PFC tutorial.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Fast-Higashi on Lee et al. dataset (sn-m3c-seq on PFC)"
   ]
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Preparation\n",
    "\n",
    "### Download input file\n",
    "\n",
    "Download the demo data from the following link\n",
    "https://drive.google.com/drive/folders/1SuzqQ_9dliAmTb-fGprFnN3aZrfWS-Fg?usp=sharing\n",
    "Please download the whole folder `pfc_500k`. This contains the cached input files.\n",
    "Change the file path in the corresponding JSON file according to the location of the downloaded files.\n",
    "\n",
    "To run Fast-Higashi for a new dataset, please prepare the same input files for the Higashi software. Use the Higashi software `higashi.process_data()` to transform contact pair files to sparse contact maps. The config JSON file you used for Higashi software will be the exact same config JSON file you will use for Fast-Higashi. It only contains one extra parameter `resolution_fh=[xx]` where `xx` could generally be the same as the resolution you used for scHi-C analysis.\n",
    "\n",
    "\n",
    "### Install Fast-Higashi\n",
    "1. install pytorch>=1.8.0 with cuda support when available.\n",
    "2. `conda install -c ruochiz fasthigashi`\n",
    "(Although fasthigashi would install pytorch when needed, there is no guarantee that it will install the correct version with cuda support. It is recommended to install pytorch separately before higashi.)"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Start running Higashi"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. Import package, set the path to the configuration JSON file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from fasthigashi.FastHigashi_Wrapper import *\n",
    "# Set the path to the configuration file, change it accordingly\n",
    "# The format of the config file is the same as the Higashi config JSON\n",
    "config = \"/work/magroup/ruochiz/fast_higashi_git/config_pfc.JSON\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Process data for fast-higashi model (pack from sparse mtx to tensors)\n",
    "The file you downloaded from the google drive above is the cached input (already packed into tensors)\n",
    "If you run the code on a new dataset, it will automatically pack the sparse mtx to the tensors format and cache it on the disk."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "setting to gpu:3\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6a984f694b3e484b92dbd8469a5b929a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "457 499\n",
      "pass 4236\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8fa8fbd3d7c347749b2322cb898b9590",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "481 487\n",
      "pass 4235\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d017fd7c6042477b99d337dafe0be7c1",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "391 397\n",
      "pass 4232\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a0cd6a336dea40bf824ba7346635634b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "378 383\n",
      "pass 4235\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ea4dd30aa1b24af4909cee6ff5dd62e9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "357 362\n",
      "pass 4232\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5c8433a8a1f04cbf87d502ba28a61faa",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "338 343\n",
      "pass 4229\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "983e066dbd944c1588c659c124783be2",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "314 319\n",
      "pass 4234\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8aa9523541f6487cbbc03b9205875c32",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "288 293\n",
      "pass 4235\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3a2cfe4bd9cb4b9a944cd2560c2a4d06",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "248 283\n",
      "pass 4231\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d3f3e04b7b124e1f89c50ea6c5cb9793",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "267 272\n",
      "pass 4234\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "82fadc7e594e41a883a644558626ac28",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "265 271\n",
      "pass 4232\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "043a3ec0411f42ec866a654a3458f093",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "263 268\n",
      "pass 4235\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "162fb53d3da44fd4879b2304e85965fc",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "193 231\n",
      "pass 4232\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "33142f0b5a3c44dbbb477b5ffbd2253a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "177 215\n",
      "pass 4224\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5b1854012d3e4f168aa43993534c8819",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "166 206\n",
      "pass 4231\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f1dee972f64b46eea4288e520adb860c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "160 181\n",
