Fastest_dimension_of_array.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Which is the fastest axis of an array?\n",
    "\n",
    "I'd like to know: which axes of a NumPy array are fastest to access?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## A tiny example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0, 1, 2],\n",
       "       [3, 4, 5],\n",
       "       [6, 7, 8]])"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "a = np.arange(9).reshape(3, 3)\n",
    "a"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'0 1 2 3 4 5 6 7 8'"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "' '.join(str(i) for i in a.ravel(order='C'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'0 3 6 1 4 7 2 5 8'"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "' '.join(str(i) for i in a.ravel(order='F'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## A seismic volume"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "volume = np.load('data/F3_volume_3x3_16bit.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(194, 299, 451)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "volume.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's look at how the indices vary:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "idx = np.indices(volume.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3, 194, 299, 451)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "idx.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can't easily look at the indices for 190 &times; 190 &times; 190 samples (6 859 000 samples). So let's look at a small subset: 5 &times; 5 &times; 5 = 125 samples. We can make a plot of how the indices vary in each direction. For C-ordering, the indices on axis 0 vary slowly: they start at 0 and stay at 0 for 25 samples; then they increment by one. So if we ask for all the data for which axis 0 has index 2 (say), the computer just has to retrieve a contiguous chunk of memory and it gets all the samples. \n",
    "\n",
    "On the other hand, if we ask for all the samples for which axis 2 has index 2, we have to retrieve non-contiguous samples from memory, effectively opening a lot of memory 'drawers' and taking one pair of socks out of each one."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x216 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from matplotlib.font_manager import FontProperties\n",
    "\n",
    "annot = ['data[2, :, :]', 'data[:, 2, :]', 'data[:, :, 2]']\n",
    "mono = FontProperties()\n",
    "mono.set_family('monospace')\n",
    "\n",
    "fig, axs = plt.subplots(ncols=3, figsize=(15,3), facecolor='w')\n",
    "\n",
    "for i, ax in enumerate(axs):\n",
    "    data = idx[i, :5, :5, :5].ravel(order='C')\n",
    "    ax.plot(data, c=f'C{i}')\n",
    "    ax.scatter(np.where(data==2), data[data==2], color='r', s=10, zorder=10)\n",
    "    ax.text(65, 4.3, f'axis {i}', color=f'C{i}', size=15, ha='center')\n",
    "    ax.text(65, -0.7, annot[i], color='red', size=12, ha='center', fontproperties=mono)\n",
    "    ax.set_ylim(-1, 5)\n",
    "_ = plt.suptitle(\"C order\", size=18)\n",
    "plt.savefig('/home/matt/Pictures/3d-array-corder.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x216 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axs = plt.subplots(ncols=3, figsize=(15,3), facecolor='w')\n",
    "\n",
    "for i, ax in enumerate(axs):\n",
    "    data = idx[i, :5, :5, :5].ravel(order='F')\n",
    "    ax.plot(data, c=f'C{i}')\n",
    "    ax.scatter(np.where(data==2), data[data==2], color='r', s=10, zorder=10)\n",
    "    ax.text(65, 4.3, f'axis {i}', color=f'C{i}', size=15, ha='center')\n",
    "    ax.text(65, -0.7, annot[i], color='red', size=12, ha='center', fontproperties=mono)\n",
    "    ax.set_ylim(-1, 5)\n",
    "_ = plt.suptitle(\"Fortran order\", size=18)\n",
    "plt.savefig('/home/matt/Pictures/3d-array-forder.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "At the risk of making it more confusing, it might help to look at the plots together. Shown here is the C ordering:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f017af8d4e0>]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA2QAAADFCAYAAAA7QDEeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzsvXmUJPtV3/mJyIhcau/uqup9X97ThgSWkEAYgTaEpAFLBg4W4nBmbIljGx8BMwNje2ZsYWHswZaxsBBHZjMjeZCMDxayMTbCEg8hIelJSGh5vWRkV9fWXZW1RGRm5RaREfNHZGTXklkZkRG/iKr34nsO56HOyrqVN355f3f53nslx3FIkSJFihQpUqRIkSJFihTxQ076D0iRIkWKFClSpEiRIkWK5yrSgCxFihQpUqRIkSJFihQpEkIakKVIkSJFihQpUqRIkSJFQkgDshQpUqRIkSJFihQpUqRICGlAliJFihQpUqRIkSJFihQJIQ3IUqRIkSJFihQpUqRIkSIhpAFZihQpUqRIkSJFihQpUiSENCBLkSJFihQpUqRIkSJFioSQBmQpUqRIkSJFihQpUqRIkRAUEb90dnbWuXLliohfnSJFihQpUqRIkSJFihRHHl/84hc3HMeZG/ZzQgKyK1eu8PTTT4v41SlSpEiRIkWKFClSpEhx5CFJ0gM/P5dSFlOkSJEiRYoUKVKkSJEiIaQBWYoUKVKkSJEiRYoUKVIkBN8BmSRJGUmS/kKSpP8s8g9KkSJFihQpUqRIkSJFiucKglTI3gU8I+oPSZEiRYoUKVKkSJEiRYrnGnwN9ZAk6QLwJuDngZ8W+hcdU3xj8xvc277H99/4/thl/772+1yfuc4LTr0gVrntTptf/+qv82Mv+DHG1LFYZd/bvsdXyl/hB279wN4X9EX4038JHUuY7OXaMo0XvpWbL/kxYTL6wbRNfv2rv87bnvc2prJTscpeMBb4zOpneNvz3harXIBPPPgE07lpXnbmZbHK7dgdfuNrv8Ffv/XXOZk/uffFT/8SbNwTJrtu1bnjNPjmH/j/QJKEyemHp5afQpEUvv38t8cq13Ecfuvrv8Ubr76R0+OnY5Vdrpf5mPYx/uYL/yZSzPr+zOpnMDsmr7r4qr0vLD8NX/xNcMTIdXAoVu4z+V3/J2cuvVKMkAHYbm7z0Tsf5R3f9A5kKd7OiS88+gJGy+C1l18bq1yA37n9O7z87Mu5On01VrnVdpUPfeNDvOOb3oEi73L7rDb88buhoQuTvdncZOXERb7pe35xz7/fXavyG5++T8cWdMCBhcoCJ/MnY78vLceiuF3k1olbsZ9vvaVTaVe4NHkpVrkAi9VFprJTzORm9v1Ri1BZESbXsi30ls7bX/vNvOYVbxImRzT8Tln8JeBngMlBPyBJ0juBdwJcuhT/QUgav/m13+Qzq5+JPSDr2B1+7rM/xxuvvpGfe+XPxSr7zx/+Ob/ylV/hBbMv4DsvfGessj/0zIf4fe33DwZkX/0P8MXfgqkLQuQ6wJnKMl+rLELMAdmX17/M+7/8fq5MXeENV98Qq+zfufM7fPiZD/ODt34QNaPGKvuff+Gfc+vErdgDsttbt3nfX7yPU4VTvPXmWx+/0NiGT/wjyM9AdkKIbLmxwTebTYzXLTI9c1mIjEH4l0//S07mT8YekN2v3Oe9X3wv2UyWH3nej8Qq+2Pax/jXX/rXvPnamzkzfiZW2e/70vsADgZkn30/PPNxmBATnNqOxc3qI76Q/+XYA7I/uP8H/Jsv/xted+V1XJu+FqvsD3zlA2w0NmIPyIyWwc9/7uf5Wy/6W7zrW94Vq+xPPPgEv/KVX+E7L3wnL5jdlbhd/jx89t/A+BxkckJkT9QecUaWYF9A9u8/t8hHn17izFReiFwHh0c7NcZVm6lsW4iMQWh0muhNi9WNR+TkbKyyt1s67U6LpbGNWOU6wFq9Qj5jMpPblxCvVcEeA0HBqYOD7UyybWwL+f1xYWhAJknSm4F1x3G+KEnSdw36OcdxPgh8EOClL32puJTHEUVRL9KwGrHLXamt0Oq0EpFd1IuAm81PQrZlW5i2iSrvChDWb7vB2E9/XYjcjXqZ8i89j4zVEvL7D4OmawCJP+vpzHRscmvtGo92HnFx8mJsMj14n/mAvtdvu/996wfh1vcIkf17H34Df+PeZ2k1t4H4AjKzY7JYWaSgFGKT6eGonO84YTs2JaPE2fGzB18s34Ybr4G3fUSI7D9f/jQv/7U3gbkj5PcfhoHfrRig6RrZTLxOsicXjtj5Xu92obzzT2D6vBDZf/jLT/Dd2+UD/35vvcoLz0/z+z/xHULk3t66zQ9+/Kd468238u5vf7cQGYPw/i+/n1/9yq/yr17z/tiT1W/52Ft4UHnAZ370S7HKXaws8qbf+0necOUN/OKrdgXfVhv+6Vl45bvgNf+3ENm/+bXf5L1ffC+v/+4/E/L744KfcPWVwPdJkrQA/A7wakmSPiT0rzpmsGyLhcoCpm1i2eKocv3gGfmm1YxVbpKyHcehpJf6yy7fhvknhcnWDI2mJKNYpjAZA2UneKEP1LdouUYycsF91tBH3+VuQDYn7pyttLYAaDfFUYn64UHlAZZjJWpPnkvn++HOQxpW46DcjuVSYgWeMc0o0ZQkMJ87+t5qbrHV3ErUnhwp2eU7kJuCqXNC5NbNOpvWDjnHPvDavbUaN+cHkq5CI0l7kpTsI+mLbmlgWzD3PGGyi3qR+cJ87NTUqDE0IHMc5+87jnPBcZwrwA8D/8NxnLcL/8uOERari73DH7exHeg0xiE7IaOzVl+jZtYOyrY7sHFXrBOjazRkCbUTLwUCknvWRsug3CgnIvsoOOl9AzJ1DKbFVO1anRbLCQVkRSPZykUSsjt2pxf4H5nzvVUC2xQckLm2TI45IHMcJ7EK2VFw0pNglAz83OXbMPeEsD7V+5X7NCSZnOPQsR7fmUbdZL3a4uZpMZRvOBrPOm7ZR9IX9aqwc08Ik13SS1ybiZf6LALpHrII4DluAM1OzF8C74vfifeL71FtIIGqyW5975atPwCrKT4gkyQUgUNDDpMN8Z8x7zknIfso0Hz6VmHnngBZjPlcMBaod/0jq1URImMQepWLmJ8zJFdBWK2t0uq0EpE98DvtVWFFVvt1jYYkI3fipV9vNjeptN1zndTd0eq0sPtUbUQiKUaJR/uGAedM8H1Zl11j1mw+7u+5t14F4FYMAVnc+vZo30nIPpK+aPmO2zs2e1OIXMdx0AyNGzM3hPz+OBHIo3Ac51OO47xZ1B9zXOEdRIBGzNnGpBzWRzuPejJjz3Iau/S9W3b5jvtfwRdMU5LI2fFSFreb22w13apJUmcMknvWccutm3VWa6v9ZZfvxBD0u6bZbBrC5AySDQlRbYyFRGQPtCdxyN6l7z0BgmfLZm8JkevRvpuSRMaMNyA7CvYEknOW4/7M9437vf9/z92xswk75VjuS4BmY6v37/fWXYaLSMpiUlXvxeoilmMlIvtI+qLl23DiCqhi+pI92ndaIUsB7P0SxElHsB27Z2zj/vJ51QNI2Ojslt0rjYtzYlyaj0yu0xEiYxASdWKOopEXjPuV+zjdeeN7ZDd0qD4USr8o6kUaXSemE3OFbGCAIBhL1SXMbpIj7md9VGzZngCh/AzMXILsuBC56/V1qmaVhiyhxDyg6KjoO867utKusN5YB47Q+Y6hF9ZjlAC0dgVkd9eqFNQM52fEOOmtTovFqlulOjL6jgFH0heNoQoLcH36ujAZcSENyCKAZmhku6NN4ywTr9RWaHaaZOVs/DS2brYvK2cTMTqevg9UyKbOQ17MFMDN5iZGy6AtZ/o2KYuEl+1L4lnv0XeM1Ngdc4eHOw/dz2w1cZz4hrfuPt97HWWvCiuuQblklOh0p8B12jVhcvbDtE0eVB70nnUrRirbQH3HJLuvPREML8HT9+4o3xF6xrxKUUvOoHTirfYnpW/Ya8viPGeJnm+j1N9+x0SLtbq2rN16XO0vrte4MT+BLIvpXVswFrAdOzF7IiGhSEoifVxHyhftmLBZjCcgm0kDsuc8PKrNk6fcAxfnBeMZ+SdPPpkI9WO2MMtsYTYRJ+bJk330XX5GeOMowPj4HHnbxrHjC8o0XWNcHefC5IVEnnUS59vLuD158kkcnFgDBE3XUGSF6zPXB2SVxZ0zTdc4f9Ll23faVWFy9mOpsoTlWIk8ay9AeOLkE8mc7372RDA82vcBffcmLIq3ZZncFGrMAVlS+tabOpvNzURke07jk6cSuKt1javTVw8mT8u33T2KU2LG3TesBiu1FWan3b205q4BRe6ERXH9Y14CMynf6OLkRSayE7FWqY6kL7rpTVgUO5xotjDLdC6+dTyikAZkIbFcXaZtt3nBKXfZYpwZEa80/vxTz4+9gqDpGtenr5NX8rF+5nKjTLVd7S237Mm2bSiLnbDo6Xt64hwZwIxxf4+n74JSiNXQVtoV1uvriZ7vA886Bmi6xpWpK0xkJw46MUoBBC1rbnfaLFWXuDb3QgCcdnwXek/fCT3rc+PnOJk/Gev59qg2SZyxgfreXoBOS7gtO5E7QSY7jhrjeGzHcdB0jZsnbqLKarzf6W7Qn9SzzmfyXJu+lsjgmOszfe5q0RMWDZf2fa5bufAGFBkNk0eVJjdPi+sfK+pFZEnmyZNPJqLvazPXYveNjqQvGkMVtqSXnhV0RUgDstDwjPwLZ10HKtashFFifmye+bF5Ok6n14MhGt6lem3mWuwBgpdpPKBvYxGshlAnpmSUmFQnmRibB/Y2KYuGZiSjby/zlcj57lKbbp64GbtszdC4Nt1H3+Xbbo+iqAmLlQU6Toeb8y9x/6EdY9BvaEhIPO+kS5WL+1kncb5Xa6s0rAZPnHgCRVZit9/Q57sVhxNjuPruKHlyMU6M3Wpuobd0rs/En1waeHfEgJJR4ur0VcbV8Vjl1s06qzur/fUtmhbb1ffFWfccewFZsTfQQ2CFTC9xafIS07npWPVt2iYLlYVEkqdH0hct3wEkOCV2wuKzga4IaUAWGp7R8bIScX4JinqRGzM3yCv5WGU/2nlE3ar3ZCdyqZ7aZ3TWxTcoF/Ui12euI2fdiySugMxoGWw0No6WvmNAUS+6TowyHqvsptVkubrs6juzT9/r8TQo35p9ASZAzFXYC5MXOJE/AcSn747d4b5xP9HzfX3mOoVMvA5UUS9yKn+KM2NngN0BWXc40awYyqK3B+zGzA1stUDOjm9A0W59J/Gsx5QxrkxdARK6q/fbE8Hwgv4DAVl9C2prYmmxRglFUrjUpYh2Wi79+t6a+1+RO8i8uzqv5LEcCzMmWu5SZQnLtlzZMT/rI+mLlp+BE5chOyZE7lp9jR1zJw3IUrjQdI2z42c5lT8FxPcl8Kg2XhY/TtleJsaTHTftZCY3w7mJc8CuxtUYentKeonrM9fJdAOyuJb2eoY2KX3nM3kuT10mI2XibYQ3HldNIL6BIh7VxpPdO2NNA6qrwgMyWZK5Mn2FliwjmfHSNL3MLsRnT5ZrLtWmd75jbEbv2bL9zzoGePbkgL7Ld2D6EuTEOKwbjQ2q7SrXpq+BmicfZy+s8Xgi2pgyFrstuz7jyoX47Em1XWW9vt47Y5ZjxcZm2T2Bbs/d4Q0nmhdXISvqRS5PXWas4PpGdndA0b31GnlV5sIJMU56j/Y9Hf/d0Tvf3e91rLbsKPqigquwHlUyDchSAI+dxrirVN7uBS8LFKfsPVnlBCh016avkcvkkJCom90em/JtmDwLhRkhcreaW2y3trk2fQ05F3NAZiSr76vTV8nImVgz2nWzzkptpdenCPGN3N/tNO75zOW77n8F02IvTl4kl8nRlDNIMX1my7ZYqCwkYssOVE1iXK2g6RrzhXmmslOxyvaoNtem++jb6+0RhN32xFHHyAJWTIGRpmtMqBPMj83HXiHz7o647UmvSjWdwF1taKiyyoXJC/tsmbciRmwC89rMNbIFt+Leq5Ct17g+N0FG0ITFB5UHdJzOXt8orrtDd2nfV6evund1jLbsyPmiHVP4cKJn08h7SAOyUOhRbaZv9AKEuDJ+3kG8MXOjl5WIU/bJ/ElO5E/EmgXaTbWRJGmvbNFOzC59K7kpAMyYdkR5VJuz42djr5B51A8g1mDQm7C453zHdM5Kuku1uTx1eW8W33NiBPb2FPVi73JpyxnkmJ71YnURy7YSsyfwOOHQttt0YqLReQMPoHu+Y8qke1SbA/q2O/E5MTPXkVSXDtxsbAuTt1/29ZnrPfsdl76NlkG5UU7EniR9V1+dvooiKwcrZOo4TF0QIrdpNVmudWnf3YCMbvL03lqVWwIHevTVd4zP+vzEeQpKIVbf