qcd-top-jets-lund-plane-toimg-weighted.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import sys\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sklearn.preprocessing import KBinsDiscretizer as kdb\n",
    "from preprocess_qcd_top_jets import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from matplotlib import pyplot as plt\n",
    "from matplotlib.font_manager import FontProperties\n",
    "plt.rcParams[\"font.family\"] = \"Times New Roman\"\n",
    "plt.rcParams[\"mathtext.default\"] = \"rm\"\n",
    "plt.rcParams['text.usetex'] = True"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this notebook I will demonstrate how to jets in the Lund plane representation into images that can be fed to a neural network."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Useful reference for the Lund jet plane: https://arxiv.org/pdf/1807.04758.pdf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "qcd_lund_jets = np.load( \"qcd_lund_jets.npy\", allow_pickle=True )\n",
    "top_lund_jets = np.load( \"top_lund_jets.npy\", allow_pickle=True )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now I want to separate the data into separate planes:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [],
   "source": [
    "# the default weight=1 corresponds to pt, 2 is mass, 3 is mass drop, ...\n",
    "def get_plane_w( lund_jet_sample, plane_id, weight=1):\n",
    "    return [ [ [i[6],i[7],i[weight]] for i in jet if i[0]==plane_id ] for jet in lund_jet_sample ]\n",
    "\n",
    "qcd_lund_0 = get_plane_w( qcd_lund_jets, 0 )\n",
    "qcd_lund_1 = get_plane_w( qcd_lund_jets, 1 )\n",
    "qcd_lund_2 = get_plane_w( qcd_lund_jets, 2 )\n",
    "qcd_lund_3 = get_plane_w( qcd_lund_jets, 3 )\n",
    "\n",
    "top_lund_0 = get_plane_w( top_lund_jets, 0 )\n",
    "top_lund_1 = get_plane_w( top_lund_jets, 1 )\n",
    "top_lund_2 = get_plane_w( top_lund_jets, 2 )\n",
    "top_lund_3 = get_plane_w( top_lund_jets, 3 )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Getting the jet masses:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "qcd_masses = np.array( [ np.max( [ i[2] for i in jet ] ) for jet in qcd_lund_jets ] )\n",
    "top_masses = np.array( [ np.max( [ i[2] for i in jet ] ) for jet in top_lund_jets ] )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(qcd_masses, alpha=0.5)\n",
    "plt.hist(top_masses, alpha=0.5)\n",
    "plt.xlabel(\"mass (GeV)\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And then flatten the arrays so that we can look at the different densities for QCD vs top:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def flatten( lund_jet_sample ):\n",
    "    jf = []\n",
    "    for jet in lund_jet_sample:\n",
    "        jf += jet\n",
    "    return jf\n",
    "\n",
    "def flatten_w( lund_jet_sample ):\n",
    "    jf = []\n",
    "    for jet in lund_jet_sample:\n",
    "        jf += [ i[0:2] for i in jet ]\n",
    "    return jf\n",
    "\n",
    "qcd_lund_0_flat = flatten_w( qcd_lund_0 )\n",
    "qcd_lund_1_flat = flatten_w( qcd_lund_1 )\n",
    "qcd_lund_2_flat = flatten_w( qcd_lund_2 )\n",
    "qcd_lund_3_flat = flatten_w( qcd_lund_3 )\n",
    "\n",
    "top_lund_0_flat = flatten_w( top_lund_0 )\n",
    "top_lund_1_flat = flatten_w( top_lund_1 )\n",
    "top_lund_2_flat = flatten_w( top_lund_2 )\n",
    "top_lund_3_flat = flatten_w( top_lund_3 )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And then plot:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA30AAAHPCAYAAADnF7PuAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/Il7ecAAAACXBIWXMAAAsTAAALEwEAmpwYAACRg0lEQVR4nO39f5hbd3knfr8/2NhJ/AN5HBtju8aRjTEhBDMehwBpmibjNE350iyME3bL09IumWG70OW52lr1stezu8+3rRnv9nrolnYZm+6mbGkTZuAb2EBDRskmkAYTzwgDJphgy45xTPCPseL8tInzef6QNPrcb/04o5FmNHP0fl2XL+vWOdI50pyPNGfOfX9u572HiIiIiIiIxNNrWr0DIiIiIiIiMnV00iciIiIiIhJjOukTERERERGJMZ30iYiIiIiIxJhO+kRERERERGJMJ30iIiIiIiIxppM+ERERERGRGIvVSZ9zLumcG3XO9TrnOp1zPc65AedcZ6v3TUREREREpBVcnJqzO+eSAIYBdABIAEgDSHnvM63cLxERERERkVaJ40lfQid5IiIiIiIiebFK7xQRERERERFrbqt3YCo45xIAkgBy3vtsi3dHRERERESkZeJ4pe8OAF0AsgASzrnBwkmgiIiIiIhI24lVTR+Qv8rnvc8FcS+Ard77bRXW7QXQCwALFizYvHHjxvyCV39hV3w5h7pckrDxa15b3+NFWmR0dPS0937ZRNevOoZE2pTGkMjk1Tt+AI0hkVCtMRS7kz5WmNzlMIAl4ckg6+rq8iMjI/ng+WfswoP31rfRjbfZeOGK+h4v0iLOuVHvfddkHmvGkEib0hgSmbxGxg+gMSRSawzFKr2z8NceNlb4Pzmd+yIiIiIiIjITxOakr3BFb6DwfyWa0EVERERERNpObGbv9N5nnXN9FWbrvB1AplZqZ5kXT9t48eray08esPEDf2TjTR+2cbJ7wrsiIiIiIiLSiNhc6SsYC6/0FWbt7ANwZ8v2SEREREREpIVic6UPALz3Q865HudcT+GudQC2qVefiIiIiIi0q1id9AH5E7+Gn2TeQhvz7Jvnjtt47iU2vuqDNu5Yb2NOBw2X83OJiIiIiIg0IG7pnSIiIiIiIhLQSZ+IiIiIiEiM6aRPREREREQkxmJX09cUlyRsfODu2suXX2Xjow/b+OC9Nl7ZZeOXc9WXqcZPREREREQaoCt9IjVks1ns2rWr1bsh0nZyuRy2bduGJUuWIJvVBMwik6HvMJHGxGkM6aRPpIZ0Oo2BgYFW74ZI20kkEhgcHGz1bog01VT9AaPa8+o7TOKglX/4i9MY0klfJZck7L+FK+y/c8ftvxMj9t/LOftvzXX238kD1f+NHbL/pKV6e3tx+PDhVu+GSNvq6Oho9S6INM3QUONdpep5Xn2HSRxM1biZiDiNIZ30iYiIiEyD4eHhWfW8IjOBju/m0EmfiIiIyBQbGhpCOp2eNc8rMhPo+G4ezd4psZTJZJBKpTAyMoIjR46Mf2Ds27cPW7ZsQU9Pj1kvm81ieHgY6XQaw8PD6OvrQzKZRF9fH9LpNLz3Zv10Oo2zZ8+OP+/w8DC2bt2Knp6e8TSEYg769u3bzb4Vl4+NjeHw4cPo7+8v2+9K+5PNZtHf34+xsTHs2bPHvIabbroJyWQSe/bsQWdn51S9rTILDQ0NIZFIAMhPjjI8PIz+/v7x+3bv3j2eQrlv377xY78ok8kgnU4jmUxibGwMXV1d5hhLp9PIZrPo6OjA2NgYkskkuru7xx9babzwOKy0LQBmP/g1ARpDMnsMDQ1h3759AIBUKgUAWLduHXp7e8fXqTYWax3TuVyu6vNms9my77BwfwCNIalfJpPBwMAANm/eDADo7u5GJpNBZ2enOV7D4y6Xy+HOO+9EOp3Ggw8+aI6PWsfiVI2bjo6OCX03xW4Mee/1z3ts3rzZjzt7xP576lv237f/f/bfvv9u/z34SfvvS79l/9Va/x9/0/7jffnFS/afVHX27FkPwA8ODpr7u7u7/cDAgFkvkUiMr9ff328ekx8mVqXnBeCHh4fNfclk0h8+fHg8Hh4eNvHg4KDv7u4u2+9q+zM8POyTyWTZ/vT391d4B+oHYMQ3YwzJjDA8PFx2TPb29vqzZ8967/PHDR87yWRyfPnw8HDZ8Tk4ODi+fHBw0G/fvt0s7+/vN+PL++rjhfeVtzU8POwBaAxJbFT6PvE+eixO5nuq2rKZPIYaGT9eY2jKDQ8P+87OTnPf4OCg7+npMceU95WPyWQy6UdHR83zRR2L1Z7L++aMm6jvpkr3zdYx1PKTran8ByABYGAi6+qkL54qDd7R0VEPYPxDobhe+EEU9RyV1k8mk2W/7HZ3d5sPlN7eXt/b2zup5y9KJBJly/hDa7L0C2u8VPpCHB4e9mfPnh3/owjr7e0dP46TyWTFP2QUj7dEImHGURHfX228hPdxHD5X+OWqMSSzWaVjdSJjsfjYer6nqi2byWNIJ30zW/j5H+I/zhXvq/T48Lip51hkzRo3Ud9NlbY/W8dQ3NM797R6B2TmKV4yHxkZGU9DC++fqEqpZ11dXSZOJBIYGxsbj1OpFHK5XNk6uVxuPN0uan927NiBgYGB8SmE0+m0eR0iRT09Pdi8eTOGh4exbds2dHd3jx8rxbRPrpUYGxvD6Ogostksstls2bE1MDCA7u5upNNpdHR0lB23QH5spNNpk75ZLVUTwPi2JjIGNYYkbtLpdM2xGGpGypfGkExGte8EABW/ByainmORNWvc1Ppuqma2jqHYnvQ557oBZAHU/9Oce4mNuXXCJQkbd6y38f67bNz1URsfvNfGa28o3Z630C47vtfG62+xMe+rTEgymWy478tkPuSKeea7du1CIpFAMpks++CI0tvbiyuuuGL8g2IiH44zzquv2Pg1sf0oaqlkMokjR45g586dGBgYQF9fH3p7ezEwMICxsTF0dHSUfcmEJ4WVFJdns9maxx2Pr1rrZjKZCbyaPI0hiZuosdhsGkMyGcXP6Uo/58m21mnkWGzWuGmn3+Vi+ZuWcy5RuHkYkznpk9jLZrOT+utOo3bt2oXh4WEz/XC9gzyRSKCrqwu7d+/G7bffri9aqar4l8NigXkul8O2bdswNDQ0PjFLNVHjI+pLrp7xVc+6GkMSF5lMBslkMnIsTvZ5qx3XGkMyGVPxO9NkjsWpGjf1mK1jKK4tG2733mt+V6kok8kgkUi0JBUllUqN/1WnKJfLYWxsDJlMZsJ/Kerr68PAwIBSaqSm4eFhc8UtkUigv78f+/btM1fsWHEmtkQiUfEqXCaTQVdXF8bGxsqO2Vwuh0wmU9dxWdxWpX3h59cYkrgoXi2PGouTfd5qNIZkMoqf05XaJ0w0c4rXm8yxOFXjph6zdQzF7qSvkNb5xVbvh8wcnKaWSqWwZ0+p3LPeS/KTfVxx/TANovgXq+KHRTEnPEpPTw+y2WzL/sols0c4jTSQ/4LcunUrAGDPnj3j02AXFeskqi0PU3z6+/uxc+dOs3znzp2mJURR1HFdbV8qPYfGkMxW1UoLosZi1DE90ZIFjSFpxJ49e8q+U6qVAnR2dpqTr7AdDzDxYxGYunEz0XUqrT8bx1Cs0jsLaZ0J731uguv3AugFgDVr1pQWcM0e182dPGDjV1628VUftDHX+K25zsaH7i/d5po9fu4TI7X3jZ9b0NnZiXQ6Pd7PKJVKmT5ixV9a+/r6xie7ADDeS6W4rNhXpXjfnXfeiR07diCZTGLnzp3IZrNIpVLjz1/sAVPsYdbT04PBwUGkUils3bp1PP1mYGAA99xzD7Zs2VJzf1hvby9uv/32KX3volQdQ1G4hi+qxo+X11pXxq1btw5dXV3jk7YUv4SKx1RPTw8SiQRSqRS2bNkyXptQ/GLm5UD+y7dYmN7b24t0Oo1du3aNp9pwj6OJjpfitnbt2oXOzk7kcrnxvyz39fWNr6cxJLPZ4OAgBgYGsGXLFvMLcK2xOJFjutLzVvoOSyQSGkMyaZW+E4qf02zPnj0YGBgY/yNj8Vi788470d/fP6HP86Jmj5uJfjclk8lYjSGXn90zHpxzvd773WEMYKv3flvUY7u6uvzISOGEik+0eOIVPulbvNrGUY/nE7PwRI5P+vgElLelk76anHOI0zEeGhoaKmtu3Sjn3Kj3vit6zXJmDNVLJ33SArEaQyIt0Owx1Mj4ATSGWmXJkiUYHR1tyVwJs910jqHYpHc65zoBqI5P2oJmShNpjMaQSGM0hkQaM91jKE5/Ik8C2OKcC+/rBpBwzvUD2Oe9r5x4HCV3tPZybulw2eU25qtzzz9Tff2FK+wyvqp47FEbJ+lycTZde3kbmWyt3ky1bds2bN26dTylrtlXKFoq6mpdrSt/av8gExTrMSQyDTSGRBrTyjEUmyt93vsh730q/AfgHgCZQjy5Ez6ZlTKZDO68804A+XzqRnvyzQR9fX3o6OjA0NBQUxr0irQbjSGRxmgMSSibzaKvrw+5XA6pVGrKZ82Mg1aOobj/SXxpq3dAWqOzsxODg4Ot3o2m0pTYIo3RGBJpjMaQhJLJJAYGBsraF0h1rRxDsbnSF3LOJQspnT0Aup1z/YWaPxERERERkbYSyyt93vssgFThX/14RkyeQZPr7rhGL2zBAADLr6r9fGHdHtfk8b5w/SDPFHruuI25xcN1fwKRGa/WbJ2A6vZERERE6hDLK30iIiIiIiKSp5M+ERERERGRGFOOVCWcMsltE7gB+tGHay9/OVd7e/MWVt82p3fyvjBOJf38jtr7pmbuMhvU06ydx1A4vkRERETakK70iYiIiIiIxJiu9Ik0IJ1OY3h4GOvWrQOQbwifSCTQ29tb83GZTAY7d+4EkJ/yGAC2bNmCnp4eZDIZjI2Nobu7G6lUCkNDQ8hms+jt7UUikRjfztjY2HiDT5HZSmNIpDEaQyKNaZcxpJM+kUlKpfKTw/b395v7h4aGsG3btqp9Anft2oXh4WEMDAyMf0gA+Sanu3fvxuDgoHnudevWIZVKVeyDs2vXrprbEpnJNIZEGqMxJNKYdhpDOumrhOvoVnbZ+MDdNuaWDdwm4bLLaz8+fH6u/+P2DutvsTG3h3j8ERuvXWDju2+z8YeCx/PrlKpSqRSy2WzFAdrT0zP+l5vh4WGzrPghwfcD+b8SdXd3o6+vb/yDoqijo6Pifmzfvh2pVAp9fX2zuzkq1+hF1fCF44THCNfwRT23tITGkEhjNIZEGtNuY0g1fSJ1ymaz2LVrV9lfhUK9vb3IZrMYGhoyj6v2V56iZDJZ9yX+/v5+7N69G9lstq7HibSKxpBIYzSGRBrTjmNIJ30iderv70dnZ6e5nF9JT0+P+StPKpWa0OO2bt1a9z4Vc8ZFZgONIZHGaAyJNKYdx5DynCqJaptw1QdtzOmcY4dsnFhb+/HHHi3d5hYKvG1OJd14G+3L35rwuYdeMPGiDTbG32wp3f6DH9hl3P5BAAAjIyPo6opOhV23bh2y2Syy2SySySTS6fSE/vLT09ODXC5X1z51dnaav0TFHo+LUK12DjIjaAyJNEZjSKQx7TiGdKVPpE6ZTGZ8hqdaih8mxUv1uVwOS5cundA2ijM7TVTxQ0lkNtAYEmmMxpBIY9pxDMXuSp9zrgdA8ZrrOgCj3vvdLdwliaGJDOR6B7tIO9EYEmmMxpBIY9ptDMXqpK9wwpfx3g8F94065xLe+10t3DWJkWQyicOHD0euV/xrTTHvO5FI4MyZM1OyT7lcLjK/XGSm0BgSaYzGkEhj2nEMxeqkD8AOAGkAYRVkGkAfgMmf9HEdHdfssdxRG3PLhnPHbRzWzvFzc+3S179i4y66NH3Bhos22HjsgI07NgbByGftwhv/1MY8NX6b6uzsnNDl92w2i0QiMT6Au7u7kU6nJ/S4XC6Hzs7OCe/Tvn376lp/xomqo+UxuHBF9efitieMn5tjtXSYchpDIo3RGBJpTDuOobj9dpMCMEb3JQAoyVyaZseOHdi8eTNyuVzNy/4DAwPYsWPHeFxszlksBq5mokXC/JgHH3ywrseItIrGkEhjNIZEGtOOYyhWE7l479Pe+0wxds4lANwOe+VPpCGdnZ3o7e3Fzp07x+8bGhrC0NAQMpn84ZdOp5HL5bB9+/bxdZLJJPr7+9HX19fU/dm9eze6u7v1F1aZNTSGRBqjMSTSmHYcQ3G70gfATOayFMBm731jV/pW0pSuh+63Mbc24NQzTufkdDJu0xA6cLeNNy6w8eLVNr6M8pNP2vAIp3feEjzf3X9tFx6818a/Tw9u43TPgYEB9PX1Yffu3ejq6kJ3dzcSiQTS6TQymQz6+/sxOjpa9rjiB8fWrVsxMDBg/kqUy+Wwe/du8+FSNDbGF7Dzdu/ejcHBQQwODjbplbVIrRYMQHn6Z5iC+eLp2o/l45TTP3nbF5638byFtZ9fJkVjSKQxGkMijWm3MRTLk77iRC6Fk7+Ucy7lvc/xes65XgC9ALBmzZpp3UeZ/QYGBpBOp3HPPfcgl8th69atyGazOHz4MIaHh8fX49SB7du3jzfg7OjoQCKRGJ82mD8kUqnU+F+aUqkUli5dOl5AnM1msWXLFrOt6aYxJI3QGNIYksZoDGkMSWPaaQw57/2Ub6SVnHPbAdzhvd9ca72uri4/MlJoss5XFbj5+jdpghO+0neSrogxvrKw/pbq6/KVPt4Xvgp54BHaFxuO3mfjzX8UXOk7SI3bN62ysa70TUgqlUJ/f/94PDQ0hJ6enhbu0cQ550a999HdSiswY6heUQ3VeXKkjvWl241e6ePlPP51pW/aaQyJNGa2jqFGxg+gMSTNE8cxFKuavirSADqdc0o0l2mRyWSQy+UA5PPB49TjRWQ6aAyJNEZjSKQxcRxDsUnvdM4lAYyivIYvV/g/CSDDj6uI/9LP+Epe1HTzfDWOhVfzuP6Pn5uvKh6kK3t8wYR+wms22vjxT5Wu7r3lGrvshYNPm3jFVXSFc9OHa+9bm+rs7MTmzZvHp/WdSB8YIXw1LryyB9grg1FX8qKeW1esZxyNIZHGaAyJNCaOYyhOV/oSAEZQ3rKhWF05sRM+kQbt2LEDiUQCX/ziF01qgIhMjMaQSGM0hkQaE8cxFJsrfd77jHOuUhVkCsDuhmfwFJmgRCJRcbYnEZkYjSGRxmgMiTQmjmMoNid9AOC93+Wc63XOrQNwBsA6AMPe+11N3RC3SYhKL9t/l4053TNMF+WUSW6bMHaItmXDFw7amOfHuEjxJZehqhW3UXuIqEllPkStLKKm4Zf2xQdmVIo0t1Go9Vg+Ljnt+PlnbBw1kUut4/g1sfoIFRERkZiK3W8s3vvdrd4HERERERGRmSJONX0iIiIiIiJCdNInIiIiIiISY7FL72wKruE5RLVqUS0buFn0NR+zMdfG1VoWNb087eoCKl968C9svOz19PQvlm6/cM4u+8X9tll7x43UvJ0de9TGXLuoqfGliGvheMzx8lo1fVyjd9nlNubxuXCFjXk8czN23rfwOOb9inodIjNBPXWrIiISC7rSJyIiIiIiEmM66RMREREREYkx5R5VwimVuaM2fpHSHF/8no0vo1YHRx+2cWJt9W1HpUA+9oiN19ptvfBNu2/X3GRXX5S08dPB7PYnj9lll1KW28m7bLzx1qftHfd91Mbv/ayNw3RPpXpKLdzSgdPPTge9SXi8cvomL2c8vvnY5HTtsGUL71fH+trbEpkJGk3nVHqoiMisoyt9IiIiIiIiMaaTPhERERERkRjTSZ+IiIiIiEiMqaavEq5P4Jqe1W+vvZxbFZwYsTFPKR9u7xDV7HF9IDtua/i4ZQOoNOrrf2njX/9A6faqa+2yRz9n4+t6bPxC1sYLOg7bOx74Ixt/8F5UpRo/CXHNELdlCHGN7JP32XjNdbW3xY8P6wWB8hrBc8dLt8P6vkq4pQO3gxCZjVTDJyIy6+hKn4iIiIiISIzF7kqfc64HQAeAdQA6AQx474dau1ciIiIiIiKtEauTvsIJX8Z7ny3ECQCjzrmk937XhJ8oahr2g9+w8eKlNj55wMYnqLXBVfR84faWr7PLDlDKZIcNcYLi5TYce8jGt/47Gz/6N6Xbb6B2Dms22vgYvaw1lA5ati8dlPY69MHS7Z677TJ+z5U+FG/ckoHTOaOE6cCc+snpmpxeHZVKzOOX0zvDfR07ZJdFfXbw635NrD6CZSYLj1t9voqItJ24pXd2FE/4AMB7nwOQAtDfsj0SERERERFpodic9DnnkgAGnHOdtChTWM73i4iIiIiIxF5sTvoKV/h2AaA5JZEo/M/3i4iIiIiIxF6sCkq896kKd9+BfJ1fbtJP/DI9lGr4nnvgjIkXXWVjJKlOL5zyHbAtHEaohm8jtWw4ZFs0/OQxu/hNK218hursOtbaeMvNwW6N2WUXqfzoh/tsfCnNPr+M20WMUS3j+ltKtw/ea5clu23MU+GrpUO8RNWycZ0e1/zVapvALVSi6gV5W1yHd+xRG4efB8vpoOfHcn0g48eLTBXV8YmItLXYXOmrpDCRSy+AbVWW9zrnRpxzI6dOnZrWfROJA40hkcZoDIk0RmNIZGJifdIHYBDAtnByl5D3frf3vst737Vs2bJp3jWR2U9jSKQxGkMijdEYEpmYWKV3hpxz/QD6vffpuh/M6Zw0Lfv5b1I6J7VJOE+zuM8/Rymb17zdxg99p3R7A+3LAZvOef6cXXyBM9foJ7qSss245cP8IIXz4km77DlK93zrFhtzOmjZRy2lh+L+vy3dvv437TKeFp/T4q76IGQW41YFjFPPOJ2Xl4fHy+mDdllUamjuqI05fZO3zY8P94X368n7bHw59T3hNOYLz1d/7krbnkc51SKtwCnUYYmCiIjMSLG80uec6wUwPKkTPhERERERkRiJ3ZU+51w3gBHvfSa4LwmMz/ApIiIiIiLSNmJ1pa9wwpcAkHPOJYv/APTphE9ERERERNpRbK70FWbqHK6yOAugUjuHiaH6ssxDdvEqqpt7/Ac27vnWTlrhMzYOawIXr6KN27YHL1BN3/pNtPoxGy6gFg7HaN8XJkq3D+23y677iI3HqMzuCMWv0nO/6d02Ph/UAM4/+BW7cONtNuZapqhp9zUd+ezGLRy4rpZ//mGdbVQd3MouG498tva+UA1v2bEZLudtnRixMdcHcv0hL+faVtXwyVSotyaPj/N619fns4hIy8XmpK/Qh8+1ej9ERERERERmklild4qIiIiIiIilkz4REREREZEYi016Z1Nx/UGHLdqbM/d7JuYavuuon90Lf7rDxAs20fbCsp1DtoYPCdoVeuipo3QH9c67nGr6fon6ALqgnvAN1KfvPNUH/oymwnnXbbRt6lfoj9t4/pogoG2V1ULRex65PKqvm8wsXMPH/er458nCmj9el/vu8bHANXt8LD3xiI2P77XxmutKt7nnH9cycc/ALHWR4b59jPd98era64tMRL21ojw+6xXWEHI9IfeyFBGRKaErfSIiIiIiIjGmkz4REREREZEYU3pnJZwudtymcy5YbBdf9pyNL1CG16uv0PNT2iSCtMcXKIWSWy488I82vvE3bXz2ZO24o5O2faF0M0frvpGybt76Phv/8Ku0/GYbO3r84aHS7XXvpf146K9tfPMf2jiqhQOn6CXWQmYQTufkFK+oFE1OawyXc+oZHxt7P23jrB3PWH6UtjXfxmE6JwAcerB0O/krdhm3muDXcWWPjTnVlNM5edu10uzU3kFqCcdFVJuTqPR4Pg45rtV6hFOi620fISIik6IrfSIiIiIiIjGmkz4REREREZEY00mfiIiIiIhIjKmmrxKaZt0ftIu/TV0VqJoMj1ILh8432ngjtTYI2ywcPWAXLT9n465ftvFLVEpx5oSNn37KxpdR54NVwc5voHo/fu6X99v4zV2gFVDTut5V47df+aZ9E+fym8i1UMuvsjHXgWz6sI15qnyuMZHW4ho+rvPhGj6uOQpbHXz5Q7Uf+wTV8G14s42P/Zgeb2v6nvv8gyZe9Ns3lYIRuwxcVrfxJhtzDR8fp1zzd5I+ELh2NcR1WFxHKe0tPD6iavh4PDKuH+WYj+tweyvpi+M0fcFyTV+99YYiIlKRrvSJiIiIiIjEWCxP+pxzg845nqdSRERERESk7cQm/8c5lwSQApAD0A1gYNJPRilVJ6iNwl5afR/F1EUBj1KK5VJqw7BsU/VdWUbZXC9QWwVus7Cow8Y33WjjZyhbbEWNbLHzlFq6iNJSj9FzrbnWxqcesvGyd5dSOufwkbd8qY0z37HxLZTeyWlvT95nY073DNOV1M5h+nHfEk7Z4p8Jp3xxum+Y4sXpYtw24YbfsvHeL9iYVvfZ8yZedAO1cHg0SOmksRy2QAEQ/bo5zZWPY35t4XHMKc9KYZaJimrZUO9nJI/XWi0dOH2T90UtHEREpkRsTvq891kAfQDgnOuJWF1ERERERKQtxDK9U0RERERERPJ00iciIiIiIhJjsUnvbCqqKXiaZlnP0OofoPgIxZ2X2vgYlT+EbRZeoDq67z9g4wUJG3PJ0LkxG3/7uzb+NZpB/ulgX1ZR/d/8NTb29D6seR8919dtvOrD6+wdBw+P3zy03y56U8cZE5+n1hPz7/1be8fNVDn5+GdszPVN4VT3qulrPa4hyqZtzLVuXPcT1vhxGwR+7r1fszGNMSy2oaP40Odtjd/6d4cL6bm4xu/+R2x87VEb8+viGj5+beFxzHWtfFzze6gWDjJRUS0beIxxiwY+FsOWDlzvx+0eomr4VPMncfDMfhuv2NSKvZA209ZX+pxzvc65EefcyKlTp1q9OyKzjsaQSGM0hkQaozEkMjFtfdLnvd/tve/y3nctW7as1bsjMutoDIk0RmNIpDEaQyITo3yfSihV5eRzdvGHaPWzFC+heAGli536uY2vCGZef5raO/yQtv0mSt98+he0Ldr2u95h40WUfrYoGQSUUjl23MaLqR3E+REbr6LMtFceOmziuRtLt9dvQk3zaVugLLiylKDFq23MU/ybHaPUpGR37Z2RxkWlFvLPgFsXsDB9d4RaMCTfaWNKgcZ624Lh/F6bvnmaxsHlNGa+f3/p9jw6LjdSinSOs96yNMBXvs7GnObKaXLhZ1OY6gmUH9cc85jh3HClf0oRHyucUskpmlEp82EqMqc0r6ZeP1HtH7hVCR/njLcnMhMonVNaoK2v9ImIiIiIiMSdTvpERERERERiTCd9IiIiIiIiMRabIg7nXALADgAJAEkA/c65NIBh7326xkPL0XT/S6nlwu6XbEyVNdhA8Rx6l4/R8hXZ4PYqu+ylp23MNURr6LnfTHV1x5+08feohcMvBy0cuKRnyXIbc63TZVSreGxv7X0NawZXUFkGqCzjOdoWt83YmKM6Lp4q/8B3bPz7t5Vuc40HTzce1S5AGsfvMU8Rz60LHv5PNg5rhDpeb5c9Rj972hQO2hq++TTj+/NP2PglKil6Q1AHe8lldtkz1EWB63nLPPmsjd+90cY8Hf3LudJtbkvC7R24zpVxTaC0j6iWDFzTx7iujmtR+fnDmj+uwePWI1zDx/WEvG0W1T5CZCZQywZpgdic9HnvcwBShbCvhbsiIiIiIiIyYyi9U0REREREJMZic6WvqWgK6e9HpHMyysjElS/a+ApafjGYOf1H1M6B5aiFw8a32fixb9m4a4uNl1Ia5MJE6fah/bXXXXZz7X1bQKmkxyhrZ2XYHoJmix87amNuD8H7Ako9xUkbenrP3YG7S8FVH7QLD95r44232Vjpns3HqS2cmhg1DXuY5kjrHolIMzbHIYBHvmTjF2hTlN2NTUFWJac8X9FDK2dteIranCxbQ+s/SqmpN95k43NBH5X9d9llfNx2fsTG3MaEj2N+z8O0OLVzmP3Cn29USwY+FjjN+DgNMn48C1uw8HHJeN845tRRPo55/fC16LNbZgqlc0oL6EqfiIiIiIhIjOmkT0REREREJMZ00iciIiIiIhJjKtQo8r6U+x/WfwGgWdkxh+JlFHPN38/O2vjdv2zjR4I6vLfQ7PPnxmz8vV/YeBGVQnE9Ek8Z/zLVumWDursraCbss1Qn98PP2fitH7Lx+XM2XmNLI3Eq2FaO6v3eRLVQx75uY24f8ejf0LZopvs1m2yMow+XbnP92KYP25inEOep7TUFeOO4TpLbC3CdJVtzXen2V237jjNUZxfWrQLAszQmeMwspgHONYGngrI6fu5n7rfxCurfcpL6tSzj/i78Z7iRB1HVhjfbmKfN5/c4fM+A8uOcl8vsxnV5terZomryws/PSs/FdXTcdiGsk6fvV6y9wcZR7SQYfx7zawnjqM9u1fzJTKT2DtIkutInIiIiIiISYzrpExERERERiTGldxY5V0rtoNQU6i6AMxSvo/h7FL+L4hFqqxBki2ExtWygTDW8ieLnqJ0EZWTimaM25pTNMKXz55R6xqmg6zfRcz9k49fT9POjQzZeFmTwcdrq9++y8WF6H97yRhu/5302/im1i3iO3rhFC4NGGtd32YVD/8bGH/yfqInT5jiuJ62qXS1cYWNOTVxOucb8nt7/1+M3Dz5sFy2idh+vo9nm2TU0QJ+l7LB/+ImNrz9Sus1JcRcpXpmx8RuoXcTw39l46//bxufpuA5TTd3cH9PGcjbmNOarKR+b2zDwcRqmNfPPSy0cZr5aLTl42fPP2JjTNXn90wdrLw9bNABlbZCMJ+iLglNDeV84RZPTQWu1fOD9jEprjWpros92aRZO4QwpnVOaRFf6REREREREYkwnfSIiIiIiIjEWuxwd51wngC4AWQAJADnvfbrmg0RERERERGIqVid9zrkkgB3e+23BfYPOuTHvfabGQ61jj5qQysuQoJgmTsciihe/1sY5arsQ1gTyTr6Vt01PPod+guuoPcSbPmjjw1Q+EU5vzy0bDu238QvUkmEFdTI4+JiNF1Nt1XNBHd8Z2s+lS2x8Nb2pL1OZxyNfsvEqKqw8RiUnbw3K+PzQV8wyd8s77cr3/q6Nf3vYxidGbLySagRr1fi1aw0I1+lwHc7LORvze0w1Qq8E9adz59lVo1o2cJ3rI0/ZmMpH8RzFYekrNZooS534BRX5jVB94PrX2fgQlUJdSuVJYQuWOTm7bNFC+qRaTP1fuNaJ6/S4bUZYE8i1TFxXxTV+r75Se7lMvXrqz3h8cq0b/7y5jQ23YajVdoGfi1uH8GOv+ZiN999l4w3vtfFlNYp46223w++hyHQJ6/hq1ftFPVYkELf0zhSAAbpvJ4D+FuyLiIiIiIhIy8XtpO925NM6Q1kA3S3YFxERERERkZaLTc5NIbUz4b03J33e+5xzDs65zgmneNJ08UfwiIm/RqtfQzFlLuJSTuekrKvQzylDazVnaFGGzgFKTbuKctN+Qlk3b7rNxoeDdDJu2cAtGvY9YON33Gjj5dSyYf5lNr7vG6XbV1Ca6nP0pm3otDG3eOCY8dT4Yd8Nt5yWHfyOjTkL6vHP2JhTmzhtitsNcLpnO+L3iFMLOZ3s8o02pnTPucnSgf96yky7SJmFj36XNkW79gaKL1D86xSHyb687k1vsfFTP7IxdVjB88/a+DTFSxbYOBEcux30Fi58wsYuSR8mX91h4+4/tDH/jMKYUz+jKJ2z9Wqlc3KbhFopkUB0+ianDp+kdiJhWiVve4yaIFG6Nu7/zzbu/ECVnSx49FM2vu5PSrejWsNEtargz6modH2l9rcPTsHkeNOHaz+eUzLrSemMemy9yye7rsx4cbrSl4hY3hGxXEREREREJHbidNJXN+dcr3NuxDk3curUqVbvjsisozEk0hiNIZHGaAyJTExbn/R573d777u8913Lli1r9e6IzDoaQyKN0RgSaYzGkMjExKnwIgcAzrmE9z5XYXlEFViAp4Qm/4HiZyi+nWKu8fsFFwIF5lN8mMpyNqyy8RXU6uDsSRtzDeBz+20czqx+KSXADlPx4m9QKcUPbGcLvJk6F3ybHt8ZtFXguiuuH3z6kI039tj4n6hO6yPvtvGx/TY+8j9Kt7ldxM3/zsbgWbr32xYPWE8tHrgehmv4wphbE6jeLy9qCnh6355+uHSbj6XvHLExDzeq2sGXKf4tiqnLAtYGt7mdy8NUw3eUlr+NYqp8wqUU516wcVfwif3Qt+2yO/6LLQB+IW33bsF19OnC09HXqneqVe9X6bmiWjrI9At/RlzLxm0TuIUD17qtvcHGPF4XLrXx6mvHb1788jfMojl/9F/q25cHqF/P8/QNzK8trCEcoy+W9/+9jY/vtTG/Tv6sj2qLoTq+9sG1blG1b1G1cuHvobd8uvqySs8VVT9YD9XwxUpsrvQVJnDJgWr3nHOJwvKJ9+kTERERERGJidic9BWkAdC8j+gq3C8iIiIiItJ24pZzkwIwCCCcE7qvcP/ErbR5iuupZcNFWp26JOAyakdw4TkbL1hs4xfOBduix3Lq2iWUZZOg9gNz6CfK23qZ5qs/cLh0+19QCuVvrLSxo6m0122ycfg6AOCGf2njz/1j6favrLPLjlBGD3v4L2x8Dbe9oPfpRdqXq4L0T97Psjn8Ob3z5l+rvXOcwlM2fXnw4jj1SPI4zZXTVTiNKnDoBzamZwJlPIM6HeBdFHOjT8oWRZi4RolnZema74xYvplifj5O9/xxkGvKlSunvszJptbPPn/exOvP/bVdgVqs4Jp/W7rNLTaOUW43t3Tgn2dUSwBpXD2phryMWzBwSuWT99VeHuW+UtsF/g5a8Nk/tnck6LHcwuFKqnF48p9t/DjFncEoo+92HH0YNfHyqBYPUT8DkaKoNgphiub9n6i9bq3U0Errc7porfYQSu+MlVid9Hnvs865lHNuO/K/uyUBDCi1U0RERERE2lWsTvoAwHufhtI5RUREREREAMSvpk9EREREREQCsbvS1xRUn/JDWnwdxfsofhfV0XXeaOOnDtr4YtBMola9HwDkuECJcI3f5VSXxy0dTLnaidrPHYWf+yUq+/jXQcuHfQ/YZaep7rFzi425tvEYvYePDtn4Ciq9+PFI6TbXInIN3zP03CuW2ynGsYlq/A7db2Ou+wjrSFTzkcfvw4kRGye7ay5/PvfU+O311Afhn6jGj7oe4HTErnGLFUdxeJhzH5iyGjyKH6d4DsVcVseVcOHzX051rU89YeOu99p42UZ6smMRG99wtHSbW43wMc41fPzzjWoB8CoNcLV4qF/UZ0n4M4n6HOLWBtwWgWs8nxy1MR1LF54s3V7AZc10aFw8buNnjtp41VVP2+fO2eXzrqHWJOGxx+0gNtAg4feFj/Oo47hWK5N2/ayXieG6vQ8/XLrNNXifStg4qkVDVP3gVLZ8kBlFV/pERERERERiTCd9IiIiIiIiMaYcmiLvS6kYL9oEME7nXEPxeoo5FfF89dnmAdhUxGWUTfLtv7fxKtrY4g4bn6N8M96XspYOwYt5hdIzOQ113W8vMHFuyCbOcepp2CYBAFywr5zGyvEFes9+ccHG/Lo4jZXflzNB6up5mjI8Rxk/q2hW7zKPUbrn9b9p42/+qY033lb5NlCeLtQuLR04XYyn9Ocp5Pc/ZcLwOP5flM7JbRCouQC+TzFnPf6MYp6cPkwuu4OWPULxGYopGQzPUkyHOX5Bcdge5nHq0PAWymo7Ri/09fTBNZ9SwXPUqyIx8rVScN1v2YWcBneaPiyifp6X07uulg7Ti9McOZ2af76cxvgopXNuojYKOZuCOe/K0u0ff9GuunoDPZS+R57P2ZhbOuSetPHy521rkouZUpL1nM6jdmVuRcHvwwN/ZOMrqa9R1PuolM72UavtwUSE6Zz8fNyCgdMvG03PrLXvUe0hol63Wj7MKLrSJyIiIiIiEmM66RMREREREYkxnfSJiIiIiIjEmGr6ipwr5d+P2JotKhfDujfb+GdUC3P1h+kB3HaB5m0P69meppnruS7uCNefUQnYIqrTOZyx8Qs5Gy8O9n0RvdB1tO3Dn7c1fKu5mJGcphYQK4PyhkuoRITrOHj5y1RYtYZKgn6J6kIyD9l48y2l29++1y67+nrURvWD2LjOxtm0jVdSUWBYH8M1fFzLxLU0HPPU+LMVvy6u8aL31NPbFh4vH3ijXfa/bfkfqBsIllHMNXw/pZjbKoRDjMqTyp47RzEfalQdCiptxSaKjwS3uQpuny1lwi+o1pHrZM88bON33mbjXPDZkVj4Bbvw2n9rY57in6f0rzWVfSXhOOF6Mmkc1w5zje2Bu228+traz3fUNif5IT38rUFh/Nx5dhl/tv/guzbmj9+f/ZWN577WxstvsPGc8KXysXTvgI03PWzjtfRkXLvK+LhPrC3dprkCyj7zZHaLql2LapNQz/PxMo65BjCqLo/XD5dH1fCpZm9W0ZU+ERERERGRGNNJn4iIiIiISIzFNr3TOTcIYKf3PhO5MjlPmUo/pOWvp1TEV2he9bHHbMxTTm/+bRufCtI9V1DKJKcicron574co/RPTsGcf6OdWnvs3tLU2q9QGupcygBasZaei9oirKJslcM0G/YTe0u319JzL6G01I7VNv4JpcSuudnGL9Dr3tBpYx/MhX/trXaZo22DM894Hv1jh228kqYrP0oT968JcpuOUQOBWqmglczW9M7X0EcNvw5OD9z/oAndlXZxImhHcNdP7LInaNNLKaYkKyyJWE6dEHA8uL2JllGiL2i3y/YtR/GvRTzfrwa3KYMZWyn+Z4rX0vu0NqLFw5pw5/lNefSvbXzDH9o4Kp2TY057Dh//Kif4ET62JK/We8zjjVMPr/qgje//YxNepM9bTs9fmLDxU8H6vC5lhmIZDcjnztqYPyEvoe/fCzTIwnY/C1ZTE5UN9OnA7wO3GmGc9lqr9Yjakkgoqg3DXTeUbt/yabus3nTNE9Qc6Bn6HeXf/8/qj+dts6i01XrbR8iUitW3pXMuCSCF/O9S3QAGaj