{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "3718ebd2",
   "metadata": {},
   "source": [
    "# Validations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "be3d42a1",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T15:34:30.308623Z",
     "start_time": "2023-09-22T15:34:30.305814Z"
    }
   },
   "outputs": [],
   "source": [
    "import os\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"0\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "aef9d75f",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T15:34:30.919845Z",
     "start_time": "2023-09-22T15:34:30.524032Z"
    }
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "from matplotlib import pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "1b79e6ee",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T15:34:32.355951Z",
     "start_time": "2023-09-22T15:34:30.922115Z"
    }
   },
   "outputs": [],
   "source": [
    "from suno_utils.audio import Audio\n",
    "from suno_utils.tasks.mert_25 import preload_models, encode\n",
    "\n",
    "_ = preload_models(checkpoint_filepath=\"/home/tony/Data/MERT/mert_test_8x_400k.pt\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "e37f87de",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T15:34:32.363575Z",
     "start_time": "2023-09-22T15:34:32.357958Z"
    }
   },
   "outputs": [],
   "source": [
    "audio = Audio.from_file(\"../audios/canon.wav\").get_segment(to_s=10.01)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "f06559a0",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T15:34:33.034748Z",
     "start_time": "2023-09-22T15:34:32.933292Z"
    }
   },
   "outputs": [],
   "source": [
    "out = encode(audio, do_clustering=False)\n",
    "out2 = encode(audio.get_segment(to_s=5.01), do_clustering=False)\n",
    "out3 = encode(audio.get_segment(from_s=1, to_s=6.01), do_clustering=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "id": "e89985f6",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T15:34:44.913479Z",
     "start_time": "2023-09-22T15:34:44.911323Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(250, 768)"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "out.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "0e55f806",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-21T15:02:07.512186Z",
     "start_time": "2023-09-21T15:02:07.361895Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f23498a9120>]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(out.mean(-1))\n",
    "plt.plot(out2.mean(-1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "a3e8e3a0",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-21T15:02:08.505846Z",
     "start_time": "2023-09-21T15:02:08.378542Z"
    },
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f21f2a44ca0>]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkAAAAGeCAYAAAB8Rr70AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8pXeV/AAAACXBIWXMAAA9hAAAPYQGoP6dpAAEAAElEQVR4nOy9aZgkV3klfGLNtbbu6r1b6ta+S0hCQjJjMJZpGJihx1gW8tgCDQMDY/HB6Bv4LCxLDDCWjYUssOQRYowHbGuQhW1ZYCwQLXY1Etr3vaVeq7qqa8mqXGO534+7xI3IiMilsiq7uu55nn6qOjMqM3KJGyfOe97zaoQQAgUFBQUFBQWFFQS93zugoKCgoKCgoLDUUARIQUFBQUFBYcVBESAFBQUFBQWFFQdFgBQUFBQUFBRWHBQBUlBQUFBQUFhxUARIQUFBQUFBYcVBESAFBQUFBQWFFQdFgBQUFBQUFBRWHBQBUlBQUFBQUFhxMPu9A0cqfN/HgQMHMDAwAE3T+r07CgoKCgoKCm2AEIK5uTls3LgRup6i85BFxi233EKOPfZYkslkyAUXXEAefPDB1O3//u//npx88skkk8mQM844g/zLv/xL6P5/+Id/IL/xG79BVq1aRQCQxx57LHT/4cOHyVVXXUVOOukkks1myZYtW8jHPvYxMjMz09F+7927lwBQ/9Q/9U/9U//UP/VvGf7bu3dv6nl+URWgO++8E1dffTVuu+02XHjhhbj55puxfft2vPDCC1i7dm3T9g888AAuv/xy3HDDDXj3u9+NO+64Azt27MCjjz6KM844AwBQLpfx5je/Gb/927+ND33oQ02PceDAARw4cAA33ngjTjvtNLz++uv4yEc+ggMHDuBb3/pW2/s+MDAAANi7dy8GBwe7fAcUFBQUFBQUlhKlUglbtmwR5/EkaIQs3jDUCy+8EG984xtxyy23AKBlpS1btuBjH/sY/uAP/qBp+8suuwzlchnf+c53xG1vetObcM455+C2224Lbfvaa69h27ZteOyxx3DOOeek7sddd92F3/3d30W5XIZptsf5SqUShoaGMDs7qwiQgoKCgoLCMkG75+9FM0E3Gg088sgjuOSSS4In03Vccskl2LVrV+zf7Nq1K7Q9AGzfvj1x+3bB34Q08lOv11EqlUL/FBQUFBQUFI5OLBoBmpychOd5WLduXej2devWYWxsLPZvxsbGOtq+3f343Oc+hw9/+MOp291www0YGhoS/7Zs2dL1cyooKCgoKCgc2Tiq2+BLpRLe9a534bTTTsNnPvOZ1G2vueYazM7Oin979+5dmp1UUFBQUFBQWHIsmgl6dHQUhmFgfHw8dPv4+DjWr18f+zfr16/vaPs0zM3N4R3veAcGBgbwT//0T7AsK3X7TCaDTCbT8fMoKCgoKCgoLD8smgJk2zbOO+887Ny5U9zm+z527tyJiy66KPZvLrrootD2AHDfffclbp+EUqmEt7/97bBtG/fccw+y2WznL0BBQUFBQUHhqMWitsFfffXVeP/734/zzz8fF1xwAW6++WaUy2VceeWVAIArrrgCmzZtwg033AAA+PjHP463vOUt+OIXv4h3vetd+OY3v4mHH34Yt99+u3jMqakp7NmzBwcOHAAAvPDCCwCoerR+/XpBfiqVCv72b/82ZGhes2YNDMNYzJesoKCgoKCgsAywqATosssuw8TEBK677jqMjY3hnHPOwb333iuMznv27AmlNF588cW44447cO211+LTn/40TjzxRNx9990iAwgA7rnnHkGgAOB973sfAOD666/HZz7zGTz66KN48MEHAQAnnHBCaH92796NrVu3LtbLVVBQUFBQUFgmWNQcoOUMlQOkoKCgoKCw/ND3HCAFBQUFBQUFhSMVigApKCgoKCgorDgoAqSgoKCgoKCw4qAIkIKCgoKCgsKKgyJACgoKCgoKneL1XcBDXwVUH9GyxaK2wSsoKCgoKByV+M4ngInngdGTgOPe0u+9UegCSgFSUFBQUFDoFNUZ+vPl+/q6GwrdQxEgBQUFBQWFTuFW6c+Xd6Zvp3DEQhEgBQUFBQWFTuHW6c9DzwKz+/u7LwpdQREgBQUFBQWFTkAI4NaC/79yf//2RaFrKAKkoKCgoKDQCbj6w/HyD/qzHwoLgiJACgoKCgoKnUBWfwDg1R8CntuffVHoGooAKSgoKCgodAKhAGlAdhiozQL7H+nnHil0AUWAFBQUFBQUOgFXgKwccPyv0d9fUd1gyw2KACkoKCgoKHQCToDMDHDCJfR35QNadlAESEFBQUFBoRMIApQFjv91+vv+R4Hy4f7tk0LHUARIQUFBQUGhE3APkJkBBjcAa08HQKgZWmHZQBEgBQUFBQWFTuCwFGgzR3+ewFQglQq9rKAIkIKCgoKCQieQFSAgMELv2dWf/VHoCooAKSgoKCgodALZAwQAw8fSn/Pj/dkfha6gCJCCgoKCgkInEG3wjAAV19GfTgWoz/dnnxQ6hiJACgoKCgoKnSCqAGWKgFWgvysVaNlAESAFBQUFBYVOEPUAAUBxDf1Znlj6/VHoCooAKSgoKCgodIKoAgQEZTClAC0bKAKkoKCgoKDQCZw4ArSW/pw/tPT7o9AVFAFSUFBQUFDoBHEKUEERoOUGRYAUFBQUFBQ6gTwLjEOVwJYdFAFSUFBQUFDoBPI0eA5VAlt2UARIQUFBQUGhE8QqQIwAlRUBWi5QBGgl4LWfA//4YWBetWcqKCgoLBiiDT6uC0wRoOUCRYBWAA7ddxPw5J2Yefyf+70rCgoKCssfaQrQ/CGAkKXfJ4WOoQjQCsDUJL0ieWm/UoAUFBQUFgzRBi95gHgXmFcHarNLv08KHUMRoBUA2ysDAIjn9HlPFBQUFDoHIQT/60ev4IFXJvu9KxRxCpCVBTJD9HdVBlsWUARoBSDjVwAAxG30eU8UFBQUOscT+2bxp/c+j2v/6el+7wpFnAcIkMZhKAK0HKAI0ApAjhEgeIoAKSgoLD9MzFHCsW+mCnIk+GvcKv1pRQmQygJaTlAEaAUgD6YAqRKYgoLCMsRsla5dDddHqer2eW+QogCpLKDlBEWAjnIQt4EsGPFRBEhBQWEZghMgABifq/VxTxjiRmEAahzGMoMiQEc56pWS+F1TJTAFBYVlCJkAHSrV+7gnDHHDUAGlAC0zLDoBuvXWW7F161Zks1lceOGFeOihh1K3v+uuu3DKKacgm83izDPPxHe/+93Q/f/4j/+It7/97Vi9ejU0TcPjjz/e9Bi1Wg2///u/j9WrV6NYLOK9730vxsdXZk22Mjcd/MdXCpCCgsLyQ0lWgEpHggKUVAJjHiBlgl4WWFQCdOedd+Lqq6/G9ddfj0cffRRnn302tm/fjkOH4r8cDzzwAC6//HJ88IMfxGOPPYYdO3Zgx44dePrpwPlfLpfx5je/GX/6p3+a+Lz/7b/9N3z729/GXXfdhR//+Mc4cOAAfvM3f7Pnr285oDo3I37XfKUAKSgoLD/IBOjQ3BGgAMW1wQOSArQyL7iXGxaVAN1000340Ic+hCuvvBKnnXYabrvtNuTzeXzta1+L3f5LX/oS3vGOd+CTn/wkTj31VHzuc5/Dueeei1tuuUVs83u/93u47rrrcMkll8Q+xuzsLP7qr/4KN910E972trfhvPPOw1//9V/jgQcewC9+8YtFeZ1HMurlIJBL844A86CCgoJCh5g9khQgQmjYIaBKYMsci0aAGo0GHnnkkRBR0XUdl1xyCXbt2hX7N7t27WoiNtu3b0/cPg6PPPIIHMcJPc4pp5yCY445JvVx6vU6SqVS6N/RgBABUiUwBQWFZQiZAE30WwFyJQKW1AZfngB8f+n2SaErLBoBmpychOd5WLduXej2devWYWxsLPZvxsbGOto+6TFs28bw8HBHj3PDDTdgaGhI/NuyZUvbz3kkw6kEBEhfYQRoYq6Op/erSHoFheWOI0oBkglQUxcYC0L0XaA6DYUjG6oLjOGaa67B7Oys+Ld3795+71JP4FZXrgL0X/7mYbz7L36GvVOVfu+KgoLCAjB7JHmAuAFa0wHdDN9nWEBuFf1dGaGPeCwaARodHYVhGE3dV+Pj41i/fn3s36xfv76j7ZMeo9FoYGZmpqPHyWQyGBwcDP07GuDV5sTvOllZBOjVSToDbazfV4wKCgoLQlQB6msatMNSoM0coGnN96s06GWDRSNAtm3jvPPOw86dO8Vtvu9j586duOiii2L/5qKLLgptDwD33Xdf4vZxOO+882BZVuhxXnjhBezZs6ejxzlaQGQC5K8cE7TvE7FoOq6qxSsoLFfUHA916Riuuz5KtT6uZaIFPhN/P58HpozQRzzM1pt0j6uvvhrvf//7cf755+OCCy7AzTffjHK5jCuvvBIAcMUVV2DTpk244YYbAAAf//jH8Za3vAVf/OIX8a53vQvf/OY38fDDD+P2228Xjzk1NYU9e/bgwIEDACi5Aajys379egwNDeGDH/wgrr76aqxatQqDg4P42Mc+hosuughvetObFvPlHpHQ6gEBMlaQAjRXc8EvEh3/CJgdpKCg0BVKNbpuaRpQzJiYq7k4VKphKGf1Z4eSUqA5hAKkCNCRjkUlQJdddhkmJiZw3XXXYWxsDOeccw7uvfdeYXTes2cPdD0QoS6++GLccccduPbaa/HpT38aJ554Iu6++26cccYZYpt77rlHECgAeN/73gcAuP766/GZz3wGAPDnf/7n0HUd733ve1Gv17F9+3b85V/+5WK+1CMWWmNe/K6TlaMAyZK5UoAUFJYveAbQQMbEusEs5mrzGC/VceK6gf7sUCsFqKCygJYLFpUAAcBVV12Fq666Kva+H/3oR023XXrppbj00ksTH+8DH/gAPvCBD6Q+Zzabxa233opbb721k109KmE4AQEyVpAJeqYahD66qh1VQWHZYrbqwIaDLxl/gd36G/FZvAmHFnEemO8T6HqMt4dDTILPxd+vsoCWDVQX2FEO0w0IkImVowDNVAKy1/BUCUxBYbliturgbO0V/Jq3Czvm7wQAjC/SPLA/+97zOO/z96V3jrb0AKlxGMsFigAd5bDdsvjdXEEeoJmqg98yfowvmF+B5xzlI0AIAaoz/d4LBYVFwWzVwYBGCcmgNw2ALJoCdN+z45iuOHjmQEoQbksPkDJBLxcoAnSUw/aDKxmDeH3ck6XFbKWBq8278Nvmj1Gcfqbfu7O4+OEfA1/YBrwek3RenwN+ciMw+fLS75eCQg8wW3FQBCUdJmmgiOqiTYTnGUOpZfOkSfAcygS9bKAI0FGOnC8pQFg5CtBspYFR0BBI4hzlOUBjTwLEB8afbr7v2X8G7v8c8JMvLP1+KSj0AKWai4JWFf9frZUWRQGqu54onbtpZfN2u8Aqk4C/ci46lyMUATrKkSPBwmGuoC6w6twMbI0uPr67/Etg/+Pbz+ADf/0QvLiWfo8RWzfmqpiXxmpHx2w7hZWH2aqDIiQChNKieIAm54N1wvFSFKBWHqD8apoSTXygPNnDPVToNRQBOppBCPISAbKOFBO07wEPfRU49PyiPYU7PyH9Z/kToDse3IMfvTCBfdMx5kwecOnGXBXz27zl/x4orEzMVh0UteC7ParN4tBc79Og5SGrblp2GO8CS1KAdAPIj9LflRH6iIYiQEcxGrUKLC2QYE14/Y2Q53jxe8B3/zvwvU8v2lMQ6crL95Z/6Y8vyE6cNC8IUMxVMb9tBUUgAKClvxtPAl77Wb/3RGGBmK06KCBcAqs5vU+DDhGgdhSg6CR4GWocxrKAIkBHMcpz4WnEFtz0K5ulwsRz9Gfl8KI9hVENHnu5EyBCiCh9xUrzogQWowB59fA2KwUvfZ+efF65v997orBAUAIUfLc3WjTaY6LHPiDZV5SuALXwAAGqE2yZQBGgoxjVuZnQ/03Nh+MeAWWww6/Sn3GKRY9g1SVytcxLYPJiHGvO9FM8QPy2lVYCa7BSoYoHWPYoVR0UJRP0FkaAeu0DCitAaQSohQcIUJ1gywSKAB3FqJZnAADzyIvbnPoRcCKc4gSomr7dApCpS+rXMlc/5MW4EacA8RKYF0eAVqgHiE/srs32dz8UFoyoArTepPMNx0u9VYBkAuSktsFL0+CTYBfZtimBigp9hyJARzHq83TxL2nBzJzGkdASPrW4ChAhBHk3IEBkmRMgeTGO9SZ4bXiAvCNA+VtKOCz+oTbT191QWDhKVQeFkAmadjQemuvt+nGolwoQv28RVW6FhUMRoKMYToUSoLIxKG5zG30+IOvzwPwY/d1ZHAWo5vgYIlLb9zI3AHsewQnaPpyvPR/vTfBTPEBKAervfqxAvDg+h5fG53ryWI7no9zwQm3wQ4R+poupAKWboNvwABlsUv0yv/g62rHow1AV+ge3SklAzRiA52gwNNJ/AjS9O/g97oTdA8xWHaxGQIC0ZX7yd3wfX7f/FGsxg19U3g1gNLxBO11gK20h5h4gRYCWFDXHw3v/8gFoGvDIH/0GLGNh19h8ErxcAiu6MwB6rwCFS2BtmKDTusAMm/5MWnt8H9CV/tBvqE/gKAYnQA2zAEejVyROv+diHX4l+N2t0TlWPcZMtYHVWkCAyDIv/7gewVrMwNI86HGm3rZKYMubBHYMXgJTJuglxcRcHXN1F6Wai0p94SnIs4wADUglsKwzAwMeDvVQASKEhAhQbOAoR1sKECdAMcfk9GvAjSfQETYKfYUiQEcxSI3K0J5ZhAcDwBFQAuP+H45FqJHPVByskgiQ5i/vk7/n+SLPyYvraGunC2yZlwE7hiqB9QWT88F3sOr0jgDJozA0EIxgvqcKUKnqhhoMFpQEDUgEKOa42/8ojQB56b4u9lShl1AE6GhGnZIA3y7CYdVOzznSCFDvy2Az5QZWI/AgaMv85C+rdrEEKC0HSHiAlvd70ClIgylAXn3RvGYKzTgsjZOo9YgAGfCQBXtcna5jq7VZjJd6lwYdnS2WaoJ2WiRBAwEBirso4WpsvTc+KYXuoQjQUQytQfMyiD0AV2MEqN+ZOEtAgMrzM8howQl/uRMg+TPz3ZjXwgcuqhIYBSEgDYn0KBVoybAYCpCcAo3hYwDQcRg1x8dcvTfl7YmImpQ6DV4oQCkEyEzxAPG/Z+uzQv+gCNBRDE6AtEwRHqgHyO23B2gJCFCjFA4f0/zl7QHyHJkANX9+LrutUW/OHPHZ1SpZSQTIa0CHdPJdoT6ghuvjPbf8DP/v3z+xZM95uNxbBahUc1HkBmjDBgY3AQA221Th65UPaGI+TIBiR85wdOQBirlgEQqQIkD9hiJARzFMhxGg7CA8pgD5/SyBNcrA3EH6O18gFiGXyI0QIH3ZK0DBZxY71oPd1qg3l3qqVXqbRvxAKTraEQ2fW6EK0KuT83hi3yy+/eSBJXtOWUnphQIUygCyi0CBjpg4NsMJUG/Ws+jjtDULLJUAMX9QnAmaE6DGHO0GU+gbFAE6imF5dJEwcgEB8voZzDXFWuBzI8G05EVIg5YHoQII2sSXKeSyV1MJjBCYoK9Pj1F5NHkBXik+oIYiQAAwW2HE2PXTTb09RK8VoNmqE2QAZQICtImPw+jRPDCuAFmGBqBVGzzbn9Q2+JQcIPk45d2KCn2BIkBHMWxGgKz8EFzWBu85fTwJ8vLXquODxWMRCJkeGbJqLPcuMKkE1lTKIsGJTfeb38vQbSulDNakAM30ZTf6jZlqcKz3oiW9HUxKClDNWTjpmq1IClBmUBCgdQYlQL1SgLhytWGIjrfw2kqC7tIELZexVRmsr1AEqM/4lycP4qHdU4vy2FmPngjs/CB81j3h91UBYhlAq44L5ugsQoeOUQsTII0s79KP7Psh0WG20hWmGUP0QretFAVIlcAABC3kADDfWBoV9HBZKoE1emyCtotAgSrHw2QGAFDuwXMAQRfYhiFKatJN0NwDlDYKI8UELauyqhOsr1AEqI84OFvF79/xKP7r3z26KI+fI/REkCkMw+ceoH52gXEFaPXxizorJ9Ogc8B42c9Y5h4g+TPzowuq9NoMvxEOliQE5kpUgKIlsBVqguYlMACo9KhbqhUm5TZ4t1clMK4ABSUwngbt9chDwxWgjcP0wizRBO17wXG0UBM0QH1ACn2DIkB9xDiTbyfn64tSo88TeuWULQ7D12kJrK/dQId5Cew4wGIK0CJ4gHIOJUBlmy6WOlnmBEj+zKKp1tICq4GEF9wUsnRUQ5XAAEQUoCUgQK7nY7oSfOd6oQCVak4QgmgXgeJaAEDRpap57Gy8LhCUwFooQPIFW7dJ0KoEdsRAEaA+YkZaLErV3p6cXNdFkS0c+eIQfOYBis2RWSpMSQSIK0CL0AU2wK4OG/l1AAB92ZfAgs+sicBGO7vkWIGourZCS2BkhZbAZqrBd6W8BB6gqUojJED2zgQtK0C0BJZ3mcqb5tVpEw3XxzRTyza0UoDk46sXCpAqgfUVigD1EfIV2kyPCVB5LhgFkR8cAeEKUL9KYI0KMMfacWUPUI9zgFzPxyCbFu0X1wNY/iUw+TNrmmsWfW0y6WkiQCujBObVw501fmW6T3vSX8xWg+9KeQk8QHIKNNAjE3RVVoAGRAnM9mvIodYTBWhS6gBbU6TEJbENnh9TmgEYKbPE20mCBlQYYp+hCFAfMSPV6OXfe4HKHJWIG8SAnckFJbB+ESA+BT47DORXSR6g3hKgUs3FKo1eVelDGwEABpZ7G7z0mUUJT/QKM6QARd7bFUKAGrXwScVboR4gWWEuL0EJbDISJrjQHCDPJ5irucEk+EyRlsGY8rJaK6WbldsEnyk2WsyI6fWJw1BFC3wu/UH5+paWBA0oBajPUASoj5BJz2y1tyenapmqIBUtD2haoAD16yQoG6AByQPUWwI0U2lgNZj6NcAUILK8CVDoM4sSnmjGkbxtkwK0vN+HduFUKQGaIQUAgF+Z6ePe9A9yWT2OAH37iQP4o7ufxsHZ3vjwmhWghRGguRqfBC95gDRNqECjKKVPbW8T3P+zdiADkxGg5BJYG4NQgfZzgBQB6isUAeoj5Br9bI9LYPX5GQBAVaNEg7ADkvSrHHRYaoEHgvp5jz1ApdIschp9X0mRKUDLnQDJvq3o5xclQEoBglOjJbAxsgoAoK1YD5BEgGIMyX/2vRfwN794HW+/6Se448E9Cx4s2msFiK+Jg7qkAAHCB7Ram00fWtomOAFaM5CBpdMgxGQTdBtjMIB0E7QqgR0xUASoj5hdxBJYnSlANT1Pb2AKkNavEpgcgggEC0iPu8Cq02MAgDpsaPlh+lTLvQQW6uxK7gIDECY90cV3hRAgjxGgQ2QYAKDXVyYBmm2hAPGOrbm6i0//01P4na8+iL1TzfPk2gVvgbeZirJQBYjv/5DOvsf2AP3JFCBaAls4AeIZQGsGMjA4AUoiVk67BIgpRL7bPO5ClcCOGCgC1EeETNA9JkBOhZaB6gYtA5A0SXYpIHeAAYuWBF0vjQMA5owhGCyMzFzmCpAmEZemyfYdmaCXtxm8XXgNSoAOktUAANNZeTOXfJ9ECFCYjBBCRGv8f33r8chaOna9ehgf+7+Pdf2cXAHaNEJV54WaoEvMxD3ACZBQgGgr/Ch6Q4ACBSgblMAWrABZwe/RCw/5OFRt8H2FIkB9hCxR97oE5lYpAXIYAQok2X4rQNESWG8VILc0AQComMMwLXoVZsHtiVegX/Bl1SdS8mrqCgsRoFpk25WhABHWBTaOYQAsH2mFBc7N1d1QS3pUASo3PHH/x952Ir591ZsBAI/vnel6LTrMCNBmRoAWmgPE96Moe4AAUQIb1WZ7EoQYKoEZLRSgdj1A8v1NBEg6RlfY9/JIgyJAfYTcpSH/3gv4NUqAXIstGowAaf0YDOrWgdJ++vuqbfSnuTgKkM8GoVbtVTAYATLhLdkwyEWBtGBqERXHi5Y0UxSgpm2PUvgNesKcJUXUCLsSX2GdYLMRRTk6CoMbjE1dQ9bSceK6ARyzipbLn9g709Vz8hIYJ0ALTYLmBChPoh4gqQTWAw8Q7wJbU8zA1OkpMVFZ4iX7VgqQLitA0c5NZYI+UqAIUB+xmDlAYATIEwSIZVb0YzDo7D760yoA+dXs90VKgmaDUJ3MKhgWJX2W5i5zAhScvLRIOa9puG2KCdpz+jgHbinBJmxXkMEsmAK6wozQURUnOgpjvsbKS1kTmkZVjzccMwwAeGzPTFfPGShAlEj1SgHiI32aPECY7WkJbO2grAC1yAFKmwQPALoOsPmLTV48lQR9xEARoD6BEBJpg+8xAWLdBYTJxhpXgPrhA5l+jf4cPoa2sQKLNgvMrFIC5OdWw+QECF5yW+tygHzFGPH8uB0oQO4KIUAaS4KukgxKhBOgmf7tUB8wE4nViHqASowAFbNBmN8btgwDAB7b23lwJCGkWQHqiQmaIOuzi6QMJ0C8C2zhHiBCCCbmJQWIeYCSS2BteoCAwAidWgJTBKifUASoTyg3vNDBG5WsFwqdEyC2aGjMELyoc7HmxpsHUQLAzB76c/iY4LZFmgZv19kk+PwoDDPwACVe0S0HSKRHj5Qw3ehok5RRGCtFAdKZqlhFBiWwLsgVrgBFk6BFxk4mKNWce+wIAKoAddoSP1d30WDH2Kbh3pigZ6sOMnCggxGpSAlsVCst2ANUqrpouPQx1gxkYLIusGQTdJseICA5CyhkglYlsH5CEaA+Ier56XUJzHAoAdI5AWIH46J5gMafBb50NvD3VzTfxwnQyLHBbYuUBJ1zZgAA+sCoKPuZ8MTivCwhlcD0aAksqgClBCH2dQ7cEkKX0npnuQK0wjxAXF1eVaAXPlETNO8AG5AUoFPWDyJj6pitOtg9GR4n0gqTrIxUzJgYztPnXGgOUKnqoAjpAslinyUbiLoKJXgL9BlNzNP1ZyBrImsZMFuZoPkFG7+AS0PSOIxQG7xSgPqJRSdAt956K7Zu3YpsNosLL7wQDz30UOr2d911F0455RRks1mceeaZ+O53vxu6nxCC6667Dhs2bEAul8Mll1yCl156KbTNiy++iPe85z0YHR3F4OAg3vzmN+OHP/xhz1/bQsAXKJ47MVNpwO9hp5Lp0gVMzw0CkBSgxQpC/MWt1M/z2k+bW47jFKBFSoIusEGo5sA6YUSkCtDyLYFpvtwGHz6R+R15gFaGCdrw6OvOFQZXvAK0cZiWauYjJbC5WjMBsk0dZ24aAgA82qEP6HCZfrdWF23kbAPAwktgdBI8+w7bReqrAYSP0NAIsl4p4a/bw6FSkAINIGSCjlXBOlGAksZhhIIQ54AFBlAqdI9FJUB33nknrr76alx//fV49NFHcfbZZ2P79u04dOhQ7PYPPPAALr/8cnzwgx/EY489hh07dmDHjh14+umnxTZf+MIX8OUvfxm33XYbHnzwQRQKBWzfvh21WrDYv/vd74brurj//vvxyCOP4Oyzz8a73/1ujI2NLebL7QilyALlk+ZOjYXA9igBMhkB0lk9elEIUGUKeOpb9He3BpT2he+feZ3+DJXAFmca/IBPT3TZ4bXiCsyCB2eBV4p9hd++AkSk95NECJDfY7/VkQrLo1fpxYGBQAFaqQRoiF5oVJJKYFkrdHtghO7MBzQpzdPKmvS0Unf9BV3UzcoKEG+BBwDDQsOm+1lkFzzdQvh/GAHiJmggYR5YxAOUWioUJbAUAkR8wOk+fFJhYVhUAnTTTTfhQx/6EK688kqcdtppuO2225DP5/G1r30tdvsvfelLeMc73oFPfvKTOPXUU/G5z30O5557Lm655RYA9Mt2880349prr8V73vMenHXWWfjGN76BAwcO4O677wYATE5O4qWXXsIf/MEf4KyzzsKJJ56IP/mTP0GlUgkRqX6Dl7zWD2aRtejH0EsfUIYRICtHr+h0axE9QI9+Paw2HH45fH+aB6iHChAhBCNsEnxheL1YgHSNwHGXbxiibFyPeoC8iL/AlwiQ34gQoBWiAFk+fd2DA0MoYWWaoPlawkMJKw0vREbmYxQgAHjDMYEPqBNMcgWoEChAwMJa4UMlsEwxdJ+TpWNOBryprh8fAKbZfvNSITdBAwmt8IIAZfBHdz+NN//pD5MbWJJM0NELEVUG6xsWjQA1Gg088sgjuOSSS4In03Vccskl2LVrV+zf7Nq1K7Q9AGzfvl1sv3v3boyNjYW2GRoawoUXXii2Wb16NU4++WR84xvfQLlchuu6+MpXvoK1a9fivPPOS9zfer2OUqkU+reY4CWwoZyN4Rw9+HrZCZb16VWFXWQEyKRkoOdzsTwX+OVf0d8tVm6YlAiQUwXmaTozhiUPkEiC7h0BKpfnkNfo4jKwel0ojdVdxid/Of05qgBFS2Cy0dmLqGv+CglCtBkBGh4aQomszBIY7wLjhmQAqEglKdEFlokSoGEAwPNjpSbVKA1CARrIIGvoACh5WIgRuu764RKYBDdLO8EGFqgA8fckbzO/oB4oQLHRGXy9snLY+dw49s9U8eyBhHNFnALkewCJkELVCdY3LBoBmpychOd5WLduXej2devWJZaixsbGUrfnP9O20TQNP/jBD/DYY49hYGAA2WwWN910E+69916MjIwk7u8NN9yAoaEh8W/Lli2dveAOwReo4byF4Tw9UHo5DiNH6JVTNs8JEL0aMXpdAnvxX4HZvUBuFfCG36O3HZY8WTN76U97AMhJ7/8iDEOdO3wQAFAnFrKF4VAYmddjr1E3mJyv46b7XsT+mc4632TfT5TARktgvtRVFyVAZCUEIXqOmP02MjwscoD8auet3csZ/GJq7WAW/JwuG6EDD1C4BLZhKIf1g1n4BHhyX/uk8XCZEaC8Cf2vt+Nbmc8CIAsyQjdcX1KABkL3uVnqAxr0Z7p+fACosayinEVVK5kAxfoGJQ8Qb6yYrSYcV8IELd0vk6EMtSegvrgX2wrJOOq6wAgh+P3f/32sXbsWP/3pT/HQQw9hx44d+Hf/7t/h4MGDiX93zTXXYHZ2Vvzbu3fvou4nl6iHcxYGc4wAJR1IHcLzCYqgClBuYBgAoDMTdM8VoIdupz/Pez+w/gz6u1wCk8tfWrC4BEnQPVSApqnSNKMN0ueSFaBG/zug/mbX6/jyzpfwf36+u6O/09MUoGgJTCp7+VEFaCUQIMlPsXpkROQA+ZWZPu1QfzAjrS8FpvLMSwRovs49QGbT33YTiDg5xxSnbA3Y9xDO115AEdUFGaEbXrIC5GXpxVTBX1gbeVUoQJQAGbICFNcKL3mA6qx9PvHCNc4ELf/OQ2FVCaxvWDQCNDo6CsMwMD4+Hrp9fHwc69evj/2b9evXp27Pf6Ztc//99+M73/kOvvnNb+JXfuVXcO655+Iv//Ivkcvl8PWvfz1xfzOZDAYHB0P/FhNigcpbGM71VgEqV6vIaixGfoAuFHwultHLyeiHngN2/wTQdOD8DwKrT6S3yyWwOAM0EBAgr96zQZX1WWquL+lU9YJuwGNfce8IKP+8ylqL52Mmc6dBTn+Ofn7RWWAy6eGG6Cqx2bb9fw8WHSyHyiMaBgt51E2qHPjVlVUC400WQzlLlLkqUidYXBcYx7nMB/RoB0ZorgCtsYPvYw6NBaVBO56PQoIHCD3qaq2w/csyBUjTtPR5YE4MAUr0AMWUwOSLEK6IqyygvmHRCJBt2zjvvPOwc+dOcZvv+9i5cycuuuii2L+56KKLQtsDwH333Se237ZtG9avXx/aplQq4cEHHxTbVCp0AdT18EvTdR3+ETQRmqs9Q3lblMB65QEqz82I37MFXgLjClAPlRCu/pzyLmB4C7D6BPr/2b1BXkacARoIR8lHo+K7hMMGoZbNYXGbC7rA+0dACODk5CQ+YtyDfG2io7/TpTb4qIIXVXXkzi9OgOZBfSBkJUyDZwpQBVnkMxY8m8VArDATND8pD+ctoW7Mx5bA0hWgdgMReQr0qB0cZ1mtviAFyPEIiohXgDS9N55GrgDJxm3eCp/WBUbMjAhQTLxwjRtAzdc63QKy7CJbeYD6hkUtgV199dX46le/iq9//et47rnn8NGPfhTlchlXXnklAOCKK67ANddcI7b/+Mc/jnvvvRdf/OIX8fzzz+Mzn/kMHn74YVx11VUAKDv/xCc+gc9//vO455578NRTT+GKK67Axo0bsWPHDgCURI2MjOD9738/nnjiCbz44ov45Cc/id27d+Nd73rXYr7cjhCYoC0RHNYrAlRlBKgKOwgD5JPRO1kwCAF+8BngkQTl7HmW0XT+B+nPwiiQHQJAgunvXAGSQxCBcJBYj9Kg/XlKLGp24DXyNLqweUdACOBvTX0Ff2B9Excfvqujvwt7gMInlGgJjEgdJoIAEV5u7P97sOhgBKgGG4WMAZJlFwAryGfRcH2hbMgKUNgDFN8GDwBnbBqCqWuYnK9j33R7x+YkaydfZQbfsRwaXZugPZ/A80lQAosoQDzYNVoS7hRcocrLBIgpQPEmaPo6Xd0WN7X0AMWVwMxMQOqUAtQ3NNP/HuKyyy7DxMQErrvuOoyNjeGcc87BvffeK0zMe/bsCSk1F198Me644w5ce+21+PSnP40TTzwRd999N8444wyxzac+9SmUy2V8+MMfxszMDN785jfj3nvvRTZLF/nR0VHce++9+MM//EO87W1vg+M4OP300/HP//zPOPvssxfz5XYETnaGcxaGRAmsNyWK2vwMAKCCPDjNMG0+Gb2DBWPyJeBnf07bOc/5j8FAVYAm684zM/vm8+lPTaNlsP0PUx/QutMTFSCiG9A0g3ZE9MoHVKGT4Bv2KnGTCwtAte9jIMpzM9ju/xTQgIzb2YInG9cNuCCEiAGWXNUpkwwKWj2cA8SuNrkC1NSOexTCr1egA6iQDPK2CS03DMwDhlel5QfTbvUQyx58bdE0SnB4h5M8DoOrQdEuMICWg07bOIgn983isb0z2MKmxCeh5nhCURoxg+9YDvWuTdCcfAQ5QGETtNajrla+f7wEBgAWnwcWqwDR/XG04HvUUgFyY0pghhUYuxUB6hsWlQABwFVXXSUUnCh+9KMfNd126aWX4tJLL018PE3T8NnPfhaf/exnE7c5//zz8b3vfa/jfV1KzEoS9VCPPUD1MvU71PRg4TL4ggEvdAJNRWk//enVqaKz5qTgvskX6c/BTeEOjdUnUAI0yTrBYgjQ3qkKdtz6c+zSM7C9Ss8IkF6jmSBeNiBAvmYCpMdjIHw/SKVtE6VH/h4b2NWs5nd2UpCvcvlgV9sME6AKsiigHjsMdZ63gq8AAtSozSELOgl+TcaAkZO8fLVZoLimb/u2VOBry0DGhKFrwgQtD0QtpZTAAKoCPblvFi+MlYCzN6Y+3xTL0rEMDXlpdEUWTtclMN5hFShAEQK0iAoQN0KnKUCOlgHYxWRnJmh2fBqZ4DWpEljfcNR1gS0XBF0agQcobR4YIQRff+A1PPJ6a2MiJ0B1iQBxBciG2/5crHnJbD7xXPi+iefpz9GTwrePMh/Q4VeARhkoM7+LRIAeeX0ah8sNVH22+PaoFd6uzwAA/JykAGn0ObxepSDX54Db3wJ87R0dRdjnnvw78bvW4aItb2/Cgyt52bgJmpe5NC94LzVOgMBKYIs1BuUIQr1CTyY1ZJA1DQzkslIW0Ez/dmwJMSsiNqgCUczQkzvP9am7nvCvyMNQZfD8oPFS6+OGl79WFzLQpJN5TuteAeL7l2SC1o3edLUKD5CsAOkpJmh2sRZSgFqaoKX3kJesTVsqgSkC1C8oAtQH1BxPHHhDeSsIQkxRgJ45UML19zyDP/ynp1o+fnWOkiTXChYNyw4mozvtzsWak2IDDj0fvm/iBfpzzcnh27kR+vBLQQZQZiiUAcSvUOvgEnFvCFCGDUIV7aVgChAA0qsk6J2fA8aeBPbsaj+/Y/xZDE89Lv7bqQJkRgiQ4wafn6wAAQjJ7ZwMlfl9K8AE3ajRk0ldy0LXNQzmzBU3D2xW6gADgHykDZ6nQANAMUEBWjdIvzPjpdbH5uH5YA6YfDLPotG1AsTVl4GENvielcAazQTIFCWwmAtFdrEm1i4As0nWBeEBko47fiFm2AGpUyWwvkERoD6At6jqGpWp2+kC44pRO0bp2jwlQL4dyP8m67qiJ9A2FaA5KbAyqgDxElhUAeKt8IdfTvT/8NdQA7tC6hEBEpPgiwEB8li3SE/mYO19KOh8A0TLdUs8+g26L4ReWXYq28vjLyy44XwStrhyn48uXW0KBYjQ+7QVoAA5VRo14OiU8A9kLZEFtFIUIDliA0CTCZr7dQq2Ecq9kbFukL5/7RAgPk9rtJgJlXNy6L4LjJP8AS0+CJGXwKJNAZ0itgssrQ2erVWNthSglBwgIxMEITYUAeoXFAHqA2akKzRd1wIPUEoQYp3N1Km3QV5cFvqmZQMCxD1AnZTA3NkDwX9aKEDVhofr/vlpfHc/M9xWp4EDj9HfEwgQz6fpFQEqePQK3yqOitsCBWiBJ3+3Adzz/4BH/ANob4ihUwOe/CYA4H7/XACdl8AMSCZozQ8vzIwcVQhdbDWZALH2+XmmgGgrQAFymQLk6PR7OJi1goGo1Zk+7dXSgh9fPGC1IEzQdA3hBChJ/QHojEIAGJvtUAEKlcAaqDa66wJreHRfiwkKkBjtw5oCuoVQgGy5BJZmgqbHV4MEBKjS8MT6HAIvgcUlQasS2BEBRYD6gOAKjR5EQ+xKreb4iVdMvJ203sYVlV+lpRmDjcGg/+GT0V002rwqq03tF7+Twy8HUq5TDdSdUUqA/vdPX8U3dr2OP9m5BxhiY0ReYXlNSQTIZwtELzxAvocioQuJPRgYXX2NKUALNQD//EtUBcvzVn9Qj1MrPP8doDqNQ9oo7vfPAdB5CcxoMkHLHiDWBRajAPHf55gCJOcJHa1w6/QzcQ16Al+JJTA5BRoACswDJBQgkQId7/8B6AgNgJqlW4UZcg/QmmImdDLPod71MNQGU4AKPAcowQNkwY3P62kTcR6g9DZ4uj+1SP9QrDIfZ4JWJbAjCooA9QHRGj3v1pDvi4ITo1obCpDGvClWfji4MTQZvT0lQJ8PSmCa71BjM8A6vAj19RRGcXi+jq/8hOb+HJ6vBz6gfb+kPyMZQM0lsB7kANVmYYC+N7khiQDpPVCAJl4EfvIF+vs7/iTwGLVDgB6lGUp/770FDcI+A3RPgEy44YXZD5ugDT8gQAZbeLkHSPN75IM6guHXWRI0y5kakBWgFVICi64v0S6wtBBEjsGsKUhBqzLYYW6CjihA2QUkQfPvuOgqS1CALHjxSk0bIIQkBCEmlMB8TzQSyB4gIMG/GWuC5iUwW3WBHQFQBKgP4Hk/fIHSNK1lKzwvfXk+gduihGU49IoiNygNHzWCA7atyeiEwGaJxROEKR7cByT8PycDmoa/uP9lYbAsNzx4q45nj8H2M8kDJEpgPfDnVA4DAEokj4F8ELLo6zGLUKf4+c104TrhN4Azfwuw2AnVaUGA5ieA3T8BgYb/2/hVuOxw0zr0LRhRE7S0MHNSw03Q8mgATobKMfcdrfAZKfVMqvoMZleeAiRHbABBi3fUAxSXAcShaRrWD7EyWAsCFKRO2yE1I7eAJGjH86HDRw7suI14gHiyvbkAAlR3fdHI2ZYJWirVV0lYPYv1AcWZoGUCxLONVAmsb1AEqA+ILlAApHlg8eREXkjSfECO58N26QGVHwjaweXBoE6jDTJQnYbJSiY/90+nt3Hfj/D/nIQ9hyv4uwdfD/1pZWBb+LEiBKgkusB4CWzhCpA3T0MQp8hA6MqWMA+Q7y1A/SgxL9QZ76XpcjY7obYyQbNgRjczgv1YAw90kdU7JECmpBhFS2D8ipSTHNOv0/Z83xdjT+ZYG7jeyzEoRyoYASJMARrMySbolUWA+EWVMEE3eBcY8willMAAYO1Ae0ZoPmOsmDFDqmhuAV1gDdcPyl9AogJkah68drtaI5DVqRABEjlAkceVLtRqfqQEFqsApZigzYxUAls5KeVHGhQB6gOiNXogMCwmlcBk0pNGgCbn6xjU6Ik5PygRID14LredVGTWATZFinjG30pvO8QVIEaARk/Gjd9/AY5H8KsnrcEIzzPKR8ZecE8QQ1AC650CVJulatU0BkLeBsJKYNpCPEDc7MwXLCsfvr3F3zk8Nl/nBKgzMmZKxCV6xasxYlfm4y4AushKipfoEFsBChDhZNoKFKBZrCwTdKAw0+9dIaELLK0EBkAoQK0IEFd/CxmzqQus6xwgeRCqbgZ+GgbDDDxAse3qbaDC9s02dKH6AEESdJO3yAn2p+aHu+fiFSBe4pfDSRNKYAswcit0D0WA+gB5ECpHqzDEsAKUvKhMzNUxwBYOXeoCg67DZQpEewSIZgCNkxG8RDazB2edYBO0BLZb24x7njgATQP+v3ecjJECfT0TtqT4ZIeA3HDooUUOEOmdB6jOBqHOYAC2GXytCVuEFjQIlF/V2oXwz1YeoEhmyEiRnpQ7bd01ZAVI8+DInz8jU2LcBUClemnRrWpcATr6PUA6I52aIECWCEIkK0QBmol6gOyIByhlDIaMIAsofb3gylLBNsI5QFr3s8AcT5oDZhep8ipBMxZeAquKSfDh02CiCZqXwMycCGrkiFXuY0tgkgmaq1q+27txQAodQRGgPiBOAeK/J4Uhyt0U9ZRF5VCpjgEwZSI7FLqPT0Z3G22oIUwBmiDDeNFnBOjwy/Qq6PDLAIBbn6Zfnx3nbMLpG4cwwgjduLY6kH+Hw2qQ4wWDGoUC1IMuMGeOlpvKRvg1E9YFtqAUZH5Vy70/7SpAjNjxFvXRIbrgdVoCsyLz21yprZZnBIVMmW5DECCPaIBFydFCQ+OWAzROpjP0sxrMWUIB8iutU9SPBpQiJXbRBdaIKkDpJTBOgFp5gMoJClAWjQUlQYs5YBH/DwAxl3AhJmh+UclnpXGYSW3w/KLCzDRFibTdBRbXBg8oH1CfoAhQHxDrAWLkISkLSCY9aa2lE/N1DGpMmcgMhu7rZCwEzwAaJyM4gNW0ldp3gZd3Ar4DYuXx7dfpwvrRt1LTMy+BTVd9YDUzQif4fwC5BLZwAuSVKQGqmBECJGToxVCAWpXA2DR2j77va4fp33XaBWZGiIvrBK+Fm6ANy0ZNKGq1ILANFmybdYitAA+QwQiQYVPSlzF1VHT6vq8EBYgQIi6wmjxAdZqZE0yCb1EC4wpQiyygsuwBkk3QCwlC9PywAhQFK+mb6N4DVInJAALkLrCoAsQImZltugiNbV4RXWBxw1BtOk+QvzYVhtgXKALUB0STWgG00QXWpgI0WwuunLJhAuRxQ3AbJbDqFCVAk/oqjBazeJlsonc8+88AgNrQcai7dBE9cS09iLkCNF1pSAQovgUekEzQvZB/y7QLrGYNh29nC6W2kJM/JzpRAtSqC4y9rjmXvu/rhujfdVoCMyMKkCd18fHXZZq29H4GQ1HrsGBn6JVolEgdjTDY+A+D+bU0TYNnD9M7k9rg9z0C3P95oLb8zaiVRqCIiC4wRoB8QvPE2glCBKQ06Lnk49Pzg1byQtQErS3ABO3JClAMATJ4G3z3HqC4DCBALoElKEBWtkkBivcAtRiGCkhhiIoA9QOKAPUBQZdGULYYamWCdtozQc+WpmFo7MCNKkDsBOm5rUtgjRkagtjIrsVxo4WgDPbCvwIAxm1KbM7aPCQmy3MP0HS5AZx5KTC4GTjlXeH9kxWgbpKgH7gFuOWC8JgOAFqNEqCGPRK6XShA3XqAfD8gOnyxstrsAmOmyVlGgDaMsBJYJwoQIbC18Pby58db6g3LDs9Wc7n/yEImwxWgo58AWR59z81sIbiRlYKNeqnZbPrM3cBfvxP4yZ8Bz92zRHu5eOAnYsvQxIk9L53gyw1XmJYHWxKgwAOUlLbMy2oAkNfdUKk5uwATtONJXWBxJTCuAGkL8QDRfY8qQFarNnhJAeL+oVQPkKy4y8NQAakTTJXA+gFFgPqAaA4QgJbzwEIeoJQS2PzsFAA2AsLKhe7jYyE8pw0ywAiGV1yPbaMFvMgVICbVvuBtBACcuWm46TVMVxzgtPcAVz8DbP2V0MPGKkCdeICe/hbtQnv1R6GbDXZ172bDBIhfKXY9B0v2+fD2d7tdDxBPjbWQtw0MF7kXp/2TAokJL/RiPEC2ZaPB02ndurjqrBMLmSxTgOAe9d0mFqHvuZmVVIMcJUAa8cJX2g/eDtz1geCqnJVRlzNmK8HFFb8w0XUtlAUUlMBatMEzBajh+onKNPf/mLqGjB8+HmgbfJejMFwfRT4HLK4EJjxAbvzMrjaQqADxEliqB4j+LSeJset23MWXnAQNqDDEPiP9EkCh5/B8ghKToMMeoBYlMFkBSllU+CBU1yrCjnROeLoJeO0NBrUr4wAAY3Ajto4WsIt3gjE8XKZpy2dtDjw3vASWlGUERBQgoVh00AXGW1Fn94ZutupsAGxuVeh23i2ycAKkASxbRpihW3aB0X2twcYxq/LQTbrQGh0oQJ5TbzpIfZkAMVXHsm3aVaeBntBZCGUdNrIZiQh7TnD1eRTC9ikBsnOBApTJFlEnJjKaC/z56cAxF9HOxCfvpBvkRujsuqOgDCE6THPhb00hY6LS8DBfd9sKQgSAjGlgVcHGVLmBsVJNKLwyuP+nkDGhRU7iOdRRW0ASdLEdBQhe16Mw+JyyJg8QV4CixMpp9gCtHcjg9cOV+HU7zQStSmBHBJQCtMSQTcCyAsTLYUkm6LAClEyA6vMzAAASKX8BgMfnYrUqgfk+cg16NZxfvRHbRvNBCYzhp9OUaMQRoOkUAlSK9QB1kAPECcnsvtDN2QYlQDofU8HBc4C6HQPBF3W7QE2LQMcKUJ3Y2LIqD9MMFu12IRueeRefrODxspaVUAJrwEQuF8kIOoqRZQpQJhecNAfzFr7ivRsNs0hD5176XkB+3nYtcP5/or8fBSehkpzKLIGTnUrDw3ybOUBAEIaY1AlWllvqI2WcnNZ9Fxhtg09TgHgOkNd9DhAvgVnREliSCZorQIEHiM9Miy+BxZig5S4wICB38nfvgVuAXX/ZwStR6BZKAVpi8Bp9MWOKWjPQWgGqhTxA8YsKIQRedQYwwpPgOfx2CVB1ShhmB9dsxtbRAsawCnPIYQBVEM3Ey95arCrY2DQcqAsjcgksAaESGPcAdZIELRQgiQB5LjIeXUD0wmh4e5MrQF2e+LnKw30/8u8tFSBKkGqwsWUkD92kC6gBD4QQUaKgr8Gh0+a3/SpwzuXiZlcih3UtA5O48CVJnbfU23YmTCiFAmQhm10hBMj3YIO+N9lCQIAGMhZucn8bxkXX4PdPrQGv/xzY9zD1p535W8DPbqYbHgUEKNoBxsFLYHM1B/ON9trgARqG+PzYHA61IEB525COBw0AQRYNuD6B4/mhta4d1F0fIwmDUAGIEthCcoBqCSUwPpfRaSqBBR4gngPECWKpRoey8r+lD8RzgBKGoQLNJbDSAeD7f0h/P/O3gOLaLl6ZQrtQBGiJEef/kf8/F3cgIUx6kurqpaqLnFcGDMCQB6Ey8MGgLQkQ8/9MkkGsXzWIrasLADS85G/CufrLKOW3wK2aOHPTUOgkLkzQbZTADF2ThqF2oAA1YhSg2gx00MXKKkZLYGwAadcKUKQFXv69zSBEWgLLwTTpPhps0eZXmgCAA48BT9wB7H0wRIA8ltnkEh2ebgEk3AXGS2AZW/YA1QICRGzkM1KK7tE8EFVS5LK54KQ5yMpBs3UCbDyH/pMhrsKXfxeYiNiIrC88DfpQqS5sYO0oQLwVfmw2/hgNp0AzAplfDVQm2RwvgprjdUyAHLkLLLUNvgceoCQTdGIQYkao8GsGgmOrVHXCZUJe5gqZoKU2eKC5BHbg8WDbsaeAE369/Rek0DFUCWyJMROTAQSECVEpxlDXjgI0MV/DAJONjUgIIhAMBiVtEqBDZAQbh7LIWgY2DmVFGWyPQX/K5S/5Nc1WncS6PF+g1w1kpJJNmwoQIeESGF/J2SDUGVJAMZ8N/Qn3AHU9BkIQIGkR7jAIsUpsHLM6D8PkV61+8/vDF8nIY3LDswMzmGzvNZfA7ExGStYOghDrsFDMWqgTdrI7ihUgwsixTzTkCxIByvKLi4TvAC8XHw0KEDu+BiMEiJfAeCnLMjRkzNbL/9oWYYi8CyxUAmOqhan5sOB1ZYR2XCkHKDYIkb4+W/PgtRgOnYRWOUDNbfBsf6wgCbqYMWkCNmJa4eNM0FECFO0CO/h4sO340+2/GIWuoAjQEmM2JgMIoFcdfJGKy5Sot+EBCqdAN5fASJsEqDpF1ZVxMowNrMS1dbSA7/lvhGvkcK97PgDgzE1hAsQ9QITEkzggIEAbhnOdT4N36RUlACoZ81yXCu18o4NQw++rxkpgXY+BiFWA2muDJxETtME8QEZ0oCkQtA9H3guu9rgwRBdfqATG/ESZTCbiAQoI0EDWhIOjnwA1avQkUoUtsm+AQOkoVRO+A3E+jGWKuJBVICiB8bleA1krXIJNAFeAkktg3ARtBGWcwhpxf7dhiK0VoODz9br8ToskaLkEtv9RXHTw69DhN3uLhAk6I9Zj29CDENuo8p2aBM3uEyWwOAVIEaDFhiJAS4xAom7uqBhKmQhfa6MLbGK+jgGND+6M8QBxAtQiE6d8mBKgaWOVIGVbRwv4of8G/Mk59+G2mQsAAGdtHg79nWXoGGDbTyWUwfjrXz+U7bwNPqq48DIYV4BQbMo26V0JTPYAtReE6NQDD9DmEdkEHaMA8Wn1EQLE57Y5MAMCJKVam8wDlM1mYoMQGzBRyMgE6OhNg66VOQHKhMYbcDWklKgALV8C1HB93PXwXtzzxAE8P1bC5Bz93IWi7NYB3xPH8UGW6tyqA4yDhyG2MkEXbEkByo0AGiUVmS7HYTTkJOiUIESgzViPGMQqQHf/V/yb1/8Sv64/2nyMlmg4LIrrxEVoxtKTM9y4yuM7NE8MCCdBA4Ad+e4dfCL4+z4pQNPlBv7j//4F/vHRfa03XuZQHqAlhjApRq7QAHrVtn+mGqsAtTMMNW0OGBBMRictrpjqUzQEsZYJDHjHjdKT/nefPgTPJ1gzkBGLY+g1FCzM1d3EVvhZdhW+cSiLVzotgUXN0rP7gPVnCgI0RQawLqIA6UwB6joEUO4C42hTAXJqFdgANIuWEQkzblIFKLK4coIWCYXkap0LM7jqlT4/g6VEZ+0M5uM8QLAxukIIUL1KTyI1ZEIeOl4C4/ETTVjGHqB/ffogPvmtJ5tuH85b9Pv5F+cCq09AfvUXAABjs1wBapcApQ9EjZ0DlinSMnFjrus06IZLgiDEFA8QEFZEO0EwDJURoNJBYOI5AMBx2kEcjh6j06/RnyNbBQGyDSM5w00iafAdQM+Eh6EC4RLY3BgwLwW8Tr5ICazZvM4uJn768iR+/vJh1Bwfv3nu5tZ/sIyhFKAlRpDT0UyAhlIGosplr6QSGFWA4sdgAAD0mK6EGBA2Cd4trBO3USM0cIAtoGdFDNAcohW+HL8o8dLYhqFc5yboOAIECAI0TQaaFnZdlMAWmAMkL8L8d49eXSeBl8CIQU8iXI0yNR+uF/k7XgIjXqAGQfIAaSaImOUmKUCsBJbNZCUPkDQKg/ASGFvkj+ISWL1CT8B1LXzC4N+JuYSy7HJWgCaY4jOQNUPf/dM2DNET9txBYO9DKGaiJbD2CND6IfrdPVyuN5dtAcxLOUDBxcKACGHNoS6IRicIJUHHtsFLBKjLOX9VMQyVHRu7fyzuO0Y71GyCnnmd/hw+VniAMqae3MFrSN9DvsYltcE35oPy15pTgOwwvSiaeL6bl7YgTJdZiGpK4O7RAkWAlhizMZPgOZKuJHyfiAMOSDFBz0kKUEwJrN3BoEb5EABAH9wgbts6Wghtc+bmZoUJaJ0FxAnQxuFs523wTSUwGoboszlgU4gjQMx300MP0Fd2HWy+P3Z/6QJOzCzfGXGXG5Xt5RKdpAJxAuTBECZoWcXhoYr5nOQB8uqhURjFjAWHmKHHOxrRqDICpIeN8EEJLEkBYseKUwmRz6XEvukKnjvYuQLF82jecfp6PHn927Hrmrdh1zVvw8nrBwJC59VRtOnFCo+oKGZat8ADwKq8DcvQQAhwaK75QqUiTNBGUALLFCUC1EAtJbcs8XW5PmyNj42IUUB0Az47ffldknpOzEQbvJQuf4w2Hm6D9xxglirjIQXI1IMMtyYCZIX/HpBKYDFBiNwAveEcqmwDffEB8bW70cXnttygCNASI6kLDEDigRQdvJfUVXForoZB7gGKU4D4AdmiIypXowQoM7JJ3HbMqjzkzvxoBxhHkAXUvCh5PsEck8xDCpDvpCopAgkKkDtPQxunY0zQhrnAQaCRHKDxUg033LcbHmFvRlonGCvtEZ4gHTJuRj4D+cQrKWKEkVUPgQIkf37cA5STPEDEqYm0b9oFFpTA3EYHkQPLDC4zQbtaEgFqoQABfZvKfflXf4H33PpzTJU7O5k3JC+KpmnYMJTDhiH2fZNey4AZPr5azQHj0HUNawd4GazZBzQfVwKzi+J4yWndK0AZpBAg0MYAAPCdLgmQ3AZPSIQARRSg2X1UnTUyQHEdGtwELStA0RBb3RBeKKH8NJXApA5E7v/ZeA6w7gz6ex98QPz8kxa4e7RAEaAlRpAD1GyCTjqQojX0xBLYXB0DYoJyHAFqYyyE72PApYpKcU1AgGxTx6aRIPTwjE3xBGhYKEDNzyF3hm0YygaKBdDeQNQEE7Q3T/d3Th+EHWnt1XvWBUav1J45MAtAQwXZ8P0x0Nhr8nugALmaKbr4hOfB96Czwbf5bGAq95wavAb9HlAFSCJAXZ4slgOcGv0sHCM8A4+rgg3Xj/ejmDbAP6M+lMF8n2DvVBUN18eL4509v+xFab4zeKxBM/y5t1sCA4KZYOOzzcdo2ATNns8uCAUoiy49QJ4Pm/nbBFmIgDcFtGrqSEJoFtjEC7RcqNH1Y5M2Cd+VjklR/joG0HVxUZoxdaHmx1kXgk4wXgJLGoY6F5TANpwNrGcEaOyprl7bQsBJuFKAFHqOt52yFu85ZyOOXZ1vuo93ZvBFhSNKeOoJC8qhuToGNXZCjjFBx0azR1GZhAEfPtGwam3YALdtlB6sG4ay4qowirR5YLy0V7ANDOasoGsJaK8TjCtA3ADJCBBhHqC6Pdz0J5wAmehukYyWwJ7eT8sUVbCFLUUB0hmR0aw2FCCZlEoEiJugPc0UJna+iMpm9nwuiwbzAPmNKnz2fjaIhYJtCA8Q7yo7GuHV6WflRghQ0TbB7WpJ8Qz99AHJXVJ7DrfIloqgIZVimiATID18PBY7IEDrB5MVoHLIA8SOlUzYA9SdCdoXqd5JCpC3UAIkd4Fx9Wfrv4Gn2zA1H4POeLCxZIAGgk7csALUxkDUaBI0L4FVp4G5AwA0YP1ZYQVoiQcYc/VeKUAKPcdVbzsRX3rfG3DqhmaFhteiKxHJuB0FqO56mKk4qR6gQAFKVkMIa/U8jEFsXBV+jG2MtEXzf2SsKrASWIwJmhOgoZxFQ9h0Ew5hV66dKECrjqM/5w4CngOtSglQwx5p+hPT4l1gXRr6IgSIKkBAhbBFOaUTTPeC4DR6Q3CV7jmRzyBuYjQCw7OLQAHi5TJHUpFymSwabNSJ59YDAqTZMA1dmiN29BIgn31Wnhkm57quYRsz8f/LUweb/g5AXwmQfMHz+lSLdPEI0glQMJuroIePx3bGYHCsE2GIzd8dHoQYygGywx6gbtrgPdeBqbF1zoy/2PJELtbCSmB5mQAd/zZU8vTCb3Vjf7DxNFOARo4FAEkBMiTrQtw8sEgaND/O+e2i/MpIzuiJVBVacwotn1Wng/b7JQIvgSkFSGFJUcgkEaCIAhRjgp6cbwAgQXhYjAeIdyGlzcUqTfAQxBGx8HFcev4WnLJ+AB+4eGvi3/MSWFwO0KyUUqtpGgq2IU2E74AADR9DyRzxgbmDMGt0EKqbXdX0J7o0gNTvZmZQpA3+mQMRBSgyATv03IwA6ZwAaRpc8Jj98AlptiK9flkBYtK5LylAhKlFcjnLtm24GguidGogjAB5rPOPD8I9qgkQy13yzWZ19T+9eRsA4PafvBq/sPeTAEnH++sdKkB8LYhNdZZeSzMBal8BWpeqAMUkQcsmaK3eVRJ0SKVOKIF5zF9Duu0C423wug+89jN643FvRaW4BQCw2pHIcoIClGmpAEU6b3kpzIwoQBwbzqY/rSwwehL9fYl9QIECpLrAFJYQOZtPbI6WwForQBNzdRRQg8E8IXEKUDtjIeYmeAji6qaryjM2DeHeT/wqLj5hNO5PAbRXAuPt/oWMKbXCt18Cc808MMj8SdOvwXIoKSG5ZgJk2pSoWHDhdDM1WrTBFzBbcbBvmu5DpVUJjBAYbLEzpBBFn5WiorL9zJz0ONLiz7fzdFPI6Rq7zW0E21mWDY9dVcoEyGe3udrRrwDxYEphOpfwW+dtxtqBDA7O1nD3Y/ub7iecANVmF3UX4yArQHumuiuBxRIgyQRd0MKfe7tBiACwfoh5gGJN0FxFibbB0+98tksFCHK3YkIJjHuAusm28nwi1tGBw0/S9yo3Aqw/C7XiMQCANTIBklrgAYQ9QPkUD5BcAiOkuQRmmID8fd1wTvC78AE15zwtJrgC5Hiku4vGZQRFgI4g8Ej2aNdEkwIUc0V1qFQLyl+6GZRdJGhiMnrygsHHYFQyaxK3ScNwykT42cicorxtBEbotjxA9PXd8+w0qnnWos9Mgj7RoMcMgDXYa7a6HZoolcCeORicHFuWwDwHOujnpGcCJc1NMCP7XrwJWnSBaVbgIfJ5CSwYlGoYOghTe4hbE8oRV4D8mAyhow7cI2Y1K0BZy8CH/g0tnf6vH78iUn4JIfjj7z6Hn7zOTkx9UIDm5RJYpx4grz0PUA7h79tgJyWwgeR5YGEFiD1fqA2+Ow8QPwYItJB3TgZXNbvxAMn7lN/H1J9tbwF0HXWmAK11ZQWIl8C2wvWCJHfb1EWq/0zVAYn6dWQTtO9BlLpkVUvuQpQH9XIf0BK2wjdcP/R9jHYgH21QBOgIAg/kipbAogpQLUaaDIUgZgaBmJBCMRcrxQPkzbIQxNzaxG3SsKoQKEDRxSBOAQrC+9pXgOb9DMbAVCjWOjqDAorZZq+AaTEFSIuZvyVj38PAC/c23y75Gp49EOS0BCboBM+GlG6tSydkj3WZkEjeTCidO9QGT2/3NTMwf/MSmEiJNqBpmlB74NTpPwC+wUtgfIzG0dsFpjGCrNnNBAgAfufCYzCUs7B7sox/ffogCCH4H99+Frf/5FUcdtl71wcCJCu+s1UnXklIgFyKab4zKM9mte5N0OuG+DywyJw6nwh1p2Dr4XIxV4C6TILmKigx7Ni1DFhYF5i8xpqvswDE494KAGgMUgVorcdSmevzQIVGbWDk2BApkE3Qnk9C5AFAuPHEk96/EAGSymA8/0f+fQlLYNEO5KPdCK0I0BGEXAIB4goQz+GJV4DSB6ECgR8mLRVZL7ODfnB92/stg5fAHK95MShFCFBIAWpnHAY7wVVhYwxMoTpI5eG4FGgAYgCpCbd5/AQHIcAdvw1883Jg/lD4Pq7wWHnh/wGkEliSAsTImk802JICxEtgUSUmUQHyuQIkl8DotlxF8nhZzZCuNpn/iN/m8Rb6dlO3lyF09h3S5LElEgoZE1f+ylYAwK0/fAV//N3n8H8eeA0AME+YYtoXBSh8vHdihG5XAcqS8AVGNx6g+bobOqZl4lYw/SDKIWqC7iIHSPPp95QY8eUvACIYdCEK0GrLgbbvl/RGRoBcRoA2eEwB4uWv7DCQHQp5yGxDR9YyxPufmAbtOWFfk1zW4z6gVceHu3c5ATr8Snrgag8R3f+j3QekCNARhEIm3gPED1ZeOor1AM3XgxDEuA4wADo7GI2UEphVox1V9lB3BChnG+JqNHowNSlAtuQB6qANvoYMXvNYx9fkCwCAaTSHIAKAxhYaG26yAlSdpuM0iN/ccSGXwFgHWDFjBiWwJA+QNAk+a0nt71o8AUrqAuPKEFWA6ONoJGyCdrUwAdLcOjSXn0Cy7O85ATp6FSCDESA9E0+AAOADF29F3jbw3MESvvrT3QBoqOcc+keAKpELhU7KYO3mABluNaQSddIFVsyYKLCLM3kqPG+BN3QNGV/aZ4kAZVHvKglac7kClEyA4oJB2wVXrn7Fep4St+FjgVXUKO8ObQUADKBM14aoAZq9HkPXYBr0PR1uNRDVrUu+pkhZj6/XcvkLAIprgcJaAAQ49FzHr7EbTEeCOI/2TjBFgI4gJLXB8wNuSBCgZlbeahAqAOhW61BAHt43OBBPotpB0jgMnsQrFKCMiRofh9FBCaxKbLxUG6a3saGfSQoQX2hMeMkeoJk9we+1meB33xPKVE3P4ZUJSobOPXakdRAiez01WMjZwWEWmKCjJbBg4eQhhvRx6Ha+bkldfPQ2uUUeAIjkN9CY3M5v84UCdPR6gAymepnZ+BIYQLsUf/dNx4r/f+49p+N333Qs5gn7mz57gIDOjNCpbfByqrVTDRmfixkTePLvge9c3XKoLwCsKrLuTukEKVKgbQMaL3+ZOWrsFUnQ3SlAOu9UNe3EbRZiguZr7Duxi95wwiXBc2fyOESG6X+mX2tugY8xnifPA5NM0HIKtFzW4+s17wCTscSBiFHvpiqBKSwZuAeo7gYmO0BSgLLpClDIAxQDo43J6CaTnjO55KvoVhgpxKdBNytARhCG2EEbfBUZPDU/ELprKokAGYEJOtHQx4eqAvSKj0MiNy9M0c9kdcHGttX51l1gnKwhE0ybBuBzBaiJAEkDUGUCxE4Evm5KBIiVxcScMPa6ucfLCxQgCALESNJRrABZPn3fTNlTEYOPvuV4/Nsz1+PPfuss/N5FW7FmICMpQEs/Eb4cLYEdbr/cUU/rAquHCVA+E3wPixkTuP/zwMN/Bfzi1pbPs6pAv0eHJQKU2AIPiOyeXJdJ0BovF7VRAuuGAFUbHvKo4a0+I0BnXSbuM3UNewjzQE6/FqMABWMwOLgRukkBkk3QIgU68pouvgo4633AOb/bvKNLPBIj2r2rFCCFJUPeDk7gcutoVAFquH6TwXhC7gJL8AAFBCh5wbAJPXFameSr6FYQ88AicmqUAOVts6McICJ5gJ6OEKBpDIgSYQiMNNia1zyBXexYAgHi5EbT8fQ43b/TNg4iZ5uoEq4AJRAgrgARG1mzmQD5kUWbSDK+25BzgFjqsxYoQNzE7gqDNDuM2UlH8+rSFTS9jYcokqN4Grzt0/fNyqaT95GCjb/8j+fh0vNpt8+aYqavHqBKw8Vb9CdwZeHnADorgTViTsYCkgkaToWOqwAlLIauBWTvZ18C5idSn2d1oVkBCkIQzaa8LKEAddEFRgiB4YcJfOx2uqSudIia4+Gd+kPIoU69N1suEPeZuh4mQJEW+KDsGLznQ0nzwOQcoGgLPMexFwO/+RWgsLp5R1lZbqnCEKMXrYoALRC33nortm7dimw2iwsvvBAPPfRQ6vZ33XUXTjnlFGSzWZx55pn47ne/G7qfEILrrrsOGzZsQC6XwyWXXIKXXnqp6XH+5V/+BRdeeCFyuRxGRkawY8eOXr6sRUHW0oUyKvuA+AIyJJ3goyrQ4XIDA1qrEljrwaAWI0dpZYRWSCqBRdvgCxlJAWrDA0SYMlIjGVSRhZcNkp+nSTFBAQreMyc6f0vs2N7g9zgFyC7imYP0xHj6xiHkbUNSgBKu1h0+i8tG1pYJUIJxU1KAfHnoK9vO160gxoB9fp4jGaQBQXZ0rwGdye16pAR2NCtAnADZ+YEWW4YxWsxgnilApE8lsJutW3G9dys2YaKzElibJmg4VeExFKUw/v1uzAE/+ULq86yKI0ChMRhSBhAQeIC0znOAXJ/AajEHDEAwGialqxWgis1zB0uhi8ZKw8N7jZ/Q/5x9eagkZRmSAjS1O9QCTx8vGEDLwT1AbZfA2oXFCGV0EPQiIaoAqRLYAnDnnXfi6quvxvXXX49HH30UZ599NrZv345Dhw7Fbv/AAw/g8ssvxwc/+EE89thj2LFjB3bs2IGnnw7kvy984Qv48pe/jNtuuw0PPvggCoUCtm/fjlotOIH+wz/8A37v934PV155JZ544gn8/Oc/x+/8zu8