      "pass 4235\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0ebaca20894347fb85310674872cdcb8",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "158 163\n",
      "pass 4236\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f862254fdb93425ca34c32ccf2b17906",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "151 157\n",
      "pass 4229\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9ce16afae388407384edeed721b83b4a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "114 119\n",
      "pass 4234\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8ce086967b65450986c528f4c2d5d797",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "121 127\n",
      "pass 4234\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0759c2ea5c824aa89fb2443d28815ee2",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "74 97\n",
      "pass 4223\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e67ce2b18bce46efb4a23a0ce6017365",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "71 103\n",
      "pass 4235\n",
      "total_pass 4145.0\n",
      "good 4145 bad 93\n",
      "merge_fac_row 1\n",
      "cache file = /work/magroup/ruochiz/fast_higashi_git/pfc_500k/cache_intra_500000.pkl\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f4c4356230a346e09f4181a8faac9837",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr1 (499, 499)\n",
      "457 out of 499 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c932048cf4cc4919880353ec72165789",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(457, 457, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4a337d99624a42fd96b2ccdc7d25e439",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr2 (487, 487)\n",
      "481 out of 487 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a5403ef6da8c447aa2477ca7b3b5ae7b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(481, 481, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "751f9bddd3c64cea91ec5532bdf79e82",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr3 (397, 397)\n",
      "391 out of 397 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a2794c3efe6d46738724f1d937868ac0",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(391, 391, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "05e94a12ed5046139eeb82e6f7e6d236",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr4 (383, 383)\n",
      "378 out of 383 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6a1474408b82417286d910206d84c54f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(378, 378, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6e82e983c13847a79baa3cd67e6c5d5d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr5 (362, 362)\n",
      "357 out of 362 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "be3e37f0ce514c90b69ac6a2feee13cb",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(357, 357, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2d65eb67be9e46dab195222879983a91",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr6 (343, 343)\n",
      "338 out of 343 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1e319dfc51064bb5ac66b0527f3ffb2e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(338, 338, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4e654633f7c149d687bd68e76a7e7d04",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr7 (319, 319)\n",
      "314 out of 319 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "207025772d5a474c906c0178fdc95749",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(314, 314, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3465368f73e94fbf9d75cb98c69e5a03",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr8 (293, 293)\n",
      "288 out of 293 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8258bb57805540eaa88870fa1e0957ea",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(288, 288, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "91a7943b99324471add7d93fe50083b0",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr9 (283, 283)\n",
      "248 out of 283 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "57f4aa8155e542b094a766f9678d03c2",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(248, 248, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "386a507acede4cffb56e896ea466b044",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr10 (272, 272)\n",
      "267 out of 272 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "77bb3dab95c34f83acf1ba7f25e5466b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(267, 267, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a33523b3adbb46e3aed19fd89f0fce76",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr11 (271, 271)\n",
      "265 out of 271 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "17a660596c0648a7b5d19e8a852e075d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(265, 265, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ec1501769c68458dabdb02f8b1fb0bf9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr12 (268, 268)\n",