6Ej6olv3wTaFVmE9X3QmLyYRHzfSgCwEVmortDqtvRdMzFllLxMD8WbdknDSN5ubVNqV3kb2nuzehEXxDcrXZq6hdMerWk3jsLdEKvva9LXYz1itXWOtvpbIs95PJYP4zndRL3Jp6hJqRqWgFGh1Wm6AUL4DSl7YhEWzY7JYWezpu51RyMS0tNcb3pKEvjVD48z4GcbV8R6dLA5HxnZsSkYpmfOtDzjf2wtgNYU7MdO5aU7lTyF3F0+3YuqHTUrfXpUqkfOta+QyOc5NnEtEtpfgKSiFx2PYvQSmwOFEtmO7FZvcCWzcibHVpslDo8kNgQM9NOMx7fu54hsdSV80jhUxhsaNmRvCfn/cSAOyENjPX42TglHUi8wV5pjOTcdqdBzH2TNmNM7G1YH6Npbc4QcCv/hFvciEOsHpsdNku4unrRgrZF4QmlfyNDtN7BgWU++m7vVkx5R1K+pFVFnl4uTF2C/V3U6jR8FodpruBTN7E+SMELnehMXHAZmKYrWFyNqPol5EQkqmJ3VXlSqfiY9u41Ftdn+34rZlu7PKrhPjLR4X26d4fdp13KQu/boVQ4Vsq7nFVnMr8bsjCUrutelrZORMrBUyj/bd93zHNJzoxvQNJFmmKUlIZj2WCYuarvVo34VMfLbM2wOWlD8IR8wX9QKy2VtC5Hq+6LXpa0J+fxJIA7IQ6GXdph9XbOJyWL2GcE8uxGPk1+pr1MzaY9mqm4mJYwfa7tI47NJ3DE6Mx8+WJIlstzwex9Le3VQbiPdZe1WTRDLaeokr01d6VBuI5zO3Oi2WqksHvlsNq9F1YgRWLnb19gBYGRWlE09AVtJLvQx+nPr2qDa9LL4anwM10J7EgJJR6tG+e0G/1dw1YVGcE7ObhpzJutSxdkt8P+xuqiS4zzrO+7KgFDg7fhZZksln4ksueStLgFiTHfcrj2nfnuym1YTGNtQeCafFZqQMl6dcNkFLziBZTe6tuQGZaMpizyeL0Z4sVZcwbTPZCvBR8kXLt102SbcKHzX2+6LPBqQBWQhousbpsdNMdKfuxfUF3L97wcsqx7EVfr+T7smOg2JU0ktMZad6U4R6+o6hQXk39cMLyLypUSLh9VIlUUHwqDbnJ84DxDoWfLe+48xoLxgu1Wa/7OZOGSrLwpvgZUnujeW2lCxqTOOa+9mTOPS9WlvtUW3ilt2jaU7HXyHb7TTKkkwuk3tcIZu6APkpIXI92veBgCyGAUX97o649S1LrssT17OutWs82nl0wJ7EeVf3KmSevmOYsKjpWo/2DW5AJptN7q1XySkyF0+KmbC4n/adhD3ZzyiJK1l95HxRwVXY/fbk2YA0IAsBTd/LX42rSblHtZnem3WLIyjaXxqPM5vuZXYlSerJ7lXIJk7D2Ekhcreb2y7VpvuZcwVXjhPD0l5P34k8a6PYm7DoyY5DbsNq7KHaxJlVPpDF78q2e8t6xY7wvTBxoee02Zkc2Y744RaWbblVKs9JlzNk5WwstixpezJbmGW62xMa1xh2x3H20DRhlwO1/kwsAz08e+INKLJa4qv9Rb3IuDrO6bHTQLxZ/IH6Fozde8A8uRDf+VZkhYuTF3uym50mTm9np9gdZLt9o3YmQ8ZqcXdN/IRFy7ES0ze4vVSebAeHVkd8H/BR80Vb5g5s3hM+AAvSgCwFLtXGo7F5iGvM6QGqjVeWj0F2yShxIneCk3k3KInLWe63kX2vEyOeD9/LAnWnRjlt8VlOTdcoKIXe3rVYn/U+fnZcToy3B8w734qsoMpqLLKLenEP1cY731KvQVlsxm+3PXGX9lrC5HlYri67VJtdo4MLaky2zNgbIMQZfO/uFfRkxyF3vb5+gGpTUAo0zR3YuCu8cgGP7w61EF8/bMko9XrX4LG+RVcQKu0K6431RJ71oARPXBWbK1NXUGW3SjWmjmE7Np31r4M65u66E4BWp8VidXHP3dGWVTKdFsX1mtCF0Pv7nmNN5hnuhMUxdSxW2UfRF1X0Jei0hbeRnMyf7PmizwakAdmI8Kg2u7MScdEgDhj5GBtXd/cfQHxGZ7O5idEy9jiN7t6genf5YAwNyp4To45hApIpvkLmjS32qDZxPesdc4eHOw+TPd/7n3VMTvqlqUtkM1ng8fnObBQhk4MTV4TINTsmDyoP9ujbVvPkYhj/vv98Q3x0Mo9qM9mlzsWa4NlFi4XHZ0x0gNBX30qefK3sTlgUXCGbzE4yW5gFIJvr9sPGVCHbf3d0nI7wJcn7qWQQry3LylkuTLjj5ZNglHjw6GTO+jNuj6KoCYtd2vfu820qKu22xIreED7Qw9sDBvHS3ftVYeOQfRR90cL2ovsDgm3Zs2mgB6QB2cjYn9mF+ChdmqHtodqoGRVFUoTL7k1Y7GN0RF8w/fjCBaXAZMMQPmFRM9w9YB7VBqApy0gxXKr7x7rGRVnc33/gyY7LyCuywsWpi3tkx+HE7HfSPX3ntu53nRgxExYXq4tYjrXHnjhKgbwd3zRNz4mBGG3ZACdG9LN+tPOIulXvGyBYgquSg+6Ok9V193+ITC517YlXpYqrH3Y/7Rvic1j3O42e7FjsiaEdoH2D+M/s0b776VveuCu8cgF7745OJstD000C3BQ80MNbgg08XqMh+Fn3Jiz2uTtEyz6KvuikseL+wKwYv6wf7fvZgDQgGxG798h4iCurvHvsfE92DBmRcqNM1azu+eLH1aS8fwIduPo+W+/uAhNI8/GCUM+Jge7UKMGUgGq7ynp9va++hTsxRnJVKs3Q9lBtIJ5gsN1ps1RdOvCdBhjbXoxtWa8HRx1DBUxT8HdL1zg3fq5HtYF49G07NveN+8me7z4VBOG2TNeYyc3sodrkM3lmaxvu/xA8OGa3vnM9+rXYgGz/FDiI91nnM/ke7duTHdddvceexPSZF4wFHJwD9nuyY5OprQnv7clImd5wIgBLybFszgNiR957tFgPiqyQkTLC9b1SW6Ftt/veHXElHI6SLzptPHQpsTkxz9rzRdOALAXgfgnmx+aZyj6ehhVH1s3rpdr95YtL9u7dObvlQgxZIF1jUp1krjD3WLZa4EKz60gIzPjtp34AtGQFWfDS3kGZXYjHyGflLBcmL+yRbdmWcIpRPypCHOfb2wO253yrBQq2TWGnLNSJ2U+1AZCy3eyu4B1R/TKNceh7pbZCs9Psa09iq5rs65uDeGzZ/gRPQS1wemcbJs9Bd89h1NhsbLLd2t4bhObdgMwWTL/uR9OM8+7YTfuGeAa41M06qzurez5zVs4iS7Lw891v4EFBKXDN7Npuwb2wFycv9mjfALaSY8k6Q1aRuSRqwqJtslBZ2POZ41qS3Nc3imnk/lH0RU/VxN6XvfM9nQZkKeiOid53GLydXCKxVl9jx9w5sJ28oBSEZ3YH0dggHqNzwIlRClxpNXHG54RNWNSbOpvNzQPP2swoZETTNI3+NE2IR9/eHrD9skUa+qbVZLm63Pd8i/7M+8ege3LjcGK8hnBPxwBy1lvauyVMbm8PWJ+ALC59J3W+T+VPMdOl7MUluzecaJ89GVPGOFevCJ98B3v1nctNYwMIHlCk6S7t+8z4md6/xfmsE7EnxsHetbgChJJRQpEULk09HtxRUApc79kysRT//fbEVvMs2Oe4NjuOkhHjdi5VlrBsK1Fbtv/ugHgqwEfJFx3P5Dm1syW80g8cCAaPO9KAbAR4VJsDX/xMgVanRUdgI/7+scUe4qBgaIZLtfH2gHlyQbzR2T8RDVx9XzNNOrM3hcqFg6NV2xkFxRK7tFfTu1Sb8b1UG0hI3zEEZAsVl2qz39DGQcEo6kVkSd5Tpcpn8lxve06M2Ol3+y82qbtQs9UUVyHzqDb7n3Vc9gT2Xqq5TA4JKR4npo89AbHfrY3GBtX2QapNQc5xobkTy4TFPQGCLNOQJOH0a0/fuxNqcdiyWrvGWn0tEXvSj+HgyRbd36PpGpenLh+gfV9vm3QyOXdhrwC0O+09e8A8OEqB+/Y5sQuh+9CQIT5btp/2HQdl8Sj6opc7DqrdEXtf9vFFnw1IA7IRsFpbpWE1BjqsIvdODNq9EEeJ2qOS7a9SgVgnfau5xVZz6yCNressm6fEla0H6dvKZFEEL+31qDaZXYMk4mhSrpt1tyF8f9YthozfICpCHHtVSkaJS5OX9lBt8kqe66ZJR86Im7DYpdrsdxqVLgXFbBpC5MIQfcdBtSnspdpIktRbqCoK3nCifrRYEDswZ5A9OWO2yTu20KxyUS8yoU4wPza/59/j6IftR0OOw5b164WFeIYeaLqGKqt7aN+e7Di+W/2oZNdMk53pc8KGEz2oPKDjdA7ouyFPseLMc2NODF0R3PO9n/YNMfpGMwPOt8BzdhR90VgYJX180WcD0oBsBAyqmsQx4MLbvXCiy/3fLVs41WZArwmIddL79R8AzLTrTDgOjRNisn3g6rugFPZQbQAsJUtWdEDWJ4ufy+QAsfq+b9wHBp9vkbJLuku18faA7ZYteq9KP6dRlmRumjZb46cgowx4ZzgsVftTbTI5N5vcFhiQ9ZuIBjFllfs4MSDeYV2rr1EzawfsSe98Czxng+6Oc43u2HmBWWWv6r3fiWnJGaH0a6NlsNHYSMyewMG7wzvftiNuiqlmHKR992QLPGNNq8ly7SDtO6/kudE2qUyfG/DO8BhUpVqx3fvz8syBt0SGkl7i3MS5PbRvEG/LerTvPsMtIBl7kqQveqXV/bzrPiQqAAAgAElEQVRzt4TI9XzR/ef72YA0IBsBXmZgYIZV4OXWb8CEJ1uk3I3GBpV2pS/9AhBaveg3RQjoTSXbmblw4D1RoagXuT59fU9DOEBHyaF2xI3HrrVrPNp5dNBJlzPkMjnhZwz6V2FBfIXs0tQl1Iy6599FZ7TNjtmXagNw3TRZHxe3fHIQtUnJuZUjqyW2QnZ2/Czj6viefxdtT2zHpmSU+l6qop/1IHvSO98CbVlRLzKdmz5AtTnTnRbrzIpxYqD/8BZw+2FFDigadL7j0ncuk9szYXG3bJEVBE3XuDF98HyPKWNCz/dCxd0DdqBi07E40+mwNTk/4J3hoekasiRzZfrKnn9fMd1hXBenxNH8i3oxEXuyUluh1WkNpvjHUHE/Sr7oxeYO5WwecmLoqZ4v+mzrH4M0IBsJmq4xV5jr7V7wINphHUS18WQLrVINyHx5AYLoCtmEOrFnDxjAicojAAyBGb/9Y4s92IrYpb39xkR7ED3ARTNcqs3FyYt7/j2OgKxf75onW6TcB5UHWM7BKhXtHc6ZbR6OiZl8B48d1qtTe6k2anfantWqCJM96Hx7+ha1JPnhzkMaViORCtnQAEFwxeb69MEq1dzOJuuZDK1sYcA7w2EQ7RugnVFRRAZkA+6OWNgVxkHadxyyPdp3oud7/+AYfQmAjQlxfTearnFx8mKPyeFhuXWKLCZzeTHrFSzb6kv7BrefKgl7EkeV6ij6oufqFR7kxNgxGGxPng1IA7IRsH85sgfRFIz1+npfqg3EZ+T7yRZNwSgZrtO434mZNFbZlGVqam7AO8PBaBmUG+W+n9lRC+RjaJgdqG/BTuMgqg2Iy7q1Oi2WqkuHBmSiAoSicXBsMQAbdwFYKgjcnaOXOD9xfk9DODxe2ttpVYXI7dgdt0rVJ4tfUAo4OMIqCIee70w+FqrN7j1gIN5+O44zMKt8srqOpqrCZA+i7gGYGVVoP2xJd2nfZ8fP7vn3WChdA+5q0Q7r/YpL+x5kv4X2zenagT1gALmtBQAejZ84+KYIZfcbRb7cmOKa9BC7Laba79G+E/GNBgQIiqygyuK+03AEfVG7w3xdp6Rm+78xAhx2dxx3pAFZQNiO3be3B8SXiQdlYjzZoo38VHaq71Qb0ZQAjza4H+P6EqWsKuxzD+JngxuQ5WyB/QfdPWDnJ84feE30sx6kb9FOzILhUm0GybYdW9gOtJJeQpbkA71rlO8A8CCXFyIX3GCw3xnzAjK7LWZH1GpttS/VBuKzZX0rwKr48z2o8gziPvNmc7Mv7RvHYabyCC2rJHJ3WEoWVWBA5ul7P+1bNKVrx9zh4c7DQ22Z8POdUIWsH+1b2bhHS5Io5wTtATuE9r28M84NaUXYgKJBVUEQPwJe0zXOjJ85QPsGsXf1kfRF9Qeodod7qpihMZ7sQb7ocUcakAXEo51HLtXmkAtd1Je/Nya6j2zRY3z77QGLQ/Z2c5ut5lZfJya3tSA0q3yY04g6hgqYgpzlQVQbEKvvullntbY6kPoB4hyow5wY0Rk/Tde4MHGhJ6eH9WewJIklRcxAD8u2WDAW+joS+YJbwbHbYmg+/cbO92QLHtms6RqzhdkDVBtPtmiqTRJVk4H2xFhCsVqURNoyQ2NcHT9A+wboZMQOKBqkb1VWUSRFeFUwKXuiyMoB2rcnWzjtu9/C3PJtHmRz1AVVvReri1iOdUDfjXaHtbrKTXlZeEC2f8IixOQbDVhQLPJZH0lfdP02AHfExWOH+qLHHUMDMkmS8pIkfV6SpK9IkvR1SZLeHccfdlTRbyO7B9FNypqucSJ3glOF/lUqy7GEVBB6y0wHcHZFZoEGZmKqD8m0a2gCK2SarlFQCgcawgHo7ohqNMUs7R3UgA9iK5L3K/dxcA4/34KMfFEv9qXaxCF7oL7Ld1grTLFji2lGX6ouYdpmf9pJoUstMsXQ9waNvId4bFkS9mS9vk7VPLgHDNwAISNlhMkeeHd0q7DFrCpW331618Dth80Kol9X2hXWG+uJPGtfd7VAe3Jl6sqePWC7ZYuS2+q0WKz2r1JRvsNiTtxAkUH61so1HCRuSct02mLo15qu9aV9g/gqVb89YB7GlLFkz3fcvmjZDcjuqnIivuhxh58KWQt4teM4LwZeArxBkqRXiP2zji68rNuhGVZBnPhBY6L3yBZg6DebmxgtY2AWSOQF06MN7pe9/gwAmqoK6zfx9oDtp9oAyN7S3nr0AVmPajPIiVELws5Y73wPoH6AOCemZJS4OHlxzx6wnmyB59u0TR5UHgxwYp5hffyk8Cx+3wBBnaADOIKqsCW9xOmx00xkD/bHidR371JNwJ4M2ksF7g40obZMLzGZnWS2MLv3hV22LIm7w1bz5DpiArLD7AmIvzuycpYLEwen8CaW4EHsZ+7RvvfLblXBWGJ1bFLo3SFL8oGE2t01Nwi7Ka0IG1CkGQdXlnjIK3lM28Syo5+KvFJbodlpJvKsj6QvWr7NTmGGHVlOxBc97hjKw3HcLnqPM6N2/09MZ/0xgGYcQrURSIPwnJg3Xn1j39d3N0jvXrQaBQ774oP75a+0dxlax4HP/DJUH4WWfeXR5/n7epUzn34fsCu7u/51ALSsyrcLdN5ecbZ/7kHuOrGt5nbkcgftAfNQyBR4ZIXXbT/0qDZTB6k2WTmLhCTUiRnUqNv3u7X4OfjGx0LL3Wnp/NRGme+491lYX9v1igPbD9i89rJEaMiSLNOUZSRBFTLN8KFvARe6R7UZdL5F0nyG2bIDsts78Ol/Be3wz+BlC3/IywDpv/3DvS/cfwpz7CSVTEbI59abOpvNzYHP2lEL5Pbt4/rYl1f4y+Xw9LJ72w9orr2JP/hCnqe+8o2Df9vKq/nM1iz/ZOvga2Hxx4tZFPOH+Kd/cOfAa9vNKs21N/HvnqrxJ1PRyrZsC017CeOVb+KfVA/+7q+Uz1Atv553f/xrj5N9m8VI7kujXeGF1dfyab3I05kP7vqjmmC+nS9U5qh9fa7v3xUWTy1LKI0f4hf/sLTn37+8pKPIcFla4y8FDCjyaN+vPPfKvq/vDr4ns9GOYg9sTyJEYF9UX4LPfxBCVsMdHN5078/cwPsP/8HeF+//KZXp80A1EV/0uMNXY4QkSRngi8AN4P2O43yuz8+8E3gnwKVLl6L8G48U+i2P9SCySbncKFNtV/typEXLHrTrYrfstfouR3ZTgz/6v0DJQx/KRhC8qNPghQ5IX/p/D7zmXHwFW/KKEEpApV1hvb4+UN+ZbkBmCsj4+dG3yD6uy5OX+1JtRFYQ2p02S9UlXnf5dX1f79uk/Mn3wMKfQR+aShBMOBZvtZoUin8C0qf3CT7B8tx1mkb0Dgw83gPWj2oD0JJkZAHn26PafMvNb+n7epL2RDTNZzo3fWDCoocD5/veH8FTvwjqOPSplPuFA/xVawdFUmDrtw+8vnPrdbDztBB9e0H/IFuGOkbecehYbTJKFtt2+D/+41fp2A5ZJVybeasjYdnfyn/5yjagH3i9bj0fA5nVh0uh5PRD3byALMl8ZOvg77YdG9N6GZ+u2nxOjlZ2x+nQtl7K16p5bi8c/N2mPYHZeRkffXoZyUsyttpA+F2HDieBK3y1n4mWztM0ZBxD4iNrAvRtnUbmLB/RD/7uv3p9HHWpg2NGX+1fri7TttvDfSOrGXlA5ueuronqAQ7qi37p38Fn3gchdeBg82azTq66Aw/v7n1Rkth48rWw8alE7o7jDl8BmeM4HeAlkiTNAL8nSdILHcf52r6f+SDwQYCXvvSlz8oKmrch/PtvfH/f173dGyIc1mGjPkVSMEpGiUl1kvmx/gslD2SBujxi/uc/gPN/JZTsN370u3nluVfynu94z4HXJKDw4W8V85kPGRMNkOku7TWbB52MKGT32wPmQXTW7XknnzfwdVEjmxcqC3ScTrDzXb4DL/4b8NfeH0r2v/3KB/jAlz/A537kcz05u7HxhX9BY/OLoWQMwqCBBx5amYyQgGy1tkrDaiRmT2Dwdyuv5Gl2mtiO3ZcuHFb2oF4qT/aBM4YE/3sRsqMH/luNTb7ro9/Fz7zsZ/jR5//ogderlSX4vTcmcnd4CY1mc5vxidOs6A0aZod/9tYX8cPfGi65+uN/9ONsN7f56P/00b6vv/0P3k5BKfBvX/9vQ8nZj7pZ5+X//uX8xEt+gh9/8Y8feH2jscF3f/S7+Ycv/4f88JM/HKnsj2sf5x98+h/zn77/P/X9bn/k9kd4z+fewyd/6JMufbVWhn9xA97wz+AVfzuU7J/65E9R1It8/C0f7/v6O/77O6hbdT78xg+HkrMfpm3yrR/+Vn7s+T/GT/6Vnzzwen1nA34RHAHBiZdwSMqWzY/NDwz08pk869Z65HJH8kXLt+HUTfh7T4eS/bnVz/LOP3onv/b6X+PlZ19+4PW1B5+AT30qEV/0uCPQbec4jg58CniDkL/miOPRziPqVn3gF1+WZGENpIeNLQaxY0693TmDnJgDn9kLyGafCCXXaBlsNDYO3TeRlL69pb2iKmRXp68e2APmQVSVqmk1Wa4uD9W3SKfR9/mub0FtDebCnTFP9oXJC32DMeiOYe8GCFGiY3e4b9w/VN9tWUEWsLQ3SXui6Rqn8qeY6Y71j0u2twcskD0pPwMnLocKxuDwFRrwuD9TlL7HlLEDe8A8SN1+2GaXft3r+TkdfvdeYH1HhGFBv8jzXTJKKJLCpan+wWyvF9ejA3v3ZQS2bNCeu55sQfpeqrh7wAZS9wrd73pb3N2RRH/90PMtaIXHSL5o+U5k9yUcTV/0uMPPlMW5bmUMSZIKwGuB26L/sKOIw/o9PAhzWA2N6dz0wN0L3pdAxICLYVn8A5+5fBumL0Eu3IU+zND2lR0RNEMjn8lzbrzPhEVA7fK2RTQpDxxb3IVnaKNeknzfcCcsJqJvXXMbwqevDJQLu853dzId84OreUFkD9M3RH/BLNdcqs1h9sTMqChW9BMeDxt5D2KdmMMGHoiUvdHYoNquBjvf5TswF/6MHTbR0pMLgvTdHXgwyInp9cM23AFF99bdKsaN+XDUpmq7ynp9PTF7AoPPt8i1DkW9yOWp/rRv6GfLvIAs3DnzaN9J+Scw2EmXZYWmJAmZGKvpGmfHz/bdAwbievs92neS+vYt22q5rSRR3JdH2Bc97vBTITsLfFKSpL8EvgD8keM4/1nsn3U0MSwzAOJ2XpT04VQbiN5p3Gpusd3aPvSL71GMegFC+XY0mZghRh5cfYvgKpf00sA9YPC4QmZH3KRcN+us1FYOdWLySh4Hh1bE+2QOm0DnQWRG++LkxR7VYj88B6r3rCPKKlu2xUJl4XB975cdEfzYEzOTRRURkOka84X5gU3XopwYbzjRMHsCAvTtw54UMrv6MzsWbNyLLKs8oU4Mpn0LDBBKeunQ8y3n3MDL2xF1b63G6akc04Vw/b8DJ+Tugnd3RA3N0A6lfWfkDFk5K2Qs+DB997VluWmYPBNK7oPKAzpOJ5EKmaZrSEiD+xTBHVAkiMY2LOiH6O3Jw52Hhw4nAnH+SWBfdFMDpwNzT4aWfZR90eOOoQGZ4zh/6TjONzuO802O47zQcZyfi+MPO4rQdI2T+ZOcyJ8Y+DMiMiIe1SaJrPLQ/oOubNuxadttd4JPhE5MQSkMpNp4skVkYobpO9c9A1Ev7fUmLA7TN4h51oqkcHnq8qGyRWWVDw0E+