{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "663642d6",
   "metadata": {
    "toc": true
   },
   "source": [
    "<h1>Table of Contents<span class=\"tocSkip\"></span></h1>\n",
    "<div class=\"toc\"><ul class=\"toc-item\"><li><span><a href=\"#Guided-Hunting---Anomaly-detection-with-Isolation-Forest-on-Windows-Logon-data\" data-toc-modified-id=\"Guided-Hunting---Anomaly-detection-with-Isolation-Forest-on-Windows-Logon-data-1\">Guided Hunting - Anomaly detection with Isolation Forest on Windows Logon data</a></span><ul class=\"toc-item\"><li><span><a href=\"#Notebook-initialization\" data-toc-modified-id=\"Notebook-initialization-1.1\">Notebook initialization</a></span><ul class=\"toc-item\"><li><span><a href=\"#Authentication-to-LA-Workspace\" data-toc-modified-id=\"Authentication-to-LA-Workspace-1.1.1\">Authentication to LA Workspace</a></span></li></ul></li><li><span><a href=\"#Data-Preparation\" data-toc-modified-id=\"Data-Preparation-1.2\">Data Preparation</a></span><ul class=\"toc-item\"><li><span><a href=\"#Historical-Data-Processing\" data-toc-modified-id=\"Historical-Data-Processing-1.2.1\">Historical Data Processing</a></span></li></ul></li><li><span><a href=\"#Feature-Engineering\" data-toc-modified-id=\"Feature-Engineering-1.3\">Feature Engineering</a></span></li><li><span><a href=\"#Data-Modelling\" data-toc-modified-id=\"Data-Modelling-1.4\">Data Modelling</a></span><ul class=\"toc-item\"><li><span><a href=\"#Isolation-Forest-Anomaly-detection\" data-toc-modified-id=\"Isolation-Forest-Anomaly-detection-1.4.1\">Isolation Forest Anomaly detection</a></span></li></ul></li><li><span><a href=\"#Data-Visualization\" data-toc-modified-id=\"Data-Visualization-1.5\">Data Visualization</a></span><ul class=\"toc-item\"><li><span><a href=\"#Histogram\" data-toc-modified-id=\"Histogram-1.5.1\">Histogram</a></span></li><li><span><a href=\"#3D-ScatterPlot-using-PCA\" data-toc-modified-id=\"3D-ScatterPlot-using-PCA-1.5.2\">3D ScatterPlot using PCA</a></span></li><li><span><a href=\"#Correlation-Plot\" data-toc-modified-id=\"Correlation-Plot-1.5.3\">Correlation Plot</a></span></li></ul></li><li><span><a href=\"#Interpreting-Global-and-Local-anomalies-with-SHAP\" data-toc-modified-id=\"Interpreting-Global-and-Local-anomalies-with-SHAP-1.6\">Interpreting Global and Local anomalies with SHAP</a></span><ul class=\"toc-item\"><li><span><a href=\"#Global-Interpratation---Summary-plot-for-Feature-Importance\" data-toc-modified-id=\"Global-Interpratation---Summary-plot-for-Feature-Importance-1.6.1\">Global Interpratation - Summary plot for Feature Importance</a></span></li><li><span><a href=\"#Local-Interpretation---Feature-importance\" data-toc-modified-id=\"Local-Interpretation---Feature-importance-1.6.2\">Local Interpretation - Feature importance</a></span><ul class=\"toc-item\"><li><span><a href=\"#Example-Outlier\" data-toc-modified-id=\"Example-Outlier-1.6.2.1\">Example Outlier</a></span></li><li><span><a href=\"#Example-Inlier\" data-toc-modified-id=\"Example-Inlier-1.6.2.2\">Example Inlier</a></span></li></ul></li><li><span><a href=\"#Populate-dataset-with-SHAP-values\" data-toc-modified-id=\"Populate-dataset-with-SHAP-values-1.6.3\">Populate dataset with SHAP values</a></span></li></ul></li><li><span><a href=\"#Conclusion\" data-toc-modified-id=\"Conclusion-1.7\">Conclusion</a></span></li></ul></li></ul></div>"
   ]
  },
  {
   "attachments": {
    "aff21132-09c0-40ea-a96b-555ac4d81aed.png": {
     "image/png": "iVBORw0KGgoAAAANSUhEUgAABRwAAAMMCAIAAAC+BvyHAAAgAElEQVR4nOzdWXMcSXrueXePPTJyTySQ2LhVd0st6ZhkOrK5mJv5aPpsYzN2bGaOpKOtW1IVSRB7JpB7xh7hPheo3qpAFoniBuD/qzJaLYyiRzGBzCdef1+XxhgBAAAAAAA+nPrSCwAAAAAA4L4iVAMAAAAAcEeEagAAAAAA7ohQDQAAAADAHRGqAQAAAAC4I0I1AAAAAAB3RKgGAAAAAOCOCNUAAAAAANwRoRoAAAAAgDsiVAMAAAAAcEeEagAAAAAA7ohQDQAAAADAHRGqAQAAAAC4I0I1AAAAAAB3RKgGAAAAAOCOCNUAAAAAANwRoRoAAAAAgDsiVAMAAAAAcEeEagAAAAAA7ohQDQAAAADAHRGqAQAAAAC4I0I1AAAAAAB3RKgGAAAAAOCOCNUAAAAAANwRoRoAAAAAgDsiVAMAAAAAcEeEagAAAAAA7ohQDQAAAADAHRGqAQAAAAC4I0I1AAAAAAB3RKgGAAAAAOCOCNUAAAAAANwRoRoAAAAAgDsiVAMAAAAAcEeEagAAAAAA7ohQDQAAAADAHRGqAQAAAAC4I0I1AAAAAAB3RKgGAAAAAOCOCNUAAAAAANwRoRoAAAAAgDsiVAMAAAAAcEeEagAAAAAA7ohQDQAAAADAHRGqAQAAAAC4I0I1AAAAAAB3RKgGAAAAAOCOCNUAAAAAANwRoRoAAAAAgDsiVAMAAAAAcEeEagAAAAAA7ohQDQAAAADAHRGqAQAAAAC4I0I1AAAAAAB3RKgGAAAAAOCOCNUAAAAAANwRoRoAAAAAgDsiVAMAAAAAcEeEagAAAAAA7ohQDQAAAADAHRGqAQAAAAC4I0I1AAAAAAB3RKgGAAAAAOCOCNUAAAAAANwRoRoAAAAAgDsiVAMAAAAAcEeEagAAAAAA7ohQDQAAAADAHRGqAQAAAAC4I0I1AAAAAAB3RKgGAAAAAOCOCNUAAAAAANwRoRoAAAAAgDsiVAMAAAAAcEeEagAAAAAA7ohQDQAAAADAHRGqAQAAAAC4I/tLLwAAADxKRhutha6N0UJrY7QwRgjzux+FEOZ3P1UKKYWUN38hpRRSCaWksoRUUioh5Re8DwDAI0eoBgAAn5cxQpg6i3W2ruNlHS90vqmzxFS5KQtTl6IqTV0ZXQshhFTSsqVtS8uRtisdV7mh8htW2LLCjgpblh9JZZOrAQBfCqEaAAB8esYYo01VmrowVSGqsooX1WZWLSfVclzHiypemiLVRWbK3JS5qUpTl0IIIS3lONLxpO1K11duYAVNq9GxWwO7s21HfRN1pRNI25G2K21XKiUk3W0AgM9HGmN++mcBAAD8DKbMdZGWi3E5O6uWk2o50UVmqlxnic5iXWa6ysVNdVrXRtffbwgXQgoplCWVEsoSypKWrWxHOb70QuU1lOtL27ObPbu15fR2nf6e8hrKC7/07QIAHhEq1QAA4NP4vjpd6CLVyarazIurN8X4ZTE9K6dnQmshhfi+rdrcROhb/hvCiFqb+g//pBZSSvV9zBZCCGG3Bk5/143nusisZs+OusoNpBt8330NAMCnRKUaAAB8EqauTJmXi4v84rtyelYtxlU818myzmKdJ+LmE4gxQtx8GPmADyTyd0PLvv9b17e8UAUtK2zbnaHd2fFHL9ztF8rxpO1+gjsDAOAPqFQDAICPzNSlLlKdrKvNrJgcZae/Ka5P6+VEl7kw+uc/0Dd/mBAuhBAmi02eiPVMSmXNB3bn3JSZqWs76llRR7mBdDwhmGQGAPgkqFQDAICPyug6WRaTN/n4VXH5slyM63ih81gXmdG1+HQfPKSUQkrXV65vhR0r6no7z72dF+7WU6c3+uPKNgAAHxGVagAA8DHcdFAXab2Zl/OL7OLb4vJlfvmyTlY3/+oTxunfL0AYU6SmzOp4KWdnukh0keoyM1VuNboqaDIbHADw0VGpBgAAH4HRtSnzcn6RvPyH/OK7enVVxQudrnVVCCE+eaL+E1JIIaVUQdMKWnZ7y+nths//1tv7lXJ8aTufcSUAgIePSjUAAPh5jDF1qbO4mJ3lF99lx/9WXB3rPDFV8TkK1LctSBhhjNHp2uRpna7qzVzarhDG6e3Zzb6wbPm7yeEAAPxMhGoAAPCzGKN1npSzs/g//u/s9Lf1eqaz2Oj6CyXqP1qYNsJUJkuqukq+/f+q+UXjz/53abtWEAlCNQDgIyFUAwCAOzOmrnW2ycevstP/yM//q7w+MXX9UUZ8fwzGGCNqo3VtqktTZipsCym97Rd2eygti/5qAMDPR6gGAAB3ZYwp82o9TV7+Q/rqn6rNXFeVEObLFqh/xBgjTF2ZZJm++qc6XkrLUUFTeYG0CNUAgJ+LUA0AAO7EGF3mxdVRfvafxeWrcjEWuhJGf+ll3coII0xVVuuZUFZ28u9SWe72c6vZk1JyhDUA4OcgVAMAgLswutZ5kh3/e/zt/1Mtr4Suv44t329njDC1Thbp6/9likwFTavRFsomUwMAfg5CNQAA+GBG1+Xispgc5ZOjajHWRSa+9Fiy92GM0WVRrafF9XF2/l/CctzujvKjL70uAMA9RqgGAAAfrq6Kyevk5T8W09M6T+9BmfqGMUbXusjL1XV2/G9CCMtvEKoBAD8HoRoAAHwQo9NNtZkVV8fF5KiOF+Lm9Kz7wgija50l5excBZHb31eur/xIWs6XXhkA4F4iVAMAgA9hRJUsi+uT8vqknJ2ZqjS6/tJr+iBGGGPKtFqOlRuUWydW2HJsj1ANALgbQjUAAHhvWpu6LCZH6at/LGfnpizuU436jxhjRFVWm1l68hshLeVH0vGkVEIytQwA8GEI1QAA4H0ZU+syL8av4m//X1Nk961G/UeMMcbUm3n25l+lVN7oF3bUE7bD8VoAgA9FqAYAAO+rXEyK6+NycWnK/B4n6t8xWpuqrDaz/PJbYVluf1+5wZdeFADgniFUAwCA91UuLtLjfykXY1PXRt/Ljd9/ypi6ruNlev5f0vWd1kAQqgEAH0h96QUAAIB7wNSVLrJqMSnOv6vXM6NrIe7DGVrvZoQxWqfrcvKmvD6t042pyq//tG0AwFeFSjUAAPhppq5MnlaLcXH5yuha3M/5ZD9ihDE6i02Rle2hTlYm6kllMa4MAPD+CNUAAOCnVevr4upNtboSxjyEGvUPGZ2s8suXwrbdrSdK+V96PQCAe4Pt3wAA4KfVq+vs9LfV6lqIh5epjTGiTtf5+GU5OzdV+aXXAwC4T6hUAwCAdzPCiGozKy6+rdfXRtfmYaVqY4QUuk5XxeSN09kWVSGE4WwtAMB7IlQDAIB30trout7Mi6tjXebmgXRT/zFjjNFZXNZltbzSZaZ0LZVFrgYAvA+2fwMAgHfRRVqtrnS2EeaBlah/xBhdpNV6qtP1gzgwDADwORCqAQDAu9R5XC4n9fc580Gnam1uniDUyfKhjDcHAHxyhGoAAPAuOtuUs7M6npu6Eg+3fmuMNrrW6aacX1Sb+QO+UwDAx0WoBgAA73KTM+t4KXRtzMOtVBshTK3zuFqM63hhdP2lFwQAuB8I1QAA4F10nlSLcZ2ujXiQJ1T/njHmdw3kbP8GALw3QjUAAHgXXaTVeqbzRDzgMvX3jC7zajOvs5hBZQCA90SoBgAA72KqUqcbU+VfeiGfg6lKk8WmyB50TR4A8DERqgEAwDvpypSpqaovvY7PQte6zE1dPoKyPADg4yBUAwCAdzFam7oS5nEM7jLa1BVTygAA749QDQAA3s4YYbTR9SOp3BohhNZCa2Ee+qHcAICPhFANAADewhhhtDDmEaVLY4Qwwhih9SN5jgAA+JkI1QAA4N2MEOYhn1B9C/P7HwAAeDdCNQAAeAcphHxU8fL7xwdSCim/9FoAAPcAoRoAAPyUR5gujaBUDQB4H4RqAADwNkYY89hai+X3BerH1EkOAPgZCNUAAOBt5Pe7oKWUj2ov9O/u+kuvAwBwD9hfegEAAOBr9X22fEyP4H8fp6UiVAMA3sdjepsEAAB3oKS07EcSraWUUllSWiRqAMB7ehRvkAAA4M6ktKTlCPU4PjNIJSxbKOtLrwMAcG88jjdIAABwV9J2lBcq2/3SC/kcpLKVF0rHo1INAHhPhGoAAPAu0nJUEEnH+9IL+SxsW3kNRagGALw3QjUAAHgX5Yd2e6j86BEMAJfKDezWQIUtKdkBDgB4L4RqAADwLsprOJ1tK2gKoYR4wLlaSimVG9jtLStsP5YecgDAz8YbBgAAeBflR05vz2p0pWXJBxw1pRTKUn7kdHesqPuQ7xQA8FHxhgEAAN5FBZHT27WijrDsB9xpfHOYluVHdmdkR13B9m8AwPshVAMAgHdRbmi3tpTfkFI+6O3fQkihXN9u9lXQpFINAHhP9pdeAAAA+Kop11eOZ/lNabtS16Yuv/SKPg2lpOUor2FHXcuLHnBNHgDwcfEUFgAAvJsUUlrNnrf7C7vZl8p6cGPApZTKClru8Knd3ZG2+7Dr8QCAj4tQDQAAfpK0m31v95d2cyDlQ5sBLqWQUllB0x0+dbojabkP7AYBAJ8UoRoAAPw0uznw9//cbm0JKR7c1mgppLCClrfzwuntSdv50usBANwn9FQDAICfZoUd6QZ2+9+kG4oyM1oLYb70oj4KKZQlXd+Kuu7gwG4PpcWnIwDAB6BSDQAAfpq0bOX4dmfb2/2F1exLpR5EvVoKKVUQOVuH9mBfBU1pOw/ivgAAnw+hGgAAvAelpO3YnW1v/1d2ayCUJeX9/xRxczZ10HK3n7lbB8pvSGXRUA0A+CBscAIAAO/L6e4IqerFJL/41hj9pZfzc0kppFJWo+3v/tLbeqoc/0uvCABw/xCqAQDA+7LCjnL8vL/ndLarzbxOVsLc285qKaXj2VHP7e+5g0O7RTc1AOAuePMAAADvSypL2J67/TzMk/TNv+osFlrfy5K1lFIpu9H1D//SP/gLq9GWlkU3NQDgDgjVAADgvUkpLdvp7gpj6s2sWox1npgyv2/1aiktW3mh3Rl6u790h0+VF4kH0CIOAPgSCNUAAOBDSGk32lLKYnDozC/LxdhUpRH6/uRqKZVSjm+3h87g0B0c2K0tzqYGANwZoRoAAHwY6fhWQ7lbT+pkZbTW6UbUpamrL72u9yKVlJZthW13+Mzb+cZuDZQX/uRVtTa1NjcPDixLWYqN4gCA7xGqAQDAB5OW7Q6fSdvTyapaXOrcCF2br75YLaWUylKub7cGweFf+ge/VkHrfS4sqzrNa2OMMSL07cDjExQA4Hu8JQAA8NgZIYwRZaWzopJCuraybWkp9a65XVJZYUtK6e3+QmfrYvKmXE6Err/uTeBSSCW90N164u//mbt1aDf7Ulnvc+VsmZ5P1nGSZ3m1M4iG/ajZ8BqBy2gzAAChGgCAx84YobWJ0/Jqnikp2pHbCOzAk/KdkVFKpbwwOPxL5Ya6Kqv1VBhjTP3Zlv3BpBDKssJO8Pxvgqd/bbeHUr3vxO+L8ep//tvJ2eXyarb59S92fv3NzjdPBkHgKsPIcAB47AjVAAA8XnVtsqLapNUqLiez9HQSu47aHYSjQTgaWOrdncNSSsuxmn3XmGA5NnVVTk/rzVzcbJIWX1PJWkoplXR9t7/vjX7h7XzjdEfS8d4nEGd5lebl6Xj57euryXQzWyaWpYqi3iT5cp32O2G76fuubdu0WQPAI0WoBgDg8SoqPVsVF9fJm4vNm4vNm4t1M3R++aQtpRx0fMf+qVOmpFSOZzf7wYu/VUEr/u3/mWUbU1dS11/TNnAppZSWZTfawfO/Dp/9jdMdKdd/zzO01kkxmW5OL5cnF4u8qIzW55eL+SI5Gy9fnUz/6pe7v3g66HeCyHY/9W0AAL5OhGoAAB4XY0StdVHqVVxeL7KzSXw6ic8nyXiWXi2yOK1sWzUCZ6vrDzp+I7Df2VothFTS8ZzWljBGZ2vpBuX1WbWZiaoUujZGf67bunVtUkolbUd5odPf97af+Xt/7vT2lBeK92ilrmtT1eZ8vPrXby+Oz+dJWmhtjBFZXhVlXWuT5aWuzWqTHow624OoHfmh79qW/IkKPwDgYSFUAwDwuBhjqkqvNsXri83rs/Xrs/XFdbKMiyyvq9oYUx1fxlHgDLuBpWTgWcr6iYgolSW9htPbVW5gd3bi3/5fpi50GusqF787huqLkFJJy1Zew25vhd/8Xfjsr62oq/xIvl+Nuqx1mtevz+b/45/ezOax1vr3t6K12cR5nperdXZ8PvvV8+E3T7ee7fe3+yr0bUI1ADwqhGoAAB6LWps4rdZxcb3IL6fJ0fn6dJJcLbLVpsiLutZGCFHVOsmqq0V2dL4OPKvb9AJf2NZPpVCllOOJqOcZbcrMbg7K+Xm5GFera1Nk5qa/+jOla/m7P5TdGtjtLac3cvr7/t6v7PZQ2u57jvsWQiRpMZkmk+l6Oo/TtPjB8rXWRan1xlR1LaXcJMV8me5tt0dbzW4raISuY1tKCeI1ADx4hGoAAB6LujbzVX46jl+drY7ON6eTeL7O69poI35/xLTWpjT1YpUfna87Tfdgp/HTx2vdkEo5nuzsNKKet/MiO/ltdvobU2R1VQqjhTHm84wuk0JKJaSUynZ6I//g1/7+n3vbz6XjScv5oIy7jvPT8fJ6HqdpUVW3TDU3RlR1vYn163Q6vlqfjZeHu90/f7H9dL+3o1rNUMmfGKAOAHgIrL//+7//0msAAACfUFHqJK8ms/T4cvOfR8v/fLM8vogvZ+kqLvJC31o/lkJUtWlFzrDne64deNZ7pUMppVLSsqVlS8dTfmQ3ulbYkpYtpDS6lsZ8utKtlFIqywoiq9n3dr4Jnv634OAvvNE3Tm9kNdrSujk96wN++SyviqqOkyIvKsuybp47aP3D/183B5LV2lSVzosqScs4LfK8KsraGCOltC1FtgaAB4xQDQDAQ2aMSPNquc5fn61/+2r+by/n/3m0HM/TVVxW1VuniNXapHndarjbvaAROM3Q+YBYKKW0XavRcTo77tYTK2yZqjBVbopE6FoY8ykSppRSSCWVZbcGbn8//Obvmn/xf3i73zi9PctrCKk+KE7fMELYtqW1kVIYLbK8qrUubytZCyG0NnlRbeL8eh7PV0mal7rWtmO7juV7tkWXNQA8XGz/BgDgwVqsi8U6H8/Si+v05HJzdhVPF3mSV1Wtb69QC+E5VhhYncjrtb1fPWmPtsIPS9Q3pJTSEo6npHIHh0JZ7uCgWl5Vq6tqeVXH8ypeft9ifbMz/MPOtZZSSnHzp5BCGOVHdqNjtwZ2a8tqb9ntLbe3bzXa0nbev4P6xzzHakfes71uFDrDXjToNSbX68l0vYnzOC3N98v+g5sJcJkoZ/NYGJGmxWKd7g3bu8NWvxt2WqFrK+unpr4BAO4dQjUAAA+TMWK+zo/O1y9P1q/O1pezZLbMf/Iqz1X9tv90FL3Ybz3fbx5sRz99WvVbSGVJ11L9Pae/p/O4jpf5+FV2+pti/FrnmdGVlNLUtdC1ENq89xgzKYWUUihLKktIaYyxo547fOLv/5m3/+d2o2s12neoS/+Y61iuY7Ua7rP97k6/Oew3vzu+tl3rYrzKirqub1mzMaYs62WVrjbZ9XxzcbUaj7qL1eDZQd91XRU6ip3gAPDgEKoBAHhQ8qJOi3q2zGfL7NXZ+vX5errMZ8s8zaq3XaKktCwZBU6v7Y4G4cF2Y3cr3OmHvbb3sfYtS9tVYcvdeqK80Nv5pt7MdZ7oPK6TlU6WdZ7oIjFVaarS1JXQlakrobUQQkgpLVsoS1q2tB1pOdINLC9UQdMK28pvKDe0Gm3rplLd6Eo3+CiJ+gfaLf+J6Hie3e9GJ+fz04vFZLaeL5OiqOv6h7vojRFCmDyv5svEaBEnxWyZXC/i0SAa9qNmw4tCl2gNAA8GoRoAgAfCGGGESfJ6vspfna6+O129OlsfXWzqWtf1W+vAUgrLkp5j9dreNwfNmwJ1v+23Gu5H7ASWlmNZjuVH7uBAGCOMrjbzcn1dzS/K+UW9mVXxQuexyVNTZLrMTZmbqhLi+8O6pONJ11NuoLzQCttW1LU72053ZLe2nOZA2o54v6On76wd+e3I73Ua+zvdQafRaQb/+VoVZSVEkWVGiB8WrY0ReVHlRbXZ5OPr9WS2uZptXjzpV7UebTUDz7YsJoMDwANBqAYA4CEwRizW+XSZn13FZ5Pk/Co+v0oW66KutX7rPDLhOqrVcPttb9gL9oeNp7vRsBd0mm7g2Z9wtJYUQijlNxwlLS90Otu6yEyZ3ZSpja5EXRtd/6FSrSxhWVLZwrKlbSvbk66vvIbyI+WF0rI/RWn6Vje7wQ9321HD6bSD3e3O2eXi8mq1WKVxUty6g10bU1b1cpXeTDJbLNO9nfb+Tntnq7UzaKqbxnAAwH1GqAYA4B4zRhhjam3q2lzNs5enq1dn69dn6+kyn6/f2kF905aspGwEzk4/ONyJXuw3D3eivWEj9D/DZwMppFBuoNxARP1P/8t9NK6jXEeFgTPsR1u95u6w/R9N33VtIeZVpavqps/6T6K11kbrerXJNkm+Wqfj6/XVbDNbpmVtmpHv2sqxlbrpEgcA3E+EagAA7rG8qNdJMV3mV/Ps6Hz98nQ9X+XLTZEVt5/8JIRQUrqOFQX2oOvvDsKnu9HuVjjo+J2m5951JtmjIqWwlGw1XCWbtq0G3Wh32Do5X5xeLmaLuChv7bI2Wossr7RJjDCbtFht0rPLxcFOe2+n3Y78KHS/yL0AAH4+QjUAAPePMUIbU9V6GRfjafLmMn59tj662BxfrKt3t08r6TpWM3S2OsHz/ebzvebT3Wir6/uuzWlP70kKYSnZCJzAd1rNYHvQ6ncbnVaolBRCrNZpmpd1fcuBW0VZFWWV5dV8mUznm6PT2V/+cqfWZjRsWZZ0bMu2eKgBAPcPoRoAgHtGG1PXYrkpJvPkbJIcX2wupslkli03xTvap4UQllL9jrfVCXa3wr2t8GCnMewFrYbrOpYizX04KYRjyWZoH2y3osDutYPj8/nL4+vzy+V0EWf57ePWtdZFIZbrLCsq+a2cLtIne90nu52DUZcuawC4jwjVAADcG9qYujZFqZOsvriOX56ujs7Xby7i+TpfJ6XWt9eolZKWkratGr59sB09GUVPR9HesDHs+lHofOZbeEikFLYlbcsOPLvT8rd6zWG/Gfi251hKyfkyzYuq+tFx1jdd1lVdx1mZptXl1Xq+TBartK6FY1uBb/uubSmpPuGkOADAxyRvnVQJAAC+QllRL9bF1Ty7uE5OLjc3Z1BvkjIv67LSt76lSylCz26Ezk4/GA3C5/utg2Gj3XSboRN4lkMT9UeijSlLnWTFbJFcXK2OzhavT6Yv30wXq7Qoq9s/bklhKWXZVqvhtVvB3rB9MOr88tnW0/1us+EFHpUPALgf+H4NAMBXzQhhtKlqU1T1cl2cTpKT8ebN+eZ0El9cJ+8YSCaEcB11cwD1Vtd/vtd8uts83ImG3cC2KIR+ZEpKz7U8N+i2gq1e1O9GzYYvhDy7XM6XcZaXZfXDkrUwoq51VetpUS1W6XQen0+WeVlV2owGUb8T+J7j2Ba7wQHgK0elGgCAr5rWJi/qVVxeTtPTcfz6fH05TW4K1ElW1W/Z8i2EkEJs94OdQXi40zjYbowG4bAbNAIn8CwpOcLpEyqrOk7L+TKdTOPvjq9/8+3lxWR5PU+qqn5byVpJaduW59q9Tri91XxxOHhx2H9+MNjqNixFlzUAfNWoVAMA8JXSxhSlTvNquS4ms+zV2frNxeZkEs9XeZa/NU5blrQt5TmW76lne83ne62DncbeVthtenRQfx6ObXWaVhR624NWs+EpKcLAcd3Fcp0lSV7V5sYfLjBCG1MUVVFU6002nq43cR6nhdaiLOtW5Aae49iKzQUA8HWiUg0AwFeqqPRklt4UqE/Hm9NJcr3IkqwqSl3r2zuohRBR6LQbzmgQ7m6FL/Zbz/aaoW+Hvu06ihObPidjhNZmkxSLVXI6Xr4+nb98c/3d0XWcFnlZGf3DM7d+z7JUI3Q77WBvu/N0v/dnz7cPdtrdtu+5NqkaAL5CVKoBAPi6VLWuapPm1WpTHl1sji82p5PNxXV6k6hvvURKIYX0XOW79rDnb/eCJ6PoySjaGzZGg/Azrx83pBSWJdtNr930oobXioLQdxzbGk8300Wc5WWel1qbH89sr2u9XKdxWqw3+XKdlqWO02J/u9XrBKHvuo6l2L0PAF8TKtUAAHxFjDFxWq3i8mwSn4w3r842Z5N4nZRJVuVlXdfvOjRruxfsbYVPRs0no2jY8wcdP/Bs37M+8y3gx4qyzvJqscpmy/S/jq5+893lZLq+nsV5UeVlJW8f2y5tWwW+02oGo63WN08Hzw/6z/b73ZbvuxZbwQHg60GlGgCAr0WSVUlWXS2yySw9Ol8fX27Or9LZMq9q/bYOattSnqsavt1suIc70Yu95u6wsbsVRqETBbzLfy1cx3IdKwzcrX7kOMqx5clFcBIu58tkvkrKsq5KLcSf9lkbU5Z1WdXruNjEeV6UcVKkWbU7bA26YSNwPc+2lCRcA8AXR6UaAICvgjHmbJKcTuI3F5vjy83lNL1eZFlxcwD1W9+um6HTa3t7w/BgO3o6aj7djULf9l3LtpRlkbe+LsYIbUySlnGaH58vbs6yPjqbLtfZepPduhVcCCGkcCzL951OK9geRE/3+798Ntzbbm/1It+zbUvw2wwAXxbPsAEA+JLKShdlHafVJq1enq5en60vrpPxNF3GRZy+rYNaWkr6rhUG9rDn7w/Dg+1ofxht94Otrm9Ru/xaSSksKZsNt9lwlVKB74aBE0Xe+Xh5ebVK0iLNyqrStdbij8O1EWVVl5s6L6o4ydOsyvN6uc4ORkWvHbQj13Nt1+E4awD4YqhUAwDwJa02xWJTnIzj4yZSOzQAACAASURBVHH88mT15mKdFXVR6J/Y8u1Yg653sN14ttt8vtccdP125Pmu5diKcHUvFGVdlPVilc9W6X8dTb57czW+Wl/NNnFSpHklhRE/+s1XUlqW8n2nGQW72+1n+73D3e6T3U6/E3ZaHr/xAPClUKkGAOBzM0YYY+Ks2iTleJpeXCcn4/hkEo+n6fUi/0Fv7R9zbBV4dityei1vb6vxdDfaG4Y7/bAR2L5rMRL6HrnpsrZtq9HwlBRR6J52Fmfj5WS2mS2SLCvzvPrBK0Ebo6u6TkxeVFVVZ1mxjrN1nO0OW7vDZhR6jdC1lGSrAgB8ZlSqAQD43LQ2VW3Or+Ljy83r883rs/X1Mpst87LSZaXfcWEjcEaDYH/YeLobHW5H+9uNKHRcWyklGQd9H2ljtBZFWaVZeXKxPL5YvD6dHZ/PJ9erxSqt69ub6aWUtqUcx+q0w61e9HS/9/ygv7/T2dtp+67lOZxGDgCfFZVqAAA+n6Ks07xexcVyXRxdbF6drcfTdDxLb+Z+33qJktKxle9ZzdAZ9oIno2h/2BgNgkHH77Y8xyZB3WNKSmUJ23I819baeK4dBk676Z+2gsur1WKVxklRVnVd/8mjFmNMWdVlVRtjiqIqq3oT54tVuo7zfifotYPQd3yPz3gA8JlQqQYA4PNZboqreXp8Gb8537y+WB+dr/NS17XWWui3vCPblooCZ9Dxn4yiZ3vRN/utrZ4ferbjKEvRSPtwVLUuy3oZF9NF8uZsfnQ6e31yfXm1Wm/yvHjryDqlpO/avu/sbncOdrvP9nvP9nvb/cagG37m9QPAo8VTTAAAPi1tTF2bOKtWm+LiOjkdx+dXyflVcr3INkn1tiwthLAt2Wy43aY77AajrfBwJ9rdCoddPwod25J0UD8wtqUsJaWUjq1cW7Uabivy+t3G+Go9XcTrOM+Lypg/GWBmjKlrk+ZVUdZKrfKiSrNitUn3ttu7w2a3HbajwFKC1gAA+KSoVAMA8GlVtc6L+nKavjpbvzpdvz5bTZf5Ki7fMd/7hu9aT0bNp7vR873mwXa03Q9aDecmdxGoH6qbIXa1NmlWXl5vTi6XL4+vX59Mjy8Wy3WqK/2WLmuhlLIs1WkHvXZ4MOocjrq/er797LDv2cqxebkAwCdEpRoAgE+lKPU6KRfr/GqenU7io/PNeJpOl1mSVUVV3/pYW0rhu1bg2Z2mO+j6L/ZahzvRVtfvNt0otOmgfvCk/H5Tt/SdrV7DtpXv2Z1W2OtGF+Pl+HoVx3lZ1j/Y4GCM0FprbdabrKrqqtLrTZ5k5XQZb/eiQTeMGl5AlzUAfBp8ewUA4FPJi/pqnh5dbF6ero8vN2eTOM0qbd5xZpaQUoa+Pej4z/ebL/ZbL/Zbu4PQsqSlpKA8/ZjYtuo0vUbo9rvR7nZnZ6v17etJXeuq0lWtRf3DV5AxQgiTZWWWVXFcTqabq3l8crH4sxfDXzwZ7ClJqAaAT4RvrwAAfExam6LUcVbOV8XlNDk635xfJeNZOl/nWV6/bb+3lEIp2W16vZa3uxXuDcO9rcZoEPZanuMotns/TlJK25KBJ3ot35LtwLXaTf/kYnF6sbiebearpPrRfoebaF1WtTZmOo+Lsqrqer5MDkad/Z32aNjqNAMpBS8nAPiICNUAAHxMWps0r67n2Xenq5cn6+9OV9NFnhXVu9unlZS2Ultd/8V+65uD1tPdZq/lNRsO4eeRU1K4tnIjtx252/3o+WH/1cnsn397/l+vJ3GaV5UW4pbXldZaa71YVctNOl8mJ5fz86vebDVwXacdBVIIUjUAfESEagAAPo6y0vN1Pl1kF9fp+VVyMt6Mp+liXbw7UIe+HQVOv+MNOv7zveZNnO40Xd+1CD74Y7atQt/e227Zlhz2Gwej9k3VOkmLoqxu6SgwwmiTF9VqnZ9eLPKsXK3zl7vdg1Fnux81G57nWl/gNgDgwSFUAwDwsxghhBHGmKyoL6fp67PVy9P1yWU8madxevvxwjduRlI1Q2e7Fzzfbz7fbz7bbR1sNz7XwnHPWEoGnu1vRTtb0cGo/WSv+8//cZ5mpRGiqrXW5pZWfSOKoi7LOkmKq+nm9el8b7v9d//tUBhhW8qxFZPkAeDn40gtAADuzhiT5HWclNfLbHydvjpbnUzi2bJYbYo4q8pK33qVpaRlqW7THXT8/e3G4Xa03Q+2+34n8poN5zPfAu4dI0Sal5u4uLxan16uXh1ff/fm+nq+mS3it32sk1JalrRtK2p4u8P2wU7nyV73cLf7ZLfbjDyl2A8OAHdHpRoAgDuqtakqvdwUk1n66nT16mz9+mw9mWd1bfRbwo2UQknpuZbv2ntbjRf7zWd7zWd7zWboRCFxGu9FChF6Tug5vXb4i6fDYS+yLGW/UUlafD8b/Efz5Y0xVWXKSmd5OZ3Hx+fzs8lqucl937Vty3OVpaRSJGsAuAtCNQAAH8wYobW5WmRX8/RkHJ9N4vPrZDJNl3FRa2Numx11I/TtduQOu8F2P3wyajwdRZ2m12q4rsMB1PhgSknHFvujluM82d9pPzvoHZ3N3pzNirIuy9r8aD+4FEIYYYRIs/L8clFV9WKVPDvovzjsD3uNVuTZFq9DAPhghGoAAD6A1kYbU1YmL+rTSfzqdPX6fH1yGS/WxSYtb71ECiGlVEraluy2vL1h4+mo+WyvuT8Mdwch5UHcmZJCWXJnEO0MotFW63C32wjdvKhW62wd52VV34wH/+G2CWPyvJzk5WKVnI4Xl9frqqqrqi9EKwwcx7KUotEaAD4AoRoAgA8QZ9U6LiazbDxLX56uji42q7jcJGVe1m+7xLKU56hOyxt2/YOdxtNRc7sX9Dp+M7DJLvhYmg1XqZZSB7vbnZfH169PptfzeL5Myqquai1v2zxR1TpNi5PzeZ6Xb87mT/d7h6Pu/qgTBY7nWrw2AeA9EaoBAPgJN/O9y1pXlb5eZONp+vp8fXS+Ob7cXFwn77jwpjod+na74e4OG8/3mk93o6ejqBnehBZSCz6awHcC32k3g2cHg1bke659fD5XSq7jPEkLrY350XhwrU2eV5N8Pblej6/W1/N4HRdGyGG/0Wl6rqMsS0nB6xQAfgLTvwEA+Am1NlVtrubpeJq+udgcX26u5tnVItskZZK969Cs0Le3Ov5OP9gbNvaGjf3tRq/lRaHt2pbFKdT4BLQ2tRbzZTJdxMcXi+PzxdHZ7OxymaRFmpVaa/2WM9N9z2lG/rAfjbbbz/Z7Lw76g16j1w4dW9m8VgHgnahUAwDwVnVtqlonWZVk1ZvzzXenq+PLzcllHKdl/JY4LeVNgVq5juq3/Wej6HAUPdmJdgbBoBMwkAyflFJSKTHsNwbdsNUMeu1G4DuuY13N4vkyybKyKKr6thOts7zM8nIdZ+PperVO87w63O1WtWmGbhQ6lqUsmv8B4C2oVAMA8FbrpJwt88vr5OwqeXO5eXO+XidVnJVVZar69jOobUuFvt1tuqNBeLDTeL7X2u4FrcgJfdt3LcaS4fMwRqR5GSfldJlMrjffHV+/OZtdTTezZZKkRVHVt3ZZW5ZybKvV9Lvt8GDUebLXOxh1DkedZsONQvez3wQA3A9UqgEA+BM3x2UVlc6LejJLTyfx8eXmzcXmcppeTpN3PItWSnqOFQV2v+PvDsKnu83DUXSw3Wg1HNtSNKbic5JShL4TeE676W/1It+3G4Fz0vD8iT2dJ+s4K8u6rn9Ysq5rXde6KKvZItnE+XKdxUlelPV2Pxr2G55juw6zwQHgh6hUAwDwJ6pKZ2U9XeQXV8nR5eb12ep6kc9XeZpXaf7WEd9CCN+zRv1wb9g43GnsDxs7g7DbdMPAdmxFpMaXorUpa72J89kyPbtcnlwu3pzNzyer6TyOk0Jr/eOPglIKKaXvOYHvbPWiYT96dtB/ut/fHkRbvchzlGPTxQAAf0ClGgCA71W1zgu9TsrFOj+bJK/P1ifj+HSyidMqL+tbn0JLKW7apwPP6jS9F/utp7vN/WE47AXtyPVd67PfBPAnlJKesrxO2Gx4oe80G14j8KKGd+I503kcp3lRVGWtzR8NMDNGGGOStLgZbzZdxElWbpJyHXfzom5HXityXccmWgPADSrVAAB8b5OU41l6dpWcXG5OJ/HpOF7FZZpXdW3qt8xMVkq2I7fX8vaG4eF24/l+a3erEXiW71qOreigxtdDa5OXVZJWi3U2vt68PpmdXMzPxovZIl7FeVXV4i1d1ralmpHfioKDUftg1Nkfdfa221vdsN30P/tNAMDXiEo1AOBR08ZUtSmKOsmqq3l2dLE5GW9OJ/HVPJst87K6fRqZlFIpEXp2FDo7/WB3KzzcaRxsR7tbYafpfeZbAN6HUjLwHM+xw8BthF7gO62mFzXc88lqMl2vNnmWFVX1w93gN13W1TJZx3lelKtNto7zTZzHO52dQRQGru/ZliVpcADwmFGpBgA8amWl47S6XmSnk/jN+ebV2frmAOq8qIsfBYzfsy3l2Opgu7G/HT4dNQ93on7H6zY9z7XYE4uvmTFCG1NVOi+q+Sq9vN6cXHx/lvX5eJmkRV3f2mUtpZSua3mePeg0tnrRwW73cLd7uNvb7keBb7m87AE8YlSqAQCPVF2bOCtXm3IyT8+vkjcXm9NJfHmdbtKy+tFU5N/zHMtzVStyO5H3fL/5bDcaDcKtbtAIbDqo8fWTUlhSWq7luZbrWJ7nBL4bhV6r4Qe+cz3bLJZpXpRl9SdDBIwxxpgs03lRVZWOkyLJytUmj9NyE+eDbtBp+oFvOzZfAgAeIyrVAIBHKs2rk3F8fLk5Ol+fjpPxLF1tirzUtX5rohZC9NveoOMf7kSHo+jpqLm/HXqO5djKUpIOatwvWpuq1mlex0k5nq5PLhYvj6+/O7qaLuL1JtNvmSNw02XteU7U8HaHrf2dzrOD/pPdzu6w2WzQ+wDgMaJSDQB4RIwReVlneb1Jy+kie3m6ujmAerrMV3FRlLd3UFtKWkqFgRUFzv6wsb/d2Bs29rbCra7fpYMa95ZS0lWWZSnfsx1Hhb4d+HboO+fj5eXV6qZ9Wv/oGdNNl3VZ3VSt6zgt0qxcx9lqk20Pmt1WEPiOxVnWAB4TKtUAgEekrs18nV8vsuPLzdH55rvT1eU0KUtdVuYdBWrXsXzX2h82Dncaz/eaz/dbrYbbaji2rWyL6ID7zQhhjKhrXdV6tcmny/ToZPrdm6vXJ7Ojs3lZVnX9lnF9SlqW8ly73fQH3Whvu/3icPDrb3a2B5HnKouNGwAeDSrVAICHzxhRa7OOi8WmOL9Kzq+S03F8fhWPZ+k6Kd/xeNl3rUZgd1tev+0/2Wk83W1u94JhL/Bcy3WYzISHQAohpVC2ujkEzvNs25KNwG03g3YrmFyvx9N1WdZlWf/gQqNNpWutta51UdRpVqZZlebVwU572G90W0HUcG2LLxMADx+hGgDw8GljilJfTNOj8/Wrs/XxZTxbZuu4rGr97g1bjcDe3Wo8GUXPdqP9YWNvu+HaymZvKx4o11a25fpuZ2fQHA1bT/a6//IfF1lRrTfZj0P1DV2bXFdlpZO0mC/T08vFk73ur7/ZeXHY89w2oRrAY0CoBgA8ZGlWxVm1WBfTZf7qbHV0vrmaZ7NVnmRV8ZaQYFnStVUjcNqRuzcMn46ao0Gw0w86La/hO6RpPGBSSksK5VquY20PIt+zlVLtVnByPj+9WMxXyWqd/fgqbYypdaZNVeu8KKu6Lst6sUqm83i03druN33XVoRrAA8XPdUAgIdsMs8m0/T1+fr12froYn12lWj9ffv0294APddqBs7uVvhsr/liv/liv9VsOIFnM98bj4rWRhuTFzor6n/+7dk///bs26Or1yfTd1wipRRS2Lblu/ZWPzrc7f7Vr0b//a8OO03fsdjeAeDBolINAHho6tpUWq82xWJdnow3J+P44jq5vE5nqzwvbq9O3/BdqxW5g46/0wv2h+HBKNruBt2m57qKXax4bJSSSkjlK8+1nx/0XFtt9Rp7O53z8WJ8vS7Kuqxq8adPpowxwoiqqpNaX8/iutJlWS9W2eGos7/T7rb8VuSRrQE8PFSqAQAPzc2hWUcXm1dn61cnq1en63VabNLqHdXpG92W93QUPd9rvthv7W6F2/3AcywpSQF47G5OtL6aJ+eT9f/4x9f/6zenyziL01IaI97yNXXzheO6tu85f/nL0f/2109eHPQOR22L51MAHhwq1QCAhyPJqlVcXC/yq3n65mLz5mIzXeTLuCjKWuvbP/tblgxcOwqdbtPdHYbP91q7W+Gg47cbjutY7PcGhBBKSUdanaavpPzvf3XQ7zRen83OxovFMo2TvKq1+VG6vnmGVRRVXeujs5kx5mKyfLLXHQ1bo0HLdS3XJl0DeCCoVAMAHo7JPD0dx69O1y9P16eTzekk+cm3Oc+xei1vdyt8sd98cdD65qDdiVzFeG/gNsaIqharOP+Hfz/9l/88e308vbxa52VV1fodVWshhZBid7vzZK/3N7/e/9u/OGhFbhRQ2gHwQPDtDABwv1W1Lkq93BSzZX4yjo/O1+NZejXPVnHxjkStpHBdq9fyht1gb9jY3Qr3tsKtXhD6NokaeBsphWWJ0Le/OeyFnrW31Tq5XJ5czKfzOMvLoqi1vq3LwghhxGqdHp/NhDGzRfxkt/tkt9NuBs2GZynJbHAA9xqhGgBwXxkhhDFFqddJeTZJXp4uX56uX56u47TMy/ptgVpKIaW0lWwE9sF24/l+68V+a28r7LY837U+6w0A95CSIvCs5/vdve3Wwah7dDbzXCWFmK9SY4qirIy+vWS92eSbTT5bxN8eTf76z/fr2uzvGMexXUe5UvEkC8D9RagGANxXaVbNlvlknl5cp6eT+ORyM13mSVZV9VtL1ErJKLDbkTvo+KOt4Plea2/Y6Da9KLSZ7w18EFvJbsuXsud7zsFO9+hsfj5ZTufxapPneVnX+tarylLHSfHd0dV6kx/udQ93u7tbre1BFAZO4PG5FMC9xDcvAMA9Y4zRRmhtlpvieLx5fbZ+dba+uE6u5llZ3f45XgghpVBSuo76fQf1s73Wk1HUb/ufc/HAg2FZqhV5zYY3Graf7PY6rctG6LqOJdV6sTR5URotfvx8q6rqqqqPTmdvzuaX16vJ9frPvtkxUm51Q9tSSknFvH0A9w2hGgBwn1S1jtNquSmuF9npOH55uh7P0vk6j5Oqfst8byGEbSnXUVsdf9jzD3ei/e3Gdi8YdPyG73zOxQMPkBSWEs2G84sn/WbD3e5HJ5eLk/PF9Wyz2mR5Udc3s8F/xBgzXyS6NklWXk5WT/a7BzudfidsNX3P4WR4APcJoRoAcD8YY6raJFk1XeZnk/j12fr1+frofL2Oy/rtB1ArJS0lQ9+OAufJKHqx33y629wbNqLQoYMa+PmkEFKKRuA0gnavEw56jW47DD0n8OzzyWq1ybKsrGp965l2q022jvPZMjkbLxbrNE6KJ3u9XWOaDe9mZCCd1gDuBUI1AOAeMEakeT2Zp5fX6ck4Pp/EF9N0usySvNZGvOPYrFbD6Ta9nUE4GoRPR9H+dthquFHgONTBgI/Ndax+O3Qsq9MMRsP20ensfLKaTNerdbZJcq31LXPBjSmKarXKXh1frzbZ1Wyzv9Pe225v96NuOwjZSwLgPiBUAwC+XkYIXZuq1kWl56v8pn36+GJzOU0XmyIv6luvklIqKf5/9u7kOY7zPBj42/vePftgBoONAElRFC2ZFq3F+uI4TuJLkktcOeWYPyR/Ry45piq+JbarUuUtouxQu0ESpLhgB2Yw+0zv6/t+