmb8MiiOXs0j55xnrnpWihu4P8UsUVVZLkbYOcEkdj/y1OlkGnF9bygwxpHGT+cS26Taeh+M88dv3q0ihX/ElvWUCOFPLjXPpBI5bFHtooTGG3lgrsWk7uc8g/HTFPFr5xWZ2Plv6Qvuoq+MNmIxdOV3SnqSO1ddK2YjEFQkgZAo8hy6rGCSTjIHQe5oMow4RGyA2Tk/0qWG0XVbeHVl/CYIm6xvj/X3rNSpLBvR1xLnDr7D3rWH8lzvJDfp4qt1iwW8slnjBM0IywXq3vW+PPV8WOXi8p5RRplYVGODyPvU6AFYoF7qeahDomeBgkVPhXYmXPQf3rtk8+HY7hZJSCOPvUi59Jhz+mNT++KixrlWY5J7Cj3XEzj9IAA/ao0rwPXz8muCgy5/5c0q9SaXoTjn04ptB41zCW9+TR2ztbGO1Ci/+HUU9zhWx3GaLUOPoR7tzTlxlsdhxQoCSAxOcknFFRcMB7jFGfqYp2wD3Irgp2N7A8U5l/LA/vXdIST71mth7sMSR0FDnUxyENglkVh+45ou3O4jsKBfsVuweeYFH9qsS+hbcINPVicjz/8A/mhIxeWaXJ5/pSAgaPARwWlLf2xUd5SwESHBI9R6UT1Eed1aURqMKpzn3NNd2bO1hx701iMBR+k5GaZKC/qBw3fHetMUOSYshWMDI+KJZqQJXP2FCBITfjlScjualRDZBGM8E5JFEhrKqpjGMnJqOV/LU7id5I68CpEvJGD9RwKiuR+HDEelZMAUK4cQjnOPYcGnBkEj5G0dPtxQzuYrFxgdacxBVycYduRQBhzIHfJB6DPcU2bd6V3dTgGjI5V8NjKrxUcnzJwD0FEEkba4Zj9C7RimKxCkn1Me5pzYMzLjgf3pqgsEXoCcmiiKu5owzY2jmnMRlgM8HApkeHZuRlmxRHwqMexPFEMJ2MMfUKEPzJDgbFUcknqacwEq+YTTMgLt5yetAPcGPyP713mhskYCimFmDcY4pi5wynjPPFA/fvxkYH+tNLBXxmuQh0G44I70jJhy3uMA0DixOGIzjpTHIfC8qB7d6KFyAXYAUB3DHAHHagIPUCS32BrgN7LjI/m9qEqlyCD065opyEbBHNVDJAoPHIpoAYMzHBA/rTGIUbiwHvU7TtGvNRkeXZ5dqoz5j8Zps0DaW0l5OLYEIn1NJ19P/erC/vLa3h/B6UZFjxiR88sfk0GfUorWNrewiAHRpj3+1VbbvLwXGCcnFIU5WCDGFH2NdQ8442Ka6iKmlApKcOlex851KBSY5pwFA4dKIOcUMen5NEXmgXA781Xz+vW044VP6cVZLgtzVanr1S6f9KQuf9K4819PR+f/AKtqlt+Xf2LmpgILDB68D5qNBxI2emM1LjGQoGMEnPvXGdPXkBfOCv4WNJfMxl5F6oewrTWng+7OlR5RkS8Km5fPPljnC/c1Xfw9SXUtZmeMh2J8zDjIOOma9gmiT09mVDkg8K5HHFefkye/gw6Ya68mw1KKyKBLZEWRQvONv6fuRUJhd3esXE0XmNqF2u8s3AiiX9P9KtprBo5L+e59ch/LDDpux1oukQLbXzzSSMSsG0v757VzldMsP4BBappf4KLyd14pYDPTcedxrT6TZQyaU11IhEg/yJO8YB6D4zTViW71OYShd6wrMhYYPHUVZybTAI2VPJHqwGwQPapcmseN0V8tvJtdcZKLGo7ls7s/2qdHepHP5CuD5ZwQe2faqlbKK8IgaX8uJg7yA+snniqfX/4h2thKdO8PRQzMv13UgBVD7Anqaz5tTgyy9NzbxIysN/qyVznjB+aiujzuoRNkSKeT1OO9UPgrxhP4hEum6gFS9CFxMo9BAI7VprwM8xt0jKF8s7A5wP8A1reN3HLk47hlqs/cSFtTggWcuWbKfOKnxuIp3MWWVG2k9mz1qu1Dy9NkXYV3ImA5/wDLHz96j6Nqo1BtyQPEH9WG6cdTj5qW6dcMQdYsBaXe6LLW04BQk8Ic1Ij5i6Hk5OastRhW+sXhJwXIkjx+nHaoK5JwD25HzXbiz3i8nJx+GXQiDA3A9OKddHM84PRc01sCAAfrIxTrsgy3bjucVthBUArgHrTkA3Hr0pIyPScdqcp5cjvVrKRF/kLTpgQvH8ppUH5aD3pbgL5TDJyENIWK6AgQINvq5oyZLgn2zQIW/KRcc0eBSX+rPFWs/U5wQqY71zkhBgZ5pzj6c/y8UsgPlLgc1jbdCJ/JkkPA+KhaozC5TH/ylH9qkXT7YBCpzkjcfaomrzwfivIJxJEFbI78VWKdK4XRrQnuzH++Kr2f1Z6gDGPapUhzpVgf0nfwf+qoJY+YQRwelbjNpjAsuCeQevvT1zlQT0OaaxBcdhSZI5/lNVk1i3m5HG40fhU2ZwCP6mutzuly5GB2pAuXXcpOckVna6JcEl8dPKXNCRgYQjDgjfjvT5gXWQg8udtNDbY9+BkDap+KFDbiM7cHPf2pnmABOBwcUpjO5Yx1yCx7Uu0SXTIMEdwO3zRCKuZHLDnGAR2oQ9DMTzinyPtmIzhepNBRi8pJ4Un+goHFyLhmx9qUkqhP854HtTMHHPWnjngdR0qhbcBJBu6jk0+8OIxGDznNCXkBsZ5xQ5nzdsxORgAD2olNPofkkD2p3BTdv69BQ2PrIGGPvQmbB55Pb4oOJzuGOTxmuOQQoPA7+9NYheSAQadjcwjAxnv7U0Hgk5C/T3zQw+WxtJAPWiLtUMDnjimso2gKCFPc0HPjGSKEcswC4HvSyuJADuwBxmpFhpmoag+y1tGIPBJHH9aAHmbVx3/1p8EU95OEs4jK+MEDov3rSW3g9YgsmrXYSNRt8mPtjtmu1HxNY6VB+D0O1SSRTgsvT9z3ptZjolv4dttLtWudTmSS5A3CPOQKo9T1m5v5wVJjiUYAQ4B7dKiXc817ctNdSFpCOdp9P9KAWTIG3GO4olpnmADav0cZ+9cWO4g9qQ5xtxyT2oixgLuZs5qsnBMiupN7LwM11BVU4dKTvTwPSa9j5xop2K4D4p2aDgMdaIopop460Cjgg+1VlqSf8TmPI8rA/c/+lWbYWMknsarbTK6PfSfz4T7YJP8AvXDk7sj08U/xb/7FSgO5ux2ipYwCSO42j7mo6f5rZ+KstMiWbUbdCf1/T71x+V6/+smv/hrorafaT3chKyzqAuB0xW8a3MruwA2lxwP9aFptvHDaLGAyFRgDFWQUB0UcKV/vXktfTwmopJrDdPMEAwzEkH5GKrTp34bUWtJCGtXUHb3BrVPbiRnOei7aHPZpNBtkUAcesfVXK12xUsdtIdY80MxTaBtbqGHQiqTVPF2k6XqItkkmmnhbEkUMO5Vc/ozmtiY41w54OPrH1ZHSqSLR7E+IDqbw4ZPznOQA7+5461z8noxxlQ/F15NZ6XBo+ngrqd9mSdwuDHGRlj9+grzC7nS1gQW/lyRKxDInJz9vetF4h1mddR1XUbrf5rwtHC4+sRZHb/esvpWm7rZZ7a2yjH1SliSD7n5rcnWy2z0u/DevHQ7savcrJDG35LsFyYlY9WH8te022p2t5Zn8LdxyoVAM69ce6ivM/CdisV5JJqkcl3Z3kbW80TJ9QJHbt2p1tbX38OvEy6LMwfTLqUzadcs3+Yn/AMnPQEZ/et8fc2836P8AqeTV67i5sgyqQbx8LE31SqvuOwzzUax8y1bMpxIxwo7bfapn4lLd5tSu2jZh+TAAeVOORjHTNVruGjXJI3DAx/3rOVduOdLd70MwX2bmgxzbrm8GOfNGPgVST3LW+0Zy5Yd/00XRb0XN5endkptJH71r89+OP7MJJKvjh5Ix2DV05B8//qNDhYyXsKqMZcGkkI/MXuea9c/j56NGfR9qIvCtUdWx6ffnNF3ZXFWs/VjEG2w89664xh/+k0+L/Liz7UyY5ckfyGs7Wq6E4UcUe1G6bPYdajqcICD1qTZA5Zf+atWs6TWGX3HtT76VLaBW/UegpzDy2LNwo6/NU99dFrjzJfpA9IrMm1oF5cmG12Y/Nkxk/HtQ9SUR6jKXBXhPSOvShoTIA0oydw/YA1M1hc6lcnaXLMoTHU8dMVrTEuzLlz+BsDsznzAQvUAN/wDYqKTkxzYLpvBKjpgnGP6Vaap5thY21kVVpyHeRx2ychR8DrVZaDzp4YGBw0g3KvGeeTVSnarD+G1WW3VQm7DqhHG34qGMMh2kkH3q21Cd7jSBNJgvBeNArEeoLtPFVbeg7duAOlNprToiVilYDDbdoojErI4BwEUL+9JEqquTgj6iPfFMkJby1JyS2WI756f0xUD3OY+B0Gc/NCjw0Sq2etOlkVFK5PtQkeQI43YB9xVDhhW3E+pfWftSIzefOy7eg/vQzgHgZVuCc9u9NO3a2FOV4bntQ0DIfUwY8gHbinW3qkVSuSRyaG4CFGz9Xalgk2zLliB0oh+4enHU54p0SsSzkAFfmmIs1xcJBBE00znCRoMk/f2FaJdAtNNCvrV/HG4G7yIzlvsabNKKJgZGPO0f61GEU0jMVt5X3HsDWnTWfD9nEwt7SSRH4wRjdQx43REEdvpZUA4wWHT74ptfH+qAaZqDDJsZyO3FPOjaowBNhKB24q+PjqYr6LDb/wDX/wClMPjm6J4tF47M9TdXUUY8P6vL/wDoJDjoDxUuPwzrUhGbQIMdGapL+MNUZHARUVunOdtRj4o1WRQTKMDjnPNN06S7fwdqBB8+4iiJ6LnJNTYvB9oBi4vGlcdYxxis++t6jJbiLzE5JO8cMKj3F9dShVe7kbAxkHBqdm42Ah8PaU43RQq6D9bbif2qtuvG+R5dnCQuMcgAH+mKyUpMj73QEj9THJoZLKFHXjH2rWkuWvg97qF5qEjfiJXwOkSMQKYvpiJHHHShqNgyCN3vTHfjAGc981WLlaJ5hJ4AyaFnBKqMk05jhgBwcUJ2O7A4Y96rI287QvelyF4Y0FSSME9O9FVdx9XOKK4ysDhVJFdTtwHAAxXUFcKdzSYFOr2PnOGcUoFdTwBRdOUU8CkHGOKfjrzSezuhXTbbeTioSqyeHpW7mbke4xUq/bFsQMZYhR8fNStN0a+1jTILawVQxId5JPoUAkE1xvee3pxlnHNfWXUDzX5HYff5/YV6B/D/AEB7l21W5twLZBshDfVIf5hQLb+F2pNqEKz39jJYbszOrHcB1wK9NtYYoUiijj8uJBmEYxgV5OXK47fS4MJbsaOJowDnP3609lDEswzgdqegGzdnB7ikZtyn7V5Pfb367A3Kv08D5oUk2dxDemmyy5TsCOM1CeUeZtJ9W01yyrrjgeZQwzjPNA1KATQ23q2B5PUfinxKWcIcbW/V7VA8aXj2Xh6K6hAZUlEbAdge9Sdzbrj1lpmtf1qze7ksrCzSaWVGiuZXAwsfsKDpVtaizt9PsMC2tk3TEc5kJ4GaiTeHJbzToNUhk/IZhGwU4LZ6n5rX6VpNvo1qyWcOQRkiU9fnPvV7rpZIsdIslSV52DZiXZsAzuai+KdMttY8N3Wl3zDylVZVmx/8Ow5yrHuKm2QjtbFJV3SRsd4P6mf+X7Vh/F2p3D3w09bxpbaLMt3CoyqZ6Lx1xXSW