      "263 out of 268 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a27f966154e7434e98753962000f053b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(263, 263, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "daa9ee78fcad44f782a25cf19352d9f4",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr13 (231, 231)\n",
      "193 out of 231 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "725b12eff68346069f0e0cc33ef8d332",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(193, 193, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8abf2087e2104322ab348181e974e747",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr14 (215, 215)\n",
      "177 out of 215 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e73d608244c54393bfbc42a4c7eddb3a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(177, 177, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5e52d66ca0b94992b1bce1801c9a030a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr15 (206, 206)\n",
      "166 out of 206 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c623ba576b1c4af5841af96b7a03bdaf",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(166, 166, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "119e54fdf7764057830825a13608633c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr16 (181, 181)\n",
      "160 out of 181 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e8dd37a3845845e6b4d7c519795be435",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(160, 160, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e71f4b9288174b6ca329912b03fa4757",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr17 (163, 163)\n",
      "158 out of 163 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0485ea92c825495288d20b62e8022811",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(158, 158, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "48cd07e1125046d9b9c7c59d9d0143a4",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr18 (157, 157)\n",
      "151 out of 157 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ffb94864c8324d28b0dd0f811eb274b1",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(151, 151, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "69c8199df237425d9eabf6f038696bb4",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr19 (119, 119)\n",
      "114 out of 119 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1d0c8e4b0aec428e853fb1440d7963d7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(114, 114, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e519492e122b4e1aa6347ae625a87cd1",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr20 (127, 127)\n",
      "121 out of 127 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9eb193474a6b4c0dac534e7f4b328dba",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(121, 121, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "bb93a0bbd6504c0786056cc8c8c40540",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr21 (97, 97)\n",
      "74 out of 97 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6c05f5cbc2fb40baaf3458312792d64e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(74, 74, 4238) False\n",
      "no black list\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "12257038b51840b58228d0c50e7b3dd0",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "chr22 (103, 103)\n",
      "71 out of 103 bins are valid\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a1e22a75008741acb1660548fd72075d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/4238 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(71, 71, 4238) False\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "33a2e762726741bda26e63c90e3af4d5",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "saving cached input to /work/magroup/ruochiz/fast_higashi_git/pfc_500k/cache_intra_500000.pkl\n",
      "4238\n",
      "sparsity 0.05111497517410393\n",
      "do_conv False do_rwr False do_col False\n",
      "recommend_bs_cell [4238, 4238, 4238, 4238, 4238, 4238, 4238, 4238, 4238, 4238, 4238, 4238, 4238, 4238, 4238, 4238, 4238, 4238, 4238, 4238, 4238, 4238]\n",
      "[[457 315]\n",
      " [481 321]\n",
      " [391 298]\n",
      " [378 326]\n",
      " [357 319]\n",
      " [338 313]\n",
      " [314 305]\n",
      " [288 296]\n",
      " [248 324]\n",
      " [267 289]\n",
      " [265 289]\n",
      " [263 288]\n",
      " [193 297]\n",
      " [177 289]\n",
      " [166 283]\n",
      " [160 280]\n",
      " [158 279]\n",
      " [151 276]\n",
      " [114 314]\n",
      " [121 321]\n",
      " [ 74 274]\n",
      " [ 71 271]] 5432\n"
     ]
    }
   ],
   "source": [
    "# Initialize the model\n",
    "model = FastHigashi(config_path=config,\n",
    "\t             path2input_cache=\"/work/magroup/ruochiz/fast_higashi_git/pfc_500k\",\n",
    "\t             path2result_dir=\"/work/magroup/ruochiz/fast_higashi_git/pfc_500k\",\n",
    "\t             off_diag=100,\n",
    "\t             filter=False,\n",
    "\t             do_conv=False,\n",
    "\t             do_rwr=False,\n",
    "\t             do_col=False,\n",
    "\t             no_col=False)\n",
    "\n",
    "# Pack from sparse mtx to tensors\n",