s5kvtCTRNE1lAlXqzAiQrHHInmOv5qLt++hgMyGMhkheMDFgJLNOcI9QukrKAeHtvSAESHqDWB7jPZj1VmWpUyW0Q901hID7bRA9uc5NO/iEFaCb4Xbqq5R1gZ2waDBOgVgoQLGSlk5NQgNzIZyApQJ6kAPHQQ6Kb0Ay64HMPlydNigcAjREgw6/DYAqQZnEFiF+JHr0eoCzo+5ZpoQBFsbpoCwLkVZc+CLFWr2NEo9+1rfoYxkq1tlWTxDZ4N9J27VQEARrIms1dSQ9/jXYbJSCeAEljMKIlsAUoQI4XzAHTrLQSWDgWIgl//C/P4Z1f+il+9EKgcmmze3Gx8Sz9z9mXhbY3DR17fK4A7Q4UIEaAGjEKkAhDbDJBS6MwooNQ2wE7ftuyCPQA0YtWpQAtADfddBM+9KEP4corr8Rpp52G2267Dfl8Hl/72tdit//Sl76Ed7zjHfjkJz+JU089FZ/73Odw7rnn4pZbbgFA1Z+bb74Z1157Ld7znvfgrLPOwje+8Q0cOHAAd999NwDAdV18/OMfx5/92Z/hIx/5CE466SScdtpp+O3f/u3FfKk9Q5AGLREgtoAUM0bQCi8ZoSvsfuEBSiqBWXwwqBvyGAkQggyhB0CrMkIa5CwgDt8nmGML5lCcAtRGezaRPEAAMGOtE/clmqClqy2vkfAcLTxAxMrj+YOUAJ2+cQg520C13S4wYsd6gEhk0SZSec6X05p9ToAs6CLJmxmjGQHij8kVPsNvCJKkRSbRH80lMP7dzRQ6U4AsQ4fGwuhIbek9QG41UGpOsA6DEGDfdHsqkJxIHL4jMqLFqaDIPEADWTNs3j/urVRF2fk/Ep8njgDNy5Pg5TlgQKgNvtNRGI5LkEHk+xsDrgBpLUj9q5P0tT43Fny2G17/ZwDA89lz6GgdCSEP0P7H2DGuAUN0Rlhggg6O6+GkBHw5ByipBJYGniGWMnS5l2g2Qas2+K7QaDTwyCOP4JJLJHe9ruOSSy7Brl27Yv9m165doe0BYPv27WL73bt3Y2xsLLTN0NAQLrzwQrHNo48+iv3790PXdbzhDW/Ahg0b8M53vjOkIsWhXq+jVCqF/vUD3AhddZpLYFnLEAedLE1ytWioRRu8CAVMaAknXkOM0liICVrMA5MWy7maK4YaD+bowhXyALUj8bJteAv6uB6kVVMTdLIHCAhMw01IJED0/azrOdRdH8WMiWNX5SMKUMIsMKkNPieVwIgY4JisAMklMLG4GxY0kePEFaBgUjwQf7LQbaYAGUe3AuS4LvJs3EOuwxIYANgFWjbWU2a7LRbksttpOfr9a9cH1Ehqg4+auZ2q8BgWs1ZAgAwb2H4DoOnAs/9MA0FjwAmQbILmZfpYE7ToAqt3XAKrex5sjSlAKR4gru5qKYo23U/6/Ifn2b4Tgm377wEAPDryzqbtLUPHHsIurngn3eAm4Ueqx3TeDSWVwIQJuhEobinG7ibwY7qdmJAeQJmge4TJyUl4nod169aFbl+3bh3GxsZi/2ZsbCx1e/4zbZtXX30VAPCZz3wG1157Lb7zne9gZGQEb33rWzE1NZW4vzfccAOGhobEvy1btnTwanuHuDRofsBlLUPUnWVmzs2IQ3orBSjIxInriKrXgkU3k+2+BDYS0xLKpeGspQsSFx6G2kZAH7sK4n+z1w1mf9EcoBgFSNPgstZzN24OllsH5qXvY0wJbN6n79upGwag6xpylikNQ21VAos3QUfb4OXxJCSkAAVt8LoRJkCBAkRft2E3EyD+mYtS4FGqAFUqgaKR61ABAoBsgfrJDLfMRhYsHXzJrHycSROHOyZAUQWoHqcASSUwKd8K604DzmEWgR/+z9jnCUzQwXdzPuQBmgseD5BygDrvAnM8AlsoQGkm6PY8QHwtnZxn+77vlxiu7kGFZPDi6l9r2t7QNUxgCDUiXVCNBNEJaW3w09EEfNkEzY+9FiWw//3TV/GVH79CH0coQEvjAeIKkMlKDcoDtMzgs4GXf/iHf4j3vve9OO+88/DXf/3X0DQNd911V+LfXXPNNZidnRX/9u7dm7jtYiIuDZovIBlTFwedLCvzWny7CpAJD07MF7smnUSyPWmDD0620Q4wgOYABaMwWh/gmhu0lgPAy/VhcV/FHIJlxH+duUcmNgMn2l0h5wCxk8S0S1/P6RupSpC3jQ6DEIP9IqILLKoABZ83ker9optLt6BHYgx8YZBmBMjMwCeBmdMlOky+2LYxB245o1YOFA87m94GH4f8YGCoT1T1FgmapABtJNQf2Y4R2vMJXFbKbiqBRc3cUg7QYDamZPWm/0p/7nsYiM6zQqAATZeD7w9fd/IZQyJUERM0G4baNCMrBY7rI8Nnl6WaoHlXZPp3mitVggA9fgcA4F/9N8LINpNly9BAIHWCAcL/A6S3wf/ytWkc9+nv4vTr7sWv3fgjTFTYOus5bZXAZqsOPv8vz+GGf30eP31pUvIALa0Jeu0AXd+UAtQlRkdHYRgGxsfHQ7ePj49j/fr4lOH169enbs9/pm2zYQM1xp522mni/kwmg+OOOw579uxBEjKZDAYHB0P/+gFRAotRgDIJJTDejTEArgAldIGxEoqlxY+FaNToIlYjFizTaLq/XQQT4ZsVIJkAFTochqpLqgoAPFOhn5FHNJBM/GsGgknofpwCJJe/gNg2+FmX7vOJ6+jJImcb7QchEivkAUq6ag0RE4kA8dBDopvSKBP6efuiRZ4+vm0ZaCBQweqwhGKoLaBleDmgWqYn9BosQO98WRsZHECdsPduiTvBDCcgXKvYBPJ2soBCIxna8AD92zM34IKtq/Af3rA5+N5yhWHV8QBYa3x5sum5VoscoOAYEjlAdkoJDHUQQjoaqtnwfKEApbXBo00PEF8fD883qDn8mX8EAPyD96tivZXB051DBGg4XQE6c9MQNg2zTkIClBsedk+W8coU2zfZBJ1CgCbmgmP/T+99Hr7BlPglUIAIIWLN5vPflAeoS9i2jfPOOw87d+4Ut/m+j507d+Kiiy6K/ZuLLrootD0A3HfffWL7bdu2Yf369aFtSqUSHnzwQbHNeeedh0wmgxdeeEFs4zgOXnvtNRx77LE40sHr9OWYNviMqQs1oS7JyqIWD7aoJShAoVDAGGZfr9GDrIF0ibYVeAlM9gDFKkC22X4QoueIK73RkWEAwMPltXDsYTxJjkcxl7zPnpiEHlP+4QSIL3AhDxArgTG1h3eZ5SxDGoYaf6VOpLEdIQKUkF6ryWUXt7kEBsMOBVkSQppKYBlTD95PUAIkvCEmD1E8OktgdWYkriHZNJuG0YGgE2ypCZAuEaBcYwpZ1PF6GwpQKgHiHiCeEeVUcdrGQfz9Ry7CBdtWhUtgAFUamMkXU682PddIgX5/ao4v1ptyI64NPmyCNjRazqo1OiBAbtAFlmoY5sGgLTxAVVkB2v8wUJvFvDGMXf5poWOTg5d/9iYqQM0m6KG8hZ/9f7+GZz+7Hb/8w0twGcuYqrPye2gYaioBCo7PZw6UcN8r7HN0KrHKXC8xV3eForhugBMgpQB1jauvvhpf/epX8fWvfx3PPfccPvrRj6JcLuPKK68EAFxxxRW45pprxPYf//jHce+99+KLX/winn/+eXzmM5/Bww8/jKuuugoAbRP/xCc+gc9//vO455578NRTT+GKK67Axo0bRc7P4OAgPvKRj+D666/H97//fbzwwgv46Ec/CgC49NJLF/Pl9gRxChAvdyWZoOlCRJAn6R4gfuAleYAc5gGqawsjQNwEXXU8Qd5iFaCM0X4QonQFNDg4hIGMiXnk8fULvo3fblyXOtsoIEAxV4qcAPHBg04lICDsJDHHCBAvIYRKYG4t1jPiNQK/UrgExruxIgoQiVeAeAlMMwIFyIQHzycgbtAhBtCTYF0ir3XYQVZJpIPsaINTpSfgutaBwVTCaJ/CEAkhMN0w2dmsTWDfVDW+U1MCvzrXteCkHdzJCEmBNQpEvWrR4EIAWHUc/RlDgIoZU7R9czNxOAmavWeZMAEC6DywTozQjufD1rgClEJodV7WTf5O+z4RXbJT5Qb8V34EAHgh/wYQ6LEKkBWnAMV4gKKkU9M05G0TawYyWM1GhzS4qhjyACV/R0WZjuHLP2ZWDOIvuno7Uw58mrxRpR0CNDZbg9/iu3qkYlEJ0GWXXYYbb7wR1113Hc455xw8/vjjuPfee4WJec+ePTh48KDY/uKLL8Ydd9yB22+/HWeffTa+9a1v4e6778YZZ5whtvnUpz6Fj33sY/jwhz+MN77xjZifn8e9996LbDY4UP7sz/4M73vf+/B7v/d7eOMb34jXX38d999/P0ZGpDr/EYp4E3SzByjUBl93UUANOtiXNVEBCk6gcV1gDUaAGgskQINZljSLoAzG54ANRhUg5gEirUpg0nR108pi0whdYJ+Y8ODApL6GBHACFBsCyEMQ150eXC1zIzQjQCWPLlgFQYBMlGWlIUYF8urxbfBE523w0RJY8HlqUn4Lv7olui1Ntvfg+kS00nMPUMbUA08VgDqxxElLM/jJ4uhUgPwG/+52R4DWDARhiFjCVvi66yNHwuWNbcYEGp6PsVL6MSF3I2lalAAxQlJkDSPR76gU8ikgCFBzHpCmaU2t8PMhD1AkCNGwRImq03EY1ATdRmaOyMVKJgY11xPCiU8AjxGgp+1z6L7FKECGrkHTgNd5JxgQLoElRQ9IEBeqMgFyW/uaDjMC9NaT12C0mMGLU9I6vcit8NyzOZK3xf638gD95MUJvOmGnfjC915I3e5IRfJZo0e46qqrhIITxY9+9KOm2y699NJUpUbTNHz2s5/FZz/72cRtLMvCjTfeiBtvvLHj/e03clZzDlBIARJdYGEFSIQg6mbo6isEphQktcF7LGiw25MIh6ZpGMlbmJxvYLrSwPqhbEsFyHeqSHUdSWMwcraJLasMPD82h2fZhPbYEEQGL20SuiiBHQNkh4HqFC2DDayTCJAt9hegHqA6LPhEg64RenWdCZspfZZa3dDskDk7kQBJMr4mlcB42U8zrYAAaR4anh+kSTMPULQE1oAZeICs1lfLyxmccDp6lwRISoNeynlg5bqLIsIE6Iz8LO4rUR8Q95XEQaRAx5n/BQFiKkbUQxItgQGpChBAjdBjpRqmKmEFiJbAYh7PygP1Uset8LQExku/KZ9nZDhwHOQ5a3nUYB58BADwmEkHj+ZiFCCAKmqv+cyrahUCIonAfhCbvs3Aj7uaLytArUtgk0xdO2ZVHr9+ylr80T8/DQ86DPiLHobICdBw3havrZUC9MTeGQDAMweWPkC0FzjqusCWO4ISWHBQc7UnawYt5HISdKXuhgehRq8GOXgJTPPQiFmQHHbSdvWFKUBAUAbj88DiCFDOCkzQvqwAvfZz+i+0c0EIYs42sJkpQLtZyFlsCzwD98jEdoFxAjS0GcgN09+5D4gt6jPMBC2XwAAtdRwGVyQ8IyLhcw9QZNGWS1O6pADxlnfolvAAWXDheiSYEyaVwBpNHiB6iBuiBHZ0mqB5dpKzgBLYHCuB+bWlK4FVGh6KWpicnJQ5DADY06IVPijFxJzEuSLDT9xuFfClk1kcYVl9PP2ZQIB4WWcqtgQWMUEDkTDEzkpgogsspVzEjf16igdIthJcoD9PydLwsXidlbfiFCCAzgN7lWzEzEWfBv79l0PG+npbChAjQIQ9vhuUwCq+gbf/+Y/xtZ/tbvo7XgIbLWbwvguOwbGrC6jytPxFVoC4Wj+St8T+t1KADs3R/S1FE7CXCRQBOsLAJzaXYxQg2gUWlwPktRyECiAyF6u5I4pPIXcXqAAB0kBUdlDFESBN08SoBnGFWp0G/uY/AH/3W2EzMJ+uTqipeMsI7TLhped2CFBTCjIhEgHaAuRGgn0AAgIUKYFZhg7L0FKzgPgJ2Y9cwYousMhgVnkRl8tUBicsRtAGb8KD6/nCR0RECcyImKBtZNgCz0MU0wbhLmdw07nbpQIkj8Oozc90txNuHXjpvuRsqBjMs/I1APH9O1anIxtaGaHTJ8EzFYt7gICwghA1LQOBAnT41dRWeF4CEyboUBK0pIQKAtShAuT5yHAPUBsmaCOF1MvNJL+iszDc494iiFGiAmTQi8jD5/4+cOZvhe7jF5+pChAjpTVfMkEzBfpQheDF8Xnc9ci+pr+TCZBl6Pjgm7dJYbFLowCNhBSg9M/tEOtam1EESKEX4LPA4trgs5Yu/CShJOiGi8EWGUAAQgTIjVFDeBnB7SSpNAHRgailGAIEADBZ6yg/uPc9QqVipxLbkl5lpmKuAHGkmaATB4FWpwP1ZnBjQIB4FhC7rxIxQQP0yjE1C0gQoIgCxEpg0flFMgEyYhQg3bCFp8ICLYEFbbX0djvaBSZ5gHQ+KPVIUIAICQ1/7QX8xsIIkGXocAyqhlTmpltsnYCHbqfE/Wc3tf0nlYYbKEDrqNdxrUdjPtpVgOIJEFeAJCOvXAbj31lZAeKdTvVZoNIcGsuPaZ4GHZ4FFjFBA6IVPqt1pgCFusDSFCCjtQJUCRGgZ+gv294iCFmcCRoIjNBxRvSg9JhctBcKUEwJjEdVjMd4vCaYujbK1LbNI7nO0vIXAH6xOpy32vYATTAFqGkG2jKBIkBHGPJiFphUAhNt8IYUhCgpQHVZAUrOw5GvptyYuVhCtejyJCJjDQvS+spPXsEvXj0cTIKPEBWeXiyMv/t+GdwppzJzBQgZ5CwDm0fC0+pTFSCdj5+IHKTcAF1Yg1/ur+JAnZGViAJUQQaGroVONHnblMZhxGS2sKttP9rFosf7FjQSfJ66pACJxd20QiZ21yNBGY15nJpM0FIOkG62PlksCXwf+MqvAre/pbdXtAtUgADAs+nFQ73cpZ9hN5suPtG+IXS+7gUeoHWnAwCG6rQxpJUJOjEFGggISXZIGqcgfU9jPTs5OvIBiC2DyWnQvk+ET7FgG/GKkqQAdTIPzGkzB6g9AkT3cRVKOE1nQ023BQpQXBs8EHTVxXklxQBaq7UHqCorQGz9cRDEhESJ4SQjFKuL9HWvLmRazx1kePnQPD70jYeFL6dTzMQqQO2VwGarzrLsBFME6AhDbBI09wBZchdYWAEKeYCSoAckwYspgfEyQpNvpQv8pzdvw8ahLPZOVfG+23+B59gw0aF8mADpNl0kNS7P73souFNOZZbGYOQsA5tXdaAAaTxrKKIAsfIXGdqMD3/jYfzgNfaeNBGgLAq2Eeq0ydtSFlAcAWLvJTEjJlY+lDTSOm9Ii7jpB5+N4XMFyBLkyWQmdhKjAEU9QBnuAeKDcJeKAPkJC2d1Chh7Ehh/Gnjyzp49HU/PXtB3lxnZ3W4mwhMCwuZoeXPxo37iUJZLYGtpeKvllDCIcsvyg9wd2gROSDIDQVOErCDEtcEDqZ1gq4pBCUwuLRVNPyDjURM0qAdIVrRbQZ4G344J2kj5TnMT9EX6swCAg5njgOIaoQAleYC4AuTGBcammc8ZuIJS9dg20jDUBgme81ApONYJIaIEtoYRoNGBDA33RLA+J+GeJw7gvmfH8aWdL6Vul4SwAtTaA0QIEQSIEIhh18sJigAdYQiGodID1POJSG3OmIbwdCR2gaUpQJomrj7SCBDpQQns+DVFfO+//Souv4BOWuavIVoCM2y6SOpenZ409z0S3BmnABFqgh7MWqHHSlOAiCAdUQWIEiC3uAnTFQfThF29RghQmWRD5S+ApUGnXJkJQmfFl8CiCpAuKUCGrADJM5EMnuTt0fdThCRyBciIdIFZ4krOYFfSS+IBmnoV+MI24P7PN99Xk8jFrluTiVKnSPBcdQIjx5LFq110gU3vhlalZaP5ifbH6JTrLgr84qW4Tnh2tmgTLVWTthSgzED8RPE4BQhI7QRbLXmA+AWaoWvI+NLjxihAWa0hLuLaQcMjyLRTAhPBoMmPzZX0X9GfAgA8ab8BQGAx4Ip7FDzGw435fgpFPk0BYp9JhZfAJBO0fIwenA1ITbnhiXV9dIC+ttUFW1xoVSrpI1rma/S1/vzlyVAFoV3EK0DJ722pGg7UnY0Ogl0GUAToCENQAqNfPPkLGFKA5CTouosB7gFKM0ED8FjLtOc0t4TzIZx+VLXoEgNZCzf85pn4mw9egE3DOawbzAjzMoeZoc+lExc49Cz1H3DIJ0vJA8RJ4BZJBUonQFwBii+BVXK03XU2iQAhKwzQHFQBSh6HoXuMAEXeSy7bI3LVKi/iJnEEMTBJ0AbP1SOaA+QLEqXpyR4gfiWqW+E5YouKl3dS9e6l+5rvkz/TyReAl3/Qk6fkhLPJc9UBrDy7eOimDV4i7vn6ZNupveW6G4ywyRRF3sxm7VBL34xQItJM0HaSAtQ5AVrFxmFMlRsiA6hgG9C4mmTmhBoJIDwPrAMFqOFKQYgpJuh2SmDcqP1mg/p/HtTOgOP5IvE4sQvM4CWwbhUgRoBkBYgTIBK8R3KZk5e/8rYhzgNZyxBNKeX59O7EqkPfh7rr42cvNY8zaQVhgi5IClDKCJNDc+ES7XL0ASkCdIRBBCGyBUZud5c9QIkKUFoJDEEmTuxYCK5apM3f6QL/5sQ1+Omnfg0/+dSvNXVdWBmJEL320/AfhkpgYQ8QAGweDv42LQdIEKBoCCBTgGYs2io8S9jJIBKEWCWZJgKUkz1AMQoQb2XX7GgJLF4BMhBZxNnfc2KkG3ZzkKXUIQbQBTnqARIKEB+EuxQE6PDLbAdiiET0tl1/0ZOn1LwEz1UHyBQoAdK7GYa6/2HxqwUnbOBPQbnhoaCx484uisThLdpES/+FCEKMzQGKK4HJClBCCSylFZ53gR0uN8IZQHEt8IAg/9QD1GUJLGUtMszWJbBqw8VmbQLHaIfgEAM/qZ0Ushdk7fhToKWnlMCk2YxJ4PcFBMiRSmASAZqVCJDUASaDf6fL5XQCJL+unc8dSt02DnzQ7XDeli60k7+D3ADNMVNdfiGrigAdYRAEiC0YXDq2DA2GriV2gQkPUEsFKHkwKJ+2ThZwEkmCrmuh2TkcdkYiCLsjBCimBFYjdkCARtpUgLjq0mSCpgRowqCdMjPgBGiabstISBkxJTBLD0pgMW3Phkf3V4+EUmpiGGq0DT5ygmBklJesdDPwANEgSxIoQLwEZjXPAuMLmck9QFGitRiYZB6EWrOXhrDPdMJYT4e47v4JcPDJBT8lH5S7kO9uboB2AVpu5wSI+3/E/9v0AYU8QJkBoQBtaUMBiptJFdwpdWWJElicAhQhLaIVPsYDxAjQXM0VmTGxc8A4QjlAHZigQ0GIrRWgposHCeW6h4tZ+/sT5HjsKRtCjTJ0LVHF4QpQbAksjXgy8OOu7PEusGAYak3yAIUUIEGAIq+ZEclqOX1AbogAPX+oY1MyL4Gtkkpg6QpQ+ByiFCCFBSNqghYp0GaQ9gtEcoDqnStAcaGAvIygJSVJLwLyWTuQhF//Gf05ehL9GVsCyyDHrtq2rAoUoDQTtDB/J3iADpLVAIAZuQQmlbUqyIoUaLHftikNRI0sTISIVvYoAeIlAi1CeJqICbtaNMAUINMWfxsoQGECZBsxw1AFAcoEz7PIQxVDBCjyXFXWYv5kYwPKJ7yb3rjr1gU/JS85EqP7725xkBIg20s/0TTBrQsSV2JhirOH2vMBVWoNqYFhoCMFqOH6OE47gOv2fhB46lvh/eHfddkD1GjDAzSyjf6szTS1wg/nLPCRY3un6WOFCFBUAeIm6E6ToEMKUDKh1Y3Wxv5KwxUG6J/7p6Ph+UK5yFlG8wgRBjPNBJ0WP8BgCwLEHl9qgxet8Qi3wgct8BHVi6nI9Wo6MZfLjJPzdTyxbyZ1exkN1xflwnZHYURLYDPKA6SwUPDab8P14fkk6PRghjsRsOWEFaDBNj1AfspkdFG2WQQFKAl5WzLucsJzwm+w/88EGzakHCCzQwWID02UFSC3AbCr9NfdVQCAGUgEiBEuTzPhwIwpgRnBPLCoAuQ5Yi6bkYkqQMy3ECFjBiILDSOjfHHXTVvqAqNt8HrEAxRVgBokUIAsZsbWQWKHt/YMTjWIF/DdpvJgo0I/4znk8MsNv0NvfPpbQOnAgp5WpGdHTecdYHCIfg9yfrOil4qxp6H5DRwmA3jCpyWk6TYJkFOTyJZdlBSgCTRcP/UqvuH6eLP+FDY2dgOP/11whzzM1S52ZoK288DARvr7VDipWNc1kQW0d4qStmLGCEpgCQpQtsMk6EZoGGqKAmS2oQA1PJyhvQYAeM44he47I29JIYgAYKWZoNO67xiEAuSybSQTtAhHBHBwttkDNDoQJkAG+4xaESBufOZq9Q+eG0/dXgZXf3SNrqXttMHLHWyAUoAUegA5mKvScIMUaK4AWWEFiOdxDLarAOlcAWougfGraD3qW1lEFDJmaII5ho8FRk+gv4dKYPT1VUgGWfYeHbuaLuzFjBmat9UEbsyUfTdzBwAQwMhgb50+Tol7gGqz4iTi6PS9GIiaoC1DyueIKAZuUGowMmHTN1drogpQ0yLOCCpXhqgCxEtgVAHSpDlhAJAxjGQPUEZ6j6OJ2L3E4VcASCftSBnMK1MFaI7k8eP5LcCxv0I/l1/+7wU9rclN5wsgQMMjVAnMkyp8rwOSyPw/T/jHYxyURFUPN6f8xsFjg1d96JQwDNOuyc3aBACSWoKou17QLTW7X7qDESCrQD1nURO07yeXwID0VnhWBuMkIm+3UwLrcBq8S4JRGCldfYEHKPmx3VoFx2mUXB/KnwgA2MMStpMM0EDQBRZrgk7rvmPg63XZbU6CrkoeoHGJAB0uMwJUCJM+M0vXEKeeTsx51eDtp1NPYyc+oKAF3mZ2hdZt8LwEZrFyoSJACgtGxtSFzFxpeE0tl1FzGvcKBV1gKW3wSElFRpBAvKQlMNtAjUgH/JYL6FBSINYEXZNM0MevKeKqXzsBf/TuU9OfJG4SujQD7NA8G9fBPUAgQpGoMwIU1wVWSRqFwQL+fKIJ5YVDY+mxIQJECEymADlidlAN8D2q2CCsAOkageO6gDBIp3mA6OPxEhjdsUVcqCZfDP8/SoCqXAHK48l9M8DZl9M7Ih6aTsHDI7UFdDCOrKIESNcIZkodZAGx8M7H/RNwiAwDAJyZ9hQtn6knrlWkM/yGtoBAQ05rYBSlVOWk4foBASrtD8qN0VTmqAnarUKQ1KgCBACrWBksxQi9l5GIxDlggFQC67ALzPOQaSMIMRgNk6wADc6/AkMjqFnD0AZot+e+6dYESOQApXiAYr1XDHy9diCV35mqKytA43N1kTY9OcdKYBEFyM7Sz8hvQYA4yXzXmRtg6BqeH5sTn1MrBINQg7mCQHobPC8lbl1N94+rSMsJigAdYdA0LdQKX3OjHqCwCZp3i7XrAeIdUSRqCEaQPxMt2ywmCrYZOmlj8wUBiQt5gHgXWGCC1jQN/337ybjsjcekPocWNzVaIkAT8zyi3kJdY4SlRK+oa+z/HXWBuZys2chG/k7jHiB5X6TfRVnNrYeUGsMKPED0Txqi+4UrQLYRDUK0xdWZLROgmM++Z+AdYBy1cNcXYf8vkTyeOVCCO7yV3jHbnmKSBB4euRD10srk4YB+t2amOmgjZuTtcXI8wE6y2nyb5QfWFeeZjIiYNjSWxkyN0MlX4LRdnH2Wjfmgw07OAAKaTdCCsGtNMQ0AUjvB+EBUrqIUMgbQYM+XogDNdxCS57hECkJMLoHpJve1JZ+k11aoH21u6GRBLHj5Lq0EltoG35YCxBQUeW3jwape8LyeT3CYrT9JXWCZHP1u8FmNSeAK0PqhLM4/lvrZdrZZBpMzgOj+NzfbRME9QCeuo5+7UoAUeoLACO02K0CREhg1rpFwJ0kKRChgjALETyLGEpbA8hkjmHUDAJvPl6ayzwS3cxM0G4baEYQCJB2g3KcytCXUzjmvs/ePnZA5AYp2geVtqQQWzQFiClANliCuHMEEa2nRlghJQIBqodsNKQkaAFynITxABs9D0TW4WvBeerotTJ62ZQTq0mKWwLgBmiPaCVYLFKC66+OVBpu/NrtvQaGIlt8D9VLTUNUoWZiZPtze35QPA9PUK/O4fzw2bt4KAMhU2ys/aEw98UMzuWQjdPLJvS53SwFBGUxOgQaaS2D8fisfmnIu0EYnWLgLrBx+Pg4+CwyNRAL0yOvTeHUi7G1xXK+tURiiBAYPJMHYv7FGX0N55FQxXmJvGwqQ2U4bfJoJ2uAESFo32Pte8cPPy31ASQQol6cEQ3NbKEBSuONvnEbLYD9osww2VQ4mwQNozwPE1s0T19LPXZmgFXoCkQbdhgJUrrvIwIGpsS9qVIaOQJTAYk6CFmE13YhvZTERUoDMHLD+zNgSmC8FIaYtXHEQgWmy6sJKXGRgAybng/diTmPvHyNAFSQpQEZ7ClBkX3Wm4ujSVaucUF0mkgIk7S9VgAIC5LsONNkgzeDp0u+Sf8IydCHHx6WA9wyH0wmQztSCOUK/Y49MZwFNpx0y5Ymun7YXChAA1HS6X/OlNgei7qcBiK/4GzCvFXH8cdS/VnTbI1C6Q0+KRJ6iHgpDTFeARAkMEKqlUIC4IsPJFfeqJRmgOdLCEPNhRaZgSyWw6ONxBUirY67WTIAm5+u47Cu78P6/fih0u+s50DVGPNogQBbc2KGlALDFoa+hsfpUQSz2TdPjM2kQKhD4Wry0NvgUAqRp1EfjyASIvU9VL/y8vBV+MjIIlSNXYJ9jyuw8QogwQedtA79+KiVAD+4+jFKtNTEJSmBcAQo8QHHksuZ44jNVCpBCTxEqgTnBHDBAmjLMbq80vED9AQK5OwG8IyquDGKzk4hpLx0BCnmANr6BnuT5VPbGvJgaThp8FlgmMbwsEXEKEFOXKuZQaPEUrfCMAHFCUmxqgzcEOWpWgKTMosi+xpXAPCfYr0pIAaKLkkc0WFaQBA3QEhhXkXgJDAiPgiASGbJNXZR33JgU8J6AkEABGj2Z/pR9XAD0BiuBgX7HnthfAQY20DsXUAbj5H2h6qVj0s+/3DYBCspfW1cXsIEpQKP+FJw2xj+YjJRosnrSpgLU8FoQIF4Ob1KAWhAg3gpfnWoKdFwVMeim5wBJClAMARqbrcH1CfZPV0MnWSIT9BQTNCf+FlyR7BwCIdjqUnXOW3c61jBiwRWcbGoJjHl4IgoQISQ9gVtCxtThQ6d5V4B4n+ZZOCLvXB0v1VBzPKGSRT1AxSL9blh+LdETVnd98LcgZxvYNlrApuEcHI/gxbH0AEVALoGFFSAgPguId4BlTF2k+ysCpNAT5KUsoKjhLmuFpclyw0WeJ8laeZE0nAheAoshQBbrvLCyS9cGX8iYQQls8/n0p+xjYgoCaQSqSloAWRx00S0ivWb2uDy3hWOad4Kxk8k8K3PFjsJICkJkJ5q61LIf7AsbSSEpQA4jQD7RgseUSmAOTNqVomnwOIlxHeEBMiRvkDzHjYQUIE1cjTqLpQDNjQGNeRBNx0/KbKp4RAGymALkWnRRf2LfDDC0hd45u6frp7YJa+NdIHl3LXoSr83PBDcefCI27BJA4P/xT8Dxa4tYvZ760TKag7Hx1mGIpssJkEQehAKUPg8sNDICCEpgrUzQSYSFI1MEitTLFFWBVkXKM7QNPvJ8HKwBIIcG5mJUCK5M+CRSanGlC7pUBYiboL14Bah0AEOYh0t06GtOFiUwjvQSWHwbvLyfaSUwIEiDFkn0zHjOPUDcPHxwtibK8LapN3Wc5pgHKKc1cLgcf/Eim8zz7Hm5obmdIaVyFxgQfm1xnWDc/7N2MCOeRxEghZ4gH+MBykZygEQXWN1DkStASQuahCAVuflA4icRK5NwZbgIyNsGHvBPp96X0/8DvdEw6QwjIFAQ2OLtm7nE8LIkaEIBkhYCZhid9ukJkx/whz12AmVqxLxP/7apBGbJJuhoGzz3ACWXwOQuMJf5sRxImUiSCdqBKaL5eZI3cRtiBpIhlcBkBUj+3dIDOd5tLBIBYuWvuewmvDDHSHSUALn0ZHncZqr6vDg+B3dwM71zpv0holHYPVKA+DHUKLP9fvofgK/8KnDfdc3bEiJKYI/7J+CEtUVoVg4ljX53Jw68nvpUrueLQaJmTlKAmAI6oFVbeIC8oF0ciFGAkkzQLRQgQCqDhbOAVscpQPPMZ8IGuQpIw1DLjWaSIpfFQh4h9r33NSP1gs6wuALkxStA4zQB+hWyEfl8sclbk1YCMxPa4GU1pB0FCAB8qXQNBK3xx7AYj/HZmiA2a4qZpvVNswMljRumo+DdwLahC/WKr1nlNghQ1AQtX2TG+YA4YVtTzIih1JWGl9o2fyRCEaAjEPzKJE4BkpOgCSFUARIEqA3iwslAhAD5fpC9YeeWjgAVMib+yvu3OLvxVyAb3xDcETVCizEdnZ/g0hSgSZeeqLeN0tc8yQkQI1wlj75fcSbo5Db4oASWRIDk+UW8JOWFCFBNeIBcGDCYJ4EHWfpuQ2QHGVIJTFZ95KtnXdfgLrYHiLXAv65vQol5fEKzvwiB7VH1YXTNWqwbzMAnwITOTpwLKIFlhHq5sO+uxoJEXT4R/ok76c+XYwa7Hn4FqM2gAQvPk2Nw4lpKnkomSxZvMRW+XA/mgBk5Kb6CfW6ZFiMkmk3Q7P2LKjydlsAAYDUjQBPPh26OLYHNM6WLq0YcvA0e9PsWNULLBKhSlzxxfCizHn6uKPj3ngaDNr9P/hglQM+RY5C3DdHBxpGqACUkQcuzsVop0YIARV7HHAtH3MoI0MHZmghBjO4jgFApcTKBAFWZ/ycrTajnSlJc+TEKrgDxEpimacE4jFgFiO7H2oEsBrIWOGdbbiqQIkBHIEIm6KgCxA5anwCuT1CpuyiKKP3WCpBoK42UwOqSoTKTXUITNDtIXV8LX2mIVvgZAIDOFaAWHqc4cI9MaGo0a8c+1KAnm+PW0JOBmAjPMOu1UQJzq+EOJqEAWaEFCQgye3QSbO+yEpgrEyBpejRVgJgpkxMgzxEBcLICRCTSQyLlg0ABWiQP0CRtgX+yugYlnqkUGWci9jk3hLM2DwOA1AnWpQLkuaIVeqEKkJGjBIjUS1RJefVH9I7p14BypDV+7y8AAM/geDgwcQIjQLUsJXSVw/uRhnLDRZFNgjeykgLUZoJyw5VGRgBBmjYnnU0KULQElkKANrFy9Ou7Qjc3KUCWAcyxVuuBdeHHkGaBAc0EqCSdLMMKED25kpQWeEDyAGkevBgC5B18CgDwvH8MChmzSQFK6ya1EmaByf6fVko0v2j19agCRNeEY1kJbLxUS+wAAyDGgWS1RqhhQ0a1QfeL+0eBYM1qJ4IgaoIGgIyR3Akml8AMXRNka3aZDURVBOgIRD7TbILmxEeuzdbZ/JY8u8JqpwQmWqkjYXjVegMZ5idYSgIkX4WFpFq5E8z3gzEdXRAgkRcSIkD0xDxWp/dtGs7BNvVgICrDnE/vL9phApS1JAUICHeCidDGtC4wWQGir81FMM3dd2piXpsDQ1yRBgpQQIBMS1pgJdKjRQiQp3H/0OKWwJ6qrwsUIJkAMdLpEh12dgDnbBmm288zstttCUxK3rYX+N2183SftPoc8PJOcTIG0BzWuIcSoF+4NGH4+DX0+PMLlAh4s+lhiOV6QIBCxy474WU0J7UNuR7XBUaIFEzITdBJJbCU9WLrv6E/9/0y1H00HOkCG9Dmg/coQQHKa3UApMkHFFKAGsHvXJ32UwzQAEJdkW7cbMPxZwAAz5FjkTF1DGbNkGqTXgLjQYiREhhX5NvwIfLIEl9LUoACD1DiIFQgpKQlKUByBxhHUAJrbcbnLewjheA95fsfqwCVuAJEPyP+vVAKkMKCwU1sFccVC2DWDHeBAUDd8VBpSApQOwSIqyERAlSXZhKZS9gGb+haqOQnIJfApBMcurjC16MKkNsQj3mgRu9bM5DBYNYMusAYOMlpHoZK84t8wq4CZQLEFKCqlFrNEcT3B4uK53IFKIgE8BpVQVQcYopofl8PSmA8AZd7IYAI6YnMdHP5Yy9WFxjrAHvV3yC6vIhMgJgyMY8c8lkTZ22mxGfXYfaZdqsASSdoa4EhntmBYQCA7szhp9/+PwAAAvYZ748QoL20ffsR/yRsGs6JE445TGdp6S3CEMuNoAQWytDhV/ztKECyCdqp0K6tJhN0ZBhqWyWw4ymh8eoi6RpgJl1p7t6gy7rEskPNY0ikTKYMnKZSjEyIytKxr3EFqEUJTM7F8qO5Zk4VxjRVJPdY26BpGjRNC5WYUmeBcQXIi5qgw7lsaeBrtRdRgGrsIoeP8qk6Hl6dpJ9JrAJk8e+Dg8MJChD3AMmviXeulhvpCtDzYyVMlRuwDA0bh4PPzBYKUPN3UC6BARA+oOWWBaQIUL/hNpoC4IQJut6sAMm12Zrro1z3OvIA6XEt4QBqNekEvoTDUIGAXIQOVKEAzQZXrgCMLoLuROcVJ0CSL2VvhS7mawYyGMha0jgMijKyyJiBsZAjb5sg0FHlHWxyKzwPQiRWjALEyJgU3uaxcqSvGWIumufUBFFxYYgFmStAjtMIyj7ywEjps4sOteXls0XxADlVYIZ2cb1KNoqcHyKHWUop0AXbxFmbhgEAj8zwctlMeJBnmyCMfNaJhYyVPBS3HRQGaDluBPM4u/ogAOAfvV+hd0pEAJUpYPIFAMAj/ok4fm1AnHOrqKk7W0sPoQspQDIBkk54aQoQbYOPnBBLB1KCEDsgQJoGbH0z/f21n4XukstgRYeVBaPqDxBKmc6hOQtI/r+s/mptlsBkBciJkvqJ56ERH4fJAMrWqLhZJhjdzAITKdDtKECsBMaPO7GvrBQ9lLNEB9XT+2eb9k9AUtImIxPYOYIQRJkAsS6wFh6gf3iEesfedspaDGZlBSh5IrwwQQ9yBWh5doIpAtRPNCrAl88B/mZH6OaclAMUmKCDjyqYB8YUIJEC3Y4CFNMRBaBRpYujCyM0cmEpwOvWIalW9gCxhbtGLGRsC53CMCIEiKsSdhHj8/S2NcUsBrJmkweoQrJNBmiAerI0DbFhiB47wdRhN3mADCsYaMq7YrgHSDZB+42aUIYcmEKS50neTqMBg4VfmrICZKURoORBuAvG1KsACOZRwCQGhQcopABJKdB528BQ3sK20QLKyMGxuy+DuSIiwUqdz9QOdGaC/o3scxjUKpg3R/A19530zv2PBhcrjAwdzhyDaQwKAzQADK+lBGjYn0r1X5TrCeqtKXmAUq7eG1ETNEDLYNEgxKQk6FYXTJwAvf7z0M2yETpXYwQo6v8B6DrCjr0cGk3t2HN1SQGS7tO5PzGlBZ5uKJn/owoQM0A/7x+DgnRSH21TATITZoGJ9biNMFZ+oepq0n5qBh18C7qOrx+kx+jLh1hzwECyBwgASvPxE+G5ep4LeYDYhWXKd9D1fPzTY7RU+95zN4f3P9UDFHSBAcCgUoAUOsbhl+mCtfsnIVOyMEE7rmSCDg44OQ263PCCHKA2SmBCgfDDC4ZTC07aSw257V9ALoGJOWDNJaV2YNgsLyRKgLJDwZXMQAYDWRMzJKoAZZoM0ABV4nJWfBaQV+ehjc0eIDm+n/sLuNLja4YYZeE7VfhMqXFgiLZcIhSgOix28jPlLjArD5fodOxFRC3zdK4ALcIixcpfL/kbAGio6iy+X1Z06gEB4u/p2awMNmOxE2gXnWBuLXi/W7UmtwRTTQyW0Gye+i48T45BhWSocsiHvTL/zzMmHcR7QowCtA7TqcMoyw1XGmEjHbuMxOoaaVY2JNRdLzBB5+gUeszuiwlClEzQhARkvV0CtPehUJlxVSHD9g+w+MiPYgwBAiJp0OHvXamaoACxQNa0OWB0B3R47BTmRQnQeNABJq8Zq9tUgHjTQbQLrDMFKIYAGUHQoKZpWMcIELcaxXuAguN4bi5eIeVdYHlLVoBat8H/9KVJTM7Xsapg460nrw3dl9QF5nq+mFy/litAOaUAKXSKCu8qIUGWBpKCEIOPSg5DrNSlRbQNAqQxBUgn4S+qU6cko6G1uOpaBMSa9WQTtDQGo+M5YAhPjSaECALkZwZE++eagQwGMhZmEH4PqyQbS4CASBq0lAXkSaGN0bA0TkBNiQD5LO3a00wxyoI4NeFrcGFAFx4g+vduQglMt/P4785H8P86H4Fhx5fAFkUBYgboV8kGnLi2CLtIS0m6Vw9OnlIJjH/Hz9hECdB+8Fb4zsMQHU44Se8IEEf2zH+PVQN5PElYWzgvgzH/zwN1OjhUJkBcDVmrzWDP4UhGlIT5eroHCEifAN5wfWT4MFSe21M6kByECELzpdopgQHA6hMosfHqIf8TL4EVMmYw9DWJAElqVvseIEZmWilAYIo1IBoGBLgB2j825N8LlcC6SILuzAPEmg4gEyD6/NzTuWEofIyuiSuBGZYIU6yU0xWgUAks27oL7FuP0guOf3/2xqZjR45ckXG43AAhlACvLqgSmEK3KEvzguYOil/jR2HIClBQAivLozDa8QCxRUWPlMAcdhXtaJ2XmBYKftDKbbGhifBcASLdKUAme80iMp95gFyLXiGbuobhnIWBrIkysvAhdaYh0zQGgyNnG6jGZAH5jAC5enOomWGyHCD4wmDJr149mPB0+njErYnbQwuopOKIKdiSFyJj6rjbfzPu8X+laUHzRQlsEUzQTAF6xd+A87eugpEpwuMGca641ZoVoM0j9AS536fZOV2VwBhJqMMW3o2uIaeQWwVg21tw6oZBPO7TGV/Y/zBVa1kA4n1lOjbihDUSAWJ+mKzmYPxQshG6UncxAK7GSATIsIXx2ktKoEakBMYnuMslsGgbPEAvJlolQXMk+IBGClI21hzLABqI8QABoZb+NA+QrP4aXJ1uw4sYJKNL32lCgDHWAk+OCbWGh0pgbbTBR2eBdaQAWVwBkkbYsAscXkLjChBHNK2ag7D3oladi029DkpgzV1gSQRotuLgvmfp9/O3ztvcdH/SQFTeATZazIjjLTBBqzZ4hXYhFCBECJCcAxTnAZJKYHW3sxKYUIDCB4Xb4ARo6RUgHkL4he89j5fG2eIdKoEFJY60q7Yk8C4pGpgWKEA1g75fq4s2dF3DQNYCoKFqBifBClIUIMukpREgZIL2GWHzjeYFnE9uNzWpBOYGJTDf4ApQXRAV3r4OQIwy8dxAAZJnhMkLc9QPw9WjRSVAZCPeuHUE+ayFOURa4euSCToTmM8B4DVXKuF0CEGAtBYlk3YgKzEnXgJYWZy6YQCPcQK072Fg7EnArcLNDONVsgFrBzKCFAAArCyqBv0OzU0kv55yzYkvgWkaXEaE/ZQBmHU5B4jP7zr8MsBTxvl6IHlxKAFqUwECYgkQV4DytgEIBSiJALEWbq3enAMUMkEHKgMnQJrZ+vN02MUBkRWg0gGgNgNfM/AS2ZSoAOXt+OMakJKg/agC1N4cMCBYs+WBqFzJycQoQAa7EIsDT4POkEYsyag6cSbo9BLYd546gIbr4+R1Azh942DT/ZkEAjQxH2QAcQznVBu8QqeQg9X4lRQCFl9uSG3wsgIkl8AaXkcmaFEOinSBufVAtVhqXP0bJ+G0DYOYnG/g8q8+SA2BoRIY3bcKMl2VwEyLviZbc2mQGSvFVDR6AuAnYd7eWzGCk2A5pQQWnggfECA+t8yLyTHhw1CpAsRKYG5QAhOpsW5NzGsLeQjYAppEgGRzZpMCxP42bg7cgjCzB+TQcwBoB9gbt65CMWs1ZQF5VTZ/DXkU2Hd8TZGeAF6sdR+GyEuOPSnfygTolHcDAE7bMBgQoEPPAi/fDwDYVzgDBDreuHVV08PUc9RPUZ1KJkD1egUmM7JHS29cCeRqYhS+T+D6JMgB4iWwiReCjeQLItkI3QkBOpYRICkPaFWsAtTCA4RG01TyUAmMnaR9n4jE9miOVRz4xYEvJ9uz7+JM7lg0YCFnyQpQex6gIAk6wQTdFgFiXVSSgstb4vk6tk4iQKsLtih1R6FJ0QhxYYhcQQuZoG2uAMVHKfDur/eetyk21DHWA0QITvjxx/Al65ZQuU6YoBUBUmgbsgLEU1wRVoDqKSWwmuNFFKB2SmDxHiB+pdkPAjSct/F3//lCnLphEJPzdVz+1V9gT5URAdkETTJNXVXtgAcFish8qRQDBHV3ToDmteDEUUWmKQSRI59QAiMsB8g3Ylr2GVkx4cFhiytfvIlmiPA34tbhSyMyxGNLCpAVUwILK0Dh94qTp6aOmYXA94G7/ys0t4pH/BMxVzwem0dyGMiYTQqQV5kBAMyRvLj6Hh2gn/NrLidAnStA3HTu9EIByq+i5NseAE78DQDAKesHcQgjOEhWA8QHHv4rAMAvvZMAAOdvHWl6GFLkYYgHm+7jcKuSodUKH7ucPBMnngDRRGIiwktFCYxHPNgDgC59/rIRuhMCNHoiUFhLs61Y2e+CbauwqmDj105Z24YCxAlQPeQBqrteSFngERiOH4Q76lYbBAhBMKgAm2A/b1FiKitAcg5Q1k5eS8xWJuguFSBeho5TgGJb4Dn4OAwtfh5YnAeIr2dxCtCrE/N4dM8MdA3Ycc6mhP2PaYOvl3DMwe/hPcYD2JwPblceIIXOkaAAtTJByyWwkALURglMlIOIK3JogMBsGadaLAVGCpQEnbJ+ABNzdXz8bjaFul4SnoUa7K48QJrkAXK8wAM0yxQKrgDxDIySRIAqCV1gADNBk2YFiBM2Eudh0AMy5okSGLv61Uz4bF81tw6fKTWe7MviZMdtQNdI6DGB8PekWQEKBqn2DL/4S+C1n6Kh53C181Gcv20VNE1DMWOixDvqWPeXzxSgipYX+5a3TRRsA/sJM0HPHWwa09IKXAFyekHezQxw5b8CH/y+GEp63JoCbEPHo/7xwT4C+O4MnfwepwBZLAzRqIyHjrPQfrN5Y46RD5MVBOXTJALUNAdsZGt4g6bJ7LIC1KYHCIj1AW1Zlccj116CT/zq5oBwJSpAwYlb9vxE/UC8BOZ4RLyujhQg+TstSub078MeoM4UoKQSWDtxCyJJmQTbutwDxL7/6yUPUGwLPAfrDMyhjokYAhSXA8TXrarTPCvtnifoBfevnrQGawfjvVaxHiDpfV5fDFQj7gGaVW3wCm2jkm6CriaYoLkKUnO8yDDUDggQwtOZA99KfwgQQKX1v/vPFwIAnp1mX03iiw65KjJdeYA4QbC46sIUiSmPnhSiJTCeBeRoNlyYKSZoM1YB0tw0AkQfi7bB04WFl6SIboDwk7hXF0QlFKTGXovmSt4QPb7s1awAMX+R1yMCNP4ssPN/AAD+dujDeJ2sF2SgkDFFGjR/vwkrPTbMcLlnzUAGkxik5T/iB1PN2wT/7vbMv7buNPqPwTJ0nLiuiMf8E8VtRDOxq34sBjImTt3Q7J/IraJX1av9KRyYTfDxMLOyazYrMXyOG0no2KOT4KWTTWaAKjXy/2VwhalTBQiQCNBPxU2apgXqj5kLm8dDzyspQPVkAsRLOI7ka9KjydIx4MdGyAMkZvFxAiR7gGy8/bR1eOcZ62PzvcRuJyRBd6YAsQtVuQSmhUtgQzlLHKexLfBih+SJ8HElMGaCtmQCJDVzNMJlsNcP0/XqouNWp+x/TAlMWjvW5QMCJCtASYT/SIQiQP1ECwWo4fli0YhTgOiXDSh0MAzVYq3RFlxhnAOCK8044+5SYnUxg6GchTrsgIyx8mA1Zrp6W5Bazx3JAzTp0scPSmB0uymfnhjqOl28Ez1Alh4bhCjISdzkelEC80WLLRFJ0JYgTbpXF0TFlwkQ8xBlIZ0Y5RJYKgHig1Z7cJXm1oF//DDgNUBO3I6bpi4GAJx3LFVNilmzeR4Y++lY4e8pJaAaqvkN9IYOy2D8u7uY5dtTZR8QgMmBk1FDBuceOxLbeWYM0teyRpvBe275Of7uwdebTqaEzezyrObjlrDjUHPjFaCmQaiGDQxuDP4fvRjiClBtNjiJtU2ApLlgMiHjBGhgHZA0GFTMsWqEPD/RTCDuU2l4PmzW2q+31QUWDAcW4LEZbOSETIA0TcPtV5yP//W756UOM02aBSba4DsogYUUoEgJTNM0UQaLbYHnkLxUcfPAAgVI8gOaQYJ8tAzGS1WcuMQhUIAk8iTNxluTa1aAXJ80ka0jGYoA9ROhLrDAAySrHLEmaPbFnC7ThawghqG2XtAsi09QdkP1Wn6lGataLDG4ydLl6cCMHFa7NEFzgiDa4NmJeJwNQl3D5tnw3IzDHl206xqfA5ZUAjNRJuz9krrAdI8SIC1ubpnOTdCBAscXb183RfaJ5tXFVa0nmZw1RmJymnQVGCqBNQdmCnCi1AsF6KGvAuNPAblVOPiWL2C+7sEyNJy8nioPAxmzaSK83mDxA3ZYLeAK3KzNCFCHrfCksfgE6JT1A3iabBN+rKd0GoB4wbbm8hcA0RZ+rFXC5Hwdf/hPT+MdX/opHtszLTbRHaoA+THKLbE4AYpXgBryIFQjQwnIkNTK3KQAse+ifNFltUmARk+kQYtujZrAOfhFW5L/R3reaBdYkgIUel1tdIHFKkC8aYKkH79pMHuiADGlngTPz1vi5XWMt8KneoCkifCxCpDDghAjCnkxoRWer/1DCV1ndP9jPEAS0RyVlrecZQj/Ybs+oO89M4ZvP3Eg1tO0VFAEqF/wXGHWA0BPEqyMYht601VlSAFiJbCpSgM6fDZtGeEskQQEfhgvHFvOfSt9VoAAYIRdlYhSCSOH3XqAeAuwpXloOJ7wpBys0dujJbBDrDRWZSGHSVJ5OAcohgDFSfgxJmixeGuGWOioAsSVIVkBYleQ8gwouQ0+xQPUUwWIddrgwv+CF+YpYdw2WoDFFsFCplkB0hv0hE+sSAmMLfyHDR6GyAiQUwPu/n3ge38YVh6iYIqbt4jf3dM2DKIOG8/pVAX6lznqB4rz/wAQBOiMwSqu/3enYSRv4eVD8/jEnY+LTfS0qexMPdS9FA8QD0HkXplByczaRIDYZ1GeYE9utUUwAFByte50+vu4RIBkBSgJUg6Q3PbOM7+irdqOJ3mb2ijH+3FdYIIANStA7UKUwBbSBs/WqrovE6CwBwgA3nryWmRMHW9MItOAVAKLnwgflwMEJGcB8Vb6wRQCFOcBkkuyq7PBe6NpmjQOo70LrD/+7nP42P99DK+lhIUuNhQB6heqU+wXLQj8mqdXVJqmhSLNAXbF8PoDwOw+wcyny43A/wO0J2nrgRoiEyBRtmmj7r7Y4ApQlbejl6g/qtsgRJkgeG5DnJD3VoNJ8IBEgBy62PCU5+QcIAMTZJjtY+BbMZhMrMcNbuUKkEbgeXTRIiyUkuiWeP8NiQDJJmiNEaAsI0A+9JCBNs0ELbJgEhSgXa8cxoV//AN8/5mx2PtD4CW/7DBeZNlNJ64NTrrFbKQLzPdhsvESfmYo9FD8/T+oRQjQL24FHv9bYNctwN/8ZviCIbQvybEDvQL3+fw/1Q9h31u+iH8onwXb0MVE+yYwAqTNj+HKi7di5//7Vpi6htcPV7Bvmr53hhsZWiqBtz1rCYb1UAgi/1yH0ghQRAFqt/zFsZZ5orpVgFBHw/VFOYUrQOtZ+acsl8A6UIBERITXfDFX9jkB6lwBMvT4JOhGJyZoTiBIcBw6TEGU//6jbz0eT/+P7Thny3Dyg0kDcifLMTlAMSZoIDkLaJaNIUlXgJo9QOVqQMhHMuH3Jq4T7JsP7cH9z8eHgfLPPLfAAcYLwZIQoFtvvRVbt25FNpvFhRdeiIceeih1+7vuugunnHIKstkszjzzTHz3u98N3U8IwXXXXYcNGzYgl8vhkksuwUsvvRT7WPV6Heeccw40TcPjjz/eq5e0cPCFKL8qqN2XJCN0xHibmX0V+Ot3And9QHwxp8oN5Hn5SyqfpEIqB81UgwOJT2Be6knwcRjJ04WvzOZJ8VJhFTZyKa2riZBmCrlOQ3iADjms9h7pAnvYPxEkM4Rf6ucACJsJZeRsA68SVraZDL5/JidAmXzzH0mGZZd1f0EyQWsiqbshiAqRS2Ds88shGJMhI80DJN6HSAo4x09fmsB4qY4fvTgRe38IgjDn8OI4PZGfuC5QMgZCJugS0JiDBrZgZuNLYHv8UXrDzF5gdj/wkxvp/3ULeP1nwF9tB6ZfT9yXxfSvjRRsrB/M4lWyEbdMXQBAw1mbh5JLspwUuDWgNotVBRuns7Efv3yNXvyYrHNQz8Yot5wI+/EGajoJnhMF9ro7UYDa6QCTsS6GAIkW+LXN24vn5QoQ3VfeCs8zgXgXVNWhJWHHJbC1ThQgVgKTc82Yb2reo8dKoRsFSE8vgXXiAapKChDPBIrGeVitkqWlQMnJufba4AF5zFBwzBNCMF+tYz0Ot0WAZA9QpRJ4HUMmfDR3gj0/VsIf/ONT+O93PRn7+Hx+WdL6uhRYdAJ055134uqrr8b111+PRx99FGeffTa2b9+OQ4cOxW7/wAMP4PLLL8cHP/hBPPbYY9ixYwd27NiBp59+WmzzhS98AV/+8pdx22234cEHH0ShUMD27dtRqzUvFp/61KewcePGptv7Du7/yY8CA+wkGtMJBtCSmD7P7pvaLRbdqUpDmiZdSDYiyuDloIgCpLspZZslBleA5hC+Sq0i0920b8kk7Dm1UCJxzjLEApkxdViGhr1kHQ7+l2fxF3gfgOQSWN42sZuwE111CqhMAZ4DneXzxCpA8r7wBGgvUIB454vpx5ugBQFiZc9QhxhaKUCsg8yPVxW4KT666MeCK0BWDi8fogrQSeuCk25TCYyRzjoxYWfDxJAToFcbw/SG2b3AD66nz7H5AuDDPwQGNgKTLwD/+5IQ2QQC9XKxDfynbKCv7+7HqdqXXrLIBmGe7Li+kG3/0O4p+D5BxmdKUAwB0pl/zPSSPUBNSolMgKIEx+YEaKEK0HPBba3GYADixD1g0NfBSzFRBQigPqCwAtQGARLRDs0K0BwjQPmuPEC9MEHTdaXmB2sWzwRqZ5p8eId4G3wDh8v1pk6rasw0eCBYu2TPVc3x8T7ci19kP4bVr96dsv9MAZLWg6p8jo2UpaMDUR94mXY5T1caTfvr+wQVp9m4vdRYdAJ000034UMf+hCuvPJKnHbaabjtttuQz+fxta99LXb7L33pS3jHO96BT37ykzj11FPxuc99Dueeey5uueUWAJS93nzzzbj22mvxnve8B2eddRa+8Y1v4MCBA7j77rtDj/Wv//qv+P73v48bb7xxsV9m5+ALUWE0WEDkNOho8jOXwqvTyLD69HTZ6agFHoBEgLxQrTbwrcSoFksMPlZghoT3pUq6bYOXouirM7TVGkAJBawZCOZ1aZomOsHmGkRcNaXlAFWRxST3rky+JBZfADBiFSCpHMdle/5TN8SsNg0EOruS9fXmEhhXgPwIAbJjugWDp2YEKMEDxCMXorJ/LHjXoJnFS4eoAnSSpAAVoybompQCHbni42nQz1eH6Q1TrwJP3QVAA/7tF4D1ZwIf2onG6lOB8iHM/PjW0N9zArTYBn5eBuPjaS5I8v9wrGZdYwceBxD4hR7aPYWq46EA+h6aueYWck6EdT+eANVdr1kpaasExhWgTgkQNX1j7iAl+kDrEERArEtDOn0d/ETMf8rzpCoN6ovL8NJeOwSIf/9jPEBzC1CAhAk6OgvMa98DZAsFSGpqIeEusLYhtcHXHD/UaeV6vtivqHUirgQ2W3Vwtv4Kfbxn7my5/3UneA9qUgksWkofiqRB/+JVSoAICY4Z8TiuB86JuvFo9QqLSoAajQYeeeQRXHLJJcET6jouueQS7Nq1K/Zvdu3aFdoeALZv3y623717N8bGxkLbDA0N4cILLww95vj4OD70oQ/hb/7mb5DPtz6p1+t1lEql0L9FBc8Ayq9KUIAi3Tz8SpB4KLIBivN1F8UO5oABEG3UlhZRgJgqoMd1Li0xVrES2JQfIUDdmqA1TVx56Ux58zULdVhN2RvcB1SqOSKdNs0EDQD7dXbiOfxSUI4hGuyULjAA8FjnhihJ6RZ0aYK7zkokoRKYGTZB+1qkVNqGB0j34wkQl9EbHShAh+sGKg3aAXbs6uCkOhBtg2eqm5wCzcHToJ8rD9AhoIyg4twrgI1voL8PbsTPitsBAHv2hdvkOXlfKgIEULH13GObE6BDOO4t9OerPwIAvJElRr8yUcaeqYqYAxZLgJhiY5NG7PBL2i0VmZo+sAFgQ1STS2BMee+UAGUGgGEa/CjKYK3GYABiqPGQRr8vc5ES2EDWFARlvu6GX1cHJbBQWZcRoFmX3tfNRZPF2+Cj0+BjZjMmIRNDgBqkuQusvR1iDRkGfW/kzqmKFGfSbILmY5WCbWarDgbZOUR77ecAi2No3v8gjoWjVpcIkBuuuAxJHiDfJ3hw95S4rywNuwWCtQZID6RcbCwqAZqcnITneVi3LnyArFu3DmNj8UbLsbGx1O35z7RtCCH4wAc+gI985CM4//zz29rXG264AUNDQ+Lfli1b2vq7rlFOL4HJX+SspYfkxgEEX9h8B5PgAYRLYJJZzWQnEcNe3JNIO+AK0KQbJhBVdGmCRjBOQq9S4lm3igA0UX7h4ARoYq4Oft5JU4AA4HWNEaDJlwQxqMNCNk7a1YJDjsf3i9lcuglDKpsZjACJ+WAAdPb5JZfAmuMSxN8mjEHh4DK647ZDgOhCuGeOvknHjRZDPoaCpAARqQQ2J80B41hdoJ9B1TfECAlkh4Bfvy603QFmWjcbc6HbjSUiQKdtCEjFKesHU/0TAIDj3kp/vvojgBAM522czMqEP3zhEAZY+VqLMUGbjDxn0AjnsDDU3RilxLACNTlJAeKZTJ16gIBwGcxzgjJ+mgLECNCgIEBO6Odg1hTHV6VOFSChbLVhghYXBzEm6EAB6r4NvskE3YECxH0+VU9SgIQJujsFaNCk743cCcaPW11rfty4LrDZqoNBjXVe+Y4g6FHEKUD1mkyAEhSgioPnxkohM3QlMo+M/z9vG4nzz5YCR2UX2F/8xV9gbm4O11xzTdt/c80112B2dlb827u386GMHaGSXgIrhAKt9JDcWPSDE0Ang1ABJHqATEIPKPNIUIAK9EDiJmWOGroMQkTQLWIwAlRjBusmApSh2x2U0nujsrK4nZ3Id4MR2MMvi4GRdF9jDi9NE8Zlz4soQIYJ2zJRZ1eJvGuKaM0KUDslsOgirbHPXkswQVdFCax9ArS7RE8QsgEaoFeeXAHS3KoovcyRXBOhtE1dRB9Uh+l8Lbz10/TYkLC3QrexvfAVq75EBv6tqwviBHNBzPyvJmy5kLazz48BE8/Tv2M+oB89PyFKYHHHLi+fZtEInYA4QpPgJZO/GIoaee8QLW13qgABAQEaf0ZqpzeBfHKaMCdARdATbtQDNJC1QidpR/YAtaMAxQ34ZaXjGtjU+i5MtmIWWKQExj8L22inC4yNNJJM0HW/2xIYXZcHDE6AgvNBRQpBjIY7DvD3thYhQAjMzHjp+7FPyXN96tJ6UK9Lqo+X5AFq4BevToXuiypA/P/9LH8Bi0yARkdHYRgGxsfDbXDj4+NYvz7+qmH9+vWp2/Ofadvcf//92LVrFzKZDEzTxAkn0Fr8+eefj/e///2xz5vJZDA4OBj6t6hooQDlQwqQEVKACl5AgPKdlsDYFZMFD7NSF5jFvAaxvpUlBu8CO1gPn9CqJNP5wsHAS0VmjRKgeaZOrB0IPwdXgMZm6SJaSLlC4e2bL3lSJ5i0+CaRtWh2iSaVwGxTF9H5hssVoEBpMMxwGzxJKYElKkAJJTChALXlAaIL6MvT9G9kAzR9bgMNecQDS3cuoRBrSuVE9NnzPgdc9rfAhf+laZvd8+zqOUKADOFfW1zybho6zmCdXBcdn3LSF3+QAY6lCdl45YcAAuP0I3umUUg5drkHKKM5qMUoQA3XR0bkAEnf4Xd+AXjHnwDb3hr+g+h7Y3dxnMsKEL9YK6xtmmMWAiNABZ9+l6MeoMFcUAKrNNzwjLM2PEDi4sBvVoBqsGHoWmhAcLsIpsFHSmBe5yWwihesH3WWCt2tCbqg09c5XZYJEJ8E3/yYcV1gVAGSCdB9QMz4Cp43V5dKbE5dIj0RE7RcAtv1yuHQfZWEElg/DdDAIhMg27Zx3nnnYefOneI23/exc+dOXHTRRbF/c9FFF4W2B4D77rtPbL9t2zasX78+tE2pVMKDDz4otvnyl7+MJ554Ao8//jgef/xx0UZ/55134n/+z//Z09fYNbgHqDAKsNh8lA6KL2IuSoAkBajgzwa/86vIjk3QLqYlBcgi9PHNTP8VIF4SGW+E98U3s13LpVwBsmo0S2bKowvKttHwlTA3QXMFKC1FlpPUF1xG5qdeFYGINWIlEiBejuNDUPnirRkmbCMgQBbPiZE6x3gJLKt1rgBxAmQklcCcTjxA9Hv30hTd9sS1zd+/QsZGibDPkGX7zJHmEhgQEKD9WAOc+u+aOhpnKw7G6uy1e+HgNJMbhReZAAHAn773TPzpe8/E9tNTyj4y5DIYAuO05xNJvY1pg+fJvwkKUCOJKKw/A3jTR4XXT6BJAeqiBLYuhgCl+X8AQYAypAYTrqQAcQ+QJU6C5YbHhqHGKFsJEBcHMR6gKrGRt43UkRdJ4ApQVA3tahaYlATNW+K7LYHl2XE/JTWwJGUAYf4QttRepL9GCRCkY2juADD+NKLgxFFeDxoNuQQWVYCYd7Ps4KHdh0OPUY6WwI4QBWjR6dfVV1+N97///Tj//PNxwQUX4Oabb0a5XMaVV14JALjiiiuwadMm3HDDDQCAj3/843jLW96CL37xi3jXu96Fb37zm3j44Ydx++23A6CdOp/4xCfw+c9/HieeeCK2bduGP/qjP8LGjRuxY8cOAMAxxxwT2odikR7sxx9/PDZv3owjAkIBWi1lhlRpjT43HDFBhz1AWXcOAP2bTsZgABCLiqn5KJUDOdMiDUAD7GwX0niPMZA1YehaYKJlIHGztdoEJwpWnR6Yhxx60jghcuIOFCD63qQNTOSf0auNIZBCjpZ6Jl4A0EoBMgEiyfY+Wxx0CxlLR51J9yZTgGTjtC5KYPRz9/Xw/qWNwtBFxlC6AtSyDd73hQHyuUkXQAYnrms+iRcyJkpuAYOoAjN7AABzyGFzzFUfT4OeiMk4AYC90xURrJj34wnQYitAAHDC2gGcsDaGsCTh+F8D7gOdpO42sH4oi2NW5bFnqiKaGWIJkJSgHKcAhUpg7eR/9aIEtvpE+l2szwIHHqW3pfl/gNCQ1AFUhPmZpkITnPDQH2FHrYhd+FWhUgTErnVJk3uAtBgPUBWZrvw/QJDLEzWgd9QGz6fBS8NQq0wB6tYEzY/7sALUPAgVAHDHb+PfHngMW7Q/R7kRKJZz5SoKfHrAlguBvQ/SMtj6MyP7zwicRMDdhnR8RkpgPAn6hbESfELXzhPWFvH43pkmBahcTyBtS4xF9wBddtlluPHGG3HdddfhnHPOweOPP457771XmJj37NmDgweD0s/FF1+MO+64A7fffjvOPvtsfOtb38Ldd9+NM844Q2zzqU99Ch/72Mfw4Q9/GG984xsxPz+Pe++9F9ls/w28bUP2ANl5caXEr6zkPAeqAMkESFKAOhiECiCkJlRqVRBC4PkEGaYAWUeAAqTrGkbyFmZJeJEmCzjB8TRlu04VoMNuDprWrAANcgJUaq0AZdnB6xEdhHsvxp4CANRTOtaCEhhTftjVq1CAWKuszQgQMeQSGDNBixJYeP/yNp3JY5t60/PrFiutkVYeoBYlMGlA57RDn2/r6uaSSjFjYo6TWEaASqQQG3zGFaCJhLlAe6Yq4rFyqFISxsAJ0JHQwdiEtafTMrdTBvY/DCDwAaWVwPjJPwOnqYUYiJkF1gpNJbAuCJBpUxIEAK/cT3+2UoAMU4zoGdQqmK+5IIRgrubgeO0ARp79W+yY/j8AaJkmXNprxwSdrADVYHfl/wHkNngSyrDpRAHi6ocjKUA8E6hbBchmx72s3seOwajPAQceAwBs1cZDHqBGWUpUP+O36M8Xm31AcQpQiABFTNA8CZpzxgu2rRKkKKoAVZ30iJGlwpI8+1VXXYWrrroq9r4f/ehHTbddeumluPTSSxMfT9M0fPazn8VnP/vZtp5/69atTUFMfYXvB1kaeWZWHNhA1Z+5A8DaU0JlAqoABV+2jCOXwDr0ANlFEE2HRnzkvHlUGh40DWLRsXP9V4AA6gM6MB/el4VkFHGlJNOg7/sc8tgykm+6EuMlsHFBgJIXUNkc7Y4cD/vQM4IAJZqgAfjgBIgRH0GAqAeIXzGKUpXcBRb1AEUUoKxl4Nb/eC50rXmRNkQJbIEmaCdQDmuwcfKagvBMyChmpTRoNipkDrnYur8gQAkK0J6pCubA5mOBAI05cdFgkZTRI/2GrtMy2NPfomWwYy/GBVtX4VuP7GuvBKY1Qh4MjobndTQyoiclMICWwSaeEyfXlgoQQD+nxhwGUcFcjfp8HI9gRKNexgypwYaDct1DIYOOTNBcHRXhnoQIgl4lNtZ1qTBYkq/J9YmYDVbvYBSGzvxHDckEzTvCOg50Zd8H229WgKpxg1DHgpLWCOZxQCqBOYwANYwC7JPfAfzrJ4F9D9FzUj7ItorzAHlOsgIU7Yp803Gr8OjrMwCaPUDBGIyjXAFSiEFtBiDsS8U7KIQRmipATSZo6ctmNwICFAxCbdcErdPJzgBWa3OYrjRQbXjihJo5AkzQAG2FLyMbzrlZCAHiU5gb9OAvkXxT+QsISmBcBSlmkludTUMXV0m1oePpjePPAKCLbysTNGEeII0REt0wkZFM0AKyB4gRIF2j+xclQADwG6etw6+f2nxlblj0hGImeIAqDQ8W3NYeIGaAdvUMCPTY8hfAwhC5isdI3lxMECLQHgGqw0KDlRBQZ40AhMBm6qV+BKiXseA+IGaEvmDbKlhwA6UjTr21Ag9QLSaWINEEnYReKEBAEIjIs5rSxmBwsNEng1oZ83VXlMFG9KCUOYiKlATdgQlaECA+VqYh9q2GTNcmW8MIfEOyEboTBQigF6+OpDPwjrCki6NEsLWPN6uEPUB0n0IztQ4+IX4d0uZDCoxXmQEAOBbLdlpzKn3PuKrHEKcAhQhQ1ATdRIBWCwVOziECAkLUbwVIEaB+gPt/MkPB1VukEywXbYOXFCCrMSN+LwoTdPsLmsZaZEe0OcxUHNRcXxCgI6WMQMMQNTERvkEM2Jnuh11yBSjrzACgJ+J4AhQ+iIstJHQuO1cGttIbOugC85gXRxcmaBsZ02gmQHIQoh6+j+jJBC0KToAMNCtAnk/wZv9hPJ35T3hbfWfT/SGwEkODeZVOinkfAZ4GHf4+lUg+1pcx2soDNFUBoAkfEOF5NlIYm9lNZ9NSgBOg/Y8AtVkcuzqPY4sSqbHjFKDAAxSnANXjhqGmoWcE6PTw/9PGYHDwLCBUMFdzRAfYOisopQ5qZZQbLp0F1oEJmg93FgTICbqbarC7SoEGAhM0ADhSubUTDxBAVZSGRICqfpcKEPv8DI++Z3FdYGEFKJi/NYz58CywKj12fO7POunt9GekHV4oQK4vKijyNPgoAbIMXbzfAxkTp28cEsd6pYkArRAPkEIMhP9HaqWNZAGlKUBmXVKAOi2BAUJ1WoU5zFYdVOsuchpPlT0yCBAPQ6yxifA1ZDq/apLga6z7jQ2XLJE8TliTrABxtLpC4Z/TXHFb6Pa0EphoXRcKEAsyM03aBk8ipEYucUS6e+IUoCQEClAzAao5Hs7TX0RGc3G6+1zT/SGwk0yFEaBEBSgrKUAMc8jHLnrteIAACB+Qw65i5dEjR0IHYyyGt9CxGMQDXvsZNE3DRVvo63X1THPHFhB4gDQnUQHqyAQdJTwLVYA42i2BgXqA5mouSiwgb60ZkJVBVFCue2i4LjIiCLENE7SYb8feC1ae9TUDDoyu5oAB4eGkcQpQ2wTINEIm6LJH/y7TsQLEyr9+Axp8TMklsDgyISlAw1oZ8w1XkBitPgMA8DPMd3oiI0Av/yBoyACQYVlHhAQz0XzZ9xMzp26YRZhcsG0VDF0TClAlMo1eEaCVDDkDiENMhD8AIGxoy0SSoA32BQYQjMJo1wQNiDrvKq2EmYqDuhxv3s5iugTgYYjzLLCw6zEYDFGiMIc8jk8pgXGkdYEBwec0mzs2dHsttQRGH9P3wyUw7gFqVoCs+N+BkDrUCiYnQDEKUKXhCWN1UolMwOHTtpkCtC5NAQqrMiWSjyWVvAtsqtxo8iC5no/90yzdlylKjfkZdif9/rtEh30EpJgnItIO/5/fSI99LW4SPCBKYJkEBShkgm6rCyyqAHXpARo+FrAk8tTKBA1IClAZczVXKEBrDIkA/f/tvXuYHGWZ//2tU/d0z2SOSWYSkkDAsElIOCVLHOAnskQSjGIUUbgiJyPsavJbDruIIoffomyUFUSQ18juq667xCj7IrvyKl7ZgPCyxgABdTkKK5pAmISQzLmPVfX+8TxP1VPV1efu6Zru+3NduWbSXd1TU1P91F3f+3vftzKBiVTWe4EtwwTtV4CyahSAkreJaTE0VXE6MYhmiFnTcgy+ZaXAJBP0pFlpFZj792tDGqPJrPM5EaMwnGtGJuk03gSAbmUMtu0GHVqadWVXROHN/FXsfJh81/M6OUgTgZ8iBz0+EzTgpsHeeyy7yW6XWhzICEWqqfsAEXmQK8AEPgXI2wna2wdITbou/rJHYQBO4NWLMQwn0kgnpKZYITGSimaIYiJ8wo5W3AUa8DYTBFg1UrAC5N2uZAVIiQMd7sUgWagKTEywNkUKjCtAWiTQAyS6PwPw+IHYi8oJgPgoiYAAKJkxnRJb3Q6eFu8gFCA7goimYkFvcOqpQ54Iz5lAPPDuuScecYZivjvu/flvjySRtWxENBXjPKDKTHhTYElESr4oNYRjz2ZfX/kZkE05KTAtyAANSH2AghWgVNZCRCnDLFyLMniAeQhlFai9FA+QqwCNp9wAqM/nAZpIZ2FJBvvSTNDcEydUTX4+ZFTeOLAKj4l/HlhK+juUmsKK+DxAE44JusxzVVLm27laLzr55yhAB1/yVMX18J4/EymmAolRMlo772auGcDAiez7t93UmdxAMpW1WHsMT8dt7ywwAPjwSXNxdF8c606c49knvwla7HOhIpOpIMQrRhMzIQahyimwQiZorwKExBEA7ENZsJQ2HyIFxhWgdIpd0CwopeXdp4BengITpfC1VoC0eJfTuVSms9wUGDceJtKmWyIMVgafL2BzUmC8Ckws3k4KzB8ASVVgOQqQPyAqgMFVBSMgBTaZNp35YkaJClACERybpwIM8E2E52QjnYGN6VRVcQbT+n1A+3j6a15vDAmVvZ+bApsmAdB7zmHpotE3ged+4A6gzKfc6oUVoMBZYIXQDG+wXEVBgRMAxftKq0CTFCDZBN2tuB29OxWWArNkk20pa5HmM0ELf5rCjkk1KRanFJ4HQGkpACpZATI0jwdIDEMtOwBSVScgnNXG9uMIN0K7HiD+c4T/h88d7NV4F+5UFsmMhXabHXcjLo1zET2AeBUr+5GKU/2WzloYT2VhKNLaYebeKH32/cfhievPxtxuFrC5nah9CpDTvZoUoNYjUAHiAdD4EGBZ3hSYTwFSbBPdKi/TrsQDxH9urzKG4ck00kl2gUkjktOBt1EID9Bhk32QmAeomgDIGyj09c0M3K5iE3TaBGa+x3k8rbiKRs6+iBSYEwBxaVo3PH2ABIoc5Pj9IuWYoKNuE0z4ZhwlMm4KzCiqALllxv4RGDJBCpBV4Dx1fUDeO0vh/1nQG0eCp0TNhFCAXNN5pWNSpgQjBpx1Pfv+iTvcqexSo8Cc7QFEFBPpTG5AmjbLTIEB3qCn0hQYAPRzI3Qp/h/AowCZlo2DoyzI6ZKGOndiApPpLGyuKpiKXnjEBkd8NlRfCizNPX/VBEDi8ytM0EIB0lUl72fbT1T3mqCFGlTRWsZvYGbHWEAmfEA5jRCFinPUCgBAj+IqQPIcMKO9233vOVwBkszTgDQPLGtiLJl1g24gxwQdRD4FSOxzpSb1WhHiFaOJCfIAdcwGoDDpcvJQYQUIwCx9EoDtBkBleYBcE/TwZAZZrgBllHCoP4CoAgMOmWzRTtiRwFk3JeNTSvpnB3sX2gzVUwFSagoskc56FCCzgHxvq9xcaPkUIMcD5Ps76PkVICXIQJuHiOF6ZGzf3Vsi7abADDsDy9cB1wO/yCQQzWkkKdPRpjtVWwAwaUfRFs3v08nXDVoOgFIaez9LBEBcAUrZRrgVIAA45VLmoZk4CDx1N3ssXyAiGYDF51MmnTXL65cD+AKgKvp9HXcOS8kcd3Zp2/MAqItfePcPs6C1w85VgMQ6Z6olrkUiABJBuzgfIBSgKlJgvnlg5ZbAAyIAcj+zaRhQFW+VWek7xP5+s4UCxAOgnBSYMEAvPAsA0AWW8hoXARAPiJRYt/vejgL0O89cMNENOp21MJrMwEBhBchPPg9QS8wCI/IQpABpBtA+i30/9raTWgFyFSAAmK0nEEEWhsJPrLI8QCIFNobhhBQAqeEwQANuCuxgmu1TosYK0PyBYO+CoigeI3SxAMirALkBUFYt4KVyBjiyxUSUpau6EVgGrxX0AJWuAOkR96KSTntVlkQm68wXiyhZT+lvDnKn3QJBqV8BGkW8YGdeoQAd8nmA5AAorbOAwRZ9gLgClEKkoqGXU4oeAc7+Evv+EBuZUiwFBgCWPH+J4x0aWmKwIPv7qkmBzToe+MJeYE2JcxV5ANTNTc/7xaBha9TZpBOsDF4Ya60S1yLF7wHiwbmYBF+Nx8Q/D0yUwJcbAFlg/wAgDR1RvbL5ZOLv18cDoMOTPgUoorG0Ou9FJoz3M+xxKLAwkTK9k+CFCRpgvYBUgzXj5V3bAVkBsjCWzLpBNxDoAfKTrwrMNUGTAtR6BHmAAI8ROuZXgHwB0Ext3B2EClTkAerhKTCTL7DZEAVAIgX2ismq4/bas6vrGioFDqN2DMf2d+XdVE6DFasCi8sBUJ+bArMKpCUcBciXAtP1PFVgctDj8zKVpQBJVVKZtF8BshwFKIp04XEYTqfdaMG0E+sE7QbmY3l6AAnyNUMUHqD5vXFkeF8oJPnF0/EAGeVP2G4Eyz/OLjaCfJ9bVXUG+GYDAiDWCFG0riix+k0EPUZ7SemlgpQadAFuAKR4FaB4VgqAlEleBcZn3JXqbXNSYPxGkF+UEzVUgMQ8sFSZJfBsW3ZOZnV27CfstsrbeXAjdF/EqwCJKrB4RAfefY19PiMdTgpMg4UOJDGeyngnwcsBkB4BZi1m30s+ILkX0Fgy61WAAqrA/BTrA0Qm6FZkMk8AJJXCR3TVMaC16VpOCqxPnXQH2ukxQC3jRHJSYKMYnpADoPCkwNr5TKv/tE7FdT334qvZi6vqAySPk8jXBFHgUYCKLKBigU1kTKD7aKfazNbyX5REACQUIN0xQfMqMJ8HSJWDKc2fAivHBG3AtPldbY4C5HqAIsgiE1B55CCZoAsFHX4FaAyxgopaKSmwrMECIJWX8tpCjbKngQIEsM/pOTe7/89XBQbA5DckQQqQZxp8ySkwrgBVk/6qBH6hneGkwNhk+IjprQLLmLazFpWqAInO6JpPAUrY1XuA3HlgXg9QuQoQADx1/A0YWvl5DKGv/CaIAv7364mw3/XwhKgCk9QU4f8ZWM5mTPKgt0sZx7ijAPHj3ua7CQzwAckeoNFExs04AIF9gPy0O52g8/UBohRYa2HbwSkwIKcUXigeUVkB4qbJHnXCLYEvx/8DOAFQRDGRSYw6C6yphqePiqIo6Gk3ACj4r4m5BYeLloSklEygHQOd+X9XOQAqpgCJtNxkOgtoOhIdCwAAVoGGkv4Bjiq4CdrgJuhCZfA5HqAyTNCq4pgwsxm/AuSmwKLIFJ4HJqUZCgUdHVGvB2jUbi94QZoZoACNJjPO4Mf5vXGYEREAsRRYNi3tS9g9QII/+6Bzd55zEyRh8iDazgSnwMo3QTc2AOrgAVAiY6ILE55NhC8lmWRrml2yAsQCHTcFxo7VpMVeX80F1k2BeT1A5QQwQkF5se88vLnsswAqGIMhEAGQwX7XI0EpMOH/EWXtMVbp1cO7QedVgOTXSAqQ+EylsxbGkhn3nANKSoHF8ipAlAJrTVJjbjAT9wVAoqcGrxARH942Q1KAeJDUo4xXNAaDbR93LtBa4l1YfNEoZNxtBKIX0EF+QazGBC0HChljRsEcvJwCKybRelJgAF5ftBH/n7kMr8VOzv8ikcbi3Wt1iBRYBKqq5ChxqlwOrPnVodIDIADIigAo7b17S0h9gCJKpvA8MMcDFC3Y0bajTYcJDeM2u5CPIVY4BSYUIKkbtEh/9bVH0BHVYUXYDYDoZWKmpkkVmIyiABf838AZVwOnXJJ3M4v/3T29cThld4IG3BRYNRVgldDWDQCI2Qlo/FyXS+ABoItflEVLDrvEtUh8rjVfADRh18ID5DNB889EOUqjCJZSWausQaqB8ACoU/cGQB4TtFBv5pzEvvIAqFsZz6kCyw2AuBFa6gUkPlOVp8BcE7W4qbIsmxSglkWoP0acSZQy4g6NBzvnLJmNo7pjrNRYBE282V43xssfhCrD1acZ1ihSvBGiVSBt0wj6eF8YUZRQjQna01OkLU/pMWdGtHQTtFCLxPT41+d+BJdkboTJF/1AHBO0CcuynYuCmNZu+arxNEOuAvN7gMoLgDL8Z2cyvgAobXlSYNlCHiCpDL7QYi6CHdENeswuzQQtK0Cy/wcAbJ4yMrI8AEqXpkaFjt6FwAduAzpm5d1EfB6VgDvttGkhInqylNq7S6w3U60ASaX+Ig2WqwDxAEgoQCWm450yeF8ANGbWQAHSfGXwmcpM0ABLISX568seg+HsEDsfOjUW+B7xlcHHdSkFNserAHVjHGPJLCYmJt3rRk4AtIx9HX2TTYaHTwFKZd20K1BSCkw+/mI/k1lXDSIPUKvhGKAD+tA4ARBbBG7/6HI8dcPZrL24TwHqgmSCriAAUqRmiBMT7G7MDskYDIFQgATVBEByoKDHuwtuK4IaXVWKqgqi5fvTbxzGSCLjfLjbCr1O8gBlLEtSgNjP9Stxil5AASrDBA24CpDpC4Am0xlnYSw1BZZAtODFQFMVxCOaM79rFKWZoMdTWUcil/0/AKDw4DWSZees8IxkFKaeNROWMDcHpMDYKIwKTdBTPTRW0501SgQ6PQqv4uProKNK8Bu9UtciUSGp+zpBj5vsPKvOAxSsAJVngnYVFKEAtVWsALG/W7vGPUCTaViWzfyHANoTbwKpERYQC0MzD4C6+KiR9IQ7RSCnB1VbF9BzDPueK0mygjXmL4MvQQGSvaziMy03Raz4WNQICoCmmqBBqALxoZckbydVI6JtrgDNwLjUBLH8OzpFaoaYSrC7sULG3UYgSuEF1XiA5CAiNqO34LYiBdYe1YuWqx43qwPH93cgY9rY+fIBR44uGKw5KbAsMqbtBkBc6bGkO3rLVmDoUtCg+lNg5RnX3QDI21wvK82Di6C0FFiihLSTPA9stIgC1BHVHX/EoTG2uPoDIJUraxFzErBtx7+WUcIVvNcCkQYKUoBSWbPyMvipVoAATzdoAOgWChC/4EaRdv4BKNnYrfDPimglIYLzCYunwKqpAlNFJ2ihAFVggubrQCpjuZPkK/YA8fEe3Kt3ZCLjVVMO8/L32UvdGyXHAzSGiXQWWd5BPa13BBfO+HxAsgI0mvR1gi7BAwS4KpAIfGT/T6NvWigAmmqCmiAKdK8C5EFE21wBmmGNuWMwyjVBA24pPMaQ5J2gYYQrAPIrQFUFQJJyMqM7v/EUcBWgYgZowdplrIv3oy8MOXd5BffVad9vIpO1AgIgqVwdmuNFkF8rUMtMgTml1T5fiZlyUxJRpVgVGE87FSmDB9gxFONMRtFe8IKkKIqjAr3FS6X3HmZfRQCkxdhdqwYTyEy6AVCIWjjUDLEe+FINlmUjY9oVNEIUAdAUe4AAR20QCpDjAeqeD4BdBGcg4ab1SlSA1DweINEHqBrfoFsF5leAyjBBe1Jg5StIHrgCFFPZ3100NhREDv+efSNSWYDkAZrAeMqElRgGAJiRPDYA30wwef9HExlvHyDb9EyPz0e7rxu0SIV9VdsKPPAJ4ODLRd+jXlAANNXkqwAD3A99UItxsQjykRkd1mhlYzAEPADrU8bcO8xSpfQpIkcBilR+usoKUGfRAEgoQKUtdGtPYEHpE79/xxnkWajSQ5EVICkFJuR82QCagQ5Nk+6SqlSATD6HzPRVgfm7DWfTBe7uJAWo2N1wR5uOfzHPxWPmyfhP89SiKYljZ7Jz+S//5Vn86Jm92PsuC8yEB8iIdSBr85+ZHHUqpMLUw6pm8BsS1XdDlDYtKLAQESXJpX5uRaPVjhIGmNYaRwHiKTAxCDU+0/HkdSoTzgVWKfG8FhWSwkfnOTc1tarKQKcTtM8DVHEKzPEAVWeCjlpJZxTHW0fY79tmqE5lpKeyUDJBjyczrNEhACvfCJY5+RUg1gjRN0ewlHEY0WAF6DS8ALz2CyCd2+l8qmisBbsVcRSgoBQYX8iyvpy/mQVsfkfOU2DtsgJUUQDE0kA9GMMRsNcrIZkEL+jxBUAVV08A6Iy7vgetiAeoK8YWVf9csHwsmTMDR/fF8ad3J/GLF1kLg1JSYIrtTYFBzQ2AstCcqdRs5/1dosv7CJtcAbJ8HiDbtwiZJQVA0aJ/k46ojiesk/CExapSipnKb/3wUvz19ufxwlujuOH/cctxF/Sxv188qmMcMZZCSbkBkNmEAZD4PGo+BcjTBRooPQV26mVMRVi6vkZ7WAbOPDAW+MziXaER62HPJUfQiUkprVdqHyCeAvMpQAlEqxudA2kWWFUeoKAqsOoaISrZBHriERwaTzlKaTyiswpjwHs94Ot8F8YxkTKhpvgImXxFGqIS7NDvgUwixwOUEwCZKQCFPWV+BUgEQjHnBn6KPWkSpABNNaIJYpACJFJQ/qhaXgB5CqzdGsOquXzhqySnL5mg23jeXQ1ZANTrT4FVsaD1dkrHKN/dD+d/HT8Ta07ox2fOXFjSeyuKgrXL2N9FLEiFAiCRjlOsrCcFJgKjHAVIraUCxD1APgOjlfYrQLnGW3en5CqwwkuIP+ApqgDN6sDDnzsDX/rgEkdFMzTF6dskm6qRHHX8ctmQ+ddqgSIUINOnAMk9gIDSU2CxbuC0KwtWntUNnwI0k08odwIg+BWg8gIg5zMkRqPYkaoHbeaUwVfSCNEISoFVpwAhk0RPnK0Db3IFKGZoQJofU/l64EmBZaHzBqJqLE8n/BlzmCpnm8DBl5zPdzqoDB4ocSCqdx6YSIG12fy11YxlqRIKgKaagh6gPFUf8knGFSDFyuK0maI5YuVl8L3KmLOYqpFwBUCsEaJLrUZh5JR/+uhsM/CdS1bivOVzSn7785Z5ty2sAPHnbBNZ02TT2eV91OUASHOqKNhr3blCgG9OWAlkebPLTNJbhix8PQK/QuTdVuoDVORiMMMXABVTgABWfXPl+47FjmvPwvqT5+LaDxzvBIHxiNRcMTXiXPCskPWwqgXihkSzvH8LzyR4KLnz4cKITwESE8pZANTNnsMkotwDpJSY1hN9sHT4FaCIk3qpFMPXCXo0Kcy7pb+vkwKTTNCVN0IU14dJRx13FSA5AJKuB5IJ+t2JFOIW20aP9wT/DEXx9AOSPUATybS7VglKCIDaffPAJtNZqJCqGBvhSeNQADTVFPQA5VOA+ImiqKx1vljsR95kXytKgbkmaNEBWI2E6y7a7wGqXR+gwgFQJZw0rwtzutzjV9AD5Jigs0jL1VgiMJLM6Blbd8pxBZbiLsBlB0B8JlE2OeZ9whd0BzXfE9hOGXzhPkAA8wDJlFOWPL83jrsvOgWfe787Yy0e0TAGHqgnR52CgbD1sKoF4oZE9wVAqYzpbYJYyWDNqcanAHUJE3S8VwqOJt2bMaPMMnifB6jYoN5SEJ87kQJ78wjb96N6Sr9RrG0jRB74Z5OOOi48QCwA4sc0QAHqUiaQzFhOAKq3d+f/OZIPSKhdE2kTWfmmiHsJS5kIH6QAOVMMAEqBtRRnfwlYd6cbZcs4AZDv4iMCIo0vdjyvi5F97GtVJmg3BaY18EQMQq4Ci2iqNxVULnIDwToEQIqiYA03QwOF1SrVad5melNRPL0l3/1moTkt+QWmFACppTbBE6/V2eKYTXgDIMXnOzMLBECQ52+VUAUmU01ZMsDSoM58sdSoYxAOWxfzWqDxQDhip51SbIArQEqZXaAbjRTkAECnzS/WcgoMbgqs9ACIbafDZB1TpfRstQGQKIM3uQK0j1ckzi8jAIoENUKs2APkZgj8ClAsrwLErhXdGAdgOwGoUqhRaz+vInvnVWdf3x1Pef0/YoZdGQpQIu0qQKLrPKA0tPiGAqCpZtEHgD//DNA1L/c5pwrMd/ERUbbwe/CoHuMH2NcqPEBdyqRTTaZHwxUAtRmas4hVNQgV8CpARTxAlXLeMjcAKs0EbXpnconHDa8HSNe8AZAtB3NlNkK0+eJopbwBkOoLgOx8C5tlQuGetFKqwHI8QFV2fvWkwJKjUgVjuNK3tUDjn8co0o56AFQ4CLXR+PoAdVj8/PN4gCadAEgrsSWHRwE1Mx4TdKltLPLhdILmCtC+I96u5KUge2icRogVV4Hxn5tJoJfbA/bLJuhABagbAKArFjqQyD8HTEYER+lx5/P97nja6/8pIwCKGV4FaCJlSlMM2huqYFIAFCZ8naAdZAUIcAMgQSUKUKwbtsL+/AMKa3tuhMwDBLgqULUVHeV4gCpl5TG9mNlRfH9F92bVyiIrp8CEMiSZ0XP6AMGbAvOboovCF0c75fUA+Y22eT1AUqosq7UVVeVm+FJg1V6UmAmaHR87NerstxWyFg61QCiybUrGUQ+ACueANRopyNFgImblU4D4gOASFSBPEYCVcTxhSURK8psVQpNM0GPJDIalobylIqfAqlaAHBN0wlkXnUGo+UzQRsy5OehWxvNPgvf8HDcTIfb/0HjKPedUwz3vypgILzxAiYyJuFCAGmiABigAChdiEbeyrPRdIE4yPU8AVIkJWtVgRrsBAHOVd/mPD5cCBLg+oKoM0IAbAGnRujV81FQF/+f8E/DB5QN436L8lTZOFZhtwsxKARAPSFXpopaFnpMC8wZA5S3yKj9XVHG3KPY9RwHKkwKTAqBSBlbKAY+iVN/6Ph7RMM49QGZixA3cmjAAEh6gNp8CVNEk+EbDL7hdyqR3Dlhbt1cB4qk9tdRRGEawApS0qw+AZBO0qLbqiRtlBfFuFVgtFCARAE0G9EiTPEBCnXGedOeBlaQASak2jwIkz54Tn/0KqsAmUlIKrMG2C+oDFCbkRTybBDQe2AifiEjjcFnTocLW9na8D0geRgfvJxQJYQAkct1VGaABVykpMgi1Wj504lx86MS5hXdFKEB21vEAZaDD4FKwbkSQtVXoioV00RRYeQqQ2sYWRzXrVn1lTAtR27uQ5TVBCwO0HUHEKP6z5YtQ3Ki+9X08ojtl8ObkiNsjJ2QtHGoCDwKiCFCAhAdomqXAupSE2wU62sVSuJJBehLs91FLvEnRJaXINtNOSrQmKTDVNUH7h/KWilsFZlbfB0jKEPh7pMUNNVgBAlgANLYf3cpE/knwMlIxjtjXsVQWs50ASFaASpgILxQgqRN0VYO8awgpQGFCvuuRI+scBcg3y6rCk0j1VaKFMQXWy/tdVB0AiS64Qd6rKcaZYA3TSYGZcH+/iK4iBT6ywg5IgamVK0B6jJ0rWsa9C09kTMQU30KW786ujDlggLcMvtqyZICpbJMqW+Ct5IhTIaWEbIxLTeCpC78HyNMIscw+UA2DX3BnYIIbcuHeyEkl8uUqW7qmOZ3BLamysRZVYI4CZFrYd0QYoMsNgAJSYBVPg3dTYP4eaR2GxTIHQG4AFHeN0F1KKSkwEWglPB4/T5PKfH7VAIJmgTlTDBqcAiMFKEyoGlMqpFw2gAAFqAYeIACqbyCrGrIqMMBVgKpOgc1eDFy0DZh5fA32qjpcD5AJi99BWYr7+0V5ANSOFKsCy1GAJOWlTAUoEmcLn2G6ClAibaINvoAnbwBU2iR4gVwGX21jOkFa7wBswE6MQrPYQqo2eCGtCzyoa1PSHgUolTWlQGGaBH7cWNuOBPoU1ozPWcckBWgC/PcpUdnSNIV9RmDBTAw7txFJRGpmgs5argI0r7e8m0Rxk5C1bEcBqXwavBQA+RSgLlW6gTH8ClA3gDI8QE4KLImI1ILDMUFrEfdaVMJE+KBZYH0hSYGRAhQ2nOi7kAJUAw8Q3Inw7s8O32La114jEzQALF4HzFxU/ftUicrHV6gwnZlccmk7U4DY752G7h2FAV8KrEwFKBpnKbCI5Q2AYvAtZPmkbX7HlyyhCzTgS4FVWQIvyPDUsJI8As1mgUHYxrjUBH4hakPa6SIM+EzQZbZBaBhS6vm9PZIBGvB5gMpTtnRVQYbfx1sJNuYhAwMW1Ko9QE4naMtyegDNK1cBktSeMd5IsfJp8PwctzLoiXnfo1MV14i23MpQfpz7lSNoE6nTUhQgM4Wo7t58ec65MkzQ/llgE6msmwLzB2tTDAVAYUOcWHJjOqcKLEABUtTK7wL988hCeDd5dB/7gPR3hm/fKsXbB4gtipY/BWbzFBg07zBUwBv0lFkFFm1nC1/Ucs+vybQp9eXglKAAldLQTU6BlTpcthjZCAvitMl3nMeaOQCKIuN0EQamaR8gzXAudhtP4OdzjgIkpcBKVIB0VUWWf3bMBFOWUgp7bfUBkEiB2RX1AALgUVDE5PaqR2EAaFfSnvfuUAvMheTHeb5yEABgQyncCkS6DrRJI1cMMXxXDoBKSIEFKUBhmAMGUAosfAR1g3b6APGTLi55gCIzKu+j4B/HEcIA6LxlA/inS1di5TF5WrdPQ0QKTIOJVJACpLkeoExAFZg3BVbeRzjewS427UiwVIqucQ+QN+BR8t3ZlTEJHvCmwGqlAJmRGcAkYKQOO49pIetiXhMMtwrMrwBNuyowgAU6mQngyB/Z/8U6xgOgNiWDDpsH5mUoQCIAspPeAKijyoBb7gRdSQ8g8R66qiBr2RjlAVDFPc2k9VnJJtHTbuDAKPucdsh9dfxwz+gCHgBljY4cVTnfz2mTvIFucGpIVWCVdoKmMngiiKBu0IX6AFVYAQYgVwEK4V20rqlYvbQf3fFpIvWXgDvA0UI6HeABMjQpANJg+EZhRCLSRa9MBUgEQHEkHUk+mclNgSn5UmBSmXEpKbCYoUHEb7VSgKwIu3tVbBYUpGwDEaMJ7+V01wMkK0Cp6dgIEXDTYCIAEutYZAZTJQDMVPi08lJngakKMk4AxF6b5OpptV3HhQn64FjS6bdzVHf5a6T4nFi2+H+FnwNFkYzQk55O+aKSt5ACJAIgM1KkD5qmOypz1HbXASPIBF1BH6DJdJaqwIg8OAGQlALL1wkaqC4Aag9/CqwZ0aQBjhmuAMmVXVFZAbJ1+CvHO2LS36nCPkDtSDl3pHIKLCX8NWYeaZunwEoZhAqwESEiFVErBcj29TlJwqi8tDjMyCmwfB6g6aYAAcDwn9hXsY6pKhQeHHWLKqUyAjtTJDJ4d/NJ7p+rVQrsjUNsn/o7oxVVo0Z9r6nqXPVMhHcDIDelFKQAseM8k5vP7VIawfJAS1aGKzVBi8/9ZMaEbdveWWBkgiY8GAEpsEIKUIUGaADTwgPUjIgUmK5YyGSFAuQzQfO7WEvRofhTnFrlKTCxQEaVDEYneTpLKoNPGVxdMTPBry8zBQa4PqBaVYFp0RmwbPeYJBFBpMoGi6HEcE3QHg/QdE6BAa66La9j/otyGeX9wgQtUmAJq0YBEFdexbytckvgBf6Ap6qWHnmaIcbBb5gLBEACNVZCAOTMoXMb8kbkACjf3MoAhAJk2+xmi1JgRH6kLpwOThUYP+GNuBsMVSMhSgFQBgZQKC9M1AxN8gBleQrMVnwmaH4XK3uDHOS0V9mjMNzzZXKMpwykMvi0wRZHNa8HSDRCLM0EDbg+oFr0AQKAWNTAONxgvZShrNMSfhfepmSQTHvL4KddI0QgN8gpFACVpQDx85AHQBO2CIBq0wfI5qmrcv0/An8AVBMFiHuABG28HUTgDbHsGQWgxbuL/xx+7kWk4ghPJ2hxLSqhEaJc9v/uONs+VsizNIU04aoxzXHc9bICJPoA8ecURcqfVxMAuSborDqNFtJpjmiEqMNyOkHLxuaoriLN72oDAyBZ9SkzBQY9woJdAIlxFgCx6cy8IzVXgFSrNo0QAfdOvFYKUDyiuQNRwRSg5kyBSUNxU27bgmmvAAnkhq7+6eRlKEBZ/hlRUlwB4gFQrTpBC8qtABP4bxQqLoMHPB4guRmiU9VZwAMkKCkAchQgdx0INkEX9wCpquI0pXxnnG1PChARTJC06O8DBEgBUBURdCQOW3SbjTVhI7mworoKUIZ3grZyFCA3BZb7+sobIQJAUmV/8+QEv2BkLOeOLMMNxmqRFFiyxEaIgHshqpUHKCaNw2D70qQKkFSUYEqjSdKm5e3KO12oswKkpNn5nEQEqlJ981R/A9J5FSpA/nOzYhM04Hpm0hOecRhuAFQ8Bab4g80g+HXIyGuCFgpQ8QAIcD/774zxqjXRuLEVPED33XcfjjnmGLS1tWHVqlV4+umnC27/4IMPYvHixWhra8Py5cvxs5/9zPO8bdu45ZZbMGfOHMRiMaxevRqvvfaa8/wf//hHbNy4EQsXLkQsFsNxxx2HW2+91am4CTWBAVBA0zNxUlfjAYLbDLHU2TtEDVDZAuhVgHxl8PwuNqsEBDhy0FOuAgQgo7FFJz3BFKCEVAWWjXSzH1FMAbIjJS/kx/cz0/J7Ztem4oMpQG5w0LQBkGY4/aFMSQFKTcdGiECRAKjb+1wZfkShkqop1mAxgQjaIwHeuTLJVYCqT4EZmgKtmnl44hgmRzweIMPMMwcMAIwYLDmgLMUEzYNvw8pjgg5q11IAkY4UClB7qzRC/NGPfoTrrrsOt956K5577jmcdNJJWLNmDQ4ePBi4/a9+9StcfPHF2LhxI55//nmsX78e69evxwsvvOBsc8cdd+Cee+7B1q1bsXv3brS3t2PNmjVIJlnQ8Morr8CyLHznO9/Biy++iG984xvYunUrbrzxxnr/utVjBARA2UIKUJUXFZEf1iuTd4kK4AqOBmkavM8E/YJ9DADgDe2YvK9nb1K+AuQEQHx2UiKddTxAZpQtjpqd52ahAhP0F89bjCeufz/ed/yssvc1iHhE8yhAKbtJq8DgpqYtyROYysqNEKfRjUtOANSd/7kyUmAiAHIUIDtatQEayFWA5pc5BkMgp7wqHoPhvEE3+5oc8VSBGWaBFBgAVS54KakKjJ13mplyAjaPCbqMFBjgKkCHhALUKh6gu+66C1deeSWuuOIKLF26FFu3bkU8Hsd3v/vdwO2/+c1vYu3atbj++uuxZMkSfPnLX8app56Kb33rWwCY+nP33Xfjpptuwkc+8hGceOKJ+MEPfoD9+/fj4YcfBgCsXbsW3/ve93Duuefi2GOPxfnnn4+//du/xUMPPVTvX7d6pDksDkJmlC92ooS9UEfPUhA+IFKApg6u2uhSACQrQFFdw7+Y5+Kk5P14KnJ67uur8QABMHU+TT3BAqBUOoUI7/Jq8QVWt/IFQHIn6NKWD11TnY7etSAe0VvDAwTA5AqPlZZSYNO9DB5gDVzl9awGKTAtzc7nJCKI16DnlCEFQJqqYKDCbvSyUlqV/wdwg8bEsKMAGZoCVQw3zndDXCjdGIS4Ic4mnI7T7TpvxVCmCRpw/X9CAQpLJ+i6rhrpdBp79uzB6tWr3R+oqli9ejV27doV+Jpdu3Z5tgeANWvWONu/8cYbGBoa8mzT1dWFVatW5X1PABgZGUFvb2/e51OpFEZHRz3/GkJgI0SfCRoAVlwOLP4QsPzj1f08cWcwne4kpztyAMQ9QHIgI5SVEXTkyPBs2yqqwABYfJE0uQKUTbrqgmiSplnFPECR6hfzCmEKkC8FpjVhGTwAU2WfS1vqC5b2NEKcpimweE/+54CylE0nBZZxU2DVGqABbwpsbnebUxZfLnJwXpX/B5BSYMNY0BdHPKLhuFkdQLpACgwoPwCS2rGIz3mHJo3CKFcBino9QM7onWZOgR06dAimaaK/v9/zeH9/P4aGhgJfMzQ0VHB78bWc93z99ddx77334i//8i/z7uuWLVvQ1dXl/Js/f37hX65eBI7CCEiBHbUCuOgBoO+46n6eGIhKAdDUwT1AmmLBMoUC5K0CE/jHYLAX8m0VtbLWBXzRsVNs0bT44mlDcbr15k+BuWXwkQovCNUS81WBpWyjYcFYvRHeDTstp8Cm4TR4wHvh9Q90lp7LKpGyxvuY3Cen8wAoaUeq7gINeFNglfp/AF8AVO15KlJgiWF0thn45fXvx7999nQgzQfM5g2AuqX3KEMBygQoQLo8C6xEDxBXgA5xBShqt4ACFAbeeustrF27FhdeeCGuvPLKvNt98YtfxMjIiPNv3759U7iXEkGdoB0FqA53e8IDFMIxGE2LpABZwgOk5SpAAILvOoVaVIH6AwAKN84r/IJhcYNtVmuDYuRWf3jwKECNUV2CFaDmXMqsgKGTaY8HaDopQN3u9wUCINUoL63nr5RMIFIbD5B0c1FdAKQFfl8RIpDhYz9mz2hjapcTAOVJgcm9gMpSgJJO0BbT5BRY6aMwAMkDNJ6CCimF28wK0MyZM6FpGg4cOOB5/MCBAxgYGAh8zcDAQMHtxddS3nP//v04++yzcfrpp+P+++8vuK/RaBSdnZ2efw0hqBN0kAJUK2Yez752zav9exPBSGXwlpWbAitZAarA/wMAahurylLSQgFiAZClxaDxi49uF+8E3SjfTWAfoCZVgGwtNyXuKYOfro0QCwZA5alapu9zkES06kGogE8BqtAADXhVn6o/M44Jetj7eK1TYAEKUFyr3ATtVIGNpdwxGEBzK0CRSAQrVqzAzp07nccsy8LOnTsxODgY+JrBwUHP9gCwY8cOZ/uFCxdiYGDAs83o6Ch2797tec+33noL73//+7FixQp873vfgzpduhwHdYKupwK0+EPA5f8vsPrvav/eRDCOAmS5LQ4CPEBAbiUK25YHQOWOweBobewuUcvyRZOntSw95lx88gdAcifoRgVAAX2AmlQBEn26FL8CNB1N0HLBRoEAqNygzq8AJe3aKEDyxPRKu0AD3qCn4knwAskE7UEEQPnaopQdALnKoxgzE1MrN0ELBSiZsdwmiIra8BRu3UcoX3fddbjsssuwcuVKnHbaabj77rsxMTGBK664AgBw6aWX4qijjsKWLVsAAFdffTXOOuss3HnnnVi3bh22b9+OZ5991lFwFEXBNddcg6985StYtGgRFi5ciJtvvhlz587F+vXrAbjBz9FHH42vf/3reOedd5z9yac8hYag3Go9FSBVA445s/bvS+RHeIBgQrFMQAMUOQWmFVOAqkuBGTF2IdKzvLcMT7faRgwaT4VGSkmBNVQBklJgzToKA3A+86oUAKWmaxWYHmGdfzOT3i7QgPeiXGZazw5Igc2scRn8vLCkwCQTtIdiKTARACkqq8ArhuEqQOJz3qZUNgsM8HaBj4vJ9UZ7WV6velD3AOiTn/wk3nnnHdxyyy0YGhrCySefjEcffdQxMe/du9ejzpx++unYtm0bbrrpJtx4441YtGgRHn74YSxbtszZ5vOf/zwmJiZw1VVXYXh4GGeeeSYeffRRtLWxP8qOHTvw+uuv4/XXX8e8ed7Uji0Gu4QVp/wwoA/QdJK7ifzwFJYBEzp4ZYUUzOiaClUBLDu3GZtn2wpTYJE4WwCj1iSypuUENbYegxaJ8n0rlgIrfRZYrYn5+gCxafDNWQUmLkTyaBKvAjSNTNAAu4BnJuurACFSk67jRq1SYHodUmCJYTakTFHY16IpMB5wRjtLK5yQAhxxcxFV+VqllzcNHvDOAXQUoAanv4ApCIAAYPPmzdi8eXPgc7/85S9zHrvwwgtx4YUX5n0/RVFw22234bbbbgt8/vLLL8fll19eya42ngDToyMzTifDI5EfyQOk8/47ii+dFdU1JDJmcApMeIAqaIIIANF2drGJI4nxVBYaV4AUIwYtIhSgPAFQ1u0E3SjVxd8HKIWI52LVTAhTupr1BkDuMNRptiZEO4Gxt3MDoEgHUydsq+x1zlL9ClC06kGogJu26YjqmNVR+c2npxFitYUDIgVmZdjNSCTObpAtrs7kC4BEuxO5GqwQAQpQ1KMAlWuCdn/vWEjmgAFTFAARZWCQAtT0SB4gRwHyBTMRXWUBUFAKrEoFSOceoA4lidFEFmo2AaiAEolDi/AqsCAFyMw6wXg5jRBrTczwmqAzarTqsQdhReHrgWb5U2DTcBYYAHQvAA69CvQc7X1cVVlwlBwue52z/SZouzZ9gOZ2x3DTuiWY1xOr6vzypsCq/MxEOgBFA2yTHatI3FV/gPxVVfNPA066GFj4vtJ+jtSORQQvbgAULV8B8qTAwtEFGqAAKHyIBS1DClDTwhdsVbGd+TpqQAAE5CmDF2pRhQqQ8ADEkcRoMgPVTDoBkM6rwCLIwrZt78IvtWZgHqDGpJ00VUFadxdPU5tmaaAyEKZ0zZ8C06ehBwgAzr8HePu3wNFn5D7X1sUu6mX+TqZvXl6yRmXwAPCZ/3Vs1e9R0z5AisKOU+IwS4N1znX9P3pb/sIIzQA+urX0n+PciCdw+RkLoWsq+iYU973K9gAFpMBIASJyCOwETQpQU6G6gUMbH0Kq+O5ihRE6MLVTpQIk7rzakcTe8ZQzB0yNxmFE+RBExUQ2m4VuSBcX7v+xoCAFo6HGY9uYAfCiFEtt3s+Fxn0SYjSJbdusDF4EQNNtTeicy/4FIXxAZQZAtq8YoFadoGtFTWeBASyNlTjsGqEd/09thg0D8IxkGjyuD4PH9QH/KKVdxd/IygCWVdRXJI8mccrgQ6AANWnpxDQmsBO0UICm2WJHBCMFLm081aT41D2xaGpBC0uVfYCcAEhJ4uBo0pkEr0faoUdcNSWT9t3d8RL4pB0BoDR0/lYkEsGEzT4P5nQLAsrA9WSlkDEtpLIW946JrrxN9LuLAKjKFFiiRsNQa4U8pqUm/aqkgagAihugKyGoIa+ciZC9ZyWUwnsUoBClwCgAChuBnaCFAkQpsKZAbnqosMVD1fIoQIEeoCpTYFE3BTY0kkKML0hqJA4j4la7pHMCILcEHqjRYl4hcjNES6+8QifsaFyRa0MGqayF/3ln3C2BB5ozACo31Z/TCDHi8Zw0Gk8foFopQIDbC4gPga2pAmS4CpCDGaAAASWlwWRTephM0BQAhQ1/J2jbrm8fIGLqkSR7oQCpujeYEYumFhQA8QCm4jsoKQV2YDThKEAwYjAMA6bNfqaZSnhfxxWgSXCfUAObD8YjGn5snoXfWMfiT8Z7GrYf9UYoQG1KGsmMiUd+97ZrgAamXwqsEELZKLcMXs31AIU1BVZbBWiYfa2LAuR6gBxkK0aZClA8yAPUKmXwRBnIHiDbdqNugBSgZkFVYUGBCtvxAJVlgj72/cD7Pg8sOreyn8/vFDXFxvDIqOMBQiQORVWRhoEY0jDzKUC2UIAad5cdj+i4K/sJ3IVP4MQQLKT1QuVm1CgyPADa7ypAilpxN/BQUgMFyLKZPy1MKbCaNkIE3OPkKEB1CIAKKkAGM2NrUXZzXsI4jMAUWIPngAEUAIUPj7SYYiazoOeIaY2laFDtLNqU4ABILJSBJmg9CvzFlyr/4ZL0PDZ6BDEl7XlcBEDZtF8BcueAsX1srAIkaNYxGACcapw2pPHMHw9j3+EE3mMI/0+TVb/NXMS+dh9deDsfsgmapWeVcClAtRyFAUgDUYfZV1EFlm8MRiU4ClDARAJxI67zAKgEBSjm6QMUjknwAAVA4UP2M2STgGW6/28mubvFsaAByDodfVXd5wEqlAKrFlVFWosjYk5ifGzEkwIDgDQvK85mgk3QCUShKHnGdEwR8oLatGMwAOemp01J49/2vAkAeP9xncAf0XyK8KmXAgMnAnNOLO910s1DAhGoSo0CjRrhHYZaCwWom33NMUHXwQMUaILm1yGnF1BxD1BEVxHRVKRNC+1kgibyohkA+IUlm3SjblUvrYU5MS2wFLYQihSY5pP93TL4+vzNTZ2rPZMjblkqV4AyYBeU3BQY+3/SZnPAGtl8UFaAGqlE1R3dTYH96n/eBQD8xaJu/lyT3RCpGjBvRdnmfr8C1B7VQ9UYs6aNEIFcE3RKzAGrgwfITLs34XIKDAiuWC6AKIV31xsKgAg/iuLtBk09gJoSmwdAUeEB8gdAwgNUJ5XF5I0EY3bSScOJ8y6jsH0x076FTVKAGj17SzZVNrUCxO/E25CGbQMzojpWHMXXh2YLgCpEkQKmWnWBriXeFFgtFaBh9rUuJuiAKi//tajMifDCB9Sh8u1DkAJr4pVjGiNPhKcu0E2JGOAY5TOdNF8KLFrnAMjmi2WHkswpS3UUoJwUmOsBanTQ4U2Bhafkuebwu2wRKH/ghH5EIY0kIDx9gMJWAg/UeBgqEGCCLjIJvhIMyYqR4VYMW4zt4dciTbpOlYD4zLb7PIeNhAKgMOJ04UyQAtSkiAAoXwpsoIudA7M662R05YbJOJI5HqAM9wBZ+TxAduPmgAnaWyYFxhUgftH48Ilz3TvyZjNBV4ji8QBFw6cAGTVuhOiYoOvoAVI1t11HNuGtRhZrlfhaYgAkPrPtang8QOE6UwiGnFsVYxNIAWoq7BwPkNf38FdnHYeT53fjzEUz6/LzVb5Ytiu5KbAsT4HZ/oVNaoTYeAWoVVJgrgeoO27gjPfMBP6HVGEPkgcoYdduDlit8CpA0yQFBrBzL5VhCpA89d2pAuPXqZInwrO/S5hmgTXxyjGNkbtBkwLUlFiqNwBSfMbP9qiOc5b0181ro/GJ8O3ITYGZ/IJi502BhcED1GIKENJYe8IAC/ZoTfCg6Lkm6DChqwpEJrumZfCZSTaNPV0HEzTg7UknK0Ai4NTKVID8JugQKEBNvHJMY+Ru0NQFuimxeQpMU2z2QKVzvSpEj3UCCE6BCQXI8pe3SrPAGh10xFumDF6kwDLYeMYx7LEsrQkefFVgYUuBKYqCvo4oVAXojtdAtYt2ut8nh+vjAQKk65BUjKMabjWyOP9KNEELBchtvNr4AChcZwrB8HeDBpqv50eLI1JgDmp5pb/VorexcRozlITUmZUpQFmVp8AywSmwUJigZV9FUzdCZGuBBguLZonSZAqAZGQFiA1CDZcJGgDuv2QFjkymMbOjBn8zVQOiXUBqhBmh65UCE6XwmYQb5MjXIccEXbwPEOAqQFHL23ajkVAAFEbEwianIGixayr8E6ynfKQBN0H3YMx9jCtAYraS7c/te8rgG2yClu7yGzmSo+7IjVEzCdaDhYYje1D8fYAi4busnbKgp7ZvGOMBUHLEDYBq2QkakKqRpRSY7Dtzni9dAdJgwhCjXEKgADXxrdM0xmlDnnRPLlrsmoscBWiKF22++MxURtzH+HlnqiIAz2+CbrQHKNYqozAK9WOhKjAAgCJdlBMh9ADVBdkIXbcUmKwABQTdTgqsNA9QT9xwDdAAKUBEHuTI26kCIwWomchRgKY4BSYWy5nKKABmfNa4CmWqecyN/AIchjL4lvEAKQoLdLJJJwB1U2B0UwT4GiGG0ANUF+Ru0PXoBA34TNBBKTCxTpSmAH1i5Xwo40PAc2CDfENwTWvilWMaI3eCDjrxiGlPbgA01QqQCICYAmRpbqrFEgFQ3hRYCEzQhpQCa+YACMgdOeCowo2/gIQB1dMJOoqOEHqAao5ohjg+5DYorEcZPMD70QVch/TyPECzO9uw6fQ57D+RDhbcN5gmXzmmKXInaEqBNSch8QD1gSlAluQ1sfmFVfFXd8hl8A0eNtkyw1AB6YaIK0DUCNGDaniHobZUCmx0v/tYrWdrFVOAykyBAQAy3K8UgvQXQAFQOJE7QVPFR3OihsMDFONNEC3pYlpMAUrakYb7buRKn6YPgPxFETQex4MiXZTD2AeoLogU2Mib7Kseq/1NlCcTEWCC1sozQQMA0mwNCcMcMIACoHAiS96kADUnIfEAOUh3ZDZf5NQ8naCZAtTYNEObZ8J2k6c85KII+SspQAC8KbBECIeh1gVHAXqLfa1HRZVzI54sbIIuMQXG3ksoQI2vAAMoAAoncidoUoCaklwP0BRfxHMCoIAUmJUvBdZ4D5CqKk4voKZXgOSGdADdFPlQda8CFLZhqHVBeIBGpiAAyubrA1TeNHgApAARJSB3gqaeH02J4g94tKlWgLwLpiIrQPxcU3M8QOFJgQFuJVijg7G6I6fEAbop8iF7gJIhHIZaF2K8r9D4Afa11iXwgHsdyuRpx+I355cCX0PIA0TkJ8h8Rotdc+FPeU11Ciw6w/vjI3IKjJ1rquwBMjOAlQXAFaAGm6ABoK+DLcbd8Sk+dlON7leARAqM1gQAUKWLchiHodYFkQIDnxRQFwVIMt8HmqArUYDqVLJfIS1wpkxDZNMjDT5sThpeBu9dgNSodEfGzzVVToGJOzewu+ww+G6+dsGJeOntUfxZ/4ziG09nxN9q8jD7SmXwHjRfCqw1FKBu7/9r3QUa8CpAQcb7MkdhAJBSYOEIgBp/G0fkogf0AaKKj6ZC8VdsTHUZvB51Rl4AgCbn5LnioHkCILbIWVCRhh4K380pC3qwYdXRUELQT6SuzFvJvv7PY+wrpcA86JqKrM3Ox7Ta+CadU4KjAHEaogBJ7VpKhVJgRFFkdz0pQM1JoxUgALa0CClSAKTy808LUIBSShSA0hoXmbDwZx9kX994gnX9pWnwHnRNwSTYsTCNjuYPiAHXBC2opwcomyrSB6icFFidBrdWSAtohdMQT/8FanvfjOQoQFPtAQKgRmewgYqApwpMnGveFBgz4LIAqAVKz8PEzOOB3mOBw39gKhDdFHnQVBW3Zi7HHOVdjHcMNHp3pgZ/CqyeClC+TtAaKUBEPQjsBE2LXTOhhEABUmTfgNSXQ+F3frodEADxO+0wpMBaBkVxVaBXf05pcR+6quAn1v/C/2Wubw0DNMCqRuVeOnUJgORp8AVM0OUEQFQGTxSFOkE3PbkKUAMUFXnRlBQglZ9/upVxn08OAwAmFbYdpcCmmOPXsq+v/cJNI1AjRAAsBSZomQAI8KpAkToUAnimwQdUI4vzr5xRGKIKjBohEnmhTtBNjzzB2lT0xgwGlH0Dsh/IYOeaRwFKHAEAHEEneykFQFPLgvcy4+vku8CRN9hjpAoDYAqQoCUGoQpkI3RdGyHKCpCUqi9zGjwANwUWEg8QrWJhhDpBNz1y+35LadCi7QmAJAWIf6/bkgI0+S4A4AjYnSYpQFOMZgCLzvU+RmsCAOYBErRHWkgBko3Q9QgoHAWohsNQKQVGFMUgBajZkVNgltKgRTuaLwBiC5sRFADZ7DVkgm4Af3ae9/8UAAHwK0AtFAB5UmB1HoURNJGgIhM0zQIjiuGRHkkBakbUMARAHg+Qe0em82DIQBqweadZHgC9a7EAiFJgDeA953irBemmCEALe4A8KbB6lMHLChC/Gaq2D1ArKkD33XcfjjnmGLS1tWHVqlV4+umnC27/4IMPYvHixWhra8Py5cvxs5/9zPO8bdu45ZZbMGfOHMRiMaxevRqvvfaaZ5vDhw9jw4YN6OzsRHd3NzZu3Ijx8fGa/251QQRAVtY9YSjf31TIKTC7YQFQPgWInX8qbGf8hehCfMgSChAFQFNOWxdwzBnu/8kEDQDQJAUo3koeIFkBqkcn6GIjmeQUmLhRKkarlcH/6Ec/wnXXXYdbb70Vzz33HE466SSsWbMGBw8eDNz+V7/6FS6++GJs3LgRzz//PNavX4/169fjhRdecLa54447cM8992Dr1q3YvXs32tvbsWbNGiSTbkvuDRs24MUXX8SOHTvwyCOP4Mknn8RVV11V71+3NsgLW2qUP0Z3e82EJwXWiAowIK8JWo9Ii5y4u+MK0EGTB0AhmAXWkohyeIDWBI4ueYA6yANUO5wb8Yw7iDfIBA24ClExQjYLrO6r2F133YUrr7wSV1xxBZYuXYqtW7ciHo/ju9/9buD23/zmN7F27Vpcf/31WLJkCb785S/j1FNPxbe+9S0ATP25++67cdNNN+EjH/kITjzxRPzgBz/A/v378fDDDwMAXn75ZTz66KP4p3/6J6xatQpnnnkm7r33Xmzfvh379++v969cPZ4AaIx9JQWoqdDCoADl8QBpEen8cwIgpgA5HiCthe60w4Qoh1c0t1FdiyMrQJQCqyFGwI14UAoMKH0eWCvNAkun09izZw9Wr17t/kBVxerVq7Fr167A1+zatcuzPQCsWbPG2f6NN97A0NCQZ5uuri6sWrXK2WbXrl3o7u7GypUrnW1Wr14NVVWxe/fuwJ+bSqUwOjrq+dcwVFU60bi0SHd7TYWiSwFQA5ogAsjrAYoYBjI2D3BMrwJ02OZVYKQANYaeo4GP3Ad8+G7vBaqFMTQyQde1EzQAJEUAJAU98veljMMwM0xNAlojBXbo0CGYpon+/n7P4/39/RgaGgp8zdDQUMHtxddi28yePdvzvK7r6O3tzftzt2zZgq6uLuff/PnzS/wt64Q/v08KUFOhSSmwxgVAwQqQrilIgQdo2STL7ztVYCwAimgUADWMUz4FnHppo/ciNJAChPoEQPKNeJKPzJFTYKrqmvJLMUKLBp5AayhA04kvfvGLGBkZcf7t27evsTvkr/qiKrCmQglxAGRobOI7ANaGITPpKEGHMQOGpkBVW2DgJDEtkD1A7a1qgq5HCgxwVSAxMzDfdSlfCuzwG8A49/sKA7SihaaCsa4B0MyZM6FpGg4cOOB5/MCBAxgYCB5aNzAwUHB78bXYNn6TdTabxeHDh/P+3Gg0is7OTs+/huLP74fkhCFqhBz0hC0FpqlICwXITDnqj61FkUCUegARoUKOxVtSAdJj9RulI9KsjhfVN7RZXJeCUmCTh4FvnwH8418wE7Xs/2lE5/sA6hoARSIRrFixAjt37nQesywLO3fuxODgYOBrBgcHPdsDwI4dO5ztFy5ciIGBAc82o6Oj2L17t7PN4OAghoeHsWfPHmebxx57DJZlYdWqVTX7/eoKKUDNjScAmvpJ8AAkE7TiOb8MTUXKlqRtHgBl23oAKFQCT4QKRVEcH1BLdYKeuQhY8mFg8HP1+xnCimFb7KvfiiGPbfJz+A+s8eHIPuC326U5YOHw/wBA3c+W6667DpdddhlWrlyJ0047DXfffTcmJiZwxRVXAAAuvfRSHHXUUdiyZQsA4Oqrr8ZZZ52FO++8E+vWrcP27dvx7LPP4v777wfATvZrrrkGX/nKV7Bo0SIsXLgQN998M+bOnYv169cDAJYsWYK1a9fiyiuvxNatW5HJZLB582ZcdNFFmDt3br1/5drgNziSAtRceAKgBikqUa5yRmd47sgMTcEYV4DsbBIKrwDJRnsBUBNEInzEIzpGEhn0tDfoZqIRqBrwyX+t78/I8aL6rkN6AQVoTPLb/upe4Px72Pch8f8AUxAAffKTn8Q777yDW265BUNDQzj55JPx6KOPOibmvXv3QpVyuKeffjq2bduGm266CTfeeCMWLVqEhx9+GMuWLXO2+fznP4+JiQlcddVVGB4explnnolHH30UbW3uH+uBBx7A5s2bcc4550BVVVxwwQW455576v3r1g75xNMioZEMiRoRBgWo91jg9P8N9Cz0PGzorgcom07BSA0DADLRHgDUBJEIH3//0eU4MJrEnC5qDVBTcm7E/SmwAh6gccmmcvh/gP/+N/Z9SLpAA1MQAAHA5s2bsXnz5sDnfvnLX+Y8duGFF+LCCy/M+36KouC2227Dbbfdlneb3t5ebNu2rex9DQ2eAIjSX01HGDxAigKc+5WchyOa6lSBmekEDJ4CS0dY4zXyABFhY92Jcxq9C82J34uaY80oMBFemJ9Vg5W/P8/VqpDMAQOoCiy8eAKgFpJ1WwUp6FFC9vc1JBO0mUk6HqCkwRQgSoERRItQzIohrlNBE+GFAnTqJex1ogdQiBQgWsnCStDMFaJ5kAKgeCxcDe00VZEUIDkAEgoQLRsE0RIUq0YuNBFeKEADy4GTLnIfD5EJmlaysCL1ZSEDdBMiBUCGES4FCAAyjgKUAhJsDMak3g2AukATRMtQVAESKbACClBHPzD4v93HQ2SCppUsrJAC1NzIlV+NMkEXIKuwfbKkFNikzhQg6gJNEC1CqQpQoRRYRz8w63h3kK88xLXBtFDThGkGmaCbmzCYoAuQVdhCxwIgpgCNa2SCJoiWIp/p2f+83wRt21IAxMdSnfc1oH0m8OdX1n4/KyR8Ky/BkAMgGoTafMhBT8hM0AAPgGzAyriNEMcU1jeITNAE0SIYRRQgPY8ClBx2ewO18wCoewFw/r0138VqoJUsrJAC1Nx4hgqGT1ExVWkYKleAxjQWAJEJmiBahGKNEMX//X2AhAG6rSvXRxQiaCULKwYpQE1N2D1AKgu6lcQR5+5uBDwFRiZogmgNZAVINXIb8uZLgTnpr+DZm2GBVrKwQgpQcxNyD5BQgPQJ3s5eb8OEzQLxiBY+xYogiDrgn0iQ73l/CkwoQML/E1IoAAor5AFqbkLuATK5CdqY5AFQvA/pLBuISAoQQbQIsgIUdB3S8nSClivAQgytZGGFFKDmJuwKEF/YIpN8IYv3IpU1AZAHiCBahqIKUJ5ZYGIQKgVAREV4FCAKgJoOjwcofAGQrfIAKHmIPRDvQ4orQFQFRhAtQrEbcS3PNHhKgRFVYRSJvInpTdhTYHyxU2CzB2K9bgqM+gARRGtgFJlJKQIkfydoSoERVUEKUHOjhrsMXihADvE+SoERRKuhFxnJJMZaJI54HycFiKgKOeghBaj58HiAwqcA2X65m1JgBNF6FGvHMvN49vWdV7yPCwVoBpXBE5UgR96kADUfYfcA+Rc7uQqMAiCCaA2KKUCzl7Cvo28BiWH2vZlxusdTCoyoDI8CRAFQ0xFyD1DOORfvdRQg8gARRItgFDFBx7qBzqPY9wdfZl8n3gFgA4oGxHrrvYdVQQFQWCnWf4GY3nhSYOELKKzAAIg8QATRUuhFTNAAMHsp+3rwJfZVHoKqhnutCPfetTKkADU3IfcAKZQCIwiiWB8gwE2DOQHQ9DBAAxQAhRfqBN3chD0F5vedSSZo6gRNEC1CKZmI/hPYV5ECmyYl8AAFQOGFOkE3N6oKKPzjF0ITdM4U6FgvUhleBUazwAiiNShJAeIpsAMvArbtEs34zQAAEqhJREFUTYGFHAqAwgr1AWp+ROATwgBIkc85PQZE4kibpAARREtRyo34zOOZ4Tk5zEZgOCmwcJfAAxQAhRfNcBUC6gPUnIQ6AJIWvngfACCVIRM0QbQUquoGPvlS9UYb0Hcc+/7gi5QCI2qAorjRNylAzYkIfELoAVLlcy7OSlmpESJBtCCiFL7QjbhjhH4ZGKMUGFELxEWIFKDmRJS/h1ABUiPeAMi0bGQtNheM+gARRAtRyo34bG6EPvASKUBEjRBdOEkBak7CnAKTqz+kEniAFCCCaClEAFRIqZZL4akMnqgJvQsBKEDX/EbvCVEPQpwC0+SSV2kQKkAeIIJoKcTNUKFqZFEKf+BFIDPBvp8GClD4bj0Jl4t/CIy/A3RTANSUiAaIIVSADF1DytYRVbIeBUhVAF1VGrx3BEFMGaUoQD3HsIxFNsH+b7QD0Y6671q10K1cmGnrAma+p9F7QdSLWX/G/F09xzR6T3IwNBUp8AXPNwleUSgAIoiWwVGACnhRVY2tZ4IZ4Vd/AFKACKJxXLQNSI0C7TMbvSc5GJqCNAwACSDWI80BIwM0QbQUpVYjz14KvP0b9v00SH8BFAARROPQI4AevuAHACJ5FCDy/xBEi9HWyb5G2gtv17/U/X4aGKABCoAIggjA0FT8zjoWPfok4v3LkDpMPYAIoiU581pgxlxg8brC282WAyBSgAiCmKYYuoorM1fjlJlteKhjFva9/hYAYPYMaslAEC3F3FPYv2LMnn4KEN3OEQSRg6EpsKFi3GLGx/9+cwQAsOyorkbuFkEQYWXGABDrYd9PEwWIAiCCIHIwNLY0ZEzW/fmF/RQAEQRRAEUBFgyy72ctaey+lAilwAiCyEEEQOmsBcuy8eJbowCAZXMpACIIIg/rvw0c/gNw1KmN3pOSIAWIIIgcDI31+smYFvYensRYKouIrmJRf/ibmxEE0SBi3dMm+AEoACIIIoCIkwKznPTXkoEZjjJEEAQx3anbanb48GFs2LABnZ2d6O7uxsaNGzE+Pl7wNclkEps2bUJfXx86OjpwwQUX4MCBA55t9u7di3Xr1iEej2P27Nm4/vrrkc1mnecfeughfOADH8CsWbPQ2dmJwcFB/OIXv6jL70gQzYoIdLKmjRd4+usE8v8QBNFE1C0A2rBhA1588UXs2LEDjzzyCJ588klcddVVBV9z7bXX4qc//SkefPBBPPHEE9i/fz8+9rGPOc+bpol169YhnU7jV7/6Ff75n/8Z3//+93HLLbc42zz55JP4wAc+gJ/97GfYs2cPzj77bHz4wx/G888/X69flSCaDoP3+0mbFl4UBmjy/xAE0UQotm3btX7Tl19+GUuXLsUzzzyDlStXAgAeffRRfPCDH8Sbb76JuXPn5rxmZGQEs2bNwrZt2/Dxj38cAPDKK69gyZIl2LVrF9773vfi5z//OT70oQ9h//796O9nZXZbt27FDTfcgHfeeQeRSPCskhNOOAGf/OQnPYFSMUZHR9HV1YWRkRF0dnaWewgIYlpzcCyJ027fCUUBumMGjkxm8NPNZ2L5PAqCCIIIN6Vev+uiAO3atQvd3d1O8AMAq1evhqqq2L17d+Br9uzZg0wmg9WrVzuPLV68GAsWLMCuXbuc912+fLkT/ADAmjVrMDo6ihdffDHwfS3LwtjYGHp7ewvucyqVwujoqOcfQbQqwgNk28CRyQwMTcHxA2SAJgiieahLADQ0NITZs72dIHVdR29vL4aGhvK+JhKJoLu72/N4f3+/85qhoSFP8COeF88F8fWvfx3j4+P4xCc+UXCft2zZgq6uLuff/PnzC25PEM2M7jM7H98/gwahEgTRVJQVAH3hC1+AoigF/73yyiv12tey2bZtG/7u7/4OP/7xj3MCMj9f/OIXMTIy4vzbt2/fFO0lQYQPUQYvIP8PQRDNRlmNEP/mb/4Gl19+ecFtjj32WAwMDODgwYOex7PZLA4fPoyBgYHA1w0MDCCdTmN4eNijAh04cMB5zcDAAJ5++mnP60SVmP99t2/fjs985jN48MEHPWm1fESjUUSjNOeIIADAUL33RsuOIh8cQRDNRVkB0KxZszBr1qyi2w0ODmJ4eBh79uzBihUrAACPPfYYLMvCqlWrAl+zYsUKGIaBnTt34oILLgAAvPrqq9i7dy8GBwed97399ttx8OBBR9HZsWMHOjs7sXSpO4jthz/8IT796U9j+/btWLeuyARbgiByUFUFuqoga7EaCRqBQRBEs1EXD9CSJUuwdu1aXHnllXj66afxX//1X9i8eTMuuugipwLsrbfewuLFix1Fp6urCxs3bsR1112Hxx9/HHv27MEVV1yBwcFBvPe97wUAnHvuuVi6dCkuueQS/Pa3v8UvfvEL3HTTTdi0aZOj3mzbtg2XXnop7rzzTqxatQpDQ0MYGhrCyMhIPX5VgmhaRC8gTVWwZA4pQARBNBd16wP0wAMPYPHixTjnnHPwwQ9+EGeeeSbuv/9+5/lMJoNXX30Vk5OTzmPf+MY38KEPfQgXXHAB3ve+92FgYAAPPfSQ87ymaXjkkUegaRoGBwfxqU99Cpdeeiluu+02Z5v7778f2WwWmzZtwpw5c5x/V199db1+VYJoSoQP6D2zOtBmkAGaIIjmoi59gJoB6gNEtDorv7IDh8bTuODUebjzEyc1encIgiBKoqF9gAiCmP6IFBgZoAmCaEYoACIIIpCumAEAOHl+d2N3hCAIog6UVQVGEETrcMfHT8TvD4zjlAU9jd4VgiCImkMBEEEQgZw4rxsnzutu9G4QBEHUBUqBEQRBEATRclAARBAEQRBEy0EBEEEQBEEQLQcFQARBEARBtBwUABEEQRAE0XJQAEQQBEEQRMtBARBBEARBEC0HBUAEQRAEQbQcFAARBEEQBNFyUABEEARBEETLQQEQQRAEQRAtBwVABEEQBEG0HBQAEQRBEATRctA0+DzYtg0AGB0dbfCeEARBEARRKuK6La7j+aAAKA9jY2MAgPnz5zd4TwiCIAiCKJexsTF0dXXlfV6xi4VILYplWdi/fz9mzJgBRVFq9r6jo6OYP38+9u3bh87Ozpq9bzNBx6g4dIwKQ8enOHSMikPHqDhhPEa2bWNsbAxz586FquZ3+pAClAdVVTFv3ry6vX9nZ2doTpawQseoOHSMCkPHpzh0jIpDx6g4YTtGhZQfAZmgCYIgCIJoOSgAIgiCIAii5aAAaIqJRqO49dZbEY1GG70roYWOUXHoGBWGjk9x6BgVh45RcabzMSITNEEQBEEQLQcpQARBEARBtBwUABEEQRAE0XJQAEQQBEEQRMtBARBBEARBEC0HBUBTzH333YdjjjkGbW1tWLVqFZ5++ulG71JD2LJlC/78z/8cM2bMwOzZs7F+/Xq8+uqrnm2SySQ2bdqEvr4+dHR04IILLsCBAwcatMeN56tf/SoURcE111zjPEbHCHjrrbfwqU99Cn19fYjFYli+fDmeffZZ53nbtnHLLbdgzpw5iMViWL16NV577bUG7vHUYZombr75ZixcuBCxWAzHHXccvvzlL3tmJLXa8XnyySfx4Q9/GHPnzoWiKHj44Yc9z5dyPA4fPowNGzags7MT3d3d2LhxI8bHx6fwt6gvhY5RJpPBDTfcgOXLl6O9vR1z587FpZdeiv3793veYzocIwqAppAf/ehHuO6663Drrbfiueeew0knnYQ1a9bg4MGDjd61KeeJJ57Apk2b8Otf/xo7duxAJpPBueeei4mJCWeba6+9Fj/96U/x4IMP4oknnsD+/fvxsY99rIF73TieeeYZfOc738GJJ57oebzVj9GRI0dwxhlnwDAM/PznP8dLL72EO++8Ez09Pc42d9xxB+655x5s3boVu3fvRnt7O9asWYNkMtnAPZ8avva1r+Hb3/42vvWtb+Hll1/G1772Ndxxxx249957nW1a7fhMTEzgpJNOwn333Rf4fCnHY8OGDXjxxRexY8cOPPLII3jyySdx1VVXTdWvUHcKHaPJyUk899xzuPnmm/Hcc8/hoYcewquvvorzzz/fs920OEY2MWWcdtpp9qZNm5z/m6Zpz507196yZUsD9yocHDx40AZgP/HEE7Zt2/bw8LBtGIb94IMPOtu8/PLLNgB7165djdrNhjA2NmYvWrTI3rFjh33WWWfZV199tW3bdIxs27ZvuOEG+8wzz8z7vGVZ9sDAgP0P//APzmPDw8N2NBq1f/jDH07FLjaUdevW2Z/+9Kc9j33sYx+zN2zYYNs2HR8A9k9+8hPn/6Ucj5deeskGYD/zzDPONj//+c9tRVHst956a8r2farwH6Mgnn76aRuA/ac//cm27elzjEgBmiLS6TT27NmD1atXO4+pqorVq1dj165dDdyzcDAyMgIA6O3tBQDs2bMHmUzGc7wWL16MBQsWtNzx2rRpE9atW+c5FgAdIwD4j//4D6xcuRIXXnghZs+ejVNOOQX/+I//6Dz/xhtvYGhoyHOMurq6sGrVqpY4Rqeffjp27tyJ3//+9wCA3/72t3jqqadw3nnnAaDj46eU47Fr1y50d3dj5cqVzjarV6+GqqrYvXv3lO9zGBgZGYGiKOju7gYwfY4RDUOdIg4dOgTTNNHf3+95vL+/H6+88kqD9iocWJaFa665BmeccQaWLVsGABgaGkIkEnE+UIL+/n4MDQ01YC8bw/bt2/Hcc8/hmWeeyXmOjhHwhz/8Ad/+9rdx3XXX4cYbb8QzzzyDv/7rv0YkEsFll13mHIegz10rHKMvfOELGB0dxeLFi6FpGkzTxO23344NGzYAQMsfHz+lHI+hoSHMnj3b87yu6+jt7W3JY5ZMJnHDDTfg4osvdoahTpdjRAEQ0XA2bdqEF154AU899VSjdyVU7Nu3D1dffTV27NiBtra2Ru9OKLEsCytXrsTf//3fAwBOOeUUvPDCC9i6dSsuu+yyBu9d4/nxj3+MBx54ANu2bcMJJ5yA3/zmN7jmmmswd+5cOj5E1WQyGXziE5+Abdv49re/3ejdKRtKgU0RM2fOhKZpORU6Bw4cwMDAQIP2qvFs3rwZjzzyCB5//HHMmzfPeXxgYADpdBrDw8Oe7VvpeO3ZswcHDx7EqaeeCl3Xoes6nnjiCdxzzz3QdR39/f0tf4zmzJmDpUuXeh5bsmQJ9u7dCwDOcWjVz93111+PL3zhC7jooouwfPlyXHLJJbj22muxZcsWAHR8/JRyPAYGBnIKV7LZLA4fPtxSx0wEP3/605+wY8cOR/0Bps8xogBoiohEIlixYgV27tzpPGZZFnbu3InBwcEG7lljsG0bmzdvxk9+8hM89thjWLhwoef5FStWwDAMz/F69dVXsXfv3pY5Xueccw7++7//G7/5zW+cfytXrsSGDRuc71v9GJ1xxhk57RN+//vf4+ijjwYALFy4EAMDA55jNDo6it27d7fEMZqcnISqepd5TdNgWRYAOj5+Sjkeg4ODGB4exp49e5xtHnvsMViWhVWrVk35PjcCEfy89tpr+M///E/09fV5np82x6jRLuxWYvv27XY0GrW///3v2y+99JJ91VVX2d3d3fbQ0FCjd23K+exnP2t3dXXZv/zlL+23337b+Tc5Oels81d/9Vf2ggUL7Mcee8x+9tln7cHBQXtwcLCBe9145Cow26Zj9PTTT9u6rtu33367/dprr9kPPPCAHY/H7X/91391tvnqV79qd3d32//+7/9u/+53v7M/8pGP2AsXLrQTiUQD93xquOyyy+yjjjrKfuSRR+w33njDfuihh+yZM2fan//8551tWu34jI2N2c8//7z9/PPP2wDsu+66y37++eedCqZSjsfatWvtU045xd69e7f91FNP2YsWLbIvvvjiRv1KNafQMUqn0/b5559vz5s3z/7Nb37jWb9TqZTzHtPhGFEANMXce++99oIFC+xIJGKfdtpp9q9//etG71JDABD473vf+56zTSKRsD/3uc/ZPT09djwetz/60Y/ab7/9duN2OgT4AyA6Rrb905/+1F62bJkdjUbtxYsX2/fff7/necuy7Jtvvtnu7++3o9Gofc4559ivvvpqg/Z2ahkdHbWvvvpqe8GCBXZbW5t97LHH2l/60pc8F6pWOz6PP/544Npz2WWX2bZd2vF499137Ysvvtju6OiwOzs77SuuuMIeGxtrwG9THwodozfeeCPv+v3444877zEdjpFi21JLUIIgCIIgiBaAPEAEQRAEQbQcFAARBEEQBNFyUABEEARBEETLQQEQQRAEQRAtBwVABEEQBEG0HBQAEQRBEATRclAARBAEQRBEy0EBEEEQBEEQLQcFQARBEARBtBwUABEEQRAE0XJQAEQQBEEQRMtBARBBEARBEC3H/w90JPUDqL1w0wAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(out.mean(-1)[25:25+125])\n",
    "plt.plot(out3.mean(-1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "dd064657",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c9cb0e2e",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "ce900c11",
   "metadata": {},
   "source": [
    "# Clustering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "bae5870a",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T03:01:18.370136Z",
     "start_time": "2023-09-22T03:01:18.366427Z"
    }
   },
   "outputs": [],
   "source": [
    "import os\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"0\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "6e588f40",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T03:01:24.869006Z",
     "start_time": "2023-09-22T03:01:18.371428Z"
    }
   },
   "outputs": [],
   "source": [
    "import gc\n",
    "import os\n",
    "import tempfile\n",
    "import time\n",
    "import random\n",
    "\n",
    "import funcy\n",
    "import numpy as np\n",
    "import tqdm\n",
    "import torch\n",
    "\n",
    "from suno_utils.tasks.data_loader import load_audio_mp\n",
    "from suno_utils.tasks.mert_25 import (\n",
    "    SAMPLE_RATE,\n",
    "    EMBEDDING_RATE,\n",
    "    encode,\n",
    "    encode_files,\n",
    "    preload_models,\n",
    ")\n",
    "from suno_utils.utils.s3 import read_from_s3, download_s3_files, upload_s3_files, check_s3_file_exists\n",
    "from suno_utils.utils.text import write_jsonl, read_jsonl, get_file_ext\n",
    "\n",
    "_ = preload_models(checkpoint_filepath=\"/home/tony/Data/MERT/mert_test_8x_400k.pt\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "51949a6b",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T03:01:34.969184Z",
     "start_time": "2023-09-22T03:01:24.870845Z"
    }
   },
   "outputs": [],
   "source": [
    "p_read_jsonl = funcy.partial(read_jsonl, allowed_keys=[\"s3_filepath\"])\n",
    "metas = read_from_s3(\"s3://suno-data/datasets/bundles/v2/music_sample/metas.jsonl\", read_f=p_read_jsonl)\n",
    "random.seed(6006)\n",