44vJZebNX2urXWq3MWk6gwEyR70kQYF2qj0yD52gZA75qfNcboT5eQ8ahinQEHoTVTb2wvdPaBLhYL2ycT2c7ciLvuHwehFWlrqCa3Z3Hmxi21K3YfjYHGdrY+tfdW6jsM1z3vt65JhfFQ3OoyNOqt6i5CBvbvUzwJcG5tdWvT+u78tfkDrVJr91Hb2kt2pGEyRj3PFXH8PI/I8H2qg5keR5WP3/APSvR+XH6+f+/PvTdWRJv4Cp43ihFhvcE87c12nuPxSFWwVORUcTY9RYZdfavTrt4PjkYGRiRwKdEwJP2oakKX5/TSRSKAMcZFVme15D/lxfaorvnnd7ipMRxDGfioLHKbe5JxWW6joTt2kY5qw02Ml0ye+cmq0D1gZ57mrCF9irEndSzH47VUjru4aXzAG9IYjPyKqDm4di5yAMAVJncbY4I2ztOZD7k0iH8uRAo4OCas6ZtCb0hHOAS2Cv71camy6Xd3N+cNeTjbag8iEfzn/aqhCEIb6sNzn71J1hzJrFz1JPY9h7UqQC+LJZ6Y8jl3ZJBuJ5Yl8k/wC9D071azDjjqR+wouqKWsdJDLnEUh4/wCuolg8cF/azyNiIMUY+xbgc0RJLZ8P3Ks3IvgQfk8VXnJ4Jzg1YahHHa6RBaGdZpjO1xIYvUpHQDPaq1Qdy7gPUeqnNUo4+nAUdDSPgSrlRgDPFOJAXA5OSM0CYlVc55UYNRDGcurPgDnpTSxKrHnnG45rp8qFQcbl/sKYCNzHPp27c1Q9QNpK4IxxTHcrFyRkjmmsSqIAQB0HzQZjsbryvX5oGs24tg9Dha55gHkk2btu1R7ZPGaTICqx+stux8VZeHoFl1BruVM29jE90y4zuI+kY+9ESZLpvDsLWFowGqSLuvLk9Yh3jX5qpMbPK0kheRn5Lyckg9vvTBI89zJcSsWkkYuzHuT70VVbHGeTiim3eBFGijcq88dqjFgRjB5okz7ppdp9K4UD5pkaksC3GKaZ7cVJ2iQentg0scSjLP0HYnNIxBkBboPalbazA88dKBGIbqDspjsoG1AfjFPlyqEsRQWZt3p9qBykA7jwftSOTt3Hhj701CSvqaklyzgA9untRCs7FVAph9O5c5x+qijYo5Ofn2qM56kd+lAknq5BoXG70inSEsMYxXICig8c1Udgk571zFFGMZNN3EOSn7k0u0N6mPNVCqR+3tRU6bR370gX0jGM0/dtHSgcEUDBIrqb5i+4rqKgGlxzXYzTxya9j52iY6U9eld+nOQq9C56A9h/60W2tri+mWC0t2nmJwVXkD5z0pbqdrju3Ro7Yxn5anL7Hdk8cCtzY/w7VkV9Q1Fg2MmOFRtA/fnNWTfw/wBEkQ+SbqLIwGD5ya4//XGV6Z+bOzbzSLTJ9au1sYGKJkNNMR/lr8fJ5r0nS9PhsLJLa3TESDA+fv71LtfDUWlWwhtBuAO4Mwyd3cnFOV0hkMc0TWxB9DfUvyeOlPLG+nXHCzqpaDA9vtxii7853cg9yaFt+nA4P0Ecq379qdjnGePas5YzKduuOdl6P/SQnqx29hQt+GDg4HTB70/fsIIOKa4Q5fr7ivDzcVx7j6HBzS9VClQrIyMcM3IFRpFYnIIJ6EVIlePKM5J6nP2qLNKpYkjk88dxXir6EtcWEJBdvR7DrVP4nv3j0aaB1VjIvCH3qyEYacZO2HIAPeq+/wBMe7T8NcSReaWLNJ2A9quMdNyMt4MtL681QSXKyJYQruVSTtY+2K9M/BPNgqqtGeV9l/aq/wANWggsGSP1Ih5AI/r9qtorq4hVWiWKROVfacD7/tXfHF5MuW20Yr5SxyMqmaMEF0Pp9hx71hk8PyWerXVzDfLbSs5kWS4A2575/wC1bjfE9q03moyYKhkGF3DoefY4rMzXVtZruv3iuJmIDdSrN7kGs8n8i/mtjPXumrpUi6k0EC2akrcxW0u6Jlb9SHqD3K9Kh+IfDutoItZ0VfxPkqEjlhPqmiPOxx34PWo/im4uNPnE0CQLFcZVo4ECxlT7gd/mqDTPE95oaQq126fhlaOJY3JCKTwGz1rnxx35c9TVUWvXrS6W8S4McsqjcT0PdcfBr0zwun4bw9Yx7dpWPk+9YfxNptrM6avMkdtFdt/w8SHmR8DLBewz1Jr0LT4zHplvE59YjUH717uGaj4/6ct5LmywLzI6KjOT8CoW8ny1XsR19jUq0A/NJ6i2cgfcVFUF5GXspAOPiurlsdnzKAehyKjcqjY6A8U4NkLzyCaaW9GO+aUlaOFttshI/RULOCFJBbk8VIVmNnGw/l4zVXAxilZs+q4bBz+kfFYbp7grNx3GcUdpxBCyoN0rjH/SKTUImjjSZRjjB+1CnQQWcMh/836ye1GfgKr5YwDknkn3p4wIwQ2cdR70IDGVA2rnAC+3uaKqYQ/2rTJETPqIGM5xUrWIx/i0zMx5GT9sVDRtuOeM81K1go2pSOhyrRR9PtzQAv50a106KIEmGNg5+WOaqizsH4wA2AB0/pUmQFlOTyMdPioTHG4AnrmqzaLGUjiboEHJVTj9qS2JS42EcL0FDtyN7gjOUP8AapFvmSWR2/lB/ehBixG444B/1qLIdyKufUzf2otw5EDDPPTjvUYt+YMY9C81FNlfdcBQSdvWmHoFP0muUZQsOC579qmafYXepTeVZwmUr9TH0gf1oaRpJEUqzkgKv7VCbzJz6I3mJ6BFJP71uYPDmnaYjzapMryEZYO3pHxinjxNoNkRDb/Uv6oo/wDei+MYeSJ0UM8MyrjaW8s8GrrTg8PhLWb2JsSSGO1UY4wTkmtPa+INJ1NzELhAcjaJV4JqL4htILHwzepHthWS4V0TP1N7j4qbXxmmK24yIyCMDJPvUlSI0Rs+ke9CiXziFUD1HDY+KdcnZATxz6cVqOaEfVIVTpu3E+9EK+g470iRlgo+njNETywoznBGWqoHt2x4HWkwxVcnjuaUH1bj0bgUu0j8skZHWgDKA/XO3vTG5fpt44HvRZHCjaB0qOxZwZGPNRCEYXB65p6oykMV/emIA6hyc0Qt0xnFAhUlc54/1oRAaTjgAftRGZyuBg5+n5oOQqld3B9+3xQMLH70jMMY7mudggwCOaRACy+k9+aqEiXcO5PajsAuFzhu4q60zR4pvD02pSuRtJAUfHeqdDlOvbI+aLogTHelwGGAOaRWyDndmuBYmqlNMZz1FdTirZ+kV1ERRjn4pcqoZiQFHGfn2pyLJLKixoJC52qvua9C8NeC0szHeaoivdJ6ooMZSI+7D9TfHavRln4vLx8N5KqNA8FTXvl3Wpb7a3xkIh/Nkz/oPv1r0Gxs7bTbfyLC3jtIxjPljLN85qUI1Y7mc78cjHH/AN/FKEHYV5M+S2vpcXFOP3HAAtk9B9OB/rTs5OTmlK8V2K5d+3XTtx98f9Ncwjk+tASRg8ek/tS44rgOelXdXpAayFtuaBm8s/VFn0/t7UNyF6dD0qy5jLHGc1HuLZXjZouuM4rthyd6rlnh9iBkcnr8UNmKEuM+n9OOtN34/brTTLmUEHGBn4Nd9TLqvPLce4rtYd4ZDj/KfBUjoD7VWQSyFJZZ2YPGchAMnHtV5NHHcWz2siny2B+4PZh9utZ6WOWznFvcbi6D8uResy+5r5fPwXDLb7v4v048mPjfa0094JZ/wdxIGllQmA54z7ff4oN5FJqEhswyxlmHmRiVVfPtjrn4qsnkLQhPIiWV235aTaXx2H8v3pbe9S0kN7HbRvLtYSSl9/lueoB7/euWEdOTcq4uby1s7dUtZALqEYbKk5x2wOtDh1K7mm838AYYimJSxAV/+kdv3rG3OpTSzqwmicvlVdefT7nFSbd4/wAMJZZY2EZHmy3JYCM84wO9dNVx1J3Wg1e9mSy/FO+yMjy4IlYIN2eMg/uc/Fed6/ql3/iYEl5HIIhhhHghh+3ei6vq11dzJI6rcNk/QfTgdwtZS5O+5bau0lfb3qzHbneW43pdnUxd2sliWEEcqn8yQ7tg98Vk0eKzffJCJ3BKwKQecfrYd/tU+2Uqvm7tz525IyBVfqwKzx7h6yDuDHaf2rWOMjHLllnPKjadc3Gqa/DLcSNLNkDPTj2GeAK9ngKxQxqcZBA9I4//ABXkPgyDztYBYcKM5r1W3l82L0887B/09zXpx9PDnbatIWIe4cnBW2YADvQ48ruP8xJNLb8xX5AyI4gifK+9ITyRnb3qgIIjcgtznkU8tmRsDIxQnyty4C5PFPDMzMQMYFKRoIj/AMFHnj0kcVCtbRmu0LfSvNS7Zg9pAOhI5zUi3XJO0d+aw6yHzxB/KhZQd6nIzVTf3CSyklfyhiMKPjvU/UrlYYXlX9fpT4NULF4pED8krg1YxRAzQy+W2CD3HPFEXcvfIxkD4qPHgEopyMZBNPUllVQTuxjiqycAd/ljqck/akc5CsjY3ZBz7dqdeJHFrDW8BYxpgEt2PemAeYxXj0rxiqAsG83b1yPtURwA555qQzlXU8ntz9qjy8PwMkmjJtu2Rn9W0j46VMgG2INnqTn7VEhyev088VYFAkMQAwSDRdIkuGUc+rqAe9ADKqSSsQpc456j7CjysBk43BeMHqT7D5q20/ToNPsItS1eLezA+TB3B7M3zSgWlaALqzW91F2gs1bPlt9T47gdql3fibZE9npEKQwheZGHJFU+qatdatckyvtjXO1I+mOKh9CoAzkbc1lfQczPOxkuJnmbOWMhyD+1R4mfzDIuRngbeBTzhlYg8IacvoXdj1Hn9qrHZsijay7V9I6/NGkvbrUIrOK4lLrbKY0BPBB5zUcnFp7s77v2p8K7iiHr9VNLuptjhRJMV6Arx71GlIJjBztXr96NEx/CtsPBagM67Cc55qpZogjMsgDHCj/SlO07lQAKDhT70PzNsezuaKqqkYG35oGSfljKDcp4+1IMRDazEs36sURBuDSH0jsKE4wdx6nrmiUCZyQAeOf61HY5fCglaPIc/vQ87FKqMvRD/wBJA+mmBmOFxge9IHYKcde9du2Y3daBzqAqoMmT9I9ql2Wkyahp9/fiURw2YHb/ADGNQg20q5Pr7VptNjaP+GmptI21pZuCPiiybZAAOQcbfvRgG2cKSenxSbkEKqq7396kq5SIhup7Uqa7a7SINvgSdG2jO5iKxy+WyDAwVXAx3rbWf/8AAsm4bRsPPvWIRQEynZeaT03kcJNgBx17GkkLKQcAE/NJgSKM8EUJxg5JrTk4u2eTXUmM85rqD0zwv4Si0lFu7rbPfMOXxn8P8KO/3rTge5x3496TGMKg+n6T8U8KBXHPO5XdezjwmE1CZHYY96cBxXVwrDeiilpKXtTYWlFNzS7ugpsOI4oSKc5GR80QnBb7U1OTzTZVXf2xys0YAHO9R2qoeXCAAc1qpAOSeh4rPalYvG5ZThD7V6+Dk3NV5eXjsu8UEzgpuLZPsOv2qPexG9ieKMCS4Vd1uf5x3j/fpQpd9s+4qGQcmhJPllw2HHqQ/wAtejPjmePbz8fPlxZf59qVtXh2L5SRhiAYNzjzY2BwQD7VX6hcT+Q4muVnE2VUkjcEH275qH42U6dqMV1Gu3S