9F81HGYOjjSOggWq4tYtpWYvuHwS9VSsigClvYOq1KJCsjW6mvsmDv+9B3xhEe/tqz3mbdKYJuRODHDlpn2AoSI9W20DMqN8qGfOdMNyNrdfth761VuhqyOgT99jyljQiZ5arrGlekrA2nfIGZ9SNNqslRdCmZP1rsJzJA0rMDnO0J4e8AG0b4BWnIG2YzWSe/RvqePtj2Jms0S2Bctu6s1wvplR90XPe6ItmP6WQ4/C+lEDFzYbG5SaVcS/RL4yUA1zAZsL7gjdiOikl2bvnZoY39eyUduaGvtGmv1tUP1ne32v9gR74gaRiUDsc/60tQl1MzgzLiI8212TBYrA5aZdlHI7PvMXhVWDmfCvOEtvs93hNAMjTPjZwZSbQA6SpZsxPtzbMd2abGH6FuVVRRZSdaeCJA9lZ0aSLWBffbEq8LOR5BVHqLvnuyIP/Ow3jUAtTugyGoZ2LZDcb0WSf+YpmvkMjnOTfSnfYO4AUXDaMgghs2yUFnwRfuGXRWE8u1IzphmHE77hsf6jjxA8OEbtWQl8gFFK7UVWp1WIsklz5YdVhXck6yOUnZQX7R8x50Se+pmKLlH3Rc97kgDMp9wHKdXqj0MIigBfjLpovpcPKrN6bHTw2V3mo97e6IqjQ8xOiJGwPuh2oyNdZe7RrCfaDe8PWCXJgdPNxP1rP04jSIyrH6oNmpGRZGUx896/XZklQsJaSgXH6KnvPg537aSJ9uJNiDzqDbDLjaRtsxPBUEERfTGzI1DG8ILSuExja03nOhWKLlbzS22mluJ3B293rVDAzKvH7bGit6g3u5EUyHrUlMH0b5hn74jQt2ss1pbTeTu6J1vPxUbqwE7G1DfiMyWHUb79mR3nE6kS5K9PWDD9G1mlMj7Yf34RmNKd4qogGc9V5jruwfMg4i7eiRftHwbTlwFNR9K9lH3RY870oDMJ9bqa9TMWiIO67CGcE8uCMhKGO5G9mFODHSbOL3SeEgnptKusN5YT0TffoyOqo5hAUS8V0XTNa5MDaHaCHjWrU6LpepSMufbGO407pHd2Ibao8hosecmzh1KtRGh747dGTq8BcBW8+TtaKc7+qV+CPluGf6oNhCtvj2qjZ8gdE8VduYyZAdXMP3Ajz05IDsilPTSUNq32q32d1oVit2BHlFVyPzo27ItTDs6Wu79ijucKKm7IyNlhtK+oXu+I5yweNheqv2yo6T5L1WXMG3TR0AW/YAiP8MtRPVnDqN9gxhbNpIvGlEC86j7oscdaUDmE343hIvY01TSS0xmJ92dJQOgyiqyJIsxOj6oH9DNiJTvuH09+XAb2nv6HiZbgL49qs35ifODf0iS3CZlAbx0P2cMojV4C8YCtmP7etYiKkWyJHNl6srhsr1nXe4uo4xoYlQS+l6trQ6l2gA4SoG83YlMLuyiafpw3kR8t4bJFeFA9ag2PuxJL0Ao34ms0g/+AjIR+r46ffVQ2ne2O6Co065xb7078n4+XEBWa9d4tPMo2N0REZLWtx/aN+wPyMLZMj+0792yheh7yLMW0Q+r6Rqnx04zkR18Xnv9xxEneJK6OwL7olYbtrRoqNdH3Bc97kgDMp/wQ/0AcRWyYVQbSZIil73d3HapNsM+s7ov4xfRPhVISN9G8dAJix5aEQdkDavBSm0lkayb7yy+gEb4ol7kwsSF3uU1ULY3wCWiBmXLtrhv3E9G3z4m/gGQHSMLdCLcRVbUi0OpNhD9d6tHtQliTyJCkCoVQLNdg427kU2LHVfHh1JtRFVshlVCcwW3Wum0a9xdqzE3mWNmLBtKrp/eNRDzrIt6EUVWBk5Y7MkWUZE0Sr4qz+Ddl3cgNwVTg/vs/OBB5QGWM3gPWF/ZEcG7qw/rpQLoZHJkI6yEgr/znZWzkQcIfiYsglh9+/aNtkpgW5FVyI6yL3rckQZkPlEySpzIneBk/uShPyfCiRk2JtpD1E3Kfp2YXkbb3HGrFxEN9CgohUMbwkEczcePvqNuUvb2gPnNKkdt5IdRbcAdrtG223QirNoMGxPtodekXL4D6pi76y4ElqvLLtXGR9UExFyqQ89Zd+BHo7sjKgoMGxPtIWp7sl5fp2pW/duTBAOy9sZd6LQj3XM3jGoT9YALj/Y97Ix5E2Oddo176zVuRURXhAB3R8QVhCtTVwbuAevJjljfrU6LxericH1nckhIbvVi/ZloJiz6WFkCfYYjRQDNcCcsjqmHL0/vKLlI+2E92vcwWyZJkjjfKIG7I7Av2ktghgvIjoMvetyRBmQ+4YdKBtE3Cm81tzBahq9Rn1HL9rKcfjN+kr4IViOSrHLJKA2l2niyW50WthNNn82OucPDnYe+9N3OKGQi5MT77e0R0aRcMkpcmrpENnN4ZjzqgQumbfKg8sD/+facmNlboScsek6M3/MdNc3n9NhpJrOHD0+Qu/1Lre5I8rBwHMc39SPqgQtJ6lvTNSazk8wV5nzJtte+4f5DRL09fu+OKPXtUZuG6Tvf7SFz2g2Ka9GMvC8ZJbJylgsTh6+l6NmyKJ+1DyoZRJ/M82jfw/QtSdLe5FIU96VeGroHDMQMzPG7rNdW8+Qi7Idd3XFp34HujojgtwIs5HwH9UXLdwAJZsNNWDwOvuhxRxqQ+UDPiQlg5KMKEIKM+oyaTlbUi4wpY0OpNp7RUTfcvzWq0rhfpxGiM3h+xqB7MGUVJUIqmbcH7OLUEKrN/p1cEckOou+oHJnFyiKWY/nKuj3O+EXT2+NnbDE8Pt9RV8j82BOp2xvRiqhC9mjnEXWrnojD2rNlw3rIBGSVNcNflao33GXDm7AYzlnWmzqbzU3/+o74Ow3DbZksKzQkiY26zE67w42Q/WPgnm8/tO+o7UnDarBcXfafcIi4cgE+72qlAPUt2FmP7L68MOmD9h3x3eHtAfOj76j7Yf3aE4i+GlrUi8wWZn3RviG68z2KL+qsPwMnroA6eHCVHxwHX/S4Iw3IfKDcKFM1q76/+ODSF6KAXyoCuHQEEZkvP1QbgPz2gvsPITN+1XaV9fq6bxobRGfwgujbVNRIm5Q1Q+Py1OXhVBuvET6iZ93utFmqLiWj7wBUhLySR25Voboazd4eXePc+LmhVBtvjHRUQb/t2Nw37vuyJ5mc6xy3I6qQ+e5dI/qBOZquMZObGUq1kSXZHR4TcUXS72cGyGwUXUpsLlxw4mcK3G7ZUdpvzdDIZ/KHDyfqoiXLrNTdytit0xFUyALQkCE6e7JgLPiasOjJjvKMebTvYcOJwHWWp42H7v8IOdAD8DWxFXZRRCOqxC7XlmnbbV/P2lELFBwHJ6IqWaAAIeoKmY+x8xD9+R7FF3XKz0Q2AAuOti963JEGZD7g9Xv4LdVCtA7rpDrJ/Ni8L9lJZPG9zzymL8PkOcgfnjUahiAb2UXoOytnuTB5ONUGoJPJokYZkPnMfHmGNqrRxQuVBTpOJzF9+6HaeLLnat1KUUQVMj/6zsgZcplcZJ95tbZKs9P0pe9Ml9JotoxIZPvtfQAxFQS/l2pvgEsE2Gxsst3aDmTLclulyOiK4N+WRV2R9EP7BmjKGZabrt0OO2GxbtZZ3VlNxJ74HXjgyY5ySXJJL3Fx8uJQ2rcne6a65v6PkOfMtE0WKsP3gHlyIXp9+3nWUrZL32ttRyJb0zXmx+aZyg6f6Bzld6vXS+W3Ekoy+i4oBRTHQdqMzpYddV/0uCMNyHwgCI0t6iZlb5eLHycmyrJ8j2oTIAs0ZaxE1j8G/pxGERWbK9OH7wHzEGWTctNqulQbH0ZHkRVUWY3sM/sdgw7Rj032GsIP2wPmoaAUOFPvBiYhz1mPauPTyOeVfGQBQpCqoNKlxJjN6AKyU/lTzHT7hg5D1E6MXxoyRGvLgtoT2XHI60uRVWHHlDHOjJ8ZLrvbCB9VgOA34QBuP+xK8ySzE1lOjEc0YTEB+10ySiiSwqXJ4QN/CkoBBydSNotve5LJM1vbgOwETA9P/h2GpcoSlj18wiII0HeAu8Pb59eMiH7t0ZD9IEpbtlZfY8fc8RUUCdO3T1/0omkh2WZkCcyj7osed6QBmQ9ohku1OZU/NfRnPY52ZD1NPsbo9mRHWJb327QKboCQlRROVNcjm0qWz+SHTliE6JtmvSy+H3TUPFk7moBsoeKfagPRPuuiXkSWZN9VKog2AA5yvi82aqAU3IW9IbBSW6FttxPRd4/G5uNSVbv7/DrtajSyAzjp3meOIkDYaGxQbQ+fsLhfdhQIWjW5YFlkOhE5MV0n3Y8TM6aOYTt2JEuSq+0qa/U13/o2ZZUlcy6SgR5BEg5R9wBrepf2fcgeMBGy2522rz1gPdlqgdM729FOWExC34Y/2jdAptsP226Ep197tO8g9iRqmr3fvmeIVt9BfNHrZteWRLFP8Rj4oscdaUDmA94yU1+XaoQDALaaW2w1t/xln4iW5hPEiQG4guLS9yIqjftpCIdoA4S6WXf3gPnMxDhKnlxETco9fSeQ8SsZJS5NDp+wCNEGwB7Vxk9gAt2ArFV3p0X5OBuHIUl9a7rGfMEf1SbbrWR1WuEDsiBji8H9zB2nE0mAEMRp9GRHqe8JdcIX1WZMGeN623NiIujt8blCA6K1ZUGqVABtWWXJmudmRCPvVVn1RfsWkeAJYk+ikv2g8oCO0wlkT842KpEN9PBL+456QFEQfctd+nUUE2NXa6u+9oB5iLJiE8Q3UmUVRVai1XcAX/R628RBcqcSh8Bx8kWPM9KAbAiCTLWBaCmLQXcvRNmkXDJKLkXMB9UG4Amr26gbgRPjl58N0Q64uG/cB/zr21ELFCJqUC7pLtVm2B4wD1FXEPwa2igpGEtV/1QbcJ/1tXYbO0JabJBzFlWTchAnJpdzAzI7goAsCNUGon3WI9myiPTt7SryRbXJ5LnWyyqHc2KMlkG5UU7k7vCoTX5lr8mz1J0CNyMY6KEZ/mnfUY5hb1pNlmvL/s93hAMugp7vE7bESbMVTeVCL3Fu4pwv2rciK2SkTCR3R5AJi/B4QFEU9OugVZMox7CXjBIn8yc50d3fN1R2RAMuRvFFr5smrakzkB1ewTwMx8kXPc5IA7Ih2GxuUmlXAn3xIZlLNeqs8vXp674awgFumN1KUUgnptau8WjnUaDSOEQz4CJoFh91DBWwIhjvqukal6Yu+aLaQHTP2uyYwag2CZ7vKQfOdjq0T/kLZg6DpmucHT/LeHfx8jBEpW/bsQNRP7IF99LvtGuhZXyWJwIAACAASURBVAeZSgbRPmtN15jOTfui2niyoxqbHIgWqxa40TapFWYgFy44GcVphGj0XdSL5DI5XxMWAe7b7qqNsAM9oKvvaX/6VmWVjJShboa33w8qD7AdO5nzbWjIksyV6Su+fv5Sq/t5I6LF+j3fkiRFZstWa+4eML/nW8m5jACzFb5CFoQ2CNEuKg4SFEF0d8covuj1tslOyB5FOF6+6HHGs/8ThkTQcmmUWeWiXmRcHfe9eyGv5DFtEyuCvqYgWXyAa+0WRrYABX9Zo0HoVS6CVmwicN6KehFFVrg4efgesB66WadGBE3KQRrCIToKRo9qk9D5Brg6NZxqAzC3407naobsH4Pg5zuqAOHhzkMaVsN/Za7gjoh3zJ3QsoNM/IPoK2R+9oB5iOpCD0q1ycpZrpsWG5OHL5D2gyTvDs3wT/sGWLTdft0oJiyu1FZ8n+89S5JDIigNOcq7Q9M1Lk5e7K3IGIbzjW7FO2S137ItFgz/tG+I7u4Ier57A4palUhkzxXmhu4B8xDVXqxelSrAgIm8ko/MP4EA9kRWuGKaVKbPRiL7uPiixxlpQDYEQcZEQ7RNnN5uEb9OTFT9PUGpNgAXm3VWCvE2hEO0lJeSXuLKlD+qDTxe2ht2R1Sr03L3gPl0GiE6ymLRCGbko66QnZ8476shHHCnkgE7J4ZPUDsMHbvje2+Ph6goL0HtSS43jQ3QjsZJD0S1iciWeRMWgwbAkerb5/mWHJurpkl5/PBdaX5Q0l2qzdlxfw5RlHeH12viF4vWWU5Q4dSEv4BiEO5XgtG+Idpn7XcPmCcXork7gur7bN2gLkkw7TP5NwBL1SVM2wxsy6IK+sF/8lTtrsOJgn7td8+dh4JSoG236YTs+V6vr1Mza8ErZFHSYn0+68naBllgeyo81e84+aLHGWlANgSarjGVnWK2MOvr56OmnQT94kch26tS+c2k4zicb9RYzIfjKcOuPWAT/srsUevb92fm8dSoViPcXpUFYwHbsQPJjqpptqSXXKqNTycmyp69ohHsfJ+orNGSoBrSWfaoNkH1nQRNU5JlmpKEFAGlK6jTGNV3y6PaJKlv37L1B+Qdh9WxcLsU4bG+/VJtotJ3UNo3wJI5xw15OZRcCB4AQ4QBwgi0bwivb4/2HUTfc7UtSqqKHXLPbdCqN0Sr7zPjZ5jI+quqRjWgyHbsQDRNcPu4IPy9lfT5DuKLjutLAJQn/NHED8Ox8UWPOdKAbAg8vnAQqg2EzwxsN7fZam4FprFFITsoPxtjibxtsZDNh5ILwak2UX3mullntbYaKOvmTY1qN8MFZEF7eyC6pllN17gwcaGnx2Ho9SCEpGB4VJsgmd3Jyir3VZVmyKl/QcbOe4iS5jNbmPVNtQFoyRlkM3yVqqT7X+kA0dmywPaE6Ho+NEMLRLWhfAeAlUI0vVRJ6NsbTuT3fDuOw3LrJE9IK9ghKUaargWjfROdLQvSmwnR6XuxuojlWIHsycnqOlpWjey75WfCooco744g9jvXrczb7XD060c7j1zadwLJpcB95kSs7wC+aG7LtQPlsXBtJMfKFz3mSAOyQ+CNiR7lIEaVGRjlQg9bOdF0jYJS8LUHDOg5McWsv8zkYQjqNMqSHInz5u0BC3KhK92m/3Yr3NQozQhGtYFos25B6QBRyF6uLrtUmwCyx7eX0FQ1kvMN/qkfEGHFJsCeOw8tOYMU8mJbr69TNf3vAYMInZhRsvhqNDvQgvausf4MAIu54RPrDkOlXWG9sZ6I/fZ6Tfzqe63Som7nuCUt0wy5I8qjfauy/7sgiu9Wu9NmsbqYiJMeVN80dMabRjS2zNAC0b4hGn0H3QMGu/phQwZko9iTqPwyTdc4kTvBybx/lkYU+h7FF1U3NZaVDFXCTYI+Vr7oMUcakB2CzeYmRssI5LjJkkwuk4vMiQnUPBrR2GRvD5jvqTbl2wDcVcLxL+pmndWd1cAOaxTVi1H07U2NskKO8fUawv3sAfMQRRBq2iYPKg+C6zuCEfCBqR+tKrnaGlpWjeRZnx477ZtqA4+znGEChFEawgHaGYVM2EyjMbo9ieJSncxO+qbaeLIdnEjOWaCG8PIdNtU824TrNelRUwM2/0M0ybysnPU9YfHeukshuyGt0GyGG1AUlNoELp0s7Ge+b9zHduxkkqd6CQnJf0Jt4y5AZBWyoNWDKPS9Uluh2WkG03e3QhZ2QNEoVZMok0tBB0xE4Z+M4otK5bssZJ9jvugxx3PjU44I71IN+gWMIiOi6Rpjylig3QtRNSlrRkCncf02tdw4ZScc3SXoMlMPUelbkRUuTvmn2qhdTrwVciT5qFWqZidcgLBUWXKpNkEv9Cj0HbAhnHLXiVEjcGICZhrB/cwODq1Oa2S5j3YeUbfqgWWbGYVMpz2yXBi99wGisye+q1REQyfTmzqbzc1g9qT8DA/HpqOj2gQcPADhaT5Fveh7DxjA3TXXft2Sl0MNKGpYDVZqKyPZ77CfOUwWPwpbdmHSP+3bq8JqarjkUo/2PeLdEQY93yjA3ZFRsrQi6IfVDI1T+VPMdO9fP4jiu+VVqYL2M40pY+HPd1Bf1O7Axl0W8+PPLV/0mCMNyA5BYCpCF1EFCEH4wuDSfCDcGN9Ku8J6PRjVhvJttidP07AaoQKEUTeyR6XvoFSbbHdqVCfEGN92p81SdSnwZx5Tx7Adm7Y9uqOe5Pku6kXOjZ/zT7XpVmHDVshGodpANM7bKP0HAO1MFtUKH5DN5GaCUW0isCcQnIYMCenbtqF8l/WJk5E46flM3neVCqKt2ATRd3G9yqRqcooK7RADiu4b93FwErHfRb0YmPatyAqqHE3FPdh9eYdOJsuqEm5B83J1mbbdDq7vCEbAj3pXNyUZKaTsIHsFPUTx3So3ylTb1UQSmIHv6u0F6LRYKUw+t3zRY440IDsEJaPEpDrJXCHYTpq8EgGlyxiNigDhshJBp8DhOFC+Q2XqLB2nE2rvREkvocoqFyaDLTKMoml2FH1nPQpGiArZQmXB3QMWMAvUm3YY4nNrhuZSbXwuM+3JjoCCEXRsMeXbOJksy4oS6jOv1lZpWI2RsvgQUt8jUD8AOhkVtRNukEnJKHFt+lqwSzUCe7LZ2GS7tT1yQBaJvv3KNhbBarAxMZsI1UaVVRRZCaXvHu07wBm7t1bj0qSJJIXbETVKFRaisydBad+e7DBnzLRNFirBhhNRfobmics4khRNwmGEuyPsGPaSUWJ+bJ7JbLBVNy1ZRrJGZxl4tO+grKUoxrCHOd9RVIAD+aLdKuyjsennji/6LEAakB2CUTIDED4jYrQMNhobI1UuIFwWKDD1o7IC7Sq1GZfqF6bfxJuw6Jdq4yGsvptWk+XqcvCsWwRTo0Y1OpFUEHSXauP9riCyw8jt2B3uG/eD6bt8G+fUTTohnZhRqE0Qnb6DUm0ALCWH2hk90eHtAQt6vtWMiiIpx1rfY8qY7z1grLtVWH3qTCQB2SijmsN+t4KOiXYch7trVS7PuI3/ZgjKYskooUgKl6aC7QmMigI9iuMWVvZSZQnLtgJXyNqn3J8P9awTbqkY5XyH7Yddq69Rt+qJVMjC2DLLsTBDJNUC+6JdRsnGxKnnji/6LEAakB2CUXp7IPzAhVH6DyAao1PUi+Qzec6N+52w6H7xGycvhZY9ysADCJ9h9ag2QfWdG/MqZKMHZEW96O4BG6FKBSED4BD6DkW1qblUm0BZt/JtpPknkSU59PmGEb5bETQpj+o02kqeXIjK80Zjw6XaBPzMEP67NWpVMJLz3c3sBnViKlNnQ33mWrvGWn1tJH2HHbgQ9O4oV1tUmhZXT7pJMLM9+o6ool7k8tTlQLRvCH/G2p02i5VgExY9hA1OAtNimxWorNA55Tq4YSpVRb3I2fGzjKvjgd6XV/JYtoU54voQ27F7FfegaGdUMiEqZD37PQJtEML7RtO5aU7lg+31imI40ii0WKYvQnbiueOLPguQBmQDsNXcYru1PZqRV8M1KY/aaxJJVkIvBdoD5mWV2yddPY36uetmnZXaykhOTNim2VGpH1l1kg7ghGhSLhku1SaXyQV6X1hKl2VbLFQWRnMawzoxgScs1kBfRJp7MhLZ84V5prJTgd7X48SPKNvbAzaKPbHVHFl79Kl/o/YKQviBC0W9yIQ6wfzYfGC5EJ5iFJQWy+RZ5LETidgT6Pb3hDzfqqz63gPmDfS4PufqO0w/bGAachcFpYBpmyPT3R9UHtBxOomcb013ad++94B1V8RI808C4XpsSsbo+obRv1sPdx66tO8RkktmRkUJ0Q8bhjYI4X2joMOJ4PHdMaq+R/JFy8/A3BPPLV/0WYA0IBuAUXZdeBhTxkJfqgWl4J9q00Uuk0MiHKUr8Nji8m0Yn0OZcBevjirbW2aaBM1H0zUUSeHy1OVA75NkudukPHpAVtSLozluIQ3eYnURy7YSc2IggJHvjokmooBsVGoTjK7vtfoaNbM2kr4dZYy8PfoumTDUj4JSCJXZ9faujUL7htH1PRLVpny7d8badpvOiEGwRyVLxJYZWqAJi97I+1tnXBqt3RqtQta0mixVl0b+zN7vGAWjOume7LD25PzEef+0724VNnP6hcDoPTY92vf06Poe9XOH8Y2skP2wmq5xMn+SE/lgy47DfmaP9p3E3RFY33YHNu7B3JPPLV/0WYChAZkkSRclSfqkJEnPSJL0dUmS3hXHH5Y0Ri3VQjSUxWvT1wLvXpAkKRSdzKPaBC6Nzz35eDP7iBdMbwz6iPoO0zyq6RqXpi6hZoIvtm7JMrI5mmyzY7JYWQxl5Ed91qP2H3iywzqNZ8bP+KfadLPKzD0Z6ll7VJtRacgQ4nyHsCeoBfKOgz2iI6PpGlPZqcBUGwhPTx2Z9h3SngQOQrsTFpl7MnRDelEvksvkRlpmms9EoO8gAz3Wa0wXVM6cDNcPu1BZGIn2DeFHZGuGNhLtG8IPuAhMQy7fBiVP9mSXsjiiHV2prdDqtBK5Ozxb5rsquAuWkg0XkI1I+w5rvzebm1TaldH0HdKeBL479AdgNXv35XPGF30WwI+WLeB/dRznecArgL8rSdLzxf5ZyUPTNSbUCU6PnQ783tCN2SOMiY5CtufE+C6NO04vq+xNMRqVgtHbA+aTarMbUTTCj6rvlpxBHtHoeFSbME5M2Kzb1angl6q3x8Z2RqvaeNQP3yg/A7IKJ6+FGtn8aOcRDasxMi0Wkskqk3UD1+aIAxe8BvygVSoI993abm6z1dwKVwEe8VkHHsttLIG549J8wn63ur1ro1Btwujbo30HsWX31qrcOj3B2Ji7tHvUfthRewUh/LPWdG0k2rcne1R9e3vAAtNiZ2+R977TIYOipO6OucIc07npwO+1lTzZEQcUhaF9S5LkPusQZwxGr8JCOH0H8kW7bSTMP++544s+SzCU1+A4zkPgYff/r0qS9AxwHviG4L8tUWiG238wihMTpkm50q6w3hh998KeL8Hin8P9p3y/V9n8Bj++bfAttz8B958e/gazAa0KzD0RmqM9yh4wD56+bccOnMlpdVosVZf43qvfG1guuFOjRg3IisbovT1R6Pv8xHn/e8D6yG5aTcasNnzxt8Dn4mLbsfnu+1/gRbMvgj/5f/wJvPvfYfYmZJRQQw/C9FL19D3ihV4ySiNRbQCkrvPWaGziKCf40J8/oGX5C4Ydx+FLd05zY+YG7/vje4Flry69GMu2RnrvSm2FVvnV/OXd87zvUbD3tzttWuVX84dfklhbuec6sjsbvt//wFjgJY3v4Hd/57/5s+ENHay3wOJlbndqtMqv5gOfXGA6F3wv1xe+McvZiReNpLMHD55PtV0Z6b3r9XVa5VfzTPEK79vw9/7bD6u8+cXnyHUnfzrm6AFZ0D1gHsIOcPGy+KMgr+Qff6e3F+Cr/8FNNPpAtbnN/7K5wWsXvwY1n7Zs9S/g+muQJTlU9SJMn2IUd8eo/klHyY/cDxuG9g3hAoRegieMvs2Ge7ae/g2ob/p+//OL/5lbtoX01C/6e8PS593/zt4i//CTR8MXDYhQCcxjjEDzxSVJugJ8M/C5Pq+9E3gnwKVLwcbeHkVousarLrxqpPd6PTaO4wQO6Hpj0EfcTr6nv+e//G+w9lXf731+9//4zPv9C1TycOnbHmeBRqR/aLrGC2ZfMNJ7PdmtTivwCPcFYwHbsUfWd5gm5ZJecveAjeDERJHFD2NoPdljX/tP8Il/5Pu9MvC3Abb+BO7+iX+hr/i7Pdkj09j00bNuYWlVRb04stModwOydlPnqaVH/MJ/vR3wN7yKv3gEf3H77gjSXTLEex+M8l6A1/O7G01glPe/nk9uwCe5i3tyggwGcX/2T323RM0Dt+DzDSADvJ5ff2oNWAsg08O3sQl87e4on/kmAO9dGl3fH9vo4FffsgTffv0UmYxKU5KQ2qPb71Fp32G+Wx7t+zWXXhP4vZ7sntw/ex88/eu+33sC+AkA/T8GkCjBtVf1ZIdxWE+PnWYiOxH4vWHuDsdxKBkl3nLzLYHfC+Co+ZH7YcPupQpDdy/pJSazk8wWZgO/d49vtPZ1+C8/Hej9PU2v/Lz/N51/KeSnjo4vGhCarpGVs5yfOD/S+48rfAdkkiRNAP8R+EnHcQ6MYnIc54PABwFe+tKX+ksxHVH0qDajfvGVPA4OrU6rlx3xizClcdjVv9axYOMOfNtPwOt+ztd7/84f/x02G5t85M0fCSBRAlkmv+M6LqMY+YbVYKW2wvfd+L7A74W9Gb+gAVlYfZsZFSVEb8/FyYuBzwiEy3J6VJtXnn9l4PfCPj5++TZkJ+FnF8CHwf/U0qd41yffxW9/72/z4rkX+xfapX/llTxG3Rjlz0YzRqfahMnie1SbN157Y+D3Asg5d/lqu7HN3bUZ1IzEV//x96BmhleDP//wc7zjj97Jr772A3zbuW8PLPtnn/pZvrH5DT7+lo8Hfu8vfP4X+Lj2cf7sh/9sJKbByz70Mt72/LfxUxe+Fz7wbfDXfhVe9AO+3vs9v/s9vOzsy3jPK98TQKJry/50+Sl+4n/8veBnFPhq+au8/b++nX/1Xe/l1SMECe/+7Lv50+U/5RM/+P+z9+7BkezZXefnl496SSWpuyX1S63uvvf2fcwwHs/4Yry2WSbAMB7Mjr0OvNgDtncXwiywC+xCGJuNIMwfG7HEboBNgAFje8ewBi/hHbwOYwzGNmtgmFmPx3g8M919W1VSS2p1610v1bsy94/MLFVJVVJm5e+XOZLyG3Hj3lZL96hO/vL8zuN7zvk3gX/2hz//w/zjL/9jPvOJzwRiGuia82yamoaYsEpVLBd5dO3RRD8bZqXEenWdrt2drDeTE0HRzmO493vgv/mXvn72H/7OP+Tv/Nbf4T9+138MxjQYsGVhArJJqwdh7o5XR6+od+sT35e2mSVrW9iWhdCCsVkmnfjnIVQAXJ6c9j2k77KbTPvT/w5unp+APmwe8pF/9hH+0rt/ie953/f4F+oyhb4ifNEJ4DHUrtKERfAZkAkhTJxg7Kdt2/6U2l8pfoQ9iINNs0Ffgv7uhQkawmFgbPLhmkMjW3xf/wI4V3Z5lQ8tfsj395+UC5Nx4r09YJNmYrz+nklkr5RW0IUeeMKih66RYqox2ajowGO5BxCmMXuz6uwBmzjzZQ70fOw443XR/eV2ViqrWELw2rVHk52zkFm3SfVtaAamZk4ke7exS7VTndieGCknIOs0yzzbrvJwfoqM6U93q5UiQti8ef1R3/EOgqlUllavMdHPrpWLvDH3GoaPwHEUcqkM7V4Tff8JCBtuvQPG+ees0q6w3drhjeuP0H18/0lMpXIIYdOxWoE/91o1pL7NcPp+OHufjJEK/LMAbU1H6wTfEdXqtVivrvPRBx+dSG6YuyMMDRkG2CyWhdh9DO983LddelYucju/RM5NmEwiexKH1bItVsurvHvr3YnlQjh9T3p3CHMKDWh1qqQDJsYKpQLX0te4nrk+kexJ9W3bNoVSIVQVFlx97zwGoTt3po9zVqishb4vPdmx+aIToFAqOL7oFYOfKYsC+AngsW3bf1P9rxQ/+hO6QgYIkxzGYtnZvRC0F8pD3+jseo2db/v6uaPOES+PXkqhsQWFrAB4Un3fy98jpU/mxPSMNKkJ9ud0rA7PK88nPmOmZmJoxmT6Dplp7AfAvaYzAdHnGQOHBrGYWySfmtyJmbRKNeki7EHZk+g7rNNopJ2daZ1WhWc7NR7d9K+7QqlAPpVnIbswkeywfRdhpmQd27KngID5N339XJix8xDOfntUm6XppYlkx6nvljZZP2yf9h2H/Q5B+/Zk9+wendpLaBzC4juBZIcZOjBpculF7QXNXjM2/wQmvzu8fthm/SDwz4ZJqMHkFcmD5gGlVknO+d59AtdfA8PfAJpL4YsGRFhf9CLDj6a/Afhu4PcLIf6T+89k3JsLgpXSCjkjx62pWxP9fBhKwEppJVQjY79MvPvY+cL8W75+Liw/29RMDDFhgODuAVuemaz3MIy+w1A/wG1SnmBq1HrFodqEclgnHHDRn9A1aSO8SzFq117B0Q4s+A/IQp/vCS/VsFSbMLLD9K4BmBknk1w7qrJ+UOfRov++kUK5MNEyUw+T0k5KzRL7zf1w+vZk7zyGaw/A9EdHDrVigPD2JMwy04yRoWt36QSkQXu07zAJh45uoHeDV8hCJ9T0yQOyldIKS/mliWjfcPys26++4Hxhwd992d8DFoMtk0Elg8n1PZ+dn4j2DU6FDKDZCDYsx0uohdH3pAGC1ITx7hPfZwwuiS8aEGF90YuMcwMy27b/vW3bwrbtr7Jt+6vdf34xil8uLnijPid1YiYdcFFtV9mp74TKAvWblHefwuwypP05b2GmNg3JniDjVygXuD9zf6IJi55cCG502r0269X1UPq2zQzpCZqUwxp5mHzARaFU4M7UnYkmLHpyATRvktuCv6yyR7WJI6sctioITrZxUprPXHpuYqqN6To/G6Uutg1vBqiQhRlbDI6+u3aXjhUsQOjvFZTxrHefBqpcFMoFMnpm4obwMJSusFn8Se+OtfLke8A8TNoPWyhPPmERwjMcwp5vAGvbHRrt05Zt1hzad9hnHSqhNind3aOITnB3BF5ZcgKa64+0W8ECMo/2HcaeTLoXK6xvZGomutDptGpwUAxchb3wvmhAyPBFLyomq0VecoSZQAeTj8j2ytNSsm47wTIxxVLRodrkJ6PaDMkOiLBO46RZoLWKQ7UJo2/byJCZYB9XoVxAICZarulhaGRzAIR1Yjx9m/uO4fR7zrZqWzR7zdDnu2N16AakiYbZkzQoO4zTOOmlmso6o/Kfu2vI/FbI9hv7HLYOY3m3ZIwtzhpZ2u0j2F8JZMu8KtWkVJtJB7jUO3W2jrZCn28IfneEpcUCdPQU5gQTY4ulcLTvSc9Yx+qwVlkL/U4DDqMkPQt5f5UIWed7kgChWHZo3zOpmYnkprQUAkG9E+x827bdH7YwKXRvQFHAnYqy9D1pAJw38yzmgkx5PYa3JDlTfgG2FYhRcil80YCQ4YteVCQB2QmUW2X2GnuhK0UQPAMlw2nMGTnanQbsvRe4NP5g9gGGFrwJ3sMkBq/ZbbJR3ZCS5Qx6uYWl7gHYqRxp26YX0JHx9oAFnQo5iKyRDZz58qg2MvSdOViD1DTM+jOcMqqCkw5wKZQK3MjcYM7dtzQJJnGgZPSueTui1ispDE1w/8aUr58L238Ak79bxXIxFNXGkz17tA9WJ5gTE2JPkicXgn/m1fIqIEnfAe+OYrno0L7zk6+c6Rmpifphw/aupfU0AhFY3xuVDbpWSNq3q2/dC/p9Jk08WxYmoRamJzXMGfOWJAc9Y9v1bY46R6GcdG9AUbcZbBBW2KoghGBXlCbfSTsoe7a85fzBp192GXzRSRklYX3Ri4okIDsBWS8+TJZVTuvpiafaeLLnW0fQawUrjZfDURE82ZNUqcJSbcLoWxMaD2YfTCxb8zjxzWAUjLBOI0ym7xe1F7R6rdBUMoCp0kYwJ6Yc47sVMtPoyQ4qd6+xR6VdCfWZ0y7VcfMoy8P5KVKGP7PtVU3i0Le3dy2sEzN/5C5Q9RmQVdtVtuvboXvXYLLPDPHpe9I9YB4m6Ydt99psVDdC2RMvQJjknYbwtG+A1P5qoOFEhXKB21O3mTL9JUfGyQ7qKHu07zhsmYwEpuH2w3ZawVaXFMoFZtOz3MjcmFj2pAFwWEaJJ3uuuu2Mo7/hbz3EZfBFW70WvYCLwGX4ohcVSUB2AjKoH5M2KRdKBV6bDbd7IWNkeKPt9gH4dGLqnbrTEB7S6ExSou7rezYcFQGC67tYLrKcXyat+5t4NBL9JmX/U6P6VJs49R2S+gEwU94KXLlYzE5OtYHJ6E3eHrDQ+tYzgWlsntMYivqRdSpkm/UZHt0MMNCjVGDanOZm7ubEsid+t2To28hw+6hMoAmLEqqCYQIEUzO5l783sexJ7w4Z+raMNKmAztPzynN6di/U+YbJbVlY2nfWyHKt18NsliKtwsJkQw+2als0uo0Le3ek3H7YXsv3xnaAPssgbIKn2WtiBWgxOGgeODtpQwYIGSPDfG0Prj0E098Amsvgi0Kw6pwsX/SiIgnITqBYLpI1suGoNhPuVQnLzwbH6LzW8QIyf6XxPtVGQhYoMLWp5FBtJt0D5smF4Pr2svhh4C3tbQWokG1Uw1NtYEJ9uw5r2PHBMz2LXLMSuLdHxvmGYBfMdn2bWqcWvgJsBte3DJqmphmUSfGqneeNxQADPcpFKVQbCPZulVtldhu7Us73UrMKc8uQ8jeARtaErknerUKpEJ72bQY/381uk83aZujPbJtZMgEDMhnnGya/O2TQvl9vB7sv+7RvCfak0W0EChDCjp3vy57w7rieuR6K9p3KOP2wVrvm+2f6tG8JxV3flgAAIABJREFUQSgEs2Uyz/fNo4PArKWL7otCMFsmyxe9qEgCshPwMjGTNoTDZJSXWrvGq6NXUjKNb7Q7dPO3wOfCShnUD0/2JJmYsFSbSfTd6XVYr6yH/sy6u1clSJOyLKdxUn2HpdpoQuNtz28LMGGxWC6GP98TPOv+2PmQF8wkGe1CKTzVBuAJd7HReDNghUyGPYFgAy5kOY0ZI8O9ZiOQE7NSWiGtpyeesNiXPeGzDlPp9+RCsPPtDSeSEpDZNnaAqbGFcnjaN0xIoZNAQ87oGV7vJzD9nbOt2hatXkuKbHAWa/uFDNogTE5ZDPuZU2613wpQIdtv7lNpV6QERTDZ3RFW9rRIs9ioBU5gXnRfNKhsWb7oRUUSkJ1AXFn8fuVCgqF9rdOhdd0/jWOltIKhGaGoNp7swM2jEi5VXdNJaalAAy48qk1o2e7S3m7TPyfeoyI8nJmcagOTNc3KON8Ab/Vs5z98XjAvj17S6DakvVtBzpkM6ocne5Lm6LBUG4CnOO/mI58VMo9qI8OeQDxZ5ZyW4n67he1zlyI49iTMHjAPQZ91vVNnq7Yl73xPou+wfRepHBrQCtDfUygVuJe/F472TfABRV2ry1p5TYq+X2936JhZmPHXL+PZkzgChJXSCgvZhYn3gA3KDnLGZAwnAkh7FbLOke+fka3voHfHlDkVivYNcL/TRscOTIu96L4oBLNlsnzRi4okIBuALKpNP0AIkhmQVRrXUrzW6VKf8z9tq1gq8mAm/FSbnJELNF611WuFbgj3kDWzgWSvlOUYecMNyII0KXtUm0n3gHkImuXsWT1pDbNvdHq0dBNm/RlOWU5jn9IV4Fl7VJtrrjMwsWxX37Zt+/p+GWOiPazYS2j0eDjvc8KiROoeBMxylgpkjSy3p26Hkn2zVScFdOf9B9LFUlGOPTGygaqCq5VVbOzY9K0LPRTtGwaX9vrvh/V6TcIiawSz3xvVDTpWJ3ySxXQSmKX8zeDDiWJwWIulohR7EvTu2KnvOLTvsFVBd4WH3fYfkMm6OyYZAe/dl2ETasstV6bPgOxS+KITViRl+KIXFUlANgAZuxc8BKWTFUoFZ/fCdLjdC3ONChnbpjrnfzpO2I3sHoJ+5rVy+D1gfdkBKUbFUtGh2ky4zNSDt7S32/I/xnelHG5MtAdP334DBI9qI0PfD9sttnOzoPkzIdKM/ARNyjJ6BcG5YCzbom35W3Gw39yn3CpL0XfRussdfc/3hEUZe3tg8gDhtdnXQlFtABaPnL7M5rUHvr7/qHPEy6OX8mxZEMdNIg0Zgut7eWZ54j1gHrSUQ4f12w/r0b7juDv6+g6b4NGzvNHusJ9f8P0zhVKBW1O3mE75pw+PlB3w3bJsi0I5PA0ZJvNPIPz5Ns0cHYAA71ahVGAmNcN8dj6UbG9lStC7Q8ZdfbdRpQcw72/C4mXwRSetAMv4zBcVSUA2AFn8bAhOCZBFtZmtvASgNOMvOy2LagPBaScyxroOyg5CRSiUCixNLx0vBp0QHife79Qoj2ojo0qVNbL07J7vJckyxs57uN+ss5X1P2CiUCown52XQrUB/7QTWRMWh2T7fK9l2pNV6w739Ze+v79QLkih2kxKoZOh7/nqLgD1OX/9YLJ6BSG4PZFFtZlkz56sqnfQpb3r1XW6djc2fUO4PWAAZrPCDctiZ+q675+RQd2D4A7rq6NXDu1bVkUypt6ehqahBQzIXp8LX6Xy+qn8nrPD5qEzYVHCZ75VL7FlpsD0N4DmMviiQe8Omb7oRUUSkA3Ao9qE2b3gIajBk0VFmDrcAOBg2l/Gz9sDJuOCyegZulaXjtXx9f2FskO1CVulggkuGEm9VN4YX79Tozarm3SsjhTZ3gXjl1olzcg3SlzvtFjP+B8MIstpDFpB2G3sUu1UpVUkg8iWVaVqdXu87C3wUHvh+2dk9a4F/cyVdoWdxo4UfV+rbLNp6Bxp/j6DrF5BmKxi82DmAaY2+XAiAEMz0IXuW3ar12K9Gn44ERxXyNo++2H7vT2S7o6gAYIM2je7TwB4OeUvUeTtAZNivye0J1JsWVB9lwpcS1/jesZ/4DoOLU1HdPwn1AplebRY8K9vWb1UAIu1AwqmfxreZfBFgw4UkemLXlQkAdkACiUnMxCWagPBshL1Tp2toy0pjkSutM62rlP16cTIchphsgrCvfy90FQbT7ZfuR2rw/PKcymfOZ11Liir5S8gk6rvgCNtC6UCN3M3yaf8V7ZGYu89ANZS/hr5ZY0thuCXqkyncZILPZ/Kh6barO4dYaHxutj0/TPSEg5aCk1o/j+zJCoZQL6yRdE0fZ/vYrlISkuFnrAIk2WVZZzvoDvQPNq3DNn9flifAVmxVAy9B8xDYH1L6l3zArKtjD+b+KL2gmavKfW+DByQybBlAVd4yLInAG3NQPcpWybtO66EGr0Oc0f7rJiG/2T1JfBFg67wkOkbXVQkAdkACmU5VAQIlpWQsczUQ+pglZWUGeglMDSDezPhp9pM8gLKGm8aRN/rFZdqI+FC7zcp+5waJashHCa40CU5jew8BqCY8hdIvzp6Rb1blyI7racRiOABQkzBoIwq1XvbTrD/htjw9f2lZon95r4UexI0QJBGQ7Z6TJVfUAhoy8LuAfMQ5DM3ug02q5ux3h0y7ImZ8fph/VfIlvLhad9wvJPLDzzatxTHbfcpdU3npeGPntWnxUq0374TauUCNzI3Qu0BG5Tdttr0fOyd86pUshzltm6gd/2N+pdNQ4Zg/knOyIXaAwbAfgHdtiik/CeXLoMvOom+ZfmiFxVJQOai2q6yU5dDtYFglBdpVATLQt9foWgGcGLKBSlUGwhWom732mxUN2LVt5QKWXoOC7Db/mmDd6buhKfaEGzAhUe1kaLv3ae0NYPnmr9hIjL7D4QQwZ51ucBcek4K1SZohlVW79rKdhUNi7fwR1mUvcslCL2pUC6Q0TPhq1SHa2i9DoUgtiymBM9aeU3KhEUPQc73SmkFTWhSqlSmt8LDZz9ssSznfINjy1q9lq8A4UXtBW2rLadis/uYzcwUDZ99RTLfrUnsiaygKMjdsdfYo9quyqlIAh3dQO/5G4zk6TuWimRZTu+aV4X1a8suhS9KcMqiTF/0oiIJyFzIPIgQbMBFoVTA1EyW8uGm2lBeR3QarKRM342r0qgfBGtIX6usOXvAYsgCFcoFBCL0MlMAoWk0hfDdpCyT+hFkwMVWbYtGtyFH37uP2Z++QT3AGQM5WTcIRsGQ1RAOwc73fmOfw9ahFHvy3naNm6kyM7ZPJ0aBLQtiT6RQbdwqrN+sske1kXnGWr0Wln3+kmRZe5IGZfumaZaKLOfDT1gESHk7onz0w3asDmsVOcOJ4NiW+VmSLJOGzO5TXuVmA9mTxdxieNo3wSpkMldoDMn28V7L7M0E6OopzK5/W5Y38yxk/U/BHIdJWipk0WJtBKum4Uv2pfBFOU44xOGLXlQkAZkLmaVaCDY2uVCWRLXZcTIxm+kpX7Kb3aZDtZGYiQF/GRGZVDJPdpAs0N3pu30DHRYtTQcfsntWz6lSSTxj4G+vSv98S6qQHeQXA+n7eua6FKoN+A++ZS0z9RDofEu0J892qtzLlsnYNpaPiZrFcpGskQ1PtXERyJbJqlK5WWW/1f7V8iog156AP+etWC5iCIPlvP/dj+fJDpJckuXEBAnINiobdK2udH37GVAkjcZWP4DaNq+mrgWrwkoOQv3I3q5vc9Q5ivXukBUMdowUpt/JwG4CU0ZCLYj9LrfK7DX2pNmyRn6RpuavF/dS+KKAJjSHXRGDL3pRkQRkLlZKK2T0jJSpNhA8q/zGrAw+vDsxKjfrS/ZqWc4yUw9BLhiPaiOjSuXJDpL5ktk42tJ0dB9TozZrm7Sttnx9+8h+eVnO0JdqswKVF5RnbtPsNn3tQJOtb7/Peq+xR6VdkZ5V9nO+ZWU5212Ltf06SznHSW75GEnu9a7JaAgHpzLox57U2jW269vSgv7ezB3qPp0YFVUq8P+s78/cx9TlUG383h3tXpv1ipwJixBsaa9sWmyQ6sVKaYXbU7eZMv1PeR2J3acA7Ofnfcm1bEsqTTMIpUvZ+fZ5d8ymZ7mRuSFFtmWkSfX8B2Sy7g5NaKT1dPRVqt2nNK45S9v9PusL74sGlC3bF72oSAIyF8VSUcruBQ9+s/j1Tp0XtReS+PBPIX+bXnranyNRlk8lA/8VhHv5e6R1f5P6/Mj2syS5a3VZq6xJ3XXR1g00H03KKqgI4N9pXMwuMpOaCSfUdWKO5pawsc+lGMkcW+zB77sls//Akwv+L9Vpc5rF3GIomat7R/Qsm3vTzqXW9LG0V9bYYg9B9S2LFmvPvwX4PN9lh2oTdg+Yh6DvVhz6fl557tC+ZQdknfMrZIVSQdqERQh+d8i5L50EZil/y5fcl0cvHdq3JH3rmk5aT/sKiuK8O4qlopThRB4sI0PaR4XsoHnAYetQ+t3hpworLeHQ68DeMzo3nM/gV98X3hcNKFu2L3pRkQRkLqRNoHPhNcKfFyCsVhyqjZyJUY9h4S3flJdiyaHa3J+5H142wSgBMqkf4Lz4fgKE9eo6XasrtWLT1gx0H70PUmmDBKeIyqSS1a/7y/h5VBuZ+vZ7vmU7MUEpizJ6157tOAMWlmcdB6bVODsgK7fK7DZ2Y9G3RyULLdvqwd4ztMX3Af7tiSyqDfh/1s1uk83aptwKsO7TiZE8JtowMrQBfFCMZNO+/erbo31LY5SYU7SmF2LRN/inkxXLRa5nrnPNpZXKkAvn69u2bafiLtE3ss0saev83kwV+vYbIBRLDu379tTtcAIPVsHq0L3xCPCfXLrwvqgnOyZf9KIiCchwqDavjl5JfQlyZg7Lts7dOyFvTLQFu+/Bwju+jc5KaYXlmWWpVBs4n3bS6XWkUm2CyJY5RtdDx/DXpLxSWuHW1K3wVBsXfj+zZVvyjPzuEzAyWO5o2vNkSzvfA/CddSsVmEnNSKPaBKFVyeqlerZdQxOwNOuY6vY5lEXZQT/4p4iulFZI6+nwVJvSc+g20RbfwdT8DfWQneDxO8BlreLsAZN6vn2OgC+UC2hCk+rENDXd14Ai2U6j33frRe0FrV5Lni1beJOMmQtEY5NVFQT/lC7ZQZHf873f3KfSrkgPyDKWhX1OUKbi7sgYGd+27LXZ18LTvt0EJotvA+fr+1L4ogOIyxe9qEgCMuTucvHglx/e370QlmpT3oDOUbAKmUQ+PPinQTyvPHf2gMk0tAH0DfBwRt6l2tNNjN75Cx896ocs+NX3q6NXNLoNeTSf+UdkUjlfsmVXqcD/GHav/0AW1cbUTAxhnCv7oHnAQfNAij15tlNl+XqOXM6Z6nbe0t7+pSrTlhkZ3zQfKVQbdzgRC2/7smX1Tp2t2las9kTmex3kfC9Ny9kD5qGpaYhz+mG9PWCyHTeI2J7sPHHOWAB9L2QXmE3Phpftwo9s27YdGrLkdxrisd+2mcMAOufYlEKpwLQ5zc3cTWmyg1TIZDJKjPl3gPP1fSl80UHZMfmiFxVJQIa60jj4ewmk7F5we3tYfMdXRrvVa0ndAwb+R+mulOWO0R2UfR4f36PayNgD5qHro0m5Z/WkGx1TM9GFfu6zljq2ePdpvwoLPs532ZmwKGMPmAc/WWXZY6KDyJZG3cOpkD26mcdITwPQOadCVigVyBpZaQ3h4L9CJtuJYeEtX/perTgN4SrsyXmyC6UCutB5MPNAquwoq7CDcJb2ni17o7pBx+rEY7/LkhIOjUOovYKFt8kaWbp211cFQba+/QQI2/Vtap2amvPt8+6QmXAQKYch0qjvn/l9nv2WlVADJzg5752utCvsNCTtAdt9AnP3yeQcloYfewIX3Bc9ITsOX/SiIgnIcA5iWk+HX2Y6gCAvgVQnZv5NX0Z+rexQbWQaWi9AqHfOznwVS0VnwqJkJwb8BQiyX3zLSJM6p0l5q7Ylj2rjQgjh61n3aZphnZhW1anEuo4y+DvfsneL+PnM+819yq2y9CZhP7JlTlhc3Tvi0eI0hpuV77TPXtorbQ/YAPx85qPOES+PXkoa6PEEZu5CZoackTu3x6a/QkNB9fm8ymChVJBOtckaWTpWh+4ZNkUF7RvcfthzBhSp1Pd5z7pQKnAzd5Pp1HQ4gbvvOf92AzI425Z5CR4VAZlfmr2K5Ol557tYKpJP5ZnPzkuTrbkBWeucAUWyacjg0oGjtCduFTZr+r8vL7wvekJ2HL7oRUUSkCGRajMAP5SARrfBi9oLeQHZ9E3IXY/UaRyEEMJXiVoF1cbPXpWu1XX2gMkOyMwMaat35vfIHhPtwZe+y5KoNp4Ts/iOr/PtUW2UGHm/l6oKfZ/nNJYLTJlToak2z/eP6Fo2b97Mk3J3uPValXNly8yugvOZzwsQpOp713FiPNm+qTYzcqk2cH6AUCwXlegbzn631qvrdG15e8A8dHUT45x+WK9qIrOXKgiFTtoALIDFt309a4/2HYfDquLu8K3vslzaN4DmBtNnrfA4bB5y0DyQru+ckYvON+p1Yf8ZLLxFSkshEOcmqy+FLzqAuHzRi4okIENdFh/OpgT0dy/Iyiq7ToyfrNtKaQVd6NKn2vihGMkeE+3JhbMpAZvVTTpWR3omxjZyZM5pUO7vAYuhWiRN37vHvT1+9L1T36HaqSox8m2rTe+MIFj23p5B2X5osTLGRD/bccaPv7E4jZlxgmmrNX4keaVdYae+I/2M+RkAIE3f/eFEx7bMj75VUG3g7PPd6rVYr67HcneooJKBO6DonH7YQlk+7duPvi3bYrW8Km9FjJGF2WVfslXp26/9vpa+JpX27fUVnXXGbNtW4htpaa8fdnxApspJ90NZlLYH7HANem1YfKfPZvHlG110X/SE7Lh80YuIKx+Q9ak2Chw3ODsrIc3o2Lbb2+NmlX00Cnt7wFJ6KpzsEzhPdsfq8LzyXMmLD+foW1GVCjNLxraxzqkgLOYWyafyUkWfV0HwLlVZu6HQ03DtAVk9gL4lP2sv43fW5VYsF8mbeRayC9JlRxUAv7ddRQh4fWGadNZxxqz2+IBMWVXQR1N4sVwkpaVYml4KJ6z0HLoNWHirL9tPFl+VE3OW7D7VJoa7o1gqIhA8mH0gVXZXT507oEj2gAmgv4/yrM/8ovaCZq8piUr2GBbeBE3zp28F00shWnsyCD9094PmAaVWSfpnNtxdmGcNKOrrO4YAuFguyqF9e1VYz5ad86wvhS96AnH6ohcRVz4gU+XE+MlKFEoFDGGwPLMcTlh5E9q1/mjVrOlkYix7fNVGGvXjBLLm2RmR9Yoaqo1ffYP8KhUuJ755xo6oldKKGn2fk4F6dfSKercuiUr2FObfBE0/5sSfQfNRZeT9XDDemGiZVBtP9ln6LjVL7Df35Qz02KmxfD1HNqWTcfcP2e2jsd+vTN+mv3dLCtVmYDgRnK/vRrfBZlXuHjAYCPrPkK3KSfd7vpfyS9L2gHmwzPSZS3v7e8Ak61sT2rnPWi4t1hlOBP4rkvPZeakTFj3ZfhJqcdwdquyJmXECsu4Z9OuV0go5I8etqVtSZfsJgKXd1f2+ficgi/R8DyBSX/Sk7Bh90YuIKx+Qqcri+2maLZQL3J+5L2/C4gDNB8a/gO1e26HaSM66ebLjyMT4zQLdmbojlWoDINwR8M3Gwci/71NtZAeC+NC3zKrg7pN+ts8PzadQKjCXnpNKtRmUfV4FQQUn3a++ZTzrle0ajxadfgsvIOOMHoRCueBQbabkTVgEn7ZMNi12/k1H9jk7udbKa9jY0m2ZJrRzs7srpRXpw4nA5/kuy12h4cEyMqR746nAm7VN2lY7lrujT/sOK7tZhurWKVt21vmWvbLEw3lO+m5jl2qnGu/dIflzmz76YT37rSqhNm5JcrVddWjfsmixs8vgTsiNS9+R+qJjZMfhi15EnBuQCSF+UgixI4T4YhS/UNQoloqYmslSPiTV5gR80apKRUkvvlcaP6YsniXbW2aq5ILRz8m6lQsIhNSGcPDXuFosS9L3CXhNyuOW9m7VthyqjYoAQT+bly5tT1KrBqX1U2fsTCem7FCbZF+q51UvDpoHHLYOlej7vMWishIOnZ5Fca/GG4sOxVU3UrSEODMgK5aK0hvC4Xx7Uu/U2TraktcLm78N2bm+7PPsCch3YuD8bHqxVGQ5vyyf9n3O3dGxOqxV5O4B82AZGdLnZLNBkb7PCYCLZYf2PeNS3ibGiQTmefZE1QoNT3aj2xgbIKgceHBeP1WhVCBv5lnMLUqVa6ad59c7g36tgoYMjr5tbFq90ZNEpVIld44TmOCvQnbhfdGTsmP0RS8i/FTIPgl8s+LfIzaslFZ4OPsQQzOk/n/PG+Pb7DbZqG7IK41PLULu+rDsMZebSiPvp0J2d/qudKrNeQGZKqoNHDcpt8dQFlXsFvHgR983MjeYc7OSE2PPm7DoODG6ppPSUmc6MSppmuDjfCsw8udN6SqUCuSMHLenboeS83y/Tqdn8+bN4/HeTaEhzqCIejRN2ThP354TI82WuY6yJ/s8fRvCUNIQ7qdio1TfY571RmWDrtVV8m7ZpjOgyB4zpKhP+46pQiYt6Idjiv8553u7vs1R50iZLbOxaVujJ1uqvqv9VL1lJ9TSbrW/N2aFR7lVZq+xF+vdEVq21XPuzEX/tuxS+KLjZMfgi15EnPvkbdv+dSHEA/W/Sjwolot81fxXSf//egGCdbQHv/Hjzgs6gHJ9m+8qV/jIy6fQ+gfhhD3/9KlMDIx/AQulgkO1kdwQDudnlVXsugBnB5qhGWNlv6i9oNVrKcm6Ge6gjqfbVX51Z/XU33/m5Rbtg6/nM0/S/Kdnp/8+DFY37vOq1OOT//xfOMtOT2D3ZYnfJT7EJ//J/xlOUGULun8INm7CgfMZuoffyG88neKTzdOfqdapsf/qA2yn3+aT/0HuZ16v2rQPvp5Pfe6Az8+c/n//5s467YOv53PvTfPemlzZz9bvsrf7/rGf6dee2GR7H+OnPr0WTo47YfHR4vEQmJamo3VGB8C1do3t+rbSAMHeL8L6b5/6e2v3C3yiXOWrVz8Lr1bCCdt9Ch/+3iHZXgVhlGOoYg+Yh7NsWbvXZqO6wR+8/weVyIUznJiyuqBImDl0oNWpkU6frkQVygVuT91mypySLjtjZGh2juC3ftrpiR6AZdv87uef54MLH4TPhrwv3/slMDIw5wTxfp1GVbRBcO5qb7DJkOxygdn0LDcyN6TLPm+FR7Fc5CP3PiJdrrfCw26N7odVGfR7E2O7G5+Fg+en/n7m+a/wPdUGd7/8CxBmqEezDL3WUHIpY2TYre+O/RHVvui4871WUUP7hnh90YsIaaG4EOL7gO8DWF6W2xioCvVOnRe1F3zbG98m/f/tBQhvFv89PPmVU39/E/hBgM9+Uo7AD3xH/z/P6+/xptqMugDC4qwsUNfqslZZ4/cu/V7pcj3Z4yo2KjMxhuu4/O3P6PzO3pdHfMcM8HH+118sSpcND4AH/NAmwLURf/9RAH5hJ6yca8D74Vd3Ae9S+UN8egs+/YVRnxng4/zCNvwC4/4+DD7OJ7ePYOT/OwN8nL/5rzYUyL0D3OGHtsZ9pg8C8EPF8J95LmfyxuJxhayt6+hjzrdXpVLpND74jZ+C55859fdf5f7Dv/0bcgQuf92Q7J7do2t1RwZdxXKRN6+9KUfuCZxlT55XntOze0rsSX/NwBj77dmyhzNyad9Af0BRq3E4MiBTRW0CR9/3D17Af/izp/5OA/4ywO4vw5d/Obywh78PXGrveX0uqqtUcMZd7fauya5SebL3m/sj/87bA6bCnuRy7pLpMfRrZdOQOQ5OZn/xrzhj6U/gm9x/+KUfCC9M6LD0u/t/PMs3isIXPc+eqOqRhHh80YsIaQGZbds/BvwYwLvvvjuaEP0VhtWyk+VWNeEla2SZK2/B3DJ83/879Hc/9oUf46cf/zT/5o/+GznZ3eyxM35exk8a9WPUr3GG0Vmvriuj2pwnW6WRNzOz2Daslg2+42uW+Kt/+J2hv/9T//pPMZ2a5oc/8sPSZf/93/77PP7tT/Ijr7bhW390yJndbezyvb/0X/PnPvhn+ZbXviW8sFTOGXvv4hP/4hM8uvaIv/71f/3Ut/7sez/Lj3z+R/i5b/s56dnd1coa3/OL380Pff0P8QeW/8Cpv//zv/bnafVa/INvCplJH4Gf+vI/4se/8A/51f/qV081QFfaFb7lU9/Cn/nqP8Mn3v5EaFnZlE7GPO4Ha+smWu9sapNK2snU4XN49IfgvxzW6/f/+vfzqr7NP/rmnwovTNMhczzJbrAhfVYfnnDX6rXYqG7wsYcfCy93BM60JxHo+yzZsveAefD6YZuNA2bmhmmgPatHsVzka299rXS54Hzum7U95w9/7jdgar7/d5/e+jTf/+vfz49+04/KqSIMBJvn6rtc4HrmOtcyoxJe4XDWwAWP9v3RBx+VLteTfZaPAGrOt5maogPQGV8hyxrZ0LTvUcgaWVKWjVlah2/4C/ANf3Ho77/957+dr178av7a1/218ML0VH+ghyd7nL6j8EXPsie60KUPJ/LkQjy+6EWEXLLqBYNK6gc4AxduVNfgzrv9/i4PX6q/ZPbaQ8z8TelyzypRd3od1ivrfNPyN0mX68kel2nsj3VVGAyeleW8NXVLCdXGTM+yyyy1js7778xwbeq4yd+yLTbqT/j2u98+9HVZuDaV4lH3gGuiBq+/C7PH5+nxiwK1VIu3lj/AtXn55yyf1bC06sjP9bK5wmzO4PXrt6Rnd+v2FMKooxvNkbLXjx7z+5Z+nxJ9X8+lEEadTLrHTGr4LK0dbSCMOh+4+ZoS2R3dxOyOD8jSepq703ely80YGQzbZrq6DR/8xClb9oWjTT4w/4FTX5clG5zqxcmR4/1SVebVAAAgAElEQVQ9YIrsScbIUB6zK6lQVkv7hrMDBFU9F5p7pkcNKNqqbdHqtZTJzhgZbtXLTnJx/hEM2I2nzV3Kus79mx8CyaPnTc1EE9qZDquyz3zGjr/95j6VdkWpvs+9qxXJbmkaYgz92luOHHoP2AhkjAwPOh2EbcHtrx6yWbV2jWetfT42/341tuyMISpR+KJn+UYqad8Qjy96EXGlx96vlFYwNIPlvBqK5ZSR4cbRwVB/l4dCWd3uhbOyEmuVNWVUG09222rTs06PTvaybrInLA7KjuNSTWWv8cxyJiM9ujm8+Pnl0Usa3YZSfb/R7mCnpmFmeNy5yiynJ/ssI//G3BvKqDYw+nx7VBtl+j5j/5qnb1Wyu7qJOaZCtlJeUTJhERx9P+h00GxrqCcCjqk2Ks83jH7WqvV91gCXQqmgjGqT0lJjA4Su1WWtvKbsM+vegKIRAZlqfWeNLEuNqnPGTtiNldIKC9kF6XvA4Owlyd4eMGVJxDN2/EWh77MqF1PmFDdz8hN5AE1NHzugSOVdnTWyvNFxF5+fsGWq9gr2ZZvZsb1Uqn3RrDl+gMtl9UUvIvyMvf+nwH8E3hJCbAoh/qT6XysaFEtFHsw8kD7VxsODroUxwonxqDYq+NlwdlZCJXUPzn4Bi6WiMqoNjB+b7FFtVF2qqfQsz2ynMvFooOcH1PKzwdH3a50O3Ruvn3JiiuWiMqoNjB964FFtVGX7zsoqq57adJbsYqmojGoD0DPSmL3RS3uLpaIye5LW07zeduWesGWrFYdqozKLD+OftS50JRMW4ewx7F4WXwWEEGNlb1Q36FgdZfakH5C1TlcGZe7YG4WsnmG5WR+ZwFTZuwbjn/VOfYdapxaLPYni7jhrcqqq3jWAtqajjQhCy60yu41dpf7J6+0OltDgxrAM1XdHVh+frFbti44735fZF72IODcgs237u2zbvm3btmnb9pJt2z8RxS8WBVRvCH+tPToT41FtVGclRmXdiqWikmWmp2SPKM2rpNp4shu90y/+1pFDtVGl70z2Bs/sJab0Fgv54Yy5aupH1sjyeqdD6/ppg6rSafRkj9K3R7VRpu8z9qpIHcE+AmcNXFBJtQEnIEuNCMiOOke8PHqp7DMLIXirZ2ODQyUbQBTnG8Y/6+UZ+XvABmWPkutRbVTeHeOqz56+Vck20+7S3ubppb3FUpFbU7eYTk2f+jsZuGFZzFg9WBjuwfX2gCnX95g7CyI432Oe9Uxqhvns/Km/kyW71Wthjdg7p7JKBdDWDfTu6V1gynupdCeBeZS/BcaJu7pcJKWlWJqWuwesL/ss30ixLzrOnlxmX/Qi4spSFhvdBi9qL5Rm3R60XGd1fngKmMqxrnDsNI6j+SxNL/WdWtkYN+a0a3VZLa8qbeAcl/FTre9Mdo5n1l2W0oenMoorpRXms/NKqDYAM90u8z2L+rVhqkOfaqM6AB5BwVA5JhpAE9pY2SulFXJGThnV5qwKsGp994wMqTHZVVB3vgEedbocZufAHN4f6FFt7uXvKZHbt2Vjzlkc9uR55Tldu6tU3+Nkq6Z9p9yBKt3W6YBMdQP+nboj01oYvi892rfS5NIYOpnqqsl5d/Xrc+qqVOOc5VKzxH5zX3FAZmKM6If1zreyKqzpJDBLM6fvB28PmAraN4y/O6LwRePyjeL0RS8irmxAtlpexcZWesEsNY/YTmWGJu2Ak3VTNdUG6Pc1jMtKKKV+eCXqE5WTzeomHaujXPaoz6w6QBBC56m9xHJq79TfqaRKAlyrbgNQnRse5rDb2KXaqSq9VMdRFqNY9jiuQbpYKip1YsZRMCrtCjuNHaVOo21myIwIyPpZfIXn7GG7zaup08vFParNyYmTsjDOnrR7bdar68rtibcDbRBR6HusLSsXuDN1Rxnt23QDspNLey3bYrW8qlTfi0cHAKeq/VHYk6w+pkJWKnAtfY3rGflDHmC8PfGqgirtyTjZqqmp4PTDGiP6Yb0Ji3em74z4qfDI2ILlTpeD/OKpv4vMNzqh7yh80bPsyWX1RS8irmxApnJssYc7jQrr6eypr3sN4aqoNrqmk9bTp178jtXheeW58tI4jDDyEej7rCzQzdxN8qn8iJ8Kj/2jNmXyLBnbQ1+Poko1W3kJQGnm1tDX+w3hiisI45yYfCrPQnZBqeyzssoq5cLp862aSgZgmzkyto1tDVOMCqWCQ7XJq6Ha0Otwt9VkKzt6UXAc+l4tryql2gzKPnnGi6UiAqGsSgXjB4qotidpt9/UOrGY+UXtBc1eU6m+52t7VDRBIzNsp1X3UsHZd4fqygWcPt/7zX3KrfKlvau7Rgqz1zn19WK5yMPZh8po39nyFjqwMz0cYNc7dbaOti6tvs8635fVF72IuNIBmSEM7s2oodpg9VislyimTh901ZcqOBmRk1N11ivrDtVGZdZtTJOyl3VT6cSMNTqKncb3tp1s8j3j5dDXXx29ot6tK5U9XdrkSAjKJ6qwqnt74LhReFQFQWVDOIyuzvWpNoozjTD+UlXpvAkzhwa0TgxcKJQKPJhV1xDOfgEDm/XMcFWm0W2wWd1Uq+8x9kTlIuy+7DHPeqW0wlJeLdVm1PlWPWERIJV1qqB2e3hHVJ8Wq1Df1yrbFEyTxokAuFAucCNzg7nM6QqtLIzSdxS9a+POWBQ05LNsWc7IcWvq1qgfkwJLH90Pq5oWq+89BeBlbriFoD9hMaa7Q6kvytkB2WX1RS8irm5AVi5wf+a+MqoNh2sYVo9nxjAfud1rK51q42FUE2ck1I8xY3wLpQK3p24r2QPWl+1+5sEAoU+1UajvlR0nm/xA2xz6ehTUj+zhOoWUSbM33CBdKBeYS88po9oA5Mwclm3RsYYznR5tUCVGXTB9Jz2CjPap810ukNEzSvaAeRDujqhW43Do66ppsew+AWAtPRyArJXXsLHV6vsMe6JqD5iH/gCXE7KV65vR5/tF7QVtq63WnmSdJe52e9iBUr0nCWCmvOXYspP6jsienJS719ij2q4q1behGZja6c8cBS123Pn2qJIqE2o9M0P6BP262q6yU99RS2PbfUoPeJke9kOiSKiddXco9UUZfb4vuy96EXF1AzLVmYGdxwA8OdEf6u1eUF2qHXWhF8oF5VSbs8ryUVyqPbs3FCBs1bZodBtK9f3edpUpUWfBGu4hi4KKkD5YpWiaY/Wt8lId9az3G/sctg5jCchU713z5MLojLZKqg2AcKfbNZvHAZnqPWAA7D7BAormcAXOcxrjovks55eV7AE7S3bH6rBWUVul8mTHcb5NM0cHoHMiICsVWMwtMpM6TVuVgtouqVbVqZANfG6vShWFvk9m8aPQtyd71PnOm3kWc6d7nWTKhXjuatvMnhpQpHpCLgC7j9lKZajZw0nEQqmAqZnKhhPB+AEXcSUwL7svehFxJQOyZrfpUG0UOzEAz0xtKECIgkoGo4ceFEoF7k7f7RtiFcjqp7NAPaunfMIijKYEqF72CPBsu8ay/grTGm5SLpQKXM9cV0e1qR+gH+1SMM2hZ616mamHUXSyKKgfMLpJuVh29oCppNqMy3Kq7l0D0NwKWXsgIPPGRKu2ZQeZPJURTowh1C0zBTA1E0MYY7P4KtFfrzAge6OyQdfqRkLzGdW7Bmqz+ABNTUMb5aRHUIU9GZBt17c56hzFZk9Avb5H0SW93jWltG/99PqQcqvMXmNPfUBmZMiO6IUFxXfH7lM2M1Mj7YlS2jej7Umz22SjuhGJPena3Svli15EXMmAbK3iUG3UOjFPqeWuUde0IWO7UlpRTrWB8Vm3KLIhwFC2cbO2SdtqR5IFAk7pG9Reqis7Ne6ZO5jd0w6r2myfw4cvpIadmL3GHpV2RbkjcZa+46ogvD77utIqlamZ6EIfkl1r19iubyv/zHraqU60m6X+1yLR9+5T9qZvjLQn92fuY+rqqDZw+lm3e23WK+uR2ZNBWxbVMtNx51s17RugpemIgRHwlm05NM0IEpgnbVnc9mQ2PcuNzA2lsnNGbiSlK467OioqmW1OYQKdzrDsjJ5RNmGRbhv2C2zlZkf7RrPR6HtQdiS+6BjZl90XvYi4kgFZFBPo2H1MddbpJxncb1IsF7mXv6eUagOnd0R1rS5rlbVIsn0w/OJHZeTHyV7MqqPa7Nda7B+1WUrtYw5kn6IYW+w5MUUzRb1z2mlUbfDG6XvanFZKtYExPWQRjNEVQpzKaEdVFTTSzvS5zsDS3kK5oHQPGL0O7D1jP784smcvirHFJ/X9vPKcnt2LzZ5EQbXJGJlTe7Gi0ndL09E6xz2p3h4w1QFZLzXFjq7Hdnd0rA5d63jQRLFUVD6cyJM9+JkPmgcctg4jqwAPnrOoEg4i5QYIjYMh2Sr3gLG/AnaPnalrQ/r2aN9R+UYjA+Co2ENXyBe9iLiSAVmxVMQQBvdn7qsRYPVg71l/We9gubZQUk+1gdMjyder65FQbdJ6GoEYyvhFMRENRtPJVDvpz9yBHkuZA9K9Y058n2qjuHKBOUUpM33qjEE0WWXg1LNWTbXxZA/KLbfK7DZ2I2kSPhkMRqVvw62Q9drHAZm3B0wZ1eZgFawOldk7Q/pu9VqRUG1ghL4jrFLB8PkulArcmb6jnGqTNbK0rTY9t88mKto3QFs30KO2J7tP6d54BOL03XE9c51r7jh+VTg54MK27UhoyOCe71709mTUWodiyaF93566rVS21w/bHhhQpHrFgJfA3MvPD52x1UoEtG/G2xOlvuhJ2VfIF72IuJIBWaFUYHlmWR3VpvQcuk1a150squdMdHod1ivrkZRqTzoxUfGFhRCnZK+UVrg1dYvp1PQZPxkeJ8vylm0pH1vcD8iy1aEm5WicmMew8CYZM3cqQJhJzSin2oyiQURFRTh1vqNoCB8ju1AqkNbTSicsAqTcXsRu63hp70ppRTEt1nFijmaXaPVa/QBhrbyGZVuxBcCa0JQtMx2UC5wKBqM4Y/0AwXVkXtRe0Oq1IpHd0U2M7nE/bH8CneJqv73wJnCaVhXVGRuUvd/cp9KuXOoEzzga22uzrymlfQNori/QcunXtXaNV0ev1FP8hUZlejEWmuaonmvlvqiLk8/6KviiFxFXMyBTPbVpx3FiutcdGV5G5HnlubN7IQaaj0fTfDijfqrNyQZpj/qhGiczUK+OXtHoNtRWyLar5NMGc5kumcgDsqew8PYpfXtBURRVKjg28gfNAw6aB5Fk3byhB5btNIZH4jQOyh7Q90p5RS3VxkUq7QRklhuQ1Tt1tmpbkWSVG9ccSmTLXa8QCe3bxajzvTStdg8YnLYn3h6wSOz3CectirHcHrp6CmNgaW+hVGAhu8BsevaMnwqBo3042kVbfAcYrlIVS8XI3ulB2VGO5fb2OXrwaN83czfVyh0xZCKKiX8AesqjXzsBWSQsmt3HcO0Benr6dJVKJe3bhZesHpIdwQRROD1s7ar4ohcNVy4g86g2UfT22ItvAcec4ZVyNGN0YXRW4u70XXJm7oyfki+7Z/XUN4QPyIVjJyaKscXPtmu8cXMakc6RtW1sNygrlJ0Ji8r2gDVKUH0JC28P6bvfuxaBoT2p7yjG/J+UPehAZY2suobwE7JPvltRnO901qFtWW2nKrtaWcXGVl8hm1vGdKtzni0rlAroQldepYLRFYQ47MlGdYOO1YnmfJsn3q0I9lJ56BopUicCsigGehg3PwAcf+bt+ja1Ti1Se9K/qyNMOGTN0056FLRvTWhDwWClXWGnsRPJu2X0BxQ5S+4juTt2n8LCOyPtyYOZB0r3gHkYlB0p7ds8cb6viC960XDlAjKPaqPciZm5S8pdsukdxmKpiEBE4sR4WWVvSXJUmRgYfgG3alu0eq3IMo1wbHT6Y6IVBt/Pdqo8WpwGb4Fty8n4KednuxMWTwZk+819yq1yZJULGD7fEF1WeVB2oVRQvgfMQ1Y/1vdR54iXRy8j0bdHWbTaR8CAvlXK3nkCC++cftblYiRUGxjOsHpUmzjOWCT6HiO7UCpwM3dTOe0boGekMXvOcItI9oDtOjs7tcV3SOvprwh7UiwVyafyzGfnI5F9KuEQwRmD4XcryvPdH1DUcgKyYrlISkupo313285Qj4W3Rus7It9oUHaf9h2DPbkqvuhFw5ULyCKhfuw+gYW3Rmbxl/LqqTbgBEU2Nq1ei67Vjawh3JPtNSl7md2omkfhuO+iUC4wn51XRrU5OGqzV2vzaDHfb1JuNQ77VJsossreOfPOWJROzKnzXS4wZU4pp9oMyR541lGeb09uVLuhALI5tyew4wRkfarNjKoJi13YfzbWlsWh7/XqukO1icCe6JpOSkv1bVmfahPBMtOR+o7IibGMDGm30t+nfatOLqXyMHN3KLkU990RBe3bk+0958PmIQfNg1iSp1HSYr2Ke6/lDChaKSmmfR8Uwer2E5jNnhMgNLoNXtRexGLLotT3VfVFLxquXkBWVky1sSzYfa9fGofhLGeUhtaTvVndpGN1ojXynWEnJsqyvCdbtb5X3IEej25O95f2thqH7NR3qHaq6gMyIwtz94cu1Sj1fbJq4jnpkTgxA8+60q6wU4+GagOMdBqjoH4YRoY2QOeYNqiUanO4Br1234kB51m3e23Wq9FUqcAZcHHSnkS1wyZrHtuyQrkQKe0bHH1btuU4MRHp2zaz/YAsEn27CUxODIQqlApcS1/jRlbtcCIYtie2bUc2gQ6O7YknF6IbeHDy7sjoGeXDieC42t9z6deRJTAX3yZn5rBsi7bVZrW8GskeMA9ZI3tM+1bti56QC1fPF71ouHoBWanAvfw9UnpKjYDSc+g2nNL4QNNsx+rwvPI80kyMJzuqMdEevIEL4BjaxdwiebeJVyX6jatu9kt1Fv+9bWe4wqObeTT387WbpWj6PXafwMKboGnDtJNykbyZZyG7oE62C1MzMTRjKOMX1W6RwWcd9dSmkzQfpVSbE2hqWn9pr/o9dw6VjMW3h+zJWiW6CYvg6rt3gmqjeJlpX7Y+bMuidNLB0feL2guavWZkd4dtZshYzrCcSIYt7DyBhbeB4bsjLnty0Dyg1CpF6rD27B5dqxvZTkMPgwNziuViZLTvPv26VXOGEx1tRRCQCbjxqE/fa3abkQfAQ/ouFdX6oifkwtXzRS8armRApnzyHcDiO/3RxY1ug43KBl07ut0Lg1mJKCfQebIHKwhRvfimbmIIg0a3wXZ9m3q3rrxCNpXSuTObQe83KZeiobG5vT1wmnby+lw0VapB2aVmif3mfuRGvtFtRO7EnMwqP5hVuAfsBJylvU2a3Sab1c1ossrzb31F2ZO703eV7wE7Kbu/BywG+x35mGhzChPotI8olArcyNxgznWgpaN+AEc7sOgEZEPVoohpyHDCnkT8rOvdOoVSgZyR49bUrchkx1E1yWSdYVdWuxaN/d59AtceQCp36lkbwmA5v6xO9gBO3h1Rsgzg6vmiFw1XKiBr99rqp9r0nZg3MTQDXeg0uo1IqWQw3MRZKBW4M3Unsqk2XuNq1FQbOK5eRKHvZztV3riZRwiBkXZ6yDqtMiulFebSc+r2gDXLUN1yaD4MNwpHeanC8YCLOKqwQP9ZZ/RMJBMWPdmDFcko9d3WDLRuMxqqze5TmF2G9PQpe6IJLZJeKnDOd9fq0rE6ke258+A5UJu1TdpWO7bzDdH0mgD9AUX1xr56fQ8MJ4JjW7bb2KXaVkz7HsAofUdZpfJkR51Q82xZtV1lu74dmb7TWSfAtztHka6IgWN917t1Vkor3J+5H8lwIk+2R/uOasIicGV90YuGKxWQrVXW6Nk99ZmY/G3Izg0tSS6UCwhEZE7M4NjkYrkY3WXOcZPyVm2LRrcRbYDgyu4beaWUxZozYREw3MEh3VbF0feswrHFu+85/3YvGG9s8kHzgMPWYbT6NrNDma+4KghR7AEblN21ulTb1UgbwgHauoHebUdDi9150g/6T9qT5fxyJFQbOH7WtXaNtUo0e8AGZUdlT07KhWN9L2YXmUnNRCJbpI8HFClfoeHRYgfOWRxUspNZ/LyZZzG3GKlsj9IVZfUgZ+RodpuRswxSZp4eQLtOoVzA1EyW8ktqhPU6sPfs+IwN6LtYisc3el55rt4XHcBV9UUvGq5UQBYJ9WP3mA8PxxnWQil6qg04Y7lXy6vRZpVPOOlRZ7Q96odKqk2p3ma32uLNm47z0m9SblVZKa0oziof9/aA85nbVpv3Dp1ALcoAYfB854wct6duRyYXjjPaUQf9AF/e/zIQ7fnu6CZGt+1MWBQG92fuqxFk9WDvvVNOTCz6di/09w7fo2t1Y6mQRTkRzZML8ejbW9q7VSpw1DlSXyFLTcOsMyk0Z+RiSfCk9TQC0ZcdxR4wD96zflV/xV5jL9bzHZVsoWn9flhvZYky2vfBKlgdWDym+AOUWiU2qhux6jvqe+uq+aIXDVcqIFspraAJTV1DuGUNlcbhuEk5aiqZVyZeKa3Q6rUizbpl9SytXotnpWdANGOiPXhNs6p3XTzzJiwuOs6LmXEqZKXaS6rtquKs8lMwMjDnOONeQ/qX9r4EREhtwh16MJDZjZJqA7DX2IuUagPHF8wX974IRKvvrm5i9pyATOkesMM16LX6ToxnT6rtKuuV9UjtiSc7Dn17FKNCucDtqdtMmVORyPUChHq3HjntW3dXeKzuuvpWOtDjMcy/Ca7d8CiLhXKB2fSsOtr3CQgh+ndH1DRkz5bFZr9d/yStpyOjfQO0hIbm0TSVspZOVGFd+/14/zE2djwV93JBrS86AlfRF71ouFIBWbHsTLVJ62k1Asobzkhq98UH5wWMg2rjNXF+ad8x8nFVEBayC8r2gI2T7dHYlC6E3nYCsjdcymLabVLeLa8BivW98xjmH4FL0fP0/aX9LzFtTkeyB8zDkL5jON9elSrShMOAvk3N5F5e0R6wEegaacxeR73T2N9zN9x38fTwqUO1idCenLRlD2eiS/DEdb69AGG1vOrsAYtQtjegaOPA6e+KqrcHhvUd1QqNQdkvj146e8AiZBnEfVd7AXCUtG+Alq5jt4/Yqm2pT2CCE/gzbL8hekZJs9dk5XBFrS86RvZV80UvGq5UQKY+E3M8YdFDxsjEQrXxHCgvqxxHxu+Le1+M/OXLGBmeV55T69SU6vvZTpVcSufunGPcM5lrABxUXwCq9/Y87U9YhGF9R0m18WRv17fZbexGer69i8w733G9W1FOWAToGSnMXje64URuckkTGhk9E7u+o9oDNii73qlTLBd5YzZaqk3WyMaib8MNyLYOV7ieuc4117ZJR+MQaq/61GsYHsoU9d0xqO+47suskY2M9u3JbnQbrBxGr++2ZtBsHmJjq99zN3cf3F2hg/qOag+Yh341dP9LkS9Hvoq+6EXDlQnIOr0O6xXFy0xPlMbhOOsG0WdiAF4eveTW1K3IqDYnZcdxqXr6VpkFerZd443FaTTNpdq4TkunWWYmNaNwwmIFKpunzhi4+o7YyGeNLK+OXgHRGlpd00nraV4evSStpyPbAwbx6tsyMqSsjrMHTPVAj5klSB/vD/TeraipNnHbk8PWIa1eK15bFmEF2KNfN+p70ayIGVEhq7Qrseo7DkbJy6OXvDb7WiR7wE7K3q5vR27LOrrZ36mofkXM8BkDR99Kad8jELctu2q+6EXDlQnInlee07W76kvj07cge5xR9Pp7IPpeKg9xVKnikj2ob9UVMq9/DEA3UrSEIGtbvDH3hroq1Z47YXHxdIUMYtD3QGNwXOcsaqqNx4mH6D+zZWbJWLZ62btPhoJ+ONZ31FSbWG1ZjM/akx017dt0BxRlbFv9RGI41XPtIS59T5lTkdK+Yz1jMeq7o5tkLBtDU7gHrNeF/WdDVdjBOyvqARNxPuur6IteNFyZgMwbE628NH7CifFe/qipNoZmYGpO5ieOqomHyGW7E9lUUm3KjQ7blRaP3AmLHppCI2spbhIe4cTEGRR5sqOm2gzKjrpJ2DtjEMMFY2bJ2jY6mvoJiwNBP8So7xjtyaDNjutzRz0mOu0GZFkrgj13Zq4/YRG+Mu6OyHvXYrQnXn9PHLK7RoqsbfFgRiHt+3ANeu2xQf9Vujuuoi960XB1ArKSs3tBGV/Ytk81KMPxCxjHZBnvBbxKVZMonMaVnSpAfweZh5auk1WdVd55DHoarj3of2noUo0p+H44+zBSqs2g7FjPd9QXjDmFDrw2vaRuD1jpOXSbY5NLV9Ge3MzdZDo1fc53S5Y9ECBEibSbyMraVgTDid4E7dhueLYsn8ozn51XJ3sE4gqAYw1CXdkpLcXStKI9YGPQM9LOfRlxG4mpmf0AMK4AWBNapL1rcDV90YsGXx6UEOKbhRBPhRArQogfUP1LqUChVGApvzSUHZGK8ia0a0OlcTguUcexe8H7rHE5MfPZ+UipNnCsb6Uj790Ji2/ezA99va3pDs1HdVZ5/s3+hEU4fs45I8etqVvqZI+AJzuW8x3Bsx4Fj/phaAb3ZqKbsAhAyrnQ38krqo7BQG/PcIUsbnsC0TsTsdrvuM63OzE2Elu2OPqMKaV9j0Fcz9rUTAwRT4AQF+0boGdkIqjCuoyS+RPJJT2eAME7Y0vTCn3Rc2RfJV/0ouHcgEwIoQN/F/gY8D7gu4QQ71P9i8mG8t0LI6hkEF/WDY6zMXFRbeIoT0eRiXlvu0bG1PoTFj20dZOspTirfAYt9vW5aKk2g7JjzbrFlFV+MPOgT8WICpo7KeyNaYX00B0vq/zm0Jfj1vedqTuRUm0GZcdhv+PKKpvmFF3gmpbmeua6GiGNElS3xtqyOO1JXLIzeibSPWCeXIjnfNtmJgJGyROYXYb0cHU7a2TjqVJ9BdiTq+SLXjT4Ie5+LbBi23YRQAjxM8C3Al9W+YvJRKm8R+69DDeuXedTlZ9VI2TnMULLV9oAAApYSURBVPS+EV7egP3N/pdXNq7TKX+I5y9u8anK5hn/A/moHbyPbOc+v/zFMlCOTO5hs06n/CG6ua/hU5+P9jM/fTFLp/whXmzd5VN1NbI/u7o/NGHRQ0c3WW61WHjvV/qLTqXC6jq77j78vUNf9rJPcRi7ODNfWSOLqZks5aOl2sSpb82lzX3d3hb89s+oEbLyK5C/A5nh6nbWyDq07wgnLIJTQdCEFosj0T/fMSSX4pItNI2WpvGhLurOWGnd+feJBGac9iRu2UvZpchp33Geb9vIkLMsPvjyMZQP1Ah58blTQT84n3s5v6yO9j0GV9GeeLIXc4vkU/nzv/kKw09AdhfYGPjzJvB7Tn6TEOL7gO8DWF5WNDFnQjxZ/xKfPvhuPn0AFFRJ+bDzz8+tnPj6DeCP8be29oF9VcLH4D8H4H9a++2I5QL8Mf7tFvzb34xadh74Y/ydrSqgTvaf+LrTZ7wxdYO3qvvwc/+dMrkALH3N0B/zZp7Z9Cxfc/NrxvyAwl9leomMnuF9N6Ivmi/ll2hb7Uj3gIGzA20hu8CHb344UrkAeZdG+M7n/wl8/p+oE/TOf3HqS0vTS7z/xvuHKIRRQAjB0vRSLOf77vRdDGHwwYUPRi57Kb/Ea7OvMecO2YgS+6ksHyjvwD//0+qEaAbcHtbrnak7GMLgQ4sfUid3DO7l73F76nbkw4nAedbvv/H+yOXeyt3C0IxYbJk+9wCTT3PzX/01tYI++F2nvnQvf4+bU9FN0vSwkF0gradj0fdSfomZ1EwsCYel6SUezkQ32fGiQti2ffY3CPEdwEdt2/5T7p+/G/ha27b/h3E/8+6779qf+9znpP6iYdBqNfidld8kZ+SYMhRSXnLX+8sHPdi2TbvXJm1ENybaQ8fqgg2mHq3DCtDsNknr6cgpdFHp++5cFkMfzmZ2Ow3s0oZafRsZmDlNa4lT381eM3InHaBn9ejZvciznACtXqtfuYkajcNVsufY7dCYWQJjWK9x6rvda6MLPfI+F4BGtxHL+bZsi47ViXTFgId24wDtaB9Dpb7TMzB1enDHVdR3p9dBCBF5cgli1LfVo3NQIK30MwtnKbQ2bKc7Vgcgcso5xKdv27Zp9VqR966Bq2+bSHe+fSVBCPGbtm2/e973+XkTNoHBzvUlYGvSXywOpNNZ3n3/N8b9a1wxXL3lf4aZPdV3ExXiMLLgVC/iuFzAWQ6tE72DDsTitHnIXosn0xinvuMIAj3Edb41ocV2zlLZ65BV1D92Dq6ivuN0VGPTt6aTno/nvowjEPMQl76FELH5CXHq+yLBT3r3N4BHQoiHQogU8J3Az6v9tRIkSJAgQYIECRIkSJDg8uPcCplt210hxH8P/CtAB37Stu0vKf/NEiRIkCBBggQJEiRIkOCSwxd517btXwR+UfHvkiBBggQJEiRIkCBBggRXCtF3pCdIkCBBggQJEiRIkCBBAiAJyBIkSJAgQYIECRIkSJAgNpw79n6i/6kQu8Bz6f/j8JgH9uL+JRIoR/Kcrw6SZ311kDzrq4PkWV8NJM/56uAqP+v7tm0vnPdNSgKyr1QIIT7nZxdAgouN5DlfHSTP+uogedZXB8mzvhpInvPVQfKsz0dCWUyQIEGCBAkSJEiQIEGCmJAEZAkSJEiQIEGCBAkSJEgQE65aQPZjcf8CCSJB8pyvDpJnfXWQPOurg+RZXw0kz/nqIHnW5+BK9ZAlSJAgQYIECRIkSJAgwVcSrlqFLEGCBAkSJEiQIEGCBAm+YpAEZAkSJEiQIEGCBAkSJEgQE65EQCaE+GYhxFMhxIoQ4gfi/n0SyIMQ4p4Q4teEEI+FEF8SQvwF9+vXhRC/LIR45v77Wty/a4LwEELoQojfEkL8gvvnh0KIz7rP+f8SQqTi/h0ThIcQYk4I8bNCiCfuu/2fJe/05YQQ4n90bfcXhRD/VAiRSd7rywEhxE8KIXaEEF8c+NrI91g4+Nuun/YFIcSH4/vNEwTFmGf9v7k2/AtCiH8uhJgb+LsfdJ/1UyHER+P5rb+ycOkDMiGEDvxd4GPA+4DvEkK8L97fKoFEdIG/ZNv2O8DXAX/Ofb4/APyKbduPgF9x/5zg4uMvAI8H/vw3gL/lPudD4E/G8lslkI0fAX7Jtu23gQ/iPPPknb5kEELcBf488K5t278L0IHvJHmvLws+CXzzia+Ne48/Bjxy//k+4O9F9DsmkINPcvpZ/zLwu2zb/irgPeAHAVwf7TuB97s/86Our36lcekDMuBrgRXbtou2bbeBnwG+NebfKYEk2Lb90rbtz7v/XcVx3O7iPOOfcr/tp4Bvi+c3TCALQogl4FuAH3f/LIDfD/ys+y3Jc74EEELMAP858BMAtm23bdsukbzTlxUGkBVCGEAOeEnyXl8K2Lb968DBiS+Pe4+/FfhHtoPPAHNCiNvR/KYJwmLUs7Zt+1/btt11//gZYMn9728Ffsa27ZZt26vACo6vfqVxFQKyu8DGwJ833a8luGQQQjwAPgR8Frhp2/ZLcII2YDG+3yyBJPww8P2A5f75BlAaMPjJu3058BqwC/wfLj31x4UQUyTv9KWDbdsvgP8dWMcJxMrAb5K815cZ497jxFe73PhvgX/p/nfyrEfgKgRkYsTXkln/lwxCiGng/wb+om3blbh/nwRyIYT4I8CObdu/OfjlEd+avNsXHwbwYeDv2bb9IeCIhJ54KeH2D30r8BC4A0zhUNdOInmvLz8Se35JIYT4n3HaS37a+9KIb7vyz/oqBGSbwL2BPy8BWzH9LgkUQAhh4gRjP23b9qfcL297dAf33ztx/X4JpOAbgI8LIdZwaMe/H6diNudSnSB5ty8LNoFN27Y/6/75Z3ECtOSdvnz4JmDVtu1d27Y7wKeAryd5ry8zxr3Hia92CSGE+F7gjwB/3D5efJw86xG4CgHZbwCP3KlNKZxGwp+P+XdKIAluH9FPAI9t2/6bA3/188D3uv/9vcD/E/XvlkAebNv+Qdu2l2zbfoDzDv+qbdt/HPg14I+635Y850sA27ZfARtCiLfcL/0B4Msk7/RlxDrwdUKInGvLvWedvNeXF+Pe458Hvsedtvh1QNmjNia4mBBCfDPwV4CP27ZdH/irnwe+UwiRFkI8xBnk8v/F8Tt+JUEcB6yXF0KIP4yTTdeBn7Rt+3+J+VdKIAlCiG8E/h3wOxz3Fv1VnD6yfwYs41z632Hb9snm4gQXEEKIjwB/2bbtPyKEeA2nYnYd+C3gT/z/7dyxSYRBEAbQbxKvAQuxBLswPjCwBTEQe7geTM1s4QLBDswNTAVBGIP9I9FADPb0fy9eloFlgm/Z2e5+m1kfv1dVJxmftxwleUqyzbhA1NP/TFXdJDnLeNL0mOQ8Y55EX/9xVXWb5DTJcZLnJNdJ7vJFHy+BfJfx695rkm13P8yom5/75qwvk2ySvCzL9t19say/ypgre88YNbn/vOfarCKQAQAAHKI1PFkEAAA4SAIZAADAJAIZAADAJAIZAADAJAIZAADAJAIZAADAJAIZAADAJB+iN4gJGf6ZwAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 1080x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15,3))\n",
    "plt.plot(idx[0, :5, :5, :5].ravel(), zorder=10)\n",