h9cagyBIAVwB8fmVyyWCmEFjQPT008/G86zAsdWC0qho83VjrqbXSmo5Dx3UADwvAs8KuqSrUrloFPKaoUuGLsuS0O7ZCKEwSpI0Iw/VlSRplqRZECWDkT+2g8HId704jFJMCLEUgec4joVGawDASQZBNQAAgJMLY2L78ciO9vrBTsdb23FaPd/xEz/M0kfPJBN4Vha4ckGuFpX5mj5f0wuWnDclTYZ3PQCeO4ZBPIfyhszPFi1drhT17dZouzXc6zqdnpOk2aGDwTEmUZz2Bm4Up44Xdfpup+9MT1lTJTNnyooE8/kBACcXXF4AAAA4cQhBmJA4wUGY7vX9ZtdfbzqbLXen442c+DEP5DlG4FlTEy1dXKgbCw1jtqo1qposcqIAHdQAvAgMgziG0VVBVwVDk/KWahqyrkmKLCKEHC/y/SjN8IHQmhCSppntZl4Q2240HPtjNxg7YRCl9dgsWIqqCCLPsSwkrAEAJw4E1QAAAE6cDOMoxq2+3+z6my13t+11hsHQif0wffwDdVUo5+VGRaP/m65quiIoEs/BhTgAL4Mk8gVL4Xk2b6rlgj5VNrebw+3WcOwEYzdkDpuGgDEJo3gwwnGSjp2gN/IbU9ZMLVevmFNFXYZd1gCAkwdOTAAAAE6QJMVxgt0gGbvJ6o693nR29ry9vu+F6aM6qOl8b1FgJYGrl9X5ujE3pc1M6eWcXMxBBzUAL5PAswLPSiKfM1RDk4qWamiiJHGtjs1ybBQmUZwS8sA6a0JIkmRpmgVh4nqR7UZjJ3C9KAiTLMMFS1FkQeA4loMmawDASQFBNQAAgBPE9pLeKGx2/J2Ot9lyd7ueH6ZhlKWHNWFSAs9qCl/OyTSiXmyYOUM0VUGCjVkAnAwcyzACypuyKnOqIlRLxubucHN3uNMaNTvjNM2y7GDOmhCask4GIy9OUjoYvNkZz00XFhqFvKmoEs9ABQoA4GSAoBoAAMBLhjHBmIRxFkTZTtvbbnvbbXe77XUGwcCOHvNAgWdlkcsZYiknN6rafF2fqejTVU0WOI6Dq20ATgraZa3KvCrzksjnLEVTJEtXVFngeXZkB7YbpOmhXdYkyeIoSb0gdrxwMA4cL05SXC8bpZyqKaIi8zAWHADw0jEH1xoAAAAAL1aUZGGUtbr+btdf33U2mu7IjcZuHCVZnDwyQc0wjKUL9bI2U9Xmanq9pFaLiqEKisSzLFxmA3BCZRlOM+KFiRcku3vjnb3RV6ud1c2u7YZ+cPgYQoZhOI4RBF6S+LylVkv6XL2w0CjO1qyZmgVTwQEALx1kqgEAALw0NJweOfHQiTaa7vqu0+r5rV4QJVmcPLKDWuRZWeRUmZ8qqoszxkxVr1fUgilZmggJagBOOI5jOQ4JAmdqsizyOUMWeE6R+FbX7vQdz4/DKDmwy5qmrNMsCaM0ilLHDT0/9vzYC6IkTfOWahkKzzInajA4xhhjTAjJsgxjvP//McYIIUIIwzAMw7D7cBzHcdzkj3B3EIDTAjLVAAAAXpr+ONzrB1t77lbL3e14u10/jLMoxhgT/Ii3J0nkcoZYySuNijZf1xcbZsGUZIkTeJbnWLgEBeBUIAghgpI0i5PMduP+yL+70bm30V3b6u917SzDGB92BmAYlmU4jlEV0VDlqYrRmLJeX5q6dK6uyLwknJSUNSEkTdMkSeI4juM4DMMwDKMoiqKIfoQG2DSEFkVRFEVJkiRJkmVZlmVRFAVBEEWR4ziIqwE4FSBTDQAA4IXKMElSHESZHyQ7HW+jScNpb2BHQ/uRO6hZlhF4VpX5vCnVSmqjrM5UtXpFq5UUBVbsAHDaMAghBokCJwqcIouWIXMco8i8poimLvcG7tgNkyQ70GWNCMEZyTDKsigIkjBOxk4QJziOcbWkVwqapoqaIrycbwkhjHGSJFEUBUFAg2caV0dRFIYh/Y8kSZIkoflqmpoWBEEQBEmSRFGUZZn+x+SDiqJIksTzPMuelFsGAICHQaYaAADACxXGmesnrV6w3XbXdp21Hdv2Ei9I0wwn6eEd1AyDeI41NaFR1edq+kLdqJfVginpKi8J3Imq+QQAHBchCBMSRonrx+2+u9kc/XFld3WzN3aCMEoOfQjDIIQYnmd5gTM02dTk189OXTpXm5u2ZqasF3v4fxbHseM4g8Gg3W4HQUDT0RT+GtmH+Rqt9N5f/k3/3zCMSqWSz+d1XReEl3azAADwjeDuPgAAgBckijM3SAbjqDMItzve1p7b6vntfhAnWfrQQh2KZRhZ4lSZzxlipaCcmTZmqvpUUcmbkipzMKAIgG8BOhtcU0RZEmRJUGWRZVDOkLf3Rp2+YzthGKeEIPTALmuEEEnSLE5xFKXDkc+yTJykI9uznbCY1/KWwnMs9/zvuNGu6TiOgyBwXdd13dFoNBgMoigiDzr0sfTjWZZ9/VI8II5jQkgURbqua5qmqqogCNBrDcAJBJlqAAAAL8jAjrZa7kbL3Wg6zZ7fGQRBlCUpfsQFJ0IICTxbyslTJWW+ZszV9PmaXsrJtH0aLiwB+JYhBGGMkxQHUdLpuzfv7d1e7dzb6PZHfoYxwYg59EzBIAYxksRLEj8zlZut57/7+vTFczVF4uXnv6yeRtTj8bjVag2HQ9o7Tbumv/6mjn2lPTm10Y5rWgReKBTq9bqu66IoQik4ACcNZKoBAAA8R1lGwjj1w9Txkt2uv7pj73b8ziAYubHjJ4fPIkKI51iBZw1VyJtio6o1Ktp0RZsqKkVL1hR45wLg24lh6GxwVpZ4geeyjCiSkDOVnb1xp+86bhjHaZpmB88bBBFEwjAJooRBKAgTQpDjxbWKUS3qhiarssAw6Jnfg8MYh2Ho+75t26PRqN/vu6476Zd+mmeexOFZloVhSPu0aem4ZVmWZdF5ZnBbEYCTAzLVAAAAnqMwygZ22OoFmy13bdde3bbHXpKmOMMke0REzTBIFnlDEebq+nxNP9MwG1XN1ARV5rkTtjUHAPCcYEziNHO9uDf0V7f7f7y9u9UcjsZBEMZJmhFCGHTIqYDjWJ5jdU3KW+ql1+oXz07N1fPVosGx6Jknd+M4Hg6HvV6v0+mMx+MkSbIse1Sl99OgpeA8OzZpfwAAIABJREFUzwuCkM/nq9VqsVjM5/Mc99zz8ACAI4L7/QAAAJ4xghDBJEoy2016o3C36zW7/k7Ha/eDoROH8eELqCldEQxVKOXkcl6erxszU1olL+dNSRRY6KAG4NXBsows8hzLCjzHcozAs5WCvt0adQfOcBwEUZIkGXnoxlyW4SzDGSZRnIoCFwRxt+/O1HLlgpY3FVnief4Z3JbLsszzPMdx+v3+cDi0bTsMw+cRTlP0mZMkSdOU9r2kaZqmKW20hnw1ACcBZKoBAAA8Y4SgNMNjN17fddZ27dVdp9UNhk4URmmGv+Fth1Z6L9SNhWl9uqxVCgrLMhy7r8sQAPDKIAQRQtKMxGm22x7fut++v9nb2Bn0R77jhVlGDu2yZhjEIIbjWUUWGlO5hUbh4rnamdlSIafoivD0Z5I4jlutVrvdHgwGruumafqU9d5Hx3Ecz/Omaebz+ampqXq9Dv3VAJwEkKkGAADwzGSY+GHqePHAjvZ6wdqus9PxeqPQ9pIgSrNHjPgWeFYUOEsXc4Y4X9Pna0alIFcKiqkLogDXiwC8uhgGMQwjsAzPs5WCjhewrorFvLazN97dGzleGARJkmbZg9v4CEEEEZxkGJN210lTHCW4P/Jnala1ZFi6rMiCyLNP0EuCMfZ9f5KjDoIgTdMXmaCiq7A9z0MICYIgiqKu66qqvrADAAAcCoJqAAAAz0yWkZETNbve6o6ztutsttyhHWGCHl8XKQmcoQnzNX2hYS41jIW6Lom8JMB0bwAAQggxCDEMyhmSqZWqJXOmXry30RV4ttW1ewMPBXGWYXTYGSbL8GDs217YGbhbzeFri9Wz86XZer5c0J9sQAPG2Lbtbrfb7/dt26YLqJ/Bd3hktBScTi+jS60RQoqiwNkSgJcLgmoAAABPK0lxnOChE/VG4faet9Px9np+Zxi6fvKoBdSUoQo5Q6wWlVpRbVS1mapesCRV5jnuSZJIAIBvMYZhOI7RFKGCVAYVTU3Y3htvN0edvtsbuH4QB2Hy8AYrQkiaYj+IuwOXZZDjBt2BO121amWjYKm6Jgr8kbZZ065m3/fpcDLf9+lYsuf0zX7jwWRZ5rouwzCCIKiqKkmSIAgv5WAAAAiCagAAAE+JEJSk2PWTrZa7uuPc37F32p4XJFH82HgaIYSQqQlzNX1pxlxqmJWCUs4r6NCRvgAAgBBCSBI5SeRyprw4W9xsjoq57tpWn2VRb+BFcYoxQg/lrAkhcZwmSTq2/WZ7vNexu42Cf6aaNQgv8DzHHT5J/CFRFLmuOxgM+v3+0+/Neho0X+37fhiGiqJYlkWj65d1PAAA7l//9V9f9jEAAAA4lQhBthe3B8FG07mzOf5qY7y6Y/eGoRukyUOrZCdEgTVUoVpUzkwbFxfzl84W5qb0cl7WVUEUID0NADgCBrEMwzKMKgmqKpq6osiCwHMcy6Zphh5OWH+NIJRmOIgSx41sNwzCJIozQhDDII59ZMsJISRN006n02w2x+MxnfX9HL+7Y4rjWBAEXdcRjHQE4CWBTDUAAIBjIwghQjKMBna0veeu7jjru06z6/fG4WMexSCEGEYW+bwpzdX0xYax2DAXG6YgsCxcCAIAjoyeL3KGnDPkYl6rV3OFnKoq4sbOIErSMEwwPmSQAyEojtN+nI6doNVxmp1xf+QvzhbPzJYI0XmOFZjD42o6Hqzf7+/u7r7IWd+PR/PVtm37vq+qaqVSmXRZAwBeMAiqAQAAHE+cZFGM++OwOwo3dt3NPbc3Cgfj0AvTRz2EYRDLMLoiFHNSraROV7TpslorqQVL4jkWKr4BAE9MlviiJTMzxZyplgtGtWzude1Oz3G9yA/iQ5PWGJM4TgcjH6Ge58f9kT9dteoVs1zQSnmVe7DJmgaudHvWS+yjfhSMcZZltm23Wq1cLmdZ1ss+IgBeRRBUAwDAt1yGCc3ZYEwwvcAkiHw9UJdhGObrQkqWRUfJGEcxHrvxRtO5v22v7Tpbe16UZEl6eOqGQYhhGY5lBJ4tWNKZafNMw1iYNso5OW9IHAfhNADgqcgiJ4ucocnTU/liTi0XtfsbPZ5n2107zbI0xRjjA4EwxiTGaWJnjhcOxn5n4PaGvu1GaYZVVVQlnuf+vMyPBtXtdtt1XYzxSQuq6dAy27Y5jhMEwTRNqAAH4MWDoBoAAL61MCEYo54ddexw4MZDN3bD1A3TOMmyjIgCK4u8LvOmKhR0sWiIFUvWZZ5lHndJRggZu/Hmnntv2769Phq7cZRk2aP6pxESBNZQhYIplfPKzJR2Ztqs5GVLFxWZZ2EFNQDgGWEZxHOoYCkCzygSXypoO63hTmu013MGI5+G1g89iGCMgiDpETfLsOuFrh+6fjRbyzWmLJquTtM0SRLXdUejEW2lPoFBNUIoCAKEUD6fj+OY53koAgfgBYOgGgAAvoUwIRkmQZz5Ubbecdfa7k7f3+kHfSfuO1EQpkmKFZnXFaFkiNWc3Ciqs2U1SUk1J6sSJ/Asd1hoTQgiBHlh2h2Gu11/o+U+6gqTZr85jjFUYaqozlS1hWljrqbPTumqDG89AIBnjGEQx6CcIVm6ZOryVNksWIqpyzzPZRnxgziMDjZaE4IIIWGUhFHiB/Fg5EVJShBSJL5eMWlQTddoua7rOM4JTFNTk83VnucFQaAoCgTVALxgcGUDAADfQm6YDr349rZ9c2vct6OBE9l+4gRpGGdhkqUpzggJwjRLcRSlIydq9v07u+Ln+nCuor4xl5spqgVdFPmDqWSGQQgxqsyXc1LeEA1FCOMsSrKHPo0ReEZThEperpfV2Sm9XtbKeTlviKIA6WkAwHPEMEiReJZh2EYhZyqlvNaYym3sDprtse2GYZgc+iic4TjOOIY1NVmRxcktRc/zut2u53knM5zejxDiOE6n0ymXy5IkvezDAeDVAkE1AAB8q6QZiTPcHoUbXe/je/2rK50wypIUY0wOLLmKsyyOM9dP+gxiWYbjWIFn56t6hgnGhGFQThUFnjnQZc0wSJP5cl4u5eSCKQ2caH9QzTCI41iRZzVFKOWkxYa5UDfm63opJ2vKA22KAADwnIgCJwqcporlgl7MqZWiriiCwHOtzng48uMkS7OHcs4MwzBIlviCpWqqSAt1CCH7g+qTHFcTQjDGruv2ej1N03K5HHRWA/AiQVANAADfKiM/3u0HNzZHX6wNtru+H6RZRgeVPfJykM4wQwRjTFp9/6Nb3Z4dffdMfnHKaBQUWTxYRqgpPMeqM1V9MI7wDhm78eSveI7Nm1IpJ9dLynRFm6vp1YJiqIIs8SysoAYAvEAMg3iOsXSZ51hZEusVa3NnsLk72GqNhmM/TbP99xlliS/ktHrVmq3nCpbKMogWewdBMB6Poyg6yRH1RBAELMuGYYgxZh+9dhsA8MxBUA0AAN8SmJA0I51ReHNr9MXq4PP7gyBKs+xIF4KEoIwQhMnQiW0/ocuxGIaxVIHnWO7BazNJ5CSRq5VVx4/HXtwZBDTrI/CsrgozFW1mSpur6Y2qNlVUDVV4Tt8vAAA8BoMQxzKaIqiyYBlKpajnTcXQJZZjeZ4d20EYJVn2pxuOuiY1arlGzaqUdF0RGQZlWZYkSRiGvu+fioiaEEI7wKMoSpKE53meh+t8AF4Q+GUDAIBviTDBtp/cazl/uN3b6nhRnB0y7PabEEKyDPXG4Rf3ByyDyqbEsYylCtxDGY+iKZ2dMYd2PHJiz0+TFE+VlHpZW5jWGxUtb0qmJjyc5QYAgBeNQTzHGJo4P50zdalSMta3+7fu7TXbY8eNkjQjDJPPaW+cq83VCxL/p7NdFEWu68ZxzDDMqQiqJ8IwdBxH13UIqgF4YeCXDQAAviVsP1ltu3d27bu745Gb4EevuXo8QojjJ16YGgo/U9ZEnlVETnkoPDY0QeCZuZpue4ntxkmK5+vGfF2fqWrlvCyJHHRQAwBOApqy5kROFlVDl/I51dJlhJAs8e2u40cpYlC9ai3Olqolnef/1KkSRZFt22EYvtyDPy46CdxxHFEUFUV52YcDwKsCgmoAAPg2IIS0hsGHK52VzbEfZU+ZVyEIYUKaw+DqrS4iqJZXZJE7kKrmWEaW+OmKxjBMFGeEkFpJLRcUTeYkkeOggxoAcPLwLGuo4vx0ztLFpdni/c2+7UUIoaW5UrWo6YowGc0YRdF4PA6CIMsOLjg4sQghWZb5vm/btmEYL/twAHiFQFANAACnXpLhIM5aw+CrHXun78cpftpaRYIIQmMvudd0pgtK34lkgVVEfn8NOMsyDMuUcrIscrTQPG9IOnRQAwBOMJZlJJaTBCVvKjlTVRXJ9WOEULWoWbokCn8uyUmSxHXd0zKijKJBdRiGnuclyeHLwwAAzwME1QAAcOp5UdYcBM1B4PhpnGBy/FbqQxCUZSSMsr4Tb3Z9UeCm8xz/YLqaQUiROIFnCSEMQgLsoAYAnB66Kp6ZyacpRgjJEi8IDzS50KFfpzE0TZLkdCXYAfgWgKAaAABOPTdI1jvuds/3wiRNMULPJq+CMYmTbGBH91qOofBVS+K5g53VAs8K8E4CADiFRIEThUd2HWdZFkXRqQtNCSFpmsZxfOqOHIBTDbIKAABw6tl+crfpbHa8IEwxfma1ioSQLCN9J/pqx97tB0fczgUAAN8CGOMsy/ATLFF42eiRn6KqdQC+BSC/AAAAp54fZc1+0B2HcfrsQmqECEEIESdIdnpe34kyuEQDALwyaH/yaQxNIagG4MWDTDUAAJx6cZoNncj2k+xJ12g9RhRnAydy/Cff0QUAAKcL2edlH8uxnd4jB+D0gqAaAABOvTQjQZzGyXNJTaQZCaIsTqE9DwDwajm9cenpPXIATiko/wYAgFOPEJRl5DmV+2FCSEayZ1lXDgAAAADw7QFBNQAAfAsQQp7RyO9DnhsR9KwGigMAAAAAfNtA+TcAAAAAAAAAAPCEIKgGAIBTj2EYjmVYlkHM83hy9KcnBwCAVwnLsgxz+k59DMOwLFzhA/BCwa8cAACcegyDOI7h2Ody9ceyDMczp/PaEgAAngTztZd9IE/iVB88AKcUBNUAAHDqiTxn6aKuCNxzyCeLApfXJU3m4RINAPDqYFmW47jTmPJlGIbjODhjA/Ainb4zBQAAgANEgS3okqkKz6NIWxa4giEZivB8EuEAAHASsSzL8/ypC6ppRH1KbwcAcHrB9G8AADj1TEVYqhm2n4zcOM0SjJ/NqG6GYViWKRji+WmzXlR4DoJqAMCrgud5RVEwxkmSvOxjOR5BEBRF4TjuZR8IAK8QuIkFAACnnqkKZ+vGXEVTJf4ZNj/TVu2iKb/WMKcL6vOoLQcAgJNJEARVVQVBeNkHcjwMw4iiqCgKz0PmDIAXB4JqAAA49VSJaxSVWkExVEEUnlkrHccxssgVDHG+rJVNCYJqAMCrQxRF0zQVRTlFM8Bp7beiKIZhiKL4sg8HgFcIBNUAAHDqqRI/XVDqednSBFnkGAY9g91aDBJ4VlP4kinNltWSIUJQDQB4dUiSZFkWraM+LUE1Ha6mqqppmhBUA/AiQVANAACnHoMQwzC1vPIXFyuX5i1V4p9yqBjDIJZlannlgwvl78znVJFDp+SaEgAAnglJkmhoSsizmVLxAtBDlSRJ1/VTV7gOwKkG7RavEEIIxhh9fc5FCNGKpv33XzHGGOMsyzDGDMOc0tGXALyCGISmcvJ750tRnK02nTjFcZI98aUgxzKiwDVK6nuvlc7VTVmEGWUAgFeLKIocx0mSdFrS1OjrDdWyLOu6DhdvALxIEFS/QjzPG41GSZJkWUbjatM0TdMUBIFOs8iybDgc9nq97e3t8XisqmqpVJqbmysWi6eo9gmAV5YkcAVdXKoZ775W/mpnfL/pxAk+7iRwmqMuWfJi3XhzId8oqKbCQ+E3AOBVM+lPNk3T9/0wDE94ypqG04ZhyLJ8ivrAAfh2gKD6qRBCaPp3Av1pCc2fnZyT2ng8Xl9fd103DEP6kdnZ2ZmZGV3XaVCdJEmr1bpz587y8nKz2czn8wsLC3Qxg6qqMEYSgBNO5FmRZxeqWpxhhNDAicdeHEQZIUe6FGQYhBAj8Iwi8Y2SenmxcGkuV83JmgS/+wCAVw7N+iqKksvlMMaTa6eTaXK0lmXRoPplHxEArxa4VHoqaZpGUWTb9mg0sm17PB5zHMfzvPU10zQ5jjshp7b19fVf/OIXzWZzOBzSS+wPPvjggw8+mJ+fV1UVY+x53h/+8Idf/epXu7u74/FYluW1tTWGYdI0XVpayufzL/s7AAB8s5wmnqsbLIMsTVjZHN/cHMUpzjKMMXpUbM0wf1pJzXNMNa+8MWe9PmNdaFi1giIJJ+L0BQAAL4Wu69VqNY7j8XiM0CPPoicBwzCGYVQqFVVVX/axAPDKgaD6CcVxHIbheDweDAbdbrfdbne73W63y3GcKIqVSqVcLk9NTVUqFcuy6LgIjuOex5Gkaeq6bhAECCGGYXielyTp0P2ErVbr2rVr9+7d29vbo0l1RVHm5+eLxSJCiBASBMHKysrvfve7STa71+vV6/VarTY9PQ1BNQCngi7zuswrIleyJJ5j7SBxgiSI0jjBcYIzTDAmhCCCCMMw7NextCSwksgpEr9Y0985X7rQMGs5RRGfy1kLAABOC7pSazAYCIKQpmmWZS/7iA7HcZwgCIZhFAoFmPsNwIsHQfWx0ZLvdrt9+/btmzdvXr9+fTAY+L4fBEEQBLTkW1EURVEKhUKlUnnnnXe++93vVqvV5xSUjsfj3/72tysrKxhjSZKKxeKZM2feeuutcrl8rOehVeuCIIiiOEmtn5zadQDAsegyP51Xf/AaM1/Wdgd+cxD0xlHXjuwgcYI0jrM0w5LIySJnqkJeE8uWVLHk6YJSL6qNopLXRYGHHDUA4FVHUxS6rudyOdd1fd8/aclqeqmmKIqu65qmSZJ0QgokAXilQFB9bGEYjkaj27dvf/TRR9euXfvkk08cx6GJ3/0YhtE0rVAoeJ5HCLl48SLP87IsP9sNBxjj8Xj88ccf//rXv86yTNO02dlZ3/cXFxefIKgWBKFcLs/OzsqybNs2x3GVSqVYLE6argEAp4UscLLAWapwtmZs9fzNrtschK1hMHDioRcHUZqkWJF4TRYKhlg2pXpBbhTV+bJe0EXuBM2CAACAl4nmGwzDKBaLWZaFYYhptc9JwrKspmmlUgku2AB4WeAX79iazeaHH3746aeffvnll9vb20EQHHpuJYREUTQYDK5evdpsNv/mb/4mjuP5+fnjxrqPgTGmHd17e3tbW1uEEMMwBEEYDAZJkhz32RiGUVX13XfflWX57t273W5X1/W5ubn333///PnzhmE8q8MGALwwLMMwLKpYkipyc2Xdj9IowXGK0wwTQjiW5XlW4llZ5FSJ0yTekHmWZRBE1AAA8DWGYUzTZBgmCALbtskRZz++KCzLchxnmmatVoNuagBeFgiqj4GOJdvc3Pzoo48++eSTtbU13/cRQrRkmo7IRghhjF3XdV03yzLXde/cubO3t6dpmqqqqqoWi8XJaujJRujJl6B7ofcnibIsS9N0/2HwPM9xHCGEdlMPBoN+v9/v9xFCSZIYhjEajejuh8mi6SNmnSRJOnfunK7r5XK52+2aplmv11977bVarXZoQzh9X0nTlK7pStOU53me5yc7uvZ/XUII/Wb3b8mmn0YISZIkiiL6t5IkybI8eZXo38ZxnGVZlmUcx9HGIbpA++gJNTqePU1T2hN1lKei393k7ZO+nvSlIITEcRxFEf1bWZbpMe9/LMY4SRL6FQkh9AvRHx/kAcGLQSeQmYpgKk9eI4MxSTFhEKKL7WG7FgDgVaMoCsdxlmWNx2Pf9+m7/8s+KIQQYhiGTtKxLCuXy0HhNwAvCwTVxxAEwe7u7p07d27fvr2zsxPHMf04nQx55syZs2fPZlkWx/HNmzdXVlbCMIyiCCEUhuGNGzd4np+dnV1aWqJhFUIoTVPHceI4nsTViqIYhrE/DJ7cFqV/pLdLNU1L0zQIgm63u7e3R6eUIYQwxnEce543HA5pmE33Kx69FojnecMwFhcX6/W6IAimaSqK8qhPpjEnndbmOI7jOKZpWpaVz+ctyzoQptLp4mEY0iEfLMtKkqTrOsuyaZoOBoNWqxUEAca4Wq3Ozs6KosgwDA1N6bfjOE4QBIqi0Lr6XC4niuLRv7U0TelsufF4TNuiVFXVdf0xT5Wmqed5URRNjpn2LCGEkiTp9XrNZjNJEoxxo9GYmZmhHfX0sfRfwmAwGA6H9A4LHQifz+dVVT1Ru9YAeLwoydwg4VhWEnhBYETotQYAvGLoINhCoYAxbrVa9ArwpcfV9F6/ZVlTU1O5XO45DcQFABwFBNXH4DjO3bt3aUQ9Go0QQnTU9vz8/FtvvXXhwoULFy5kWRZFkWmakiStra3RquwkSba3t0VRXF9fv3jxYqFQ0DQNIWTb9q1bt7rdLs1kIoRqtdrZs2dzudxkx+DOzs6NGzcmCV6e58+fPz8zM9PpdPb29tbX12/evEnjZ4QQzY3v7u5+/vnnnU4HITQ9PX3hwgXLso4SfGZZ1m63t7e3wzCkbxiGYdC0qmma+4dJxnHs+36/3+98zbZt27bpjdJKpVKpVKrVaqlUEkWRtpGHYbi2ttZsNmlcLYpiqVRaWFjAGO/u7u7s7NBaeozxW2+9NTU1hRBK07TX63W73Waz2Wq1bNueRMKlUqlYLBaLxXw+n8/nNU07NEylCXDHccbj8Wg0Gn6NlhLouk6XT1QqlVqtVi6XVVXd/226rnvv3r1OpxOGISFEFEV698TzvMkx06D6vffem56epncBgiDwPG/yyvR6Pdu2syzL5XLFYnHy4tCXFG4qg6dHCMoyHKeYYRDPsRzLsE+dTU4zHEVpGKdhlNhePLRDRRJyppIzJMGQ4JYQAOCVwjAMLbGmReBBEERRNLl4e1mHJAiCJEn0uktVVTgzA/ASQVB9DKPR6Isvvrhx44bruvQjdCzEu++++9Of/rRWq+VyOZpZnZ6e/s53vvOzn/1sZ2eHxsO0v/ru3bsrKyuXLl2iQfXe3t5//dd//fGPfwyCgOZCr1y58g//8A/nz5+v1Wo03Pryyy//7d/+jZZGI4RkWf7pT3/6F3/xF7///e8//fTTra2t7e3tdrtNjydJkuFwuLy83Gq1aC36+++/r6rq4uLiUZqifd+/du3ar3/9a5oWRghNT0//9V//9ZUrV5aWlvZHm+PxeGtr65NPPrl69Wq73bZtO47jNE0FQRAEoVgsVqvVH//4xz/4wQ9KpZJlWQgh27avXr360Ucf9fv9OI41Tbtw4cJPfvIT27Z/+ctfrq2tOY6TZRnDMFmWff/734/jeDgcXr169Te/+U232+31erTInG4Cp+8ir7322htvvHHlypXFxUVJkh6+ceB53t7e3u3bt5eXl+k9jiAIJiXZtFLdNM1KpfKXf/mX77zzzszMzP6m9263+6tf/erzzz/v9/uEEE3Tvvvd7/7kJz9ZW1v75S9/2Ww2HcchhDAMo+v6Bx98gBAihHQ6ndXV1atXr3722We2bbuumyQJjclp/f/rr7/+4x//+MKFC8ViUZblJ/33CABCCBGCCCF+mA7siGMZQxNkiZeEpyqFIAj5YdLpua2u0+yO+0N/ZAfFvDZbzy3OFE1N5Di4dAMAvFpooTVCKJ/P00sUejX4UuJqmqPWNI2mFmA+GQAvHfwGHoPruuvr65ubm3SHM0LIsqxz585dunTprbfe2t/KYhjG1NTU9evXP/30U8dxPM9L09T3/d3d3c3NzYWFBfppNPX92Wef0U9ACOm6/t57783Ozk7O0a1W67PPPptE3aqqfv/73x+Px2tra8vLy9vb291ud3KEWZbRG6iTMLtQKIxGoziOj3LST9N0Z2fn+vXrrVaLpuLPnDlz7ty5s2fPTmrd4zgOgmB1dfXjjz/+wx/+8NFHHw2HQ5rInTwPXZNIV3NfvnyZ1niHYbi5ubm8vEzr1VVVHQ6HmqaNx+OrV69ub2/TmFySpMFgQAhpNpvLy8tXr1793e9+Z9s2DfL303W92+1GUTQ7OzszM3PoWPXxeHz//v3PP//897///fr6+s7OzoENk/Q9Mp/P058dz/P5fH7S8+z7/vr6Or1Jkaapqqr0yNfW1j788MN+v09T7pIkjcdjevfa9/0bN25cu3bt97///ZdffjnJ+VN0gmi326XrLi9fvkyr3OHuMnhiaYb9MN3rB+u7DkKoYEqWLpiaKEucJHICzx59M1aGSRynXpCMnKDbd5sdp9UZN7v2yA5cL66W9CzLLF2erecZFkFnNQDgVcNxHL2nT6e00D6vF5+vpjlqURQtyyqXy5Zl0WuJF3kMAIADIKg+hiRJBoPBaDSaTA6rVCpXrlw5e/bsgSFVmqYxDDM/P3/+/PnV1VXP8xBCWZaNx+Ner0cbrR/j8WfnFzx28sDX8jxvZ2fn448//tnPfra2tjYcDmkadv/n0Gbvq1evdjodWZYXFhYORLy0V3ltbe3nP/95kiSTAnhZlovFomVZgiDcu3fvP//zP+/fv9/v9w+MaqOiKNrd3V1dXR2NRlmWHfqa9Pv95eXl5eXlO3fujEajhzef0Xljw+Hw448/dhxnamrqUUlvesxfffUV7R4fj8f0iyqKUiwWaSd8u93e3Nz8zW9+8/Of/3w4HPq+f+ArZlnmOM79+/dpiXi9Xs/lckefJAfAw8I46wyCOxujT2/1kiQrWHKlIFeLaiUvVwqKqQmmJh7x31eSZAM72NgZLn/V3G6Oxk7g+XEUJ0mK0wzzHCOKXGMqH0YZy7IiD/9oAQCvHJZl6agaWvJGi+zQC8xX0xy1LMuGYZRKpWq1KknQkgPAywdB9ZHQ+5FBEDiO4/v+JNtpWdbi4uLU1NSBuIjneVVVK5WEpRqeAAAgAElEQVTK7Ows7W1GCGVZZtv2cDj8xqD6G3EcV61Wz549SxuJ6QQv+nGaCy0UCrSu+MyZM7Qo6KlqQfe9VfR6vS+++OLTTz+9ffv2YDBACJmmmcvlTNM0TZP2LdPk/NbWVhzHd+/effPNNw8sEqMTzkajURAENC1cqVR0XS8Wi+VyuVwux3HcbDavX7/e7/fDMKRbIhcWFqanp7MsS5KEfrxSqUxPT5umKQjCoc3J9D6uIAhhGMqyXKlUNE3TNI0+CW14ppvJms0mz/MbGxvtdrtcLj9cLU9vS/f7fdd1WZYVRTGXy9G3tHK5XK1WGYbZ2tqiCer79+/THwedgkZb04fD4Xg8dhyHxtvVanV1dbVQKJTLZVpRBsATiOKsMwx3Ot5uxwuidGBH/XHYHYblvFIthLWSUi+rmiJoyjef7TEhcZwN7WB9e7Cx0w/CJMv+fFfI8SKu7/VH/sgJWZYR+ScfJw4AAKcUwzA0LVwoFAghdPQs7SxDzzm0phdydN1MLpcrFAqT0afP74sCAI4IguojwRjT1GIYhvsTs4qiVCoVOun64UcZhlGtVmn7NPo6SzkcDvfXAz8ZURQvXLhgmmaxWFRV9e7duzSo5nnesqwzZ85873vfo3Hs0tJSpVJ5VncxCSGtVuvXv/71559/TteJIYSq1ep3vvMdWiW+srJy/fr1+/fvu64bx/F4PF5dXb1+/fpbb7318AFkWRaGIZ2/PTMzs7i4WKlUcrlco9Ggq8JGoxGttOc4TlGUH/3oR3/3d38XRdFoNPrss8+azeYbb7zx5ptvLi0taZp26NDLqamp999/v9vtfv7557lc7vz58/Pz87Ozs57nOY5z9epVWndAY2zHcba3t9fW1ugN4ENfgTRNMcb0lZ+bm1tcXKTD0mjF/q1bt/77v/97Y2Nj8mNaXFy8ePHiuXPnZFm+fv36ysrK/fv3B4NBGIadTmdlZaVQKBiGAUE1eGI0U90fR0mKowRnTuwG6V4/0BXX1ISlGdMNrLmarin6Nz4VgxDLsoigKE6iKMX4gavDKEqHI783cHsDVxJZQxMgMwIAeDXRSeB0m6Yoir1ejy4NpfNKn8dXZL5Gp/nQca2SJEFEDcAJAUH1kdDwj96J3F/QKwiCrusHar8phmFobDaZ70Ujczq26imPh+O4RqNRKBTCMLRtu9lsTj6uadr09PT3vve9ubk5hFCxWMzlcoIgPH1QTdd0d7tdOsQ7SRJ6fq9UKt/5zneWlpbm5uboturRaLS+vk73e3U6ne3t7aWlpYfDVLpVa2Zm5u233z537tzs7KxlWaqq0rUQNNCdVFXR5HaWZfl8vlwuC4LgOM6ZM2fm5+dLpdKj5nPouj43N/f222/3ej3DMObm5nK5nGVZw+Gw2+3Sn85kgTbdgNXtdic984e+8pIknTlz5u23315cXJydnTUMQ1XVcrnsum6z2bx79y6t9qfVWXNzc2+99dbc3JwoirQdvdVq0WZs13V3dnba7fbT32QBr6YsI1GSDe1orx/0x1GSYYxJjEmcYoRQEKZjN7Z0cbqqxcnBxodDcRyrKYKhS6oiiiKPoyTL/nx1mGYYR0lv6K3v9BWZnyrqCKJqAMC3BSYEY8IwDHeEiRE0X82yLCGEZVmO42RZnuzgfLZterTem15+0MlktFGO7j151KMIIVGSBWE6doI4yfKWoquSKHBH+e4AAE8AguojwRjTpdMHWmRpGfCjiqvpkOpJBpXOAJ8san4aHMeVy2WO4+iqZDroG32d0a1UKhcuXHjttdf2H8PTB9VJkoxGo36/PxgMPM/DGNOzfKVSuXjxYrFY5Hmefuk7d+7Qh2CMbdumYerDQTWtoH7jjTf+6Z/+6bXXXtM0jVZxY4zp0LLJZ9KbGrdu3ZJlmW4vu3LlCl08Rm8SP+qY6U/nvffeO3fuHH01hsNhr9ej27wJIZIkTRq2Mca+74/H48dEuZIkFQqFy5cv//M///PMzAzNkLMsGwQBfWVogzf6OqienZ09f/68oigY49nZ2eFw+Mc//pE+VRRF/X5/f4s+AMeSpHjkRO1B0OzRTPUD13BJRjKcJhlmGXTEX3+BZy1DLufVUl7r9Nw0zfaXfxNCMEb9offVWqeUV19fLCOIqgEA3xY4I3GCOY5hBe6IV0wcx9FaM8MwhsMhHfJKR5ehZ1QKTiNqnudFUSwUCvV63TRNTdN4nn98jhpj4npxu+/eWevYbnjxbG1uOm/qEifCLmsAngsIqo+E5lQf3ipM05uPGvxIU7uTEJre15Rl+dBC5WOZPJWqqvvjdrpHkZ7f6SKrZ4gO9BqNRpNvit6LbTabH3/8MV2s7Xme67qT2eOEEBqmHtpGbhjG+fPnL126tLS0NDMzM4n8kySJ45g2jTebzV6vR5PJ29vbGOPRaLS9vT0zMzM7Ozs7O1utVjmOe9RLSnPp9HZAp9Ohq7l6vd5gMBgMBjs7O3EcT35AhJAwDH3ff0yUm8/nL168+MYbbywtLe3PkNPF1I7jTMoQ6D2Ue/fu0VcmyzLP87a3tydjzJMkoQu3Hh6fBsBRxEnWG4WdQTB2ojDKDlRrCzwjiZyli6WcrMpHOucwDCPwjKFJ01WrP/RdP4riB34XCCG2G263hp1+2fETWeREAa7PAACnG8YkzcjQjjqDgGGQLPGWLuYMkWMZ9rF5XYZhaHBLb68jhDRN833f8zy61WWSsj5WgD3ZCUIv9ugsGFVVLcsqFAqKojy+/JAQ4oeJ7UYbu8P1ncHqVs/1Y4KYKE7PzBRKeZXn2Md/XwCAJwBB9ZHwPG8YBp3wvL9hJkkS13UPLJSiaExFW4vpR1iW1XWdbj54oUf/jERRREdt7U/tEkJu3Lixvr5O3wPoRLdJuzWteN8fau6Xy+XeeuutS5cu0XrvyTsEHftBy8K/+OKLXq9Hn6rdbtNd36qqTk1NnT179ic/+cm7775bq9Ue9ZLSpPfGxsZvf/vbL7/88ubNm7Q6i1aSx3Ecx/EkpqW3SB5fSlAul99++22aV98fydOgmhZ+T55tPB7/7ne/u3bt2v7nn7w4dJz4/rl3ABxLlODuMOwMQi9M0yw7cBKSRa5gSdWCUi+rpnaMc44iizO1fG/ob++NCIOYB5/WC6IwTroDd2iHeUOCoBoAcNqlGQmitNX3V1aHWUZUmV+YNmTJkkXuKMHnpGhR07QgCGzb7vV6rVYrCAJCCL1SQkeOq+nVFB1DyzCMqqrVarVYLJqmSbMy9K8e8wyYoJEdbu+Nl2+3bq+2+yM3TrIgTFw/UhRB0yRV+oabBQCAJwBB9ZHQM6aiKJqmybI8qQOnbbFTU1PVavXhRw0Gg62trUlmkpYJ5XK5xwTVk/PvC955eBSTruYDx5ZlWRRF+0/xkiTRyVuiKJqmqev6oUuk6a0KWsW0/+H0jWRhYeFv//Zv8/m8YRitVmtvb49GpL7v27ZNbwNzHOe67v/7f//v/PnzdC32gS+xt7d3586da9euffTRR6urq1tbWzSCpY3ZdK2053l0zBv6esT3Y158QRBo5dWBinoave/POdMfIu083/+tTQbXWZZlGIaiKDBlBBwXxiRJse3F7UHQHYVxfDCiRghpijBVVEs5SVcFUTjGvzFJ5Et5vVoycqbiBVH84MQynJEMZ+2ec+te++x80dRhxh4A4LQiBGFMbC9udv372/a9LTtOM0nggjglBFUKcjkvCzzLc99wCqV1gjTipYlrURTDMEz3SZKE/gdNP9CLjUm/NMdxNKMgCIIgCDzP0z+qqkovhBRF+cY6R4KQ40ZDO7iz3r273l3fHrR7dhAmGcadvstzrKnLGOOl2WLeVJ7dqwgAQAiC6iOip0tZlnVdV1V1Ej45jrO+vk77Zg88hBDS6/XoFmX6EY7jdF1/fFCNXvga6mOZlCTt/6OqqoVC4VEbrcrlcqFQOHS6NX0veVRIeebMmenp6VqtNjU19X//939hGE66nelyMtrG3Gq1qtXqzMyMrusPv99sbW394he/+OSTT1ZWVlzXpa3aoig2Go1z5871er1Op0NvJx/9FXjUMT/84vA8b5rmowaJ011cdN3lEb86AFSGSRClQzve6wX9UXjoHDJN4etltWDJsshx33RFuJ8ocIWcWi4YhZw2doJxgjF+oJiCIajdc768tWNowtJs4Wm/GQAAeEkwIWlGBnZ8b3v81cZoremEUcogxgvSKMKvzVuqxKmK8I1B9QTP8/RCsVQq0Wq4KIqiKKI38cMwpAN6aBtdlmU0lqapCDomRlVVRVHoR0RRpAmDx1wsPYCQ/ihY2x58cXP3+p1mGMZRnBKCCCK2G9IwPkvTcl6DoBqAZw6u5o+KTvOu1WqVSoUu1kII9Xq95eXlqampCxculEqlyRhwGqqtra21Wq1JSTDP8+VyuV6vT+aKTe5QToIxui+KztCanIUfkzud1AjRP9Jl2kEQHL2i+OgBPN35TGvg9398fn7+Bz/4AR3cfeAhPM+XSqV6vT41NXXcImd6I6PRaLAsWy6X33jjjWazubOzs7a2tr29Tfuux+Nxu93e2dnZ29trNBr771bQ1HGn07l169bGxobjOLRq/fz589/73vcWFxcXFhaWl5eXl5fpKLVjHdvDVFUtFouK8sC7FB2rdunSJVVVH76Tout6qVQ6c+bMJHcNwBElKe6Po/bA79uR66fpg235PMcIPJs3pHpZtXTxGyoFH0JngOctuVLQBiPP86KHa8ttN9xujTp913YjSeQlmHwDADiF/CDd6wf3tsZ3N+1m1w/CNEkxQqg3ChkGEUQwIfWyWiuposAJ/DeHtZPibZp8FkWRzkPVNC2O4yRJ9uerJwNfaV6a5qhFUaTF5NQ3FntPYIKyDLW69q37e7vtkeuFNB1O/5ZgHIZJp++sbgmFnBpGyfRUTlfEZ7FuFQCAEATVx6IoCo0P9/b2aKhMFyDPzMy8//779M4iPfe1Wq0vvvji3r17rVZrErUKglCpVOjIaPoRmsw8EFT3+33HcehCqV6v57ruwxXXEwduXtJpWJ7nHTGCPVafjyiKdIvDgYTwuXPnfvrTn87NzZXL5YdP/fQIeZ7f3Nw8yleZSJLE87xisViv1y9fvpwkyd27d1dWVn7+858Ph0Oado7j2HGcVqvVbDYLhYJpmpOH06L0fr9///79Vqs1Kcx+8803/+Vf/qXRaNC9XDQmP9aBHUrTtP07ySlFUd55551//Md/pBnpAw+ZJL0hUw2OK0lxbxS2esFgHHlhcuA3mOdYTRaKljRd1nK6eNzJ/zzHcDKfM6RKUW/3nG7PCaMUoQe+hutHYZx2Bt7ADvOmDEE1AOA0coNkdce+vT68uzkae8kkBB068diL4wT7YZqkWFcEQ0NHCaonJhd49Jb6gcutA9ddk7P0ZD7ZgY8fBSEozUizbd+82xqN/SR58DqQoCTJRqNgDfcxQUGU6ZqsKiILWxwAeEbgav4Y8vn85cuXh8PhxsaGbdt07pRt28vLy//xH/8xPz/faDRognRlZeXmzZv37t2jsRzHcZZlzc7OLi0tLSws6LpOn5BuwFIUZdJ3TaP0LMvoSLDt7W36x0OPhy5t2t+xnKap4zjb29uffvqp53mSJBWLxampKZpCp4PB94dw3W739u3b5XJ5dnb20Art/Wg7cbFYLJfL7XZ7MmFrc3PzV7/61fnz55eWlmgjEN0sJUmSruuLi4vT09PHep3p7duVlZVr167R77HRaNRqtSiKDMPQdV1RlIcLtg+892CMaTabrhafvIH5vt/pdIIg2NraWl5e/uqrrwaDwbEO71CyLBcKhWKxWCqVJk3aYRjeuHHDNM2lpaVarcbzfJIkg8EgiiJZliuVyuLi4sPBNgCPQRDCmARR1htF3WEYHdZNrcp8taBUCkrelBSZe4IrJoZBqiLWq1an7241h36YZNkD9TK0qbs/9Na3+8xMwYLOagDAqeKHaW8U0oh6a8/1H1ygQAjJMmR7yV4vqJdVL0xl6UluHT5ZePxkGAZxHFMp6ecWyvc3e24QE3ywnRATEoTJXsfWFLFgqXGS1sqmIkEsAMAzAL9Ix5DP569cuTIYDD788MPd3V0aP6dp+uWXX965c2dhYeHs2bNJkvi+f//+/bW1tUl2lOM4Wuh77ty5hYWF/WOu6ZqESe630+l8/PHH7XZ7Y2Oj1WrdvXuX7j0+NJnMsqyqqoZhHAiqNzY2/vCHP7Tbbcuyzp8/b1kWTaHTvp39mxj29vauX7++sLDw+uuvf2N0NwmqK5WKZVm0I4gQQg/yzTffvHz5Mh3kdufOnbW1tVwuV6vVBEF4gqA6SZLl5eV///d/T9OU5/kf/OAH7777Lr19QPuO9t8aoEVWhwbVByJqhNBoNFpbW6N57GvXrl2/fv3QyeTHpSiKLMvlcrlSqdCgnRDied61a9eazebly5eXlpZUVfU8786dO47j0MnnpVIJgmpwPARhTMIo6w7D3iiMkkPuuKkKP1VSKgU5b4hPnENWZWG6au11HUUWWTbcv7AafT1SsTd072/2CpYyV8892VcBAICXwgvT9aa7sjZaWRv1x+GhFXt+mLYH4dCO/TAxtUPmrZ4oLIMEDjWmrCietr1oZ2+MySHNg3GcdQcuwzKSxLMsyhkKBNUAPBPwi3QMdFr1mTNnfvjDH0qSdPv2bcdxaPDm+/7u7i7dxkSzkTSLSzuxC4XClStXfvjDH05PT++P/RRFmZ2dbTQa3W6X1pPTmmf6VI7jjEajJEk4jqPJ2wPHQ2P1RqORz+dlWU6ShH7R0Wj05ZdfbmxsyLLc6/XK5TLLsnR0ZKVS2d7enhxDv9+/e/fuvXv3aLR/6IzuCfqo6enpH//4xyzL/u///u9oNMIY+77f7XZv3LjR6/VoI1Cv1xsOh/TLnT9//ty5c8dqG242m3fu3Llx40a73U6ShGXZjz/+eHd3VxRFhmHu3bvX7/fDMKQj2XO5XL1er9frsiwf+GFpmmZZVi6Xo59PX8C1tbX/+Z//obM3NzY2aNR99GN7/Ivz+uuv//3f//2HH344GAxo1GHb9vb2dpqm9+/fFwQhjuN+vx/HMa0duHTpEp0lDhXg4IiSFA/sqNn19vr+wI4OjChjGMQyjKEItbKaN6SnWZoiClzelEs5tWCpjhuOnSzL9l2cEcQg4npRu2cPx4EXpKLAHqs2EgAAXoowzvrjaLPp3F4fbbZcP0ofjqhZhmFZVM7Ljao2XzeqBVVVTnpQjRBCDFMwFdzIO16NY5nNncFe1ybkgWpzQgjGyPWird2hJHAiz8038lMlQxLhOgSApwK/QsdAh0bMz8//8Ic/TJKETo2m0ybCMNzb29vb29v/+bTiWtO0qampK1eu/OhHPyqXy/s/QVXV2dnZmZmZr776iq6/pqlv3/ebzSYdd0FnVxxY1zQ5nmKx2Gg0isWipmm0+xohNBqNJiPH0zS9fPlyoVBQVZX2/ZqmuT+otm37/v376+vrhUKhVCo9/hVgGKZer//VX/2V53nLy8u0sTmKojAM+/3+rVu39n8yx3HFYvH9998fDAYHlmY9XqvVunbt2q1bt2j8SQhpt9uffvrpgSPheZ4OHqdDwg8MCaOF6JZlFQqFdrs92Ue9sbGxsbGBvp4m8gwrshiGuXDhgqIo7Xb7yy+/pOvHXNd1XXd3d/fAZ9Kl5bu7u41G40DiHYDHSDM8GIe7XX+vHwzt6MDfMgzDsYyhCrWSmnvqoFoUuGJeLebV/tBzvfBgGwpBnh+1e87ADlw/0TUBgmoAwAlHCArCdLft3dkcf7Ux2uv5cXrIjXWWZQSeqRTk1xdyC3W9Ujgds7IZhHKmrGsiQkSRxThOe0M3TfGBuwaEEM+PtpppmmHEMIhBOVMRBR6GlgHwNOBS/tgsy3r99dcZhikWizdv3lxZWen1erZt062D9HN4nqfbpKampt76WrFYPJBNNQzj4sWLtHc6yzIaQyKE6DTIRqMxPz9Pw9HV1dXV1dWHD4bn+UKh8MEHH6Rp+vnnn29ubtKq7MknRFHU6/XG43G1Wp2amnr33XeHw+H9+/dHo5Hv+wghQojrut1u94hrpWRZrlar77zzDq3QXllZ6Xa7w+Fwf8xPv/2ZmZlz587Nzc2p6v9n70yb47iyM31z35faC6jCDhLgJkokZZotySOPZtRuO8JuR8x8mfkz/jUOOzwx0w5Pu1vdHqlNSa2F4i6SIABiR6H2Jfc958PpLoMFiAIligSp+3ySgMysmzcvC/nec857RBCNR1Sw4+Pjb7/9Nijezc1N8PoG726AoijILT9//vybb745NzfHcdzBfloEQUxNTf3VX/2VpmlffPFFv98HiY4Qgi2Gqampcrn84MGD4dx+T6WtqurU1NRPf/pTRVHu3bu3srICnmr7h0TTtK7rExMT586dKxQKPM/jPtWYowO+382u6/nRwd+KHJ1RWfCqVWXme2jqP6Crwqm5kudHnYEdHEg19/xoYHrtrt3omCQpSzz+g4LBYI4vQZj0DH+nYT9Y76/vmoYVRPHhTUyzGlctiKdm9MVpPau9ZJ4RJEnkdDFO0l7fjuJktz7o9OwRc7Q0RXGc9AbOw7UmRRICx0yM6YWsdPTmYRgMZgT8DvTUqKqqqmqxWDxx4sTExAQo3r29Pd/3h8JPEARRFE+cOHH69Omf/vSnb7/9Ns/zB5sqKYqysLDgOM7Kygq0zoJmWlA5fPLkyStXrjAMA50Moe0T1O5CCTFEwnVd/9M//VOWZcH327KsoXSEAyzLgosXi0We57e2tm7evMkwTKfTSdOUpmmo/g2CgCAIcBdTVRVEsizLEEcd6kzoo3jhwoVKpQLF0uDHBrF0OIzneUmSTp8+fenSpaGohkx4RVGGPcZkWQbjtBERWy6XM5kM3OD169c9z4M+YQihNE0JgoBG06dOnXr//fevXLlSKpW+yWWtWq3+l//yX5IkqdVqFEWZpgkzk81mC4XCm2++eebMmSRJYG4VRYGa8KE+Byc5mBDYF5AkCQT8ocIbrvDuu+/Ozc3967/+a5Ik29vbJElCXffQK65arV68ePH8+fPwRLCoxhyRJE2DMGn3vWbX9YNDqqkFnirnxLG8WMzy8rNIVtQU/uRModN37j+qO26QJo+9ffpBnKR+p2c32oYmsygz2lQPg8FgjglJknp+VO84j3aNla3BdsN2vCiKR8PUJEGQJFHI8Isz+sKUPldVnoPH2LOFJIiMKggcY9t+GKW+HxmmGz1uNpmmaRynA8MdGC6JCEUWaJpSZZ7kCfJlu18M5piARfV3RBAECNgWi8Ver2eapmVZpmlCkFlRFFVVM5lMNpudmZk5NI6KECJJkmGYmZmZv/mbv7lw4QJ06gqCQBRFiMRWKhWSJNM0PXPmzM9+9jOEEMMw586d29+9iWXZcrl88eJFVVXfe+89CEFDJbYgCFNTU2fOnKlUKizLpmmqqupPfvITXdebzWa73QansRMnTpw4caJarUqS9N57701NTdm27fs+QkjTtPn5+YmJiaFjOQCmXH/yJ39SKBTa7Xa/34cZgC6LiqIoigKxcejLTdN0JpP5i7/4i8XFRbhHhFAul5ufn69UKiOZ29CCYm5uThCEc+fO/exnPzMMw7Is2LYQRVGW5Ww2WyqVZmZmoD34Nz0mSZIqlcq7775bLpdbrVa324U5h0czNjYGLbv+4i/+AmZybGwMuqbB6WNjYz//+c8vXboEU4oQKpfLYOX9TfXnBEEoijIxMfH+++8vLi72+/3BYACDZ1kW3Nph8NDznOM4LKoxRyGOU9MJ6h2n1nKaPc8LDklZFDi6lBMyKkd9/yA1QgghnqXzGbGUl0s5JQoTy/YOGOQm7Z69sdMt5+UnXAeDwWBeIFGc9M2g3nYebg5Wt432wPOCOE4OiVHrClvOCadn9DNzmWJWQC+twqRpcrykJQg5XhBGcatrmdbhfmw9w3mwWqdJQhTYYk7SZf77lA5hMD9aiCP2KMY8AXC96vf73W6XoiiGYXRd1zTtqS7ieZ5hGNCNSVGUQqHAsuxTya00TT3Pg3xjz/OG4laW5f2SHgy0er1et9uFtOpcLpfJZFiWfbJR2TcBJtvdbrfb7YI1t67riqI8w4rlwWAwGAwcxwmCQNd1VVWhK/gRTx8+oHa7DVkAqqo+7QP6DqRpGgSB67qdTsd1XQjga5omCMJLt/ONeeEEYVxrOavbxtUb9YebgzBO9utbsChbmNbeOl86OalNjSks88w2a249qP3205XVzXat0Y+iEWs0opRXqmP6+2+f/MmFaZJ4Hp1jMBgM5ugkSer40XbdXtsxbq9213ctww4OJvuQJEGRxExFOT2jn53PnpvPsDT1Un+fJSkybP/61zu3H+wur7X2WkYUPdY5DGAYiufY0ydKl1+fmpvITY5pLENhXY3BPC04Uv0MgLReWZYhkxkCrU97EZqmIdc6iiLQt9/h3ZRhGNBscRyDOTbDMAeVOUmSkiSBqXiappCafmgs/ShALrSu65A4DQXVz9YDDHKkFUWJ45jjOPAYP/rpwwdEURSkEhxdkH9PaJoWBCGfz0dRBBsu32FtYDAIoShOe4Zf77iWG44k8iGEOIZSRKacEysF6RlGqgFdFRZmi64XtjpmHCePF+alhu3ttYxW1+4ZniwwPO7OgsFgjg1xkppO2O65j3aMla1Bo+1aTngw6xshpIhMMcMvTGqnZ/WxvECTJHrJdSVBIIGj56fy9B/efohmx7LtUYfLOE48P9ytD766u+16IU2RWV3QZP6l3lDAYJ4/+O3nGQACUhCEkTTmpwKsxb/nMOAiT+5fBVr3e452PyRJgpf1SIr4M4TjuO8jg5/JA/punws9tJ+bhse8qiRJGoRJZ+DXO67tRgezFjmWyulcKSsUs4Iqsc82yKDK/Oxkrtkxlx7RYRiPvI86ThAnaaNt1VvmWG0EwOIAACAASURBVEHBohqDwRwTkiQNwrg78HYa9tqOubZrdg3fPeDySJIETRFZjZupKPMT6lxFVSSGfLmj1AghRCDEMlS1pEkC6/lhEMZ+EHl+mDxuz5YkaZJErY5pOR5FEqrMpygjCSxF4fpqDOYpwG8/GAwGc6xJ0tT1477pN7pes+u6/iEWZTxHlbJCXucFjqIP99H77vAck9elQlbOZaQ4SU3LH4mTJ3Gyvdf7enmPZcic/nL0nsFgMK82aZo6ftQzgvWatbI52G5aPSPww0Ni1CJP5zVurqKentEnSpIo0NSzzfZ5cRAIkQSSeObEdIEgCNePwjAxTNcPRncWojhx3HCr1iMIFEWxpgiSwAh4kxSDOTL4XwsGg8Eca9IE2W4InbTafX+kFBCqqWWBKeXEnM5x7LOvhWMZSlOEnC4VcrLrhbYdxAdEdb1lcCw9Oa7PVDMUSeJ6PAwG8wJJkjSK074Z1NrO+q75aMdo9jzLDUcOgxh1RmEny/JsVZmpqjmV41jqVYrQEgTiOLpa0iiS7PRtzws34ySKof7vP77J4ziJ46TZNm03kEWuVFDH8jLPyoh42bPgMZjnBBbVGAwGc6yJk7RnBrWW0zV82x2tBqQpkmepjMKWc0JW5X6ILqMUSRA00hSuWtZNy291rCR9zOQySZFhefWW0erYPcNTJA7HNzAYzIsiRSiIEtuJtuv28uZgo2Y2up53IDaLEOJYKquwU2Py4ow+Pa5kFJZjqVdPRRIEoikio/KvLYwzNBWEURBGrh/G0WglURDGqeWvbXUokrxwpprTJZIkcO9qDOYo4PceDAaDOb6kKYripDvw99rOwAr8cDT3m6FITWJzOl/MCprM/hBJiwSBKIpQZb5S0ppti2HoJAmj+LHeWo4bdgdOo23WWxZFklhUYzCYF0KaoiRJTTtsdt2Nmrm6PdhrO4YdjBxGEgRNE5rEVIrS9LgyM66UcoLEMxT1qilqhBCBEEEiSWSmKxmEUKtjBWHc6Ji244/UV0O8eq9p+EGsyny5qGkyp8os+SzNZzGYVxPq7/7u7170GDAYDAZzOHGc2G60vDVY2R60+77jjQZbZImpFMW5qjpf1TSFpX+wvMUoThAie4a71zKjOImiQ7LQBZ5laCqjCRkNV1ZjMJgXQJykQRhv7JlLm/3lzcFm3XLcQ1pSsyylyexEUT41o89V1fGCpIjMD/f9eRwgEAHp7qLIsixtu77vR3GUJAd66yZp6vtRnKaW47MMldEkgiRelTJzDOaHAgcTMBgM5piS/sGiLGj1vGbXO8S0liAEjipmhWKGl0WapX/ALD2RY4pZKadLmsKHYeR6j0V+0hRFUdLqWlu1/kw1G4QxTeHKagwG87yx3bBn+Bs1a3lzsNt0+uZojJogCJoiZIEZy4tTY/JMRakURFVimB/y+/M4AEngqszNTuQQQqblxnG6W+/HbjjiPRmGcRjG27We50c8x2iKmNOFjIr7mGAwTwKLagwGgzmmJCnqW8Fuy6533M7ADx8PDpMEQVGEJDClrJDT+R/6jZBjKZ3kcrpQyEiuG/QNZyS8kaSpafutntU1XNMOZJHlWOoHHRIGg8GM0Oy6y1uDpY3+ox3zYGoPQoiiCJFnillhfkI7MalWCpIms69k1veh0DSpiEylpJ4/VSEIYmC6nh+hJD0QrkaW7UdR8vBRgySI1xbKGbXwIsaLwbw0YFGNwWAwx5Qk+YN7bc88pLcqRRGyyGRVrpQTMj+MRdnjH0dSFJnVhYlx3bC83cYAocfewtIkdbywb7jtnt0ZOAxNYlGNwWCeG7YX2U60XjPvr/W3G3bP8EcOIAhEkoTI0aUsP1mWZitKtShpyo9r+48kCJahMgo/U816ftjpOzRFtrt2EEaPf6OjIIzCKN7e6ycpkkVWU3hV5iSBfUEDx2COO1hUYzAYzDElTdHACupt13EP861lqLzOjxfEck7QZZamn0ekJatJJ2eKzY6FVg6MFqEgjEzbb3bMRsvUJE6VcbogBoN5TrR73tquee9R/8F63zmwC4kQIgmCY6iMws1WlJNT2vSYnNP5H7Rq5tjCsnROE05M5SmSlEXuyztboRGPRKvTFKE07fRszwtFnkEInZorzlSxqMZgDgeLagwGgzmO+EFs2GGr5+61HfuwJEaWJQs6X8oKqvT8Ii2KxE2M6YWsLEuc74fBPjfyNE3jOPH8sN219lrGeEnNJyn2jMVgMD80jhfZbrhRM++v9bbqVvdAjBqBomapvM5PjklzE+r0uJzROIH7EcWo90ORBMXRhYxEUaTrh62etbNH9gdOFD2WCJ4i5PlhEEbr2x2EEM/SksAoEifwzIsbOwZzTMGiGoPBYI4daYpMJ6x3nFrL2es4njfaSQtBpDrD5XWOZZ5fpEXgaYYRC1kpn5G7AzsI3f2/TZM0DKNO36m3DNMOwihlaERhVY3BYH5Ieqa/uWcubfTvrfXNA92z0B/7AkoCMz0mL87o8xNqOSewzI9UUQ/hOLqQEecmco4bcgx9f6VuO34cj1ZXp2naaJuuH4oCw7DU/GQOi2oM5iBYVGMwGMyxI0WpYQV7badjeLYbJY/3g6FIgmOpjMqN5cT8801fpCmSpshSXlmYLa5stAzTG8kYjONkYLiNjtXtO5YTKCJD/ZjqFTEYzPPE9SPbjbb2rKX1wWbd6gy8g5oQIURTZE7jqiV5flKdqyp5jRc4/AKMKJKgWLqQlRZni1EUW45fbxrdvj3yFydNkeeHcZJs7HQ5lmZokmUoSWR5Fs8hBvMf4H8PGAwGc/xI0cAKdlu2YYVxPNpGlKbJjMKWc8J4QcrrPPPc4y3lvPLGmYrt+JATuF9Xx3FqWF67a7d6Vs9wOEb+UZkAYTCY54lphztNe2XLuL/W7xp+GCYpOkRUswxZKUqnZrTFaX2iJOEY9X5UmeM4OopjLwgJAvVNN0kOyY2K42S3MQjjhOcZjmMmyxoW1RjMfvC/BwwGgzlehFHiBXF74O+1XMsJD3Y6YWkyq/GlrKDJLM/R1HNvB60p/HQls7KhaArveqHnh8NfpWkaRYnl+LXGoJiVVYlVJGxsg8FgnjF+EDtetNO0H24ONutWq+95/oENSIQoitBktpwTTk5qJya0go5j1KPQFElR5FhBCcNSHCeOG7R7tmGMZiGlaeq4QatjLa+34P8pihR5Bu9QYDAA/mbBYDCY44UXxAMzaPW8Wtu1D/P9Zhgyr/PFLC8KFEW+ACcwSWR5ji7lpHxWanft/aIaIZSmqeeFW7WeKnOTY1opJz/v8WEwmFcd24vqHWdt13yw1m90XdsN4+SQGDVDkeN5cXFaPzOXma0oWAEeCoFQThcViUOI8IP44VrTsDz0eLg6TVEUJZbtray3TMsXWEYSuHJexlOKwQBYVL+y+L5v23av12u322ma8jyfy+WKxSLDMCT5Y2wgcShpmlqWZZpmu902TZPjOFmWi8Wipmkk+SLEyh9xHMdxnDiO4zhGCBEEIYqiIAgURVEU/gP2imM54V7Hbfc9wwr8MNn/K5IgGIZUJbaUEwoZgWdfjAsYRRIUSeUy0lQlG0ZJ33BG4kNhFHd69l7T6A7cshvyLEX9wG20MRjMj4QgTDw/2ms7j7aNzZrV6LqmE8bxaJCaJAlVYgoZ/sSkdnJKK2Z4gaOxbeI3AX4ZlZIahONpkpq2Z5ie5QT7JyxN0zhObcdvdc2Ha02KIpO0RFIkz1I0/obH/OjBovqVxfO8er3+8OHDO3fupGmayWTOnDkjy7Isy1hUD0mSpN/vb21t3blzZ3NzM5PJVCqVixcvSpJE0/QLlK+WZe3t7QVBEAQBQogkyVKpVCgUOI7DovqVZ2CHO0273fcsJxypDyRJQuRoKKguZHjuhYYIspo0P1Xo9p2t3e7jhdUoipL+wG20rXbPMSyfUnksqjEYzDPBD+LOwN+uWw/WeztNp2f44QFFjRAiSaKQEU5Mqqdm9JOTqsBjRf3tFLKypop+ELW61jbqm46PEPG4rkZxnFiW/2Ct6fqRJHGKxFMaj0U1BoNF9StImqZpmtZqtf/3//7f3bt319bWkiSRJGkwGLAsOzMzU6lUsK5GCKVpGgTBrVu3Pv7447W1tUajIUlSpVKBCaxUKpqmvaixra+vX716td/vG4aBEGIY5uLFi+fPnx8bG8vlci9qVJgfmiRJ4yTtG36taRt2eLA+kKaIjMqWskJW5WSBebFKVZG4clHN6qIscZ4fBcF/ZKqnaRrFiWl569sdTea52TyP6xgxGMz3I4oTP0gaXXd911jbNWstd2AFUZyOVP8SBFIlNqNw8xPqwpRWygoiz9AUltTfDk2TFEVOjOkXzk5yHOOFkeMErheOzF0cJ7bt19vG0qMGRRAnZwrFnMTSJPncDT4wmOMDfst5NUmSZHt7+5e//OXNmzdN04zjmCAIx3FkWaZpemxsDItqhFCSJL7vX7t27R/+4R8Mw3AchyCIcrnMcZwoirIsv0BRvbKy8s///M+7u7vNZhMhxPO8aZqyLIuiiEX1K0ycpEEY9w2/1nQsJzx4AEOTOY0bywsZlRVfdOBFErlSTsllJE3h09TdL6oRQmma2m7waKstiWy1rGZ18UWNE4PBvBqEYWrY4V7bWdocbNTMRtc9NEhNEERG4abH5YVJbWFKUyWWYbDaOxIEQgSBxouaoohxmrb7dqNtOl4wEq9O0jQIw17fvr9Sj6JEVQRZ4imRwKIa82MGi+pXEygGTpIkiqIwDKMoQggFQRCGYZIk6SF5Ui+M3d3dW7dudbvdNE0FQVBVdXJycn5+nmGYH/qjCYIgCCJN0yiKYKIQQr7vw4y92FmKosjzPNd1HcdBCCVJEgRBFEVJknzruUcHAvXLy8tpmpIkqapqoVCYm5srFArP8FMwR8dywnbfq3fdZt91vMc0KoEQSZI8R+V1vpgVjkNxIM9SmsLldLGQU/wgNi1/5IAwjLt9p9EyO303q0sChyurMRjMdyFO0jBKOgNvY89a2Ta263Z34IfRwWweJIuMKjGzFeXklDZeEFWRZWks9Z4OlqVUkpifzCVJcm95L44S1wv9UUNKFIZxf+Bs7XZviqzrBidm8jldZGiCfOF/nDCYFwEW1a8mBEFQFMUwDMMwL9Bt6yhsbm7+7//9v1dWVqIoyufz1Wr1z//8zycnJ5+PqCZJkqZplmVHQvfHfNKeFZ7nffrpp7/4xS/iOGYYZmJi4uzZs7IsY1H9ojDscLdp77Wdds8bfVckCIoiBI7O63wxw/PHoPkzx1I0TeZ0sZRTun0nJRBxwK6sN3CbHavdd0p2wNA8NgTAYDDfgThOPT9u9b2VrcHK1mCnaY9sOw5RRWa8IM5NqIszelblRAG/6D41DEUwFDU7kS3l1ShKWh273bNGRDVCKIoTw/LCKHb9wPECTRVkkaNImsTf85gfJfi75hUEBGGlUvnZz35WrVbX19fjOBYE4eLFi2+++Wa1Wj0mTldxHIdh2Ov1tre3NzY2kiRxXZckyX6//2zjsU8AapVt297Y2Gg0GoIgjI2NvfXWW4uLiy8w9/v5EASBZVn1en19fT1NU47jCIIolUqe573oof14sZxwt+WYdngwTYKmiKzKjRfEcl7MajzLHIuQL0GgrC5NVXP1trm910+SJN3X1SZNURTHpuPv1vtZTZAFmjsGewEYDOYlIk1RnCQDK9iqWytbg7Vds9XzguiQlwSRp0WenhqXT0xq1ZKUUVj8hfN9YGmSEOi5yVwQxLcf7JqWlySj5esIoTCKB4a3XevffrAbhuHCTEGVuRcyYAzmxYJF9fdi+OWSpukwtnmUIOf+E4enHDE6CqfsP/HguQRBVCqVn/70p7Ozs7dv3wb379OnT7/xxhuiKD6hoPrgxb9pVAd8QYj9vxr+FlKsD71CHMeu6/Z6vVqttr29jRCKoojjOMMwhjnqT56To4/2UAiCYFn2woUL5XL59u3bQ/fvN954Y2pqiqYP/9fxnT/0e472iNc/4keASZtlWc1mEyZfEASWZdvttu/7+x/fUT7uqRY/5ptIU2Q6Ya3lGPYh1dQURWRVdrwglbNCRj0urywEIrKaOFXJPlxrMjQVhmmM9ovqNEpSy/F3G/18Rpooq/hlC4PBPBVJmoZh2jP9RzsGlFKbh/lNEAiJPF3U+Zkx5dSMXswImsw+/9G+SjA0ydDk7ERWlvhO317bbodhHEXxyGFRlBiWh+p9gkgpMq0UVUVi8csA5kcIFtXfkSRJkiRxHAcMrhzH4XkeGlYpigJ5xd90YhAEnudZluU4DkQF4VxVVTmOo2n6CUIojuPBYGAYhuu6QRBwHAd1yAdbQJEkybJsoVA4e/Ys9KnO5/NPkNNxHEdRZBiGaZqu64JhONwRz/Mj2hhuAUp84WZ5nhcEAS7S6XR6vV6SJBRFFYvFbDY7cjqUMVuW1Wg02u02dI2CMXieZ5pmp9OBCmdRFEXxEH+jJEnCMDRN07Is27bjOBZFUZIkRVF4nqco6uhf6AzDSJI0OzubyWQEQdA07dBPhJ0Cx3EGg4Ft267rUhTFcZyiKKqqQqb9odeHE4MgGI4WqsdhqUCLrO/55wdi/q7rDhcVPBFFURRFYVl2ZFHB8Z1OZ3d317Ks4TiDIHAcB3qbEwTBMIwoigfvCxa/ZVmw+F3XlSQJJl+SpBfb3/vlxXYj0w5qLWe3aR/6ykiTZEblSlnhWNloEwSSRaaQEfK6qKuCYbmOEzx2QIrCMO4O3HbfNuxAkWOOwQ6xGAzmSKQI2W6007BXtgbLm4O9luOHo6IOIcQxFM9REyVprqpOjcvFDC/yOEb9bBA4Oq8L506WUZo+2mrv1vthGEfxaKaAF0SNjrWy2dYU4YRdmBjTRZ7B7wKYHxXH6OXsZQGUbRAEruu22+16vd7pdPr9vqqq+Xy+VColScLzPMdx+9UF6JAoiiA82O/3W61Wt9sdDAYEQWials/nx8bGdF0Hg+6RDO00TcGq2vO83d3dvb29Xq8Hbt66rkObJUmSeJ6naRo+F04RBKFarUJQEUTvwczqJEkgaAy9kZvNZq/XgwrnQqFQKpV0XYcrD7Wx53mtVst13TAMoSw5l8sxDOM4jmVZa2trm5ubSZKwLPvaa6/puj4itEBRt1qtjY0N6MY8/Lnned1ud3t7G/YmSqXSiKSP4xi0t23b9Xq92Wx2Op0oinRdz+fz5XJZ13VBEBiGOaJYhQnJ5XKQ782yLEVRcRwPHwGo4jAMgyBotVq7u7vtdrvf77MsK8tyqVQql8uyLMO+xkH5OtwpaDQaMNokSYajBQ3PMMzTalEYFVzf933TNGFR9Xq9fr/PMIyqqsVicWxsTFVVWFTDLRUYz+7u7tra2mAwGM5DEASGYdTr9e3tbYIgZFkul8v77wiWse/7rus2m816vd7tdvv9fjabhaVSKBR4nocadSytnwrbDesdd6/t7LXd8MD7CkkQDEPmdL6UE4Rj9rIo8gxJEPmsmM9IYRiNiGqEUBjGvYHT7tkDy8/qEUMzJMJrA4PBfAuQ6mLYwdqusbTRf7Rj9MzRrxeEEEEggaN0lZsoyadm9PGCmNP45z/aVxWBozmWOj1fVGQ+TdOB6Vm2FyejFnFBELW7IUKIIqkkRZoqMAzFUNizDPMjAovqpwbKUB8+fHj79u1ardbtdl3X9TwPgsa5XC6fz1+8ePHChQuapimKghCKoqjX6zUajdXV1fX19Vqt1ul0IKLo+z4kISuKkslkFhYWLly4MDk5WSqV9qcfg2S6fv36rVu3ms1mq9UCk2qWZTmOk2U5l8tNT0/Pz8+fOnWqVCqxLLu7u/vb3/4WVBNEfc+ePXv58uWZmZlqtTrUV2maNhqNWq1269at+/fv93o9wzA8zwNBrqpqLpebm5u7dOnS9PS0pmksyyKENjY2fvvb3+7s7IB+EwThzTffvHLlyrVr1z777DPYZUAIaZomy/L8/PzIHDabzY8//vjOnTubm5srKyu9Xg9+7rpuq9X65JNPms0my7IEQbz33nt/+Zd/KUmSIAgIoTiOG43Gzs7OnTt3lpeXYVcC4uowD5qmTU1NnTt3bm5urlqtwllPIAzDr7766pNPPhkMBhCzzWQyf/qnf3ru3LlqtarrOkLI9/3BYPDo0aPr169D6bVlWb7vQy4ABGmnpqamp6dPnTq1sLAwjFqHYdhoNLa2tu7cubO6utrtdmFuoYAZRjs7O3v27NnZ2dlqtcpxT5EZG4Zht9ut1+srKyubm5u1Wg32WTzPC4IAxqaqaiaTOXPmzIULFyqVSrFYhOd+48aNTz/9dGNjY2NjY3l5GS4Yx3G/33/48OH/+l//6+rVqwih+fn5v/zLvzx58qQoirAaDcNoNBr379+/detWo9Ho9Xqe53meJwiCIAiwqXTp0qXXXntN0zRJko5+OxjLiXZbTt8M4gP+/CRJqBJTzgnjebGcFYTjFKkGKIqYGNPPLYwHYdTp2+jxl60oTmzH7/WdZtvMKILIUTT2AMdgMN+G68d7befRjnF/rb+5Z7n+ITFqmiZZmiznxdmKMjehVEuSLPzgLqc/NghEKBJXKSqn5kpJgla3Wrv1fhwnSTJSX03YTrCx22EYkuPo+cncRFnj2GP3BwuD+YHAa/0pgIhuu91+8ODB7373u1//+tfb29uO4+yP/aqqquv6YDCQZXlubk6WZYIg4jhuNpv379+/evXqtWvX1tbW2u32yMVZlhUE4U/+5E+g/VUmk9kfaO10Oo8ePfq3f/u3X/7yl6B7959LEEQmkzl//vxbb72laZqu6zRNNxqNq1evfvnll/V63fd9hNB7770HOn98fBzEFYTct7a2bt68+atf/erjjz92HAcOHo5K1/Xz588jhMAlG+zEd3d3P/zww7t37+7t7cGOgGmamqZ98MEH//iP/wgtqWiaLpfL/+k//aeDthbdbveTTz756KOPtra2hunHCCHf933f73a7N2/ehJ9IkvTWW2/RNC0IAsSxNzY2rl+//sEHH3zxxReWZe131aJpmuO4xcXFTqcTBIEsy2CB/oSQaRRFX3/99b/8y79AxgFCaHx8HOYzm82CqHYcZ3t7+4svvvg//+f/PHz4sNfrwSbF/vk/c+YMzNLk5CQkTkdRZNv2+vr6tWvXPvjggxs3bpimuX9uaZrmef7s2bOQ6w4r5wnJ/yOAYr979+7Vq1dv3LixtrYGGxn7gY2eP/uzP4NHkMvlIOb/4MGDX/ziF1tbW3t7e8OD4zg2TdM0zfX1dfjJ5cuXT506BWqfJElYxnfv3v3www9/9atftVotx3H2V19rmlYqlWzbhnR6URRxsPoopClK0tSwg1rT7pt+fKDnHUUSmsSWc0I5J+R0jqKO3axSJDle1OIYbdW6a9ud5PGXrThObCfoDZxGyyxkpcIxy2DHYDDHjTRFUZyYdrhes+6v91e2B63eIQ6aBIE4hlJFplIUT81oU2NKKfstO+mY7wBBIFlkeZb2ZgokRTp+0Ddcxw2CYLTvo+uFrhekCNE0RZFERhUoiqQoXPCD+VGA32yeAoijXr9+/Ze//OXt27cbjcZ+SyfA87xer/fZZ58ZhvG3f/u3kD0bhuH6+vr169fv37+/tbVl2/bBi4NiXF1d/dWvfoUQAqMsEIRpmn799de/+MUvvv76a4gNjpybpqnnedvb2+vr64ZhxHH8hB7L+3/V7Xb39vY++uij3/zmN5ubm47jgKQfAkJrdXX1//7f/2ua5t/8zd8Mg5ZDoFr4xo0brus+ePAAWmETBKEoSqFQgOO/v7hK07Tdbu/s7Pz2t78FNW7b9shoIYF5e3v73/7t38Dt7MyZM5OTk08VAR5+3PC/6/X6b37zGwjtmqYZx6Ob5TC2Bw8eXLp0CerM4zhutVrr6+sffPDB1atXNzc3Lcs6OFrf9zc2Nn7zm9+Yppmm6eLi4tF7ibmuu7q6euPGjfv37+/s7Liue/CYKIpc111aWorjmCCIubk5SZK+W68y27abzeYXX3zxy1/+8v79+91uNwiC/bOUpinkhH/88cemaf785z8vFAoURR0Tq/njjB/Gjhe1et5uyxkcZlFGkkRGZcfyoixCMt2xez8hCEKR2GJOLOakXEY0Tc9x/+NG0jRNEmQ7wU6jn89KM9XMCxwqBoM5/gRh3Oi4G3vmg7Xexq5lu4d0zyJJgiKJgs7NjCsLk9pMRcXOZD8oJEnkdSFNs6blJUmyttVpts2RYwiEUkRYlr+22eZZWhK4qXG9lJdxdhLmxwAW1U+BZVmrq6vXrl376KOPNjc30R89rhiGYVk2DEPf90Fh3r17d2tr6+TJk5cvX5ZlGdzF9vb2arVau91mWVbTNBDMYLjlum4URb7v7+zstNvtSqXy7rvvQo0u1M0uLy9/8MEH4CIGPahVVRVFEfQbxD+hzPvJzahGxHa9Xr99+/ann3760UcfIYQIgoC0ZBhbEARwRzs7O51Oh6bpixcvlkolWZb3XySKoiiKHj58uLW1BSXWHMdxHFcqlarVqqqqB93RaJqG5GdZlqHOHIZNkuTwdNDhEG2GiarVajdv3vzkk0/+/d//Hf2xy/Tw4CAIYCTNZrPZbAZBoKoqWLVBJvl3eOIw+c1m87PPPvviiy8MwwjDkCAIMOUiSTJNUyh0RwjZtj2MYEdRtLOzA6P9+OOPDx0txPPr9Xq9Xo/jGNzOIO3/KKOFmgJYVL1ej2GYYaAbNhdc14W1sbGxUa/X5+bm3nvvPZIkGYbheV7X9V6vNxgMYN6GI4T4OeybQLY/BLcHg8HDhw+//PLLDz/8sNVqIYSgpJxlWZZlIcUgDMN+v3/r1q1Wq3Xq1Knz58/LsvytGfgYP4h7ht/quY2OY9nhyIYYSRIsTWY1biwvSAJ9PC2+CAKJPJPVhFJeKefVKEz2i2qEUJqmthvUGsZYQXW9ME5SEpfaMfEdXQAAIABJREFUYTCYA/whRu2EWw1reXPwaMestZ2DtlgIIYYmRY4ey4sL09psVRnPi8fz6/GVgSQJTeE5lna9IIoTy/ZNywvCOH786RAIOW6w40csQ4s8RxKEInGiwGBdjXnlwaL6qKRp2mq1rl69+vnnnw+9nfL5/Pz8/NTU1MTERK1WW1lZ2d7e3t7eDsPQsqz19fWbN28uLCyUy+XTp0/btr23t9ftdqempqrVarlc5nnecZytra3bt2+3Wq2hqVWz2VxeXuZ5XpIk3/ctywL3KTD0AkX9/vvvX7p0KQiCfr+/tLRkmuaJEyfeeOON6elpSZK+NTwI6vrBgwf/8i//srq6Cj9kWXZubm52dnZiYoLjOKgA397eNk0TRnXv3j1N0xYWFg5eMAzDNE1lWc7n85VKpVKpZLPZSqVSLpcPHqxp2oULF9I0zWazjx492trachwHIQRm2rOzs4uLixBQff3112FzAfK0//Vf/3WYnEySpKIo8Ah4nt/c3Nzb2+v3+yBxIftdluXTp0+DE9h30NVgqNbtdnu9HniMg+x87bXX3n//fVEUgyBYXV1dWlqampo6ceLE66+/rqoqTdOu696+ffvXv/41bL7AU5NluVgszs/P0zQNtdm9Xg+eaa1W+93vfqcoytmzZyHH/lvHJknSa6+9ZlnW1taW53mwqEqlEtjFra2t3bp1azAYDBdVo9F4+PBhmqaqqs7Nzf3n//yf79279+DBg62trUajsX+Ei4uLpVIJIQSV3mAUV6vVPvzww+vXr8OTQgiVy+W5ubnJycnJyclHjx6trq7u7OzUajUoQV9dXb179+7CwkKlUnnaaf+xYbvRXttp933LjUaarxIEElhKV7hSVizlRJE/1t/YNEVVSvrJGde0/XbfIR7fHgiCqN2zGh2zM3BVWZBEBqcEYjCYEaI4afXc7Ya9tDFY3TH7lh8dLIlBCCGky+z0uHJqRl+Y1rMqh6uNng8MTY4XVYIgPD8kCGKnPugbTpqm+600wCi30TFv3N+Ok1gQmPGCktPxrgfmFedYv6IdH8C4u91ug5sX5G8TBJHP58+fP3/27NmFhYWVlRWe5yGuC7ncu7u7S0tLxWJxYmJieno6juOlpaUkSRYXF+fm5sbHxzmO6/V6giCsr69DDyqg3+9vbW1VKpWZmZkoiqBl1zBpHOLJc3Nzb731FkEQpmkWCoXBYADDGBsbO0q2M/S+Wltb+/3vfw8V2uCXNjMzc/ny5YWFBUEQdF0nCGIYz+x2u6urq3Avh14zTdNyuXzixIkTJ05APXk2my0WiwePVBTlzJkzFEWRJBkEQbPZBKkGMfyFhYX33nsP+loNS3Mh2/nzzz8f1pOzLDs+Pr6wsHDx4kVFUe7duwcqEUR1r9e7e/fu3Nxcp9PJZrPfLQUd5ChMPlwWxlwqlS5evFgsFgmCmJiYyGQyJ06cOHPmzMzMjCAIYRg6jrO8vHzt2jXT/ENyFMdx1Wp1cXHx0qVLDMPk8/n79+9DIgBCqNPp2La9sLDQ6/UURTmKqBYEYW5uzvO8Bw8eKIpy6tSpmZkZqJbv9XoEQayurlqWFf+Rbre7vr5eKBQQQhMTE1euXGFZFtL7QVSTJCnLcqVSuXz58uLiIkIom82Wy2WGYeI4rtfrX3311cOHD2EewJj9woULp0+fXlhYuHXrFsdxnufVajUoJt/e3l5ZWSmXy1hUP4E0RWmaWk6413Y7hu/6o31KCIIQBSavc4WMkNd5nj3WufQURRZzynQ1WNvuMEw/ieL9ldVhFIdW3O7ataapyjzHKtTxvh0MBvM8gRi15YS7LefRjrG+a+42bS+ID7hhIZoiOZYs54WFaW1+Qq0WJYbGUdDnBEWROV3kWMp2gzhJgij2g8j3w/3xavjTNjDcvunyHJ3RRAIhUWA4lsbxaswrDBbVRwIaOHc6nW63C0KFJEmKovL5PDg5sSyby+UWFxe3t7fhlDRNTdMEuUiSpCiK09PT//2///d3330XVJPnef1+33Vd3/ehfHqYPOz7fr/fBxco8OjaX8Ycx7Ft23fu3JEkCcT5T37yE47jMpnMN/VYPojruoZhQAwWPpckSUEQJicnFxcXQU5PTEx0u92lpSU4JQgCsPUeqQ0GwCr8nXfe+eu//utMJqPrOsMwHMeB19cIPM9XKpUgCNrt9vb2NjiKoz+m01er1ddffx2M06GTs2VZMPlDJ3OEkCzLV65ceffdd0HAnz59Ggbc7XZhliBkWq/XC4WCIAhP6NH9TUDCNjSIgp9Auv7m5uaHH354+vTpubm5+fn5kydP6rqeyWRkWYZe1tCpG9LF4URd199555133nlnbm6O47izZ89+8sknrVYL3MVAvcNoj/gQQQOfOHHif/7P/2kYBqSjQ0k/tDoDE3LIIEAIDdcbQiiXy9E03e12G43G0P0bWlvn83nYp0B/3ONI03QwGMD827adJAks/mKxePbs2VKpBP+9uLi4sbExnCXDMKDp2tPO+Y+KJEnDKO6bQa3l9M1RgwaEEEmgjMKOF8SsxkrCcX8dIUlCkbhCVs5n5KwqGJbneaMl4n3Dubeyx9JEThM4LKoxGMwfSZK0Z/h7HXdly1jeNFp9zwviOD4kRq3JTLUknZnNnJnNFLM8znl5/nAsPVPNUhQZhHGapHuNgWn7Bw8jUtRsW1/d3Q6jmOfofEbKargiDPPKgkX1kQjDcDAYdLtdSIRGCA3dsz3Pa7fbILY9z9vvYmXbdqfTcV0XgsCqqs7OzmazWdM0oRq28Uf2m0LDx8EHDUU1NMG2bRuqpl3XXV5eTpLENM0gCKampnRdV1X1KInfgOM40Gx5mM0LgVz4aEh1BgU7fNGHVHMwQjt4wUwmMz09feHChT//8z+Hds1P+HQoAC4UCvl8XlGU4ZgpioLdgcnJSVVVh8dDS3DDMIYijSAIURRPnz791ltvQW/kmZmZIAh+85vfgPM2xPwhDAstyo4yLSOAqIaQu6qqYOQWx/He3t6XX35p27bv+zMzM1NTU2CrzrJsmqaWZR0crSzL586du3LlSqFQ4DhudnbWNM1f//rXW1tbw9FCj+hyuXxozvzBsbEsC0HywWBgmman04FFBU3XRizKoQc1rDRJkjiOy+fzuVyO5/nhIBmGgd7UU1NTwxNN0xwMBr1eDxqJwUeDpz04k8HNhmE4XBgw891ud2RhY0YIotiyw87Aq3cc40A1NUEQNEXqCjueF3WZ5ZjjLkFJghAFJqMKhayUy0hBGPv+6E2Ztr+21c7r4qm5osAzNO5iisFgEAqjxPXjRtdd3zXXd82dpm05YRiNllJDjLqUExcm9RMT2kRZEjgaf4U8fxiaymckiiJNyw+COAiiMEqCMDqYVjAwXdcLRZ7RFD5JU0lkaYrE+yCYVxIsqo8EmEL1+/2hUAELq4cPH/793/89WE+FYQhdjkCFgiO3YRhgkhzHsWEYN27cuHXr1urq6u7urm3b0FUYelBDa+jhx7muC4nBNE1DlPvNN99cXl5+9OgRQgh0nWVZm5ubv//978fHx0+ePHnlypVTp07l8/mjxDkdxxkmXQMwwt/97nf37t2D8YPArtfrw1FZluU4zqGiulqtvv3227OzswzDfIeY8JOxbRsaOA1/MpSUuVyO4ziKonieVxRF13Xo7wVPCrY8BoPBk/3bvgkQ+RCS7fV6y8vLkM49GAxWVlZardatW7fGxsbGx8cvX7585cqVfD6fyWRAVO/frRiONpvNQtx7OFpZlk3ThOA/2MuDE/i3jg0WVbfb/fLLL+/evbu6ulqv1y3Lgq7phmH0+/397vSwqEaU9lGAJmdQng0/SZIkTdM7d+50u13Iqw/DMAiCYYMuWPzDp4D5Jhwv3uu49Y7b7nu2G448d5okOJbKadxYQZSEl+C7miAQTREiTxeyUrmgGqZnmB5C6f778v2o3bUabbPVcwWeU2WWwqoag/lxk6Sp5YStvvdox3i4MdhrO6YTRgcUNUJIEujxgrgwpZ2e1ccLIktjcfbCIAgkCeyJ6QJJkkEQJUna6lrugeykKE6SNNze6ydpmiRpRpNkkZF43Esc8wryEryoHQcgKLffWztN0zRNO53OYDAYKdYFu2NwtIJfgcnzo0ePPvvss88+++zhw4eNRiNJEohCQ9yPoqhhKyzIMYb/haZEs7Oz7777Lth9G4ZhWZZt24Zh7O7uMgyTzWY3NzdB1Z8/f55l2W+NVwdBYFkW6PbhHfm+v7m5ubOzM3Iw3BFo12+qTM5kMnNzc7lc7odooQSjHSo08KMGIS2KIowK2mjzPL8/Wxsq0g92PjsicNlCoXDp0iUwNt/Z2YGNkna7DSbYmqZls1nI1T9//rwkSa7rjrT7gusIgjAcLcTAYbTDKYVq5COOFkTs0tLS73//+y+//HJ5ebndbqdpyjAM7KrAChwqYehJfuiGyJOJ4xh2f0YWf7PZ7PV6IwfDUmFZ9pn0UXvlcbyo3nZaPc90wiAcNVDlOUqX2UJGKGYF4SVp7EyRBM9S+YxULqi1xoCmySiK9y/nMIojO2n3nN36QBZZSaAp8rhH4DEYzA9HFKdBGLd63mbd2qhZW3WrbwV+MPqniqYIjqWKWWGuqs5V1WpJ1mSGetY7+JijQyDEMVQpJyOELNtL0zSKkzixo8fdNNIkjZK007ddP1QkXpGFqXF9oqySBIF9yzCvGC/Hi9px4KBCIAiC53lwez54PEmShUIhl8sJguB53rVr1/793//9xo0bKysrpmmCooaUaZZle71ep9Npt9vflC578uTJTCZTLpfHxsbu3LkDdt+QXQx52isrK5DRrSjKUYpyQZeO/ISiKEVRJEk69JRcLpfP57/pfuH0H0hHgQrdf/FhOgAAv4Lg7f423XDi94yc5/P5t99+O5/Pl8vlmzdv3r17t9lsQmk9Qghywn//+99vbGz8j//xPyYmJqIoGvnE9HFGRrv/NsEI7SijMgzjs88+u3r16o0bN9bX1yG+TdN0LpebmZlBCMGiarVah9bAPxUjqwX+WxTFQ/ulIYRYls3n89lsdphbjjkUx4vqHbdn+gdT5hCBNJmtFMVSTshp3Etkw0PRZFYXxwqqKvMcSydJmiSPvR+naTow3ZWtliKzYwWZPfZp7RgM5ofD9aO+GWzsWUsbva261TX8g1nfCCGeo8s5Ya6qnJ7RJ8qyJNA/2BsH5qgQBKIoIqcJF85UZJFx/TCM4oHhBo9/5xMIhWFsJf6DR/VO337nzdmcLnEsxWFRjXm1wKL6SFAUJQiCIAjDMCzIjHw+f/bsWegwPPLtTpJkpVKZmJjIZrO2bS8tLX366adbW1sQ3oTGUXNzcwsLC9CGOk1TyNc9dABgghXHMfSsKpfL0I0ZnK7AQsw0TUEQNjc3p6env9U+Gqpnhw5hALTUmpmZYRjmoHLWNG18fHxmZuZQpUT8kSd/7qGnDM8Cd7EoikZCtSzL7vfEBmkKUWjLsqBjFjiug/HbMKYKtwlNoY4+sBEEQRgfH6coClK48/n8zs5Oo9FotVqtVisIAtu2oaHXm2++2Ww2oygamVsYrW3blmUNO4RB7sOwR/e3jnY4J9CswjCMr7/++vPPP9/c3ISIcSaTmZmZOXHixMmTJ23bXl5ehvzwbxLVI5MPCf/7xwNQFCWKIs/z+/UzQRDQKE4UxZFVBDcC3vWZTOappvrHQxQnfpB0B36t5fSN4KCoJhChyUylKGVV7mUJUwMUSWgyX8hKGU2URDaK4igaDTrZTlCrD6olzXFDhqZonMOJwfz4iJM0ipJO399t2es1c6NmdQa+649+XVAkwbFUXuemx5TZijpRknMaxzG41f2xgCSQwNMCr0RRvNeygjDeSFLD9EYaocVxEsdJs211B245r44X9XJeLmQlkkD4MWJeGV6md7UXCMMwmUwmk8mMiNWZmZm//uu/npmZyWazIyE7iGNDfrJhGDs7O2tra8M620Kh8LOf/eytt96qVqtQhg3lqd80gCAIXNctFouZTOb06dO9Xm9tbW1paenq1at37txxHAfkkGEYtVoNHKSffEeSJBWLxZGgtCAIly9f/q//9b9ms9mD8WqGYaASGHy5nwkjYWTo7+267oioVhSlVCrtH1KaprCV0Gg0xsbGSJK0LAt81PZXffM8XygUNE37PknpkPzP8/ypU6empqbefvvt7e3ttbW1zz///OrVq+C2PbSs29zczGQyxWJxmCwA9+L7frvdbjab4+PjCCEopIfRDkWsKIrFYlFRlG/aAhjG56HcfWtra2NjY7ioKpXKz3/+80uXLlWr1bW1tTAM2+32E258JP4PBniWZY2IcI7jstmspmkj6/PkyZP/7b/9t7GxsWw2O3JlKBoXRRGL6m/C9eN239tt2btNu2f68WGRalVmX5Zq6v1QJCmLbE4X8lkxo4mOGxyssguCqDdwOz2nZ3osS8vUd2kjj8FgXmqCMHbcaKdhLW0M1nfNescNDotRswyZ07jJsrwwpU1XFF1lOQYHqY8dmiKcPTlGEITjhmEU205w0Lk9jpMkTdd3OgxDXTw7oasiTSGawo8S84rwkr2uvShomobSWVVVBUEYxkLDMIQmQ2CnTFGUbdu2bUN9L2TAgiUYGEcNtSIUQmcyGZqmQQyD4fbBj4aw5MbGxvLy8lCojI2NIYSSJFlaWhIEIQgCsEODDsmg8Z58R6Io5vN5MMqC09Ef/aU8z4P7JUkS2ndFUUTTtCzLxWIRimafCVAUzXHcUFcHQTAYDPb29lZWVmALQ1VVXdcFQQAPMEVRoLczdK66d+9eNpudn5+XJKler1+/fr3dboMmpCgK7qJUKsG9fIcRQkgc+opZlgU27NlsdnJykmXZWq2mqurQkh0h5HmebdsgjMEyzfM8cFC3LOvOnTuyLM/Pz/M8v7e3d+vWrWEYmaIohmE0TSuXyzBaqIgeSmLoa1Wv18HTLgxD13UHg8FgMBguKujrpus6RVHQTmy/Yh8BHOn3d2uDZ91qtdbX16vVKjjkaZoGVu3ZbFZRFOjEDtMCIXqEEPi3UxQFPnbQSq1QKGBF/QTA57bVcwd24Aejrx4cS4k8XcwI4wVRFl8yQxeCQAxNyiI7VlAbRavbdwzTGzkmiGLT8nsDp9mxRJ6ReBrh9yoM5kdDkqZpgnpG0Og66zXz0Y7R6rmON5pURZIEx5AZlZssy7MVZaIsF3Re4CgKf10cP3iOHi+pfhD1DZekiN29vu0EI/HqNE2TOG12zCRNVZlXJK6YlbKaQOB4NeaVAIvqI0HTtKIouVyuWCzqug69gtI0XV9f/+d//udHjx6dO3cOJMf29vbGxoYsy5qm/dmf/Vkmk4miCAyiHrfA9Wu12v379xFCq6urX3311draGjTrGiGO4yiKvvrqq3/8x38EFfT666+fPn0aspEhGG5ZFhw8LCH+1k1cnufBgqtUKnU6naFM/fzzz3u93muvvTY9Pc3zvOd5Gxsbtm3LsrywsPDOO+88Q1FN07QkSeDdNZwW8FT/8MMPC4WCLMuLi4uqqvI8Dy24SqVSu92GLQDTNK9evbq1tXX27FlJklZWVh49ejQ0oIZU6lwuVy6XdV3/zqI6iqLNzU14yizLzszMvPHGGxD65jhOEARwGoOHC0pYkiQIVheLRfAeRwj1er2PPvpoY2Pj3LlzDMOAkTvUAiCEWJaF0Y6NjamqCsZm8IzggCRJdnd3v/7664mJiYmJCbCyAxfu4Whd193Z2YFkiq+//vr69es7OztP8N8GR4Bh8naSJJZl7e7uXrt2LQgCWZYrlcqZM2dgLymXy5VKpXq93ul0YCPg4cOH//RP/3Tu3LnXXnsNHNfA5U5RlEKh8M4772BR/QQ8P2p03O7Af9zG6w8oIlPMCuNFqVKQjn8nrUPhWXpiTDcsf2O3U28j4vGbjOLE9oLuwNlrGrrCF7PiS3mTGAzmOxHHaRglex3nwVp/ZdvYalgjTo0AQ5GazFYK0slJbW5CKWV5WaSxOdnxhKZJRWQnxjREEILABkEUJ6bj+CN7xgRCpu37Yfz18l4YRhfOVFWFp0gC75NgXgGwqD4SJEkyDFMoFF5//fXBYHDz5k1Qob1eb2lpybbtdrsNxaWNRmN3dxdCfOPj4wsLC0mSgPoSBCEMQxAklmXdu3cP9Mne3t7W1pZhGIcGFdvtdq1Wu3Pnzs2bN0FIDwaDnZ0duNrOzo7ruuCMBXIul8sdTFM/yNBU/Cc/+cnNmzcNw4BAN1zQMAzQkCD+XdeVJMlxnBMnToDEfSazCsF86AEGPtVJkgRBsLW19fHHH4PdWhzH1WpVEASe5+fn569cuXL9+nUYLdhfQ2oAx3E7OzudTgc2JgiCyGazp0+fPnXqVDabhZ5n32GEjuPs7u7evXv31q1bKysrFEXVarVOp5PJZARBWF5eHrZMg9bcmqZBmFoUxZMnT16+fPn69etgIRYEQa1Wg9GCi3i32x2ONp/Pnz59enFxUdd1qKnmeT6bzQ77mcVx3Gg0lpaWFhcXp6endV2HRcXz/LBBdL/fv3Pnzs7OThRFW1tbtVrNsqwnRKpVVS2Xy6CroVd2FEX9fv/27dudTkeSpIWFBUgT4Hm+XC5fvHjR87ybN29algXW9xCsbrVaHMcxDLO7u9tsNiVJKpfL1Wp1dnYWfv4d18