9Sy64HpjmH1Lnse9Z1bs3MJEjgFlGSf0Yr5OfH4Zafan6PPGWe/qc97LDLG6OOo2ADnaOuakHxIhjKQwq4mkyytyOO9VqzGC1Mrwh92fKQnjHuKrIo5GRpWIz3GMftVkYyztGvLuSV2eNCZZMjK8KBVbZzli6OTuDbUA6bvk0ssyhnjCjfIpCj2NW72UWn2UTeQqStF6znqccGtzFwzz0obV2t9QlO4LCTiRc59XuKj6xdpdXymM5EaYye5qFI7GTJJzjnHSmJnAB96lml8utNZ4K9N8rH6TG2TXoujZNlNKf0Eqg9/esB4XzHG5C9Iya9B01ClnbAD0AbiPmu2Pp58r2u48LDfKBjKqufbNDjQeYq7vSpGT7/FOD7be9z+rYaYhdrYHdgtz9q1pajXcpa9lYHAc8fFJ5jBdoH7+9JcrucY6haHGfytwPqFTSNJprFoY8j6anufLhdiMFjxioGinKKM+phk0/U7tEDgtgLgn71j66y9IepOJbm2tQfSn5jGgXP/ENwMOv0H4+aDHJlXuG5aXgZrmYog5w7dvYVrTnQo2O8gDnvT2JQqwPBYf60J3CPlDkHg0yZvyiM/SQ396ImXx/9uXeeMSE/ehKzBtynBORTtSJ/wAalcdGAcD70xmVgyrxt5BqgEhDOC+cdse9DmDEPg48sZ3UXPrIUZ45FJKnlhlJ5x9XvQdDGoVgxzuXg/NS7gokaA5GwYPzQowdgOceoJ8Ke5/pV3FDb28Q1LUMKi//AA0Lc+a3sR2BosR7S1g0qBdV1ZMtjFpb9yezfvVRf39zqNw09yxEh/R0A+MU6/vZ9RvVnuMK2eIweIx/KKjtzkAck9aiWhqOCw4J7VzEKQ2eCTilfLYz0WgyeqUDHAOf3ohqAi3VWxlsk/NNmZvJJPHO0Y9qPMm2M7gQAQN2OvHSgzqBLFH1A5J96BsvMaxqMEY/anowEgweg4NMcu05C/vRYIssAffmgMx8uFQBgKMNUYpu4x6Tz96mXBAtnIGVBxiorYBCZJJAxiqU0IrEnBytE+vaADx1phRkVlyetJuMcWTvGT06n70ZFdwEzkf9NRZXGOCSxFLLIECoyMB29JpkaPJLk4PfI7UWuK7IgSCWxmgepVLYX1epgDzin3D7HOxiW6detXfifTbTTjpotgQz2ytJls8ke9E19UG78kDv/tXLz6n9sUu5VG09a6PbICD2ohyDADgbq00TBf4Z3OfVumbI9uazIl/KwgwvQ1qYAE/ho52+kzGjWLMQRrHF5j9BwBXM4ZyGUgYz/akaVWbaFppDAM+eAMGnxK2cDf8AuEzpl/y8YP3rFRoVQ44yea20Nu0XgVlJyrocAViUY7BuHPzSemuQ/bwMUx1LUhbB+oKfYd645OearmAQQcc11FKDPU11B7kGz06e9PX3Jyfehqvf+1PbCgZNed7jq4UMyqBwc0qmVwSsZA9yaAlIW7UEq+fVIAfiu8gZy0xPxUUXLfqwB96Y08Sg4YsR/IMmm7YlPClvhjmlEu3052n+ULx/WiGG5aKI3EsbbDxtAyw+SKJFcRSr5kUqSIf1LyB8H2NNdyeASD7Co9tbxW0sksGIy3+ZEPok+W9vuOfignkts54H3qPKFdCM8mlBXAGCvspP9gaazDPTgf6+1axuks6Z+8s1LlWUE+xqsa32yk/p960t0nnReYpG4dap2XJPTFfQ4s9x4OXj1dqbWNIi1zQLzSnIRpwZYz/8uYfSR7E14rHLLbtJbSxlZlcrIhHKt0K17uwzkk4XIJb2NYDx14ZeW/h1u2dUluSI5oz1eUdH9hniuf6OPf8Apv8APyeM1WbklHkqucyYxuz0HtQ3EUKxxQq8s27bsA5apn/hu6WWaC5uoMwpvYRnOW+/cVaaClrpcgulXzJBGfU/O1u2K8uPG9GXIjadpzWK3t3cWivc+V+XC4yUweSf61UeIb1zdGBcFBGDuHckVoRcSF7+4mdzIYsEk/Vk9Kw+vy5vmZeAQMAVrKajE/1VMSBjuacoyy4/vTeAR9qNarumTdjFc5Nul6jb+HYgts8mCQsYQj3Oa9AtSkNkoX/NRM5PTFY3w/EV0yLPBaQufntWsUgIVPQD1j4rvHnt7T0xJa3L87gyHB6EZ7UquDFuyNrZ4PUU0Fm0p+MeZMNh9gKGzDeMrkHBIqt2n3AP4eKQLyRjPvUJfSpPQVY3efwoKnAj6iq1hsViXyT2qI0Wi/l2rzZ48vg/NRNRYPc+WD6cBj81KsE2afljwYhx+9V7SLI0s3ULxxWJO27dQ2Vg8saDASPmkJDyHBBYjqen70BiI49rkbhyTQzP6BtT09z71thIkhCW+VYEg9BUWUjYR1JXJpjSFBkZx2Ap6pvi2/qJ3MfiosWGokfjLeUg5e0jbPzQ44+Q3LDJyD2ot2Gl0/TZsjmN4z/9PFDVisrnorjg/NEoKqfN359WeBTpYy0uUYFmAB9sDqa6NSZNuDk+rPXirDTNPF7NJKxEdpGD5jk4yPYH+1FE0yzidRqFySLOA7VQf+c3tj2+aBqtxLc3DT3BQzdCQMgfAH+9FnvvxlwjIu23iHkwxgYAX5+fmo9zw0YwAE6ZqKgMo80NnjHB7n713Oxs4O0ZH3pzrklj3PFMVMzffrRmh4YZU92wTSRgNIMfz7z+1EJI9QIHpzz/AEqMzEBwvHpI/pxQi0uWI8M6aeN0148qg/ygkVTEZuGYscID/WrzxCFhvNPsVBCWdoobjoWGf9TVIBiNuuX4/oKFNiZlErAZOMfv71LhwjkNnAXv71GgPCn/AOYc1J3qsRcksWyKBvmFpBk8c4HagqcAtnOOVJ/vQy7LGqfqYnBHalm5ijUdRwRVSuzvZ3Y9eeK0miRWWl6Qdc1OJnaWQxwxnnePcfFZYHarEdcAY+9X3iWTybrTtPHMdnZptXpgtyalWXS2j8XaPdJs1DSlhjY4VlTJH3PtTbrwnFLE11ot3GyyDOQcoB9+xrFFzIxbcMYxuYcGiWepX2mShrK6kt2P/l4zH/Smjy37XVp4M1Ke8hS7lhSBZNzSBucDrUHxRqS6jr8z2xBtIAIUC9Dtp954q1i7tmhkliUlcM69TWfHCrtzyMkr3NIXWhVAYlgcn2PanK+1WZR8GmDCbOfUetHhh3bh2PJNaYNRW2DJwDzWuYEfwxO4kEy5xWW3rvXK5C1qkx/+zPc53DfUaxZVAVyQBj3prqDE3q6/UK4vuO0cJQnVuRnkjNPjNbmIeb4NPq27UyD71go8vEMnk5PFb3TwLjwY0ZyMRk8d6wCZONvpAGKT01yCcrjacH3rmd2zzSehcHkmkaWP2warBNpPOa6hlmJ4NdQe7SXKK2Fznuab5ck59LbVHXNAhj8lXkuX4Ay1R7nWY0C4BK9AAOvzXme5ahooeEGW7k038Rv53ce1VsLT3B3ksqjsRU5Iwox3qghbuBTCS3Wnc+1IWqBeOp612/CjJzTCfeuyABjmgUknpgUjH7ftSF89xTGbDbf9BQIcBQHDMM8Adc0nmkzhJGQMwwjjp9j805shTjIPeoyuith8be+R/wDf9aofuAlZccAHd/zf9qrbuLypmReEcbl/7VIadg6tIoVlk2tz9Q7V2pQiS1bYcbDuB/2rvw5ay048s3jtWOoI3kdD0+9YvxTdfjrsadbuPLtSFL+8h+qtVPqAgsJr1x6oVLAMOGbtXmkUsc1rKxdzc3EpkdcdM9xXq5cprTx8c3baMrRtFdmCPbGkZXPx0P8AegjKQKv8yDPzUhBjStSLHAQIg2985z/oKb5UhSA4zlRnHTFcHadol8xhs5GzgkDPNYO/l82ZsnJ961viObyYmt8fUNwPxWJmbe+cYFcuSuuE0aPqqZYx7p9g6E1DHXNW+jx7pS//ADVjH23l6eg2KFEhjxhQnFX0Y4CZy0jcn4qrtIv+GhdUK+nGSaswfKUO+cjpxXoeZM83/wBl4HqxNgDNMjcvLGjH1Z5I9qBHMkemxR8Zac7qVZArswH0jAY1F2sLho2xEu4s2ST8DpUVo9yoxGPVhs/6U3ziyySD6Y8AHviu81d7op3mQh0J/vUqy9rJpiNJYq2G+kfIqA7elY09AxlsUYyhbMoVwQcCoUjY3H6j0qNUhKq3fP8Azd6Y7nZyfWegHSnKpOAPq+aOLbaA8nqPYe1ERFWULuYY/wB6JG2eMNzwQKI4ZscekHtSONoPPfjB7UVbQxm50G7hRd0lpL+JVV/l7iq8yg3AYMrLjczY4welSLW9fTpo7yHkpwf+YexHerNZvD4LXP4EKyYYJvOA32zz9qlXUqLZaaJVa5mdoLAHmVuN/wALRNR1SOeIW1tGI7OMYVPf5oWoajLqM8ZZNig+lf04+FqLclWI6Bc7QB7UU6BsKr49CjOO+abctkQr75NFHpUgA5GAOOlRmdWucvwqDGBzzRCShmVeaGoJZlGdxHFEP6eCM+9PjXZICW53cD9qIBcgCJv5Sdo/1/1pum27Xeo2ttt3GeZQx+M5P+lJcDhUzxk4qf4cby9VN4RhLW3klK/OMf70WRF1m8E+sand5LRF2jX5AOBUK+tprKdYZWG7Yjj4DAGhH128SkZaUjIz7mrXxMgPiZoFH0xRKT7YUZqbTSHFa50afUeiQTCJR8kA0yU4jUDG7OTVs6BfBm4HcJNROfYgKKpSQ0544AqpTedwAGcCmZfIc/pNPckITkj2pibzEvH1HrVQwnBJ78HjtiiX8011dzTysGfaqg+47UzdslwBhjxQSxIJ285+kdj70QoKBijZKgdOwoYYLtZhjPSmtJtBTIwT25Yn4FW2meFtb1N1kgsnSLH+bJwPvRJLVQ53SEjHHxSLnG58gDjitqPAkMOP8T1yJM9VQgVX69pXh+wski026e4umPUPuovizaxs7gH6T0qWzCJQgPFEjt5Ei2gYH6vvTHjSMAqdzULEVg5wBwprXxf/AMsJcAkhzn+tZV1IP5hIwM4rfaNo73fgWKyuJRDHc7nDZ7VK1jHn4G0JlvWRxGBzVhZeHdX1JiYbUxqVI3ycD+tasHwf4aAZ7mO5uFHGPUx/aqvUP4lTbGj0zTUjBU4aQ9D746VN08dXtew6fc6b4VktpiokSNhkdDXlpdVDKSWI646Zr0zw7f3OraE897Kst0+4MQuAOPavMpPRLKmRuDkZxx1rU9LyavooVz34FOBCdcGmAPIfSCzM21QO5o0lrLZyYuYisp/Qewo5augjJzwK6uLrn6M/vXUHoGq+KfOn8lJ1HGcZ+o+1Wmh2NzOvnTB1XGVDVW+E/B8cKJqOpwrJdOMxqx3KtbcyEg7TwOmBjIrzvcRI1jX1OST1pd2DxihMcnJz+9N8zsBmqCkj3NMYlegyKH5mPqFDL7/p4+9QGzzjv7Um4gfYVH3Ls6gjPWkaQIpB6npQPyxPPFFXHI64qOGLBQCB8UWFdueMUD5HCRlj/eq/IkcggFXGMe/2omq3Kw2vLDPzVfp8wl9Sk7vc9BRKdqoMX4SVf1ybGJ+3FWb4ktA+P04x81T69IsVrbSAPgXiA56HNXYUNA244UnIrUurss3NMF4vuXt9FFpIF5uRnHXaOaw0OBE2SccqpI5HtW4/iBaMy2OodYT+TKvyehrFWeNy7j6Qea9O9vJlNVKuAYNLuIgvL4Lg/brVvaeRb6dA0/FtDEW3HqzEdKpb5zNHMRjLbQN38vvVfr1+JY/JRmaCOMbRnHIrOVXCM3rl+19M878AkhQOwqg7mp17Lu3AnLA4561COM1xr0QoGe1X2lJxGBxuaqRFyQB1rTaRCWkiOOFarPbGXp6bpEcc9l5DHlFyDUGeZzKkO4/l9T70a2k/DWrYOGK4zVaSVQkdW5JNdnC1KeXFup7mdjREmxExJzUKY7IIyxyfMeiWYMjMWPookSll3u0edqsM/FP3MXiVBtTGA3saAyhpsp2FSkkhfBHUdR81GhizmJYyclTy1KAPLdgp3Aj96YzcengEVM05BK0jp9IXJz7io1OzoIdgLzDLdl7ikUPK4I+kfUPenM2PzCevamSMIvVLnj6AKLo2Z9oYgYUDAFQYwWkBOSBxipRZmBkbkHse1DU7Qe2TzigMo9OOgJxinxRiS42LghupPSlVvVtzwBkGpECiN5367QCDQhZFYuQGAbGOe32oCbXuw2PQg5osrAIztzIfppLUgWkpI6k7j7GiumLRyEHKluR9hQbnEdrayr/5qsWPzR9VmkEyK23CqqDHXmoV3kaRpxOQNrDmoE3MSCzZXH9DRlYAhmOeM/aoO9fNAJOM1I84Asqjr6aJDSQVXf1zkf1qdZDyfD+s3Q+uTFup+Ccn/SquV92AB9HP3q0nUReFNNjztN1O0r/IFTaq+0i87UbOEAbvMTj5z0qRrLmbxPq0r8FGYLjvgYonh8ebr+nLwMT7z9scVE1BxJqWotuyXmfn9zRVlcZX+H+l4G3fdO7fNUMRLbm4GWOM1otTRl8EaMu4f5j4FUCAJKvmD5xVjOQUw3MCx7dK6Pc0KnPQ03PnSFgPSKcrEJ6RjFVlzRrnJIIHJqfZeHb/AFGPz2EdlYjlrm4bbx8Dqar2JYB34cdKbf3d1fIpvJpJwgwqSHAQfGKC+j1bw94d/L0qz/xK7X/9RMMqD8VX6j4k1zU4xLNdmOP6RFEdoHxxVKzKIcD0jP8AWpE1xFLbQRRejYPUfeieX8AlU/rdnbrh2J5/fNWen6fHbRrNcRKjMd4x1qJbx+bMZCMjIFTmnjYOzMTtO0DsKiwlxOZCWyRz0XvUOQ7lAb010l5hvTjpUJ5y781UtHacbgWG47cCtbo5af8Ah3qRd2Lcqp3fQPisRjJU5rc+FQJPAmqIe241FxYGNQYUIGwEcnG5jSttXjAcdSK5AXiUDpiuyFVs/V0pWe9vQfAMkkmkSowACzHBPyOlYO8QQ3txHkFllIP7nNbP+Hk7PDc25YemQH/aslr0Xka/qCY6PmkbynRulHdrWnq2QPxHStJ/EWHytbt7jp5seCexrL6e+3U7Vx+mUGtr/EiMNaWE567tv9qb7Nf5YPdGoAwTXUm2urTk964CrnGF7DigtOAey/3pHDOcucCmBY1bIGT815nvIZC7ZYl/bjpSs5xncK6SQ56kZ7Cgk+rjP2NA8lSDkE07IHPQYpqdPVt+cDpUe6uQoPQMq7iB3FA/zBLKseFQ/wAo6GgzEPdd8oMcGlsiRF5hG4IpclvntQbTc4aYqMsx4+KCXES5DYqauAoB7g5qPEp29APtRJD5aeoncBmgzXia7xIUbkIgIX3rtLuPLjMp4jkbcRjP7VA1Wb8XJHMSFJkMWak6UZHiCGWMxk5ABwQe9AbxOktxo4it5gHMilJGOADnitDY+ZPaQyzJsmeNVkTsCOp/esvrMiPph8xygWVQMxlwTnj0/wC9a2B2ksYWZWztXdj46ZFDSl1yx/xLw7dW+31Mu9D/AM6nivH4Xwq7xj0sjD5r3e0HmWILerbITj3ya8a8SaZ/g3iK5tScqzmVD2IbpXXC/HHkiE5IspXY9AFFZu9kO0hjwOlXl/L5djHEp5PLZrMajLnNayrOKonbdIQaEuf6VzHMhNIDjoa5Ou0q1wZ13dM1vPD2nuyhwNyBuo7Vg7MZkBrfeFb6SwZM+pOrKehreHtzz9NHcgK6wAcjlqgSv9Zz6dpxRfPeWaWd+Nx4+1RtpaZkboB0rtXnEly0NuD3d+aPD6bYAclW6iuK4tLYdeprgQseBwM9KiiJw4Ynqa4nyDJtGd0lDcjA7U2VyFGCfeoqXA3pjVmJOP8AervTwwsZnIA9WBis1by7iB0HArSaaB+Dvl6hFBGajeJ0m2SRV429WqNchZJ3csfTgCpZVDGx4HANQwVYn2K9O1GnMMq2emePvTHQ7VPHu/3p643AMTt9vmmOowxPTOTRBo4yJIFySZACcdgBUmPnqwEQ9QJ70NWVEaRiQWQCPHaiCEBVVjkKAcHpU2s6AumZ5FReCzcKOx//ABRoGVZpYmOBtB6fqocOJL0St0DE/wBBQ4ZCZGlJ3b3H+tNh2rOW1ORI1JK7efegXbOdK0nIH0v/AL0upyD/ABqbJb6lHFDmH/snSj22t1+9NgCA5PAp2VwCSfamKyg8dKNvjyAuM/NTaREnby1OM8jFXmuYjbTbQdLayBYH+Y1TyRvPc28S4JklRR/UVYeIZPN8Q3xXG2LbF+woqT4aC/8AiG33jGFJOPfFURkBnuSOS0kn/wDsavPC8gfxD39FvIScew4NZ6LLpKSTkZP3zQ21GssU8JaLnBO56zhYmYMQMVotXUN4R0fPGHP+lZ9tplZQBgCrEyDVlV3YZoXmflhQfqPJpxY7OAOtJKQD9I246CqzTnlAUKBkd2qPKS2V/SBmiHBjO04/5TUzTNE1LW2dbKESIpw0o6L96GtqjJClFGQ1NXAZVzgk7SAMk/bFbq2/h4gPm6lqBRB1SDHH7jipq6l4U8OsUtEjkuRwSo3O379qhMNM3a+H9XvgiWVk8MIIzLLwHqwn8J2OiacbvXdRMpCnybWPgb/kdTT9R8aajLGxsrdLdOzM3qH7Vj55bi8kElzOZpS2cuc4qrbIjHpu9znHTH7GlCn6uxp5RVXJJLE81xbAYAHOR9hRi1wUlN2K3PgmPd4Z1aMnhhn+1YcNngtjHUDpW48Abjo2qq+G9RGB7YqNYMCMqhVTkDFIwLMWbgLkj54p7p5MskJBU7yDntzQ2PIBIIFVm+2z/h4Y/wAZe8EMdtUfjGIweKLjjhkB/vU/wDK/+PSLg4aPnJ6nNO/iDb+X4him3cSR4xUjpl3jGXhbbdQk8DzB/rXoX8Qk8zw5bSg9HA/sK86GRJEW/mB/vXpvieJbzwTM45MaK4+OlPVMe8XmRJGK6l2u3OR0FdWnDb2s7i2TnBp5OxcHjPc964gRj1MB7d6hzTqH5LEDjPbNeZ9AV2PBHI9waVFLHJyR8mmwxSM3r47jHSnXMywhQzjJ74zigZdXCW0TMWQfGao/OlvboICBvPJHtUS/1MT3HlLIrYOGwg4q20S32sZZecdOKCdfsttYiIZ3ykKCPinwxhQqr1UcH3qPO7T6gOMpCvT5/wDvFTovUeTnjAxQHVQoA9qr9cuRb6bPLkhtu1D7H3qxLELWU8Z3SiwWD1BnOA2etBQaRu1PS7iMEiaP8wbvcHp+9WFo6P5bxxqhxuZT757fOartHzYahFKVDqw9WTipjKlnqTW5VWt5D5sTgnqeoFBYalN5tsskM08UhYbktziQ89M9s1pdPYNZIkkUkUpXIUvufH/Me9ZG7dZYwj723sBhfSwAPvWmsFaCzXbbPBCB1dsyP889qirLS/8A4XBOB5jLj/WsN/FHTPM0+y1aFRm2fypcdQrdCa2OlNiwdnYLuZssx+jHUn9qodNjl1jw5KdQYGPU5G8pG/l5Ckf0rWN1WLNx4xqUxluM9PLG0/JrN3r9RnNXN4k1rPPbXAImt3MUinqWBP8AfGKz103qI/uO9dL25yIueaUCm09etYaqdZRkkHPFbDTcLbK5438Y9qy1jHnHya19hCZLhIgOAPVW8Z255rmE5U56EcU0AtIWbjj+tFRfLYqR0oaLiQk9DxXauKV9UEG0cAHmo78Dr1NSWbbDAemVJxUaTnGT1qAcj8dacR+WST6cVHzuYcd6mxJuRx7jv2qKZagMFAPJYDFajTiJLa/VUx+Xjr0+az1nGDsYr6h3rQ6cMGULn8yFlPzUreJxjZoCcjAUdO/FRSQHK7cLs65qW4WK3ODjCgE9ulQmDCUtICd69KjRqHDgjninPsEKkk7mOMD5ppJGAhHpFLGQCrueR0+9CJ93bvaR26MAo8nec85/7UNyfw6bjy3JPsKt7y3jm0o3e7zJUhEbA9A3tVRc8RkZAwo//FRaBuCwb1GeqgUEt5SR4PVhx7c0t25QEEYEfXHao8Fvc3d5Fb20bzzth/KT6iuc5og2pEjVrnIBwy/6UsuH0jSCinnevJ9s1YTeH9bvtXuGt7F0jLKBJJwOlC1jTpdN0fTobi6hklSRxiM9KKqxnjcAPtT9kbNndj9qCrIFJByBT19Z4PFGUzR7cT+JtMjAziTeT8KCf9qhXU4uNTu7pyT50rYA4q20EbNXluVBAtbSVi3scYH+tZy1MkiZGC5weeg5oX00fhzMdxqMgOSlixLe2apbaGaciO3hkncrjEa5OT3rZ6Ro1tpEVxd3V/DdrNAcxQ9cH3qluPEdxOBb2rrYWRHEECAMcDHLdaKm64rJ4T0pGUBlkbcoPINZybG0uoI3nYoHJ3exrR6pt/8ACGlllbHmkEHv85qkjDSyRQRlVeZwit7ZP+tWJk600DVdRtvPsbTz487MhuS3eoJs5VkZJg0bKSBu7+4qfe6pLp9850uWW1htiqpGr/WwPqapPi1v/eV3Uf5kCOwU5AJFUs6UZTJIcg/aptnq15pEM5064MIJHAHBqAJDu4U0yX6I165yTUY2be6je38mbu8nnQ/+UzHaP26U1W2nZGoQ45KjFcsZkkPRcV3RsjGOmapsRckESEHjqev9aY5C5AXGKY3BG5vkUskiuc9SeKMh78jaF71ysvlsr5yT2rjsCYYn9qb+WqjGT96BQ2Vc49OKvvDPiGDw2961xBLKsy74wOhqiDjBG3IIxihsw7559z0osuhL27N/fz3BQIZWLBR0TPahF+OQMqAMimtnbuA4+K4jBGOh7UFz4VuDbeJLXbHLJ5mVIjGTVt/EWe2fULOJJAZ0T1J3H3rN2l1daZdpc2chikAI3fFRZX8y5knctLJIdzSNyc0jXl1oMnaA/RV7GvUA5vfAF3MVCl7cDb7V5k0YZwzHjqBWgt/GF5bQzWclsHtZYvL47fNT6sy6Z1MlFxkcCupY2ESlQCec811befVevXN/GrbVdVDdkGSfv7UsELORhR5Y6D2pllahvW2Fz7danySRW8OBjIGSF615X0g7ieO2iJkdVJGOTWO1XVxO4trclTnDZ7+1JrPiKRpjFAr8+lgy8EfftT9F0fzFWWVxKrcqrDDLQE0fSJfMDyoA/V3PQj4rWxKIbdWYgjaWHHUe1R4YVQLHkjZ++akXCt5WxcbiMAH2oI1kpbdPJEVeRicfFWcQXbkgH4oFvE0UagIMAcgtmjoQpxtXP3orpNgUMV74ry/xdeG88