    "random.shuffle(metas)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "e0c74dd8",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T03:01:34.975767Z",
     "start_time": "2023-09-22T03:01:34.970629Z"
    }
   },
   "outputs": [],
   "source": [
    "# load first minute of each file\n",
    "def _load_audio(filepaths, sample_rate, max_duration_per_file_s=None):\n",
    "    with tempfile.TemporaryDirectory() as tmp_dir:\n",
    "        t0 = time.time()\n",
    "        if filepaths[0][:2] == \"s3\":\n",
    "            # if filepaths are on s3 then load them to a temp dir first\n",
    "            tmp_out_filepaths = [\n",
    "                os.path.join(tmp_dir, f\"audio_{n}.{get_file_ext(filepath)}\")\n",
    "                for n, filepath in enumerate(filepaths)\n",
    "            ]\n",
    "            print(\"  downloading audio...\")\n",
    "            confirmed_downloads = download_s3_files(\n",
    "                filepaths,\n",
    "                tmp_out_filepaths,\n",
    "                chunksize=100,\n",
    "                n_cores=16,\n",
    "                joblib_backend=\"threads\",\n",
    "                silent=True,\n",
    "            )\n",
    "            time.sleep(5) # make sure things close\n",
    "            local_filepath = [\n",
    "                filepath if b_confirmed else None\n",
    "                for b_confirmed, filepath in zip(confirmed_downloads, tmp_out_filepaths)\n",
    "            ]\n",
    "        else:\n",
    "            local_filepath = filepaths\n",
    "        download_duration_s = round(time.time() - t0, 1)\n",
    "        # remove Nones\n",
    "        safe_orig_idx, safe_filepaths = zip(*[\n",
    "            (idx, fp) for idx, fp in enumerate(local_filepath) if fp is not None\n",
    "        ])\n",
    "        print(\"  loading audio...\")\n",
    "        t0 = time.time()\n",
    "        audio_arrays = load_audio_mp(\n",
    "            safe_filepaths,\n",
    "            target_sample_rate=sample_rate,\n",
    "            max_duration_s=max_duration_per_file_s,\n",
    "            num_workers=32,\n",
    "            force_threads=True,\n",
    "        )\n",
    "        load_duration_s = round(time.time() - t0, 1)\n",
    "        # merge back into None list\n",
    "        out_audio_arrays = [None]*len(filepaths)\n",
    "        for idx, arr in zip(safe_orig_idx, audio_arrays):\n",
    "            out_audio_arrays[idx] = arr\n",
    "        assert(len(filepaths) == len(out_audio_arrays))\n",
    "        if len(filepaths) >= 10:\n",
    "            assert(np.mean([arr is not None for arr in out_audio_arrays]) >= 0.5)\n",
    "    return download_duration_s, load_duration_s, out_audio_arrays"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "eb78c40a",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T03:34:10.662091Z",
     "start_time": "2023-09-22T03:01:34.978245Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|                                                                                                                        | 0/20 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  5%|█████▌                                                                                                          | 1/20 [01:31<28:49, 91.01s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 1/20: 21.9s downloading, 4.8s loading, 62.7s encoding -- 2.1h processed\n",
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      " 10%|███████████▏                                                                                                    | 2/20 [03:12<29:13, 97.40s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 2/20: 31.6s downloading, 4.8s loading, 63.1s encoding -- 2.1h processed\n",
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      " 15%|████████████████▊                                                                                               | 3/20 [04:55<28:14, 99.66s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 3/20: 29.6s downloading, 5.4s loading, 65.5s encoding -- 2.1h processed\n",
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      " 20%|██████████████████████▍                                                                                         | 4/20 [06:32<26:20, 98.77s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 4/20: 25.8s downloading, 5.0s loading, 64.8s encoding -- 2.1h processed\n",
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      " 25%|████████████████████████████                                                                                    | 5/20 [08:11<24:42, 98.82s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 5/20: 24.1s downloading, 5.0s loading, 67.9s encoding -- 2.1h processed\n",
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      " 30%|█████████████████████████████████▌                                                                              | 6/20 [09:47<22:49, 97.85s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 6/20: 26.8s downloading, 4.8s loading, 62.5s encoding -- 2.1h processed\n",
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      " 35%|███████████████████████████████████████▏                                                                        | 7/20 [11:25<21:10, 97.73s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 7/20: 25.3s downloading, 4.9s loading, 65.4s encoding -- 2.1h processed\n",
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      " 40%|████████████████████████████████████████████▊                                                                   | 8/20 [13:04<19:40, 98.36s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 8/20: 29.5s downloading, 4.9s loading, 63.6s encoding -- 2.1h processed\n",
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      " 45%|██████████████████████████████████████████████████▍                                                             | 9/20 [14:36<17:38, 96.20s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 9/20: 23.3s downloading, 4.6s loading, 61.9s encoding -- 2.1h processed\n",
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      " 50%|███████████████████████████████████████████████████████▌                                                       | 10/20 [16:10<15:57, 95.78s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 10/20: 23.9s downloading, 4.9s loading, 64.2s encoding -- 2.1h processed\n",
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      " 55%|█████████████████████████████████████████████████████████████                                                  | 11/20 [17:43<14:13, 94.82s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 11/20: 23.4s downloading, 4.6s loading, 63.0s encoding -- 2.1h processed\n",
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      " 60%|██████████████████████████████████████████████████████████████████▌                                            | 12/20 [19:22<12:49, 96.18s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 12/20: 26.7s downloading, 4.9s loading, 65.9s encoding -- 2.1h processed\n",
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      " 65%|████████████████████████████████████████████████████████████████████████▏                                      | 13/20 [20:56<11:07, 95.42s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 13/20: 23.3s downloading, 4.7s loading, 63.8s encoding -- 2.1h processed\n",
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      " 70%|█████████████████████████████████████████████████████████████████████████████▋                                 | 14/20 [22:34<09:37, 96.25s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 14/20: 23.8s downloading, 8.1s loading, 64.5s encoding -- 2.1h processed\n",
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      " 75%|██████████████████████████████████████████████████████████████████████████████████▌                           | 15/20 [24:23<08:20, 100.12s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 15/20: 32.5s downloading, 6.6s loading, 67.6s encoding -- 2.1h processed\n",
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      " 80%|████████████████████████████████████████████████████████████████████████████████████████▊                      | 16/20 [25:59<06:34, 98.67s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 16/20: 23.4s downloading, 5.0s loading, 65.2s encoding -- 2.1h processed\n",
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      " 85%|██████████████████████████████████████████████████████████████████████████████████████████████▎                | 17/20 [27:37<04:55, 98.56s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 17/20: 26.7s downloading, 4.8s loading, 64.7s encoding -- 2.1h processed\n",
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      " 90%|███████████████████████████████████████████████████████████████████████████████████████████████████▉           | 18/20 [29:16<03:17, 98.76s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 18/20: 24.5s downloading, 5.0s loading, 67.4s encoding -- 2.1h processed\n",
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      " 95%|█████████████████████████████████████████████████████████████████████████████████████████████████████████▍     | 19/20 [30:51<01:37, 97.57s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 19/20: 23.8s downloading, 5.0s loading, 64.3s encoding -- 2.1h processed\n",
      "  downloading audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "audio needs conversion, will be slow without using joblib\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  loading audio...\n",
      "  embedding audio...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████| 20/20 [32:35<00:00, 97.78s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 20/20: 31.3s downloading, 4.8s loading, 66.2s encoding -- 2.1h processed\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "chunksize = 250\n",
    "tot_steps = 20\n",
    "train_data = []\n",
    "val_data = []\n",
    "for n in tqdm.tqdm(range(tot_steps)):\n",
    "    filepaths = [\n",
    "        fp\n",
    "        for m in metas[-(n + 1) * chunksize : len(metas) - n * chunksize]\n",
    "        if (fp := m.get(\"s3_filepath\", m.get(\"audio_filepath\", m.get(\"filepath\"))))\n",
    "        is not None\n",
    "    ]\n",
    "    download_duration_s, load_duration_s, audio_arrays = _load_audio(\n",
    "        filepaths,\n",
    "        SAMPLE_RATE,\n",
    "        max_duration_per_file_s=8 * 60,\n",
    "    )\n",
    "    t0 = time.time()\n",
    "    print(\"  embedding audio...\")\n",
    "    encoded_arrays = encode(audio_arrays, do_clustering=False, normalize=True)\n",
    "    # transpose\n",
    "    encoded_arrays = [arr.T for arr in encoded_arrays]\n",
    "    timing_encode_s = round(time.time() - t0, 1)\n",
    "    n_hours_processed = round(\n",
    "        np.sum([arr.shape[0] / EMBEDDING_RATE for arr in encoded_arrays]) / 60 / 60, 1\n",
    "    )\n",
    "    # subsample audio array for better diversity\n",
    "    stacked_arr = np.concatenate(encoded_arrays, axis=1).T\n",
    "    idx_list = list(range(stacked_arr.shape[0]))\n",
    "    random.shuffle(idx_list)\n",
    "    keep_idx = np.array(idx_list[: int(stacked_arr.shape[0] / tot_steps)])\n",
    "    stacked_arr = stacked_arr[keep_idx, :]\n",
    "    if n == tot_steps - 1:\n",
    "        val_data.append(stacked_arr)\n",
    "    else:\n",
    "        train_data.append(stacked_arr)\n",
    "    print(\n",
    "        f\" {n+1}/{tot_steps}: {download_duration_s}s downloading, {load_duration_s}s loading,\"\n",
    "        f\" {timing_encode_s}s encoding -- {n_hours_processed}h processed\"\n",
    "    )\n",
    "#  1/20: 25.0s downloading, 4.5s loading, 61.6s encoding -- 13.0h processed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "67b92052",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T03:34:10.665581Z",
     "start_time": "2023-09-22T03:34:10.663464Z"
    }
   },
   "outputs": [],
   "source": [
    "# TODO: why is the above ~2x slower than before? torch version?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "cb95ea16",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T03:34:20.232474Z",
     "start_time": "2023-09-22T03:34:10.666567Z"
    }
   },
   "outputs": [],
   "source": [
    "np.save(\"/home/tony/Data/MERT/mert_v2_25hz_norm_val\", np.concatenate(val_data, axis=0).astype(np.float32))\n",
    "np.save(\"/home/tony/Data/MERT/mert_v2_25hz_norm_tr\", np.concatenate(train_data, axis=0).astype(np.float32))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "d758e1d8",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T02:50:05.160442Z",
     "start_time": "2023-09-22T02:50:02.891997Z"
    }
   },
   "outputs": [],
   "source": [
    "# import torch\n",
    "# x = torch.randn(20, 100, 40)\n",
    "\n",
    "# n_x = (x - x.mean(1).unsqueeze(1)) / (x.std(1).unsqueeze(1) + 0.00001)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "29942363",
   "metadata": {},
   "source": [
    "## Faiss clustering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b76ae3f8",
   "metadata": {},
   "outputs": [],
   "source": [
    "# conda activate faiss\n",
    "# ipython"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "aedeebae",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-23T02:04:41.153212Z",
     "start_time": "2023-09-23T02:04:39.862603Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1144573, 768)\n",
      "(12298, 768)\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"1\"\n",
    "\n",
    "import numpy as np\n",
    "import faiss\n",
    "from sklearn.metrics.pairwise import paired_distances\n",
    "\n",
    "X_arr = np.load(\"/home/tony/Data/MERT/mert_v2_25hz_tr.npy\")\n",
    "y_arr = np.load(\"/home/tony/Data/MERT/mert_v2_25hz_val.npy\")[::5]\n",
    "\n",
    "print(X_arr.shape)\n",
    "print(y_arr.shape)\n",
    "# (1144610, 768)\n",
    "#   (12299, 768)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "d990d443",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-23T02:04:41.157507Z",
     "start_time": "2023-09-23T02:04:41.155073Z"
    }
   },
   "outputs": [],
   "source": [
    "# import os\n",
    "# os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"1\"\n",
    "\n",
    "# import numpy as np\n",
    "# import faiss\n",
    "# from sklearn.metrics.pairwise import paired_distances\n",
    "\n",
    "# X_arr = np.load(\"/home/tony/Data/MERT/mert_v2_25hz_norm_tr.npy\")\n",
    "# y_arr = np.load(\"/home/tony/Data/MERT/mert_v2_25hz_norm_val.npy\")\n",
    "\n",
    "# print(X_arr.shape)\n",
    "# print(y_arr.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "e8492b0f",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-23T02:05:29.725961Z",
     "start_time": "2023-09-23T02:05:27.698561Z"
    }
   },
   "outputs": [],
   "source": [
    "x_mean = X_arr.mean(axis=0)\n",
    "y_mean = y_arr.mean(axis=0)\n",
    "x_std = X_arr.std(axis=0)\n",
    "y_std = y_arr.std(axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "2ddc9906",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-23T02:05:35.281604Z",
     "start_time": "2023-09-23T02:05:35.278071Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.18639381, 0.14269768, 0.11550962, 0.15448196, 0.16461523,\n",
       "       0.3115032 , 0.14381032, 1.0048118 , 0.46973795, 0.2367955 ],\n",
       "      dtype=float32)"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_std[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "694889ef",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-23T02:05:32.701721Z",
     "start_time": "2023-09-23T02:05:32.698341Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.18639381, 0.14269768, 0.11550962, 0.15448196, 0.16461523,\n",
       "       0.3115032 , 0.14381032, 1.0048118 , 0.46973795, 0.2367955 ],\n",
       "      dtype=float32)"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_std[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "b68a5410",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-23T02:07:02.612362Z",
     "start_time": "2023-09-23T02:07:02.610543Z"
    }
   },
   "outputs": [],
   "source": [
    "# # If post norm\n",
    "# X_arr = (X_arr - x_mean) / (x_std + 0.0001)\n",
    "# y_arr = (y_arr - x_mean) / (x_std + 0.0001)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "a21d1807",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-23T02:07:21.882540Z",