    "model.prep_dataset(batch_norm=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3. Run the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dim1_0.6_rank_256_niterp_1_\n",
      "time elapsed: 0.01\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "604749891fb74c8ca7632427cc2337a8",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "initializing params:   0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rwr iters: [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]\n",
      "time elapsed: 42.02\n",
      "finish init\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "645061f9c2954745b1d7051c11313453",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Starting iteration 0\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "59030c80d82d4e7d9fcda0506d04967b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "424d9d856e724d9c987dca13112936dd",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; "
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "cd30eb98cf27489fa2c43453dd5bc996",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "PARAFAC2 re=5.575 takes 19.1s\n",
      "Starting iteration 1\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b3903709e35d4406acc6e02ef9f10985",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5bcbb367add14385996bbff501efe47a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; "
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ad490fca855e4ae1b39addcf891520e3",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "PARAFAC2 re=0.452 9.93e-01 variation min9.7e-01 at chrom 20, max1.0e+00 at chrom 1 takes 16.9s\n",
      "Starting iteration 2\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5f66e0ad35874d46a76c15913703caf0",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "12b0501e6ed04c86b8867ff2f54e6b32",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; "
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "213ed21fcb624f2eb547c87821e5cccf",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "PARAFAC2 re=0.443 3.55e-02 variation min1.7e-02 at chrom 17, max4.5e-02 at chrom 5 takes 18.5s\n",
      "Starting iteration 3\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "fd049b546c26413db0eaaeeda341668a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "176eb05189d94617b7944b9e40e360d7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; "
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "96fd61fad1494d028b552b33b1edbf8b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "PARAFAC2 re=0.434 4.25e-02 variation min2.7e-02 at chrom 21, max5.1e-02 at chrom 2 takes 17.5s\n",
      "Starting iteration 4\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "79e6eac83cf14c6f954cb1c0fddddc17",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f71da71b606f49caa9fe1ad74f0daf71",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; "
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "670df150448147c7959c4fe399406bd8",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "PARAFAC2 re=0.430 1.69e-02 variation min1.5e-02 at chrom 3, max2.7e-02 at chrom 7 takes 17.6s\n",
      "Starting iteration 5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "625255c89d3746969610136194db7e21",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "101b4c59a97a4d2d9efe6393c39d98b9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; "
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "69be9b8aadb345ada02203836f9afd85",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "PARAFAC2 re=0.429 6.63e-03 variation min4.8e-03 at chrom 19, max1.1e-02 at chrom 7 takes 17.0s\n",
      "Starting iteration 6\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c485e3c039c04312b6d93dc66448e120",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6a6dcc894b814326a4d3663f1e463122",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; "
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "77254b60ed00453ca8409d06d60cc1bb",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "PARAFAC2 re=0.428 3.38e-03 variation min2.0e-03 at chrom 19, max5.6e-03 at chrom 12 takes 16.7s\n",
      "Starting iteration 7\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "42740e9e65024aebba321cd067d6f024",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0ac62c4e2afd4eccb8f32192f601e384",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; "
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3867fbeacb764dac82aabf3c6896b989",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "PARAFAC2 re=0.428 2.06e-03 variation min1.2e-03 at chrom 19, max3.5e-03 at chrom 12 takes 17.9s\n",
      "Starting iteration 8\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e1c65a9f22c542eab618ce13097d624d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f6032f22168a41b09198b1a3a7710c6b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; "
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5a13001013c54c8fb3d8e03ea016774a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "PARAFAC2 re=0.427 1.39e-03 variation min9.0e-04 at chrom 19, max2.4e-03 at chrom 13 takes 16.6s\n",