crShynQZT0zGCv7XSNQxLkEEKKxIwXxJzKiTz9ktYN0jSV16XxgpLVRElgfT+K4/9YikmSpmlsWF6tMSjl5DDKUhTCddUYzI+BNEWGHbT73tqOubw1qLcd2z0QoyYIhiZVmakUpZmKMjUml7KiJDA0hRX1MYUkCJImVJmbHNOCIDQtl6bIrVrf9cKR95AoiqM4qTUHSZLIIsfzTDErZVRsa4p56cGi+qgQBFEsFt99990wDDc3N/v9fpIkkIhrWdb6+jqEiCGVGozN3njjjcuXL+u6rqqqpmmqqhqGAaK63+9fu3aNYZhhzvZI1HEIKMx79+61222wlep2uzdu3CBJMk1TyD0GUc3zfCaTqVarY2NjHMd96+0ghE6fPg0B6qWlpSRJkiQBU/F2uw3tpkHlQmcvhmGuXLlSKpWeUPv9VAiCMDY2BrFZjuM8zwNxWKvV+v0+yH5VVc+fP18sFrPZ7GuvvUYQBFidgwIPw3AwGCwtLREEEQQBKEO4+Pj4+Hvvvffmm29CVPm7iWrYQPny/7N3p71xZee96Ne057FmFudBUktyd6uP252+cZBzXyTtGEE+VD5UgIsEBox7rmEgzj1O7LR7ktSSOJM1V+15r+G82DJNUeoWKVHi9PwasAGxWHvV5iar/nut9Ty/+93m5ubh4WF18h89esQYI4RkWVbtQFZK6bpu23an01ldXa1ezieffFItHX/8+HF1bouiGI1GSZJgjKs5eSllNbClpaVf/OIXP/3pTx3HqSbVq2i6ublZjaRqYSWEePjw4cbGhuu6xy+q6rwNBoPf/va3jLGqO1q1KP2HunNV9x2Wl5erAvVHfdGjKPryyy+/++47Sune3l7VJq3dbi8sLHzxxRdlWT569KjatV59SxRFjx49qpZIFEVRliWltNls/vSnP33w4EGj0YBQfULJ5TQqeqNst5cMpzl/1V0P39EX247vXuFSPIziwDVadafdcHcPreEoSY+FaqSUQiqKs52D8fJ8mBdSY4qwK/paAQCnpRRSSvVG2Vffj799On60NX1lrW9CsG2yVmhtLPp3VvyFthN6Omy+vfx0jYaeubpQwxhTQoaTtOSiLNULH0UUQkiNJkmWl6ahcSk//clC6Jvw0wVXHYTqM3AcZ319/bPPPuv3+3/84x83NzfH43FV0vloATYhpAoVnU7Hdd2qSZLjOPfv3//rv/7rr7766tmzZ5zzakqwWkIcBMHKykq1drfqkoWONTFyXbfb7a6srNy+fXswGAyHw+l0erw/MKW0Kgl+69atzz//fHl52XXd41XKf+QVdTqdatY0y7Ktra2dnZ04jqtt1UffTgip1jPX6/XjTZXfXnVmlpaWPv/8c4RQNfHLOc/z/KgK+sHBwcHBQTUD3+l0lFL/83/+z2pJdjVNnef50Qrnaq6+2WxWz/nZZ5+trq6+dpr6R75a7VJeWlra2NgQQgwGg6qU1/G3B8uywjBcX1+/devW/fv3wzCs+oR3u10hRNX0+9mzZ/v7+3EcV7Gz+saqoXSr1VpaWvqbv/mbn/3sZysrK0dnuNPpfPLJJwcHB9999121jbxKsFW1s1u3bvm+/9FHH+3v7//pT3/a29urKtUVRcEYMwwjDMONjY2qQl6Vll9ui2UYRq1W+/DDD3/+858/fPhwa2uLcy6EmM1mx0/+eDyu1Wq+79++ffvzzz8fDAZff/311tZWtWe7KIqjx1f72KuLv7o7ALVkXpaXoj/JeqN0NC3S7GSm1hipqvLMt2zP1q7u6cMYM4Y911hZqI+maZbxLOcnPlelWdkfRv1hPJ6mjCJGdbheALjeZkk5muaPtqZfPxnt9JKX91FjjBgljsnmW/bavLc27843Hc/WrlBbwZusmq8OPHO5G86ibDBOnu0MD/qzshQn/v7zUsRS7eyPCUY133RsPfRMxzrZTASAKwRC9RlU5Zd++tOf1mq19fX1X/3qV99++21VtOzoj0W1Qnttbe2TTz5ZW1tzXbea8v3Zz36GMS7LstfrHZWnrmpTbWxsfPrpp71ebzKZVKH6qCwzxnhpaemoG3BV6PtoRreiaVq1ePiLL774/PPPq2B29O0//oqqumtffPHFysrKr371q9/85jc7OzvVdvHqARjj6vkfPHjw8ccft9vto9ZKb5+XqkEuLy//wz/8Q7WsvVpOfPwvbxzHBwcHnU6nWq5sGMYvf/nL1dXVf/3Xf/3tb3+7u7t7vAmZruv1ev3DDz+sZn3v3LkThuFp5tV/6LVU8+RH67FzryesAAAgAElEQVT/8Ic/9Hq9LMuOj9DzvPn5+b/927/94osvbt26ZVlWdearEl//+I//uLKy8m//9m///u//vru7W+Xb46P95JNPvvjii08++eTOnTtVxa/qq4uLi5qmPXv27Pe///1oNKq+sVpKUK1yX1xc/PnPf04ISZKkWjdRzdLruh6G4QcffPDpp58+fvz4aAv60Qk/GkB1O+azzz4jhPzLv/zLYDA4XjgdIVQUxXA4HI1GS0tLQRDouv7555+3Wq3f/OY3v/rVr6oqAMeXdVULNG7duvXpp58uLy9Xq/pfe/JvmiwXh8PscJhFSZmXJ5cSmDr1Ha1dM69iibKX2aa+sdxM0nJ3fzqaJAih4y83yzkXSW8Y90aRaVDX1hCCVA3AdTaYZN8+nXz1/firJ+NX7qPGGOuMBq6+Ou9+sBqszHut0ITKZFeLqbNW3V5fauSFwAiNp6kQ8uW9TlLK/jDKCx4Glmlod1abEKrBlQYfec+gWuBdr9cZYxhjx3G2t7d7vV61Brva1Wzbtm3ba2trq6urd+7cqVooY4wXFhaqld5ra2tVAeeq/W+j0VhYWNjY2JhMJgsLC1VRq263W/WLrqb+NE378MMPdV2/ffv2Z599FkVRVZG72q1dNcdaWVl58ODB2tqa53lV6J2bm/u7v/u76pmrNed37969f/9+s9k8qtpFKa36aVfbpOfn53u93mg0qpa1VzXMHcdptVqrq6u3bt06CtVLS0u//OUvP/zww8lkUsWqjz766M6dOy93V/pxVfisdp47jrO3t1ftUhZCVEf/2c9+trq6WnUIq6qFLywsGIahlFpaWqradFfRzjAM13Xr9fr6+vqDBw+Wl5erWPvjA9A07cGDB1Ul7aqWdRAEn3766dLSUvWzC8NwbW2Nc97pdD766KPBYBBFUTXhXJW5bjabrVbro48+un///vEy49XPd2lpqZp8Xltbq0Zb3bM4Gu2tW7c++eSThYUF3/ePR1DXdQkhP//5zxFC/X5/NBoJIXRd//DDD+/evVur1QzDWFpaqtaQ379/v3rmo331S0tL6+vr9+7dW1paqu7UVPPty8vLx08+Y2xxcbFK2tWlUlWhr2qDLy0tffTRR51Op7pNQwg5Ws7ted7e3l6/36+qzVcXquM4nuetrq6ur6+vr68fLy0OEEJSKSHULCn3B8lgkhUn7twjhBByLa3bsJuhGbj6FS1Rdpyu0Xbdme/4zboznMRJWnB+fGe1LEvVH0WPnvZMnXYaDsxUA3BdRWk5mRWPt2dfPxnv9OIk4y9vTqpaUneb1nLX3Vj0l+ecwNFhjvrKIQQTQhuhfXu1mRc8yYq9w2l/GHMhpHxhvrrkIorzJ1vD5zNBCAe+aZtX/oYyuJnwD225BD+k6nFVlmWe59VK6dFoNBgMqjbOQRBUC4A1TdN1vYrf6PnHxzJJkul02uv10jStYlWtVqvCm5TyqFNXFaR1XT+qzFwt7q0KiY/H42rlLec8DMMgCFzXtSyrKgp1NBtZHa7K3tVPWdf16jEnoo6UUghRFEW1ETeO42rS0rIsx3HCMHQcp3pFmqZV+5OLoqhmNY8mNg3DMAyjqtd1pvN5dPQsy8bj8XA4TJIkz3Pbtqui041G4/jTVo+vTn6aprPZrGpMVbWwqk5FNdrTLD9WSqVpWsXy6uRXu9Orn121cV0IUf5Z1Q07juMkSWzbrg5aNVSreoOdKDNeVf+qFrQnSVKNVkp59I3VD6461vHRVpdZmqZJkvT7/cFgUK0bn5ubq+qim6ZZ7dOO47i6qIqiMAyj6uZ1VPvtaN1BdVEZhnF0UVWq8vLVz73aXJBlWbUuIAzDozJy1euqRlVdKtXFPxwOB4OBaZqWZQVBUN3IqM5/dfHDCvAjXMi8EN8+m/yv3+8/3Jr2RunLEzV3loN7q+H/uNv4+E4dn8NykAumFJJKfb85+H/+v2++/Ha/N4iy/GQ5+k7LX+qGf/vZ2t/9X7coVCEC4Jra6cWPt6Z/fDT6w3fDKCnT/BULvzVGA1f/cCO8vx5+sBLON21C8ZWtLHHTSam4kI+3Bl9+t/+nhwd/erifZSXnJxu+Yow1ndUD+7OPlz+5O39rpdFuOBcyYADeEswjnVm1TbqaNfU8j3MeBEGz2axKeVmWZds2emlFcfUt1cSv53lFUVTZ4/gSWc/zfuigRwFbKVX1Xqqmc23btizr5ZyM/ry2/DSvqIqC1RZf3/eLogjDsCiK6qBHk+2vHM/bO370KqZW88CGYVQNw04UXTt6vOM4ZVnWarVGo1GdiioxHq2gPg2McXXcH3lAFemrXmKe5wVBUG1yNgyjKvn+I6eCEFI9oCotVo1WKVWV1K5mgH/ouNWec9u2q0rgVVc213WrJmTVA6rzU11U1fR+NarTn4SjK9N1Xdd1q0UK1XL3yvGnqg5albKvbqkEQdBqtaobLtW/X/kg+M4UpRxM8oNB2htls7gQ8oUbmpRgSkno6Qtt23e16/E5EmNEMXZsfb4d9IfJLMrygp+4kxsn+d7h9LAfDaeZa+uWAe9KAFwrScYnUfFke/b1k/H2fjxLyrI8mawQQpSQdt1c6rh3VoL1Bb/mGQzmqK8yQrBOaLvu3tvocCGTrDjozQ77sxMPU0qVBZ/M0u83+4zgqrCs7xqGDu8F4IqBS/atVOmuKsJc/cuPT81VEeuof+8bzONhjKs8dvxfzivGVAGyimpH/4LOY/v0KY9+FKGVUq99XdV+7yqNv7ehVj9updRpRnjc0Wgdx0GnHm11iGoGHiGklDqxob16wPFbD292PVQ/et/3j9/Z+fGnqvYOVBf/0cuBRP0j8kIcjtK9ftIbZ7OkPLFIiFJi6rTmG/Nt23Ou1b4y09DmO0FvmGztjaZRduKFJ2mRF/xwGPdHKcEYQjUA10yUlFsH0cOt6VffT4bTvCiEfNUaSUbxfNO+vxbeXQmX51wCTfauhVpgBZ7JhYjTUnDZG0Qvr5BVSuV5+XR7kOWlbeuWqWuMQqgGVw5csm/rrFnu7YPHu44uFxiNznro95n53/6Ib/y9p7m/cC4n4Q3OP6To08tL2Rtl/XFeFOLlj5SOxVqh2albrZrlmNfqL7Ou0UbotBtu6NuzKM/y8njPaqUU5/KwP/vTwz11q90IrQscKgDgHGWFmMXls/3om6eTZ3vRJMpfXquCECIENwOj27TvrYV3VoKab0BlsmuDYEwo7ja9/3FvXnAxnmXTWRrF+YmHKaXygg/HyVeP9pVSmkYMg+kapXBvBVwd1+qjGwAAXFp5IXrDrD/Oilf1ZXUtbb5lzzWsVmhesykaTaP10G41nNC3BqO4KPmxGvNVPXB1OJj98ZvdemDd32hf2EABAOcqzfj+IH2yE339/eRgmM6S8sS2lwohuNu0f7JR+8l67fZyALdqr592w2s13Dgt9vqzrV30qlCNOJeTWfrVw/00LefbYavuUYIpufIFO8HNQf/5n//5oscAAADXWVHK0Sx/ujv778ejncM4zV+sgIoQQmihbd9ZDpc6TrtuXbPPlBhhQlCa8cEoiZI8SooXUjVCCCFCiJSy0/S67QBhBPV+AbjSilJO42L7MP722eT77elOP4lTXr7UVAkj1AiN5Y57f712by3sNCzHgsmeawhj9Ly4t2chhZKsVAgVLxUtQwgphZRCBRdZznWNGTojBF+zG83guoI/XgAA8G4VpeiPs+1esnMYDyYn79AjhDBGrq3Nt2zfvVa7qSuUYptqoWe2Gs5e39rvzxTCGL3w8TrNCiFlfxQPxkkjtGFnNQBXl0IoL8Rgkm/tx98+nWwfRqNpXr60QgcjhAluhdad5eDuanBnJYC7adfb8ny4PB8ipYaTZGt/HKUFQs9XK1WUQkLI0Tj5jz9sjiaJbemOrTNmMZiuBlcBfHABAIB3q+CyP877o+zlHloIIcdinqMttOyFlh24+jWbpj5iW9riXNgfJs+2h3GCpVTHY7WUqizFYJQ82R4yiuuB+cPPBAC4vEou01zs95NHW5OHW9P9QTKLX73qO/T0mm/cWfbvroXtmqUxAhOSN8HCXPhXD1ZMQysKESd5lp3ssyiVKrkYjOL//nZXKfWT251mzdEYNFcDlx2EagAAeIeUQnkh+uOsP86KVzWSsU0217C6TbvbtA392k7UWKa20An2+5Ft6pQkSuLjqVpKqZQaTuKnO8NWzVaqBh+fALhylEIll+NZsdNLvtucPtmdHY6yvHi5NTEiGNd8Y33Bu7UU3FkObJMxaFN/M8y3A9e1krToD6NDpfKcoxeL1ymlSi6Gk+TrR/sY42bdtU3dtTXC4F0BXGoQqgEA4F3hQiVp2Rtlu72kP8ryH5ipnm/aNU8n5DpXU9cYDT2zGVqNmj2J0tks4+KFVX8IqSjO9w+nw0kSZ6WuUR3WggJwdQip8kL0xtn329OHm9Otg3g0zbl4xR89z9YCV99Y9O+thfMt2zYZLPy+OXSN+LZ2/1bH0LXf/2krf7if57wo+fHHYIQ4F+NZ9mR7oGt0Ok0++qAb+hYh6Nq+R4KrD0I1AAC8K0LIaVz2RtnBIB1O8/LFuizVdI1na/NNO/SM6907RGNEY3ojtNtNdzhJkqQ48WlbKRTF+X5/Opyks6TwbR1CNQBXRTVHPUvKg0HyaGv6eHu6P0jilJ94GMEYExS4+mLHWV/w7qwEgaNbBmyZvUE0RjRGNpabzZo3i7PDQTQcxyUX6MUJa87lLM7F/jhJcillp+UbumaZ0GQLXF5Q/RsAAN6VNBc7vfjpTvR0bzac5lK+UJ7L0Klna7cW/Q83au265VjsGs9UV7iQGJOiEL1B9PKiUIUwl7JZdxuBY+jUsa5h2TYArp+q2/wkKp7szL59Nnm0OT0YpK9sc2CbLPSM9UXv/lpttevNNSxDp9CV+maiBCuFTEPLCzGNM6XQiQsGY4QU4lwihLhQmKDAMzVGr/v7JLiqYKYaAADeCalUUYj+KKsatL5c/NbUad03WzWr07A8R7v2iRoh5LvmrZXWaJJ+/eggTgvO5fG5iaLkXMj+MN47nPqOjurOBQ4VAHBKQqg4473R84Xf24fJNC5OPAZjTAjyHG2uYa92vdvLfiu0XFu7kAGDC8coJgZdW6wZhjaL8+E4mcySWLzwjoAU4lxwLnYPJgghSnGr7lJCTINC0TJwCcFMNQAAnD+pVFHIwST/9unkye5sNM1f3lBd943lOXdj0V+d90zjRqxqwxhrGplG2X5vmua8KPiLFWowQso0NUJws+50W96FDRQAcDpKoSTnz/Zm3z6dfP10snMYJxl/udy3oVPX1la67r3VcH3RX+w4UJzsxsOEYI1RKRUhOMt5nORInShbhhB63rkaY4wxYowErglXDriEYKYaAADOn5Qqyfh4lh+O0sH4ZKLGGGGMHYt1m3YjMEyd3pCPCNXO6mbNnu8EsziPokwcXwOulFJoNE6e7gzvrrc4l4Rg6LIDwKUlpUoLMRhn3+/Mvn022TyIhpP8xGOq4hGupbVq5vKce3vFn286oWvAXOMNhzEyDUYpWV+qY4zSrEjSYjJL06w8kauLUhSTZHN3pDNKCQ49M/RNU7/+G6bA1QKhGgAAzh8XajjNd/tpf5xNk4K/uPabEqIzErj6XMMKPeOm5cbANW+vtmZxvns4ebnN2CzOUW86GCeTOHdMzTTgfQqAS6oo5fZB/Ghr+tX346e7s5crkyGEGCWGRuca1p3l4PaSv9h2XZtBGgIVSnCzZmOM8oJrGvv2+8Pdg4kQUsq/vGkqpTDC01n2eHNAKaYEb6w01xfrDJpsgcsEPqwAAMD5k1JFSTmeFUnGy1LKF++7a4x4jtYIzE7D8l3tpoVq1zaWu7Xdw4lt6UUh+ItF0fO8lEoNRnFvGJOGA6EagEtIKpUXYjjJn+zMvnk6frYX9cbZicdgjDDClsEavrHUce+s+Esdtx5c804H4EwIwa6ta4xwIQnGcZLHaTGLsjw/0R5CJWmR5ZxRjBAmlNQD23N0Q4c3CHBZwLUIAADnr8rQhCBDp5ZBs0Icb8tsGbTTsOYaVjM0HUu7aTVXDJ3VQrtZc5s1t+RiNsteKPqKEVJqOEk3d0aWwWq+dXEjBQC8Gudqr58+3Z19/WT0eHs6S8qXH4Mx1ihphOadJf/Oir867wWuftP+3IHT0Bht112k0GSWlaV4vNnP85evKCWlHE4S/qRnmVrgWUvdoNvy4IoClwSEagAAOH8UY9tijcBYajsYoWlcxhkvSlFyybmyDNqpW+266dnM0G7EburjNI16jlEP7XbTjZM8jnOllFKIMaoxYhiaY+mmwbhQL1esAQBcLKVQVohpXGzuRd8+mzzbi3qjTIhX/KqaOg1cfant3FkJVue9Zmhq0HwevAoh2HN0jN3bq42y5HGS5wXPcs6FOGpEqRRCSMVJkWb8yfbQcUyMkO+YpkE1Bq3OwcWDUA0AAOdP08h803ZM1gjM1WF6MEx7w6w3zkbTfBIVpsE6DavmG/Rm1Cc7gRJsaCT0zPmWP56kh4OZlApj5Fi675lzLb/b9tcWGysLdd+FaWoALhcp1WCcbR5E3zybPNycjmaFeNX9L4yR72gbS/7dteDOSlD3DehHDX6cabDVhRolOM4KodTu4XQW5QpJdLzNllJSyv3etCgFJagW2q2aDQuawGUAoRoAAM4fJdi1NVOntsXqgdGqmb16djhMe6NsMMkX2vZix6n7N3RvIcaIUew7xnwnGI6Tw/6MEGybWqPmturOXMufa3mtutusObYJbWwBuCwUQnkhoqTcOogebU4396KDYVqUQr6UqHWNuJa22HHurgTrC36rZpo6zCWC12CUhL6lELoXZfL5jZppkhQn6m4opWZRlmZlPbACz9xYamiM6hrVNbjGwEWCUA0AAO8Kpdi1NEOjNc9YbDtRUk6iYjjNfVdfX/B8R2c3eOrGsfWFuXA6yyaztFFzlrq1uaY313QNgxk6MzSqaewmnx8ALhul1HhW7PaSh5vTb56OD0dZVgj5UktqhJBra+vz3v212k/Wa626qcOqb3Bqrq3fv9WxLb26V7OzP46FPLEUQkpVlvzp9nAW50lSWKbWCO16APPV4CJBqAYAgHcFY6wxrDFim8hztJpnNEPeSUtdozXPuOHbCy1Ta9bstaU6wahRc+Y7QT2w4FMRAJcTFzIv5P4gebw1ebYf7Q/SOONCyBMP0xhxLLbYdu6sBBuLfrtuORZ81ARnoDFaD2wp1TTKOZdpVgoh84Ifv31TleEYT9M4yUPfch3jg7WW5xiU4JvWTQNcHvCXDgAA3geCsa4RSjXLZARjmIM1NEpdw1ppLHZ8TaOGzqDYDACXVpaLSVxu7kVfPx3v9dNZUopXzVHbJlvqOHdWgvvrtfmWY+g3+tYheGOuY95d7yiEB+Mky0s+SaQUJx5TtbPe3BkWBTd1baETGjrVIVSDCwKhGgAA3geMEcaYEAy7hCuEYJ1QXaOeY1z0WAAAr6eUkkpxoUouS35yjpox4pis27RvLQUbi36nYXm2djPLRoC3p2u0HlirC7XJvXmNke++PxxP06IUx9eBK6WkQuNZWnLxdGfYbfvdlteq2xc4bHCTQagGAAAAAAA/hlFsGtRz9LpvDCf5yw8wNDLXsG4tevfWguU517M1RjG0EAZvhmCkMdJtuu5PV33HiKK8KAWXGRcKH1shgREqCiFE/nRn4Ng6pfMQqsFFgVANAABnphRSSGWFSAsR5yLOeMkll0qjxNCIYzLHoKZOdUYwfKo8I6WQVM/PbZTxpBAll1IqjRFTo45JbYNZOmUUEzi1ALyKQoqLkoui4HnBs4JnBc9LUQhZCikQUgRTSjSNahozdGYazNSZwajBqEbwqxdsU0osnXabdl5UZaPUZFZEKUcIUYptg7Xr5vqCt77odxt24Ooag19Q8FYwRobBdJ2tzNd+8kGXMvL91iCK8/LF+WrT0FzHaNbcVt21Tf0CBwxuOAjVAABwZkopodQ0KXvTfGeY7gzTWcLTQngWCx19sWF1a2bTN5lNCFEQq8+kWl86jovDSb4zTPdH2TTlJZeexeqevtiw5kKz6ZuuyTBBcGoBOEEhpZTKeZrks2kymKSDSTKYJsM4n6VFVPBcKakzw9Bsx/Q9MwzsZmDXA6vhmIGNPfwDdwIZJZTg5Tk3dHWNEUbxd5vTOONKIY2SRmAsd9w7K8H6gl8PTEOnEKnB28MIYYyadednHy1pGoviQgjJ+V+KgSuMfN9cnAvv35r79CdLUBUPXCC4+AAA4Mz6s2JvlPaneX+a92f5YFakuci5jDI+S3la8FFUNP28HZhzNTOwYRv1aSmEDif58XM7jMo0F1yqKCujjCc5H86Kpp+3Q7NbMx0D3sUAeC4v06xMxkl/HPeyMk6LOM6ncT5L8mmSR3mZFjzjolRIMqJpzJilo4neH0YHtuE7hmfrrqk7vlWvOS1L90zNPp6uq6oQlkEpMdYXPI1iSoih0TTnOiNrC96tRX+h5dR8w9AIJGpwjiyDtRvuxlJ9FmWGTh9v9tOsFFJ6rhn61u3V1p219upiPfAM2MMPLhB8HAEAgLNRCO2N0t89HA6jYhKXhZAll9WCcJ6qOOejuNjqp6GjLzUtQ6/7tgbv86ekFNrsJ79/PBonxTQpCy65UNW5LYWMMjGYFVtGGjjaxpzrmsw2GJxbACppEQ3jg0f7f3y8/8dCFKUohOBCCqWkVFIppZBECiGEhBQFz5M8whgTTAihjDBGNY0ai41bt+c+bgdLOjMpPlmQnxJs6nR5zq0HhsaoY2n9cYYx+mAlWF/0O3XLtRiszQHnS2M0cMjKfMgYxRgd9GclF6KU9dC5vdp6cHf+wd0F22Kwhx9cLAjVAABwBkkhoowfTvL+NJ+lPM55tQe4+qpCSgokZRUFlamTw0nuW5prMv1md6U+jTjjs4z3Jll/msW5SKutm3/eOqeUUkgJpYRUQirfyg8nmaFRz2TQnwzcWFVgniT9YXQwiPYHs/2DyeY4GUgphBJKqeO7T499W/Xf8wreGOMCYUIIJWx//EwhNYwPG+5c3Z1ruB1C2PGN1hgjjRHX0pbnXNOgw2muFFqec5qhaRkMugSDc1etknBsvdvy7q63i1IMR3GU5KuL9Q/W20vd0Hd1SNTgwkGoBgCAM4gzvjdMe5NskhR5KV/ZqbXaFRznaBwVB+MssDSNYgjVrzVN+e4w7U3zaVqWQr18bhVCSqpCSaXUMMr3R5ltMFMjjEKDa3BDSSWFLA+n29/s/v/96d4wOixFwXmp0Kuy9A+o7lhJIYUQvenuKO4FdqPutm/PPfDM0NAweql6mcbIfNtu18045UIq12aGDrEGvEOmzgyNMtqq15zhKBmN4+X58PZaU6OEwLJvcAlAqEYIoWd/tru7e9FjAQC8V7Zt37p1a21trdvtBkHw2sdPk/LJYXw4zQuuxMlGrX+hFBJSxbnYHaa2zkJX8yzYWf0ao7h4vB8NZkUplHzV3YqKQopLNEv51iC1Ddb0DEuHUA1uHKmkVKI/3d0dfr87fnI42YmySV5mUskzJerjFFJCcqlklI2VEozqeZnN19fma+uUUPLianBKMNaIg5hUSmdQmQy8W9V8tW3pbYxdS2vVrNC3DA1u5YDLAkI1Uko9fvz417/+9a9//evf/e53Fz0cAMB71Wg0/umf/ukXv/iFYRinCdWTpHxyEPeneSnkK5dVVhRCQqokFzvDzNToxpxznoO+pkaz4tF+NEs5Fz8WCZRCQqko41v9xDPZ7a6LENywADeOUrIU5f548z+f/q9JOsjKpNpB/dZPq5QSaREXPEvyaHf45BP5ty1/ESGdvLQkhGBswC0t8B6ZOjU0WvMMhBBCEKjBJQKhGiGEpJRCCCFEWZYXPRYAwHtVlqUQQsofnnQ+8Xguo4zn5Y8l6iNSqazgSSH4D8+7giMFl1HKC36qVCCESp/vu37X4wLg0qm2PW8PHj3rfTPLxnmZCsGlOu3fsdc/v1JVMTOl5M7wsUb1xcbthfo6RgghyDHgImGM4CIElxCEagAAOINSqCTnBT/Vh1cpVVbIrBCQ/E6jEDLO+aluVyAklEpLkZXiRxaKA3AtKaUkEqPo4Oud/+hNt5N8yiU//Z3BUx5CKcVVIRXfHT2ZpkON6d1whWACk4PgSqjeSZ6X5Hvpq1Uwx9X/wxUNzgOEagAAAACAKyPOp8Nof2/8ZJaN8jITUry6xPdbUwhJKQuexTk+nGw/7X3dcOdCp/UujgXAeZFSFVz2Z3l/lo/jcpKUccbTXFTLoHRGLYM6JgtsLXS0hme0PENnBArXg7cEoRoAAM4GY4wwOlUlIIyqx8J79WngqhSNwqcss4SPfhYA3BgKqTifbg8f742fRek456mUb16Z7DXHUkohVPJCKnE43dY1kxEtsJuw/hZcTlIpKVHOxSzhTw/j7w+izX6yO0z703wcl0lWIoQsQwtdreUb83VrqWmvd1yNEM9ihkYJQVBvD7wxCNUAAHAGGiOOwUou0+L1W38JxqZGTZ1Cv4/TqM5tVsq8fP25pQSbGjU1qDkMbhAheSnySdLbHT3pT/dynimp3lGiPiKVRAKN4x5CqGa328GCRg1G9Xd6UADOSik0Scr+NP96e/pwdzaY5cNZPonLWcrTQuSlKLlECClVSiHTlI+m+XYv/nZr+jvfuNX17i35Tc+oORpscABvBkI1AACcgcGIb2tZKfJSvnb3LyXYMZlnaYxCk+rXMzUaOvo4LoryR8t/I4QR0ihxTc0x4YYFuEG4KONsMowOD8abo6gnpXjXiRpVRcuUmCSjtEjma+vTdOSZNQjV4FLhUhVc7o+zR3uz//dPh//70bDkgnMl5ckGjZzLLONjhPYxphQzhjVGf7pRxwTd7nqmRnWNMHhbAWcHoRoAAM7At7X1joOQSjJeSvRDVbIwQoRg26DzdfDLcmIAACAASURBVHOxYdnQdeYUaq5+q+s+OYiTnAuJfuieBcaIYuyabKlpzdcsQ4MbFuCmSPLZs/53u6MnWZmot+hH/QaUUkLy3mT78cGXa617tuG9t0MD8FqHk2xnkP7Xk9F/Px1v9uI8F0JW5fZ+5HdESYnKEkkhvt+bcS73R9lsjS82rG7NglQNzgpCNUIIWZYVhuHc3NzKyspFjwUA8F7VarVGo+E4DmOn+nsY2Gy1bUdpuTdMVSl/qLUWxlij2DXZfN1aqFuWAaH69Wqutt5xZml5ME4LrpR4xbnFCBGMNYY9iy02rLmaqTMI1eAmUEqpuJjujp4cTLbyMnufiRohpJQUkg+ifUJo3e10whUMJQ3AJSClElLtDNL/ejr6j4eDPz4ZCXFydvqVlEJKKYSUEGinn+yP0rQQVQ2UpmcwgqF0GTgTCNUIY7y+vm7b9scff9zr9S56OACA98owjKWlpW63W6vVTvN4x2TdmnUwzn1bwymvyqKcSNaEYI1i22Cho3dCsxkYMJt6Gp6lEYx3hqlnaXHO01xJ9cK5reb/dUYcg9Vdfa5mNTxdg1ANbgAhZSnyOJ9Nkn6UTYTk76ji949QSiVFNI57UTbJy1RjOiPaex4DACfEhZgm5Vdbk99+1dsfpUL8+Oz0qymlhEDb/ZgLqTPSrVuBrXkmpCRwBnC5IIRQt9vtdrsXPQoAwBVgatTQaDsw5uvWMCq0iGSlyLmUUimFCHleQMsyaM3RuzWz5RuBrcHt7tOwdWpqtBOa83VrHBeTpMxKWZRSqufnllFiasQxWOjoczWr6RueBecW3AhC8qSIomwcZZOsjKU6z67Up6SUKsosIpMoG0f5xMUBhGpw4UZR8eQw+mZ7+tXm5OUd1KdUzVr3J/lwVszVzJW2s9ZxIFSDM4HLBQAAzgYj1K1Zf3Wb9KdZb5ofTvLeJE9ykZfCNphrsXZgtAOj5RtN36y5OqS+08MYLTVsU6P9adab5IeTfDDLk1yUQtkGDWytHRitwGj6Zss3PIvBuQU3RCGyYbQ/ig9LUVzsSKSS03TYm+5QwizNudjBgJtMIYSUerQ3+9UfDx7vz57ff32bJ1RKSvR4L/rX/9r7+4/nlhs2glLg4NQgVAMAwJnVXM23Wc3V6q7uW5prsijlaSFciwW21q2Zc6HZ8AzP0qA29ZlghBqeHjpa3dXqru7ZmmexKOMlV67Fao7WrVudwGh4hm0yCp92wI1R8nwUHY7iXiny97uZ+gUKKSnFNB32Z7uh3USQqcHFKbjMS/GsF//n96M4Ld++u5xSSCG1P8pmubjT9WYZNzRqwA4jcDoQqgEA4MwwwpSgwNJMjdY9Y33O5VwKqRglGiO2Tk2dGBqlBOr4nBnBGBMUOrql06ZvZF2vFFIqpFGsM2Lp1NSpwQjFMIMAbpCSF+OkN4l7RZlLJd//huqKUooLPktHg9neYv3WhYwBgMokLvbG2eEkTzNecnk+N5sUKoVMMn44zr8/jOdDsx2Y5/G84PqDUA0AAGeGMcIImzo1dRrYsKvwPGGMMMaWTi2dhjAPBgBCCCEuiyibRNmkFMVFJWr0fH2sSPJomg4Lnl3UMABACA2j4rvd2cEoywoh5blVGRBC5grtj9PvdmcmIxCqwSnBkgYAAAAAgEtNSJGVSc7TCylRdpxCquBZWsRclhc7EnDD9af5l5uTvVEihHyz+mSvJKUSQu4N0y+fjXvT/LyeFlx7MFMNAAAAAHCpKSVLUZS8UBceqpXikhc8k1Jc7EjADTeJy6cH0XBanGOiRn+uBD6aFU8Ookl8wXUBwRUCoRoAAAAA4PJSSEmkpORSiQtc+31ESiGlkEoqpKBuBLgoSc57kyzO38mKiTgv+xOU5HDnCJwWLP8GAAAAALjc1HMXPY6KkkqiyzIYcEOVXMZpWZTvZO1GUcoo5SW/4IUh4AqBUA0AAAAAcImpY/97qVzCIYEbQyrFhTrftd9/eXKp+Llu1QbXHoRqAAAAAIBLDB/7XwAAQgghpZCU72q9hEJInVOXLnBDQKgGAAAAAABnBzkfAAAQQhCqAQAAAAAuM4wwxhhjgjHB+OKDLMaEYIoxfIYEFwljRCl+R78RGGNKMbn43zZwZcAfRAAAAACASw0jQgmjhF70QBBCiGBCn6cZyBzgwhCCGSX03UQZShCjmFyCe1jgqoBQDQAAAABwqRFCdGZqzCQXPT+MMdaobmg2JdCWFVwkU6c1Tzf1d3IdmjoLPcPQL8VtLHAlQKgGAAAAALjUCKamZhnMJPiCP+VjhDVmWprDiHaxIwE3nG2wZmA65jsJ1Y7JWoFpG3DnCJwWhGoAAAAAgEtNo7pvNwK7oTHjArdVY4wpZa4Z1Jy2oVkXNQwAEEJ1z7i3GHRCk9LzrDWAMaaUdELz3qLf8PXzelpw7UGoBgAAAAC41DRmhHYzsJu6ZhByYeXKCCaMMt+qNbw5U7MvZAwAVFq+8ZPloFu3GSPk/EqKEYIZI3N168OVsOWb5/W04NqDUA0AAAAAcKnpzKi7czWnpdELnjojmPhWveUvWLp7sSMBN1zoaBsdtxMalkE1dm6hmjFsGbQTmhsdN3RgjwM4LdgqAAAAAABwqenUrLudcdK78FCNMfXtestfsDTnYkcCbrjQ1nxLawembTIuJCoEUm/9pBhpjFgGawfmWtuB4t/g9GCmGgAAAADgUiOEGsx0DD+wGo4RXEjlbYyxpTuh3XSMwGRQ/RtcMIwxIXhjzv37B3PrHfftd0VgjAjB6x337x90NuZcQqClFjgD+IMIAAAAAHCpEUx0Zjq6HzjNSTrMy1RIrtTbT8ydFsYYY2IZXui0XDOAKmXgMsAIrc+5nq1FSfnt9pQjKcSb/1JUja83uu4vPum2AwPyNDgTCNUAAAAAAFeAbXjLjTtClHE2LUX+XkM1woywtr+w0fkotBvv7bgA/DjHYMhH95eDUVw82Y+2erGU6Ky/GhhjQtBi01mfc+8vhS3fcKCZFjgjuGIAAAAAAK4A2/AWG7eyMt4ePk6KmVJcKfl+Dk0I1ZnR9hfXWvegRBm4PCydWjq9u+hLhZRC/WlelKI8yzIOjJHGsK7R1Y7z1/da9xb9pmfAum9wVhCqAQAAAACuAEqoqdmh3ZqvrSGERvFhwXMp322uJphggmtuq+HO1d2OpTuMQklkcLm0A5NgTDCqufo325Mn+7GQUgil1A+ma4yrltSYErLace4tBfeX/J8s+03PhEQN3gCEagAAAACAK4BgqjMa2I35+nop8qSYVTur3+k6cEIIo6zmtJeat+tux4D21ODyabh6zdEck7YCAyE0y3iWi7wUnEsulJRKKaSQQghhhKuCZIxixoihUdOgd5f8//uj9mrLWahZ59jyGtwoEKoBAAAAAK4M2/AX6xsFz0bxoZBcqURI8S5yNcYYY6wzw9SdTrC42rofWPVzPwoA5wIjHNr6ehuzB+SDeW9vlB6Ms/4sH0XFLOVJzotSIoV0jdgm8yxWc/WGZ8yF5lzNWmrayy0nsLW3LSAObjAI1QAAAAAAV4ap26ZmxfnsYLLFRclFqZCS1WTc+cEIE0wIIY7hh06rHSx1wxWMoRUruKQwRq7JXJO1A/MnS8H3B9GzXrwzTPZG2WBWTOIyzUuEkGWwwNEbnj5XMxfq9mrLWeu4tk4phTwN3gqEagAAAACAKwMjrDCqu537C39laU7B/zPJZyUqpJTynOqWYYwJJhrVNGrM1zfudD9p+4sYE4wgd4DLjhBsaGSxYQW2tjHnJrnIClFwyYVECDFKdEZMndoGdQzmWczUCIGG1OCtQagGAAAAALhKMMKeGVq6w0U5Tvqj6GCWjUtelKJ4+/lqjDEjjDHdNXzXDBfrGxudjxjVIFGDK4FgpDPS8IyGZ7zxkwippFQE46qeGWRu8FoQqgEAAAAArhiCKSN4LlxmlG0PHj7rfztJh1E6kVK8zXw1xoRgYhqOZ4aL9Y3l5t2WP69RHRZ+gxsly3mccUOjukY1RhiFVA1eA0I1AAAAAMAVgzGmmIZ20zNDRjWhhDXbH9KDJI+SIkJKKXSGquC4+g8Tg1m24YZOs+62V1r3Ntofasyg5IWPi0IqIWTJpZRK1yhjhMBUHni/lEJcSC4kJZgSggl6+xXcJRdFKYpC5CUfz/LxLA9cM/BM39FdG9rIgdeAUA0AAAAAcDUopRD6S4LFmFDK2v6ipbvDaL8/298ePNoePBSSCykkOlX1sqrKNyWUEtbyu0uN2y1/oeHNeVZdYwZ5cY5aKZUVPE7L4SjJct5quDXf0jWYygPvj1JKSjSLy9Esd0zm2pquUV17qytQITSNit4w2u/P9nrT8TSdzPLlbrg0H64v1iFUg9eCUA0AAAAAcAWUXKYZlwpRijVGdI0QjDGirhm6ZugYvm81GGEYo4LnBc9KXpS84LLkshRSSCmUkgohgjHBlBJKicYoY1TXmaExQ2dGN1xdaX7Q8LqB3SCYHj+6UijNyyQthpOsP04OerM0K9dyLqSq+6ZtaRh2XYP3oihlloudXvx0Nwpcve7rnqO7tqZrRGeEMUJP3WuaC1mWIkqLWZTv9Wa7B9Odw/HOwWQW50lSZHkhpKj71lzLg+UY4MdBqAYAAAAAuAKSjG8fxAWXhk5rnt4MTcL+8jHf0l1GNVNzFuobaRElRTRLR1E2nqXjKJ+keZyVCReFUpJR3dAsS3ccw/es0DVDz6o5umcZrmP4juHrzDyxiVohJJTa70ebu+Pt/fHuwbQ/ioWQUZLnRUlWmrrOGIXUAd6HKOUHw/TLR6P//U3fd/S6b3TqVrthtUKzERqerdnmaQNOmvPhJHn4pP/ld/u9YTSeJElWpnnBuRRC9QaRaWjrS3lRKo0hWI4BfgSEagAAAACAS40LmeVif5B+tzmJU24arFM380L4jm4aVGNEY0SjukZ1S3cbai4r46yIZ9n4eajOJmkRZWVSlQfXqG5opqW7fwnVZmjpnqnblLAT881cSM7leJYNJ8n3W8Pvt4Z7B5OD/ixKcqUQpZgQHPiW55q2QQmDembgnYvTcreXbO5HT3ZntsF8R++Ps4NR2qlbcw1rrmG3aqZlUkOjr30qzmWa8b3e9Mvv9kaTJM3Kox0TGKPxNNV1Npwks6TwbMYo5Cbwg+DiAAAAAAC41LJc7PXTh88m//XtYDTLNUbnGtb+wFvqOAttJ3B1Rp8vTq1KjhnMYkS3dLfudLgsueBScfGX5d+EYEIJo4QxqjGiUapRQik+magRQlnOZ0nxh693//vbvcNB1BvFWVZkORdCIIS398YY44VO0Kw5GjU1CNXg3YsSvteLJ1GhFEpzUfJslpQ7vThw9JpvfLAS3Fr2F9uOEb4+VBOMKCFSqjQri1Icr0GgFIrTcjBKesN4MI416lgG5Cbwg+DiAAAAAAC4pJRCUqo45bu9+Nl+tNtPxtOCUpxkPMn4NC5mSdlt2nMNyzaZZTCMEUa4CswIvXmf3iNcqCznB/3Zd0960zibxbmSSimFEcIYTaPscBBt748bNVtnRGP0KN4DcO44l3kph9Nsp5dMohIpxKXkAmWFwDGaxeUkLkJP7zSskp+qsRxj1LE129INg7GUCCGOl/YrSxEn+WF/urU7ckxW88139cLA1QehGgAAAADgkpJKFaWYRMX2QbzXT7JccCmFwsNJHqXlYJLvHibri166HMy3bKNJCTrnSKuUElIVpcwKXnKhpEJI4edfQlKqJMk3d4aurbu2YVu6ZVDYegrekbQQo2m+P0h3eukkLhQ6NrGMUMFlmgkh1enba5kGY9Ru1exmzSkKXhT8hedUsizF3uH04dPeXNNZ7gbn/HrANQKhGgAAAADgkiq5HE7z3X6y3Yt7oywvRdWBOpciLwXnMsm4ZbJWzaz55zAv/TKNEcfSfc+oBZaUMs+5kugoeCil8oIf9GeubbSbvucYNLTp6YsvA3AWacYPhmlvnE3jIitemFXGCGmM2Cb1HS3wdEN//dpvhBAlmOq0FthL3VqaleNpisQLK8CFlINxvLk3Gow7acY1jTAKexzAK0CoBgAAAAC4pPJC7PXTZ7vRzmEymuZcvNB3Oi8lF0WUlHkhpVQYo3NfeW2Zmq7R+ba/slCTUkVxIZAQx4ZRctkfJ4Yx6R5MaoHlOrppMATttcA7EGd8r58MJ1mec/Hi7wLC2LFYs2Y+3w1xlv3PgWduLDfH0/TpzhCJF74kpZpEGTmc9kfxJMp914BQDV4JLgsAAAAAgEtHSpVmfDjNtw6i7cN4GhcFl1K9ECR0RnxHq/lGq2Y61ivKjL09SrCu0W7Lu39rrtvyLVOjL4YKpWSe8/E03dwdPdsZDcdpmnMp1Q89IQBvQEiVFWI8K/Z6yWhacKHUi78LGCPf0bsNu+4brq2dqWaeYxndtt+su65j6PrJNM5LkST5zv7k8dZgGmXn8GLAdQQz1QAAAAAAlw4XapqUh8N0cz/a7SdZIV5+jGNp8y17peuuLXi+o7+7AmFzrcAyzcN+9P1mnwuRlxz/OdEohZSScZJv7gxNnXWanusYGjUoOdX6WwBOo+QyTsvhJN/tJeNZceLuEkKIYBS6erdlB56ua+RMN5gsU2vVvWbdrQeOkqoo+MtHf7Y7cmy97ptzTfdtXwy4jmCmGgAAAADg0im57I3S7YP4cJhNo4LzF1IEwZhREnr6ypw717A8W9O1d/ihzjJo6JvzbX9tqVkPbJ0x8uK+6bIU40l60J/t7E8OB1Gcci7US8EHgDeU5aI3yg5H6WCaJxk/cW1pjNgGa4TGfMt2bY2csQK9rlHf1RuB1W64rmO8/M1CyMEo3tob90dJnJZcnKq0OLhRIFQDAAAAAFw6RSn2B+nmftQfZ3HKxYuf4wnBukZrvrEy77Zqpq6Rd1odjBBsaHi+7d9d77QbvqEzSsnxWCOkjNO8P4y29sZ7h7MoKUt+coEuAG8szfn+ID0YppNZkebixJWlM+LaWis0uw3bMc+8DldjxLVYPTDnWq7nGIScrE0gpZrM0r3e853VRfmKZSPghoPl3wAAAAAAl4gQapaUu71kcy/e6SVxxl9e7GqbrFUzlzrOUsep+QZ5x/W2MUKU4FbdKbnsDWeHg+l4iriQSKEqOVc1yZO02NmfeI7RCG1DJ8w33/XAwLWnFJJKRSnfH6TDSV5y+fLNGs/W5hpWMzRrvmGeru73cRgjjHHgWcvdem8QP6GMI3H8NpZSqhQyyYr9/mxnf6Kzmm1qb/vCwPUCM9UAAAAAAJdIKeRgnG3ux0/3Zru9JM1P7vBECLk2W55zVrvuYscJPf30jXnfGMa4WbM3lmsLHb9Zc2xLp+TkYdO83DkYP9sZ7h1OJrMUVsmCt6eUEkJFSbk/SMezQryqBp7v6gttp1UzfVfTtTfczO+75upivVFzmUYJORmRlFRZwfd6083dYZzkb3YIcI3BTDUAAAAAwKWgEJJCpRnfH6RbB9F4VhSllC8mU40RQ6ftmrky57Zqls7oe0jUFUKwobPFueDueocLFSdFVpRCiqPDCyGzrByOk829kePorm1g39T19zdCcP1khRjN8v1+ctBPJtHJUI0xIhiHnj7fsj3nre4u2abWrNmN0K759mSWlOWJm1lKcDmZpAeDaBzl7Vy86z0X4GqBUA0AAAAAcDkoxYVMMr7XT7b241n8ipJIGiO+rbXr1sq816qZ7/ljPaVkYS5UiPTH8X5vWgpelAr9eQuqlCovyvE0frY7dB2j3fAMg2mMIigEDt5UmouDQbrTS/YH6TQqTsxTE4wpwaGrL7Qdz36rJdmWyUyDNWt2PbSKopyd6J6lkORyMkt7g2gyzZKME6JBqAZHIFQDAAAAAFwKXKjeKHu2N9s6iA+G6SsXfjsmW+w4i22nGZqOxfD73clHMHJtvVWzFzrB4SDCh0hwKaU63pg6zflBbxZ6VrcdWAZzLI1QCuEDvJks54fDdDjJy1et/PZsLfSMbsvuNCzHettcgzHqNNyP784rpQ4HsxNflVIlaTGapr1hPBzHOnMNTX/LI4JrA0I1AAAAAMClUIXqJ7uzzYPocJS+8jGOzRY7zmLHaQSGffZCx28JY+xYGkb2fNvvDeMkLaI4Lwou0V/yTp6XvYK7trHQmdR9q91wNQ1CNXhD1Uz1cJq/cou+52gLbbvbtDt161wO1256D+6R/d70v6tL9liOl0olGR9P094wGozimm96DoRq8ByEagAAAACAC6YUSnM++j/s3dlvHNmaIPazxb5m5J5JilLd29c9d257BoYN9EMDNmD4f24/DIyBgbHbxjx0u+d2912qJJHJ3DNjX0+c4weqVGSSUkm3KJJJfr+HQqGYYIUiM1LxxbfF9WyVnS+yLG9uv0ZXqWerr4b2m4kzCAzGHm3cLGNkOvLLWiRZGadlnBSt+GlC+dU48CSr3l7sLFP1HL0fWI6pwiRw8FXqpk3yZrHJZ+t8F1dte8eIMsdSTgaWY93bLG7LUIZda9iz+10ny6q8qD/+SEoppSirZrGOZ4E1Hbr39T8FzwAE1QAAAAAAj0xKmRV8E5azdX6xyrPyjsJvXaODwDgd2q/Hdt/XHzH7yxgd911CyWwZLtZxVTVVzRGS1+PqNKveX+4/dFarzNQVCKrBV6kasYuqxbaYb4owrnh7OKIMI+xa6nRg/sJu6utMXTF1ZdB1Rj13IaPrQTW6mhpQNYtNHPhmXg6llBiG8AGEEATVAAAAAACPrhVyE5ZvL5PFNo+yum5uVLpijAjBtqlMB+aoZ5iGwihB6NHu5jFCCsO2oZwMvX2Ucy7qpq1qfn21b9O0UVIsN8n5fG/oimPplsEYg8AafKmqblf7chOWZcVbIdHN9dSmxmxTGffMk4Ht3ncZ9rBr/+43o7ZtN7vsYC12w8U+yhebeL3Lhj3XMhTl8WpGwNMBHwIAAAAAgMckhKwbsd4Xb+fJel+medPwG0E1IVhTqG+rJwNr1DNNnVH6mBkyjBGj2NLZZOC+mQbdjmXoV3H+TxreJnm13qfn83C5SZK8rprrndcAfJJESEhZlHy1KzZhWdZtK+TBZ8c02KCjjwJj1LuHEWUH+oH9198N+oHNKD6osGhbEaXlepsut+k+Kuqmvd//NThSkKkGAAAAAHg0UqIorVf78t08fb9Ik090U4+65uuJc3ZV+M2eRLpXUeiwZ9eczzfxPs55K3grhJAfM3tYoiyrzueha+v9wKYEmZoF5bLgZ7WtKKt2F9eLTbGNqoNnTFdsU5kMLN/VVEbuvbPAMtVB1+kFVsc306zKixtXJZYoL+o/v9vYhmoaE8u4t+JzcLwgqAYAAAAAeDRCyjCpL5bZxSqfb4rbhd9XA7enA/NsbE/6pu88lYHDjJLAN6SU06G32WdF0VQVr5v2erlsUTTLddxxjXHftQwl8HSVUWiuBp/HW5nkzTYql7tiH9e3u6kJxq6lTvtmx1G/xcQ+U1cxpr2O1QustpUHQTVCqCibtxc721S/exX0fJMQAg+LXjgIqgEAAAAAHocQknO5Dsu3l8kuqupGiJu7eCnBhsa6nnY6tEddQ1PoYx3qnTBCmsZOx35W1HlRF2XdSiGujXZqhagqvguzt7OdZaqBZ7q2Bpk98HlV3a735WJb7JOqqPjBRaEyqqu052mTe537fR0lSFNJ1zdPRn5eNPsol/JGb3XD212YLzfJZpf3OrZtqtBZ/cJBUA0AAAAA8DiqRmQ5X2yKd/N0n9yxiZdR4tnqMDBOBtYgMFTlyd24awodD9yy4st1HMZ51bScCynlVfu0EFKINoyL8/ned41h16GUGLoCuWrwKVKism5X+2K5LZKsqW41LWsq8R215+ujrmHf39zv6wjBCkKBb56MOstNSmnUtuJ6VN22IsnKzT5bbJJ+19YUCkH1CwdvPwAAAADA49jH1fez+N08uVhl6V3d1JpKR13j1cgedg3XUih9cndujBLP0gZdazLwhl3HNjWF0YNa2Lxslpv0chlfLKJ9XLRCCphYBu4ipeStyAu+2pWbsKzv6qa2DGXcN/uB7tmqrn6r2g2MkWcbk6Hf8SxDVxm78T+SEkkp07z+4Xz3brYvqjsuXvCiPLmvZgAAAACAZ08I2XCxCcvvZ/HlOt/FVVnfyMgRjFWFuJYyHVgnQ6vjqobG6NPL8BKCDZ0FrjEdedOR33ENQ1couXGHWdc8iovVJjmfh5tdlheH480BuNIKmZd8n9SrXbGLy4PPCUYIY+yYyqRn9jzd1Nm3zA9jx9ZGPSfwTNvSFHYYvUuJirKZLaPZIkrzmt8eUA5eEij/BgAAAAB4aFUj8pIvNvnbWRKl9e0XMIY7rjYdmK/H9qRv6uqTvmfTdeV03CnrdhvmaV7XTcuvlctKKdtWhEnx9nLnuXqvY/U6puYbj3vM4AmqG7EJq8t1ttwVYVLzg6CaYEaIaymTvunZ6jd9woQxMjUlcGWvY3R9s6qaLK8OXtM0bRgXm322j0vfNSzjcLEceDngjQcAAAAAeGhZ0ax2xeUmn63z22u0MEIqo31fPxlY454ZuJr6tDs2VYX2OtZ04I0Hbi+wTF1l7KdV2h9qZbNqvoovl9FsFUVJyVspoAoc3NRwcTX0O0zqomrbwxFlxLaUrq+Pe6ZrK996PZuiEMtU+h17MnBt846p+7wVSV7tonyzz6K0am/NRAAvx5P+ggYAAAAAeJbCpL4q/N5GVVEdjmLCGF91U58Obc9RNeWpr6EiGGsK8R3t1bhzOup0PNPQlYNjLusmjPPlJjm/3G/CvCjbg1VJANSN2ITlel/enk+GEDJ1Nu4a45457BqOqTzAFitC8LDv/NXrgefcUVghhCirJkrL+Sra7lJoanjJIKgGAAAAAHg4DRdp0Sx3xfezZLUv8oofDP2mBFsG6/n6ydCaDEzHVCj91jm5XwpjRCm2LXU6dE8nfi+wbFO7nqxGCLVclFWz0syTMQAAIABJREFUDbP383CxTnZRXlT8EY8ZPClCyLJuo7Re7Yr1vqzqOwJU02CjntHv6I6lqA+yXo5gHHjm6dgPfEvXDicFCok4F2lWLTbJep+VNYfaixfrSffnAAAAAAA8M1nBN1F5vky/nyVhcmc3NekH+pup/XriTHqm9s3mG987XWOTgVNWzXwVxUmZZCXnrbi2ighLFCfl+9k+8Myub1GKPPuOqlrwAvFWhkm12Oazdb4Jy/quTLWls1HX8G2VPNRDJkywY6mDrtkPrI5vRnGRF9euWSklQkVZLzfJepfmBXctobAHOzrwhECmGgAAAADggUiE4qy+WGbzdb7Zl3l5I1WLMWKUWDobd81XI7vv68c1+ohRYptqr2OejPzxwPEcXdPYQZK9rJpdlC3W8fv5frPP8xImgQOEEOKt2EXVYpvv4yor+EFrAKPEMljgapO+5Tvqg3VDYIRUhTqmNhl6r0+6rq0fBsxSVjXf7rP1LguTsqg4TAp4mSBTDQAAAADwECRCSKIord/OktX+6v77xgswxrpKfFud9s3TgWXqR3mfZhrqq2knK+pdmFcNr2qOroUZnAsh5Hafvb/cdTyj17E8R1MY5Ktfuo8jyrLiqrzhxk91jXZdbdQzpwPLd7QHzgUzRs8mnbxoojhfbZMPGeqPR960YVysd+lml3Zc3dAopUdTXQLuy9E8+wQAAAAAOGpFxddhOVvnP1wm26jkrTwIHSjBvqON++ZkYA0CQ9eO8tZc19ggsKdDbzxwA8/UdYUygn4Mgq7Wa0VpebmKL5fRfJ2kWS0lguzeiyUlqhuR5s1qX652ZVnx2wufDY0OAqPf0T1L1VT6wPXVjOJ+YL2edvqBbZkauxkzt0JUNQ+j4v18v9omDb+jcB08exBUAwAAAAA8hCRr3s2TH2bJD5fJPr6jm5pS3O/oV63UHU/THmQU071TGe24+rBrTUfeoOvYhqoyitGNMCjLq+U6uVxGF/N9nJatRLfjKPBCCCnzku+Terkt17uyvGtEmaGxYdfoepqmEvrgk/AJIZ6jj/r2oGt1PEM9GHMgkRQyTos/v1tfLMI7u8HBs3eUZUUAAAAAAEekFZJzsY3Kt5fJYptnJW9vtoxijHSVdhxt2rdejWzfUZ/4YurPIAQTgj3HeDXppHm9DXPORdW08tofuWmEEPV6l57Pw8A3O57lmIpxnOXu4BdqWxml9XJXbKMyzg977CnBCiO+o477ZuBqjD7CGDCMkcKIZajDnjsZeEXZ5EV98BgoL5r5KlmskyipDE3R1Cc/sh/cq2P9vgYAAAAAOBaci6zk63359jLZRdXtrCzB2DGVYdc4HVmvRpZtKo9xmPfJsbTX0+7raXfYc2xbO5gsJYTgvN1Hxft5eD6P5us4K+5I3YOXgLdiF1fzTb6Pq7zk7c1JA4wR21C6nj7pmR1Xe/g09bUjocOe82oa2KaG0OFhVDXf7NPVNt2GeVY0MK7spYGgGgAAAADg28pKPt/ks1U+3xRxVstbUTUhOPC0VyNrEBiudcRp6o8YI86HSeDesOs4tqaq7HrqTkqZ5dVml87X8WwRhUlZN6KFUOSFaYWs6nYTlsttkRX8+gK2K7pKe77W7+iBq5k6w493ZTBKOp456jkdzzRvjeUXQjRNuwvzP73dXK7ig+Xz4NmDMhsAAAAAgG8rzZr38/R8mS63RVG3d2SqCe75+uux0/M1Qz/sQD5GlGCiYN/VT8edOKs2+0y0smladG0kWVE1XIjFOvYcvRdY/a5jqPQRU5HggUmJOBdFxbdhudwWRXVHN7Ku0kGgDwPds1VDe8yaakqx5+j9nhP4pm1pbSuvR85SIinlLsx//6eFY6lvpr6mQpz1gsCbDQAAAADwrdSNyCu+2BZvL9P1rqy4OKgLxRi5ltrz9ddj+9XYdi31EXpGvw2MkW2q05EXp+V6m3IuirLhvP14BqSQnLdhnM8WYS+wXUcfdS1V0R/3sMGDEULGWb3YFstdsYuq6uaIL4wxIdg22Khrdj1NVcjjdikTjE1d6Th6r2MFvlmUTVk1B6/Jy3qxTpbbdJ9UlFH9ZnUGeMYgqAYAAAAA+FbKut1G5WydvZ0nm7C83WmJMQ5c7c3EeTN1z0Y2eV55WstQNU1J0nKxjtOs2oa5EFKIn2InKVCUlARHgW+5lmYbiu9AUP1StELu4mq2yhbbYpdUB5XfBCNGiW2qw64RuBqjj3xpEIINjfmO1gusrm+vd5nECN+8oMvqY2d1YeiKpsC4spfi6Dt2AAAAAACerKxoLpbZ5SbfJ3VRtwdhA6PE0OgwMF5P7I6rMkaeWVBNCFYo6bjG2TSYjryubxr6jRlsUsq65klWrbbJxSJcbdMwqWAp0QshhAyTevGJbmpVoT1fG/WMYdf0HI2Sxw9bMEaqwoY9dzL0bFOj5DB5LoSom3YfF+8v99t9BuPKXo7H/3QCAAAAADxXWcHPl9nlOo/TurrVTa0wYmps2DVejx3XUh/pGL8hjBDByLX11yfBycjvBbZlqgdxSMPbrKhXu/RiES022S4uKwiqX4ZWyH1cL7dFUfHbP9VVOujok5456hqupdLHzlRfURU67Dono45j6bf3ewkhm7YN4/ztxXazzwSsX38xIKgGAAAAALh/eckv19kPl8kPs2S9L++cBtxx1TdT52RgDbqG+Xy3NBsa7XrGqOdMR17gW7rGbkxOlkgIkaTlchOfz/ezRZjDeq3nTkoZZ81im1+u88W2yMs7gmpNpYPA6Pq6wp7QnAFGsedo/Y4R+KZj64zRgxdgiYqyWe/SbZgneVM3MAb8RXi2X98AAAAAAI8oL/lsnb+9TH64TPbxHWu0EEKBq303tU+G1qDznBuJNZVpKht27ZOhH0bFfBVJia4/ZRCtTNNKCnnh7j1HmwycQYCeThwF7p2QKE7r+Tq/3GSLbXFnmbSu0UHH6Hqa8pQ2zFFKXEvtdcxex/RsvSzr+tYjoKJs1rtsF+VpXqsKfVIPBcA3AkE1AAAAAMB9aoXkXIRJfb5Ml7uibsT1PVJXTJ2ZOpsOrDcT17OfYeH3ba6tn007cVrso2y1SeuGS/nTowYhZNO0+6i4XMbnvUjXlI6rmzcbsMGzIYUMk/pyk6c5F1LKmxeIqhDbUMZdYzIwu57+pILqK6ah/vqsl+VNVtRZUR9c33XD47TY7rPFOlEosQ2Gjn9JHvg8CKoBAAAAAO5TK2RRtbu4Ol9kq11RN3cspjZ1Nujo04F1NnEc80WEjo6tvWIsjIvNLqsqvotycW04lZSyadowzudr9n5uW4amq9TQFEjxPT9SIi7kPqnmmzzNm9tpaoWRwNNGPWPSMwNPe4Kry01d+e60lxfNn9+v1zsixY3u6aZpk7ba7vPFOvZtbdSz4GP87D25Bz8AAAAAAEetKPlsnb2bJ5frPEzqtr2jtLXjqN+duKPAcExFVV7E/RijxNBpzzdfTTqDnuO5uqbeyO4IKYui2YfFfBnNVuE+rsq6bWF+8rNTVHwfV8ttMd8UWXFXN7VC+77e7xi6Rgl+io0AjBHf1QZda9C1fdc46KyWUgohwjh/e7Hbhpm8VagCnp8X8SUOAAAAAPAApERCyKzgs1X2fpEtd0WcNQdhISFYYSTw9O+mziDQLYM9werWb4FRrKs08M1Xk86o73Y8U79Z3S2ELKsmjIv5Op4to22YZ0Vz54A3cLykRHnJd1G12hfLbZHfmvtNMNY12g/0fkfXVfo0l8wxSlxL63esUd/p+qaiHATVSAgZJeXb2W4TZpzLO0cqgOcEyr8BAAAAAO4Hb0WaN4tt/m6eLrbFnYN/XVMJPO31xH4zdXxHe/iDfFyurRHibcN0G6ZN04ZxgSS6VgSOGt7uo/xyGb3t7HSNvp74mnI4YBkcMxmnzWyd7+KqqDjnN64RRrGuscDVx12z5+vq037eZBrqm5NelNa7qMhvdVbnRb3aJOttto9L11JN40V0ebxYT/qTCgAAAABwRGoudnG12OQXy3yzLw+CaowRxti11Vcj+3RoTfqmY724+2zLUIZdazJwT0Z+xzVUhV1PRUopOW/jtFyuk/PL/WwRpXndtrDu95mQErUChWl91Rlxu7yfUuKYSs/XBoEZuJrytDsjDF05HXdOxx3L0gglBwvYy6rZhdlqm87XSZLDlrhn7kl/UgEAAAAAjkhVt/NNfr7KdnGZl/xgAhMlRFNo19PfTJ2er7MnWtn6EHzHOJsG03FnNHBNU5Xop6hZStQ0bZbXq21yuYqX2yxMqoa3j3m44J60ragqvovKy3WW5nd0U6uM9Hx91DU6jmpo7IlfIQqjvmv0A6vfsTxbp/SOwGq5if/rfzufLaOHPzzwkCCoBgAAAAD4pYSUdSOitJ6t89kqj9K6ag4TrKpCXEsZBvrZ2O64GiUYP8ERTA/CsfXp0J8OvXHfdSydkJ8W+Uop21YUZX01PHm+SrZRUTcwr+w5qJo2zppNWC62RV42Bz/FGGsq7fvasGs4lqIq5Invd2aM2Jba9c1BYPuOcedwhM0++/0f5/NVVDcwde85g6AaAAAAAOCXqhux2hfv5ukPs2S+ycv6jsyqaymvJ/bZ2J72Ldd6EbupP8XUWdc3xn3ndNzp+pahq4zeaJxuW5Fk5XKTvp1tZ4uoqJrDXcbgCMVpc7HKlrtiF1VFdeMawRgzik2d9QOj39E19QiCFIKxyohjqqO+0wtsVb0jtZ7l9WKdrLbZJiyL8o7kPHgejuDzCgAAAADwlAkpy7pdbov3i3S+yfdJ3fDDid8qIx1XOxvbk74ZuJqhvejhW6pCHVMZBPbp2B90Hc8xDtZrtUIWZbOP8tkymq2iXVTcLqcHRyfOmtkq24RVWjTNzRFllCBTZ76jDjpG19PVY5hOhzFiFFumOuw5vcAyNIWSw87qqm6iuFht09kyTrLqsQ4VfGsQVAMAAAAA/OWElA0Xad5cFX5nBRft4QYdTSGeo466xuuJE3j6065pfSgY+67xatI5nfiTgWuZN1L3V5t+y6rZhvnlMn5/Ga62Sd1AZ/Vxi7N6ts7T/LDwGyHEKOl62qRvDgPdd9Qj2jOnq2zQdQZdx7I0RaEHV7eUSEq53qW//9Nis08f6RjBN3c0n1cAAAAAgCdItDLNm/W+nK2yxTbPS357VrWhsUHHGPfMSd/ybPXFtlJfhxGyTHXQtSYDdzryfNfQVEavVc8KIeuaR3Gx3CTvLveLdVJUDSSrj1TDRVbwbVTNVnlyV1CtMBJ42rhn+q5maDc+CU+cqrKub/Y6lufohq7crgCXEoVx8fZiuw1z6Kx+riCoBgAAAAD4y9VcLHfl28vkYpmt92V1VzbVMZVXY3s6sAJXNbTDXNaLpVBsaKzfMU/H/rBre46u3iwC50LkRbXZp+eX+8tVnGZNwyXs1zpGSd5cbrLZKrtcZ3cG1YyRrqcPOrquHkHh93UKI66t9XyzH9iOrVNyR3iVF81ml+3CIkyqqobO6mcIgmoAAAAAgL+ElKjmIs2by3X+fpFtojIvedveCPkURmxT6Qf6m7E96hqWzo6orvVb+9Bq7pknQ2889AZdxzRUfG0muhSy5m2alattcrmKFuskTIqDMwyOwlV/xDoso6yub47xIxjrKvVsZdg1BoFxdEE1JVhXqe/q05HXDyx21wVe1k0YF9swX+8zGFf2LLGffwkAAAAAALhFCJkVfBNWF8vsYpXld90rGxobBPrZyLnaTf3E9+4+PIyxa2sIu5swi5IyLeo4LdtWfGxKxxJxLuKkXKziP75bM4ZtU1HYi56dfoySvJmtsjirb9cZUIp9Rx33rGnfGgSGdmxB9RXH0n512tuF+R/fbhCqD37aNG2Kqs0+u1zGtqF0feNRDhJ8O/CsFAAAAADgL9EKsY+r2Tqbb/JtWFX1jWnGGCNCsGMq04E16ZtdX7cMhiGovgljpKnMtfRhzzkZ+73Asi1NYYfrtfKy3obZ+/n+chVFSVXVXEIR+JHgrSgqvouq2SqPszsKvynFHVcb9YyOq1k6Y/QowxNTVycDZ9RzPEfXNHYwN6FtRV3zXZTNlmGclteeGoFn4ig/tQAAAAAAj65p5WpXvJsnq12R5DXnB0E1VihxbeV0aA27hqFRCnnquxCMGMWBZ56M/MnA6we2pt0opRRC1A2P0nKxii+X8XKbxmkN056ORVm3+7he7crZKkvuCqoZIYGnjbqGpVN8uJHqaKgq7bjGILAGXdu1D2tSpJRSyCjKL+b7MC45RzBy75mB8m8AAAAAgK8jpSyqdhuWF6vsfJFFaX2wdBchpCnUd9RJz3w1snu+rjBypNHCA6AEO5Y2DOzp0IuSIsurLK+F+LCZTEokhCyrZh/l83X8drbXFOra6pGmNF+avOCrXbEOy31clze7qTG+GjrA+h19GBiGzo73EqEEU5UGvvndq25Z8SSthLgxU09KmWTVcpNswzzOSttQDAqB2PMB7yUAAAAAwNdphdzH1fkyezdPz1dZdlc3tWmwk4H5euK8GtmBq905ExhcIQRbuoJ842ToJVm13qa7q+VD7bXO6kYkWTVfxbqm2KbyauJr0Fh9DNKCz9b5Nizzkh+kZzHGpsY6rjYKjOERjii7reOZv/31OMvq97N9VXOEbvx5s6KWW7TZpdt9Roll6BCIPR/w/Q4AAAAA8BWElHUj1mH5fpGu92Wa84M0NSVY12jgaqcje9w3HVNRFUhT/wxKsa6xbseaDv3J0OsFtq6x6ydNCNk0bZJWi3VyuYzn6yRKSqihfcqEkA0XUVpfrvMwrduD1C1ClGDPVkeB0fV0x1JU5egDE8tQT4bueOD6nmHoykEtO+eiqJrFOvn+fBun5WMdJPgW4AEJAAAAAMBXEAJVjVjtyveLNEzqhh8OHWKM2IbS7+ivxs4wMGCH1heilPiuMRmKzT5I87qsmrxoPub6pJQ/FYGvku/f7zDClqFCo/qTxVtZNXwfV5frLE5rdOsBCCG442rjntlxNfPWcK9jZGhMVcxhz+4HdlE2ZdUcVIC3rbhcR+b3yqhvv552Hu9IwT2Db3kAAAAAgC8lpAyTarbKzpfpbJWnRXN7jK+u0mHXmPatUWC4tkrp0YcKD4NgrGvMd/TxwD0ZeYFvmcaNxmkpUcPbLK9X2+T78+1ik+QV5+1hNzt4IoqKr/flal+s92VW8IPrhFKiq7TraaOeYRmMHO2IsusIwQqjnmOcjv2ub1F6WKIihAyj4mIRbXZZnNV1037iN4EjA0E1AAAAAMCXEgJto/L7Wfxuns43+Z27qXWNTvrmydDqd3THYBRSqV8GY6QyYpts2LNPRn4/sF1bZ7fWa5VVsw2zHy52l6s4y5u6gRLwJyor+HydL7bFNioPrhSMkUKJqStdXx91DfN5dRe7tnY2DXqBzShB6HAMeJKVy3Wy2mX7qKxqCKqfCQiqAQAAAAC+SNWIJK/nm+LdZbqPq4aLg55ehRHfUcc982xsT3qGocEWra+DMVIZ7bjGZOieTTvjgWcYNxpTrwpo86LehdnlMvrT+81mn0kIq5+krGguN8U+rhouD7qpMcaupQy7xjAwAk/Xjn9E2XWmrg57Tj+wHFvX1MPnBW0ry6qZLaI/vF2FSfEoRwju3bN6LAQAAAAA8O2UFQ+T+nKdv5un8V0bd1WFdD3tZGC+mdijrqEd/+Clh0cZ8R1dIhQlZZLVy20SJiUSCP24XgshWdU8SorZMrRNVVfpdOAc5APBU5CVfL7Jo+SOFgmCse+o0745DIzA1Z5D5fc1hq70u3a3Y3mOUTdtVR92VjdNez7f67oy7FqnI+/xjhTcG/iuBwAAAAD4eVKiMKmvqr63cXVQt4kRIgQbGhv3rOnA8m1NUymGNPXXuzqTusp6HXsy9MYDt+ubyq0i8Lrm+6h4Pw/n62QblcVddfjgsRRVu9qXl+t8tsrDtLpdSYAx8hx13DNt85l0U1+nKNS11K5nDrq2bWro1h9PCBkl5XIdb8M8TqsGOquPH2SqAQAAAAC+hLzqpp5v8jCpDnNvGBGMTINNB+bJwHYsVVWeVUXrQ8IIqYwGvjkZuidrL82qomzqhn8MTYSQtWjDuBASTYfeZp8TgmHr79ORl3yxyWerbLbKsrK5lahGhGDfVsc90zSUxzjAb0tlRGFqxzOGPWcX5WiNpZT42kkQUiZZtdqlm30eJhUhWIGviyMHmWoAAAAAgJ+R5s18U5wvs3eXaZTWt4MEVaGDwHg9ts/G9jDQNRVusX4RSrFlKD3ffDXpnIw819ZVhd7a+tvmRXW5in//p8Vqm0Bf9dORl3yxLfZJzQ9XUyOMkGMp464x6ZvjnmE900chGCHP0V+fdHsdmxFya0+9bNu2LJvlJpkto7yoH+cowf2Bb3wAAAAAgJ+R5M3lOns/T99eZnF6Rze1ppBR13g9cc7G9iAwNMg7/TKUYEtnPd84G/un447nGJrKDqa+NVwUZTNfRb//42KxSYS8vQgZPI6roDqMa3FH5TdyLWXSN6d98/nN/b7Oc4w3J0E/sBR2GFRL+WFc2WqbXizCNIeg+uhBUA0AAAAA8ElSIiHkLqq+nyWbsCzrw8XIV63UHVc7GVrTvmlozzZIeHiKQn3PHA/c1yfBqO8eVNRfTQLP8nq9Sy/m4Z/fbfYRzFJ+ZHUjwqRe7YrLdbZPqttBNUbYs9VJ37RNBePn1k19na7Rrm90fbPrW3dVuUveijDOl5skSsui4m0LD4WOGHzvAwAAAAB8kpCSt3IbVT/Mkk1UVrdGChGMTZ11Xe1kYI17hv68lgM9LoVR3zXGA/d1GKRZtd1nZcU/jpKWUratzIpKSHGxiP70bksIDjzjcY/5haubNkyq9b6cr/MoqW8+gEIYf+imnvQtx1SecUSNENJVpqus1zH7Xbtp2iStrv9UStRyEcbFapvs4zIvOTEwpfDtcawgqAYAAAAA+KQwqdf74ofL5N3ik2u0hl3jbOKMe6bvaCqDMsB7QzBWGHEtbTp0wzhfrmPO27xoWvFTrMavisDXsapQz9FPxx1GMYW564+kqNrFtljvy6zkNRfoZke1oTHbUIZdc9o3LfMZjii7bdh3/4d/f4oRXm0TKdH17WJCyqxodlGx3mXbfaZQ+5nt635R4HsfAAAAAOAOUiIh5T6ufpgl7xfp5TpP88OgmhCsq3TUNV6N7EFguJbKIKi+PxgjhWLbVEd9dzryBz3HsXVKbwTMbSvKmq+2yR/frhfrJCuauhGf+oXgm5JS5iVfbotNWOYl5+1h8behsa6vDwN92H22I8oO9AP7b/56Mhl6jNGDoQBCyKJs9nGx2qbrfVbWsBbuiMH3PgAAAADAHRrepnmz2hXfz5J9XN1+AaXYs9Vx3zwb2ydD04Ru6m9DYcRztFHf+dVZbzrydV05CE6wRE0t0rx+N9//079drnfpYx3qS9a2sqzaMKnmm3wXV3d2CLu2cjIwA1c3NMboiwhDTF0ZdMxBYA16jmmohz+Wsq74fBW9v9zDDPCjBt/+AAAAAAB3KOt2n9SLbfF+nobJHfe7lOCOo056xsnAHAaGrkHp5jfBGLGo2g+sNyfdOClni7CqeN201ytpG97KAs0Wka4xx1QnA4dgTKAI/AHxVmQF38fVYluESXW9RB8hhDHCCHuWOu1bvqO+nDpnXaW6SvtdazL0RCuStLz+Uyll3fDVJrnwjCQbtkKSZz287Rl7EY+IAAAAAAC+Vpw2by+Ti1W2Csu8vKMyU2FkEOivxnbH1XSVQR/vN4IRwhgZmtLv2pOhdzLyO755WAQuRMPbbZi9m+0vV/E2LMoKimkfVFm3q32x2Ba7uMqKgxn5iFFiaDTwtHHftF9GN/V1/Y79218Ne4Elb31J8KtxZZtkF+VpXh+eOHAkIKgGAAAAALihFbJuxDau3l6mV0OMy/rG0G+MkaoQx1TGPfPVyPYdTVUI5EW/HYyQrrKuZ04G7qtJp9+xFUavZ/SkkLxt46ScL+PZMp4tY9j9+8Culmntk/qqrV3eHFGmKdS11K6nj3qGZb64UtnAN//qrN/r2KpCCT3orBZZXu2ifLPPw7iob+0XAEfhxX2mAQAAAAA+r6h4lNSzZfZ2luziSqLD7lBKcN/XX43t76bu6cB+ISOXHhej2NRZP7B+ddZL82q+jnkrOecfYzcsUcPbvKzfzXa6xnSVDLrWox7yC3L1LkiEKMGGynKFV40Q1+Jq21SmA3PY1buupr6Y2u+PTF3tBXYvsLsdO0nKLL85owGjpmmXm+RyFdumat2x1Bo8dfB3AAAAAADADXnB59v8cp1fboo4a26m3BAhWFXoIDDeTJxJ3ww87ZEO82UhBKsE+45+OvY3++yHix1vRZzc6KxuW1FWfL5OCMGvxv7ZtKMqTIF57A+CUmwZrOtp04GpKiTOm6puGy54K9tWOBabDsyer7/A2m+EkK4xH+NexxoEdstFXlRSIowxo4QxoqrMtXUpUVW1tyamg+MAQTUAAAAAwA1J3rybp/NtnhZ1zW8VfjPiGMrVGi0bckoPS1NZxzUnA+/XZz0pZZpXQvz0Bkkpr4pp17vsYhGNes6o7wSe8YgH/EJgjCyDvRrZls76gbHcFqtdsQnLbVRFaZ3kjWup04HlWi/0emEU6xoNPGM6dNOs2u5TjBEh2LY13zWHPWc69N6cdqcjz9Rf6Ck6dhBUAwAAAAB80HBRN2K9L99dpqtdUd5KHGGEbUPpd/RJz5z0TMuAW6kHxRixTHXQtb877SZZOV/HuZQt/3G2k0QSobpus7ze7LPFJnUdHYLqh6EpVFOoqTPf0QYdY7UvlttiuS22UbmNyunAnA4sx7q1U+plIAQrGHc8Yzryt2G+j3JNY5ah9jr2oGuPB+544A66TuAZqgJfKUcJ3jYAAAAAgA+ygi93xQ+X6fezZBdXQh6WYmKCur72ZupMh1avo6sKlBY/KIIydcnVAAAgAElEQVSxQlHH1c9Ogk2YvZvtkZRpVn14ozBCGBGCCcFVxZOsbBqYAf6gFEZcS9FU2vW006GVZE2YVLu4ejWyTwaW/vK6qT/CGPmu8WoSRHHBuTgZ+ZOhO+o5Xd80dKZriqYqqkIZhXmHRwmCagAAAACAD3gr85KXFeetELeaGxVGTJ0NA+P12O77ugnzyR4cxghjZBrKsGtPh97pxBdSZnmDkEQIqRrTVOa7Ztc3B13btXXI+z0wSjBVqaZSZCkdR625SIsmyRrfUT37haapP3JMbdxHRdGzTPV05I+HXtc3XAuGMjwH8EUDAAAAAPDB1YjpwNOmAwtjvNoV7fUJxobS87XToXU2sV9sIetToCqUEDwduv/u16O6bherWAhECPYdoxdYb06C1yfd05E3Hri2CW/To6EUa5hSii1DUSjUdCBDZ4wRQxv+6lXX0BVdY6ryclP3zwwE1QAAAAAAH6gKcS1lOrB4K21D0VWa5E1R8poL3grPVk4G1rhn9nzd0OBu+NFQggmhgWe+OQn2YT5fxwQTTaGTkTcduq/GnenI6/qm5+iPfaQvGsaYUkQp1WD2FkIIIUYJo8TQIP56huBNBQAAAAD4QFNp4Gq6RoeBMe6ag8C4WKazVR6ldZw3gae/mTqDwLANhRBofXxMGCHLUMd977vTMs0qQ1ddWz+bdk4nnqkpusZUaE4FADwUCKoBAAAAAD646ghVGLEMhVFs6tS3Fd9Rt1EVJvWvTpzXE6fjagxWHz8BmsoIJq/GHkLS0BTL1AZdqx9YGCMIp++dlKiVsqrbsmnzqi3rtmmlREihWFOIqTFDpZpCGCEYzv5XkhJxIapGFHWbV23VtE0rMUIKxbpKTY3qCtVUSjGGc/tkYXlrrCUAAAAAwEsmEUISNVzUvA3jeheVu7gO0/q7qfN64ugqhaHfT4GUSEpZNW1Vc0IwJURlRIEm1W+jFbLmYhNX67i83JXrqEpKLoR0DNZ11ElgDDy952qGSgnEfl9DSiSkLOp2G1fLqLzcldukSgpOCHZ01ve0SaD3Xb3naiojFApknioIqgEAAADwPEmJWiHjokmKJq/avOJlI6qmbVspkWSUKIzoCjU0amnM0ZljKAo7LOsuKl6UPCt4XrZdXwtcDUPMAF4YLuQqLFdRtY7LTVyto2qXNUXVCilNjXqm0ve0vqv1XH3gaUNP1+Cp0xerGrGMyvW1cxvmTV61BGNDox3rp3M79LSBrzOIq58kKP8GAAAAwPMkpKy5mO+Lt6tstisWYblNqjBtiopLKQ2NOabSddShr08D41XPejO0HYIxuREzaypVGLFNVUjJKIZWavAC8Vb+sMr++X0UZk1SNFUjmlYIgRBCTSuyst0m9YVe+Jb6m4ntGoqqwHXyRSRCRdP+cZ7+YZaEWZ2WvOKC/3huK95mJV/HlWMovqX8+1dex1bZC971/ZRBUA0AAACA5yareFry801+vs7n+2KxLzZJtUvqpOBZyaumlVKqCjU1tjLYYlfMNvm7VfaneXLSM097pm8qls6uekMJxoRiBPex4EWSEmUV32f1MiyXYZlVbVm3Qv5U6ipa1Iq2aUXTipoLz2TruKIEmxqFWuXPE0JmVbuOqlVYrqIyq3jViNvntuaiaWXVtH1XW8dVYKuWxqBY5qmBoBoAAAAAz02UNxe7/D//8+r/+tdtWfGybhsuOBdCSiHR1U1r28qy4mGC59tcYURTqaGxv3nt/y+/G/xqZOsqJTA9Grx4Eskoby53xVWjb9OKVhy2jn6YYdYIKZttWs/3hcqIpugQVH9eK+Q+ref7YptUadk0XH763LZCyE1SzXcFJdhUKbSgPDUQVAMAAADg+QizepfWvz+Pfn8e//48Wu6LthVCSCEOx8h8CK2R5C26mrub5PwPLGYE75L6t6fuwNMDW4VZxkdKSMRbkVdtVvG04HnFay6ERBojhkZtnVk6szSmUIzhPf40KdE2qd+u8l3aNK0Qt6K+j4SUDUdR1rxf54ZK+66mQpzxWVzIVVS+X+dR3jRctp8edCWEbJDYp80Pq9xQ6dDToWf9qYEPOwAAAACej3Vc/ess+S//svm//21T1W3dtAihz09lvRoiLYTkWLxbpot9sY4qLuTvTpFnKpCvPkYSISFk3YhtUs/DYrYplmEZF40Q0jWVnquddM1xRyc+JhpjGMEOrk8RUm7i6vtFFuU1bz93IUmJuJRx0bxb576l/ncTgaBr4rN4K5dR9W6dRznnn35agRASEolW7tNaCNR3NSElfGSfGgiqAQAAAPAcXK3P/cNl8l/+Zf3neZIV/HZ2+vOkRA2XreDn6+z/+TdMEAoc1TdVU4PY4JhIiepWvFvlF9t8n9a7tA6zOs552bRSypqLsmnTkq+ismOrk8A465vQpPpJEpV1mxRNzcWXvLzhMi15Ud8uZAaHhJRF1aYlb77s3NZcJEVTNu23PjDwF4CgGgAAAADPQV61m6T614v4//xvy4bLtv2i+9QDUsq2RZfbfBWVts7+auIwSgyVQsR1RISUdSP+OE/+6/f74sfBWh/nP1VcxAVfhKWmUFNjvz1xu86P25Uf+8ifIIlQxUVWcfFlD6i4EHklr2YBfutjO3ZSyoq3+Ref26uG9vpLXw4eFBTkAwAAAODoSYTerbP/9E/LP8ziuhF/WUT9kRCy4eKHVfZ//H/LP17Gn+l1BE/NVQ/w98t0HVd51VZNy4W4XrMgJRJStkLWjSgqvkuqd+tsGZX8l31mwAG4Zr7EX/DVAk8rnibIVAMAAADguEmJJJLvVtl/+qfFNqpaIX/hHb2USAr5bpXtstq31f/x1wHFkMg8DkLKTVx+v0y3cVU1bfuJjKmUqBWyasQurd+uUpVi31QVKPMHTx58DT1NkKkGAAAAwHGLi+btOluEZVbwL+z8/HkScS6Ksl1H1Q/LbJc19/NrwTcmJdqlzfmmiPJGCPmZ5ysSyVbKtOSzbbFJ6lZApvoOGCGVEUtjCv2iqIERYqhUUwiMVP9ZGGNdoYZKGfmic6tQYmpMVaAb5SmCTDUAAAAAjltcNG9X+SIs87LlXNxX4SlvZVG166j88zJjlHRt9X5+L/iWJJJhVs+2Rc3v2Kh845USSSmzkjdcTIPq86OtXy6MdIXaBhNSfskTK0axrTNDhcFvPw9jpKvU1pkQzZdMH1MYcQymKZATfYrgXQEAAADAcdsm9e/Po9k2r3n7mT26X0sI2bZiGZb//C5chuV9/VrwrbUC1a34wk54IVHTSpj99CkE456rfje0OpbKCCafDpUxRoxgz1TO+tbQ174ws/2SMUqGvnY2sFxTYZ/tLyEYMYI7lvJmYPVdDaoAniD4uAMAAADguO3T+t9m8WJX8EbcY1AtpWxbsQ7Lf7mI1hEE1Ufiag5ZK79wnpOUshUShpR9Csao52ivB1bHVhSGyaejaoIxo9gzlVd9c+BpDBa8/xxG8MDTznqmbymM4s+EyoRgheHAVt8Mra6jfebRBngsEFQDAAAA4LiVNd/GZVo0v3RA2V3yiq+jMqv4vf9m8E1gRAlW2Jf29GKCGcUQAX4KRtizlElgdF3NMRRNobeXj2GMKMEqI7ahBI467ugdW6UQ+f0cSnDHVkeBHjiqbSiaQj51bjVGHUPpOtokMHxLgaGJTxD0VAMAAADguFV1G6Z1XvFvUcJb1i0Xsqi+oOURPA0qI5ZG8woJ8TN90hghhWBDo9oXB+EvDcbI1hmjeOjpa68KszohuOKC8w9FIRgjRrHGiKUz31KHvt73dMdQ4Gz+LEqwYyhCooGn79MmzEhWNhUX/MeR9QQjxojKiGsw31IHvt53NV2FJ0BPEQTVAAAAADhuQiLeym80vFlIxNv7rCoH3xTBuGMrJz1jua82iRCfLgPHGBGMbYMNfb3nQrny5zBK3gwtU6ObuFrH1Sau9mlT1K2Q0lCpZyl9V+u7Ws/VBp5uQtT3xTBChkL/aux4pvLx3EZZk9ctxthUiW+rH8/t0NMVKAB4qiCoBgAAAMBxk1K2rfxGo6akkK1EX9igCx4dRqhra2d9q6pFmDf80xPLrnqAHUM56Zp9V2MQrXwaI3jk64GtbuKqG5Uzg1l6lRZcSOnorOtq08AY+HrP1UyVQs7/q6iMnHSNrqNu4ioIS1tn26ROCo4xcgw2cLVp1xh4etfRrurDH/t4wd0gqAYAAAAAAM8ExrjnagijbVLPwxIhJOUd6WqC8VWramCrbwbWqKMzBpOGPodgrFDSsVVDpT1XK2vRtEJKpDCsK9TUqKEyXSEEQ8Pv18EYEYQ1hXQdzdTo0NfLpm24xBgplOgqsTSmq1SFDoWnDYJqAAAAABw7jDHC6BuMKUMIY4Sufjs4Bhgj11QURiaBsUvrtOBZxetGNO2HtdUEY4VhlVFTo46uTAJj3DE8U/nMXGuAfuydtimzddZD2mMfzrNyFT8rBnEMCM2OFbxzAAAAADhuhCDGCCG4/Zm5VH/ZL8eEYgo5ouOBEdIY+c3Y8U11k1SbuFpH5T6t87oVQpoa8yxl4Gk9V+s5Wt/VbP320GUAAPgKEFQDAAAA4LgplFg6SwteivsfAK4womtUUaA2+Ghc5VQHnuZbSjdRA1v1TLZN6qzkQkpLZx1LHXb0vqv1HM1Q79gRBQAAXwWCagAAAAAcN12jXVfLSl5W7b3XgBsa63mapSv3+2vBt3a1ObnrqLZOxx29atpWSCkRo1hlRFepxoh619ZlAAD4WhBUAwAAAOC4uYbyqm9lJd8nddve26RujBHGuGOpr4e2b6n38jvBg7nKVzNKTY0+9rEAAJ45qGUCAAAAwHELHO23p960a6qM3uO4KYwxpWTY0f/mzB/6+n39WgAAAM8MBNUAAAAAOG6+pfzV2Jl0DF2jlN5bUE0p0VQ68PTfTJyeC+OOAQAA3A3KvwEAAABw3FxDURkZ+rplsLpp66a9h8ZqjBjDhkZ7nvbd0LI0uGUCAABwN8hUAwAAAOC4KYw4Ojvpmf/9a38cGAT/0tFTV93UA1f/mzP/9cByDUVToC8XAADA3eCxKwAAAACOG0YIYfxmYP1v/2GEJXq7TH/huDKMMSH49dD6X//D8DdjhxKYDw0AAOCTIKgGAAAAwHPgW+qvRvjixD3fZMt9uQpLKb86sv4w8dtRh77x1yfub8ZOz9Vg5xIAAIDPgKAaAAAAAM+BqVJNIX994paN+H//uN0m9YfFxF8DY0wpngTm//Sb7u/O/KGv68r9jT4DAADwHEFQDQAAAIDngBBMEB53DCmRlFKh+HyTz7eFEFJI+Zms9VV2GmNECO65+mnf/N2Z/x/fdE57pq5Qen87ugAAADxLEFQDAAAA4PnoOZpvKZZOx4Hxn/95tUvrhgvOhRCf6bLGhCBKicLo65H1P/9u+NdT56xvGSpE1AAAAH4eBNUAAAAAeD4YxYzSoacrlLSt7LvaNql2SRXlTZLzom6rpm24kAgplKgKMVVqG8w1lY6tdR3tV2P7t6fuwNMtjUFEDQAA4EvgXzIbEwAAAADgCRJSCoHKps1K/qdF+v0yOd8Us12xiat9WhdlIyQydOaZas9VJ4Fx0jXfDKxfj52OpeoKoRTDwG8AAABfCIJqAAAAADxPQkreynVcbeIqzOowb7KS5xWvGyGlVBWqq9TWmWcqvqV2HbXv6rpCCCSoAQAAfA0IqgEAAADwnF3NLZNX/4IkQujqHwhf/ePDwqwP48oe7TABAAAcKwiqAQAAAAAAAACAvxB57AMAAAAAAAAAAACOFUz/BgAAAAD4JCGREKgom7ysbVO1DOWxjwgAAMDTAkE1AOCRSXnV7fjTCln8o+uvEUIIIa5eQ370KAcMrsgfffwvV+/IkTal3v7j3P4cgifrqmsa4Z8apO+FkFIIyVvZcLmLi12Uj3u2qbOj/ZgDAAD4JiCoBgA8sjzPoyiq67pt26uQxnEc13UVRWGMIYSEEGEYbrfbi4uL3W5nWVan0zk7O+v1epRSuLd9LHEcx3HMOW/bFiFEKXVd17IsRVEopY99dF+tqqowDIui+PjsxrIsz/M0TSOEwMfsyboKp7OCJ3mjKsTUmcKIwn7pEzchUSvkLsyXm2S9y9a7LMnKrKj/47+bKgo1dcXQ4A4KAADAB/BXAgDgcz6m74QQV//lY/ruvhLFSZK8e/cuSZKqqq6Cmel0enp6atv2VVDdtu1yufzDH/7wj//4j+/fvw+C4OzsTFEU0zQNw1AUKMV8HPv9/u3bt2VZ1nWNEFIU5dWrV6PRyLKsYwyqi6I4Pz/fbrec86tP+2AwePPmDcZY1/XHPjrwSUJI3oowqWbr3DJY4GqmzkyDUYwJ+UtKDYSUopVV05a1OF9Ef/xh9cPF7oeLXVk1nLemoY767iAwIagGAADwEfyVAAD4nKZpsizLsizP86uEpKZpuq7btu04zr2k796/f//3f//35+fn+/3+Kpj527/927/7u797/fq1aZpCiDzP/+Ef/uF////ZO7PnNq4r//e+YgcaCzdx1y5rtbzJrjixJ6mZqUlNzdPU5DH/2LylZjJ++CWxYzseWYsl0ZZkUpQoijsJYgca6H3v38MZITBISdRirffzJAKN7nv7QmR/7znne/761+3tbVmWWZYdHh4Ow9DzvMnJyVQq9fRjQDwBCwsLX3zxhSzLnU4HwzBRFD/44IO333573759r6IKrdfrX3/99e3bt7tf9aNHj37yyScTExP5fB72dxAvIbrlNTvW3fX2rSWZY8l0jE0nuEyCS0SZRJQVefJx1a+mO3VZ3yy114qtakNttLSOanY0y/eCMAxLlc7dlSpJ5FJx/meaEQKBQCBeOdBTAgKBeBiWZdXuAwHJWCyWTqcHBgaeVUCyUqlcuXJlYWGhXC6DmMFxfGRkJJlMgni2bfv27dt/+9vfNE0zTRPDsHK5nMvlcrlcoVBAovpFsb6+fv78+XK5XK/XMQyLx+ORSGRwcDCTyUiS9KJH99jIsvz9999fuHBBURTXdTEMUxRlamoKpoNE9UuLYXmVhrm0qdxcbNAUkYyy+TSfzwgDklDICJkkR+A4RRIkudcdQNNyqw3tznJ1Znazo5qO43YL7XEcr9SVe2tMQYoGwylUcY9AIBAIAD0lIBCIh9Fut+fm5ubm5u7cuaOqKoZhIyMj09PTZ8+ezefzzyHLF/LMaZqmabqbcA5Psuh5FoFAGJZXbhht1Q4CzHL8lmJbjl+Tra2KnoyxE0PRscFoLsWn4uweT0iQOMtQJIF7nu/7QY91HRaGYVsxN0tyQ9Z102MZkqGRXSICgUAgkKhGIBAPRdO0tbW169evX716VZZlDMMOHjzouu6+ffu6VdY/KziOUxSVSqWGh4dZlu10OgRBZDKZZDIpiiKKHyIQbyxBEPpBqGhOuWHIquOHoecFjhuohothWIU1RI4OsZDnyIhA711U0yQp8AzL0jixy7adZth+ENaaWqNtpOM8QzPPckoIBAKBeDVBz6MIBOKlBmyi3n77bZIkl5aWqtWqKIpDQ0Pvv//+wYMHY7HYix4gAoF4Mdiur2hOpWWWGmZbdYIg7H3X8QLc9jw/eFzvdo6lpaSYTUXSSdHzfFW1eqPVQRC6rrdd7SwsVw9NSrEIEtUIBAKBQKIagUA8Nb7vQy10GIZgtkuSJEEQ4Bnuui50XSIIgrrPYz3k0jQ9NTUlimIul6tUKrFYLJ/PHz58eGBgYNdINViIe57nui6MjSTJB106DEPf9yHqDuMnCALS2sELDXp9wTA4juu6CcO7MLsgCOBTcIkn6MAEXbh934d7BWMmSRL6iu1qYdwdOQwbw7Dekbuu23VTZ1mWZdmdfb89z4PLhWHYvT9P3z4KBuY4TvfOPHLdn/n0HcdxHOdB04dLwNcDPsgwTK+//SMn2F19GEN3gjt7vD3NOBEPx7L9Rtuqtcx629JMr7fHOIZhOIbhGMbQpMBRNPkYSdoMTcaiXDopZtMRw3A03cb8v585DEPPDyoNdXmjns+Iw/kEQaBSFAQCgXjTQaIagUA8LZZl6boOuggUlCiKHMf5vm9ZVr1e73Q6lmVRFJVOp+PxeDQafdw+WBRFRSKR8fHxfD5P03Q0GhUE4UEHg4zpdDqyLKuqqqpqJBKJx+OJRCKRSBAE0dsMDNzFLcsCbUmSJMMwkUiEJEnXddvtdrlcNgwjDENJkiAFHcdxEKXQPVtVVcMweJ4XBCGVSkFn48fNS/c8z7IsRVE6nY6maYZhiKIoimIqlUokEt2W3X0f0XXdtm1QgwRBgCs7hmGu6zYajXK57LpuEARDQ0PDw8O9XdBA9LZarXa7rWma7/vxeDwejyeTSZ7nn0xXdxutwS1tNBrtdtu27b2sO2wBdDqdTqej67ppmoIgRCIRuJ+7Th986WHLA67L87woivBWvV4vl8ug6oeHh4eGhrrTD4LAtm1d1+v1uqIoOI7zPJ/NZsGHby+N4mA/QpZl+IJZlgV3L5FIwDen9+498TgRj8S0vUrTbLYty/b6ip8xDBM4Khlj82m+kBFE/jH+P+I4RhF4TGSGcnFFMWsNJQiwvsrqjmqWah1ZMQ3b5RiKppCoRiAQiDcaJKoRCMTTUqlUVldXLcuybRvCucPDw8lkstFogGZotVqmadI0LUnS4ODgyMhIJpPZe0V0EASNRmNzc7PbEllVVfAti0ajLPv3UknHcSzLkmW5Xq/XarVqtdrpdBRFicViyWQym81KkpTNZjOZDDifwUc2NzeLxSLoaoZhkskkdCculUrb29tbW1uwZXDkyJFMJoPjuO/7rVYLVGupVFIURdM0EIGZTCaTyaRSqWQyGY/HRVF8iEAFJalpmqqq7XZbvo+qqrquQ9OyXC6XzWbz+Xwmk+E4jmH+nmuq6/ry8nK9XrcsC8MwmqZzudzo6KhhGKVSqVgsbm5uuq4bhuE777wzODgIuQOwA1Kv1+EWNRoNRVF8308mk6lUKnufSCTCMMxjqTvQ0rVaDcfxtbW1SqXSarUsy+qu+759+9LpdHfdu9NXFKXdbnfvAEw/Go32Tj+dTvf1JNc0bWVlBaaP4zhN0/l8fnR0VNO0Uqm0tbW1tbXleV4Yhu+9997AwABJkkEQaJrW6XTK5XKlUqlUKu12G8dxURQHBwdLpZIsy/CRB83RsizTNOGLXavVQJYbhtF79yRJSiQSDMNAOFrTtOXl5Uaj0R1noVDYt2/fQ8a593v+xgKtpFXDrTTNZsd2HL8v9xvDsIhA51K8lORSMZamHuObjOMYjmNRkR3MJSp1laYpPwh9/+8WEmGIGYbTbBvNttFWrGSMoymUBI5AIBBvNEhUIxCIp+XWrVufffZZs9lst9s8zyeTyffee294ePi77767ffs29MGC9G+O46amps6dO3fixInp6eloNLqX89u2PTMz89VXXzWbTUVRMAzL5/O//OUvz5w5MzEx0SuqVVXd2tq6fv36pUuXQO7atu26Lk3TDMOkUilJkn7xi1+cO3cunU4nEgkMwzRNu3z58rfffttsNm3bFkVxamrqk08+cV33888/X1paUhTF8zwcxw3DOHHiRBiG7Xb7u+++O3/+fKVSqdfrkAEO2b8sy8Zisf379x8+fPj06dNTU1MPiVrrul6tVhcWFubm5tbW1rq7BpCVTVEUTdOxWCyXy3300UdnzpwZHh7OZDLdj9dqta+++urmzZvNZjMMQ1EUT548+cknn6ysrHz++eflcllVVdCHkUjk/fffxzAsDMNqtbqysnL58uXr16/DdgAIb4ZhRFFMp9MHDx78+OOPDxw4kE6ne+/tI/E8b3t7e2ZmBjpXa5oG+xQQm52cnPzwww+PHz/eXXdN06rV6p07d+bm5tbX1zc3N23bdhwHcqopimIYJhaLFQqFDz/88PTp08PDw+l0unu5arX617/+dXZ2ttlswhxPnTr16aefLiwsfP7559VqFczqMQxLJBLvvvsuJBesrq7eunXr2rVr8/PzhmHAHg3LsvF43LKslZWVbpPqXWm1Wpubm5cvX758+bKiKNB8C/ZiOI5Lp9Ojo6Mff/zxiRMnYPugO865ublmswkC/vTp05988slDxrn3e/7G4vuhYXmtjl2qG7Ji+zsUNYZhMZEZzAqJKMvQBLGb5djDEQWmkItlShGBZzw/+KmoDl3PN02n3tLLNYWlyYiARDUCgUC80SBRjUAgnpZKpfLjjz+CwuQ4LpFIuK47ODh4+fLlhYUFqJLF7qcHV6vVbrwOArmPPL/v+8Vi8ebNmxD8xDBsZGRkbGxsYmJieHgYjnFd1zTN9fX1mZmZy5cvX7x4sdFomKbZK1Gi0WgsFoNK3ePHj0ejUYIgHMfZ2Ni4ceNGpVLRdV0UxVqtJgiC4zjnz59fW1sDfcuybL1eD4KgWq3eunXr4sWLf/vb32RZBpHfC8/z1WpV13WIzT4k0b3T6SwtLf3www+XL19eW1srFot9cg5M2lKpFMS6aZpOJBLd7GIIgV6/fr1SqXieJ4qiYRiCICwtLf3v//5vq9XyPA/EnizLOI5blmUYxvz8/LVr1y5evHjz5s1u5B+AvPpqtcowjO/7x48fT6fTew9Wg6g2DKNYLNbrdUhpxu6ve6VSgXWHDucEQbTb7e7019fXt7a2+mticZzjuO4+AsMwkL0P01dVdWlpCaYfBIEgCJZliaJ4+/bt8+fPt9ttz/OgSrndbmMYput6u92enZ399ttvZ2Zm7t69C3XOGIaRJAlZ/ZCkvautvW3bpmkuLi7OzMxcvHjx8uXLpmnatt09gCTJaDS6trYGeygnTpzgeR7H8V3HKQjC/Pz8znHuWj2O2Inj+m3VabTtettSdLdPVJMEThJ4MsoMSGJcpKnHKajuwrF0JimmE0I8yjmu19utGsMw3w8s22vKeqmmZFKiFIpo3RAIBOJNBolqBALxzADjJVmWr127xnFcq9WCKJYUnGgAACAASURBVGj3XQzD6vX6lStXotHo6dOnIQ37mVwatNzMzMwf//jHxcXFZrPZ9X/qYpqm53mXL1+uVqsURU1MTPTFkH3fNwxjbW3t888/B/0MhdYsy0JVME3TKysr//3f/72wsFCv13sVaRfXdUulUjweh1zihzQeazabc3Nzs7Ozi4uLnU5n55FwP1ut1tWrVzVNy+VyY2Nju4a+Pc/TNG1hYQHSmyGjG8MwQRDS6XQ0GiVJUpbljY2Nv/3tb3/5y19arZZhGH1X9H1f07SlpSXLsjRNKxQKMOU9yjzP80qlUr1eh/vct+6NRuPKlSuxWOzUqVPpdJqm6Uaj8eOPP87NzS0uLiqKsjNCC9NvNptXrlwxDAOypnedvuu6mqbdvn1bURRIIO+dPtQ5VyqVu3fvXrhwAXZDuooau19ljd13INt1dp1Op1gsXrx48Y9//CNsmvTtgARBoOv61tbWV1991W63k8lkLpfr+3rDOO/cufOQcSJRvRdM26+2zErTaCuOaft9q0ZThMBS6QQ3KIkR4Ql/w7AMSRJsOsFLKdG0HEW1MOwndmWu5zfbeqnWGR9JByFGYBhaOgQCgXhjQaIagUA8S8BXuVKpQFg1lUr5vg8eVJZlhWFoGMbm5ubm5maj0RgcHNxjBngfO5VPq9WanZ39/vvv5+fnq9UqhmGRSCSRSEBpLhQtq6qqadrm5qamae+8806xWOzNpsbum2a3223LsqC0FaQO5I3ncjnf90ul0tzcXKlUMk0TwzCSJPft2zc0NATO1SBWoQwYfLkeEumlaVoURZIkDcNgGCaTyQiCIIoi2ErX6/VmsxkEgWVZxWKRJMn19fVarSZJEriR9QIucY1GQ1VV8LKOx+NQ4w2FvjiOb2xsXL58+ebNm/fu3YORQ+E3ROzb7baiKKqqwhRyudzq6moymZQkqbeQ+yGAqoR1FwQBlGp33UFwbmxsNJtNWHeYPkEQhmGwLCtJkiiKgiCA43q9Xm+1WkEQmKZZLBYZhllfX6/X65lM5kHTr9frqqriOM6ybDKZ7J1+GIYbGxtXrlyZn5/f3NzEMAzHcYZheJ5PpVIcx7muaxiGLMtQoL6Tcrl87dq169ev37lzx/M8DMMSiQTcPZ7noSYc6sNXVlY4jlteXt63b58kSTvHWavVwCBt5ziRRdkjCTEsDEMQ1Y22Zdie5/fvRgkclY5zUoLLJDmBe8LnHJLASYZMJYTRoZSm2+Wa8lOvMsz3g7ZiVupqR7FM2+MY4slC4ggEAoF4DUCiGoFAPGPAvlgUxYMHD46MjNi23Wg07ty5U6lUMAwLgsBxHEVRqtWqLMs8zz/9FcMwrFQq58+fn5mZ6VaoZrPZo0ePTk9PT05OLi0tzc3NLS8va5rmOI6qqqurq7Ozs2+99dbOAYCOheDh4ODg+Ph4LpdLJpMjIyOgu9rtNiSWQ9nzhx9++Nvf/ta2bUVRbty4sbGxceTIEbh0JBJ5iEzK5/Pvv/9+vV6/ceNGIpHYv3//vn37hoeHDcNQFOXSpUvff/89JM9DhHNra2ttba1r8b0TCIzHYrF0Oj0yMjIxMQGWaSMjI2EY3rlz509/+tPa2hoczLLsxMTE4cOHJycnOY6bm5tbWFhYXl6WZdk0zWq1Oj8/n0wmo9HoHkV177oPDw87jtNoNG7fvg0yFTS2qqrddS8UCufOnavVaj/88EM2m92/f//IyMjw8DAE2y9dunT9+nWYPizZ5ubmI6dvmmYsFkulUqOjo+Pj4zD94eHhIAgWFxe/+eYbUNQYhpEkKYriwMDAyZMnC4WCoiibm5tQxYD9dNcGwterq6t//vOfIWkcJjsyMnLkyJHJycl8Pj83N3f79u3l5eVKpWLbdrPZvHv37uDg4K5F6Q8Z517u8xtOGIa+H+qmW22arY6zU1FjGBYV6cGsIKW4RIQhyacKH6cT4sGJfKOl312p9uUxBEGoaFatoTY7uqrbBM4iUY1AIBBvLEhUIxCIZwxFUaBF33333ampKcdxIMYINckgUSCGCc2cnvJy0Lir0WgsLy8Xi0XHcaAwVZIkULajo6MURUGTrdXVVVD1tVptY2NjfHx8p6gmCIJl2YGBgVOnToHQTSaTgiAkEgmKono7V2P3FReO45lMplAoMAwDMn5sbCybzT7c3jwajY6Ojp4+fbper8disdHR0WQymUgkwL08FosxDNPtcmzbdqvVqtVqY2NjDzoh1AaPjY2dOnVqYmJiZGQEeo9JkqSq6vb29uLioqZp2P3+W6Ojo8ePHx8dHaVpGnp6gVW77/uqqhaLxWq1umuK+67Auk9MTLzzzjuw7mtra/V6vdFo7Fx3EP9jY2Nvv/12o9EAeQmR81arVa1WY7EYTdPQvLo7/Xq9Pj4+/pABMAwzPj5+6tSp8fHx3unrul6pVMCHDA6GI48dO3b27FlQ8nfu3CmVSq1Wq69wAFqXVSqVxcXFarUKy02S5ODg4PHjx8fGxjKZDHwrwBAeigjK5fL29vb+/fsfOc59+/ZFIhFBEAYGBvZ4q99kPC9QDbfetqpNU1Zt3++rw8dwDI+J9FBWTEZZhn5alRsRmMF8PJuOxqK8bjiW7XbfCsPQcXxVt8s1pVhp04OJJ46KIxAIBOJVB/0BQCAQzxiGYY4ePfrhhx+ePXt2amoqDMP5+fkbN27cu3evW2INQqXPSOzJcF1XUZRWqyXLMlS64jhOUZQkSYcOHYK6VkmSDhw4sLS0BB8Jw1BV1VqtBincfdA0HY/HDx069G//9m9Hjx4VRRGaS4G+7Q3bgt67e/ful19+eezYsUOHDp06dSoej/M8z3HcI62zIUn7vffem56ehutC42vHcTRNA0duyDTG7idXK4ryEJXLsmwqlTp16tTvfve7wcFBqNElSdKyrGazCbeo2y2Z47iRkZH9+/cLguD7/sjICKTQw6kgv6Ddbu9914Nl2aNHj3700Udnz56dnJwMw3Bubq5v3aFps2maQRDwPE9R1AcffLB//37wMJNlGaav6zqUskMmfO/04ccHDSCZTJ45c+Y//uM/wAYPpg8p/ZDf3p2OIAgnTpz41a9+deTIkUKh4Pt+NBqdmZlZX1/v8yqDJm1w9+AbS5IkRVEDAwNHjhwRBAH+rev6nTt34CPQxRr0+d7HyXHcHm/1m4zlBLWWVaoblabR0foj1TiGkwQeF5nBrBgTn4FfA8/RDE1KaTGbjtYwtVdUYxgWhoFlu1slOR5hUzEukxSf/ooIBAKBeBVBohqBQDxjoFh3aGioUCjk83kMw+r1ejKZFEURsq8xDAuCADpRPb2oBms0EDwgQXEch95RN2/e7FbtQgMn+AiUdrfb7V735i7QVevYsWMQpqYoCryjXNd1HCeTyUxMTEBtMERfi8UihmGdTqdUKkEC8/Dw8E6Tqp2AuPV9X9d1aBkNDZBbrVar1SqVStCuqTtmaDHdldk7SSaThw4dOnLkyPT0dDqd7sbJwT9MVdWuIoWg8erq6tWrV6HzdrfaHA6ArYrHSiUgSTKVSvWue7VahSC/ruuw7uA9BusO04ftFUhkgOnLsgwhX9d1u+IWiqu7S7wrqVTq0KFDhw8fnp6ehrQCeB0SznVd7xXkLMuOjIwcPHhwYGAgmUzCxzmO29kmGvIgOp1Ob4ZCEASbm5vXrl2D9H5N02q1GtiMYxgGof4HbQHAOGGZeseJ2Au249dls94yddNzvf7cb54lIwKdTfEDkvDEFmW9kAROEmQ2FZkezwZB2JT13nfDEHNdv9ZUN0vc9JhkOz5FEeTjt+9CIBAIxKsO+luOQCCeMRB1ZFm2q08oigLDsG4PJ2ga3OvA/MS4rgt1zr1BXSgh3tragg5MkEWs6//3QByGoWmaqqruGkiMx+PHjh07duxYMpnsKmqYBUmSQ0NDp06dCsMQor4YhtXr9Xa7vby8LAhCPp8fHx//9NNP3333XcgGf8jIgyDwPG99ff38+fOzs7Pz8/O6rtu2DX2qbdu2bbtXVTqOA52fH3RCSZLOnDlz4MABQRB6xSGI9u704WydTud///d/r169Cq9A6XL3GNC6Ox3CHw4kNnfLyGHdY7EYNKDG7q87LDpMf3V19fz583Nzc/Pz86ZpwgQhF91xnMeafjabffvttyH23jt90zRbrVafAxnkMgwODgqC8PBJgRe6qqq9gwmC4Nq1a7dv3+6+AkF4+BEywB+0A5LL5XYdJ2Iv2I5fk61m296pqDEMiwj0gCQMSEI+wz/DCudsOnLswIDcMe6uVfGf/sby/aApGzzXaXVM3XIjPE0SaE0RCATijQOJagQC8ezB79P9ETJme9sFPb2c7p4H7Kz6Tuh5nmEYvVekKCoej8M/YrGYKIq7BpO7WwB93aRgRvv27fvlL38ZjUYjkUipVKpUKqBFdV2HJkmdToeiKE3Tzp07d/DgQZqmHyScqtXqvXv3rl69eunSpeXl5c3NTVCMiUQik8lAYNYwjG6O+iO3Ibqm332dmcDqrFceg7jtFe0Az/NQZB6LxSKRCMdxj+VHTRBEt4803DGKovrWHbu/9OVy+d69e1euXLl8+fLKysrm5iYMJplMdqev63pXDO9l+olEAnITdk6/V43DFxK81h8ZKAbx3/txCLP3qmgA2k1jGBaLxcAVfNelf9A4EQ/H8wPT9httq1w3Gh3rwaJaTMZYln6WyjYisoO5eDYdSUR5y3Zt++97JUEYOq6vaFax0smloyMDcYZ+BuaLCAQCgXi1QKIagUA8D/pk9jM/+c5rcRyXSqV2VTUURWUyGUj33fWEfeKwl7Gxsfx9rl27du3aNUgMxjAsCAJVVU3TbLfbW1tbmUxm3759UDG761U2Njb+9Kc/QRswTdNc16VpmmGY4eHhqakpyIgul8u7Fn4/6D7AsHe9IX2vdPcOdj2VJEmSJMVisacJpcJFH7Toa2tr/+///b8ffvhhfn7eMIze6U9MTHSn/6AeV7teDhbukUcSBEGSJFxuL9/JvhsIP4qimEgkdj0+Go1KkpRIJBiG6Vqj9Y0TyenHxfUCWbErTbNUN5qdB0SqeWpQ4mORZ5D43YvAMRRFS6lIJiW2ZKNXVGP/V5vgbmzLsQiXjHPJGBLVCAQC8caBRDUCgXi1AV+xaDTaF3IcGRl59913U6kUz/N9AoYgCEmSCoVCoVDY9ZwP0f/QBXp4eBhOcuDAgVKptL29vba2trW1BXXXsiyD+XOlUhkYGNjpWAbBz1qtdvv27bW1NUVRIE94enr6xIkTk5OTo6Ojc3Nzc3NziqI0Go29341dh83zfDqd7vM55zju+PHjR44c4Xl+Z5p6JBKRJGliYuKR2dFPAORyQ9eu9fV1VVUhDnzgwAGY/tDQEEy/0+k0m809nra7an03gWXZaDTauwqQ3QD+dhCQD3roOy3LspFIRBTFPl194MCBM2fOgCndzo9ks9mRkZFEItEttN45VKSrHwvHDZptqyabquE6bhD8NGeBZUieJbMpfjArxsU9NYHbOySJsziejPGDuYTr+m3F7Eu1cT2/0dJKtU5Hley0T1MEgSqrEQgE4k0CiWoEAvEysvfkcJqmk8kk1D/3vj45Ofmv//qvExMTkiTtjF5CrJKiqK572R6BpN9kMpnL5Y4fP+553r179xYWFv7yl7/IsqxpGlREa5oGujqRSOyMZ/q+D/baS0tL0IQJXj9+/Pjvf//7wcFBSZJomq5Wq9vb2481vF0RRTGbzYqi2PfiO++889vf/hYi0n0fgWgqNOJ++gH00Tv9SqXSnf7Jkyd///vfFwoFWM1qtQomcE8Jz/OJRKJXVIOqbzQalUplaGiIZVnIhwff777vHnw8Go32fotwHH/rrbf+/d//XZKkVCq186Jg6E3T9Pr6+tNPAYFhmOMG9bZdly3D9na2p+ZZMpPg8ml+KBsRhWf8bEPgGEFiiRg3XEjIbWOb6AQBhmG9fdf8VkevNlRZtQzbixA0EtUIBALxRoFENQKBeMGQJMkwTG/lbaPRWFhYyGazo6OjD/f6wjCMpulYLJZKpTKZTDQaNQwDRNr29vbFixfr9frk5CTLshRFgUk4y7KCIIyPjw8NDT3WOCGMubi4eO3atTAMOY4bGBgYHBx0XTeRSEQiEZ7nIVc57OFBp4KYNhhKdw+zLKvVarmuWyqVbt++vbS0JMvyYw1yVyBSDXSLtE3TvHPnTjKZnJiYyOfzNE27rgtdoFiWzeVyY2NjP0eYGruvabvT777edRTb3Ny8ffv28vLyrmHex4XjuEQiAekMXcs0TdOuXLnied7+/fvT6bSiKPPz88VicWedOXTAgrsHDmQwheXl5W+++WZiYgIs4sMwbLVahmHA8ePj4+l0+ukHj8AwLAhC2/Xbql1tmg3ZctxdEr9FjiqkhXSc4zmSfnYWZb3EIvxwIblVbjMM5bq+5/39qxuEoWV7HdWqNbRGRqelCE0942g5AoFAIF5mkKhGIBAvGIIgQPR2X6nVarOzs2NjY0eOHNkZRO0DXMcymUw2m00kEl2Zury8rKrq0aNHT548Ca5RS0tLy8vLiURCkqTf/OY3TyCqXde9devWf/7nfxqGQVHU2bNn33vvPTA84ziO53lVVeFgsMLqMwzrOxV0luoV3rIsr66u+r5vWdbVq1fn5uYe0pZ570CCtyRJ2WwWunOHYajr+pUrV0ql0smTJyHNW9f1xcVFTdPi8fjx48fT6fSDKq6fEpg+eKf1TX9lZcVxHMMwrl69euvWrWc1fQzDEolELBbrFsBrmvbtt98uLi6eOnVKkqRSqbS5ubmxsbGzhJtlWbCOy+VyhmEYhgGrNjs7W6vVTp48CSn0QRDcu3evWq3G4/HJyclIJIJE9bPCD0LD8mTFKTeMetveXVTzdEESUnGWoUmS/FmixPEoN1RIphI1jqXDIOwT1Y7jq7pdbai1ZjQV5yICEtUIBALxBoFENQKBeMHwPA8ZyN302mazee/evaWlpdXV1dHR0YdLO1Ct+Xz+o48+CsPwwoULzWYzCALDMGq12vz8fKvVYlmWpulGo9FsNnmeTyaTU1NTBw8e7MuIfjjVanVpaWl2dnZ7e9s0TfCaqlQq0Dzs3r17zWbTsiwcx8GFu1AoDAwM7BrspShKEIRoNBqPx+FToC3X1ta+/PJL0Jzr6+uO4zwrj3SCIA4ePPhP//RPFy5ckGU5DEPf9zudzsbGhuu6S0tLNE1DRrTjODzPEwRx9OjRaDT6EKO1J4aiKFEUY7FYPB5XFKU7/ZWVlb/+9a+e57muu7a21ttS62mAHm8TExPvvffe7OzsvXv3MAwD8/NGo3Hr1i1RFBVFURTlQZ5wOI6PjY395je/uXDhwsWLF13XDcNQVVUYdrFYpGk6DMNGo6FpGs/ziqIcOnSoUCg81hcM8SBcL2h27ErTkFXbsFz/p7nfNEkwNJGKswOSEI8wP1/aNcdSySibjvPphNAIQ8vu2fEJsRALHdutt9RqQx0ZSCRDDMcxlAKOQCAQbwhIVCMQiBcMz/PZbDYej3dFdavVarVaIKqhnPXhZ8BxPJ/Pf/zxx5Zlzc3NgWi0LMuyrGazeefOnb6Do9HoO++802g0HqtfVLlc/u67727dugXBXgzDKpXK999/33cYSZI8z8fj8YGBgQeJapqmob8XtFbqZkGvrq6urq72DvWZiGrwxDp8+LAoiuVy+caNG1A8rKqqqqp9dctwsCAIxWIRXNae3gC8D5qmI5FIPB5PJpPNZtO27a6oXllZ6f3ss5o+SZKTk5Pnzp1rNptLS0sQanZdt91uPzLDHG7IxMQEx3Htdvvq1au+70Pbc13Xd9bkEwRhGMa5c+cOHDjwc1Skv4GARVm5acqKbVj9rb8pihB5Oh3jBjJCPLInO/cng2NIhiJScT6TFA3D7vvehGFoO169qVUaimm5QYiRGIZUNQKBQLwhIFGNQCAewR5bND3yDA86JpfLvf3225D622w2Ib0WwzBd1+v1OsjXRz4ocxyXzWZPnz5tWdaPP/54586dWq3Wbrd7S3bBOGpoaGhiYmJ0dLS3T/VeHsTz+fy7774LB6+vrxeLRcdxevOTCYJIJBLZbPbYsWOnT5+enp5+SJ9naHn9j//4j8lkcmZmBtKSYeLQP3lkZCSbzd69e3dtbW3X2/i4DtLxeHxkZOTXv/51NBpdWFhYWlrqdDqapvUOiaKoRCIxNDR09OhRSZL20qf6iZd+bGzsn//5n7/77ruZmRlVVW3bhtcFQYhEIiMjI5lMZmFhYWNj4+mnTxDE0NBQEASdTockyaWlpe3tbdhZwO43jobm0oZhFIvFvgbUGIaJolgoFD766COCIG7fvr2wsNButzudTu9MSZIURXFoaOjEiROFQgH6VCOj76chDDE/CE3Lq8lWvWXaD6qmzvDZFJeIMhxL/qw6FsexfDZ2eLpgmE65rvTt+Xh+0FbNektryHomKUZFhqZ+lupuBAKBQLxsIFGNQCAeBkEQUDAsCALIHugh1FsC3a0ohqgsVPD2lhND1bQgCIIggIzhOI6macigliTp5MmT29vbN27cgBAiHON5XqfTgYvSNM3zfO8luh+HS7Asy7LsyZMnBwcHh4aGCIJYWFjwPK+rVHEcZ1mW47gDBw6cPn0aRDW4ozEMA2cGhbzz5EAul0skEhC2FQTBMAxd16E+GbuvSPP5/PT09CeffHLu3LlcLrezmVYvIyMj//AP/xAEwdbWFoZhkE6M43gymZQk6cyZMwcPHgyCAGKhgiBwHMcwTFflQnMvWBp4BdblQYFlyOX++OOPJycn//znP3uet7W1BVXNMAWKoliWHRwcPHny5LFjxx4pqvsWRRCEnevOMEzfusMIYU/h17/+te/7m5ubBEF0p59KpbLZ7Ntvvz05Oel5Xr1ef8j04epwRajM33XAOI4PDAyk02nHcWAnRVEUx3E8z8NxPBKJDA8PDwwMZLNZWZa76egwWji5KIqiKL7//vuTk5NffPGF4zjQQa179+BLnslkjh07durUqYGBARDVO8cJp32sLIk3ljAMXS/QTK8uW/W2ZTv+zmMEnsqn+WyKi0UYjnnGpQr94HguEw0xfG2rSRBEGIS9nb08L+ioVr2l11t6QbI4lkSiGoFAIN4Qnk1yHQKBeF1ptVqrq6vFYrFcLoOHUzqdzufzo6Oj4+PjIK2hpbCu62CMzLLs9PT02NhYt1dTq9WanZ3d2NjQdR1sorLZ7PT09MDAgCRJOI7btr28vDw/P1+tVhuNhud5vu9PT09PT0+DN/Xs7Ozdu3d1XYcxRKPR6enpkZERSZJ6y1Zt2zYMY3t7e3V1tV6vy7Ksqqqu6xCjjkQi0Wg0n88XCoXBwcFcLkfTtGmas7Ozy8vLuq6DgE8mk9PT08PDw6AquyfvNpcuFouVSqVSqUD6tG3bnufxPC+KYjqdzmazYC0uiuLDRbVpmoqirKysLCwswGihi1UqlUqlUoVCIZ1Ow82Hu1ooFIaGhsbGxjKZDHbfzg3CqrAdUCgUpqenC4WCJEl9DcaAMAwNw4CLbm5uyrIMwWqQmjzPRyKRVCqVz+fB2DwajdI0/aAo6xOsey6Xm5qagp5hruuqqrq8vLywsNBoNFqtFiwTWG1DY63V1VVoKsZxXHf6YABWqVSgvl3XdejyPTg4ODU1BdPfdWfB933oc7a+vr69vQ1bNvDFkCQJUvEty6pUKjCj4eHhQ4cO9W6mmKap6/ra2trq6qosy+12W9M00zRpmoZ21olEAr5gQ0NDyWSSpmlYpscaJ6KL4wZ12VwuKpduVhc32qbtu97fg9UEjhMEvn9f/MzhzP59icnh2HMQsabtKZr9xYW7386sKKqpG87f38MxAseldPTEocG3Dgwc3Z9PJ34W/3wEAoFAvGwgUY1AIF4KIEDdbrebzSakVafT6VQqBR5jT3BCCFPLstxsNiGiG4/H4/H4M8zFVRSl3W4bhmFZVjKZjMfjPM8/XEjvBNojy7LcaDRAm8E4n8kIH0IYhuCz3Ww2TdOETQGYwvPMVYbpt1qtZrMJKQ/xePyRlu9Pf1FwKdN1HQLp0Wj0cRfOcRzbtpvNZqfTgUB0PB4XRRElez9DdMtbKSq3V+Rrt2rrZa3vXYokWIZ8ayr10an8SD6STfI/k+93H7bjfXN15dIPa8Vyu9ZU+96NiOxwIfHWwYGP35kcyifQdwGBQCDeBFD6NwKBeFkgCEIURYIgQGiBQH3iUB4krsfjcZZlIQLMsuyzFTwcxyWTyUgk4vs+5J/vGh9+OFCLG4lEICUYunY/qxE+HDAhx3Hc8zyKomiafm6X7gLTj0ajkBZOkuTjitsnuyhN0xCahok/wcLBUFOpVPckDMMgRf1s8byg1bHrsmXtlvjNc6SU4AqSkEvzMZF5bgn1JEEMZGMHJ3Om5TRkra8hve141YZaqimyYqeTHsuQ5M/nSI5AIBCIlwMkqhEIxEtBt9y0N+P6aSAIAnR1JBJ5JifcCcMwT69Cu27bu/qE/3zg9ztpPwcR+/BhPP/pQwH8U3bhhrv3/Lch3hBCDAv80LL9ZtuutR4gqlkql+bzaT4T50T++T3PEASezUSnHH9ju0VRpO8Ffvj3pHTX9Zsdo9rQ6rKezURTJEsSKMkfgUAgXnOQhQYCgUAgEIiXi8APO5pdbhrbdb0mm5a9m0UZS+ZSfCrGPp+s7y44jkcEVkpHMqlIKiGwbL+ex0NM0+3ljfp6sWn2trNGIBAIxGsKEtUIBAKBQCBeLrwgaKtOpWHUZFNWbMf7STMtgsBpioiKdCEjJKMM+Xyt1HEcEzg6GROkVCSTEjl2F9MH3XTWi63NkmyYbhAg9xoEAoF4zUGiGoFAIBAIxMuF74ctxa40Td3w/CDsM1VlaSIdZ/NpYUASks89Uo1hGIFjDIWnE8JANh4RmR0N+DDH8ZqyXm/pHc0xHT8MkKxGIBCI8CUYAAAAIABJREFU1xkkqhEIBAKBQLxE+EFoOX6zY1eapm65QdDfqIRlyEyCy6X5TIKPCDTx3M3hcByjKSIVFway8ajIkWS/qnZcv62ajbbe7Bia4fhIVCMQCMRrDRLVCAQCgUAgXhaCMDQtr6061ZZZa5mmtUs1NceQ2RSfSXA8S5LkizFcJwg8GecLuXg8xrMsRZI/eaAKgsC2vY5ilWtKU9Y9b5dZIBAIBOK1AYlqBAKBQCAQLwthgGmmV29bddlqKY7t/qSaGsdxiiQiAp1L8+k4y9DE8w9TAwSORyNcNhVJxvmIyFLUT0V1GLqer+pWpaY0ZN31UF01AoFAvM4gUY1AIBAIBOJlwQ9DWbFLdaOl2Lrpev5PRDVF4iJPpWJsISOk4hxFvrDHGBzHBZZKxlgpKaYTIsf81K4sxMIwNE2nXO80ZN12A5QAjkAgEK8xSFQjEAgEAoF4KQjC0HWDlmKXG4aiua4X9IlRmiISEUZKcFKSi4n087co6wJl1SLP5KXYUCEpCDs6loeYZbv1llZvaapuO66PotUIBALxuoJENQKBQCAQiJcC3w9tx2917GrDMCxv5wEcQ2aTXC7NJyIMz5IvKve7C0OTw4XE/jEpHuF2vms7vtw2Gy291TY0w0bRagQCgXhdQaIagUAgEAjEiycMQ8P0mh2r1jLrsmXZPxHVOI4RBM6zZDbFSUlO4CjqxWtqjCSJdEIczMXTSVEUmL7Kat8PTMtptfXNstyQdf+nqewIBAKBeG1AohqBQCAQCMSLJwixtmZv1/RK02gqtuX8xDEbx3CKJESezqb4dJylqZfiAYbAcZGnUnEul45IqQjLUL3vhmEYBKGsmPfWasVK2/WQqEYgEIjXk5fibxICgUAgEIg3nDAM26pTahhtzbEdv6+3M0URMZHOJLhcWkjGXhZRjeMYRRIizwzm4yODyYjI9jX4CsNQ0+3NslyuKZphO66PCqsRCATi9eOl+JuEQCAQCATiDScMsY7mlhumYe5STc3QRCbBDUhCPs0noiz14izKdsIy1MhAcmpUikd5giD6dLVpOZWaWmlorY5l2l6IVDUCgUC8diBRjUAgEAgE4gVj2p6s2NWWWW4Y+m4WZQxNSAkum+JFnqapfuH6YqEpMpMUB7LxVEIUBaav0ZfnB5btNmV9rdiqNXUf2ZUhEAjEawcS1QgEAoFAIF4wqu6WG0aprpcbhr5bpJqlyUySyyQ4hn7pHl1IkohF+EwqmklG4jGepsmfvB1iYRC22sbCSm272kZ2ZQgEAvH6QT36EAQCgXihBEGg67qqqs1mU9M0juMikYgkSdFodGem5fPEMAzDMIIg8H0fwzCCIHie53meJEmCeOme+xGIl5YwxBTd3a4bsuJYdn8/Z5LEeZZKx9lCRpCSHNMnWV8CoGd1hKezaVFKRTTNMi23L81bN+1yrVNvJXTTJUiCeTlqwhEIBALxTECiGoFAvOwEQSDL8ubm5q1bt4rFYjKZHBoaOnnyJM/zNE2/QFGtqmqlUnEcx3VdDMNIkszlcul0muM4JKoRiMeiozmlmqEa7s7saJokklGmkBEGs4L0UkaqcQwjCYxjyWwqkk9HK7VORzUxDOuV1Zbl1ltqs613NJuhSZpkXqYEdgQCgUA8FUhUIxCIl5owDC3L+vHHHy9fvry6ulqr1URRHBkZwTAsCIKBgYFYLPaixra6unrp0iVFUVRVxTCMZdmTJ08ePXo0n8+nUqkXNSoE4tXCdn3L9uuytV3XVcPdeQBNEek4m01xUYGm6ZermroXlialVCQvxSKbHN02XPcnnmSuHxim2+qY5brC0ITIv8gNQQQCgUA8W5CoRiAQLzVBEJimOTMz84c//EFRFMMwcBwfHBxkWZbjuGg0+gJF9eLi4meffVatVuv1OoZhoiiqqioIgiAISFQjEHvEsv22atdaZqlu6LuKappIJ7hciudZiiReXiFK02QmHcl3zGiEo2nS9/2gJ+ru+4Fle3LHLFY68QhbyEQw7OWdCwKBQCAeCySqEQjEoykWi3Nzc51OJwgCnudjsdjIyMj4+DhF/ey/Q3Acx3Hc8zz7PhiGmaZp27bneUHwIi1/XNc1DAPqvbH7QfVnPirLsubm5lZWVsIwJEkyGo1KkjQ+Pp5Op5/hVRCIF4VquKW60ehYmuE63k/+7xAEztJkIsLm04KU5Dnmpaum7oUkiQhPJ2OclBJrDb7h+V7PdMIQC4JA1axStTMgRb0AIwgMxaoRCATi9QCJagQC8WjW1tb+67/+a3193fM8SZKGhoZ+8YtfjIyMPB9RTRAETdMMw5Ak2X2x7x+vMYZhXLhw4S9/+Yvv+wzDDA0NHTt2LBqNIlGNeD1QdXe7ZrQ6tmF5feXUBI4LHJWMsfk0LyV55iUX1QTOc1QixmZTYjkhdBTTxHoC72EYhpimW9vVzvhI2vNDisRepmbbCAQCgXhykKhGIBAPw/d913VbrdbGxgYESw3DwDCs0+n0edv+fLAse+rUKV3X19fXG40Gx3FDQ0Pvvvvu9PR0NBp9PmN4IYRh6LqupmmlUmllZQXSBMIwzOfzlmW96NEhEE+L74eeH8iKXazpqu7u/IVCU3gqzhYyfDLGChxJvfQaFMcwkWdGh1Idza41NFW3+35PmpbbaGmNlt7qmMkoGxHoFzVUBAKBQDxDkKhGIBAPw/d90zRlWS4Wi1tbWxiGeZ7HMIyiKM8n9RrHcY7jzpw5UygUZmdnt7a2UqnU0NDQiRMnnk+o/MXiOI6qqtVqdXNzE8MwQRBomm40Go7jvOihIRBPi+8HtuO3FHu7buxqUUaRRDrGFtJ8MsoI3Kvxn13gmNHhdEezb98rEwQeBFivrjYt13a8Rktrtg2WJpCoRiAQiNeDV+NPFAKBeP5A+2VQdI1Gw/M8eN33fdu2FUVpNpumaeI4zvO8IAi7nsF1XVVVdV3XdR0CraIoRqNRlmVJktx78jZN05FIZGJiIp1O8zyfSCR2vWIYhhBL73Q6hmGYpklRFMuykUgkFovRNP0gEQ4fBAULo8UwrDtayDx/ylRziPmbpqlpmmmalmWRJAk9t+ESfWPzfd9xnGazWSqVYDzY/di1YRiyLDcaDRzHaZoWBGHnvIIgCIJA0zRVVeFWiKII0xEE4cX290YgANVway2z2jSbbcu0vd63cBwjCULgqEySy6Z4jn2pE797IUk8KjCZhJCXoh3V7Cim4/o974dBgNVa2vxiicDy2dQuv8cQCAQC8cqBRDUCgdgd3/cNw2g2m1tbW9VqtRsa9TwPYtelUonneRzHc7kcz/NYT4UzCHLHcTRNq1artVqt0Wj4vp9IJCRJyufz8Xic4zjoMr1HdUdRVD6flyQJwzCGYWiaDoKgGwICVQwXbTQa29vbjUZDlmVQrfl8vlAoiKIoCAJJkn2SEkZr2zbsINRqtWazGYZhMpmE0cZiMY7jKIp6gu7TMEg4v6ZpsizX63VZltvtNsMw8Xg8l8vl8/loNCqKIkEQ3UvAeMrl8sbGBhihwdng9Vqttr29jeN4NBolSbJX84dhGASB4ziWZdVqtXK53Gq12u12JpORJCmXy0mSBDdw7zcfgfg5UA13u25UmmZLsfvewjGcpgiBoyQQ1S93NXUvFElEeCKd4AtSrCHrhuH0iuowxDAsbLS0+aVyNiUcmcqh/4UIBALxGoBENQKB2J1KpXLp0qXbt2+vr68vLS0pigKvm6ZZr9cvXrxYq9VomsYw7NNPP/300095ngdp7XletVotFovz8/NLS0utVktRFNM0gyDgOE4UxUQiMTo6euTIkbGxsaGhIZZlHz4S27ZnZmauXr3a6XQgZptOp99+++3Dhw8PDQ1BSy3btjudzvLy8s2bNzc2NqrVqq7rEA1mGEYUxUgksm/fvtHR0QMHDkxOTnYjw47j1Gq1zc3N+fn5lZWV7mgxDIMQdyKRGB8fP3LkyOjo6ODgIMMwe7+HjuO0Wq1yubyysrK5uVkul2VZNgzDsizbtgmCYFk2Ho8nk8kjR44cP368UChks1kcx33f/+GHH65cubKxsbG2traysgIn9Dyv3W4vLCz84Q9/+Oabb3Acn56e/vWvfz0xMQH7BRiGtdvtWq12586d2dlZEPBglg4JBZlMplAonDx58vDhw/F4fNeAPwLxfFANb7tmqPouid8kiSdjzGBWyKX4ZJRh6FdGVAMRgZ3Yl9EMp9nSDcvtq6zWDbtSU+otXVYtgaN5Fj2MIRAIxKsN+j2OQCB2p16vf/vttxcuXNjc3OymH2MYZlmWZVnNZvP777+HV6LR6NmzZwmC4HkearA3NjauX7/++eeff//995qm9bpqURTFcdzhw4c/+eQTx3FisVgymXx4crXrurOzs5999lmlUmm1WhiGjYyMhGEYi8VSqRSIal3Xi8Xi1atXP/vss3v37smy3M1XB3AcP3bs2PHjxzEMGxwcJAiCoiiIxq+trc3MzPz1r3+9ceOGpmnQtQugaZpl2ePHj8uy7LpuPB6PxWJ7TwW3bbtSqczNzV24cOHHH39cWVnpdDp9x7AsKwjCL37xC/gxk8kQBBEEwa1bt/7nf/5nc3OzUql0D/Z9X1EURVGWl5fhlQ8++GD//v35fJ5lWRzHgyCo1+vz8/Nff/31n//851arZRhG94Eex/FEIjEwMKBpmiiKo6OjkGuwl7kgEM+QIAyDAFM0Z7umK/ouBgEkiadi7IAkSEkuFnmMnayXBFFgxobScsdcWKm2FdPzg15dbZiuZXu1ll5vGVJSQKIagUAgXnXQ73EEAvHMAEW3sbHx5ZdffvvttxsbG7quu67bd4xt2xsbG1999ZWiKGEYHjhwYHh4+LEiwDsplUpffPHFd999t7m5qWnaThO1MAzr9frdu3fPnDnj+34QBJ7n1Wq11dXVL7744tKlS+vr64Zh9ElxSNteW1v74osvVFUNw3B6enp4eHiPBmmmaS4tLd28eXNhYaFYLO5q2e15nmEYCwsLvu9jGDYxMcFx3JPdBE3T6vX6lStXPv/884WFhXa77ThOv/mwaVar1QsXLmia9i//8i/pdPrJ0toRiKdBN72O5mzX9e26ruwWqaYIIhljcyn+FRWcNEXGIlw6KUqpiKpZHdX0ftIvLAwCrN5U55fKRyZz6QT/wgaKQCAQiGfBK/m3CoFAPAcoihIEIRKJCILguq7neaBUoe4XQrgQ5OwaX/m+XyqVZmdnL168eP78eex+l2mGYRiGwXEczuN5XqVSqVQqEKmmaTqbzUKJ7x7H1isUoWK5UqlcvHjxhx9+6HQ6nufhOC4IgiiKcE7bth3H8Tyv0+nY9v81ufE8b3t7++bNmxcuXLh06VJ3tCzLwmAcx/F9Hw7b3t4OgiAajdI0nc/n9xisdl232WwWi8VisdhqtWiajsViIMjBF82yLNhlWF1d3d7enpqa+uUvf5lIJHie5zgOCq15nof71h0hRVHd9HUwOQNV3Ol0FhcXZ2ZmvvzyS1mWMQyjaVoURZqmaZp2HMdxHBjSDz/80Gg0Dhw4cOTIEVEUn1jGIxBPhm565YZRaRq1lun5/b20SAJnWTKTYHNpgX91LMp6oUgiIjDJuCClI822rhuO5/19p+//KqtlfXG1lkuJU6MZAsdQzggCgUC8uiBRjUAgdieZTJ4+fZogiHQ6vbq6urW11a00jkaj4+Pj+/fvB1137NgxMKB2HGdubu7zzz9fX1+Hk5AkGYlEMpnMxMQEwzBQ7dxutyHFulwunz9/PhKJHDp0qCtlH3ec0MkZvLjAYxxk54kTJ371q18xDOM4zvLy8uLi4vj4+NTU1LFjx0Aba5p248aNL7/8ElqFYRhGUVQkEslmsxMTEziOb2xs1Go1SPzGMKxYLH7zzTexWOzIkSMURUE9+cOJRCJvvfWWqqqbm5ue542MjAwODuZyOYIgTNNcWVmZm5tTFAWkexiGlUrl7t27k5OTIyMjk5OTH3/88Z07dyDKXavVuvczl8vt378fPNumpqYGBgZgg6NYLH799dc3btzohsQLhcLExMTw8PDQ0NDKysrKykqxWATbOShBn5+fn56eLhQKj3vbEYinQTfdUt1oq04QYn2SmiDweJQZyAiDkphP8/wr0kmrHxwjCEzk6XwmWm9p9aZmu14Y9FdWV+tqs210NFvgKI55NWeKQCAQCCSqEQjEg4jFYkePHgWvbHDz6orqRCJx8ODBX/3qV+AxNjk5KQgCjuOGYdy7d+/KlStdt2qGYQqFwv79+0+fPs3z/Pz8/J07dxYXF0FUN5tNXdcnJydlWY7H4xRFPYGoDoLAsixd17vF2zDmgYGB9957L5FIYBg2Pz+fSCQOHToE7mgQe9d1/e7duzMzM93Rsiw7ODh44MAB2E2Yn59fWFgwTRNEdb1eVxTl0KFD7XZ71y5WOxEEYWpqyjCMO3fuwE0bHx8fHBzEMKzdbhMEsbKyYhgG2I/7vt9sNtfW1jKZzOjo6L59+8IwZFnW931N03pF9dDQ0Lvvvjs1NYVhWDqdzuVyNE1DmsC1a9fu3bsHVu04jhcKhdOnTx88eHB6evr69esMw5imWalUYCdiY2NjaWkJrNEf97YjEE9GEIZBEHY0p1Q3Opob9kepMZLA4xGmkBGyKT4ZY4lXM36LYxiOYwJH5aVYua4ybIuwXD/0e7cQDMvF2nqzbcgdk8B5JKoRCATi1QX9BkcgELvD8/zg4CB4km1ubnYDs5BRPDw8fPz4cbD7jsViDMNomtZsNmVZVhSlW0cdi8U++OCDc+fOTU5O8jx/6NCh7777DqLKGIb5vm9ZVqfTKZfLD2o9/UggYZtl2W5hMPTHXl9f//rrr/fv3z8xMTE9PT09PZ1KpZLJZDQaDYJA13UYraqq3dGmUqmPPvrovffem5ycpCjq8OHDFy5cqFaroLohmNxut8vlMnTneuTYQAPv37//d7/7naqqsViMIAjbtsED3HVdaJ3tuv9nDmxZFph1YxiWyWRIkmy1WtVq9d69e93J8jyfyWQOHDjw1ltvYRjGcVw8Hg/DsNPptFotWZa74XqSJHO53JEjRyRJgn8fOHCgm0Tg+36n02k0GrtWeiMQPxO2E+imW5Ot7Zre0ZwdgWqMIPBkjC1k+IhAE694uymBo/NSNJeJRgRGN+zeLoAYhnleYJpOs60Xq22GJhJRVIWBQCAQrypIVCMQiN1hGCaVSmmaJkkSWF7D6yRJsiybTqf37dsniiK8GIahZVmNRqPT6YAmxDAMx3FRFI8ePfrBBx9IksRx3MTEhGVZX3311draGriFBUEA3aEHBgagVvlxxwmiGjp1RSIR0zQh6ru9vX3lyhVN0zzPGxsb27dvXzwej8fjNE2HYQiiuts9C0YbjUbfeuut999/X5IkmqYnJycVRfniiy9KpVJ3tIqiVCqVbDa7l+gujC2TyVAUpSiKpmmNRqNcLler1Wq1Cr27e4+3bVtRFNu2cRyPRCIMw0iSlEqlujXPOI7TNB2NRgcGBsbGxrofVBSl0+nAHgFkAUCVO/ixwS5Gp9PpfaYPggDS5nvdzhGInxvL8Zsduy5bddnSTa8vUk0SOEsT6RibzwgCR73SihrDMJahMklRSomJGKeoput4vaFq3w+CIJTbxnalnUkIYRjHcOwVnzECgUC8oSBRjUAgng1gPd2rUcGiLJFIpNNpiCSD+VYikYhGo6qqgvkWNL5WVXWnZfdeAJGfy+WOHTumadrS0pKmaRiGtdvtpaWler3+448/FgqFgYGBs2fPnj17Np1Ox2IxVVUbjUbfaDmOg2g2WH/xPB+NRpPJZCQSUVUVBLBhGPV6XdO0cGfe6g7ARK3RaPzwww/z8/PLy8u1Wk3TNGhVDTK4a5yGYZjruhDBftybYNt2s9nsdDpdlQ76eXZ29v+3d2fNcVXn/sf3PHXvngfNlpEcWxhjBgNFQaWSOlVJblJJblKVd5PXk4u8gH8lVakDVIUizPAHbBMLyRp7Hnfv+Vysg44sT/KOwVj+fi5Rd2t5gbB++1nreXq9ngjYQRD4vr+7u3v0gul0Oh6PM3w7ILOpF+61p+3+bOJFQXTHQyVZlmxTK7tms2LPVW3HeipblB2nqUrO0squ1ai4g+FsMvGPP0YTP/i9obe9N1hdrEaJpCrS0/4cAQCeTYRqAI9Bmqa+74/HY3GbV5BlWbQQdxxHtMsWlVvbto8aVkuSJK43z2az08TUu4nibaPRuHbtmjhHvbe3NxqNxD3wg4MDSZJKpVKlUplOp5IkvfjiixcuXPA8TxSxH7pay7LEakVeFbeRT1nd9X3/4ODgq6++eu+99z744INvvvmm0+kkSaLruuM4aZrKsqyq6vEkLFqOP+omiLlcx/cwTVOxFa1W68SLRd3bNM0M5wKAzNI0TVJpNA33WtPuwPfDOL6z77csy3lHq5etWtmqFEzLeOr/+1QU2VDUUsFamS/1h95he3RnrVpK03Q0nu0eDnsDz5uFlqEaOvPtAODpQ6gG8HiIFHr8BmR6J/ElUbyN4/go/onbv//hqORms/nzn/+8VqvNzc19+umnn3/+ebvdPqoniwHU77333vb29p/+9Ke5ubkwDO/+jmmaigLvA1YrYvApL3r2+/1333333Xff/eijjzY3N8Vcbl3X6/X66upqkiS9Xq/T6dx9DjyDE0s6mnZWKBTuuVrLsur1erlcFt3mgB9anKRBmPRHwW5r0h/5SXLyIZosS6W8sVB3KgXTsTRVOSNF26JrX3yu0RvNbmy2PD888fRw6oWd3rjdG7d7k2rJNnR+HgHg6UOoBvAQsnxHsyBRTT1e4xWvMU0zn88f9TMT0TSO4+l0OplMRLvsKIpms5nneUEQHP1mqWmaGJX8n7QkchxneXlZ0zRxhLtarW5vbx8eHrbb7U6nI8rLt27d2t/fv3btWrvdTpJEDHA++oQ0TUW9dzKZiLvi91ytYRj5fP6hQVTk88Fg8Nlnn/3zn//c3NwUvdkqlcrq6ura2trFixeHw+GNGzfiOO71evcL1Sc2XywyDMMTR+U1TROjrY+/WJblubk5Ma7MMIwTn2wYxuLi4vr6erFYfOj2Av+5IEwGY7/dnx10veHkZN9vRZF1VSkXzIW6U8wbunZ2CrY521hoFufrbrFg+0E0m92Rq4MwSifpQXu0tds1tGoxT6gGgKcPoRrAg4j68/GibhAEo9Ho7tPaYsLzUesy6fsz4e12++DgYH5+3rKs8Xjc7/eHw+F0Oj2KkbZtNxoN0Rk78zrjOPY8TzQYP3fu3Ntvv721tfXvf//7/ffff+edd8R3FCH58PBwa2urWCw2Gg3RvVz6PgPPZrNWq9VqtcQY6qPVep53FGIdx2k0Gq7r3u8RgNiWNE3DMBwOh999993m5uZkMhFfXV5e/sMf/vDyyy8vLy9fv349DMNOp/OAP/iJ+v/RRegTDzVE6zgxluz4P7906dIf//hHUZE+8cniDnk+n7/7S8APYTqL9trefmfaGfjTWZTc+T8QTVVsU6sWrfmak7PP1C8nuqa4jl4rO4vNYhhGB0F0/Nx7mqZRlOweDP7/zYNywV6a4yEXADx9ztTfWwAeO3Gv+PgV6CAIxBCsmzdvihRXKBQKhUIul6vVaqIFt+/74nrzZDL58ssvy+WyGKm1v7//ySefdDodkQlVVdU0rVgsNpvNzKFalMQ7nc7169c9zxMjpmq1mri3vLe357rubDYT46BFrhYtzUXLtOOrHY1Gn3/+eS6Xu3DhgqZp+/v7n332Wa/XO7Haubk5caZaURRN046P8hoMBgcHB6KDdxAEnucNBoPBYHD0AMKyrGazWavVjgZQH0/sJ8iybBiGbdtHUTmO48lk0m63Nzc35+fnxSXwQqGg63qpVBIDwyzLEle+xUkB3/cNw6hWq+LhyGQy8TxPlPTr9TplavxoprPooOO1+/50FoXRyf/mHUurFK162aqXLdt86m9THyduVleKznPL1akXdLqTJPm/RwppKiVJ2u5N9O3uxfP1mR/pmqKqZ6dQDwDPAkI1gAfRdT2fzzuOcxQdfd+Poujrr7/++9//Xq1W8/n8xsbGxsaGZVnlcrlerzebTTFbS5KkwWDwj3/847vvvrty5Yppmjdu3BAdsI8+3HXdarU6NzdXKpUyh+ooim7duvXXv/51e3tb1/W1tbWrV6+6risagzuOo+u6LMsi2YrGZrlcrlAoNBqNer0uzodLktTpdP72t7/dunXrhRdeUBTl+vXr3377bbfbFd9IHPyu1Wrz8/OFQkGSJBFNj9p9xXF8+/btL7/8cmVlZX5+Pk1T0XXseEl/Op3evn1bUZQ0TT/77LMPP/xwd3f3RNn5OMuyRGY++hbj8fj27dvvv/++53n5fH5paWljY6NUKhUKhVqt1mg02u12t9sVn/nVV1/95S9/eeGFF8T+67q+ubkpHjQ0m8233nqLUI0fjTeL9zteb+jH97pNXcjpiw2nWbWrRUs7i5GyXLAvPdfoD72bm+0wuuN/C2maDsczWZbbvWl/7BdyhnMWdwAAzjBCNYAHEWOWq9WquBQdf29zc/Odd95xXVec915YWLBt23GcCxcuvP766x999NFoNBKpcmdnZzwej8djTdO2t7e73a4YZCXLcrVavXTp0sWLFyuVimma2e5UT6fT3d3dL7744uOPP75165aiKLu7u61Wq1wu27Z98+ZNMTUqTVNN00zTFHVssfKLFy9eu3bt448/Fi3NZrPZ9vb2eDweDoeSJG1vb/d6vdlsJlZbr9cvXbp04cKFUqkkbi/btl0ul4+eEcRxvL+///XXX29sbCwvL5dKJdHq3LKsMAzFcfder/fpp59+9913URRtbm6KzblfpVpRlKMyvq7rYlZ2GIbdbveTTz5pt9uO4zz//PPFYtGyLMuy5ufnX3311TAMP/nkExHmDw8Pfd8fjUYUaA4tAAAWG0lEQVSHh4eGYWiatrOz02q1crnc4uLi4uLi6uqqYRjH75YDj10YJZ4ft/qz3da0NwruDtWSJBXzxmLdKbmm+fQ3/b4nxzbmG4VG1S0V7CRJvFlw/Px7GMaTqb/fGm7t9s8tlByLH0kAeJoQqgE8iG3bCwsL8/Pzruuapul5niiw7Ozs9Ho9cfi5XC6/8MIL4izx1atX0zTt9/vffvttkiQiBPZ6Pc/zFEWZzWZRFCVJIvLz0tLSr371q2vXromqcrYV9nq9Dz/88IMPPtja2jo8PJRlud/v37x5U6xtOp0eDcQ2DMN13bm5udXV1WKxqOv6K6+8kiRJt9vd3Nw8Wm232xXDt0RN/mi1586d+81vfvPyyy8f1e1zudz8/PzOzo5YSRzHBwcHaZreuHFjdXXVdd1isSiGcg+HQxGqW63Wu+++q2mayPDiUPr9/miKolQqleXl5Wq1atv2bDYTE8vEMfVvvvlGVdVWq7W2tlapVOr1+tLS0q9//esoim7cuCH+TXmeJ5q03bhxQ9zNFmfdVVXd29t75ZVXrly5UqlUCNX4QXl+dNDxbh9Obh9OesN79P2WJLmYNxbqubxzZn8tMQ2t5Cq1cq5Ry4dhNJuFqXTHPkRxsnMw+PrmQSFnNKu5+30OAOAn6Mz+7QXgsRCtuVdWVt544w1FUUThN4oiz/NEwVmSpP39/cPDQ8dxyuXy/Px8kiRbW1txHG9tbR0cHEwmkyAIxPnqow+s1WpLS0tvvvnma6+9JoqlDy5TP+Cr4jrx/Pz8+fPnRUKeTCZieNXRa2zbLpVKzz333IULF8RhaXFLfHFxMQzD27dvy7IsMvl0Og2C4GjatriYLfLq22+//dprr62srIjD5JIkzc3NvfTSSwcHBzdu3BDvEmVh0e0sDMNisXjlyhUxqnp/fz+KIt/3fd8XNfNSqXT+/Pl+v7+zs3N0Avx4u29Zli3Lqlarly9ffuONN27evCnOikdRJE7XS5K0t7e3v7/f7/fL5XKpVLp48WK73W6329evX9/e3h6NRtPp1Pf9o9eLm+Hi9MFRnzbgB+X58UF31urNJl54921q21QdS2tUrIW6k7fP7PMdVZFNQy261ny9MJ747d5ESu6I1UmS9gbe7f3++mrN8yNdU87kMXgAOJPUP//5z096DQB+uo66TxcKhTiOd3Z2PM+Louh4ZF1fX19dXRXFUnEHu1arLSwsiJZmvu+LE9SCSIlXr1797W9/+4tf/OLy5cvVavUopt6T7/sffPDBF198Idp6SZJULBYvX768trYmLmOXy2VxCj2OY9Ek7MRN5mq1uri4+Mtf/vL3v//95cuX6/W6mDUtateNRqPZbIrkOZvNRJcvwXGcarX66quv/u53v/v5z3++sbFRLpc1TROr1TStXC7v7u7eunVLBGZJknRdn5+fbzabq6urzWbTtm3DMPb397vdrji/LTahXC5funTp7bffVlV1e3tbPHRYXFxcWVm5cOHC+vr68f0Xu9rtdg8PD6MoOj5/q1arra2tNZvNer0ubr8Xi8XV1VXTNMXmHw3rFsRHPf/889euXXv55ZfPnTtnWdaJnuHA49Xp+ze3h1v748Pe7O5QXS6Y8zXnxfXK88+VbFNVzsp46hNkWZJlaToNpn40GHkHnXF6Z8VeliVJlhRFnm8Ua5W8rilnaa4YAJxt/CIF4CFkWS6XyxsbG2K28/7+/nA4FC24dF03DOONN95YXl52XVc6VtkWd6RXVlZarVa/3xcjuMQs60qlsr6+LsZKiWPYD16AYRhXr14VnbTFbKpKpXLt2rXFxUURWUV38TiO5+bmLl++3O12x+OxGKat67q4Ft5oNK5evXr58uVisXjUEU2s9ty5c4ZhqKp6/vz5w8NDkUXF9y0UCpVK5Wc/+9lLL720uLhYKBSO509xav2tt96SZbndbvd6vSRJLMu6cuXKhQsXCoWCZVmrq6si4W9sbAwGgyAIVFV1XbdSqaysrKyvr6+vry8vL4vnDmKE9dLS0vHNNwzj3LlzooK9trZ2NJBMdFw7f/78xsaG6HauKIphGI1GQ4zRdl13f3+/1Wp5njebzUSNWnRoO3/+/Nra2urq6vHW4sBjF8WJHyTdgb/XnnaH/vFRUkcKOX2xnisXTMtQJelsJmpBlqR8zlxslm7vDSxTm6VpGP7fA7I0lXw/Go5n7d641Z3ommyb/GwCwNNBPjFpFgDuJoZR+b4/nU77/b5oNjabzfL5vOu6tVpNhLrjk5/E8CpxbXg0Gok6rbhgnM/nRfsuXddFxfih3306nYrrx6LSq6qqbduinbXopB1FkTiAHYbhaDTq9/uTyWQ6nebz+WKxKL6pZVmmaaqqerzNuJhQLVbred7RatM0PbFaTdNOrFa8dzKZjMfjVqvV6/WCIFAUZW5uTsyytixLDLUSI6/b7XYYhqZpuq4r+qgZhhGG4dHQb/GQQqzz+HeJ4zgMQzGhutvtimq8oijic2q1Wi6XE1shtitJEnFAQPyhut1up9Oxbdu27WKxWCwWje+JMwLZWsQBD+X5UW8YfH6z+98f7d8+mIy98O4uZS9frL5yqbpxvrS+fPZ70Y+nYX8U/PcHN//fO18PxzPPC45/VVWVfM68dmX5peeXNp6rLzYLT2qdAIBHwkNQAA8nSqBiplShUKhWqyK+HnW3Ph4CJUlSVVWMsyoUCmEY+r5fr9fTNHUcR0y9fqS2ZIqi5PP5fD5/vxeImq0YlCVJUrFYrFar4vayZVmO44jF3++9R6t1XVek60ajIVZrWdbxGd33fK9oJJ7P5xuNhjiYLY5hiwQuorhlWSIAi/K+GPR1yk0QHyLGdxUKhVKpJNqPKYpife/4R4l51KIoLZ5uVCqVubk5kdjFH4oUjR+HHyTt/qzVmw3GgXdX229DVy1DrZetpUbOde79E3rGGLpazBvVktOsuUmSngjVSZIGYdzuTfcOh8tzxST530PjAICfOEI1gEcgWmeZppmmaZqm8vce8HoRdx3Hkb7vwvVDJzpRgD06hnP68dfi9rKIo9KpVyuidT6fF++6+43iBeLpw2k27QHfSNM0UTy/5ze6myjL67ou5lH/OPsPHJkF8WHXO+x6Yy8Mo/jE2TjbVEuu2azYS41czjmzLcqO0zVF0+RqyV5oFiZTv9UZHf9qmqZRGHd7k92D4XgSxomkqmf6QDwAnBWEagCP5lFTmXhx5olZGfwnuTHzex8a3R9Lms3wIaRoPBFJkoZxMhwH+x2vM/DDMLn7tlne1hdqTrVo5Rz9GWnKJcuSLMm1cv759fnhaPbtVvvOuVpSkqTD8azVHbX74/7Ic3OGdUYHdwPAWUKoBgAAj1mSpL4fD8bBfnvaGczC+B7z2POOvtjIVYumbWnP1IOfeiVvmebm7Y4kSydCdZwk46nf7o7b3XFvODV1hVANAD99z8SDYQAA8GPyw6Td9/c7Xnvgj6fRib7fqipbhlouGAt1x80Zz1SiliTJ0BU3p9crueX5csG1Tnw1SdNZGO22hls7vcmdl64BAD9NhGoAAPCYBWHc6s/2O9POYHZ3029NVSxDK7vmQsNxc8/EberjdE3J2Vq9kltdqhRd+8RX01Tyw3jvcPjdTm889Z/ICgEAj4Tj3wAA4DFLJUmSUkWRbUMzdTWMkuOXqvO21qzYzapdK5qO9Yz+KlIr5zfW5qbTcO9wmCRJmqaGrlmWXnTtWiV3bqGyulTJO+bDPwgA8KQ9o3+TAQCAH4iIz7IsG5qSt7WxpU1mURQnaSqlUpqmUs7WF+pOs2pXCqaiPGunv/9XtZy79Fzz9l5f19U4lqU0zTlGqWgvz5WX5kvL8+XFuQKhGgCeCoRqAADwOMmy5JjacjNn6Eq5YB50vPbA7w39wTgYT8PJLMo72kLdKeaNZ3lglG3q1ZKz2Cyun6tZpu5YxlzdbVTzpYJdci03b+Yd06RLGQA8DQjVAADgMbNMdd50yq45V3Vavdlua7rbmuy2pq3+TBrMygVzoZ4r5nX5GU7VlqnpurY0V7x4vlEs2NVS7txCeaFR0DRZU5/dbQGApxGhGgAA/CA0TSnkdE2VCzl9vmavLwfdod8d+itz+ZW5fCGvP8OZWlJkSVaklYWSY2umoVumVshZmiY/s+fhAeDpJadp+vBXAQAA/AfCKPHDeDQJB+OgmDcaFVslPQIAzgQq1QAA4AenKrKpq3JetkzV0FVFJlEDAM4IKtUAAAAAAGSkPOkFAAAAAADwtOL4NwAAeGRpKsVJEiVpGKVRnERJGiepJEmKLGuKrKmyrimaIquqzEnvR5KmUpKmUZxGcRLESRSnidhYRdYUWdcUXVU0VVZkmX0FgJ8IQjUAAHg0SSolSTryoqEXtkdBdxQMp+HUjyVJMjSl4OjlvF51zVJOd23N0GhJdlqpJCVJ6kdJbxJ0x0F74PcmgdhYx1RLOaNWMCt5o5wzTF1RFXI1APwkEKoBAMAjSFNp7IVDL+qM/M4oOBzMDgdBfxJMZpEkSaaulHJGtWA0i2G9YFTyZjGnu7amq9w4e4hUkuI47U3C3jg4HM4OB/5Bz+uMgvEskiQpZ2oV15grW82iNSlElbxRyus8rwCAnwIalQEAgNNKJSlO0q+2h9/sjnrjoD8JvSD2gjiMkihOJElSFNnQFFNXHUN1Hb2S11cbuY2lQsHRiX8PlqTSZBZ9utm/vjseeeFoFnp+7IdJlCSSJGmKYuqKbah5Wy/Y2vq8e3W16No6qRoAnjgq1cCzKE3T/vdGo9GTXg6AH5VhGNVqtVwuu65rmuYjvXcWxBM/2ul6N/fGIy8cz6Ikle54QJ+kQZRM/Kgvy9Y46I51RZYbRVNTZVtXFSLgfYhE3Rr6W+3ptwfjWRCHUZIe29kojv0wHnlRbxLahmroykLFkmU5bzGfDACeMEI18CxKkuTGjRsfffTRhx9++M033zzp5QD4UZXL5bfffvu1117b2NhoNpuP9N7eJNztevv92WAaBGFyR+w7Jk2lREqDKBl7YWvob7U8RZbnK7ZJqL6PJEkPB7PNw0l35PthHN9ra1NJktI0jpNZKHVHwa2DiZRKdt1RNHYVAJ4kQjXwLEqSpNfr3bp165NPPvnXv/71pJcD4EfVbDabzebq6upzzz33qO/tjYN/H0zaA38WxHEiPeAOWZpKUZKkQdobB1vtqW2q9YJpatysvrc4TTuj4LvWtD8JoyiJ73M9L5WkOE2lKOlPgq3WxLW0paqt/9iLBQDcgb/bAADAaXVHwb/3xt1xEN2vSH1Mmkpxmg69cKs1OejPwpg2LveVJGl35N/uTIdeGD9wa8WujrzodsfrjHwxyQwA8ARRqQYAAKc1DaLuOAjj5JR9TtNU8sOkn4TjWZQQ/+4vTaVpEA+n0Wn2Nk2lIEqGXjQNyNQA8OQRqgEAwGnFSepHyWkjtSRJkpSkUhgnUcy4kYeIEymMT/vkIUmlKE6I1ADwU8DxbwAAcFqyJCuK9EjtxmRJUmSZDmUPJcuSLJ+2kbcsSbJ8+pcDAH5AhGoAAHBamiKbmqI+SkRWZEnXFE0lAD6Eqsi6Kiun+9VMUSRdlVWFXQWAJ49QDQAATqvg6IsVp2DryimqpKJGnbO0uZJVyZsa1er7UxSp6GiNopkztQfvrdhVx9AaRbPo6AypBoAnjlANAABOq+Iaa/P5imtoiqxID4lzsiypiuTa+krdaZYsjXla96fIciVvLlcd19ZURZbvv7eyLKuK5NraYtWuuobKpgLAk0ajMuBZpCjK8vLym2++WalUXn/99Se9HAA/Ktd1X3zxxbW1tXw+/6jvLef18w1nOA3aQ98LYj+892QtceNX12TbUGsFY6XmNIqmrlJTvS9VkesFc9bMDWdRbxwGUZIm99hZWZZURTY0pZw3Vuu5RtFUT3leHADwgyFUA88iVVUvXLhw/vz5//qv/4qi6EkvB8CPSlEUXdc1TdO0R/41oJwzcqa235/tdGepFIRxmiTSvcKfpMiSoamurdeL5krdqbkG8e8BFEWuF01dU3a6s9sdL0nTOJWlO7dWliVZkjVVNnW16prnm/mqazzS/XYAwA+BUA08o3Rd13X9Sa8CwFNGUWRDU1ZqjiRJ3XHQHwdDLxp5kedHfpRIkqSpim2oeUsrOFrJMcqusVixXUtTCX8PJEuSpip5S1ufy+uq3J+Eg0kw8MLpLPKjREolQ1dyplZw9FJOLznGSt1xbU2cFAcAPFnyvQ5tAQAA3FcUp0GUdMZ+e+jvdGZ7Pa89DEZeKEmSqav1gtksmYtVu1E0q67pGCrZ75TSVPLDZOJHB/3Zbm92uz056PtDL0zTtODojaK1XHMWKnazaOUtzdAVHlQAwE8BoRoAADyaNJWSNJ368dSPBtNw5EUTP/JDUamWc6aWt7Wio+ctzTZUnUT9KOIkDeN07EVDLxxMw7EX+mGSSpKlK3lLK+YM19ZcS9c1WWFfAeCngVANAAAAAEBGtAwBAAAAACAjQjUAAAAAABkRqgEAAAAAyIhQDQAAAABARoRqAAAAAAAyIlQDAAAAAJARoRoAAAAAgIwI1QAAAAAAZESoBgAAAAAgI0I1AAAAAAAZEaoBAAAAAMiIUA0AAAAAQEaEagAAAAAAMiJUAwAAAACQEaEaAAAAAICMCNUAAAAAAGREqAYAAAAAICNCNQAAAAAAGRGqAQAAAADIiFANAAAAAEBGhGoAAAAAADIiVAMAAAAAkBGhGgAAAACAjAjVAAAAAABkRKgGAAAAACAjQjUAAAAAABkRqgEAAAAAyIhQDQAAAABARoRqAAAAAAAyIlQDAAAAAJARoRoAAAAAgIwI1QAAAAAAZESoBgAAAAAgI0I1AAAAAAAZEaoBAAAAAMiIUA0AAAAAQEaEagAAAAAAMiJUAwAAAACQEaEaAAAAAICMCNUAAAAAAGREqAYAAAAAICNCNQAAAAAAGRGqAQAAAADIiFANAAAAAEBGhGoAAAAAADIiVAMAAAAAkBGhGgAAAACAjAjVAAAAAABkRKgGAAAAACAjQjUAAAAAABkRqgEAAAAAyIhQDQAAAABARoRqAAAAAAAyIlQDAAAAAJARoRoAAAAAgIwI1QAAAAAAZESoBgAAAAAgI0I1AAAAAAAZEaoBAAAAAMiIUA0AAAAAQEaEagAAAAAAMiJUAwAAAACQEaEaAAAAAICMCNUAAAAAAGREqAYAAAAAICNCNQAAAAAAGRGqAQAAAADIiFANAAAAAEBGhGoAAAAAADIiVAMAAAAAkBGhGgAAAACAjAjVAAAAAABkRKgGAAAAACAjQjUAAAAAABkRqgEAAAAAyIhQDQAAAABARoRqAAAAAAAyIlQDAAAAAJARoRoAAAAAgIwI1QAAAAAAZESoBgAAAAAgI0I1AAAAAAAZEaoBAAAAAMiIUA0AAAAAQEaEagAAAAAAMiJUAwAAAACQEaEaAAAAAICMCNUAAAAAAGREqAYAAAAAICNCNQAAAAAAGRGqAQAAAADIiFANAAAAAEBGhGoAAAAAADIiVAMAAAAAkBGhGgAAAACAjAjVAAAAAABk9D9ec43hyLI6mwAAAABJRU5ErkJggg=="
    }
   },
   "cell_type": "markdown",
   "id": "44b6e29f-6c7e-4b17-9aa3-04d01ceac57a",
   "metadata": {},
   "source": [
    "# Guided Hunting - Anomaly detection with Isolation Forest on Windows Logon data\n",
    "\n",
    " <details>\n",
    "     <summary>&nbsp;<u>Details...</u></summary>\n",
    "\n",
    "__Notebook Version:__ 1.0<br>\n",
    "__Python Version:__ Python 3.8 - AzureML<br>\n",
    "__Required Packages:__  Msticpy, Msticnb, matplotlib, ipywidgets <br>\n",
    "__Platforms Supported:__  Azure Machine Learning Notebooks\n",
    "     \n",
    "__Data Source Required:__ Yes\n",
    "\n",
    "__Data Source:__ SecurityEvents\n",
    "\n",
    "</details>\n",
    "\n",
    "**Description**\n",
    "\n",
    "In this sample guided scenario notebook, we will demonstrate how to hunt for anamalous user logon activity using [Isolation forest](https://en.wikipedia.org/wiki/Isolation_forest) model. \n",
    "<br>We will start with reading historical windows logon data from Microsoft Sentinel workspace, then we will\n",
    "prepocess the dataset using series of data preparation steps such as aggregation, summarization, data type conversion, deriving new fields etc. Then we will perform [Feature Engineering](https://en.wikipedia.org/wiki/Feature_engineering) and select subset of features \n",
    "from the data prepared from previous steps to create isolation forest model. Finally, we will run the model to score the results and identify anomalies with higher score.\n",
    "\n",
    "<br>The isolation forest algorithm will split the data into two parts based on random threshold value. It will recursively continue the splitting until each data point is isolated. Then we will detect anomalies using isolation (how far a data point is to the rest of the data). \n",
    "To detect an anomaly the isolation forest takes the average path length (number of splits to isolate a sample) of all the trees \n",
    "for a given instance and uses this to determine if it is an anomaly (average shorter path lengths indicate anomalies)\n",
    "\n",
    "***Disclaimer:***\n",
    "Some of the sections in the Notebook such as PCA plot visualization , interpreting SHAP values , customizing anomaly algorithm score will require prior knowledge and understanding of data science algorithms and interpreting results or visualization etc which is typically known to Data Scientists. We have included one liner notes about the context and reasoning where possible but may need to refer other resources to grasp the concepts. \n",
    "\n",
    "![image.png](attachment:aff21132-09c0-40ea-a96b-555ac4d81aed.png)\n",
    "\n",
    "Image Credits: [Detecting and preventing abuse on LinkedIn using isolation forests](https://engineering.linkedin.com/blog/2019/isolation-forest)\n",
    "\n",
    "***Please run the cells sequentially to avoid errors.\n",
    "<br>Please do not use \"run all cells\".*** "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9119d174-cec1-49ce-820f-78a10279494b",
   "metadata": {},
   "source": [
    "## Notebook initialization\n",
    "The next cell:\n",
    "- Checks for the correct Python version\n",
    "- Checks versions and optionally installs required packages\n",
    "- Imports the required packages into the notebook\n",
    "- Sets a number of configuration options.\n",
    "\n",
    "This should complete without errors. If you encounter errors or warnings look at the following two notebooks:\n",
    "- [TroubleShootingNotebooks](https://github.com/Azure/Azure-Sentinel-Notebooks/blob/master/TroubleShootingNotebooks.ipynb)\n",
    "- [ConfiguringNotebookEnvironment](https://github.com/Azure/Azure-Sentinel-Notebooks/blob/master/ConfiguringNotebookEnvironment.ipynb)\n",
    "\n",
    "If you are running in the Microsoft Sentinel Notebooks environment (Azure Notebooks or Azure ML) you can run live versions of these notebooks:\n",
    "- [Run TroubleShootingNotebooks](./TroubleShootingNotebooks.ipynb)\n",
    "- [Run ConfiguringNotebookEnvironment](./ConfiguringNotebookEnvironment.ipynb)\n",
    "\n",
    "You may also need to do some additional configuration to successfully use functions such as Threat Intelligence service lookup and Geo IP lookup. \n",
    "There are more details about this in the `ConfiguringNotebookEnvironment` notebook and in these documents:\n",
    "- [msticpy configuration](https://msticpy.readthedocs.io/en/latest/getting_started/msticpyconfig.html)\n",
    "- [Threat intelligence provider configuration](https://msticpy.readthedocs.io/en/latest/data_acquisition/TIProviders.html#configuration-file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7ec5c212-8a70-4d47-b154-434326acee9d",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:49:29.176028Z",
     "start_time": "2023-02-24T11:49:18.345234Z"
    }
   },
   "outputs": [],
   "source": [
    "from pathlib import Path\n",
    "import os\n",
    "import sys\n",
    "import warnings\n",
    "import shap\n",
    "from IPython.display import display, HTML, Markdown, Javascript\n",
    "import matplotlib.pyplot as plt\n",
    "import plotly.express as px\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from mpl_toolkits.mplot3d import Axes3D\n",
    "from sklearn.ensemble import IsolationForest\n",
    "\n",
    "REQ_PYTHON_VER = (3, 6)\n",
    "REQ_MSTICPY_VER = (1, 7, 0)\n",
    "\n",
    "display(HTML(\"<h3>Starting Notebook setup...</h3>\"))\n",
    "\n",
    "# You may need to manually install msticpy with\n",
    "# %pip install msticpy[azsentinel]\n",
    "\n",
    "import msticpy as mp\n",
    "\n",
    "mp.init_notebook(namespace=globals())\n",
    "\n",
    "WIDGET_DEFAULTS = {\n",
    "    \"layout\": widgets.Layout(width=\"95%\"),\n",
    "    \"style\": {\"description_width\": \"initial\"},\n",
    "}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b1d5c8e5-71ee-4165-8692-263ebbc63a05",
   "metadata": {},
   "source": [
    "### Authentication to LA Workspace\n",
    " <details>\n",
    "     <summary>&nbsp;<u>Details...</u></summary>\n",
    "If you are using user/device authentication, run the following cell. \n",
    "- Click the 'Copy code to clipboard and authenticate' button.\n",
    "- This will pop up an Azure Active Directory authentication dialog (in a new tab or browser window). The device code will have been copied to the clipboard. \n",
    "- Select the text box and paste (Ctrl-V/Cmd-V) the copied value. \n",
    "- You should then be redirected to a user authentication page where you should authenticate with a user account that has permission to query your Log Analytics workspace.\n",
    "\n",
    "Use the following syntax if you are authenticating using an Azure Active Directory AppId and Secret:\n",
    "```\n",
    "%kql loganalytics://tenant(aad_tenant).workspace(WORKSPACE_ID).clientid(client_id).clientsecret(client_secret)\n",
    "```\n",
    "instead of\n",
    "```\n",
    "%kql loganalytics://code().workspace(WORKSPACE_ID)\n",
    "```\n",
    "\n",
    "Note: you may occasionally see a JavaScript error displayed at the end of the authentication - you can safely ignore this.<br>\n",
    "On successful authentication you should see a ```popup schema``` button.\n",
    "To find your Workspace Id go to [Log Analytics](https://ms.portal.azure.com/#blade/HubsExtension/Resources/resourceType/Microsoft.OperationalInsights%2Fworkspaces). Look at the workspace properties to find the ID.\n",
    "</details>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6a5f29bd-216e-4531-a475-adf7a8bd1194",
   "metadata": {},
   "outputs": [],
   "source": [
    "# See if we have a Microsoft Sentinel Workspace defined in our config file.\n",
    "# If not, let the user specify Workspace and Tenant IDs\n",
    "\n",
    "ws_config = WorkspaceConfig(workspace=\"Default\")\n",
    "if not ws_config.config_loaded:\n",
    "    ws_config.prompt_for_ws()\n",
    "\n",
    "qry_prov = QueryProvider(data_environment=\"MSSentinel\")\n",
    "# Authentication\n",
    "qry_prov.connect(ws_config)\n",
    "table_index = qry_prov.schema_tables"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "caedfba4-f532-41e5-8a7b-2d416fd62410",
   "metadata": {},
   "source": [
    "## Data Preparation\n",
    "In this step, we will prepare the Windows logon events and do some preprocessing before we do data modelling. For this case, we are primarily considering logon event ids 4624, 4625 with specific logon type. \n",
    "\n",
    "4624 and 4625 events are related to Successful sign in and Failed Sign-in. You can check more about the event Ids in below links.\n",
    "\n",
    "- [4624(S): An account was successfully logged on.](https://docs.microsoft.com/windows/security/threat-protection/auditing/event-4624)\n",
    "- [4625(F): An account failed to log on.](https://docs.microsoft.com/windows/security/threat-protection/auditing/event-4625)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7d1fff2d-7d23-4d01-ad90-a77282a38bd0",
   "metadata": {},
   "source": [
    "### Historical Data Processing\n",
    "For this model, we can consider upto 21 days of historical data. If you want to generate this anomalies on recurrent basis then depending on scale and volume of the data, you can set up intermediate pipeline to save historical data into custom table and load results from it. \n",
    "Check out the blog for ways to export historical data at scale using notebook [Export Historical Log Data from Microsoft Sentinel](https://techcommunity.microsoft.com/t5/microsoft-sentinel-blog/export-historical-log-data-from-microsoft-sentinel/ba-p/3413418)\n",
    "For this demo, we are retrieving data from the original table. We also have provided demo dataset if you want to test the notebook without connecting to your workspace."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0424ab25-f400-4045-96c4-fb4ddb0b7458",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:49:44.778556Z",
     "start_time": "2023-02-24T11:49:44.636324Z"
    }
   },
   "outputs": [],
   "source": [
    "win_agg_query = '''\n",
    "let start = 21d;\n",
    "let end = 1d;\n",
    "SecurityEvent\n",
    "| where TimeGenerated >= startofday(ago(start)) and TimeGenerated <= endofday(ago(end))\n",
    "| where EventID in (4624, 4625)\n",
    "| where LogonTypeName == \"3 - Network\"\n",
    "| extend DstDomain=TargetDomainName, DstUser=TargetUserName, Date = format_datetime(TimeGenerated, 'yyyy-MM-dd')\n",
    "| where not (EventID == 4624 and (DstDomain =~ \"NT AUTHORITY\" or DstDomain =~ \"ANONYMOUS LOGON\") and LogonType == \"3\")\n",
    "| where not (EventID == 4624 and (DstDomain =~ \"NT AUTHORITY\" or DstDomain =~ \"SYSTEM\") and LogonType == \"5\")\n",
    "| where not (EventID == 4624 and DstDomain =~ \"WINDOW MANAGER\" and DstUser startswith \"DWM-\" and LogonType == \"2\")\n",
    "| where not (EventID == 4624 and DstDomain =~ \"NT VIRTUAL MACHINE\")\n",
    "| where DstUser !endswith \"$\"\n",
    "| summarize AttemptedLogons = count() , MAX_TimeGenerated = max(TimeGenerated), MIN_TimeGenerated = min(TimeGenerated), SuccessfulLogons=countif(EventID == 4624), FailedLogons=countif(EventID == 4625), ComputerDomainsAccessed=dcount(TargetDomainName), DistinctAuthenticationPackage=dcount(AuthenticationPackageName), ComputersAccessed = dcount(Computer), DistinctSrcIp= dcount(IpAddress), DistinctSrcHostName = dcount(Workstation), ComputersSuccessfulAccess = dcountif(Computer, EventID == 4624),  ComputerDomainsSuccessfulAccess = dcountif(AccountDomain, EventID == 4624), SrcIpSuccessfulAccess = dcountif(IpAddress, EventID == 4624), SrcHostNameSuccessfulAccess = dcountif(WorkstationName, EventID ==4624) by Date, DstDomain, DstUser, LogonTypeName\n",
    "| extend ActiveMilliseconds = datetime_diff('millisecond', MAX_TimeGenerated, MIN_TimeGenerated)'''\n",
    "\n",
    "win_agg_df = qry_prov.exec_query(win_agg_query)\n",
    "\n",
    "# Uncomment below lines if you want to test the notebook with demo dataset\n",
    "# dataset_url = \"https://raw.githubusercontent.com/Azure/Azure-Sentinel-Notebooks/master/src/data/iforest-demo-data.csv\"\n",
    "# win_agg_df = pd.read_csv(dataset_url)\n",
    "# win_agg_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e425a713-0741-48fd-8aeb-ce2a1d7e633d",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:49:47.557120Z",
     "start_time": "2023-02-24T11:49:47.543978Z"
    }
   },
   "outputs": [],
   "source": [
    "#Displaying the columns of the data loaded from previous step.\n",
    "win_agg_df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9b61be6c-a82f-40fe-be80-3ad6908b7950",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:49:47.948219Z",
     "start_time": "2023-02-24T11:49:47.944294Z"
    }
   },
   "outputs": [],
   "source": [
    "# Displaying the no of records\n",
    "win_agg_df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b2a61cc7-adcb-44d4-a812-c305d15e3313",
   "metadata": {},
   "source": [
    "## Feature Engineering\n",
    "\n",
    "In this step, we are creating additional features/columns. \n",
    "\n",
    "We have selected 4 columns(features) with numeric data points \n",
    "- FailedLogons\n",
    "- SuccessfulLogons\n",
    "- ComputersSuccessfulAccess\n",
    "- SrcIpSuccessfulAccess\n",
    "\n",
    "and also deriving additional columns by calculating mean, standard deviation and zscores on each of them. Converting to zscores is not necessary for numerical features as Isolation forest are scale invariant but this pre-processing is done so as to use these features later in the visualizations such as PCA.  We have also done log scaling as part of data pre-processing steps which is not required but based on various data studies in production environment we have seen it gives finer results. You can skip or add this step based on data study and analyzing results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "92dc7b4e-f10d-462f-b605-c22148bcf48e",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:49:56.765633Z",