5AREREREQk5mJ10ue9zwLoAwDnXE/E6iIiIiIiIrGnmj4REREREZEYi9WVvmZ5gWrbbqQagxzVGNz4ARtnqd4hSfVroFmjl20KAipt6bzRxnP4J7bJ1iQsPG5rFuZzvwmq8+gIawQXL7CrPm5bNLxI78tl9DrcmI25FCdss/AcrbviQ7YNwrHP2bq5BNXd8ZTg3OJh3TU2Ph9sj6eu5zYZG6mcBdR6AlwDuJgKELl240SwAZ4anWvbuA6Hp1bn5Vw7NVu8nLMxvy9dv2HjB75mwv1BfdomemqaER7UwQE0w3tZnd1aiqkji9legpbR8AWV95a1WeDhTF0W8HqKw0osbmQwSvEmiq+k8qVL6QmW0mfFqWOl28v4mKcfF+77Cxtfd9TGPJU9//x5+vqwDpZr9qJq/NpF1GcB10mG7Qq4Fu3RT9m4rCbIhlyXt5RqqPdRW4U3BW2O+LP8JW7HQt+vjA89rsk9ddzGYX3hGeq4sOgk1fiRedc8bO/gz2OuAeQ4/C7g9g/cPmK2fpZLXq22B5XiKEEt3dgf/K5Z1PHfqAbvD95B26LnWvk6G3fWuS+h+z9ReznX8DX7fZGG6EqfiIiIiIhIjLX1SZ9zrtc5N+KcGzl16lSrd0dk1tEYEmmMxpBIYzSGRCamrdM7vfe7AewGgK6uLl+8v+N6u96ub9n4D99lY0etC9bxFDKr327jzPdsHD5+tU2xnJexKZZlDtj0lA5KBwWl4WCM0lnCDKGTdltzF9tVTRoqgFf223guZV2tp/XdZaXb5yi9E/ttOmeY3QUAP/yqjRfRe16GUtXC2el/vt8u20htMF6g5XNplJSlzB79jo355x2mYXF60AnKLeWUn6ip7jklaJpbOlQbQ2U4JY9jTv87STnSlBJ24y+Xbn+Fxiendy6jmCavLkupTFDMaZRhiiWvSz95XEHxT1Hbzynm1xK2dOB0zkUU7+flNPQvz9XeFxf8SfASyoFdlKSV+U2KatnA6bz88w+nyk+srf1YTq+eZSY8hlhUOiCnmYfvKX/ucFoit5J5/hsmfJkaIT1F/UOW2K8xvBykcGaetsvoawbcY4m6DpWlSP9vivFdG4aJbYsoF3Tz79n4wpP0XAfO2+Uv2lE37720Pn++12rTwO0geF2lf07YpMdQM0WlLUa1MqB47Ho3frvj937Frkspk4foqdffUnkXq26b9y0TtHRYYbdVtm4U3hanh9ZqCRHVmkLqNutO+pxzo6Dftbz36yqvLSIiIiIi0t5m3Umf935zq/dBRERERERktmjrmj4REREREZG4m3VX+qbFeps//U48YmKe8r9sCv/LqKDhONXwvZvysw/a5w+5tTZ+br+NF3E7iA7KdL0sZ2OurekI6gSWR9SLUTz3g1T38fWvmNBxMUZQx7fqVrvo/H4bz6dCD55OfgVtGtfY9/SFu+17uiBYfw29rDGaxruDZo/HVfSe8nTzXCf5ffp5Xx1Uep2j+cT558G1Nfwz4MdzHUi4b9Nc31cXfg855toYquN67bxSrU2CnprLcjimzgV4H8X/D8UXKA63t5aWcc0eb5tr9Li+kMqFkaP408Htt9Iy/hjiaqI5FO+/aOMbqczy50Ht1Wmasv/dCRu7F+nJn/+Cja+lFhxctMtxWNPJNXzc3qFWawKpjN8zruHjz6HHbU3fKmrRcPjR2ps7EhxLPEbOU8x1sGyAYi6xps5CWB3U8X2Z2kF8lTqNbKV+LlfQxxC3qnjrSqqs5c/raz9Rus3vKX/GPXmfjbnFg2r6ZheuR+PWBVzr9qmECTs21Xguqm3roJLoL1DLlDH6pvl4xraAKKsJ/L2gEvZ//KVd+N+oaDaqhQOrVcPHVMPXdLE66XPOJQDsQP73siSAfudcGsCw9z7dwl0TERERERFpiVid9HnvcwBShbCvhbsiIiIiIiIyI8TqpK9pNt5mwqvfbFMF5/C7lqX4esprnEspefspnfP63yrd5hSQhYdMuGh9xDTOnP5Hr6Us5W8seH5Oizp0v41P0lzbJyjusmmQrzxk2zDM3RQutA+dv8bGvJyv8p8/Rqtnq6dzAsArQbbYXMoH6qAZ389TKtv8xfZ14OY/tPHdlCPEeXZhis+xiDwoTmXj9M+otMgOzqmt8VzTmS4U1XqCp+XncfCiTQI7EYy5d22xq57ZZ+NLaVeowwNeovgNEcvD9DROTZtP8ZUUc+IhdUIAZ0lS4rBpy7CflvF+5yjmLie8LxdpzIUF35fRys/yx1TCxnM5j/Xg12zMP39OdXtN8CHLqYhRxzy3A3mNvuYA2LYY3C6AP/sP3G3Ci/S1wulgnPb4Teo0FH6qvYV26x8o5hYq1GUBv0QxZSmDjhbkgpROKryIlKDP8lUb+MkpPvrPNg7f8w3U3yHzORsnu23Mn4Gcyh+1XKZXVCoipzX+gTMh/7oVtl14/M+5EMD+vvOd5+zSX3+TjTn9M7LFQ6a0vf88aBf9x84bqq4LoPx94LTWqNYWYbpoPamgMiGayEVERERERCTGdNInIiIiIiISYzrpExERERERiTEVO1Sy0CZAr/t9mrK/rPaNpujnujqunbrxX9s4nJ6c8/qjasB4qm02ZmsCa9Z8sZNUUbTp12zMtTWPf8eEc3+PXudjf1u6vZiqK16hIpCV9J6fsHV18zeusvFBqi9c+067L9lg31ZzZYfd9vzr324Xc00R/0y4ho9L5f4+mB751k/aZeHPHij/eVJtTdnU9iys2eSfDx+HXFc1lXjMcH3Lw//JxryvVDd7afB0+6mGj6eE5xoi9jcUv5tifr6whcPraNmbKaZSp7I6Oq75+xnFXEL07eA2HwmeYq51olJV8CfBt39u47C1xeZNdllZTQgtX8QFhFz3fFXCxns/bePwM5iPea57jWgtUxa3S0uHWuObP9MizLnWNjpZeNR+N/zsiF3/YXp8OE744/EnFFNFZtn4pfKlsvW5si2s6Y1qD/H8L2x8mr7KF9CAvHCUnmC/Deed+FIpWH0tauK6ZsZ1mPR7isxupk0COf2crZu79pfpsbT+GH1EdvyebWm193dtTSDX/IXxf/w4PTnX8P3ev7NxVGsKxu0oZErpSp+IiIiIiEiM6aRPREREREQkxpTeWQlPfbyY5vTntgfLKQeErb3BxpxuFKZpcDrfBUp7XEGph5zCd+7p2jG3cAjTC3m/uDUBT6t+7+/a+L2UzslTSq8PUi459WgjPfdjX7DxylUUUxokv05Oa+WUTbOflJ77DMVrbWpE2b5zvtI8Sh9dGPwMOY2N03k5/ZbncWac6hY+no8NPm6ns2VDVDoZjzFO+aBdDaeIz9FTcbeACxExZ+dy+tgNFH89uP1aWsbTzd9B8WMUH6X4Noo5tfRXg9v/h5bdSjFla+JtFHP6J6fJhe0jPv+gXfbLr7fxEcpSvppz7LroAZwiPUZ7e0Pw2XP6oF12ZQ9qikrvbBf8usPPY04N5M+R/T824Qv0FcetC57P2biX0j0zwW1ORaOOK6CH4h0Uczo2FzicpfhUcJuOpLJj/vMU/+4PbXxrwsZL6WPr6SdtvOq9m0vBo5+yCy/faGNe/v6/tzGXArCo9FCZWvydxa0IeHmn/b1i7H/YlMu9QW+hWynF8vG7bMzpmR+nnOmPUzonVURgH/Ux+o/bSre//ld2GaeW7v31vzQx7ytW7Lcxp3vy+xIuD9s3VHpslKg2Gm1IV/pERERERERiTCd9IiIiIiIiMaaTPhERERERkRibdE2fc+4jyE/EnfXeH23aHjXIOdcDoAPAOgCdAAa890MNPSnXWXGtxK2fsfGJkdoxC2sI199il3EeP9dl8fTjvJz3vVY7Ca6D49pGrmXr3mljfjzXlIV1ePy6uHaN20Pwtvm5iX/cTinubgu2/dg37MpXU71f1Os4TjV/qyNaPIT458Gvi38+XOvG63Mdx8F7S7ev+qBdxnU7fKxMJT5O+T3i18Hx+vkmXLqyNPk6tz2YS+VhVK5QVsu2mGJqVAIqZ8MTwW2uN9pMMR1pZVPG/wHFdORhDsVh5er1qO1dFPP7wKjqDmHTlM1UUnuASmi5jnJdzsYLRuiHwi0duLAynJ6eP4eiapt4jPGxxMfiq1zZFXhNjMreucY6xO0A6C1fQIWwF2n4ck0f17aGNX38bfgExfyO76b49yl+iGL6lDO4nvBHFPN4Xknx//lnG3NNLzcDWXj36Pjt191i216U/V7AbWwep98romr2wuM66vNVmi+qho/d8mkTdmyy62+4vTRnwpuoru4n22z8V4M25vpu/h6i0lNwM5EvBM/3zjfZZf+Zvkg+Tsv/ivb14x+39YR45gYb0/tg6vh4GVPNXt0a+Ub7EwDbAaxzziW99/zZO+0KJ3wZ7322ECcAjBb2b1dLd05ERERERKQFGjnpG/Lef7lpe9IcHcUTPgDw3ueccykAgwB00iciIiIiIm2nkZO+Grki0885lwQw4Jwb8d6H2SSZwvJOur86bgfAaRg83TWn3HFaHacnHX3YxmFKJy/j9D5OLeUp/3naZ26zwKmLtab4X3Odjfl94FRETh/ix4evhVMLuz5qY043qvWeVeA6OZEnsHGdjTn1hX+++yk5ous3bcwtOR77rza+5mOl25x28+R9NuafN/+8+D2vlUIbpnoCke/ZtOIxw8c1T81839+aMBGkA/7gu/TUtCmaXb4sXfPXKeZ0s5MU/3+C2/fQshMU76c4QTGn3byF4mcpXoTqvkkxp1zeRDG3ruD2EGFSzgfosF1ts22Ro7zVF8/ZmDMoL+UfP7d4CD9L+HOFxwx/zjA+1jjmz+da6Z6zWfg6+TMvm7Zx5wdsnPuSCUfoz72X08fWQu6bEOBP5tdRzL9YcLomN0jiX2L+H4rDo4NTuemTviyd+qsUc/omp4O+RPG3gr4q701S4vg1lAzO6ff8M+Hvev4Zht+R/J0krbf/Lhtz+ufXbRrkk0HbhS/QB/8heuh/ok1dQzE3fuLseo7Hwuem9g+8rS/Qcn4urLQjfOzr9K32ddv6y7Sf+DqNsC96fvbaotpotKFGJnJZGr3K9Clc4duFfJ1hKFH4n+8XERERERGJvUZO+vqccz9xzt3jnPuIc25ttRWdczc2sJ0J896nvPc5uvsO5Ov8+H4REREREZHYc97Xebm0+EDnvghgGMBWAN3IZ2nkAKQB7AOQ9t7vL6z7R977/1r5maZOYSKXIwA2h7V+wfJeAL2F8CoAEdPCNdXlmL4U2encVty3F+fXBgBv9N4vm+jKGkPa3gzfViu2pzHU+m1N9/bi/Nqme3t1jR+gpWMozj+H6d5enF/bdG+v6hhq5KTvj733/yWIr0C+RULxJDAJwCN/Epj03r+p4hNNIefcMIB+7316AuuOeO+7otZrluncXpxf23RvL86vrVFxfm/i/Nqme3txfm2NivN7o9em7U0H/dxn5/bi/Npasb1qGpnIxdRreu+PIH9V7UsA4Jx7HYAuADcX/m8K59woaD4E7/26Cuv1Y4InfCIiIiIiInHVyElfN4Ad1RZ6759FfqK8B51zk7ucWPl5uf9xmcKl/mGd8ImIiIiISLtrZCKXhwqTuGyawLpnoldpDudcN4CR8ITPOZcstHSoZffU7llLtxfn1zbd24vza2tUnN+bOL+26d5enF9bo+L83ui1aXvTQT/32bm9OL+2VmyvoknX9I0/gXPvAHDFTGjUXjjhS6DQmy/Q571PTf8eiYiIiIiItFbdJ33OuTsBbAcw6L3/97TsAwBGvfdHm7aHE9+vBIBqLWGzler+RERERERE4m4yJ32fQv5qWhfPRFOYvOV25Gfu3Om9P9ek/RQREREREZFJmNRELt77j1a5/1kAewDAObcTNSZ6ERERERERkak3mYlc7nHOvZ/vdM4tds7d6Jx7o3NuMYAlje+eiIiIiIiINKLukz7v/XcBPBue+DnnbgKQAzAMIIt8bd2hJu2jiIiIiIiITFLVmj7n3D0AHgfwoPd+f4Xl7wCw2Xv/OefcCIB7AKSRb9qeLTRrFxERERERkRaqddI3hvyELcUVMshfyUt77x8qrPMOAFcA+KD3/vYp31sRERERERGpS62TvkMAtgJYV/j/JgCdsCeBaQDdAAa895+b8r0VERERERGRutQ66bvTe7+nwv3dsCeBQP5EsOxKoIiIiIiIiLRW3X36yp4gP4nLzSg/CRzw3v9+Y7snIiIiIiIijWj4pK/sCUtXAl9XrZ+fiIiIiIiITI+mn/SJiIiIiIjIzDGZ5uwiIiIiIiIyS+ikT0REREREJMZ00iciIiIiIhJjOukTERERERGJsbpP+pxzVzjn7qlj/fc7595f73ZERERERESkcZO50tcJ4FNRKznnXuecOwOgD8DNzrlvTGJbdXHOJZ1zo865Xudcp3Ouxzk34JzrjH60iIiIiIhI/NTdssE5dwXyPfj2O+duBADv/UMV1vvvhef/aCG+qfC4Lze+21X3LQlgGEAHgASANICU9z4zVdsUERERERGZyeq+0ue9PwJgq3PuEPInVWnn3Bnn3Ntp1W4Anw0e9yCAZCM7O0HbvPdLvPfOe79VJ3wiIiIiItLOJjuRyxYAKQBLCv/uALCL1kkCyNJ9z05yeyIiIiIiIjIJc+t9QCG98x7v/ZeCu4tX+24MUz299+fo4fXlkk6Scy6B/ElnznvPJ54iIiIiIiJtYzJX+hIADle4/zDyV/1a7Q4AXchfZUw45wYLJ4EiIiIiIiJtp+4rfd777zrn7nHOfcR7/1ywqB/5lM/i1UBX4eGJSe3lxPct65zb6b3PFe7KOOeGAewBsI3Xd871AugFgAULFmzeuHFj4YletSteeN7G55424cvPvGji186zq89Z+yZ7x/zFEa9EpDVGR0dPe++XTXT9qmNIpE1pDIlMXr3jB9AYEgnVGkN1z94JAM65HgBfBDBauCsJ4EHkJ25ZAmAHgBEADxRn6yzM9LnOe7+n7g02oDCj52EAS4KTwTJdXV1+ZGQkH/BJ3rFHbfzQfzDhjz81auLXr7GrJz73T/aO9bdE7LVIazjnRr33XZN5rBlDIm1KY0hk8hoZP4DGkEitMVT3lT4A8N4POee6ANxeuOtPvPcPFq7wLUF+Bs0jzrlPOee2AsgB6PTe/9pktjdRzrle7/1uunus8H8SgGbyFBERERGRtjKpkz4AKLRCyNB9RwAcCeI/cc59AECH937HpPdyAgpX9Aacc+kqk7doQhcREREREWk7kz7pK3LOLUZppsyjvJxm+ZwyhXq+vgonfLcDyNRK7Szzyss2PnnAxs/sN+ETR+zi36f4c+/7dRNfsfendoXFqye8ayIiIiIiIvWYbJ8+OOc+4pwbA3AW+dq+w4W2Df+6aXtXv7HCFT8A460b+gDc2bI9EhERERERaaFJXelzzn0R+at7Kdi0yc0A/o1z7vaprt+rpFBr2FOYaAYA1iFfX6jUThERERERaUuTac7+EZQ3Zy96EMAu59wfF1o6fK7hPayT936o6U96ScLGYxdN+AVa/ZsU/9mPbPy5u26wd3xkb+n2ZZfXuXMiIiIiIiLVTSa9c0lUnZ73/r9gZjRqFxERERERaWuTOek72+T1REREREREZIpM5qQv0eT1REREREREZIpMZiKXZ51zN3rvH6q2gnPu/QDc5HerxbiGj1o0/PNn7eJ99PBXKOaO8N//88Mmvvqy95aC307blectrLKTIiIiIiIi0eq+0ue93wPgo865e5xz/8I5t8k5t7bw/0ecc98A0Feo6xOZ1bLZLHbt2tXq3RBpO7lcDtu2bcOSJUuQzWoCZpHJ0HeYSGPiNIYm1afPe3878r35vlT4/zDyF7R2I98IfdrbNYhMhXQ6jYGBgVbvhkjbSSQSGBwcbPVuiDTVVP0Bo9rz6jtM4qCVf/iL0xiaVJ8+APDe70K+PcMVADoBZL33323anrXS88/YeP9dJhw8bxefoIdzXis3Yfjkz228/aPfGb/9y13UcWLTh6vspEyH3t5e9Pb2tno3RNpWR0dHq3dBpGmGhoawffv2aXtefYdJHEzVuJmIOI2hSV3pC3nvj3jvv8QnfM65tY0+t4iIiEhcDA8Pz6rnFZkJdHw3R8MnfTX0T+Fzi4iIiMwaQ0NDSKfT0SvOkOcVmQl0fDdP1fRO59x/B9BIXk13A48VaUgmk0EqlcLIyAiOHDky/oGxb98+bNmyBT09PWa9bDaL4eFhpNNpDA8Po6+vD8lkEn19fUin0/Dem/XT6TTOnj07/rzDw8PYunUrenp6MDSUT9Et5qBzSkJx+djYGA4fPoz+/tLfR2rtTzabRX9/P8bGxrBnzx7zGm666SYkk0ns2bMHnZ2dU/W2yiw0NDSERCIBID85yvDwMPr7+8fv271793gK5b59+8aP/aJMJoN0Oo1kMomxsTF0dXWZYyydTiObzaKjowNjY2NIJpPo7u4ef2yl8cLjsNK2AJj94NcEaAzJ7DE0NIR9+/JzfadSKQDAunXrTNpYtbFY65jO5XJVnzebzZZ9h4X7A2gMSf0ymQwGBgawefNmAEB3dzcymQw6OzvN8Roed7lcDnfeeSfS6TQefPBBc3zUOhanatx0dHRM6LspdmPIe1/xH4AxAA8A+Owk/52p9twz8d/mzZv9uGd/av49CJh/19G/t9C/t9O/99K/q2r883+xyv777v+0/2TCzp496wH4wcFBc393d7cfGBgw6yUSifH1+vv7zWPyw8Sq9LwA/PDwsLkvmUz6w4cPj8fDw8MmHhwc9N3d3WX7XW1/hoeHfTKZLNuf/v7+Cu9A/QCM+GaMoSgXf2H/1btcJmR4eLjsmOzt7fVnz5713uePGz52ksnk+PLh4eGy43NwcHB8+eDgoN++fbtZ3t/fb8aX99XHC+8rb2t4eNgD0BiS2Kj0feJ99FiczPdUtWUzeQw1Mn68xtCUGx4e9p2dnea+wcFB39PTY44p7ysfk8lk0o+OjprnizoWqz2X980ZN1HfTZXum61jqNZJ3wPVlk3kH4AvNvL4ZvxDvkH8wETW1UlfPFUavKOjox7A+IdCcb3wgyjqOSqtn0wmy37Z7e7uNh8ovb29vre3d1LPX5RIJMqW8YfWZOmkL14qfSEODw/7s2fPjv9RhPX29o4fx8lksuIfMorHWyKRMOOoiO+vNl7C+zgOnyv8ctUYktms0rE6kbFYfGw931PVls3kMaSTvpkt/PwP8R/nivdVenx43NRzLLJmjZuo76ZK25+tY6jW7J19ta4QTsCdDT6+Gfa0egdk5ileMh8ZGRlPQwvvn6hKqWddXV0mTiQSGBsbG49TqRRyuVzZOrlcbjzdLmp/duzYgYGBgfEphNPptHkdIkU9PT3YvHkzhoeHsW3bNnR3d48fK8W0T66VGBsbw+joKLLZLLLZbNmxNTAwgO7ubqTTaXR0dJQdt0B+bKTTaZO+WS1VE8D4tiYyBjWGJG7S6XTNsRhqRsqXxpBMRrXvBAAVvwcmop5jkTVr3NT6bqpmto6hqid93vsjjTyx9/7ZRh7fKOdcN4AsgPp/muQpiudR/DaKr6SYWzokKH5rcHv4D582y7b+2cfsymuus3HHekj9kslkw31fJvMhV8wz37VrFxKJBJLJZNkHR5Te3l5cccUV4x8UE/lwnPFeednGr5l0NxkJJJNJHDlyBDt37sTAwAD6+vrQ29uLgYEBjI2NoaOjo+xLJjwprKS4PJvN1jzueHzVWjeTyUzg1eRpDEncRI3FZtMYkskofk5X+jlPtrVOI8dis8ZNO/0uF8vfrJxzicLNw2jCSZ/ETzabndRfdxq1a9cuDA8Pm+mH6x3kiUQCXV1d2L17N26//XZ90UpVxb8cFgvMc7kctm3bhqGhofGJWaqJGh9RX3L1jK961tUYkrjIZDJIJpORY3Gyz1vtuNYYksmYit+ZJnMsTtW4qcdsHUNT2bKhlW733mt+V6kok8kgkUi0JBUllUqN/1WnKJfLYWxsDJlMZsJ/Kerr68PAwIBSaqSm4eFhc8UtkUigv78f+/btM1fsWHEmtkQiUfEqXCaTQVdXF8bGxsqO2Vwuh0wmU9dxWdxWpX3h59cYkrgoXi2PGouTfd5qNIZkMoqf05XaJ0w0c4rXm8yxOFXjph6zdQzF7qSvkNb5xVbvh8wcnKaWSqWwZ0+p3LPeS/KTfVxx/TANovgXq+KHRTEnPEpPTw+y2WzL/sols0c4jTSQ/4LcunUrAGDPnj3j02AXFeskqi0PU3z6+/uxc+dOs3znzp2mJURR1HFdbV8qPYfGkMxW1UoLosZi1DE90ZIFjSFpxJ49e8q+U6qVAnR2dpqTr7AdDzDxYxGYunEz0XUqrT8bx1Cs0jsLaZ0J731uguv3AugFgDVr1pQWnDtu1ltIj7uJ4kso5vVfG7EfB4Pb1/LCsRdsvP8uG19p+1xhxaaIrbWfzs5OpNPp8X5GqVTK9BEr/tLa19c3PtkFgPFeKsVlxb4qxfvuvPNO7NixA8lkEjt37kQ2m0UqlRp//mIPmGIPs56eHgwODiKVSmHr1q3j6TcDAwO45557sGXLlpr7w3p7e3H77bdP6XsXpeoYisI1exy/+srE47k8AqVo3bp16OrqGp+0pfglVDymenp6kEgkkEqlsGXLlvHahOIXMy8H8l++xcL03t5epNNp7Nq1azzVhnscTXS8FLe1a9cudHZ2IpfLjf9lua+vb3w9jSGZzQYHBzEwMIAtW7aYX4BrjcWJHNOVnrfSd1gikdAYkkmr9J1Q/Jxme/bswcDAwPgfGYvH2p133on+/v4JfZ4XNXvcTPS7KZlMxmoMufzsnvHgnOv13u8OYwBbvffboh7b1dXlR0ZG8sHxvWbZ4C+9y8Q/ocdGnfSdjtj2k8HtD9OyG/6Q7rj1kzbWSV9NzjnE6RgPDQ0NlTW3bpRzbtR73xW9ZjkzhhrFJ3m1luukTyYp1mNIZBo0eww1Mn4AjaFWWbJkCUZHR1syV8JsN51jKDbpnc65TgCq45O2oJnSRBqjMSTSGI0hkcZM9xiKU3pnEsAW51x4XzeAhHOuH8A+733lxOMIK+fb+PR5G181p/bjv3XRxm+k5ZcHtxe+jhZymu/IZ228cIWNl19l4zae+n6ytXoz1bZt27B169bxlLpmX6GYVi/nbHxJovb6UVf+RCYgVmNIpAU0hkQa08oxFJsrfd77Ie99KvwH4B4AmUI8qRM+mZ0ymQzuvPNOAPl86kZ78s0EfX196OjowNDQUFMa9Iq0G40hkcZoDEkom82ir68PuVwOqVRqymfNjINWjqG4XwZa2uodkNbo7OzE4OBgq3ejqTQltkhjNIZEGqMxJKFkMomBgYGy9gVSXSvHUGyu9IWcc8lCSmcPgG7nXH+h5k9ERERERKStxPJKn/c+CyBV+New93zExpd/3cYLEzY+/qSNb6Z3eQ7FPzhTuv2OG+2yizl67IEz9o5LPm3jlTRhz+qyJhAircc1fFE1exeen/hzR7WDEBEREWkzsbzSJyIiIiIiInk66RMREREREYkx5T1VctnlNqbwzZQxeZImhrzqOhvvf9jG73kfbS7oI/rdh+yyrvfTvq2l/hC5ozY+fdDG3LhazdtlJuB0Tk7B5JYOnN45b2Hp9skDtbfFKc8iIiIibUZX+kRERERERGJMJ30iIiIiIiIxppM+kWkWh0bxIq2kMSTSGI0hkcbMxjGkmr5KuH6IZ4unMrnl71tg7zjwggnf89v0+Jdt+OZ/Vbr9NLWDwEqKL99Iz5Wz8WP/1cbXfsLGi1fbOJw6X1PbT4uhoSFs37691bsxs3CNH8e1avrC2wCwcIWNzx23Mdfsct2rzHgaQyKN0RgSacxsHEO60icyzYaHh1u9CyKzmsaQSGM0hkQaMxvHkE76RKbR0NAQ0ul0q3dDZNbSGBJpjMaQSGNm6xhSPl8lnP71vv/bxvvvsvErlK+5idLFwhRKAHj+GRsfLaWDrrqV9iVnw/P7f2ji+R20/iaKH/5PNub0znDfN94GmTpDQ0PYt28fACCVSgEA1q1bh97e3vF10uk0stksOjo6MDY2hmQyie7ubgBAJpNBKpXCyMgIjhw5Mv6Bs2/fPmzZsgU9PT3T/IrqUG+LBsYpm+Fxy+mbnO7JbUzWUE+VqH2TGaOtx5BIE2gMiTRmNo8h572fsiefTbq6uvzISKFhHv8SySd5USd9HNdx0ofltGM5G54/auPyk75VqOl9n7OxTvqmnXMOlcZd8YOkv79//L5du3YhkUiMf5jkcjksWbIEg4OD5oNh69at2LZtm/nQmeS+jXrvJ9XYzowhVu9JH6/PwuOWT+q47jXqpI/3RSd9M15bjiGRJpqpY6iR8QNoDMn0mY1jKHYnfc65HgDJQrgOwKj3fnfU43TSB530TZNqHxRLlizBkSNHkEgkat5f6fGZTAabN2/G2bNnyx5f577ppE8nfTNeW44hkSaaqWNIJ30yW8zGMRSrmr7CCV/Ge7+r8K8PQJ9zbnZNryNtJ51Oo6Ojo+IgTyaTkbnjnZ2dAAB92Um70hgSaYzGkEhjZvoYituftHcASANIBfelAfQB2DXhZ+H6ITZ22MZdfTY+QT8svvK3kk7ALzlUfVnWHiDzz9l2EOALIieetvFVv2LjsUM2DusXXzxdfZlMqWw2W/OvOhPpB5NMJmdu3xi+esZX8qLaJuSO2ji8Gs/j9eSB+vaFrwQuv6r242VGiv0YEpliGkMijZnpYyhWV/qQP9m7h+5LANAnkMxImUwGuVwOyWQSuVyu6nrJZLLqsqJsNjuh9UTiRGNIpDEaQyKNmS1jKFYnfd77tPc+U4ydcwkAt8Ne+ROZMYp/Ferq6sLY2FjZh0Uul0Mmkxmf9amaTCaDRCIRuZ5I3GgMiTRGY0ikMbNlDMUtvROAmcxlKYDN3vvGrvRx6tk1/9bGnBZ5zcdsfPBeG294r41HPlu6zZPIvEjpnBsX2JjSPc9TZtv8E4/YO45SvOl3Src5TY5TTXkqfJmU4qX78C85iUQC/f392Llzp5nxqRhzusDQ0JCZ8SmVSmHPnj1Tvu9NU+9kKZwOGrYe4fRMXnf1tTbm9Gs+rnkMcpsTaTmNIZHGaAyJNGY2jqHYzd4ZKpz8bQWQ8t7nKizvBdALAGvWrNn81FNP5RfwL417P21jri/ik771t9i4npM+dmzUxpfVedK3mJ5vA8XhSd+mD9tlOumbEplMBvfccw+2bNmCZDI5XrgL5IuAM5kMkskkxsbG0NHRUdazxTmHw4cPI5vNIpfLYd++fdi6dWtT/jJU78xpVcdQvaJm6+QTu/Cksd6Tvqi+fjzbrk76ZhyNIZHGzNQxNJnZOzWGpBVm4xiK9UkfABRm7rzDe7+51npmml+d9OXppG9GqjZNcJOeuzXTzeukT6ZRLMeQyDSaqjGklg3SLloxhmJV01dFGkCnc64zck0REREREZGYiU1Nn3MuCWAU5TV8ucL/SQAZflxFUfVG3PyZrwTwlYSO9TZ+Zn/15dykuoOarY9QSwba1fk023xur40T/NJWBpcGD9xdfb+A8tpGNbGedrVmhZpV+GrchedrL+e2J2FdXlSLlaMP117OV7QZX0kMt8dXBWXGi80YEmkRjSGRxrRqDMXpSl8CwAiAMbq/WGE5sRM+kRkqk8ngzjvvBAD09fWpF5JInTSGRBqjMSTSmFaOodhcqvHeZ5xzwxUWpQDsbngGT5EW6+zsxODgYKt3Q2TW0hgSaYzGkEhjWjmGYnPSBwDe+13OuV7n3DoAZwCsAzDsvd9V1xNxahmncHEKJrvschvzRBHHKedyeZCTeZJmYuGUyg7aFl3X5IlcEpts/N2/t/E7FgcTxXCaG09tH9XSQemeMlH1HivLr6q+LJu28YpNNn7+GRtzOjanf/KY44lgwvHPY1tjQOKIvxN1nIuIzDqx++T23u9u9T6IiIiIiIjMFHGq6RMRERERERGikz4REREREZEYi116Z1OUtU2g1gVP3mdjrgGKejzXJ4WP5+fimOvo0veYcD7IizZ8x820/EJwm1tJ8DT5V33Qxvw6ufZRdR9SDdcIcW0ct3DgurzE2tJtHl/8WI65BpCXb/qwjbm2de0Npdv8Ojjm8SsyE9Rbo6fPchGRWU9X+kRERERERGJMJ30iIiIiIiIxppyNSqLSNTnFknGq2ounbczpaOeOl24fuNsu41RQbgfB2WP0E71wzMaH9tv4zcFLmfPAd+xCTgXdf5eNeep7pnRPKYpKg4xSK93z8o12Gacpc6sRXs7jldM/xw7ZOPx84G3zMc/b1hiQmSDqOKw3/VMtHUREZjxd6RMREREREYkxnfSJiIiIiIjEmE76REREREREYkyJ983A9Ubc6oBr33gK+LDGj9sicH0R1xcuX0D78oIJn37ILr6UypfOnCjdXrrSLpvzENX43ZqwMdc6cf1hvXVb0r6iaoC4Nu7kgdLt0wftsieGbBy2WABsuwcAOL7XxjzG+DgP6275uVdfi5qi6mBFpkOjNXj8nafWJCIiM56u9ImIiIiIiMRY7K70Oed6AHQAWAegE8CA936o9qNERERERETiKVYnfYUTvoz3PluIEwBGnXNJ7/2uST/xsUdtzGmMHHO6Gaemcfpn+Pyc+pnstjGnnh2gFEx66iXLbbyow8ZzwqwcSv08f8LG8x/6hr2DW0+EKXcAcN2fQARAdPoYp4vxcV4rhZpbLvBxya1Gjv7QxqvX2ZjH4H6bMo3rgn3hsc0tHBi3XFFanMwGnA7Kxy2Pb7VwEBGZceKW3tlRPOEDAO99DkAKQH/L9khERERERKSFYnPS55xLAhhwznXSokxhOd8vIiIiIiISe7E56Stc4dsFIEuLEoX/+X4REREREZHYi1Wivfc+VeHuO5Cv88tN+Im4XmHDe23MU8JzDdCKTTbm+qRLEtXX5xoh9s2/tjH/BFfOMeFlJy6aeA7V+OHF4DYte/b7Nl7Oj+VaqSt7bPzkfTYOX9vKLrtMNR/tLapG6PlnbBzWxj2zv/Zzc1uFF0/b+ORhG69+u40v/57dlftK6y98N41H/izo/lTtbXM9In82iEyFqBq8etfnmGtww+Ncn/UiIi0Rmyt9lRQmcukFsK3K8l7n3IhzbuTUqVPTum8icaAxJNIYjSGRxmgMiUxMrE/6AAwC2BZO7hLy3u/23nd577uWLVs2zbsmMvtpDIk0RmNIpDEaQyITE9s8C+dcP4B+73267gdzSwWOORXt3HEbczpnVBxOMX/gbruM20EsXlBzX3JfP2PixEq7+nna1VeCrJwz++2yC/SyF1ALh2e/b6eyX3l1xPsSpq7xe8o47U3iJSqdjHE68MF7S7c5JXrskI2PPmzjEz834UvUYeXSYzadk/80NqfWp2ZirY2z9PGTO2pjfl1K75RWqLelCuPvRD6Ow+88Xqb2DiIi0yKWV/qcc70Ahid1wiciIiIiIhIjsfuTmnOuG8CI9z4T3JcExmf4FBERERERaRuxutJXOOFLAMg555LFfwD6dMInIiIiIiLtKDZX+gozdQ5XWZwFUKmdQ2VcbxbWD1XCdTlRNYFcpxe2PuBlXP/3oq2jw/M2TnzonXb5N79jwvnX2JYOF75Zaumw5mb70GMP2PjAozZ+54dtjC8P2LjrrTbeeBuq4tcdNYW/zG788+TjPGzJAJTX6YUxt3PI/K2NN/2OjQ/YFg2XLqZ94+F7wcaXrg02/ZhdtvDEPfaOa2g8RtUf8vvC66vmT6YCj78ofJzycbx49eSfS0REpkRsPm0Lffhcq/dDRERERERkJolVeqeIiIiIiIhYsbnSN6U49eUY5Tkmu20ctmCotD6nwoTpobwupzlyp4OVS23M6WCUunZs6KKJX7+mdPs8TV2/ZiPF19K2edb99dRO4jQ9YTh9Pb8ujjk9iF+XUoJmF56WnaeA5xYdL56uHYfHA4+ntb9i4/1/Z8KXqJPIpRts/NwBG58bs/HPHi7d5jHClcOLLrPp1ThJT37rZ1CXWlPfi0wWf55yyjTj9Gv+zuMxWetYjWrfos96EZGm0JU+ERERERGRGNNJn4iIiIiISIzppE9ERERERCTGlCxfSVTLBa534Nq1J4ZsfMuna68f1rNx7QNve7FtuYATZ0x4ce9XTDznls0mXjM2imq4zApUZlf2J4Jz9Pi9tn3EHDq65qwMAq5d5NfNNSL8PnANmMxsUXU5XNPJY+zEiI3DOiA+Vh6gtgkrbThCrUgu/pONN1FJ4It0nD/7bOn2M0ftsqW0rWP0UfDWD1PLlSfvo41R7eLaG2wctjbhWijVPkkttWrn+POUa6h5+TkqjOXvNK6DD5+Pxza36+Hl9bR/EBGRqnSlT0REREREJMZ00iciIiIiIhJjygeqhFPNOJXlgm17gBWbbLzmOhtzChc/fzi9NS8L2xxU0mHTPeckaN+esOmcFylL8oePlW5zatoqmpUbL1JM2UL+VRvPuZrWPx5MX794lV0WlU50ZY+No6b8l5kt6jjndF8eU2F68IG77TI+FOi4fEPSxkuW2/irj9R8ON58aen26Z/bZRdofHW9jx7MM+GvjpjanlPfws8KTnvjMaB0T6km6vOz1ncUUJ6CmVhrY27pEG6Pt507Wm0v85TGLCLSFLrSJyIiIiIiEmM66RMREREREYmxWJ70OecGnXOdrd4PERERERGRVotNcrxzLgkgBSAHoBvAwKSfjNsD8HTUJ2n9+2lT1/2mjTfeZmOuUdj76dJtnuqace3bwadNeP6EXTz/tneaeM6J75j46o8EAb2uizR7fI6WL+2mbVGZx7Ev23hNVxBQqVJZTRfXMvGU/VzPxLVQXFMiMxsf11y3w3V7Jw+UbnNNLXn6rsN2Uwm7/MmMja98nY0fedbG33mpdHsRbWsBrZs8auMOPu4z37Mx105xfM3HSrd5DHDtMddZiVRT1q+HcI0ftxbhmkD+DuVjs9ZjGX8W1Go9UWl9kdmAx5R+h5EpEJtPR+99FkAfADjneiJWFxERERERaQuxTO8UERERERGRvNhc6WuqqBRLzoS5+VdsvP8rNuaWDSu7bBxOR7/Xpoqe329XnX89pdnkaPlG2rcRm85Z9hMPMwooa2bOBlqV9mXpMXouyujhqfARxpyOmf0hxZTuGabzAcCNf2pjToPj51fKz8zGKV48JTxP6x6mi/Gxce4FE65abxc/sdfGx1+i5RTTaMX94XPTsjkUf+FbNr4xa2NONX3jQrvvAKU1hyk/x+mF8DGu9E6phtMvOb2avwOj1q8nRZM/qxk/N6/P24pqN6HPfpkNaqVzKvVTmkRX+kRERERERGKsrU/6nHO9zrkR59zIqVOnWr07IrOOxpBIYzSGRBqjMSQyMW190ue93+297/Ledy1btqzVuyMy62gMiTRGY0ikMRpDIhOjZPdKuBaGaoQuUnnDxS8/YuJ577NtEnCQ6upW0vaOPVq63WGrhObfmjDxyc/b2rflt9oqoov7L6KWOVfa+LmgLGjRjQvsQmqb8Ob329dx4mG7+kp67kUdtPGwncQJquFbTOtyDntZDSDV/IV1kUB5m43Lg2JH1XhMP55mnX++XMPH8eVUrBq28OD2Hk/Y4/TQflr8Uxv/woZ4I8U/oTispqC9xCaKeaL6o7bDCq7k434exQepxu/xz5Rub3gv7Ri9R1yHxdPyqy6kvYTHA9fBcUsFPla4xo/HZ1SrkfDzm4873jbX9PG+cM1eVAuHWuvqu0BmA31WS5O09ZU+ERERERGRuNNJn4iIiIiISIwpt6GSMN0SAE7acA5lxszh7BJO51y+1Ma1pqQ+Sflfz9t4ObVRwEmbzjnnalpOU8SbFEsAlwXpZc899AIts69jzhr72Ll89CQpfpLi4H3iFNk5m2jye05N238P6sLpoOF7rlSJ1uMULU4X43SyDuq7EKZlpe1x+hxljrJVl9r4PLVosMnaAI1WnA23Rcu+TTENGTxL8eiPbPwExZu30L5cG6Qtcwpd50dszGNo9bU2Vqpb+4pqa8DHDuNjh5+Px2vYVoW/Xzk1nz8LeF+4/IJjXj/cV/7uFZkuUW0XapU8cAmDWjjIJMXmW945lwCwA0AC+dOPfudcGsCw9z5d46EiIiIiIiKxFZuTPu99DkCqEPa1cFdERERERERmDNX0iYiIiIiIxFhsrvQ1FdUnnKeavudzNl7KhTubfq3283NNQ1j/wD+RiJ/QyREbL6fyCH780/ttvCrY9CJ+HettC4czX7c1f1xfePH7Nn75RRtfclnp9hwqAXnu67Z28TVzbbzgRqqLrPUeAkDXR2lncpgw5cc3H9cMcc0l1wRxvdqT91H8vaqbevZ0REw1fP9Aj6fqCRyj+GKV20B5iwbaFLgk9+1UyvoSzU7/3JiNV4StTrp+xy58Zr+NeVp+Pq5r1UlK/IRjjNsgcI0e19UxPrZO0BcRH4vJ7tLtzOfsMm49Etb/AeXHae6ojbmGj4/zsOZPx7hMl3rr7njM1ao/jaoH1O8wUoWu9ImIiIiIiMSYTvpERERERERiTLkOlVC6yHzKLjlDmYVlrQqyNFnoBUoC43edp6yu4ce07Td32/jEwza+jLJwVr2bnjBMwaR0Tpyw6ZxLeTepHcQrFzBp8y+zsX+VVjh0hjZGyzdQesPXP2bjqz5Yus1TfPN0yDL1OP2LU1s4pWvxahtfCNI76Vi5SMfGw2dtzJPRc1bzLyimDEucCm5zd4j3UMzpnZw0dxd1aPl1fjxl4a3YGwR7/84sW3TL6+3KnDLL7yGn6HHqW5hepLS42S9MAeP0L06R5JjTQTn9mlMyef29ny7dvnyjXcap23xccjpnkr70+DjnbYepp7xtTqHTcS7NEpViyWOGfw8Jl9eT+gnU3x5C6aBtQ1f6REREREREYkwnfSIiIiIiIjGmkz4REREREZEYUwJ7JefsxOs/+KJdvKHTxmcetvHSf8U1Cz+0MdejHfhK6faazXbZE6MmXHslPTZnw5VcX0jrc1uFMydKt5cvtzV8oFKJ79JM2+/4/77VxPO/b1/nXGp1MSeYr/48pbMffcLG6zfZ+CTt9/L3Uf3hVwds3EXvY5gfz/nw3M6Ba0SY6j7qxzVCXLPHdTwP/JGNuU5nTdDC46St9zxMw20+7Qo9E2jLeIpirp4I1z9Ky56k+LUUc9nrEYqvofg0lQMvDTZw5bW88s9tzNPocx3H+ltsHNa9AnYafx3zMx+PEcZjLMQ1Q1zjw4/l9bnujh+/YlPpNtcDcg08f1ZwrRN/fvNxzXV7XMMd4vcsqs5VpFn4OK5V48fLOObxFjVGpG3pSp+IiIiIiEiM6aRPREREREQkxmKXu+Cc6wTQhXxDgQSAnPc+XfNBjNKiFibs4vmLKaZsEpw+aOMNNg0SJyn/7FhwO2fTOUEpko5O08P0TAB44ZyNnxi08Y3/l41NmuQxSu980YbveL+NL6bt65iz3C7/+TEbzw9ey9JNdtlaG2IOHZnLaV79F75p93XB9ZTuyWk6YToEp29GTQnOU92/Svm5SgGKFvUecYoXp6cc32vj06WUzhyNkc5fsfE/PWJjSiwFJZvhWYq/TfHbUB2ngtJRiQco3kDxNyh+C8UjPy3dfu08u+wcvYWbP/Yde0dUGhy/x2F6Ef/8dMzPPLXSNwH7ucXpmSyqZQN/x/F4Pfg1G1/2utLtl2mE2WoK4MpVtffloX4br3yzjbkFUngcR7Vs4FTQqPdU40CqiWrJwHj5458p3eb0Tf4dhb8/o55b2lasPrGcc0kAO7z324L7Bp1zY977TAt3TUREREREpCXilt6ZAkAzemAngP4K64qIiIiIiMRe3E76bkc+rTOUBdBdYV0REREREZHYi016ZyG1M+G9Nyd93vuccw7Ouc4Jp3hSHv9LVM7w1X+08fu22RhJmmc9SzV8SarxWx/UjD1kq3ouUDnDeaqz43rDpdSyYc27ad9epXh/UBvHRwPV6PHc9nPofTm538aXr7TxvKB84iTNJs81fFw3+cIYauIaP8AWci24cSmq4vz4g/famKe253YDEo3bYnCdzmruP0C4PiIoT5pLtW2PDNuYKoTwC4odxdxW4Z0Uh6ObWzLQYQ2qqsP1FNOu41KKl1H84+D2m6ie9yKVmj5xl42v/GjENPv8Hoe4Djbq5yWtx7XHtfCxwD9v/kzk2re9f2tj/rx+Jajj44443Gni+NM2Pk3xVfRZzrWqXG8Y1mRzLWPU1Pb8OcXrq75bqomqo+M6vaMP25hb6IRqfVYD5ccxxzwOVPPXNuJ0pS8RsbxjOnZCRERERERkJonTSV/dnHO9zrkR59zIqVOnWr07IrOOxpBIYzSGRBqjMSQyMW2di+C93w1gNwB0dXX58QWU0pGgNMcrOGuG01WWU0IZp7bxNNAHgpROak0wj1NfyLwr6Q7+iR6rvTxMF51P6ZjcsgHUDoIt76LnPkorXFa6uSBhF72Qs/F3aW77d7zPxjmq3ExwSweaxv/Co6Up/ue9j1Ib9t9l44230YPphzBGk/xzqlMbpfhUHUPskoSNeUxwSha1TSlLPwucoinfOeWSE395snoeztyygSaEx0PBbU705eHKqaE8ATwPT0pkKytQDrO1j7xkl3Fa6i3cP+JFeice+zsbd9Heh6lsPNX9OXrT1dZk0iY8hhrFqYoh/vny+OOUdv4MXP5GesKnTPRCUFixgA4V/ux/6jEbX0mZxK88dsbEc699nV2BP1vCtDlO5ed1o9KW+bjndE9piWkbQ83E6ZysVhsGTtfkVNGoFg68PqeLhs/PqZ/8WJlV4nSlLwcAzrlEleURVWEiIiIiIiLxE5uTvsIELjlQ7V7xJFB9+kREREREpB3F5qSvIA2gk+7rKtwvIiIiIiLSduJWaJECMAhgKLivr3D/xFF9w7dH7eIP/MuIx/N0uFwvwTViK9eVbnONwSW2foHLAV/Yb+PX0hzw8yLSr01rBCrDKaubo3KHM4/aeCHV/HHbhZN7S7eXUw3eXDoSL+XXSdPTv0hvYYLKVRasBa0Q3Oa6DM6P559B10dtHFWfVGuqdNU25UXVDPFyGjNj3BshcOkcG19CHVS4BJdbOFCFkGmTAADzg9tcJUXlSlhE8f+i+DqKf4li2nV8K7jNJbdUUov9+2x8y1pagXeW6yqfCD5GN7zXLltDex5V4yfTj2v4wjHE30H8GfjMfhtzTW3Z+raG7xU6lF4JCk7P09fjWYqX0aHzDG16CdXYz32CqnAXUxx+B/LnDH/W83HLtcgc8/vIy6V98RhhXBvH9aHhcq6548fytrimj+vyGqnTi6onlBktVr+Beu+zzrmUc2478nMgJAEMKLVTRERERETaVaxO+gDAe5+G0jlFREREREQAxPCkrykoLYayFHFkv42vuJVWGON8MspfueqDNg6n7uWUHN44ZQ4uoFSX5ygNsqzlA7eov6rUXuLCw3bC+AS3cKDnXkrLz9DyuZRqGra+OP9i7XVfimhVsfJquoPz7Gq1rsjRxPhdNLH+XEoA5FRDTo3glB5N412O01p5avTv/72NOUWa2g0cCbJdnj1Lm6JNL6P4pxRTJlpZG4YrKA6zJr9Pyzh1lFsw/L8o5hRNPoy549R7gts8HfF3KH4XL/+6jV991cabbvihiS/92A2lgNM3D91v4xWbbMx56JzyrJSg5uP3mH8G4ecar8ufcZzqyz8vWv/ik3axo0H4uuDh/D3BLVfekLTxTyhPZ8VaG8+7pPbyV4LuEgte/Vu7cONv2Dgq7ZW/n/k95uW1jnOl+rcX/k7j3yPuvs3GYXsRXnfvp23Mv1Ne/SEbc3uIqBTNcPnaG+yyqN9vlP45o8VtIhcREREREREJ6KRPREREREQkxnTSJyIiIiIiEmNKKq8knKocwOtfbxcv5ro4alWArs02XkmTqR+428bhtNFHqR5wLc0/n1hr45OHTXjxmF1cVmTE6dj7S1VH866kZdSy4SKVK8yh5+YavxP0+MuD5aeprmPVu23MU93PoToP0OvkOpGlqIHfg8epGur6hI05J53rNnja7zAXv9LjpbymiKdS57qAS+abcEPn+fHbx6j86BQV0tlqQIAqK7CJYl7/f6M6rpt7nGI+bI/UeC6gvOR2HcUPB7dvomVUXYQfULyKPu3feTOtQPXDL33mr8dvX3ojfQhe/x9szPUqjGv+eAxxHa3UL6pGLBxTXIvGMbcZ4pqgMduiwVN96LM0fOcEuzb6z3YZHzlP2C5FeAMt5+c+SR0a+PNgUTCoNvFH8StfszF/VnOrivW3oCZ+H8MaQV4m8RJVy8ZtFHh9rtsLffpXbZykb4b/Rn3EPvJXNuaa+fd9rvq2gOjP81BUawqZUXSlT0REREREJMZ00iciIiIiIhJjSu+shNIxE8tt+t9SToPk1gacxvHwgI27ftPGxx4t3b6VUgN5Km1KN7n4hF2coEzSstRTboWwPkgfPcStJmw45zJ6bMKGz9GurnwfpaYGFpyz2zp/1C5/kqbpxl4bvo2ybJZSZsR5mgb8QpBNtojaXGDjKhvze86paJyue/JA7eVhqgSnTbXLtN2czskpIRxzuufYeROG07T/4oJdldskUDcQvI3iH1PMKZWbKH4guM3tHXjbPPy4/QMPR3qXOIsZYRYdNU0AJZWXGaU0uNfRGDtiM8Vx87Yg4J9H+JkFlH/m8RTiY4dszONA6Z6N4zHGwnQzTt/iVgX88+A0R1p9Lv34l9KgG3kAVXH6JrcpoY4sOEbHMX+cJ6mdT9j+56n9dtkb19CD7/s7G99yB2raeJuN+bsgPM6j2jm0y3dBXNVqewDUbtUFlJeFhCmdnM5JZT24ipZHtYfgEiNeHu47vw7eb35srTRVaTld6RMREREREYkxnfSJiIiIiIjEmE76REREREREYiy2SeTOuUEAO733XB0W7eC9Jvzpk3bx66kOYP4177R3cE1Yx4La2wvrV/bfY5ctp1ztV20t3JzruB6Nq4rIxndWX3aSWhdwWU3yrTbe/0MTLuL6CKrby9HbUstKmuu+rI6Si6Forvv5VH84N6gxuUCtJObxe3Ydvedc73JJwsYb3mtjruvgeigpb3PBMbVN4X4E4c/3/It2GXfk4Lo6rhm6lGKegPpLFF8f3N5AyzZRPIbauAqLt/0WisPOJNfTMm4tsZViHjLHqCxk+bLy/RvHY4A/47heheuTOI6qP6u1XLVPlUW9L+EY4zodrrnkGk2usdxIFaSnR034wKBdvPk9pdtPR7Rs4DpWbt9DDwcdebj/u7Tt4Li+nOrvn/qqjd94LT3ZV+n7+P2/Y2OeCv/KHhuH7zO3W4r6XuExoON+domq8ePlXCt3VfC72lH63ewy+p0ySx/mJ/+s9vpcP8j7Fh7XXB947SdsHFU/yPj3I9UATqtYfYo455IAUgByALoBDNR8gIiIiIiISMzF6qTPe58F0AcAzrmeiNVFRERERERiL1YnfU2z6cMmvGyhvVQ+n1s0nKP+ALd+xsaHaHJ1vhweTvHPaW48zfM8ukzPaTecgvl9m4KJA5QmsDx4PkqTQ5JSR3m/F9P6nEe32LZsSNyaKAWHzth1qbq07D3m6lN62RcpZXMOHdlzgud7mTOZOum5+WfAaRjZtI2jppcP03Z46nNO6eE4ymxN+Vm4ona8YpONy6aNLh3XV1B2yH46xCnTDP+GOok8Sp1KOL2MD48w5tS0d1P8CMV8GNOmQbPRgw5V07mEEyB/iWIaEvgFxdyB5Rp6wgeDN27Tie+ZZUvX0oPX0Gcgf+bRZ2pZOiH//OsdB9KY1ZTXyKlmUS01EjZ09PTpICfzCsrX/NFzNuaPfqquKEtr/inF5yn2QT73VurvcJwOwxVrbTyfc8Uz1NJhw6/ZmFPVwuOav0f4mOd0T06xlZmNv6P496V61w/TID/yf+yyf6CSkmttG7CTn/6KiZ85+oKJxz5pf6e94c/o98ruT5Vuf/Xjdhmn8jNO3+T4mo/VfrxMKU3kIiIiIiIiEmM66RMREREREYmxtj7pc871OudGnHMjp07xnH4iEkVjSKQxGkMijdEYEpmYWVoQ1Bze+90AdgNAV1eXH19A9QtrNtIDN1KtW1iTB5S1fCirEeP1Q1wfyLUUa66zMdeXnaNauVfp+Tf9SvXnn0sFDpy7vf8uG3Mrilds3jgu2Iql575Z2rdF9DKQo5iOzAtU2DGPyiPmXLvU3vEivQ9PlG4u4Gm56S1Glh67iepX+OcXNX192H6g8yN2Gdd1RE1tP8Nq+KqOIcb7zTVdXMPFY2bxahNefL5U03eCite4dq2X4keokI7Ldl5P8dso/lFwm+uPfkwx1+gxfu7XUXyE4nB2qidoGU9tz0Yp5te95qyNXx/szM+p0LGspu/A3Ta++kM25s8prn3ieqawzUnMp7Kf8BiqF79PYY0Y14udPmjjiJrL3Odt44RT9LXFLyIcco9QDR+3YOBvRzosuXsLrqixLcDW+P0jlbjz+H2VSlF/maeDo7cJK+iFc6uZy4NfHrhOOeozUDV9EzZlY6geXLPJMbvvozbm3xseKtXdfeeqXzWL3vlp23qLa/iWf8LW+D3zCbv8hk/aTX3nz+3vbu/En4zffm6vXXfRjRE1e/w6omoAubYxFPUeSt1m3belc24UVDbuvV9XeW0REREREZH2NutO+rz3m6PXEhEREREREWAWnvRNC7rcvHQNLR972saclrH2BhtziiancIYpm5w6mPmcjXmqbF5/IT3385TPcpQmkb8+uM7PaYo8bfcGmiaYpxi+8A1absNFHw6mt/4mrcvVpXRVn9sqXPy+jefwe8wZA+HbRvuFqyjl9bJHbcwpmJRqWPYz4TTY60qpEmXvMafwNNrCIdTKNLiobXOa8uPU5oTfl4w9XubMK91edzVtm46NZ6jEg9sqcBJVhmJOFwunjP9ftKyDYprQvawFA083P49ifr5Xayzj1NKtFN9E8ZUUU2Ybbg/GyfpNqO1Jmih/Q46W32dj/tzi9M5wPEe1UJHKeMzx904osbb2upkv2dVvsIvdN2n1wzYOvzG5pQK3GuGUZh5/+ymmTZelTC8LbnN2Jqd3LqU79j9k41XrbbzsYfp+3UjxhuBz8DjlyXGKM+PjPuZpzrMepyny5xQv558/pUleCL7HruSSFGq99QS1KeJ0zvnUn+fH/2Djs1SZE1r0QZtKiq/adg/cLgKf+5c2TlIi3o1/Ssu7bazP9ynV1hO5iIiIiIiIxF2s/jTknEsA2IF8zV8SQL9zLg1g2HufrvFQERERERGRWIrVSZ/3PgcgVQj7WrgrIiIiIiIiM0KsTvqapoum0n3472y88q02Xn+LjY9RTdiVNO8z15+FNWBc68KP5XoynvaZ86G5RQML88jD6aUB4ARN8h4xjT6u+bc2PkRVQoeCuqyoxGKuyUvYcM77qVqKa+U2UB1e+J5zjd4BqnNcQznoGZ7snvDPk6cozh0t3eb3kGtnuLapkfz2VtZ4RLWe4HrQDiqW4fVv/nc2vu8vS7fpWFmYsPGSF238iws2nk89HjbZELQ6zgW3qXlLWZ3cjyjmOjz+Cc2JWD+sjeKKoBzFVF1Utm9UmQyqTMX3gmKqIzREfnsbrUw1I2Vjgn3/723MddChWlN6A9F1r+1a61SrZUPUz4dbOPCPYINtkfPcmG1zw8dtiA6lsuOSuoPgWxTzcc9tTrhmMCzppcY+ZiwDwD9Qke07eFv0cbyMvgJrfq9xff7N/9XG/NnPteK8nH++jdR/y8SEn0VRLRq4lQH/3sGfgVTbNu/aUkus8w/Zorvn6Kk7qM/QiZ/bmNui0BQJZe7/ZGl7t/wZ1RpSi6TzB2z94KLP/R+7Arfr4bkiuKYvbP+jer+mU02fiIiIiIhIjOmkT0REREREJMbaNO8lAk8X//7/28acosdpF3xJmi/z83T1G2+r/twcc8oHbytMJQTKU6P4UnuYcsD79e4/sjFPOc1prAfvtTGn8G36ner7wamih2gO4k00LTCntXJqywjNGR7mG3XQfi2npB9+rmspRfYxSlBaT4+v9b68j1J8ntlvY/55Rk0DzcceH7szFafcRaUmcWpUkJ12kbKcTtWYmR4ADlI6JyXdlKU9cnpnmDZ5DS3jtgeLKOZUNjrK8SaK+aWE6aL83PyTp6Rz0MsGvW1laXWh36V59Y8/SdtO2Dix/2v2Du5FsYkewOmGYYsHbhXDx3zUsdMuU9tHpQfWg8sM1tNn2kmbzsljkDLATFsGbslAmWlleHx+hWJOseaMy+8Gt3+Dlp2hmMc6Z7W+RG/xd6mlw5WUSj5/TfAATmHn3ws41Z/LBPjny+mCMvVqpRfydzXH/PO6+kM23vtpEz73UPU+CukHbZyj5fTpiyso5jHYTYNwMEgPfemTdj/eQx0cltOvLBf+4FdNPO8Tn7Qr8O+FYTonYI97XpfHTK2yAKlIV/pERERERERiTCd9IiIiIiIiMaaTPhERERERkRiLaXFDgzgXm+vqEmttzPUPXNPF9Sqcp1zruTnPn5ff/wkbL6cJrbn9BOek16r74G1zvdi1tO0nhmzMNQlhHR7XD3It4q1URznyWRtv+rCNw7YXAHAztXQIc8H5NfO2+efDdXZdVFvB9YhcuxHWYoS1SpX2JWp6esb7Gh4fM6m+j+uo+DjkY42njKfj/uJYqbLu9Am7aueNNj5CZQBLn7Ixl5vxlO+8fFlw+1nUxlPX06sqa9HAtRYLKP6XwW3+8H4PxRmK+XVx7RNPfZ8Ibh+nqezfvNjGL9Dc9wkuEKQhUVbTu2KTjcNjl+t3+TOOa/aixLWmL2q8h8v5c4Y/A/nnw5/Xx23VELdBeYk2Hf6EltCyJyim4Yo1FP9fFHNDnaspDutgn6NlXMPHfwXnI4vHwaJLbfwy1/Q9FEyW3xXRnqdWqx+g/DOzVr1/XI/xmYx/PvzdzGOqLLbzECy6sfTp/50/t3V13TfZh+6hGj9q9IXfp/hvKP5navHw8aBu70tUeP4ExWdP2nhl0sbn//zPTLxoOW38I/R5/tB/KN2+5mN2mVo2NExX+kRERERERGJMJ30iIiIiIiIxphyASvgSMqdvMp5Gltfn9CQWpnFwykZUSkf3p2zMqaS8bzy9ebhvnKbIz3X5RhtzCt6JERvz+xCmn6ynCeU5fZNTJDm9iNMieVu8L7Wmdef0TV6X30Ped06V4lTF8PG3fsYu42OD33PGy/l9qJWuO5OmrufXwalpPBUzvU9zgtmv53DOJFlGh/UcetmcmnglDbmnfmTjsHEJp2P+FsU/oJhbPFA2GL5NMaeHhtllr6VlPP38Boopmaxsqns+GsKsnXto2R0/tDGnwF5KP84Onl3+clqhVqsaHts8Hjk1lI/zmZTmPJWi2rfU+mzhx/LnLbfjoY+ZVyhPkt/xU8FtTr+kBjhlxxKnWP5viinTGPxtG04gT5ngoGxNrKP4RxHxL1Me6yvftPHbgrcxkaC8cj5OV19rY/585nRB/owMv4f49wale04/TkNn3MKBfo947quldM93fsSu+t//0sYr6an/PcV/QTGnRHPK9e8FKZzvpmUd1N7hZ/QleITiWx6klg2c1sq/P4XHea12DpVEtbiS+J30Oed6kP9daR2ATgAD3vuh2o8SERERERGJp1id9BVO+DLe+2whTgAYdc4lvfe7WrpzIiIiIiIiLRC3mr6O4gkfAHjvcwBSAPpbtkciIiIiIiItFJsrfc65JIAB59yI9z6crTxTWN5J91fHudhco8X1J5yLz3j9WrUVXJPFed88jTPXo3EONK+/4b02DutleF1uD8H50bxvV9JEwVyvFtbKffNP7TKuIeG2F7xvUdO48/sY5oLv/bRdVqv2sNJy3reoms+wbQY/lvc76mfA+GcQ5sPzspmE32OuF+X2H1zj+ZrSVOi/RMVrz+ds/BKVK12kUpqFCRtzfdIiKnhYcrZ0+y12EbjhBlcU/Jhi+mTBmyjmCs1wunquR7qCYppZu6w9xFGKuX4wnHmbp82/8u02foae7EWqk+zgufHHqL6Jx0VYX3zofruMa2q51omP+6g6Wa53mq31T/W0bIiqezy+18an7ZF76jG7eBEdPDddZuNPB8VwF2m3qEQI9NMGf7rSN56pFwTKx1hYm8p1rrwutz35Z4r/FcU/oXg1HTrhdPYv0Hu26pWn7R2n/9jG1/yOjfm7m2ujwt8For5HZusxPpMdfdjGXHvMylo6UMuG4FeDh6mGj1uifI3iKyn+PYp/o/aemTo+6sCAUWrvcJaWf4x6qtx/k23ZwPv+L+6kb6q1QaMi/v2I31P+rOcxEVUD2IZic6WvcIVvF4AsLUoU/uf7RUREREREYi9Wf+7x3qcq3H0H8nV+uWneHRERERERkZZz3vtW78OUKUzkcgTA5rDWL1jeC6C3EF4FIOJ6fFNdjvJssDhsK+7bi/NrA4A3eu+XTXRljSFtb4ZvqxXb0xhq/bame3txfm3Tvb26xg/Q0jEU55/DdG8vzq9turdXdQzF/aRvGEC/9z49gXVHvPcRxVnNM53bi/Nrm+7txfm1NSrO702cX9t0by/Or61RcX5v9Nq0vemgn/vs3F6cX1srtlfNrEvvdM6NolSnBwDw3nNPVTjn+jHBEz4REREREZG4mnUnfd77zVHrFC71D+uET0RERERE2t2sO+mL4pzrBmDaNhTaORRn+Kxm91TvWwu3F+fXNt3bi/Nra1Sc35s4v7bp3l6cX1uj4vze6LVpe9NBP/fZub04v7ZWbK+iWNX0FU74Eij05gv0VZnZU0REREREJNZic9JXmKmT+0QWZSvV/YmIiIiIiMRdbE76REREREREpFzsavrq5ZzrBNAFIIt8amhuKieAcc4NAtgZ1hxO4bZ6AHQAWAegE8CA935oCreVLITrAIx676clh7lwlbffe983Rc+fBDAIYADACPKvcyvy7+eU/Bzp/cxiio/LyZru8VPYpsZQ87edgMZQS2gMNX1bGkPN26bGUPVtagw1f9sJTNEYasX4KWx3Zo0h733b/iv8IAbpvkEAnVOwnQEA/cinoHZPw2vrAZAM4gSAwwC2T/W2CveNTsW2qmx/kH+OU/DzO1z42XkAw80+Rmh72wH00vt7eDrey0m8L1M+foJtaQxN3WvVGGrBP42hqdtW4T6NoclvT2Oo8rY0hqbutU7ZGJru8VPY5owbQy3b8Ez4Vxi83XRfJ/LtHqZqm4en6YOit8J9PQD8FGxrFPm/zoT39U/HwQ2gu7Ctqf6yndIPB9rWMN2XmI5jZhL7Ou3jp7ANjaHmbltjqEX/NIaaui2NoeZuS2Oo9nY1hpq77SkdQ9M5foLtzbgx9Bq0t9uRv9wayiJ/8M1ahcvYA4WUh1CmsJzvb1QKwD10XwLl721TFVIBgPyHb1wMIP/XrnHe+5yfgSk1iOn4ATSGZjmNoRlAY2hW0xiaATSGZrUZOYba9qSvMJgSnnr3ee9zheXNHkzTpvCadqF8oCYK/zd1AHvv0972RUwg/0E81W0ybp/OAeScSzjnOot9H6dIF/L55nDOdU/xtiYtzuMH0BiaKhpDJRpDGkOToTFUojGkMVSvaRo/wAwdQ2170ofSoKmmYzp2Yqp471PFD77AHQAyFe5vCudcj3NuO4AdADb7qS2O7Qbwxal6/gruQFAo7pwbDP461RSF50sUbvcg/4GRcM4NNHtbTZCIWD6rxw+gMTQFNIasRMRyjaFJ0BhqjMbQzKIx1FRTPn6AmT2G2vmkr60UDrReANumahve+yHv/S4A+wCkpurgLg6oqfrAY4W/tu0s/BUsV/gAHAawp8mbKv4lKFl4L6dyW1InjaHJ0xgSQGOoERpDAmgMTdY0jh9gBo8hnfS1j0EA2zgNYir4/FTChwE8OEWbuN1P0XTF1VT4UEoD6JmiD8Pp3JZMnMZQAzSGBBpDDdEYEmgMTdo0jx9gBo6hdj7pywGmgJSNTdueTDHnXD/yMzJNZwFpGkBns3PqC883rYWwzrneCncXj49m5mnnCv9XrE9APi1hpsgB7TF+AI2hJmxTY6hcDtAYmmIaQ/XLFf7XGJphNIYa2t50jR9gBo+htm3O7r3POudyyOd854r3Fz88pjKHeToVDvThqfqQKBSnjiKf9x0e4LnC/0kUZppqkiSALc658L5u5POl+wHsa+ZfjoLZs9JV/rLWtL+2FY5JoHqdwpT/ZW+i2mX8ABpDjdIYqkxjqKnPrzHUJBpDM5PG0ORN5/gBZvYYatuTvoI08v1cwh9AF6b5L3hTpVAgO0KzMSWB8fzmZkggX6TKf1Er/uWkqR+6hQ8B80FQKDje4r1v+gxThcHbV+H9uh1TU0g9hPwxGP7MOgHkpiOdo06xHj+AxlCTtqcxVJ3GUHMkoDHUTBpDM4jGUMPbmu7xA8zQMdTO6Z1AfhraHXRfH6Z+etopV/iQSADIOeeSxX8AKh34kxYUp7IUgN3TdHAvneLnH3PBdLuFvyL2AbhzCraVQvnxt6PCfTNBbMcPoDHUZBpDlWkMNYHGUNNpDM0QGkNNM53jB5ihY8j5fJf4tlUYUMW/EiWRP+tv6l+ICgfXDuQHbi/yZ/5pTNGl+sL2zlZZnPXer5uCbfYCWAfgTOH/w4XZn6ZM8YMPQA/y6R27AdwzFSkdhWl3ix8Y65DPq5+SD8HgdZ1B/kOwqalCzTQd46ewnQQ0hppOY6j1NIaauk2NoeZtS2OofDsJaAw13XSNoekcP4Xtzbgx1PYnfSIiIiIiInHW7umdIiIiIiIisaaTPhERERERkRjTSZ+IiIiIiEiM6aRPREREREQkxnTSJyIiIiIiEmM66RMREREREYkxnfSJiIiIiIjEmE76REREREREYkwnfSIiIiIiIjGmkz4REREREZEY00mfiIiIiIhIjOmkT0REREREJMZ00iciIiIiIhJjOukTERERERGJMZ30iYiIiIiIxJhO+kRERERERGJMJ30iIiIiIiIxppM+ERGRKeCc63fOdbZ6P0RERHTSJyIiM4JzLumcG3bOnXXODbR6f5qgB0B2Iis657qdc8OT2YhzbsA5N+qc84V/o4X3sfhvtLBOcjLPLyIis59O+kREZEbw3me991sBjADoaPX+OOcShZOmnkk8thtAxnufm+BD+gF0T+bKoPe+z3u/uRBmvfebvfdbg3+bARwGcNg511vv84uIyOynkz4REZlpJnR1bKoUrooNAugF0D3Jp9kG4J4Jbq8bQPHK5o7JbCw4WUxXWu6934X8+zrgnEtMZhsiIjJ76aRPREQkULhytq1wojRZt3vvhya47jbv/W4AuwH0TDINs3hyWitFtHgyPdkTWRERmaV00iciItJEhXTQL05w3U6UTtRShf/7J7HZrYX/K17pKyieTGYm8fwiIjKL6aRPRESkue4AMDjRdYtXBAv1f5O92teNfD1frtLCQgppEkDae9/S9FkREZl+c1u9AyIiIhNVmIgkASBX/L+QGllt/c7COtlCnATQifwkK00/+SnUy3V672tdcSuumwSwj+7uR76WMAWgb4LbrFnPV9inAeRP+LZWWkdEROJNV/pERGRWKEyukvXe7/Le7y7U3I0V7ud1e4L7O51zg8657cifKGYxuRTKibgdwERr+fq47q9wIjoEoLeOCVcq1vMVWmBsB/AggH6d8ImItC9d6RMRkRmvcPICvoLmvR9yzu1wzvUWr/gVrnwNAlhSSHfMOOeA/IlPcXKWbVO0q9tQqs2rqnCV73CVxTuR7/G3YyLPhVI935YgLXQd8ieDmaCdg4iItCnnvW/1PoiIiIwrNGbv8N5vC+7zyM9yWXYVrZDy2e+9X1KI+wF0hyc7hatmZwFsnUjq5US2W2HdBIBR7/26Cazb772vekJXaNTejdKJa9Q+Zittt/A8yYnsk4iIxJfSO0VEZEYLrl5Vq8HLAkjMgP5zvSj126uqsJ/VrvIVFU8IazZTj6rnQ/6KZ7IwkYuIiLQpnfSJiMhMVzyxyUWsVzw5HA5uF3UhP6HLhK/yTcIdmFg9X2+tyWcAwHufQb61QlSz9qj+fInC/5Pp/SciIjGhkz4REZnpilf4EhHr5YDxur+0c27AOZcoXA1LAbhpqnaweDUyakbQOq9GppC/grm9xjpR/fmKy9WmQUSkjemkT0REZrTCVS8gf7WukkptGQaQP2nqRr4+cGvwPFOhDxNI7UQ+XbPmVb6iwslr1NW+mv35UHrPdNInItLGdNInIiKzwS5Un3HzDthZLjsBJLz3Oe/90BSndBb1APjiRFaMmpiF7ET+al9Zbd8E6vmAwtXR8ApkpecSEZF4U8sGERGZ8bz3KefcsHOuJ5xJs5D6OEY1chkAg865ZjVg76i1sHDyVetqW3G97QByzrmeSexDCoUrhIUrmQmUeg121NiHLIJ6vsIJ33ScBIuIyAyilg0iIjIjFOrd9qA0OUkaQIquUhXr23LIn/jkKk2KUmj7wFe0ssFz5mrsx3bk+9x1oZA6WnjcWKXHFrY1HNXWwTl3FtF1ibVsK/Ql7K3yPEN8khukuuYA7EO+b59O+kRE2oxO+kREJFYKffr20RXBBPJXvPoA3F7s6dek7Z1t5vOJiIg0m2r6REQkNgqpkwm+6lao78t47/sAjEwyxbLS9rqhdEkREZnhdNInIiJx0oHofn7NnMlyorN2ioiItIxO+kREJDYK9X2dhStwZQr3J6Pq7+rQrRo5ERGZ6VTTJyIisRNMdpIL7k4gP8NlU074CimiWwspoyIiIjOWWjaIiEjsVJrRcwpsgVI7RURkFtCVPhERERERkRhTTZ+IiIiIiEiM6aRPREREREQkxnTSJyIiIiIiEmM66RMREREREYkxnfSJiIiIiIjEmE76REREREREYuz/D0y7SQZXF0sLAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 864x432 with 8 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig,axs = plt.subplots(2,4,figsize=(12,6))\n",
    "bins = (50,50)\n",
    "rnge = ((0,7),(-2,5))\n",
    "cmap = 'gist_heat_r'\n",
    "\n",
    "xt = [ i for i in range(rnge[0][0],rnge[0][1],1) ]\n",
    "yt = [ i for i in range(rnge[1][0],rnge[1][1]+1,1) ]\n",
    "\n",
    "labelfont = FontProperties()\n",
    "labelfont.set_family('serif')\n",
    "labelfont.set_name('Times New Roman')\n",
    "labelfont.set_size(20)\n",
    "\n",
    "axislabelfont = FontProperties()\n",
    "axislabelfont.set_family('serif')\n",
    "axislabelfont.set_name('Times New Roman')\n",
    "axislabelfont.set_size(24)\n",
    "\n",
    "tickfont = FontProperties()\n",
    "tickfont.set_family('serif')\n",
    "tickfont.set_name('Times New Roman')\n",
    "tickfont.set_size(20)\n",
    "\n",
    "def get_x( arr ):\n",
    "    return [ i[0] for i in arr ]\n",
    "def get_y( arr ):\n",
    "    return [ i[1] for i in arr ]\n",
    "\n",
    "axisfontsize = 18\n",
    "labelfontsize = 18\n",
    "axs[0,0].hist2d( get_x( qcd_lund_0_flat ), get_y( qcd_lund_0_flat ), bins=bins, range=rnge, cmap=cmap )\n",
    "axs[0,1].hist2d( get_x( qcd_lund_1_flat ), get_y( qcd_lund_1_flat ), bins=bins, range=rnge, cmap=cmap )\n",
    "axs[0,2].hist2d( get_x( qcd_lund_2_flat ), get_y( qcd_lund_2_flat ), bins=bins, range=rnge, cmap=cmap )\n",
    "axs[0,3].hist2d( get_x( qcd_lund_3_flat ), get_y( qcd_lund_3_flat ), bins=bins, range=rnge, cmap=cmap )\n",
    "axs[1,0].hist2d( get_x( top_lund_0_flat ), get_y( top_lund_0_flat ), bins=bins, range=rnge, cmap=cmap )\n",
    "axs[1,1].hist2d( get_x( top_lund_1_flat ), get_y( top_lund_1_flat ), bins=bins, range=rnge, cmap=cmap )\n",
    "axs[1,2].hist2d( get_x( top_lund_2_flat ), get_y( top_lund_2_flat ), bins=bins, range=rnge, cmap=cmap )\n",
    "axs[1,3].hist2d( get_x( top_lund_3_flat ), get_y( top_lund_3_flat ), bins=bins, range=rnge, cmap=cmap )\n",
    "axs[0,0].text( 6.5, 4.5,  \"primary\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[0,1].text( 6.5, 4.5,  \"secondary\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[0,2].text( 6.5, 4.5,  \"tertiary\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[0,3].text( 6.5, 4.5,  \"quaterniary\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[0,0].text( 6.5, 3.5,  \"QCD\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[0,1].text( 6.5, 3.5,  \"QCD\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[0,2].text( 6.5, 3.5,  \"QCD\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[0,3].text( 6.5, 3.5,  \"QCD\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[1,0].text( 6.5, 4.5,  \"primary\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[1,1].text( 6.5, 4.5,  \"secondary\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[1,2].text( 6.5, 4.5,  \"tertiary\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[1,3].text( 6.5, 4.5,  \"quaterniary\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[1,0].text( 6.5, 3.5,  \"top\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[1,1].text( 6.5, 3.5,  \"top\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[1,2].text( 6.5, 3.5,  \"top\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[1,3].text( 6.5, 3.5,  \"top\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[0,0].set_xticks( xt ) \n",
    "axs[0,1].set_xticks( xt ) \n",
    "axs[0,2].set_xticks( xt ) \n",
    "axs[0,3].set_xticks( xt ) \n",
    "axs[1,0].set_xticks( xt ) \n",
    "axs[1,1].set_xticks( xt ) \n",
    "axs[1,2].set_xticks( xt ) \n",
    "axs[1,3].set_xticks( xt )\n",
    "axs[0,0].set_yticks( yt ) \n",
    "axs[0,1].set_yticks( yt ) \n",
    "axs[0,2].set_yticks( yt ) \n",
    "axs[0,3].set_yticks( yt ) \n",
    "axs[1,0].set_yticks( yt ) \n",
    "axs[1,1].set_yticks( yt )\n",
    "axs[1,2].set_yticks( yt ) \n",
    "axs[1,3].set_yticks( yt ) \n",
    "axs[1,0].set_xticklabels( xt, fontproperties=tickfont )\n",
    "axs[1,1].set_xticklabels( xt, fontproperties=tickfont )\n",
    "axs[1,2].set_xticklabels( xt, fontproperties=tickfont )\n",
    "axs[1,3].set_xticklabels( xt, fontproperties=tickfont )\n",
    "axs[0,0].set_yticklabels( yt, fontproperties=tickfont )\n",
    "axs[1,0].set_yticklabels( yt, fontproperties=tickfont )\n",
    "axs[0,0].set_xticklabels([])\n",
    "axs[0,1].set_xticklabels([])\n",
    "axs[0,2].set_xticklabels([])\n",
    "axs[0,3].set_xticklabels([])\n",
    "axs[0,1].set_yticklabels([])\n",
    "axs[0,2].set_yticklabels([])\n",
    "axs[0,3].set_yticklabels([])\n",
    "axs[1,1].set_yticklabels([])\n",
    "axs[1,2].set_yticklabels([])\n",
    "axs[1,3].set_yticklabels([])\n",
    "fig.text( -0.03, 0.5, r'$\\log k_T$', va='center', rotation='vertical', fontproperties=axislabelfont )\n",
    "fig.text( 0.5, -0.05, r'$\\log 1/\\Delta R$', ha='center', fontproperties=axislabelfont )\n",
    "fig.tight_layout()\n",
    "#fig.savefig(\"qcd_top_all_lund.pdf\", bbox_inches = \"tight\")\n",
    "#fig.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now I want to convert just the primary Lund plane into the image format.  We'll go for the same idea as with the jet images, i.e. a 40x40 image.  This is not ideal, since about half of the image is guaranteed to be empty.  But it'll do as a comparison to start with."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To do this I'll use the scikit-learn function KBinsDiscretizer. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Also, to start with we will use just one-hot encoded images.  Afterwards we can think about filling these one-hot encoded entries with data, like the pT of the subjet at that splitting or something."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First we cut the data down to the range we want it at:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "def cut_to_range_w( data, cuts, dim=2):\n",
    "    if dim != len(cuts):\n",
    "        print(\"ERROR: dimensions of cuts should match number of observables\")\n",
    "        sys.exit()\n",
    "    new_data = []\n",
    "    weights = []\n",
    "    for jet in data:\n",
    "        new_jet = []\n",
    "        new_wghts = []\n",
    "        for splitting in jet:\n",
    "            keep = True\n",
    "            for i in range(len(splitting)-1):                \n",
    "                if cuts[i]==None:\n",
    "                    pass\n",
    "                elif splitting[i]<cuts[i][0] or splitting[i]>cuts[i][1]:\n",
    "                    keep = False\n",
    "            if keep == True:\n",
    "                new_jet.append(splitting[0:2])\n",
    "                new_wghts.append(splitting[2])\n",
    "        if len(new_jet)>0:\n",
    "            new_data.append(new_jet)\n",
    "            weights.append(new_wghts)\n",
    "    return new_data, weights"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To calculate the transform we also need to flatten the data so that all splittings are in a single list."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "qcd_lund_0_cut, qcd_lund_0_wghts_cut = cut_to_range_w( qcd_lund_0, ((0,7),(-2,6)) )\n",
    "top_lund_0_cut, top_lund_0_wghts_cut = cut_to_range_w( top_lund_0, ((0,7),(-2,6)) )\n",
    "\n",
    "qcd_lund_0_flat_cut = flatten( qcd_lund_0_cut )\n",
    "top_lund_0_flat_cut = flatten( top_lund_0_cut )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now I define the transform, and then fit it using the a concatenation of the top and qcd jets."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note that because we're using a uniform strategy, the one transformer can be used on datasets with different amounts of signal and background.  However if we used another strategy which looks at the density of the data then we would need a new transformer for every dataset we use."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "transform = kbd( n_bins=[40,40], encode='ordinal', strategy='uniform' )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KBinsDiscretizer(encode='ordinal', n_bins=[40, 40], strategy='uniform')"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "transform.fit( qcd_lund_0_flat_cut + top_lund_0_flat_cut )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we use the transform to transform each event individually."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "qcd_lund_0_ord = [ transform.transform(i) for i in qcd_lund_0_cut ]\n",