     "start_time": "2023-09-23T02:07:08.165568Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "calculating codebook 0...\n",
      "train codebook 0...\n",
      "\n",
      "Sampling a subset of 256000 / 1144573 for training\n",
      "Clustering 256000 points in 768D to 1000 clusters, redo 1 times, 25 iterations\n",
      "  Preprocessing in 0.80 s\n",
      "finish train codebook 0...arch 2.96 s): objective=1.57206e+08 imbalance=1.171 nsplit=0       \n",
      " score: 24.895\n",
      "----------\n",
      "calculating codebook 1...\n",
      "train codebook 1...\n",
      "\n",
      "Sampling a subset of 256000 / 1144573 for training\n",
      "Clustering 256000 points in 768D to 1000 clusters, redo 1 times, 25 iterations\n",
      "  Preprocessing in 0.79 s\n",
      "finish train codebook 1...arch 2.99 s): objective=1.5292e+08 imbalance=70.449 nsplit=0        \n",
      " score: 24.511\n",
      "----------\n"
     ]
    }
   ],
   "source": [
    "n_codebooks = 2\n",
    "n_clusters = 1_000\n",
    "\n",
    "X_arr_resid = X_arr.copy()\n",
    "y_arr_resid = y_arr.copy()\n",
    "y_preds_prev = np.zeros(y_arr.shape)\n",
    "centroids_list = []\n",
    "models_list = []\n",
    "for n_codebook in range(n_codebooks):\n",
    "    print(f\"calculating codebook {n_codebook}...\")\n",
    "    faiss_model = faiss.Kmeans(\n",
    "        d=X_arr_resid.shape[1], \n",
    "        k=n_clusters, \n",
    "        niter=25, \n",
    "        nredo=1, \n",
    "        seed=n_codebook,\n",
    "        verbose=True, \n",
    "        gpu=True\n",
    "    )\n",
    "    print(f\"train codebook {n_codebook}...\")\n",
    "    faiss_model.train(X_arr_resid)\n",
    "    print(f\"finish train codebook {n_codebook}...\")\n",
    "    # score preds\n",
    "    y_cluster_preds = faiss_model.index.search(y_arr_resid, 1)[1].squeeze()\n",
    "    y_preds = faiss_model.centroids[y_cluster_preds]\n",
    "    y_arr_resid -= y_preds\n",
    "    y_preds_prev += y_preds\n",
    "    print(\" score:\", round(np.mean(paired_distances(y_arr, y_preds_prev)), 3))\n",
    "    # start stuff for next round\n",
    "    X_cluster_preds = faiss_model.index.search(X_arr_resid, 1)[1].squeeze()\n",
    "    X_preds = faiss_model.centroids[X_cluster_preds]\n",
    "    X_arr_resid -= X_preds\n",
    "    centroids_list.append(faiss_model.centroids)\n",
    "    models_list.append(faiss_model)\n",
    "    print(\"-\"*10)\n",
    "codebooked_centroids = np.stack(centroids_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "b15a69da",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-23T02:08:18.475783Z",
     "start_time": "2023-09-23T02:08:18.474041Z"
    }
   },
   "outputs": [],
   "source": [
    "# np.save(\"/home/tony/Data/MERT/cluster_centers/normalized_2\", codebooked_centroids)\n",
    "# np.save(\"/home/tony/Data/MERT/cluster_centers/mean_2\", x_mean)\n",
    "# np.save(\"/home/tony/Data/MERT/cluster_centers/std_2\", x_std)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "a4dca183",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T14:38:20.951185Z",
     "start_time": "2023-09-22T14:38:20.886531Z"
    }
   },
   "outputs": [],
   "source": [
    "# np.save(\"/home/tony/Data/MERT/cluster_centers/normalized\", codebooked_centroids)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "b9f884f2",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T14:44:28.922098Z",
     "start_time": "2023-09-22T14:44:28.920438Z"
    }
   },
   "outputs": [],
   "source": [
    "from suno_utils.tasks.mert_25 import ClusterModel\n",
    "from suno_utils.tasks.mert_v2 import preload_models as mert_v2_preload_models\n",
    "from suno_utils.tasks.mert_v2 import encode as mert_v2_encode"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "ab7418fe",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T14:38:40.041062Z",
     "start_time": "2023-09-22T14:38:38.691776Z"
    }
   },
   "outputs": [],
   "source": [
    "_ = preload_models(\n",
    "    checkpoint_filepath=\"/home/tony/Data/MERT/mert_test_8x_400k.pt\",\n",
    "    centroids_filepath=\"/home/tony/Data/MERT/cluster_centers/normalized.npy\",\n",
    "    device=\"cuda\"\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "a177faca",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T14:45:43.811317Z",
     "start_time": "2023-09-22T14:45:41.772840Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Some weights of the model checkpoint at m-a-p/MERT-v1-95M were not used when initializing MERTModel: ['encoder.pos_conv_embed.conv.weight_v', 'encoder.pos_conv_embed.conv.weight_g']\n",
      "- This IS expected if you are initializing MERTModel from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\n",
      "- This IS NOT expected if you are initializing MERTModel from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n",
      "Some weights of MERTModel were not initialized from the model checkpoint at m-a-p/MERT-v1-95M and are newly initialized: ['encoder.pos_conv_embed.conv.parametrizations.weight.original0', 'encoder.pos_conv_embed.conv.parametrizations.weight.original1']\n",
      "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
     ]
    }
   ],
   "source": [
    "_ = mert_v2_preload_models(\n",
    "    centroids_filepath=\"/home/mikeys/bundle/2x1k_centroids_mert.npy\",\n",
    "    device=\"cuda\"\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "23451002",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T15:22:25.160210Z",
     "start_time": "2023-09-22T15:22:25.158571Z"
    }
   },
   "outputs": [],
   "source": [
    "x = audio_arrays[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "de42f7e7",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T15:23:16.983073Z",
     "start_time": "2023-09-22T15:23:16.980470Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([1, 1440009])"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x.reshape([-1,]).reshape([1, -1]).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "4da17918",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T15:21:36.097688Z",
     "start_time": "2023-09-22T15:21:36.094542Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([1, 1440009])"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "audio_arrays[0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "7ef5f4c7",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T14:33:39.022247Z",
     "start_time": "2023-09-22T14:32:35.382266Z"
    }
   },
   "outputs": [],
   "source": [
    "encoded_array = encode(audio_arrays, do_clustering=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "64ee2de3",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T15:29:48.488466Z",
     "start_time": "2023-09-22T15:29:48.484425Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[ 0.0444, -0.0568,  0.0112,  ...,  0.2346, -0.0816,  0.2552],\n",
       "        [ 0.1970, -0.4084,  0.0516,  ...,  0.4091, -0.2209,  0.0271],\n",
       "        [ 0.3686, -0.1719,  0.1171,  ...,  0.8551, -0.2016,  0.0154],\n",
       "        ...,\n",
       "        [-0.0656,  0.1930, -0.0224,  ...,  0.2773, -0.1341,  0.0468],\n",
       "        [-0.0414,  0.1820, -0.0237,  ...,  0.3057, -0.1424,  0.0415],\n",
       "        [ 0.0185,  0.1435, -0.1189,  ...,  0.2951, -0.2176,  0.0619]])"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.from_numpy(encoded_array[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "026bebf4",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T15:29:16.827083Z",
     "start_time": "2023-09-22T15:29:16.824042Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(250, (4708, 768))"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(encoded_array), encoded_array[1].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "id": "cf4ae30e",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T15:46:25.441639Z",
     "start_time": "2023-09-22T15:46:25.439417Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1499, 2)"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.array(encoded_array_cluster[0]).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "f79c60e4",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T14:36:08.322642Z",
     "start_time": "2023-09-22T14:35:02.780735Z"
    }
   },
   "outputs": [],
   "source": [
    "encoded_array_cluster = encode(audio_arrays, do_clustering=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "e33a7999",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T15:27:50.957073Z",
     "start_time": "2023-09-22T15:27:50.953762Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(250, (1499, 2))"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(encoded_array_cluster), encoded_array_cluster[0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "dc7ed381",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T15:52:52.903170Z",
     "start_time": "2023-09-22T15:52:52.900963Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1, 1499, 2)"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.array([encoded_array_cluster[0]]).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "id": "2f36928c",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T15:52:14.417429Z",
     "start_time": "2023-09-22T15:52:14.404895Z"
    }
   },
   "outputs": [
    {
     "ename": "TypeError",
     "evalue": "reshape(): argument 'shape' (position 1) must be tuple of ints, but found element of type float at pos 0",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
      "Cell \u001b[0;32mIn[99], line 1\u001b[0m\n\u001b[0;32m----> 1\u001b[0m \u001b[43mtorch\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mfrom_numpy\u001b[49m\u001b[43m(\u001b[49m\u001b[43mencoded_array_cluster\u001b[49m\u001b[43m[\u001b[49m\u001b[38;5;241;43m0\u001b[39;49m\u001b[43m]\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mastype\u001b[49m\u001b[43m(\u001b[49m\u001b[43mnp\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mint32\u001b[49m\u001b[43m)\u001b[49m\u001b[43m)\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mreshape\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m1.\u001b[39;49m\u001b[43m \u001b[49m\u001b[38;5;241;43m-\u001b[39;49m\u001b[43m \u001b[49m\u001b[38;5;241;43m1\u001b[39;49m\u001b[43m)\u001b[49m\u001b[38;5;241m.\u001b[39mshape\n",
      "\u001b[0;31mTypeError\u001b[0m: reshape(): argument 'shape' (position 1) must be tuple of ints, but found element of type float at pos 0"
     ]
    }
   ],
   "source": [
    "torch.from_numpy(encoded_array_cluster[0].astype(np.int32)).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f5bba161",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "2885bc85",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T14:40:05.636462Z",
     "start_time": "2023-09-22T14:39:02.525226Z"
    }
   },
   "outputs": [],
   "source": [
    "encoded_array_cluster_normalized = encode(audio_arrays, do_clustering=True, normalize=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "fce57d61",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T14:48:21.549588Z",
     "start_time": "2023-09-22T14:46:08.891485Z"
    }
   },
   "outputs": [],
   "source": [
    "encoded_array_cluster_75hz = mert_v2_encode(audio_arrays, do_clustering=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "0dfe41f4",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T14:48:50.446528Z",
     "start_time": "2023-09-22T14:48:50.443430Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[818, 233],\n",
       "       [258, 839],\n",
       "       [147, 226],\n",
       "       [ 43, 167],\n",
       "       [462, 386],\n",
       "       [920, 167],\n",
       "       [920, 167],\n",
       "       [920, 962],\n",
       "       [349, 605],\n",
       "       [542, 167],\n",
       "       [ 48, 167],\n",
       "       [349, 637],\n",
       "       [349, 581],\n",
       "       [349, 976],\n",
       "       [349, 144],\n",
       "       [349, 976],\n",
       "       [349, 167],\n",
       "       [349, 214],\n",
       "       [950, 867],\n",
       "       [364, 287]], dtype=uint16)"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "encoded_array_cluster_75hz[0][:20]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "f2b7be97",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T14:40:51.633758Z",
     "start_time": "2023-09-22T14:40:51.631572Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[116,  63],\n",
       "       [361,  40],\n",
       "       [361,  40],\n",
       "       [676, 231],\n",
       "       [676, 231],\n",
       "       [252, 231],\n",
       "       [951,  40],\n",
       "       [ 38, 231],\n",
       "       [ 38, 857],\n",
       "       [826, 264],\n",
       "       [527, 264],\n",
       "       [250, 816],\n",
       "       [349, 177],\n",
       "       [962, 448],\n",
       "       [582,  47],\n",
       "       [983, 575],\n",
       "       [949,  96],\n",
       "       [864,  47],\n",
       "       [ 34, 865],\n",
       "       [337, 231]], dtype=uint16)"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "encoded_array_cluster[0][:20]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "d6d14a92",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T14:40:52.313490Z",
     "start_time": "2023-09-22T14:40:52.311280Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[135, 359],\n",
       "       [706, 823],\n",
       "       [706, 823],\n",
       "       [706,  33],\n",
       "       [633, 731],\n",
       "       [650, 320],\n",
       "       [203, 743],\n",
       "       [609, 320],\n",
       "       [186, 743],\n",
       "       [349, 743],\n",
       "       [723, 964],\n",
       "       [314, 743],\n",
       "       [498, 743],\n",
       "       [316, 774],\n",
       "       [617, 774],\n",
       "       [508,  52],\n",
       "       [540,  52],\n",
       "       [940,  52],\n",
       "       [626,  52],\n",
       "       [670, 392]], dtype=uint16)"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "encoded_array_cluster_normalized[0][:20]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a435caad",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Codebooks 2x 1k:\n",
    "#   score: 5.953\n",
    "#   score: 5.745\n",
    "\n",
    "#   score: 12.57\n",
    "#   score: 12.20"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b92374cf",
   "metadata": {},
   "outputs": [],
   "source": [
    "# save cluster centers\n",
    "# np.save(\"/home/georg/notebooks/gpt/data/cluster_centers/mert_v2_25hz_1x10k\", codebooked_centroids)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "2564eea1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "upload: ../gpt/data/cluster_centers/mert_v2_25hz_1x10k.npy to s3://suno-data/georg/tmp/mert_v2_25hz_1x10k.npy\n"
     ]
    }
   ],
   "source": [
    "# !aws s3 cp /home/georg/notebooks/gpt/data/cluster_centers/mert_v2_25hz_1x10k.npy s3://suno-data/georg/tmp/mert_v2_25hz_1x10k.npy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "507e4272",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-09-22T13:25:12.809634Z",
     "start_time": "2023-09-22T13:25:12.805404Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1000, 768)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "codebooked_centroids[0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c517acd6",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "e2cf7999",
   "metadata": {},
   "source": [
    "### embed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c9a3df44",
   "metadata": {},
   "outputs": [],
   "source": [
    "modal run ~/code/glockenspiel/suno_utils/suno_utils/scripts/gpt/modal_encode.py \\\n",
    "    --embed-type 'mert' \\\n",
    "    --data-type 'music_sample' \\\n",
    "    --version 'v2' \\\n",
    "    --chunksize '500' \\\n",
    "    --min-duration-s '30' \\\n",
    "    --max-duration-s '800' \\\n",
    "    --output-name 'mert_v2_25hz_2x1k'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "50fc5e07",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cb5fbf08",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ee5d4f12",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "40e335f6",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2d092649",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a296a4ca",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "54c66630",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.12"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}