    "plt.plot(idx[1, :5, :5, :5].ravel(), zorder=9)\n",
    "plt.plot(idx[2, :5, :5, :5].ravel(), zorder=8)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This organization is reflected in `ndarray.strides`, which tells us how many bytes must be traversed to get to the next index in each axis. Each 2-byte step through memory gets me to the next index in axis 2, but I must strude 72200 bytes to get to the next index of axis 0:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(269698, 902, 2)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "volume.strides"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Aside: figure for blog post"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x216 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axs = plt.subplots(ncols=2, figsize=(10,3), facecolor='w')\n",
    "\n",
    "for i, ax in enumerate(axs):\n",
    "    data = idx[i, :3, :3, 0].ravel(order='C')\n",
    "    ax.plot(data, 'o-', c='gray')\n",
    "    ax.text(0, 1.8, f'axis {i}', color='gray', size=15, ha='left')\n",
    "plt.savefig('/home/matt/Pictures/2d-array-corder.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Accessing the seismic data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's make all the dimensions the same, to avoid having to slice later. I'll make a copy, otherwise we'll have a view of the original array. \n",
    "\n",
    "Alternatively, change the shape here to see effect of small dimensions, eg try `volume = volume[:10, :290, :290]` with C ordering."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "volume = volume[:190, :190, :190].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_slice_3d(volume, x, axis, n=None):\n",
    "    \"\"\"\n",
    "    Naive function... but only works on 3 dimensions.\n",
    "    NB Using ellipses slows down last axis.\n",
    "    \"\"\"\n",
    "    # Force cube shape\n",
    "    if n is None and not np.sum(np.diff(volume.shape)):\n",
    "        n = np.min(volume.shape)\n",
    "    if axis == 0:\n",
    "        data = volume[x, :n, :n]\n",
    "    if axis == 1:\n",
    "        data = volume[:n, x, :n]\n",
    "    if axis == 2:\n",
    "        data = volume[:n, :n, x]\n",
    "    return data + 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18.1 µs ± 3.65 µs per loop (mean ± std. dev. of 7 runs, 100000 loops each)\n",
      "43.3 µs ± 7.91 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)\n",
      "136 µs ± 4.85 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit get_slice_3d(volume, 150, axis=0)\n",
    "%timeit get_slice_3d(volume, 150, axis=1)\n",
    "%timeit get_slice_3d(volume, 150, axis=2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's check that changing the memory layout to Fortran ordering makes the last dimension fastest:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "volumef = np.asfortranarray(volume)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "131 µs ± 5.28 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)\n",
      "36.6 µs ± 762 ns per loop (mean ± std. dev. of 7 runs, 10000 loops each)\n",
      "23.9 µs ± 766 ns per loop (mean ± std. dev. of 7 runs, 10000 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit get_slice_3d(volumef, 150, axis=0)\n",
    "%timeit get_slice_3d(volumef, 150, axis=1)\n",
    "%timeit get_slice_3d(volumef, 150, axis=2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Axes 0 and 1 are > 10 times faster than axis 2.\n",
    "\n",
    "What about if we do something like take a Fourier transform over the first axis?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "from scipy.signal import welch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/matt/anaconda3/envs/py36/lib/python3.6/site-packages/scipy/signal/spectral.py:1773: UserWarning: nperseg = 256 is greater than input length  = 190, using nperseg = 190\n",
      "  .format(nperseg, input_length))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "32 ms ± 205 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit s = [welch(tr, fs=500) for tr in volume[:, 10]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/matt/anaconda3/envs/py36/lib/python3.6/site-packages/scipy/signal/spectral.py:1773: UserWarning: nperseg = 256 is greater than input length  = 190, using nperseg = 190\n",
      "  .format(nperseg, input_length))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "38.2 ms ± 4.34 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
     ]
    }
   ],
   "source": [
    "%timeit s = [welch(tr, fs=500) for tr in volumef[:, 10]]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "No practical difference. Hm.\n",
    "\n",
    "I'm guessing this is because the DFT takes way longer than the data access."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "del(volume)\n",
    "del(volumef)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Fake data in _n_ dimensions\n",
    "\n",
    "Let's make a function to generate random data in any number of dimensions.\n",
    "\n",
    "Be careful: these volumes get big really quickly!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "def makend(n, s, equal=True, rev=False, fortran=False):\n",
    "    \"\"\"\n",
    "    Make an n-dimensional hypercube of randoms.\n",
    "    \"\"\"\n",
    "    if equal:\n",
    "        incr = np.zeros(n, dtype=int)\n",
    "    elif rev:\n",
    "        incr = list(reversed(np.arange(n)))\n",
    "    else:\n",
    "        incr = np.arange(n)\n",
    "    shape = incr + np.ones(n, dtype=int)*s\n",
    "    a = np.random.random(shape)\n",
    "    m = f\"Shape: {tuple(shape)}    \"\n",
    "    m += f\"Memory: {a.nbytes/1e6:.0f}MB    \"\n",
    "    m += f\"Order: {'F' if fortran else 'C'}\"\n",
    "    print (m)\n",
    "    if fortran:\n",
    "        return np.asfortranarray(a)\n",
    "    else:\n",
    "        return a"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I tried implementing this as a context manager, so you wouldn't have to delete the volume each time after using it. I tried the `@contextmanager` decorator, and I tried making a class with `__enter__()` and `__exit__()` methods. Each time, I tried putting the `del` command as part of the exit routine. They both worked fine... except they did not delete the volume from memory. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2D data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_slice_2d(volume, x, axis, n=None):\n",
    "    \"\"\"\n",
    "    Naive function... but only works on 2 dimensions.\n",
    "    \"\"\"\n",
    "    if n is None and not np.sum(np.diff(volume.shape)):\n",
    "        n = np.min(volume.shape)\n",
    "    if axis == 0:\n",
    "        data = volume[x, :n]\n",
    "    if axis == 1:\n",
    "        data = volume[:n, x]\n",
    "    return data + 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape: (6000, 6000)    Memory: 288MB    Order: C\n",
      "19.9 µs ± 1.9 µs per loop (mean ± std. dev. of 7 runs, 100000 loops each)\n",
      "73.7 µs ± 5.58 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)\n"
     ]
    }
   ],
   "source": [
    "dim = 2\n",
    "\n",
    "v = makend(dim, 6000, fortran=False)\n",
    "for n in range(dim):\n",
    "    %timeit get_slice_2d(v, 3001, n)\n",
    "del v"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape: (6000, 6000)    Memory: 288MB    Order: F\n",
      "71.9 µs ± 1.77 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)\n",
      "29.8 µs ± 2.76 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)\n"
     ]
    }
   ],
   "source": [
    "dim = 2\n",
    "\n",
    "v = makend(dim, 6000, fortran=True)\n",
    "for n in range(dim):\n",
    "    %timeit get_slice_2d(v, 3001, n)\n",
    "del v"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This has been between 3.3 and 12 times faster."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1D convolution on an array"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "def convolve(data, kernel=np.arange(10), axis=0):\n",
    "    func = lambda tr: np.convolve(tr, kernel, mode='same')\n",
    "    return np.apply_along_axis(func, axis=axis, arr=data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape: (6000, 6000)    Memory: 288MB    Order: C\n",
      "705 ms ± 24.7 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)\n",
      "301 ms ± 6.17 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)\n"
     ]
    }
   ],
   "source": [
    "dim = 2\n",
    "\n",
    "v = makend(dim, 6000, fortran=False)\n",
    "%timeit convolve(v, axis=0)\n",
    "%timeit convolve(v, axis=1)\n",
    "del v"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape: (6000, 6000)    Memory: 288MB    Order: F\n",
      "301 ms ± 3.68 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)\n",
      "698 ms ± 67.5 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)\n"
     ]
    }
   ],
   "source": [
    "dim = 2\n",
    "\n",
    "v = makend(dim, 6000, fortran=True)\n",
    "%timeit convolve(v, axis=0)\n",
    "%timeit convolve(v, axis=1)\n",
    "del v"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Speed is double on fast axis, i.e. second axis on default C order."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## `np.mean()` across axes\n",
    "\n",
    "Let's try taking averages across different axes. In C order it should be faster to get the `mean` on `axis=1` because that involves getting the rows:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([ 6., 12.]), array([ 3., 15.]))"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "a = [[ 2,  4],\n",
    "     [10, 20]]\n",
    "\n",
    "np.mean(a, axis=0), np.mean(a, axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's see how this looks on our data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape: (6000, 6000)    Memory: 288MB    Order: C\n",
      "24.9 ms ± 318 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "19.5 ms ± 923 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n"
     ]
    }
   ],
   "source": [
    "dim = 2\n",
    "\n",
    "v = makend(dim, 6000, fortran=False)\n",
    "%timeit np.mean(v, axis=0)\n",
    "%timeit np.mean(v, axis=1)\n",
    "del v"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape: (6000, 6000)    Memory: 288MB    Order: F\n",
      "24.4 ms ± 3.03 ms per loop (mean ± std. dev. of 7 runs, 100 loops each)\n",
      "33.9 ms ± 700 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
     ]