    "top_lund_0_ord = [ transform.transform(i) for i in top_lund_0_cut ]\n",
    "qcd_lund_0_flat_ord = transform.transform( qcd_lund_0_flat_cut )\n",
    "top_lund_0_flat_ord = transform.transform( top_lund_0_flat_cut )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This gets the data into the 'ordinal' form, where it's the bin numbers that are specified:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 1., 12.],\n",
       "       [ 6., 15.],\n",
       "       [ 9.,  9.],\n",
       "       [11.,  6.],\n",
       "       [15., 20.],\n",
       "       [16.,  0.],\n",
       "       [20., 18.],\n",
       "       [22., 11.],\n",
       "       [27.,  5.],\n",
       "       [29.,  0.]])"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "qcd_lund_0_ord[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist2d( qcd_lund_0_ord[1][:,0], qcd_lund_0_ord[1][:,1], bins=40, range=((0,40),(0,40)) )\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "j=1\n",
    "plt.hist2d( top_lund_0_ord[j][:,0], top_lund_0_ord[j][:,1], bins=40, range=((0,40),(0,40)), \n",
    "           weights=top_lund_0_wghts_cut[j] )\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "But we need these in the one-hot encoded format:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "def ord2onehot(data_ord, lims):\n",
    "    img = np.zeros((lims[0],lims[1]))\n",
    "    for i in data_ord:\n",
    "        img[ lims[1] - 1 - int(i[1]), int(i[0]) ] = 1\n",
    "    return img\n",
    "\n",
    "def ord2onehot_w(data_ord, lims, weights):\n",
    "    img = np.zeros((lims[0],lims[1]))\n",
    "    for i in range(len(data_ord)):\n",
    "        img[ lims[1] - 1 - int( data_ord[i][1] ), int( data_ord[i][0] ) ] = weights[i]\n",
    "    return img"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[640.3985140886667,\n",
       " 637.4636631640077,\n",
       " 607.9230544971394,\n",
       " 446.8304232928609,\n",
       " 421.70911197237024,\n",
       " 351.2886465821619,\n",
       " 280.985523632888,\n",
       " 255.2015760903274,\n",
       " 213.7350875782884]"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "top_lund_0_wghts_cut[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x125f2e370>"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "j=1\n",
    "plt.imshow( ord2onehot_w( top_lund_0_ord[j], [40,40], top_lund_0_wghts_cut[j] ) )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can see above that the one-hot encoding has worked, below I just test the reshaping etc.."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "testimg = ord2onehot_w( top_lund_0_ord[j], [40,40], top_lund_0_wghts_cut[j] ).reshape(-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1600,)"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "testimg.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x127215b80>"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow( testimg.reshape((40,40)) )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we can one-hot encode all of the jets and save them out separately for QCD and tops."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I want to save them both as h5 and as npy."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "qcd_lund_0_ohw = []\n",
    "for jet,wght in zip(qcd_lund_0_ord,qcd_lund_0_wghts_cut):\n",
    "    qcd_lund_0_ohw.append( ord2onehot_w( jet, [40,40], wght ).reshape(-1) )\n",
    "qcd_lund_0_ohw = np.array( qcd_lund_0_ohw )\n",
    "\n",
    "top_lund_0_ohw = []\n",
    "for jet,wght in zip(top_lund_0_ord,top_lund_0_wghts_cut):\n",
    "    top_lund_0_ohw.append( ord2onehot_w( jet, [40,40], wght ).reshape(-1) )\n",
    "top_lund_0_ohw = np.array( top_lund_0_ohw )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "qcd_lund_0_ohw_2 = np.array( [ np.concatenate( ( qcd_lund_0_ohw[i], np.array([0]), [qcd_masses[i]] ), axis=-1 ) \n",
    "                                 for i in range(len(qcd_lund_0_ohw)) ] )\n",
    "top_lund_0_ohw_2 = np.array( [ np.concatenate( ( top_lund_0_ohw[i], np.array([0]), [top_masses[i]] ), axis=-1 ) \n",
    "                                 for i in range(len(top_lund_0_ohw)) ] )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save( \"qcd_lund_0_ohw.npy\", qcd_lund_0_ohw )\n",
    "np.save( \"top_lund_0_ohw.npy\", top_lund_0_ohw )\n",
    "\n",
    "qcd_out = pd.DataFrame( qcd_lund_0_ohw_2, columns=[str(i) for i in range(1600)]+[\"__signal_col__\"]+[\"__mass_Col__\"] )\n",
    "top_out = pd.DataFrame( top_lund_0_ohw_2, columns=[str(i) for i in range(1600)]+[\"__signal_col__\"]+[\"__mass_Col__\"] )\n",
    "\n",
    "qcd_out.to_hdf( \"qcd_lund_0_ohw.h5\", \"table\", format=\"table\", complib=\"blosc\", complevel=5 )\n",
    "top_out.to_hdf( \"top_lund_0_ohw.h5\", \"table\", format=\"table\", complib=\"blosc\", complevel=5 )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now to test loading them in."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, numpy:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "qcd_in = np.load(\"qcd_lund_0_ohw.npy\", allow_pickle=True)\n",
    "top_in = np.load(\"top_lund_0_ohw.npy\", allow_pickle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x12a3d1730>"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow( top_in[0].reshape((40,40)) )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now for h5:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_qcd_in = pd.read_hdf( \"qcd_lund_0_ohw.h5\", key=\"table\", start=0, stop=10000 )\n",
    "df_top_in = pd.read_hdf( \"top_lund_0_ohw.h5\", key=\"table\", start=0, stop=10000 )\n",
    "\n",
    "data_qcd = np.zeros( ( df_qcd_in.shape[0], 1600 ) )\n",
    "labels_qcd = np.zeros( df_qcd_in.shape[0] )\n",
    "masses_qcd = np.zeros( df_qcd_in.shape[0] )\n",
    "for i in range( df_qcd_in.shape[0] ):\n",
    "    data_qcd[i] = df_qcd_in.to_numpy()[i][0:1600]\n",
    "    labels_qcd[i] = df_qcd_in.to_numpy()[i][1600]\n",
    "    masses_qcd[i] = df_qcd_in.to_numpy()[i][1601]\n",
    "    \n",
    "data_top = np.zeros( ( df_top_in.shape[0], 1600 ) )\n",
    "labels_top = np.zeros( df_top_in.shape[0] )\n",
    "masses_top = np.zeros( df_top_in.shape[0] )\n",
    "for i in range( df_top_in.shape[0] ):\n",
    "    data_top[i] = df_top_in.to_numpy()[i][0:1600]\n",
    "    labels_top[i] = df_top_in.to_numpy()[i][1600]\n",
    "    masses_top[i] = df_top_in.to_numpy()[i][1601]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4920,)"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels_qcd.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(5080, 1600)"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_top.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x12eb10220>"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow( data_top[0].reshape((40,40)) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x12eca01c0>"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow( data_qcd[1].reshape((40,40)) )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It works!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This code has been summarised in lund2img.py"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Perfect!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now let's look at the averages with this weighted data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_plane_w( lund_jet_sample, plane_id, weight=9):\n",
    "    if weight==None:\n",
    "        return [ [ [i[6],i[7],1.0] for i in jet if i[0]==plane_id ] for jet in lund_jet_sample ]\n",
    "    else:\n",
    "        return [ [ [i[6],i[7],i[weight]] for i in jet if i[0]==plane_id ] for jet in lund_jet_sample ]\n",
    "\n",
    "qcd_lund_0 = get_plane_w( qcd_lund_jets, 0)\n",
    "qcd_lund_1 = get_plane_w( qcd_lund_jets, 1)\n",
    "qcd_lund_2 = get_plane_w( qcd_lund_jets, 2)\n",
    "qcd_lund_3 = get_plane_w( qcd_lund_jets, 3)\n",
    "\n",
    "top_lund_0 = get_plane_w( top_lund_jets, 0)\n",
    "top_lund_1 = get_plane_w( top_lund_jets, 1)\n",
    "top_lund_2 = get_plane_w( top_lund_jets, 2)\n",
    "top_lund_3 = get_plane_w( top_lund_jets, 3)\n",
    "\n",
    "qcd_lund_0_cut, qcd_lund_0_wghts_cut = cut_to_range_w( qcd_lund_0, ((0,7),(-2,6)) )\n",
    "top_lund_0_cut, top_lund_0_wghts_cut = cut_to_range_w( top_lund_0, ((0,7),(-2,6)) )\n",
    "\n",
    "qcd_lund_1_cut, qcd_lund_1_wghts_cut = cut_to_range_w( qcd_lund_1, ((0,7),(-2,6)) )\n",
    "top_lund_1_cut, top_lund_1_wghts_cut = cut_to_range_w( top_lund_1, ((0,7),(-2,6)) )\n",
    "\n",
    "qcd_lund_2_cut, qcd_lund_2_wghts_cut = cut_to_range_w( qcd_lund_2, ((0,7),(-2,6)) )\n",
    "top_lund_2_cut, top_lund_2_wghts_cut = cut_to_range_w( top_lund_2, ((0,7),(-2,6)) )\n",
    "\n",
    "qcd_lund_3_cut, qcd_lund_3_wghts_cut = cut_to_range_w( qcd_lund_3, ((0,7),(-2,6)) )\n",
    "top_lund_3_cut, top_lund_3_wghts_cut = cut_to_range_w( top_lund_3, ((0,7),(-2,6)) )\n",
    "\n",
    "qcd_lund_0_flat_cut = flatten( qcd_lund_0_cut )\n",
    "top_lund_0_flat_cut = flatten( top_lund_0_cut )\n",
    "\n",
    "qcd_lund_1_flat_cut = flatten( qcd_lund_1_cut )\n",
    "top_lund_1_flat_cut = flatten( top_lund_1_cut )\n",
    "\n",
    "qcd_lund_2_flat_cut = flatten( qcd_lund_2_cut )\n",
    "top_lund_2_flat_cut = flatten( top_lund_2_cut )\n",
    "\n",
    "qcd_lund_3_flat_cut = flatten( qcd_lund_3_cut )\n",
    "top_lund_3_flat_cut = flatten( top_lund_3_cut )\n",
    "\n",
    "qcd_lund_0_ord = [ transform.transform(i) for i in qcd_lund_0_cut ]\n",
    "top_lund_0_ord = [ transform.transform(i) for i in top_lund_0_cut ]\n",
    "\n",
    "qcd_lund_1_ord = [ transform.transform(i) for i in qcd_lund_1_cut ]\n",
    "top_lund_1_ord = [ transform.transform(i) for i in top_lund_1_cut ]\n",