QR2uBiJsEg16PeTfh7cyOMqqlvtXj0Uv4nXpbsetXckA1UrRTqZbdPV64/pI7ferC0jOqWfluAZouUYDAb4PtUGNyJAcjJHt1rre7aC4ccjJ/T2pQR7jfNbwkqXjkAK59IOfetK93BE+TqDz3AUkq4wkQ9qxs92k+v2jFVUNuYqoxjjqR3qR4u1u7stLXNtaR2rDb5hI8127cCoqJr3iyebSR4cscNe6jJtLof/AC88g/Naye7hj1vR/D9sQyWNsHcIc7SR0z715DoV2bC4k1WVm/EAEoCAQpPtWv8AAnnTatNfTl3klG9j89hSe2fin/ivo5sNbg1WJcQ6iCZCv6ZhwT+/FeXXDZdsYAzwPavpPxtoR8Q+C9Qstv8AxUI/FW2P5l6gfcZr5ql5O4jaOmPnvWtsyB54okYyaHjFHgGWA+apV1p0fK/etvp8flxJcY+rgmsjpkeZAK2mnkvakdlO/FdMZ04Z3tMiVX9Z53dKWWAKoJGDuotvh87cZAy32qU0I/Ds555AXPtW9s6V1woWG3HfZUZzwpqber6Lb/8Ap1EZR5dRNI6Idw5FTIyRlTzkUKJATkVKSM7xUWQe1i8sxYOcDkVeaerSSDYNrBW/b5qujg3dEPA7VZ6NCXuZScsgRuBxnjpR0xhJ0VocH1JgFl/3qLISnBIOR37VIKzy2cewgFB19hnv71DkTYykHcMZye9QBfiE+/vRYguYlxxkEihsr9SQQegFFRT+JjU8tx0pSLmOOW28PiN2zJcylz9qqGZJZd7tiNBu+5FW+t3DRi1gzt2x5+3vWZlfzJVjHEQ5Y/PtWVp9xIHj3MxLS+tvsO1DtrqS1ulnjMnmOwj3IdrKKbKxZt2046Yx0pjCTcrKw6qc5qsrDVL+8bVLqKS7vHiVlUKJMY4zQX2J4dtWEYMiXbgs5JOKXVCF1m8DNg7k6f8ATSNdWP8Ag4tC0rS+eXPHQUVHQsVklWMbIjhj9+lGhbY/qUEbugqVB5a6NqpVf1RY3n70MLvjyNpc9NvbpRKl6a8y6ZrkkYwy2oQke7N/2rOQkxKmRuHQ+9ai3AtvC+onOJZriMH54NZhlKTkk8e1DL00WgxEWus4XDi1wGLdKp0t1S3jTzAFYep+pq38NbTpmvyZyRbc57VTRlPwsBYelhk4+1IrU32xvBWnbiSizHBqjs5I7W4gutoYRSBsOenNWl1ewReCdPgJJMs7KPjFUwdFt3YjfnqMdqsKvNS8Mwf4g92dVtINLkw7EndIAeoAqo8Q6jDfa5PcWhJtygSNsYyAPaq51UzbwgLbc8nOKbJKxIZcAnrVShCRsjGeRTJ87x254xTxId/T4BoU7OZDuIOCelRzpqFuSDnPvSNwdvt1ApAcR5YUzePfiqQ5stwWGKbkADJwvuaaQucYNKQuwBlOM9KJRMLg9xSDG0ccVyxFiSDwO1PSI7ueh4xQKMZHShPFtkJPqp7BVk59WOwpruyy5UYB7GgYIwgDDke1OzGSAgw3cmmhwNwP1ZzmlO4svQCgWVQepOaafqAXp0NKeJOelDJYuSKAqovQkcUhVQ2c5/euRhuw3Wk2s/GMGgUoPeuppYA4Irq0aezs0VvGXLBCPfpWP8ReIHyYYhH6xjzGYcn4xUrxHrDKrQwtkdCQM81V6HoLXMwuLhYi4ORkH1D/AL15XuSNC0SR2S4umcq3JjPQ/NaoDYmFxtHXIwcewrhHHbINq7QPnNCVjNPkNuGeMiglWwZVLsCI2+kY9Q+9JNGJ7tQSfSOualRqyBpMgt/L2IoUKb5GYj5H/agKsTKTzzRDwDwBxSBM8/70khCrhjj3xzQZfxrqn4Pw/KinbLOPL49vevPtGJMQm+kcjB9qu/4j6mr3cWnxpgoAzEe1VOlvAbdFLtnPGR1qic84tioyzKF4anWswltZC0nlnkq7d6rtZnkif1n0bewxim/iTBo0aKgljdPUD1wfagiwaip1sySHcIItp+eazms6rLrGss4dlhhOFVjn9669uFRrhWuFQyMMunJC9Mfeosd5BbBVtbfOON78k/JqG2t8O2um+WsOoOfLU7/+qtpZeKvD9iyQW8e3JzkV5PE17dMFB69AB0rYaH4SedFkuXwScrUNPV7HUoLkRzwsHw2459jwRXzf490tdH8banaKgWFpTLEB02tz/wB69z02wGnqVRjt+fivKf4wLG/i63uImDGW1UED4zWojz4ccVKtlO4YqKvJ9+9WVpGWxir9ZtXmmqRg1rNLcq4QEYlHlD71mdMG0c84NXcJZM7fqJ3j/lNd8fTz5tPpUYjuTC5XB9LfepNx+XMIjyVPOOlDtm3T2t0mP+IAHI4B6c06dt90zZ9JcZ+2aLPSLfsrpBhCG2VAkH5fTJqfdJ6oyScbOKEoJjJGCMUNBW0ZdBngfFWMMJEnpA6dTUSHC26KBy1SN5D7dxGBzioaS/pba821COwq10JlMc/lqF6ruYHJ461n/wATuVuCVAPbrV34edhO6EuB09R5II6D/vUrePsK3UtIF6pgjCnioU8hZ2BYEhfVVhA2yRk2+WAT1qBdRKZm4BC9GHepCg7iYxtHFTLCHzb5SHBYAYzURdw2AkAEdqudHgWGSSVyPSueaE7R9cmLX77Y90ipsHPBzVNBbym5W2Q5JGWHUL+9TJ5DdTktg5b3xRtNiAuJZEO2FBz7k1Yk9oNypWQxLJgZ5NRLiOJF3GXBHt8GitLJcXkgG1QG4z3oUsTgkb48kEHvg0RP1MW51y43B2PobODjlajkJhR5GTjBake9ZpRLPOGLYDYHYdKYLm3SMDcznr9qUTLfJ0XViFJUGIYxnHWolncsy7CTtJOe3FS7W4iGh6yyKx2GJyc88mqaa5aZSDgDHpKjGPvWTbTXMkS+FYWcgGW7Iz3wB/61mbl492VbJFXE0Es3gHTp1IKRTy72z39OMGqRFMjEekfJqxL/ABo/DO3/AALxM4zkQDP7gVSW4EltFk8BcAVbeGpAuleJYN/LWYYKPg8mqux8oIp3kq20rx8c1Vt6TprndpEFl5IZY5DN5vdeOlBCE27qxGeo+1EeG3eKQBnJ42844pYoITbuiSH6c80Ii+X+duQ8bcUMR7GAIyGHJ9qlGJlVugwKjSg+k5GD80SwMLmQLxtzxUa49Ln049RFSP8ALmAk4IoV2hWRznILZFGaiMx2460PAz8ntRkJc8gUbCLtBUE/6VUAWJsdeaIEyq5oykA5VSaaHJ3cADtQIp2MScYbiufdvVB0ByPmmBir5K7z2BooQnaTjrk89PiqgLsVc54pXJYrkYGKRgFkz2+aWbcSuFAx15oB7OpAFOUE9vvTWZdituAFLztB3D5qB0gUL8dqB0U9R96M7IFGB1oTERA7vXk0HLKDIMJnHenyzAMNpxn2qP5mOR07ilLGQYVML2qhGmAJGSa6m+X79a6g3Gm6T+KuxcSRuka/UCSQ5rYRRraweoKrdwvQH4plvGsaKFPAHQ0y4LEDD7VrzPcFJIZpNobIJ5qxtYBGMYyMdaFaWpJLNjHvVgqhVx70ApsJFgHk9K6MeWgyckjpTJ3O8L7VzSHIz7UDgxIzjC0OZgqkhuO9I0gH1c1Fmm9z05oPIvFFybrxXcMPSEbZz3osUZMSiNArDqxPBHxUGVvP1y/kJzmYnn4o8r7XXaepyBQRtRn8yeOMZ9huNQtSvHisSISd6+iLHv3/ALUt4DcXxUNgj1O/ZFqrmuTLMNgxAvEfz80RBi02QsSTlz6h/vVzZaWDhiKHE8pYEDrVpapcA/SME/poulzpNhFC4bbzWvs5/KKqVBHbNUul2EkqBiZNuPsKm3UlvZIVEhL45yelZrUi0m1IQKS3lgjPfivFv4gXDTa7GSynEWOPk1sbzUC0TDHB7Vnb6CC+mdZFG/K/fFWJpikALcfFW9ihXkjNM1HSv8PKzo6lGY4TPSptsuCnzjOK6SOeU0u7WPy1DfpqyhJx9OMd6TTYla0nZxkKMc/PSiJC8chjYYk6nHtXadR5cuq0uh7bmwmhdtskDpcRfIzgj+gpZXVpWKD0Z6+5qr025eyxKq7+qv2O01Yzzj0MMLG3IxRvG9Eus+XEW7xjFRSwSIrnkinXTtLBBnK5SkMQEKgEE471lXQl8pn0r705BtckZYk9aNHExh6jAHSuiAGAVGB2FNro0Oywv3J4+1W2ly79UjmEZchcf9IxjNQUZDG0Zt+M5zmj2MsS3SOiSK3AwG460pOlhHtM8mSThiPV7VBvvSHwMccOOg+KmSFIp523erONn3pj/nCSWT/JjIHl+5xSNVGt4AgWWRdigAhT3qytpD+Cu7h19K+kD3qvkYySeZIcAYxH7UWe6c2pVEwCcBexHeqk6QgVRpQwXIPX71PstyWTq0W2OQEs3tUeXESZUZhYcnHqJ+aZcOo0Z29TM4A2qfT0rNPihUvvyp4LEftTHVRI4ZjntSLjeAA5A/t9qktCkYMtywUH6QvU/ekYRE3IhbjA7GhknOT0PtR5drhSADjtQypzymBVS0W2uvwiXCogeOePbIGP9KjbvLRQI1LADHOcj5+KcBxleT803bggjjtRElbi4/ACyNz/AMJC5kiiT6WY1HQZJfkfFEEbFMEbQOmKfjaQQMtQGsr+4015Lm02sXiaJ1fuh60CFvLgCqP8v6cUpUheeWbt7U1QY1ClcY4BovtOVzIoUDqcH7U5pdpOfSycAe4ptuhMeD7YFPkG4bxyycEUUjMSvByNuBQSCWVCOvelBXIXaVGOnaklIAXGAM9qFNmjD5AO/BH7U2eMSJx+kc/NOIwHx3pUGYSe9VlCSME0URZO0Nyf7U5FHmYPpx2oxdQPSBycHNVKFIGUBR1qOq7jIGOMH+tGkH5jDJwOlCQqCXzzRD+VUenPtSqwJCg4OaaH35JFDOMnFB0yFpSV6DrXSc5PxSzBxt4yDTS4JA6DHNANvspFNPAziile60whm5JxjqKAZkyMZpBJt4IyT0NK6gjIGF7mmxg4O4entUCKq7sk80QqWAA4FKFYAcDbTjGzjrxVAWAVsE811OEeBjdXU0PYQuQDnn2+KUQpJIu3lR701tzHCk1LtbfcPVXme4aOPHAziiOwVMDtzzRTtjUgdRUG5mCRknOeuKARKtMXJOG4AHauYk/UwB7DHWhWz7lL8gueeKKxXdjBIx3oIs8pVCQQTVe18qyIHH5bcMe4H/5qXOcK2EyQM46VjddvDErKWKBhjk0GZkgNnq19HIOkrAN9+c1Dc77g7cll7VfaoiajpkGtRc7cQ3u39Kj6WrMzXH4eF5xxI7lIlPce9BEvZfqgQ+pz+YRQY4twx7cD7UkasfUfrPXPvUqJD7ZJqA8EahRk9KtbadLdSyAmq2OKTstSlibbvJxTap8ur3k0YRHMcY/l4NQJZ3dgxJ+ST1pjOT6QcCo7ybRgHNRrZ8lx6hkkgVltYN5/ikk0KyhSAAR0q7lZ2bHQHvR2tbqcCOC0lkL4AI6Zqzpmd5MYDMZYzcbyM4wwrR6dHvCYH6uSew96rdQF4mrm3vYxFJCPp9603hKwGq6zb6fI5KSP+ZjuvUiumN125545ZZeMaXRdHvtW0udNNtnnMk6pkjaoUdwe9bG3/h3FKRNrF++AuDHbjkfBNae3mghtYrWBBFaRrtjhTjbz1z3o3pkYSOWZh02nA/8AWuWXLfj0Y/nk7rOXXgnQWgCWktzaysuEZ23BvvWd/wDCWvR3Bt5bIyLG2InRhhx716SrptxwPkCnA25BByOowTgE+9THls9l4ccr086vtNu7JLYX1s0UaptdgchTVLZ3KX0krpzHG20cdR717BepaSWM0MqL5Ei/mZ52j4rw/Tri0j1DUIbSfiKQhlP8me3z8V1x5PJx5OLxX4Awyr+k+qjLErbMrgMpOaFB6Yzx9XvUwY2RsTgKMEV1jloBVIEeMdcGlhQm6SMAD80f61IUBiwxgEg/amSBHLRhserO4dRRB0j36ncqChUPl5CcBaHc3SKkixkEM3L9m+aE8ktyWRiuxDnaBjd96gztKzgFVJzQ2N5ig7ml3HFSLidUKKzYyN1QN0xk2lFxjimSSSmYMUycbeaINLfqAQrHcR1qR50c2li1WIGUjlR0IqGDlyjKpIHXFN27pQ28o2MBlpoR1uCmRFFsZTg+9cQZImJTLj3pyp6mAPqz9R70R/SpC9QOTVkZ0hy5RlAQjAzTWZsZIwalyLuw3wB/WhMjFeRzRlGbn0huackTHJPOB0pw9UnC8/ajepUII2k8ZzUAkbYrkhjgcZ6UolGCMDOMg0UIVgl3HoP60xY97kleNuKGjYgxJZyMkZo6rvtmBBJBHPtQo1WRyTxgYAokbbCyM3L8HFVUu0P1RtwdvH3pzpsZRtIPc+9Nt2MjgYG9Dkfapk35qBiCH9jUaiukBCs3OM4waGI12leMDnIqcYwV5+rpUOVRHlFHFGaYuR7GlQHeVYECmSROqgg8mlG7kgkkDk4rSOiXFzkjdnjmgyMN5UpwOBUmLGfM59PLfFRWXexbd6WPHzSFITuULupsSbtxxnnGBXMrI2MfemI7KjdstnIqsjlVLYb047UFxkkYwR0+acrsVbuT3NMAwR6vV3FQJKzMinoB7UIfUc8g0ZSpG1jxSDbjDcUCZPYYFNYjAUck96X7dB700vgdOe1A0Hcdv6OhFPUAKVHODgZpsQB5zjNPyAaaHbN3BHIrggxya4Pjk96TaWOAKqR2V+K6lKqpwa6ivaorQ7hngVJJWGI9AB3NI7iMHJ6Cqa/u2KjBy5HC9sV5Xuh97rEUOepPYioVxPJPCp59R7dqFDaLK++Y455X2qReOsJUAAIKgcY7mGICMsRiq68v5rNM3ETjI+qot94phsQ6ruZx0UmsVq3i+9ulcqY4Y2G3LtkD7VRc6l43t4Aw2NIR6doPNY/V9cbUiMYhjftIeazlxeKZD5IaWXoSR/cUax0HVdTbfFayybh9bH6f2olaHQrsW0jxSFZLS4XyZk6Bx2P7VTXuRfPB1SBiiHPtVpaeGTpbLe6hOIYoxu3M+Ax+1QY0W7LzKVO5txw2TzQgUXQc5NSYiPY/0pwsmAJGT+1FigwAOSfapVEWQjoxFK85De49qa0LDsRTRAXfBOKyobuWJwMUgQtU+GzU8EjFTDaoi4XbWhUCAEKeMg0DVNTbTJ1TzJvUucJmrc2T+auR9Wenao+sXw0ZLG7aKJ/qjZmRWZftnrWtM45au2MuNROqao9yy4IjCBTyT816N/C2yM2tX18SNtnb7QexY/NY7wx4e1TxlrcsWlW43yOWeR12xxqOpJ6fsK958N+Bk8O6AunR6gtxK8nmPKq7VZuwHfH3rOfWLphZ57qRGWfa2AFA2nn+lS43AHqYigPCbWVUkDeZj1DHAp6qp9SPu+DXmm5Xtusp0mrLb+Xjdt/egu0J5SZHbHAJ6VAldCd0YABOCDVVfxxyN5gjCMv6g+KZUx40u4mvrq6FtNMn4dv0p1ql17wXJDqE2v6VbLKJIgLuFRjbIB/nKPnuKlaSEjuw8kjsO5NXWr+JbHw9olzey3AyI2VYw3JJHQitceVjP6MJYxmmzfidOt5ty8DBJHNTLj8uPsc88VS+Hrg3Wi27uoDyLvMa9snNWUgIXPOD1Fe6PmWjhyFJIxhaC7kpvxt78DOacGJibJ4AHFBZsZ2nKjsarB9sm+YxhiSwJzUZgrSE4IIPHNS7XbHIsjIcLnp9qjKoKEnr1+9UciksTk5ApSu9dxIz2p6jEeT09q5FDLt+aAaxl3YnuMCh7G3ADgKeakCPahO/HNBk3ICM5zVASv5rkg4zx/SnlTuIAyCKfKWDDIwoFOWTEa+WASaiBlMSdsccUIIQWJOakEtvbdgE0EknOCMniiaRvLLElm2j460qQBfr3HPRiaKqMMKrjOeaeYgc9QOxNE0QoEjmycjGajklU28gH9VFYuA6tgjbQhEssaFmYnuPahssRUjauQfc8UrgDbzyfaj+Q0iEEYB4HxQwFLbGHqXuKIfC5XZtOMHof+9aOC2gvYxIretRyKypOX2DoeTVlYXLW0odTgHjHxRrGiXSugdQcYPBNR4wJYckfmr396vdRt1ltVniHBGSe1UEwMLq0Z9WOlQs7d5OIwTku3AGKurXwuqW/wCN1m8/C2+3IHc/t71I0K2SWaXUH2mG3jzh+m6qzWLuTV5fNmLncciP9OR049qrWpIkro2kag7rouok3CruNvINu/8Ar3rNSRukzwupR4myUIxt+KnWnmNcNcIdk6esPjHqXt+9WHiKNLyOz1mFNguBi4QDo4/71qM3uM2zOSz/AM3SlRcRBTg5/tUhtuzHHqb0nsDU/W9GbR3tBIykXEXmEDsajGlUECq2OoFBCnaDgZzyaOyenIPBroYfOl2Kyq3QbmxmiI/l84PWnMiuM56VZyeHdXSMOdLm8tugBzu/ehf4HrCv/wDu6bFNxZjaq3wYyoOMdaQLvG7HpA5PYVdDw9rTKwXS5Mnnkiu/8Ka8+1vwKoccqzfUPkU3DxqkjB8vI9+MinEllGCOfitafBV68MD20yM5+tX4Cn2oZ8CaoDl5oETkkA08oeOTKY5BIJPuKIp4POCKPqNhLpF+lldExXB9S/yuPg1DZjnb0Heqz6c0ZZid1dTSG7dK6g9iu7otkJioMcBY72BOec+1SCF3ZO0Gl81UbB+nHUe9eV7zgm0YH1dqgXOlvqU53uUx3B6VZI4cYOc9z70exZE3bRkknhqgytx/DpL+Tdc3sig9Cvai2/8ACrw/G/mTma4PcSHANbHzgUzljjvnpUWbVIIFLSyptHU5xiqItr4X0GwBEOnQLjoSCTVP4p8TaT4ZiWHyo3vX4htIV9cjH3qg8WfxPWxAtNEAnv5OAw+hKzHh1Pw9zLqV6/4vVJQxaWTny844WgdfaXqmsa9pkGrLHJc3khk/Bg4S2j7LJ8mvRR4I0C4g8iXTRbyKgBmtWJdG+AeMU3wxoo8xdXuU3XLkiMn9CdxVpdauouls4NrkuCD3UZ5qbGei8BypM4lvYig/y3QEs3/VmnJ4GZ3ytxbtH+qSNckfetnLIUiWTIGCQD2amCJY9QiuoTtWQbZYx0IorKv/AA6if0pqKhj0/LNVd5/DzU4CDAYZUJxlGIUfJr0KZvKjBJCsj+gj+U1K3Bow6tsbHOO9B5zH/C26aDfLrkYnPIiiiLKPvRYv4WOsZx4hX8RjLL5OEJ9q3qSkNnIRuwH6qfv2OCvCN2oMZo/gaaJZDrTLIB/lwW/RvnPvVmvhTwvfgW9xo6PJE2/ZdM3B+cda0cuBnkhQfUydaiXu6a3/ABKH/iIcNkfqHsau6aGgENtb/hbS2t7aJTnyreMIpH7U4zLtPbPUf9qDNJlI7lBhmGcfFMEobJCjB5zmpe1h8yJNHsY8Hoe9VEsbwboCwHcE1MkudnKsf2qHfQi4PnM4jMRBkZznafkdq454/wAejiy1dKq6vkDbA5bsmB1aq+dpnRzcgIMZFWVxeaTYRpI4wrMRE57t74rN6zrdpD+ZLMGd+iINzf8A9vWuU/le2WT2iXV3IsbeRKy7RnHZqprjw/cazdG2nnkXz09CdeOuanp/iN+6kmK3hKnaG4Zh8e1dNfXNlC8WnSM12EKAvyUB6jNXG6Y5O4fo9u1lpNlGcZiiCekdSO9WjsFUqDuJGSKrdHWSPR7OB2y6QDzif5qsAy7Wx9RH1HtX0Y+LfdKwBQkrtwB3psrLnP1BiA2KeDlOxOBQ8qswz9PcVplI3FLF9zHnp/pUcpho1Y5A+kUaff5MR6xvnbntUdimAx3ZFARydxGOBSqcRls/akRlLqWOF7nvTSufpIZM8UBMERrnvyaayqxVSfUoyf8AammRlUnGQBgCkk3JCVxlzhc0SnzDLJyMFN1D2uFXbilu3CXfp6KoX96aJMKAfeiHMCZDnhgcU0Flb1Ac9KUuoncZPIpxbMYV+c/TjvQDhQyMwyBS8E+pc44pUxEWZwEYjGKbvwXBPUjtQdNIm5gYQPTRIrsBcKkaj5oCkyT4JBGMUMFcMMAUQW4uJH9IdQPYUONNqll6nrmgowdzleF6UVmUxE85NVAidkmerU5DxuYnOeKjO5LhU6HqaKq9gS2PbqaG2g0rUgyNZy8hz6M+1F1OzMeXjGFYYHH9v3rPoAJFjPpB5HvmtPYXn4y3CTMPOXgbvjpUrcu4DPe21n4SGmxecl28m+RWHpUe2apxcb8HaDuG0/HzWxsreKSJ1uNrxH6y31A+/wBqpNQ8NvFN5lo0bQkEjDc0lavpCgQIr59qtNIhS90+90qUkeYPNgP/AD+1Ucj3FtxLC4B4o1rcGMoyOyyKdyuO1ViWIiW80F9BZSrsleZQUYdgeTV346kz4ghtlOUjgG2pVr4gMEyS3tkk8kYwGP1E1Ra9qMmsag07w7GCgKPaoXWlaQTEueBQwpUjk49xSsxwqscv3ApHIXIIyuOtVhOtda1W0G231CdQOmTkCrBPGGvRgBrxX+XFUPmoAGjOMdRQ/NLE5NE3WifxxriBjvgP2FEsvGOu3d7b2pMIEp9XHJWsnI524B71ovBVkLnWnuZPpt0BH3qVvC23TW+IdSk0jRZJUZBIMbAR81ntC8Q63q2t29iZIvLBLygDkp1pfHUhku7e280+Xy7L7VL8C2H4ayu9TkXEjehCey+9Zsbluwv4hzxTGyssI0qZbeB6gPaseYctjGARVpqk5vtVnm8zcQdqE9qGYDuQ8EHvXRyvaAtsuOorqsTZuDwK6iaaWbVVByX3Z6bRmj2qy3ZBIIQjI+/vT9F8K3Ueya9wuf0r2q/c2mnwFndFTu7nBryvchxWckQUs273LUkt7DYQb5pQgJ9+tY7xN/FbStPUwWRN1OOOOgryTWPE+q+ILhpGd448+lUPSmh6l4i/ifp9iXggZ5JwDhk5Fec6l4k1zXn3yTPFCeAo4zVLFBHCwZjul6neeKmLcDryU7nH9gKM7S7C08ib8tfMfqWzzXo3hLw9Ldz7mB8skFnPP7CoPgzwZc6k63FxGYbbqu76n/7Vr9f8X6X4Wt/8N07y5rx12rEnIjYdyaNRaeINdg0S2W3hXMjLtVe47dKh+FLI3E7ahPuYwgkMeQxPtWB0ddQ8Ua0Vjkaa5bme46pCM9K9VvLyx8M6OkcjiNIwTGpPLH5qRomp3aCa3tkJEjHOD3q0AMMUf8/cisL4buZNa1ObW7g4gU4iStms/mSqobOeftVQLVJQqIgIGakRXASFeecDmqTUrpTfGPAYDjrUgXSsojG3IHvUF0CHAmHQdRQ7iUKCR2GaW1P/AA3qwKiX8wQHpyKonrMHhRzz6cmhpMokw3RyQ32NQrKYy2RGRwaFLdCNwXPJPQVBJsm3WLxSDJgd0GOpHaoAk8mR4GOWRQwweNp65p+kXazx6gRxtmwfjIqNcvKtzbTRLmIbkmYDOV+/3xUX0f52Y8FirHBG3rj3qHfOZbGa2CRszRsYIS2Bv7N8H70CV5n9cs7RqRt64yPbFQTdW9pIoiG9l9z9Hz81Pi77ljL/AOGardSLNrU5kuR+VFbjohHep8Vtp+nhprhx+IfqOuDVbqF/PZ3N1JJOzgybkyemazMmspC7SSyM8j/Uue9cfC2vo5cuM03M+sw+V5VvhSRgtjkD71GvPEtloFjLOqxyXCKBHlepPvWEl1+QoY4I8Fh9XsKzs9xNfXih3YqXC4J4rpjx9vPyfpmtR7Fas8tnA05/OkjDyEd/ipnljGff6ftVXDJsKLkbEG2p6Tj6eoxhcHkV7JJHy97vY6AhWOKESrPu25JIGKkKy/hZJPM9I7Y5qOkqK4baTu5xVB7k+pEGcIuBk96ZNtbyw2ecdKZLIryD0nrmlYghhg+k8UDW2iUgAAAf1p7AEcn0Y4x2oSLmUliRx3p8bKoKkjBoGAMLdju78U4h965IOWzTGIwV7V2WjWQg+qMd/tRKjtIWlbcDyf70/BxnOQpph6jr1p+/DHFEELN5xxtIxSGVo1J2KR2OehpoYGU5GeKR2XYAoRFJwc0Cs7kDcPncO9K87DzGKgYA/emebG5ZUYNjsKa0qupbpjsaAsKksMgHzB3oBSNd4DEhTTkclwRwrdPim7Aq7jyO+KqGBvrIXAxSMNoLYyPanqU29Tlu1MMj7QgTBPU0ZKEjUAbOG60SM7cgLhBxzQXAUAknIp49WQM560DmxtyDg9iasdPEl22IZ44J1I5lOAaqGBZcEGjRnO3G0ge9RZW8traZATLLZMpGGUSDOaKlt5HmIuoWwkUjarNxg1jIX58xQGwfV6ql30CyxreyPgqNp2mp4usyamfTGvV/Ne2dc4yHFU0/g+/Moa0a3QnhVEmeaqrG4t2ga1uj5TO4IctjArQWGoIsD2Uu9oYGys8RprS6xqnu9N1Oy9V7aO2OsiDioJUTlm2gv2yMVp5rrUrFFktbwSwNlsyc/sao9SvGv5luIrZEmC5mC8D7irGMsP4qLuE7FeHgg+smqp5RvOAc/NXrSxkbgjYI9a47VWXtqqSK6ndG3IOeftVc0LzGY+xoqxMBliRRYYi+/G0AdM09gkSD1bnPY96IivHvPo4JNeieELNLTRpbljzL68/ArCxwvMoCjaXIBx+nPFeg37LpPhcRYZVhjEQPuTWb706cXvbIajcPq+rzsgyWOxftW4li/A+HobToWTY5HX71R6DpcEwtZ45RvDZKkc1fahNFPqSW0R3TYzsJ4OKN4ye2TOkJFhkUOvcnrT2hzbNtTbsParyS2HJHA7qD0NRJYkQOC2Qwyx9jWts6VjErgZ7V1WMdvG6BjXVU0pvEv8ZNPtmeHSYHu5sY8zPpFeT614t1vxBKTdXTrD/JHkAUGLSoxbPMzk46KBioTzkZAGB7CvPp32YkCpyzZb3IognEYwmAe1RyS2ckmrPQtKTVLsQySsigjOB1q6UPT7a71a8W3tYzJI3GcZAr1jw94G0rw9bDV9euFklXnaei/tUaa4g8HaUP8Os4t5HLt1z715xqviLUtWneS5uXKtx5YPpH7VKum78U/wAUpbyFrDRg1tbD0+YBgt9j2rI6JpmqeJ9SFrZh3DNiScr9A75NVWj26X17HHLnYv6fevQdS1qTwxpKW2jxi1WRRllOTzUSVtP8U0H+G2iLp0UiSXTDLIn1TH7+2a801LxHqHinVB5xJBO4rnjFZu4mmlnLTStK5BLM5yT9varPSEHoxkM5HqHUVGpXpmm6vHbWkNtEgjCqWUDvjr+9avSLstpUuqMQiuSFLdGA7ivLYY5LrUrTTvOZEmvEiLj6gMc4rb+PdRbTLMafaxKkEBNvGAeiqOKKqLjXDNfSEDd6s5FW+lyXNzcjCnymI9XtVN4U0iG8TzZnZjjNeg2dlDaKpiXCgfT2qCxz5VqB6RleprJa3rKJcCBJF3ntVvrV89rZuyqCdp6151pkEut6vNJNPsEPRVXNUbXSLlzYSM78Z6DriqDUNXf8SVilLKvO5hir63tVi0nbuOZDjI4IqoGhWruPVJjzMEM2c1KsW/hcNFok91eOsAnl3s7c4Ud6zt//ABDhv5p9F0UeXbQndNcn9eD+mrPxlKNO8L3EyKGWNPKWMnAGe9eH2VxJbmUqfUUG4+4NSJa9PvdQjSR3V8s3ILtng84H9ayl74jc4CkgFscHrVTcX0sqSN0wAOvx/wClUzSGSLzjwScYzW7GZe9LDW9TN7qZ9TlNijYD3qmkuADxyfnqK0V1GkNkFRAPQrZxzn71lnPqJ+adJ5WzsR53UKVJAING0mEzarbxnsdxqM4BjQY9zVx4XG7Vw567cVqe0tmmy/EFZFUtgEmpcVwVkVlbcen2qGAJpI1cDHIp5gVka4UlCDtwOldXnl7XcMuYZFII6H780vmq4DAnO7bVTBcSxqULlhgdfvR0dlJYHod370XazDOzkDHp96R5HMZKgjngVFtpmkkcHr1zUhnYpnPIPWibPSUZI2sTjnNEWTEeCB/So5OZmyMnb1pvmsRt7UNitIjDnt7UOZlKbFdtznlq4Y2lcD3zQHHOM9qJaKeMndk+2aaNxzkgVH+kGjcgcnNE2eCQ2VyxNNnfbJGpVvc0+Egv05proXLMWPAIxQ2QShWVguOcmjjyhJliCW5C46VF52KQcYFEhmk3JuIb1Y6UWVI6Ar6cfao8wDKw3bRU2YBpMn26VHuo0iQ+ncD2NVaCsjRxKrgY7NTVdA2OSRyD70EOZcKeAKfE+JHGAdvSjIhxI2QeO5PvR0DkZwFY8ZoDxogTjIfk80UuyR5B9OcbaBXVQcSS5I6haCriIsY1xgHBIzTJHyhGB1oaKzws3mMNoyAPihEpJpI5FyGUnllKYPwcVZW8hdHt2ICzDBDcfvVzomm2PiZXub6BvOCrHuR8cVdJ4N0a2GPJkc+7SVK3rbzG4CpI0UjAkcDqc4qbYandRxPbxyMIm/zAqdRW6uG0vT3Zk0eB2XgM7E0N/ESQBPL0u0Abg+moswZiC/e1VVfeImbjd+r4PxVxe6cuon8VppRZlA8yNeFx8VG164TWHSN4EhESbgY6h6dNJHMoRyoLBeO1K1/442S3JlXDR3CHlT0NV0lrNCcYDKThs9jWu1O28xFu0cxyqQCQOuKqbzbMqTMoDMdpx3qztnKM0IR5mJVZQx4OetPnjXeFQZAq3kjjltTI6AknAHtVaqYuGTPFVixJ0mFrjU7S1RerZf5q88cXzbo7FDhcZz1ofhGBTrkpPJjjBX96rPFkxm1mVWH0RgDFSe1vWG1/4IMk9rLcOoARgqfJxT7yBpLuS7RtlxCcK/yaleF4kt/DC7B9W5j96otG1aW+vHspkBQMz7gecjpUnp0vUi70yZZYmWR28z9Xo5LU2/tJhMZmAyV/Y112ChW4iYo444qTbSNeWyPKc46j3qbXXSot47posggDNdVm6IHIVMD2zXVraaf/2Q=="
    }
   },
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "![hard.jpg](attachment:hard.jpg)\n",
    "*문제가 발생했을 때 이해할 수 있을 정도로만 파악해두면 좋습니다.*"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## NONLOCAL"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "- nonlocal == global?\n",
    "- nonlocal == free-var!\n",
    "- python3에서 추가된 keyword. free-variable로 유지시켜줌."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "ename": "SyntaxError",
     "evalue": "no binding for nonlocal 'b' found (<ipython-input-17-2984eeff3ced>, line 4)",
     "output_type": "error",
     "traceback": [
      "\u001b[0;36m  File \u001b[0;32m\"<ipython-input-17-2984eeff3ced>\"\u001b[0;36m, line \u001b[0;32m4\u001b[0m\n\u001b[0;31m    nonlocal b  # 갈 곳 없는 b\u001b[0m\n\u001b[0m    ^\u001b[0m\n\u001b[0;31mSyntaxError\u001b[0m\u001b[0;31m:\u001b[0m no binding for nonlocal 'b' found\n"
     ]
    }
   ],
   "source": [
    "# global의 반대말이 아니에요/\n",
    "b = 6\n",
    "def f2(a):\n",
    "    nonlocal b  # 갈 곳 없는 b\n",
    "    b = a\n",
    "    print(b)\n",
    "f2(3)\n",
    "print(b)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "ename": "UnboundLocalError",