     "start_time": "2023-02-24T11:49:56.698175Z"
    }
   },
   "outputs": [],
   "source": [
    "def get_zscore(value, mean, std):\n",
    "    # calculate z-score or number of standard deviations from mean\n",
    "    if (\n",
    "        std == 0\n",
    "        or std is None\n",
    "        or str(std).lower() in [\"nan\", \"none\", \"null\"]\n",
    "        or mean is None\n",
    "    ):\n",
    "        if value == 0.0:\n",
    "            return 0.0\n",
    "        elif value != 0:\n",
    "            return np.log10(value + 1)\n",
    "    ans = (value - mean) / std\n",
    "    # only interested in increases\n",
    "    ans = max(0.0, ans)\n",
    "    # take log to dampen numbers\n",
    "    ans = np.log10(ans + 1)\n",
    "    return float(ans)\n",
    "\n",
    "\n",
    "data = win_agg_df.copy()\n",
    "\n",
    "zscore_columns = [\n",
    "    \"FailedLogons\",\n",
    "    \"SuccessfulLogons\",\n",
    "    \"ComputersSuccessfulAccess\",\n",
    "    \"SrcIpSuccessfulAccess\",\n",
    "]\n",
    "means = [x + \"_mean\" for x in zscore_columns]\n",
    "stds = [x + \"_std\" for x in zscore_columns]\n",
    "zscores = [x + \"_zscore\" for x in zscore_columns]\n",
    "\n",
    "ind = [\"DstDomain\", \"DstUser\", \"Date\"]\n",
    "\n",
    "zscore = data[zscore_columns + ind]\n",
    "zscore = zscore.fillna(0)\n",
    "\n",
    "# getting means for user, domain and logon type combination\n",
    "zscore[means] = zscore.groupby([\"DstDomain\", \"DstUser\"])[zscore_columns].transform(\n",
    "    \"mean\"\n",
    ")\n",
    "\n",
    "# getting standard deviation for user, domain and logon type combination\n",
    "zscore[stds] = zscore.groupby([\"DstDomain\", \"DstUser\"])[zscore_columns].transform(\n",
    "    \"std\", ddof=1\n",
    ")\n",
    "\n",
    "zscore = zscore.drop_duplicates([\"DstDomain\", \"DstUser\"])\n",
    "\n",
    "zscore = zscore[means + stds + [\"DstDomain\", \"DstUser\"]]\n",
    "\n",
    "data = data.merge(zscore, how=\"left\", on=[\"DstDomain\", \"DstUser\"])\n",
    "\n",
    "# Calculating z scores\n",
    "for column in zscore_columns:\n",
    "    data[f\"{column}_zscore\"] = data.apply(\n",
    "        lambda row: get_zscore(\n",
    "            row[f\"{column}\"], row[f\"{column}_mean\"], row[f\"{column}_std\"]\n",
    "        ),\n",
    "        axis=1,\n",
    "    )\n",
    "# Display top 10 record\n",
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9947753c-305b-4c16-934b-63d0b89e0859",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:50:00.039528Z",
     "start_time": "2023-02-24T11:50:00.025460Z"
    }
   },
   "outputs": [],
   "source": [
    "data.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "78b21fc5-fd6c-451f-a5e5-6f591b7e9335",
   "metadata": {},
   "source": [
    "## Data Modelling\n",
    "In this step we will specify features to be modelled and run isolation forest algorithm against the data."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "15f15cc7-72f0-4a01-aca4-de5831d1681c",
   "metadata": {},
   "source": [
    "###  Isolation Forest Anomaly detection\n",
    "\n",
    "In this step, we will select subset of features generated from previous step and use it for data modelling. We will also use Isolation Forest model on the data with selected features and calculate the anomalies. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5ba8ef8d-5586-49f9-9652-f8fc9b16eae1",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:50:05.673411Z",
     "start_time": "2023-02-24T11:50:04.476195Z"
    }
   },
   "outputs": [],
   "source": [
    "def apply_isolation_forest(df, n_estimators, contamination=0.01):\n",
    "    \"\"\"Applies Isolation Forest to a given dataset and returns the predicted anomalies.\"\"\"\n",
    "    clf = IsolationForest(\n",
    "        n_estimators=n_estimators,\n",
    "        max_samples=\"auto\",\n",
    "        contamination=contamination,\n",
    "        max_features=6,\n",
    "        bootstrap=False,\n",
    "        n_jobs=-1,\n",
    "        random_state=42,\n",
    "        verbose=0,\n",
    "    )\n",
    "    clf.fit(df.values)\n",
    "    pred = clf.predict(df.values)\n",
    "    scores = clf.decision_function(df.values)\n",
    "    return clf, pred, scores\n",
    "\n",
    "# specify the metrics column names to be modelled\n",
    "features = [\n",
    "    \"FailedLogons_zscore\",\n",
    "    \"SuccessfulLogons_zscore\",\n",
    "    \"ComputersSuccessfulAccess_zscore\",\n",
    "    \"ComputerDomainsSuccessfulAccess\",\n",
    "    \"SrcIpSuccessfulAccess_zscore\",\n",
    "    \"SrcHostNameSuccessfulAccess\",\n",
    "]\n",
    "\n",
    "data[features] = data[features].fillna(0)\n",
    "\n",
    "X = data[features].copy()\n",
    "if X.shape[0] < 500:\n",
    "    n_estimators = len(features) * 4 + X.shape[0] * 2\n",
    "else:\n",
    "    n_estimators = 100\n",
    "\n",
    "# n_estimators = 100\n",
    "print(\"Number of trees\", n_estimators)\n",
    "\n",
    "clf, pred, scores = apply_isolation_forest(X, n_estimators, contamination=0.01)\n",
    "data[\"anomaly\"] = pred\n",
    "data[\"score\"] = scores * -1\n",
    "# excluding users who do not have any successful logon history.\n",
    "data = data.loc[data[\"SuccessfulLogons\"] > 0]\n",
    "outliers = data.loc[data[\"anomaly\"] == -1]\n",
    "outlier_index = list(outliers.index)\n",
    "print(f\"Outliers at indexes: {outlier_index}\")\n",
    "# Find the number of anomalies and normal points here points classified -1 are anomalous\n",
    "print(data[\"anomaly\"].value_counts())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "52356ce5-7288-4fe2-98dc-bf86b0415450",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:50:07.909498Z",
     "start_time": "2023-02-24T11:50:07.886358Z"
    }
   },
   "outputs": [],
   "source": [
    "print(\"Number of outliers:\", outliers.shape[0])\n",
    "print(\"Top anomalies by score (top most has highest anomaly score and so on)\")\n",
    "outliers = outliers.sort_values(by=[\"score\"], ascending=False)\n",
    "display(outliers.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "592e9c28-6cb4-4d96-8071-53a74411d856",
   "metadata": {},
   "source": [
    "## Data Visualization\n",
    "In this step, we will explore various ways we can visualize the outliers identified from previous step."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eae742e7-5d49-4827-acaa-9706c53b7996",
   "metadata": {},
   "source": [
    "### Histogram"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9f4a2928-6050-48c3-ac03-38164bc14f4f",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:50:12.215932Z",
     "start_time": "2023-02-24T11:50:11.796016Z"
    }
   },
   "outputs": [],
   "source": [
    "data[\"anomaly_label\"] = data[\"anomaly\"].apply(\n",
    "    lambda x: \"outlier\" if x == -1 else \"inlier\"\n",
    ")\n",
    "fig = px.histogram(data, x=\"score\", color=\"anomaly_label\")\n",
    "fig.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1b787838-18ee-4cd2-b3ae-f0077e254929",
   "metadata": {},
   "source": [
    "### 3D ScatterPlot using PCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f83326b6-d039-44dc-b03e-d8b9f1da8137",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:50:18.460137Z",
     "start_time": "2023-02-24T11:50:18.378494Z"
    }
   },
   "outputs": [],
   "source": [
    "pca3 = PCA(n_components=3)  # Reduce to k=3 dimensions\n",
    "scaler = StandardScaler()\n",
    "# normalize the metrics\n",
    "X = scaler.fit_transform(data[features])\n",
    "X_reduce = pca3.fit_transform(X)\n",
    "\n",
    "\n",
    "# Add destination user to label\n",
    "data[\"labels\"] = np.where(\n",
    "    data[\"anomaly\"] == -1,\n",
    "    data[\"Date\"].astype(\"str\")\n",
    "    + \" \"\n",
    "    + data[\"DstDomain\"].astype(\"str\")\n",
    "    + \" DstUser \"\n",
    "    + data[\"DstUser\"].astype(\"str\"),\n",
    "    \"non-anomalous\",\n",
    ")\n",
    "\n",
    "total_var = pca3.explained_variance_ratio_.sum() * 100\n",
    "fig = px.scatter_3d(\n",
    "    X_reduce,\n",
    "    x=0,\n",
    "    y=1,\n",
    "    z=2,\n",
    "    color=data[\"labels\"],\n",
    "    title=f\"Total Explained Variance: {total_var:.2f}%\",\n",
    "    labels={\n",
    "        \"0\": \"Principal Component 1\",\n",
    "        \"1\": \"Principal Component 2\",\n",
    "        \"2\": \"Principal Component 3\",\n",
    "    },\n",
    ")\n",
    "fig.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "866cd36b-39d8-4a16-9f56-518712184735",
   "metadata": {},
   "source": [
    "### Correlation Plot\n",
    "\n",
    "Correlation plot gives idea about how differnt features are correlated to each other. Eg. You will observe linear correlation between FailedLogons with DistinctSrcIp/ DistinctSrcHostName etc. If there are similar correlations between multiple features you can experiment of removing it and see the imapct on outlier results. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "44b9cb6b-041d-4ba7-97c8-bc64dbaa07e8",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:52:23.844538Z",
     "start_time": "2023-02-24T11:52:23.383355Z"
    }
   },
   "outputs": [],
   "source": [
    "corrmat = win_agg_df[\n",
    "    [\n",
    "        \"DistinctAuthenticationPackage\",\n",
    "        \"SuccessfulLogons\",\n",
    "        \"FailedLogons\",\n",
    "        \"ComputersAccessed\",\n",
    "        \"ComputerDomainsAccessed\",\n",
    "        \"DistinctSrcIp\",\n",
    "        \"DistinctSrcHostName\",\n",
    "        \"ActiveMilliseconds\",\n",
    "    ]\n",
    "].corr()\n",
    "\n",
    "mask = np.zeros_like(corrmat, dtype=np.bool)\n",
    "mask[np.triu_indices_from(mask)] = True\n",
    "\n",
    "f, ax = plt.subplots(figsize=(50, 29))\n",
    "sns.heatmap(corrmat, vmin=-1, square=True, cmap=\"coolwarm\", annot=True, mask=mask);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "45072061-f90f-4f95-bb10-65c43b52fee0",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:52:28.062279Z",
     "start_time": "2023-02-24T11:52:28.047079Z"
    }
   },
   "outputs": [],
   "source": [
    "def get_redundant_pairs(df):\n",
    "    \"\"\"Get diagonal and lower triangular pairs of correlation matrix\"\"\"\n",
    "    pairs_to_drop = set()\n",
    "    cols = df.columns\n",
    "    for i in range(0, df.shape[1]):\n",
    "        for j in range(0, i + 1):\n",
    "            pairs_to_drop.add((cols[i], cols[j]))\n",
    "    return pairs_to_drop\n",
    "\n",
    "\n",
    "def get_top_abs_correlations(df, n=5):\n",
    "    au_corr = df.corr().abs().unstack()\n",
    "    labels_to_drop = get_redundant_pairs(df)\n",
    "    au_corr = au_corr.drop(labels=labels_to_drop).sort_values(ascending=False)\n",
    "    return au_corr[0:n]\n",
    "\n",
    "\n",
    "print(\"Top Absolute Correlations between features\")\n",
    "print(get_top_abs_correlations(corrmat, 10))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3a8d0f95-3686-4e62-b645-7a6c4a218c17",
   "metadata": {
    "tags": []
   },
   "source": [
    "## Interpreting Global and Local anomalies with SHAP\n",
    "\n",
    "Global interpretability and local interpretability are two different ways to understand how a machine learning model is making predictions. Global interpretability looks at the model as a whole and tries to understand which features are most important for the model's overall performance. Local interpretability, on the other hand, focuses on a single prediction and tries to understand which features are most important for that particular prediction\n",
    "\n",
    "SHAP values are generally the difference between the expected output and partial dependence plot at the features value. You can read more scientific details about SHAP values at the documentation.\n",
    "[Reading-SHAP-values-from-partial-dependence-plots](https://shap.readthedocs.io/en/latest/example_notebooks/overviews/An%20introduction%20to%20explainable%20AI%20with%20Shapley%20values.html#Reading-SHAP-values-from-partial-dependence-plots)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "615e878c",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:53:22.053798Z",
     "start_time": "2023-02-24T11:53:22.044272Z"
    }
   },
   "outputs": [],
   "source": [
    "def get_index(df, domain, user, date):\n",
    "\n",
    "    # Use the \"loc\" method to select rows that match the criteria\n",
    "    rows = df.loc[\n",
    "        (df[\"DstDomain\"] == domain) & (df[\"DstUser\"] == user) & (df[\"Date\"] == date)\n",
    "    ]\n",
    "\n",
    "    # Check if there are any matching rows\n",
    "    if len(rows) == 0:\n",
    "        print(\"No matching rows found\")\n",
    "        return None\n",
    "\n",
    "    # Return the index of the first matching row\n",
    "    return rows.index[0]\n",
    "\n",
    "# uncomment and replace user values based on your dataset and dates\n",
    "# outlier_index1 = get_index(data, \"CONTOSO\", \"SRVACC-04\", \"9/16/2022\")\n",
    "# outlier_index2 = get_index(data, \"CONTOSO\", \"SVC-ACC-02\", \"9/25/2022\")\n",
    "# print(\n",
    "#     f\"ourlier1 is at index: {outlier_index1} and \\noutlier2 is at index: {outlier_index2}\"\n",
    "# )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "45b52a9c-f13b-475c-b35b-88b9f9e33d15",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:53:26.323784Z",
     "start_time": "2023-02-24T11:53:25.385351Z"
    }
   },
   "outputs": [],
   "source": [
    "explainer = shap.TreeExplainer(clf)  # Explainer\n",
    "X = data[features].copy()\n",
    "shap_values = explainer.shap_values(X)  # Calculate SHAP values"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "16a61bc0-084d-437e-9631-069b00caf053",
   "metadata": {},
   "source": [
    "### Global Interpratation - Summary plot for Feature Importance\n",
    "Below we have 3 different visualization for global interpretation. \n",
    "<br>For this visualization , we will pass entire `shap_values` matrix instead of single entry.\n",
    "\n",
    "Here is a short description of these visualization types and how to interprete it.\n",
    "\n",
    " - ***Force plot:*** A SHAP force plot shows the contribution of each feature to the final prediction for a single data point. The plot has a horizontal axis that shows the SHAP value, which indicates how much a feature contributed to the prediction (positive values indicate that the feature contributed positively, and negative values indicate that the feature contributed negatively). The plot also has horizontal bars for each feature, which indicate the value of the feature for the data point being analyzed. The bars are colored to show whether the feature value is high (in red) or low (in blue). The width of each bar shows how much the feature contributed to the final prediction.\n",
    " - ***Bar plot:*** A SHAP bar plot, on the other hand, shows the feature importance for a set of data points. The plot has a horizontal axis that shows the mean SHAP value for each feature across the set of data points being analyzed. The plot also has vertical bars for each feature, which indicate the feature importance (how much the feature contributed to the prediction on average) and the direction of the effect (positive or negative). The bars are colored to show whether the feature had a positive (in red) or negative (in blue) effect on the prediction.\n",
    "\n",
    "To interpret the SHAP force plot or bar plot, you should look for features with high absolute SHAP values or feature importance. These are the features that have the greatest impact on the prediction. The direction of the SHAP value or feature importance indicates whether the feature has a positive or negative effect on the prediction. For example, a high positive SHAP value or feature importance for the \"number of failed login attempts\" feature indicates that a high number of failed login attempts is associated with a higher probability of being an anomaly."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "08f9df95",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:53:46.635679Z",
     "start_time": "2023-02-24T11:53:46.587358Z"
    }
   },
   "outputs": [],
   "source": [
    "shap.initjs()  # initialisation of Javascript to load visualization\n",
    "shap.plots.force(explainer.expected_value, shap_values, feature_names=X.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8341b58b-522e-4787-b2fa-6dcb4014c7da",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:53:52.686706Z",
     "start_time": "2023-02-24T11:53:52.464646Z"
    }
   },
   "outputs": [],
   "source": [
    "shap.summary_plot(shap_values, X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "344d5753-d970-4c1f-9132-77a80ce38daa",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:53:56.321852Z",
     "start_time": "2023-02-24T11:53:56.192759Z"
    }
   },
   "outputs": [],
   "source": [
    "shap.summary_plot(shap_values, X, plot_type=\"bar\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "20de9b2f-488a-4348-8d7a-5dd84a670d92",
   "metadata": {
    "tags": []
   },
   "source": [
    "### Local Interpretation - Feature importance\n",
    "Below we have 3 different visualization for local interpretation. \n",
    "\n",
    "For this visualization , we will pass  `shap_values` with index of outlier/inlier instead of whole matrix."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d1c51463",
   "metadata": {},
   "source": [
    "#### Example Outlier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "64a533c8-6974-4ea0-b140-fee1605fbf3d",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:54:40.954741Z",
     "start_time": "2023-02-24T11:54:40.948266Z"
    }
   },
   "outputs": [],
   "source": [
    "shap.force_plot(\n",
    "    explainer.expected_value,\n",
    "    shap_values[129],\n",
    "    features=X.iloc[129, :],\n",
    "    feature_names=X.columns,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "32335fa4",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:54:42.055337Z",
     "start_time": "2023-02-24T11:54:41.934482Z"
    }
   },
   "outputs": [],
   "source": [
    "shap.bar_plot(shap_values[129], features=X.iloc[129, :], feature_names=X.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "47e3868b",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:54:46.759121Z",
     "start_time": "2023-02-24T11:54:46.753268Z"
    }
   },
   "outputs": [],
   "source": [
    "shap.force_plot(\n",
    "    explainer.expected_value,\n",
    "    shap_values[103],\n",
    "    features=X.iloc[103, :],\n",
    "    feature_names=X.columns,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cae97122",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:55:12.235175Z",
     "start_time": "2023-02-24T11:55:12.088799Z"
    }
   },
   "outputs": [],
   "source": [
    "shap.bar_plot(shap_values[103], features=X.iloc[103, :], feature_names=X.columns)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ad330293-4f65-4264-b51a-18ee788f2b75",
   "metadata": {},
   "source": [
    "#### Example Inlier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "21ddc6b7-df4f-4caf-8d85-27f4cff439ec",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T11:54:53.273371Z",
     "start_time": "2023-02-24T11:54:53.266660Z"
    }
   },
   "outputs": [],
   "source": [
    "shap.force_plot(\n",
    "    explainer.expected_value,\n",
    "    shap_values[100],\n",
    "    features=X.iloc[100, :],\n",
    "    feature_names=X.columns,\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f9f62f6e-e049-4899-b3cf-2665ec3fa588",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T08:23:58.502847Z",
     "start_time": "2023-02-24T08:23:58.363524Z"
    }
   },
   "outputs": [],
   "source": [
    "shap.bar_plot(shap_values[100], features=X.iloc[100, :], feature_names=X.columns)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "531350cf",
   "metadata": {},
   "source": [
    "### Populate dataset with SHAP values\n",
    "\n",
    "Here we will apply SHAP to the data and concatenate SHAP values with the original dataframe so we will have SHAP values with each row in the dataframe for easier analysis without visualization if required."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "619e2759",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-02-24T08:24:02.541127Z",
     "start_time": "2023-02-24T08:24:02.314753Z"
    }
   },
   "outputs": [],
   "source": [
    "# Apply SHAP to the data\n",
    "shap_values = explainer(data[features])\n",
    "shap_df = pd.DataFrame(shap_values.values, columns=features)\n",
    "\n",
    "# Concatenate the SHAP values with the original dataframe\n",
    "result = pd.concat([data.reset_index(drop=True), shap_df], axis=1)\n",
    "result.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "98fcdc8e-5178-4a0d-a100-f70180e11ed4",
   "metadata": {
    "tags": []
   },
   "source": [
    "## Conclusion\n",
    "\n",
    "In this notebook, we started with windows event logs login data with the goal of finding users with anomalous login patterns. This notebook is targetted towards Data Scientists who can use it to tweak at various stages of the executions from Feature Engineering to Data visualization to explore the data in various ways. We have released another version of this Notebook , targetted towards SOC Analysts/Threat hunters who want to investigate the anomalies resulted from this model and triage to investigate any malicious activity."
   ]
  }
 ],
 "metadata": {
  "hide_input": false,
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.10"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": false,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": true,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "211.4px"
   },
   "toc_section_display": true,
   "toc_window_display": true
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}