    "\n",
    "qcd_lund_2_ord = [ transform.transform(i) for i in qcd_lund_2_cut ]\n",
    "top_lund_2_ord = [ transform.transform(i) for i in top_lund_2_cut ]\n",
    "\n",
    "qcd_lund_3_ord = [ transform.transform(i) for i in qcd_lund_3_cut ]\n",
    "top_lund_3_ord = [ transform.transform(i) for i in top_lund_3_cut ]\n",
    "\n",
    "qcd_lund_0_ohw = []\n",
    "for jet,wght in zip(qcd_lund_0_ord,qcd_lund_0_wghts_cut):\n",
    "    qcd_lund_0_ohw.append( ord2onehot_w( jet, [40,40], wght ).reshape(-1) )\n",
    "\n",
    "top_lund_0_ohw = []\n",
    "for jet,wght in zip(top_lund_0_ord,top_lund_0_wghts_cut):\n",
    "    top_lund_0_ohw.append( ord2onehot_w( jet, [40,40], wght ).reshape(-1) )\n",
    "\n",
    "qcd_lund_1_ohw = []\n",
    "for jet,wght in zip(qcd_lund_1_ord,qcd_lund_1_wghts_cut):\n",
    "    qcd_lund_1_ohw.append( ord2onehot_w( jet, [40,40], wght ).reshape(-1) )\n",
    "\n",
    "top_lund_1_ohw = []\n",
    "for jet,wght in zip(top_lund_1_ord,top_lund_1_wghts_cut):\n",
    "    top_lund_1_ohw.append( ord2onehot_w( jet, [40,40], wght ).reshape(-1) )\n",
    "    \n",
    "qcd_lund_2_ohw = []\n",
    "for jet,wght in zip(qcd_lund_2_ord,qcd_lund_2_wghts_cut):\n",
    "    qcd_lund_2_ohw.append( ord2onehot_w( jet, [40,40], wght ).reshape(-1) )\n",
    "\n",
    "top_lund_2_ohw = []\n",
    "for jet,wght in zip(top_lund_2_ord,top_lund_2_wghts_cut):\n",
    "    top_lund_2_ohw.append( ord2onehot_w( jet, [40,40], wght ).reshape(-1) )\n",
    "\n",
    "qcd_lund_3_ohw = []\n",
    "for jet,wght in zip(qcd_lund_3_ord,qcd_lund_3_wghts_cut):\n",
    "    qcd_lund_3_ohw.append( ord2onehot_w( jet, [40,40], wght ).reshape(-1) )\n",
    "\n",
    "top_lund_3_ohw = []\n",
    "for jet,wght in zip(top_lund_3_ord,top_lund_3_wghts_cut):\n",
    "    top_lund_3_ohw.append( ord2onehot_w( jet, [40,40], wght ).reshape(-1) )\n",
    "    \n",
    "qcd_lund_0_ohw_flat = np.sum(qcd_lund_0_ohw,axis=0)/np.sum(np.sum(qcd_lund_0_ohw))\n",
    "top_lund_0_ohw_flat = np.sum(top_lund_0_ohw,axis=0)/np.sum(np.sum(top_lund_0_ohw))\n",
    "\n",
    "qcd_lund_1_ohw_flat = np.sum(qcd_lund_1_ohw,axis=0)/np.sum(np.sum(qcd_lund_0_ohw))\n",
    "top_lund_1_ohw_flat = np.sum(top_lund_1_ohw,axis=0)/np.sum(np.sum(top_lund_0_ohw))\n",
    "\n",
    "qcd_lund_2_ohw_flat = np.sum(qcd_lund_2_ohw,axis=0)/np.sum(np.sum(qcd_lund_0_ohw))\n",
    "top_lund_2_ohw_flat = np.sum(top_lund_2_ohw,axis=0)/np.sum(np.sum(top_lund_0_ohw))\n",
    "\n",
    "qcd_lund_3_ohw_flat = np.sum(qcd_lund_3_ohw,axis=0)/np.sum(np.sum(qcd_lund_0_ohw))\n",
    "top_lund_3_ohw_flat = np.sum(top_lund_3_ohw,axis=0)/np.sum(np.sum(top_lund_0_ohw))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 8 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig,axs = plt.subplots(2,4,figsize=(12,6))\n",
    "bins = (50,50)\n",
    "rnge = ((0,7),(-2,5))\n",
    "cmap = 'gist_heat_r'\n",
    "\n",
    "xt = range(0,40,5)\n",
    "yt = range(0,40,5)\n",
    "\n",
    "labelfont = FontProperties()\n",
    "labelfont.set_family('serif')\n",
    "labelfont.set_name('Times New Roman')\n",
    "labelfont.set_size(20)\n",
    "\n",
    "axislabelfont = FontProperties()\n",
    "axislabelfont.set_family('serif')\n",
    "axislabelfont.set_name('Times New Roman')\n",
    "axislabelfont.set_size(24)\n",
    "\n",
    "tickfont = FontProperties()\n",
    "tickfont.set_family('serif')\n",
    "tickfont.set_name('Times New Roman')\n",
    "tickfont.set_size(20)\n",
    "\n",
    "axisfontsize = 18\n",
    "labelfontsize = 18\n",
    "axs[0,0].imshow( qcd_lund_0_ohw_flat.reshape((40,40)), cmap=cmap )\n",
    "axs[0,1].imshow( qcd_lund_1_ohw_flat.reshape((40,40)), cmap=cmap )\n",
    "axs[0,2].imshow( qcd_lund_2_ohw_flat.reshape((40,40)), cmap=cmap )\n",
    "axs[0,3].imshow( qcd_lund_3_ohw_flat.reshape((40,40)), cmap=cmap )\n",
    "axs[1,0].imshow( top_lund_0_ohw_flat.reshape((40,40)), cmap=cmap )\n",
    "axs[1,1].imshow( top_lund_1_ohw_flat.reshape((40,40)), cmap=cmap )\n",
    "axs[1,2].imshow( top_lund_2_ohw_flat.reshape((40,40)), cmap=cmap )\n",
    "axs[1,3].imshow( top_lund_3_ohw_flat.reshape((40,40)), cmap=cmap )\n",
    "\n",
    "tpos1 = [37,1]\n",
    "tpos2 = [37,6]\n",
    "tpos3 = [37,11]\n",
    "\n",
    "axs[0,0].text( tpos2[0], tpos2[1],  \"primary\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[0,1].text( tpos2[0], tpos2[1],  \"secondary\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[0,2].text( tpos2[0], tpos2[1],  \"tertiary\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[0,3].text( tpos2[0], tpos2[1],  \"quaterniary\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[0,0].text( tpos1[0], tpos1[1],  \"QCD\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[0,1].text( tpos1[0], tpos1[1],  \"QCD\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[0,2].text( tpos1[0], tpos1[1],  \"QCD\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[0,3].text( tpos1[0], tpos1[1],  \"QCD\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[1,0].text( tpos2[0], tpos2[1],  \"primary\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[1,1].text( tpos2[0], tpos2[1],  \"secondary\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[1,2].text( tpos2[0], tpos2[1],  \"tertiary\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[1,3].text( tpos2[0], tpos2[1],  \"quaterniary\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[1,0].text( tpos1[0], tpos1[1],  \"top\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[1,1].text( tpos1[0], tpos1[1],  \"top\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[1,2].text( tpos1[0], tpos1[1],  \"top\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[1,3].text( tpos1[0], tpos1[1],  \"top\", va='top', ha='right', fontproperties=labelfont )\n",
    "\n",
    "axs[0,3].text( tpos3[0], tpos3[1],  \"weight=$\\kappa$\", va='top', ha='right', fontproperties=labelfont )\n",
    "axs[1,3].text( tpos3[0], tpos3[1],  \"weight=$\\kappa$\", va='top', ha='right', fontproperties=labelfont )\n",
    "\n",
    "axs[0,0].set_xticks( xt ) \n",
    "axs[0,1].set_xticks( xt ) \n",
    "axs[0,2].set_xticks( xt ) \n",
    "axs[0,3].set_xticks( xt ) \n",
    "axs[1,0].set_xticks( xt ) \n",
    "axs[1,1].set_xticks( xt ) \n",
    "axs[1,2].set_xticks( xt ) \n",
    "axs[1,3].set_xticks( xt )\n",
    "axs[0,0].set_yticks( yt ) \n",
    "axs[0,1].set_yticks( yt ) \n",
    "axs[0,2].set_yticks( yt ) \n",
    "axs[0,3].set_yticks( yt ) \n",
    "axs[1,0].set_yticks( yt ) \n",
    "axs[1,1].set_yticks( yt )\n",
    "axs[1,2].set_yticks( yt ) \n",
    "axs[1,3].set_yticks( yt ) \n",
    "axs[1,0].set_xticklabels( xt, fontproperties=tickfont )\n",
    "axs[1,1].set_xticklabels( xt, fontproperties=tickfont )\n",
    "axs[1,2].set_xticklabels( xt, fontproperties=tickfont )\n",
    "axs[1,3].set_xticklabels( xt, fontproperties=tickfont )\n",
    "axs[0,0].set_yticklabels( yt, fontproperties=tickfont )\n",
    "axs[1,0].set_yticklabels( yt, fontproperties=tickfont )\n",
    "axs[0,0].set_xticklabels([])\n",
    "axs[0,1].set_xticklabels([])\n",
    "axs[0,2].set_xticklabels([])\n",
    "axs[0,3].set_xticklabels([])\n",
    "axs[0,1].set_yticklabels([])\n",
    "axs[0,2].set_yticklabels([])\n",
    "axs[0,3].set_yticklabels([])\n",
    "axs[1,1].set_yticklabels([])\n",
    "axs[1,2].set_yticklabels([])\n",
    "axs[1,3].set_yticklabels([])\n",
    "fig.text( -0.03, 0.5, r'$\\log k_T$', va='center', rotation='vertical', fontproperties=axislabelfont )\n",
    "fig.text( 0.5, -0.05, r'$\\log 1/\\Delta R$', ha='center', fontproperties=axislabelfont )\n",
    "fig.tight_layout()\n",
    "fig.savefig(\"qcd_top_all_lund_wkap.pdf\", bbox_inches = \"tight\")\n",
    "#fig.show()a"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_qcd_in = pd.read_hdf( \"../../toptagging_data/qcd_lund0_ohw_pt_100k.h5\", key=\"table\", start=0, stop=10000 )\n",
    "df_top_in = pd.read_hdf( \"../../toptagging_data/top_lund0_ohw_pt_100k.h5\", key=\"table\", start=0, stop=10000 )\n",
    "\n",
    "data_qcd = np.zeros( ( df_qcd_in.shape[0], 1600 ) )\n",
    "labels_qcd = np.zeros( df_qcd_in.shape[0] )\n",
    "masses_qcd = np.zeros( df_qcd_in.shape[0] )\n",
    "for i in range( df_qcd_in.shape[0] ):\n",
    "    data_qcd[i] = df_qcd_in.to_numpy()[i][0:1600]\n",
    "    labels_qcd[i] = df_qcd_in.to_numpy()[i][1600]\n",
    "    masses_qcd[i] = df_qcd_in.to_numpy()[i][1601]\n",
    "    \n",
    "data_top = np.zeros( ( df_top_in.shape[0], 1600 ) )\n",
    "labels_top = np.zeros( df_top_in.shape[0] )\n",
    "masses_top = np.zeros( df_top_in.shape[0] )\n",
    "for i in range( df_top_in.shape[0] ):\n",
    "    data_top[i] = df_top_in.to_numpy()[i][0:1600]\n",
    "    labels_top[i] = df_top_in.to_numpy()[i][1600]\n",
    "    masses_top[i] = df_top_in.to_numpy()[i][1601]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x13a671640>"
      ]
     },
     "execution_count": 146,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow( data_qcd[1].reshape((40,40)) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10000, 1600)"
      ]
     },
     "execution_count": 141,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_top.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10000,)"
      ]
     },
     "execution_count": 142,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels_top.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10000,)"
      ]
     },
     "execution_count": 143,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "masses_top.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Looks like it works!  This was all saved in lund2img.py, which has been updated as described below."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "UPDATE:  The lund2img_wght.py code has now been updated to include more than one Lund plane.  In the script you can specify how many of the planes to include using the 'planes' variable.  I've tested this and saved the .h5 files here as \"qcd_lund_p0_ohw.h5\", \"qcd_lund_p1_ohw.h5\", \"qcd_lund_p2_ohw.h5\", \"qcd_lund_p3_ohw.h5\"."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "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": "3.8.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}