    }
   ],
   "source": [
    "dim = 2\n",
    "\n",
    "v = makend(dim, 6000, fortran=True)\n",
    "%timeit np.mean(v, axis=0)\n",
    "%timeit np.mean(v, axis=1)\n",
    "del v"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We'd expect the difference to be even more dramatic with `median` because it has to sort every row or column:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape: (6000, 6000)    Memory: 288MB    Order: C\n",
      "1.49 s ± 30.4 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)\n",
      "671 ms ± 9.59 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)\n"
     ]
    }
   ],
   "source": [
    "v = makend(dim, 6000, fortran=False)\n",
    "%timeit np.median(v, axis=0)\n",
    "%timeit np.median(v, axis=1)\n",
    "del v"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape: (6000, 6000)    Memory: 288MB    Order: C\n",
      "32.5 ms ± 721 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "27.2 ms ± 316 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
     ]
    }
   ],
   "source": [
    "v = makend(dim, 6000, fortran=False)\n",
    "%timeit v.mean(axis=0)\n",
    "%timeit v.mean(axis=1)\n",
    "del v"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3D arrays\n",
    "\n",
    "In a nutshell:\n",
    "\n",
    "C order: first axis is fastest, last axis is slowest; factor of two between others.\n",
    "\n",
    "Fortran order: last axis is fastest, first axis is slowest; factor of two between others."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape: (600, 600, 600)    Memory: 1728MB    Order: C\n",
      "303 µs ± 62.9 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)\n",
      "958 µs ± 103 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)\n",
      "9.07 ms ± 345 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n"
     ]
    }
   ],
   "source": [
    "dim = 3\n",
    "\n",
    "v = makend(dim, 600)\n",
    "for n in range(dim):\n",
    "    %timeit get_slice_3d(v, 201, n)\n",
    "del v"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Non-equal axes doesn't matter."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape: (602, 601, 600)    Memory: 1737MB    Order: C\n",
      "427 µs ± 18.1 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)\n",
      "882 µs ± 27.5 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)\n",
      "8.85 ms ± 398 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n"
     ]
    }
   ],
   "source": [
    "dim = 3\n",
    "\n",
    "v = makend(dim, 600, equal=False, rev=True)\n",
    "for n in range(dim):\n",
    "    %timeit get_slice_3d(v, 201, n)\n",
    "del v"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Fortran order results in a fast last axis, as per. But the middle axis is pretty fast too."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape: (600, 600, 600)    Memory: 1728MB    Order: F\n",
      "8.92 ms ± 310 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n",
      "848 µs ± 17 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)\n",
      "442 µs ± 19.1 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)\n"
     ]
    }
   ],
   "source": [
    "dim = 3\n",
    "\n",
    "v = makend(dim, 600, fortran=True)\n",
    "for n in range(dim):\n",
    "    %timeit get_slice_3d(v, 201, n)\n",
    "del v"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For C ordering, the last dimension is more than 20x slower than the other two."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4 dimensions\n",
    "\n",
    "Axes 0 and 1 are fast (for C ordering), axis 2 is half speed, axis 3 is ca. 15 times slower than fast axis."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_slice_4d(volume, x, axis, n=None):\n",
    "    \"\"\"\n",
    "    Naive function... but only works on 4 dimensions.\n",
    "    \"\"\"\n",
    "    if n is None and not np.sum(np.diff(volume.shape)):\n",
    "        n = np.min(volume.shape)\n",
    "    if axis == 0:\n",
    "        data = volume[x, :n, :n, :n]\n",
    "    if axis == 1:\n",
    "        data = volume[:n, x, :n, :n]\n",
    "    if axis == 2:\n",
    "        data = volume[:n, :n, x, :n]\n",
    "    if axis == 3:\n",
    "        data = volume[:n, :n, :n, x]\n",
    "    return data + 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape: (100, 100, 100, 100)    Memory: 800MB    Order: C\n",
      "1.64 ms ± 12.8 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)\n",
      "1.75 ms ± 38.7 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)\n",
      "4.4 ms ± 54.6 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n",
      "25.1 ms ± 692 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
     ]
    }
   ],
   "source": [
    "dim = 4\n",
    "\n",
    "v = makend(dim, 100, equal=True)\n",
    "for n in range(dim):\n",
    "    %timeit get_slice_4d(v, 51, n)\n",
    "del v"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape: (100, 100, 100, 100)    Memory: 800MB    Order: F\n",
      "22.5 ms ± 3.05 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "4.23 ms ± 81.4 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n",
      "1.7 ms ± 27.5 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)\n",
      "1.63 ms ± 55.2 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)\n"
     ]
    }
   ],
   "source": [
    "dim = 4\n",
    "\n",
    "v = makend(dim, 100, equal=True, fortran=True)\n",
    "for n in range(dim):\n",
    "    %timeit get_slice_4d(v, 51, n)\n",
    "del v"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5 dimensions\n",
    "\n",
    "We are taking 4-dimensional hyperplanes from a 5-dimensional hypercube. \n",
    "\n",
    "Axes 0 and 1 are fast, axis 2 is half speed, axis 3 is quarter speed, and the last axis is about 5x slower than that."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_slice_5d(volume, x, axis, n=None):\n",
    "    \"\"\"\n",
    "    Naive function... but only works on 5 dimensions.\n",
    "    \"\"\"\n",
    "    if n is None and not np.sum(np.diff(volume.shape)):\n",
    "        n = np.min(volume.shape)\n",
    "    if axis == 0:\n",
    "        data = volume[x, :n, :n, :n, :n]\n",
    "    if axis == 1:\n",
    "        data = volume[:n, x, :n, :n, :n]\n",
    "    if axis == 2:\n",
    "        data = volume[:n, :n, x, :n, :n]\n",
    "    if axis == 3:\n",
    "        data = volume[:n, :n, :n, x, :n]\n",
    "    if axis == 4:\n",
    "        data = volume[:n, :n, :n, :n, x]\n",
    "    return data + 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape: (40, 40, 40, 40, 40)    Memory: 819MB    Order: C\n",
      "4.21 ms ± 51.1 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n",
      "3.71 ms ± 593 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n",
      "6.14 ms ± 166 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n",
      "11.3 ms ± 273 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n",
      "52.2 ms ± 2.86 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
     ]
    }
   ],
   "source": [
    "dim = 5\n",
    "\n",
    "v = makend(dim, 40)\n",
    "for n in range(dim):\n",
    "    %timeit get_slice_5d(v, 21, n)\n",
    "del v"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape: (40, 40, 40, 40, 40)    Memory: 819MB    Order: F\n",
      "48.6 ms ± 2.69 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "11.2 ms ± 192 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n",
      "6.06 ms ± 25.8 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n",
      "4.47 ms ± 37.9 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n",
      "3.82 ms ± 635 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)\n"
     ]
    }
   ],
   "source": [
    "dim = 5\n",
    "\n",
    "v = makend(dim, 40, fortran=True)\n",
    "for n in range(dim):\n",
    "    %timeit get_slice_5d(v, 21, n)\n",
    "del v"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "What about when we're doing something like getting the mean on an array?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape: (40, 40, 40, 40, 40)    Memory: 819MB    Order: F\n",
      "128 ms ± 28.2 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "105 ms ± 12.6 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "74.1 ms ± 3.18 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "89 ms ± 6.08 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "110 ms ± 6.72 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
     ]
    }
   ],
   "source": [
    "dim = 5\n",
    "\n",
    "v = makend(dim, 40, fortran=True)\n",
    "for n in range(dim):\n",
    "    %timeit np.mean(v, axis=n)\n",
    "del v"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 6 dimensions and beyond\n",
    "\n",
    "In general, first _n_/2 dimensions are fast, then gets slower until last dimension is several (5-ish) times slower than the first."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_slice_6d(volume, x, axis, n=None):\n",
    "    \"\"\"\n",
    "    Naive function... but only works on 6 dimensions.\n",
    "    \"\"\"\n",
    "    if n is None and not np.sum(np.diff(volume.shape)):\n",
    "        n = np.min(volume.shape)\n",
    "    if axis == 0:\n",
    "        data = volume[x, :n, :n, :n, :n, :n]\n",
    "    if axis == 1:\n",
    "        data = volume[:n, x, :n, :n, :n, :n]\n",
    "    if axis == 2:\n",
    "        data = volume[:n, :n, x, :n, :n, :n]\n",
    "    if axis == 3:\n",
    "        data = volume[:n, :n, :n, x, :n, :n]\n",
    "    if axis == 4:\n",
    "        data = volume[:n, :n, :n, :n, x, :n]\n",
    "    if axis == 5:\n",
    "        data = volume[:n, :n, :n, :n, :n, x]\n",
    "    return data + 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape: (23, 23, 23, 23, 23, 23)    Memory: 1184MB    Order: C\n",
      "26.6 ms ± 498 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "32.2 ms ± 3.94 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "25 ms ± 525 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "34.5 ms ± 1.82 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "44.9 ms ± 1.55 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n",
      "113 ms ± 5.95 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
     ]
    }
   ],
   "source": [
    "dim = 6\n",
    "\n",
    "v = makend(dim, 23)\n",
    "for n in range(dim):\n",
    "    %timeit get_slice_6d(v, 12, n)\n",
    "del v"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}