     "evalue": "local variable 'count' referenced before assignment",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mUnboundLocalError\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-35-fc05b8794a57>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      9\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     10\u001b[0m \u001b[0mavg\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmake_averager\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 11\u001b[0;31m \u001b[0mavg\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m10\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m<ipython-input-35-fc05b8794a57>\u001b[0m in \u001b[0;36maverager\u001b[0;34m(new_value)\u001b[0m\n\u001b[1;32m      3\u001b[0m     \u001b[0mtotal\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0maverager\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnew_value\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 5\u001b[0;31m         \u001b[0mcount\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      6\u001b[0m         \u001b[0mtotal\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0mnew_value\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mtotal\u001b[0m \u001b[0;34m/\u001b[0m \u001b[0mcount\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mUnboundLocalError\u001b[0m: local variable 'count' referenced before assignment"
     ]
    }
   ],
   "source": [
    "def make_averager():\n",
    "    count = 0\n",
    "    total = 0\n",
    "    def averager(new_value):\n",
    "        count += 1\n",
    "        total += new_value\n",
    "        return total / count\n",
    "    return averager\n",
    "\n",
    "avg = make_averager()\n",
    "avg(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'count' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-36-1d5fd60ce2a0>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     10\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     11\u001b[0m \u001b[0mavg\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmake_averager\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 12\u001b[0;31m \u001b[0mavg\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m10\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m<ipython-input-36-1d5fd60ce2a0>\u001b[0m in \u001b[0;36maverager\u001b[0;34m(new_value)\u001b[0m\n\u001b[1;32m      4\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0maverager\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnew_value\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m         \u001b[0;32mglobal\u001b[0m \u001b[0mcount\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtotal\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 6\u001b[0;31m         \u001b[0mcount\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      7\u001b[0m         \u001b[0mtotal\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0mnew_value\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      8\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mtotal\u001b[0m \u001b[0;34m/\u001b[0m \u001b[0mcount\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mNameError\u001b[0m: name 'count' is not defined"
     ]
    }
   ],
   "source": [
    "def make_averager():\n",
    "    count = 0\n",
    "    total = 0\n",
    "    def averager(new_value):\n",
    "        global count, total\n",
    "        count += 1\n",
    "        total += new_value\n",
    "        return total / count\n",
    "    return averager\n",
    "\n",
    "avg = make_averager()\n",
    "avg(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5.454545454545454"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "count = 0\n",
    "total = 0\n",
    "def make_averager():\n",
    "    def averager(new_value):\n",
    "        global count, total\n",
    "        count += 1\n",
    "        total += new_value\n",
    "        return total / count\n",
    "    return averager\n",
    "\n",
    "avg = make_averager()\n",
    "avg(10)\n",
    "avg(20)\n",
    "count = 10\n",
    "avg(30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "20.0"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 우리가 원하는 결과\n",
    "def make_averager():\n",
    "    count = 0\n",
    "    total = 0\n",
    "    def averager(new_value):\n",
    "        nonlocal count, total\n",
    "        count += 1\n",
    "        total += new_value\n",
    "        return total / count\n",
    "    return averager\n",
    "\n",
    "avg = make_averager()\n",
    "avg(10)\n",
    "avg(20)\n",
    "count = 10\n",
    "avg(30)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "당장 nonlocal을 어떻게 써야 하지 걱정하지 않아도 됩니다. 다른 사람의 코드를 읽을때 당황하지 않는 것부터 시작합니다."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "좀 더 이해하기는 쉬었다고 global을 사용하지 않기. 전역변수는 오랜 프로그래밍 역사동안 권장되어 오지 않았습니다."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "source": [
    "**종종 class보다 읽기 편하고, 성능도 좋을 수 있다고 알려진 function으로 구현**하려면 잘 이해하고 있는게 좋겠습니다."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "## EXERCISE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "slideshow": {
     "slide_type": "subslide"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "After local assignment: test spam\n",
      "After nonlocal assignment: nonlocal spam\n",
      "After global assignment: nonlocal spam\n",
      "In global scope: global spam\n"
     ]
    }
   ],
   "source": [
    "# 다음 실행결과를 예측해봅시다.\n",
    "def scope_test():\n",
    "    def do_local():\n",
    "        spam = \"local spam\"\n",
    "\n",
    "    def do_nonlocal():\n",
    "        nonlocal spam\n",
    "        spam = \"nonlocal spam\"\n",
    "\n",
    "    def do_global():\n",
    "        global spam\n",
    "        spam = \"global spam\"\n",
    "\n",
    "    spam = \"test spam\"\n",
    "    do_local()\n",
    "    print(\"After local assignment:\", spam)\n",
    "    do_nonlocal()\n",
    "    print(\"After nonlocal assignment:\", spam)\n",
    "    do_global()\n",
    "    print(\"After global assignment:\", spam) ##!\n",
    "\n",
    "scope_test()\n",
    "print(\"In global scope:\", spam)"
   ]
  }
 ],
 "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": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}