      "Starting iteration 9\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2e26853b18794e64a001922418bb1239",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "46317f5f41374b798567d9fe90276956",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; "
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f2dcf64d9a6c4c25a23a7cd418ec0d03",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "PARAFAC2 re=0.427 1.01e-03 variation min6.8e-04 at chrom 4, max1.8e-03 at chrom 13 takes 16.4s\n",
      "Starting iteration 10\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e95a9cacfa8040b1b9002c14fab1897f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ce5f2c054ce04664b0d1728c49947bed",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; "
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ae4b64d422b049c98aaa784e120fe1d0",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "PARAFAC2 re=0.427 7.66e-04 variation min5.0e-04 at chrom 4, max1.3e-03 at chrom 13 takes 16.1s\n",
      "Starting iteration 11\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "22d3a3cf65e04c029f2997a77ff4e107",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "41437ee4dccf45b587b085bf793b110a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; "
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f49fb37bc71b4f7fbccbc8e4bd4c9783",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "PARAFAC2 re=0.427 6.04e-04 variation min3.9e-04 at chrom 4, max1.0e-03 at chrom 13 takes 17.3s\n",
      "Starting iteration 12\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "07254f6dc3fd40c5ac85c228a0eb8a02",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5d9fa7ef23cd4d05affacf4c5b19d247",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; "
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e6e64df1a29d46eba6f6833957a1d97e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "PARAFAC2 re=0.427 4.88e-04 variation min2.9e-04 at chrom 18, max7.9e-04 at chrom 13 takes 17.0s\n",
      "Starting iteration 13\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a6c4560b53da46d183dcf1d7f66e6afe",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "483898b490934e2ab755aec836fa85f3",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; "
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b033bd1fe24a452bab5b6891fab5530e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "PARAFAC2 re=0.427 4.08e-04 variation min2.1e-04 at chrom 18, max6.2e-04 at chrom 13 takes 16.5s\n",
      "Starting iteration 14\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "bb89f7465eec45ed95f67932f1a74ae1",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "594c736e4fa84b01b1bece73f533db2a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; 1, nan; "
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9a0e87cd1b61423bb50a4b84fd96d7df",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "PARAFAC2 re=0.427 3.50e-04 variation min1.6e-04 at chrom 18, max5.2e-04 at chrom 14 takes 16.4s\n",
      "converged in 14 iterations.\n",
      "start transform\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "014f34c8d39344b68262c8f7965933a7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/22 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "takes: 300.91 s\n"
     ]
    }
   ],
   "source": [
    "model.run_model(dim1=.6,\n",
    "                rank=256,\n",
    "                n_iter_parafac=1,\n",
    "                extra=\"\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Finished in 5mins on gpu! Let's see how it looks!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4. Visualize embedding results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABC4AAAKwCAYAAABXkz6vAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nOzdeXxTVd4/8M9N0jRN93RJN0j3lbZCoci++VNUUEQdZeuIo07FB3hk8XFGRIYZx10ZUAbHERxAEDdAGEDF0YKgYFGW0pVKCy0t3femWe79/VHKsBRoSdqk7ef9evXV5txzz/kGwZx+71kESZJARERERERERGSPZLYOgIiIiIiIiIjoWpi4ICIiIiIiIiK7xcQFEREREREREdktJi6IiIiIiIiIyG4xcUFEREREREREdouJCyIiIiIiIiKyWwpbB9Aeb29vKTg42NZhEBH1SEeOHKmQJMnHwjZ8FQrFPwEMQN9JcosAMkwm02NJSUlltg6Gbh7HEUREluFYos/oMWMfu0xcBAcHIz093dZhEBH1SIIgFFrahkKh+Kefn1+Mj49PtUwmk6wRl70TRVEoLy+PLS0t/SeAe2wdD908jiOIiCzDsUTf0JPGPsx8ERFRewb4+PjU9aWBhkwmk3x8fGrR+mSIiIiILNPnxhI9TU8a+zBxQURE7ZH1xYHGhffMz0YiIiLL9cmxRE/TU8Y+dh8gERH1XevXr/cQBCHpl19+UV2v3rPPPuvXXTERERERUfdi4oKIiOzWRx99pBk0aFDDhg0bNNert3LlSv/2ykVRhNls7prgiIiIyK7l5OQoIyIi4jpaf+XKlV4FBQUON6qTkpLSv6Ntzp07N9DPzy9BrVYPvLQ8Ly9POXTo0MiYmJjYyMjI2C1btrgDQGlpqXzo0KGRarV64KX91NfXy8aOHRseEhISFx4eHjdnzpzAtmu7d+92iY2NjVEoFEnr1q3zvLSf1NTUoPDw8LjQ0NC4Rx55pJ8oih0N3a4wcUFERBarMGzTZDZOiT/eMCops3FKfIVh23UTDR1RW1srS09Pd1m3bl3B1q1bPQGgsLDQYfDgwVHR0dGxERERcXv27HGZM2dOYEtLiyw6Ojr2nnvuCcnJyVGGhobGzZw5s39cXFxsfn6+8t1339VERkbGRkRExD355JOBN+qbiIiIutfGHws1yS/ujQ959t9JyS/ujd/4Y6HFY4lOx7Bxo/eZM2eum7jorClTptQcOnQo68rypUuX+k+dOrU6Kysrc/Pmzb8uWLCgPwCo1Wpp+fLl55YtW1Z05T0LFy48f/r06ZMZGRmZhw4dcvn444/dACA0NNSwbt26gsmTJ1deWv/rr792Pnz4sEt2dvbJ3Nzck0ePHnXetWuXqzXfX3dh4oKIiCxSYdimKTGs0pmkSiUAmKRKZYlhlc7S5MWHH37oMXbs2NqEhIQWDw8P8/fff69eu3atZsKECbXZ2dmZWVlZJ4cOHdq0evXqYkdHRzE7Ozvziy++OA0ABQUFqtmzZ1dmZWVlKpVKadmyZYHfffddbmZm5slffvnFecOGDR7WeO9ERERkuY0/Fmr+vDNTV1bfopQAlNW3KP+8M1NnjeSFyWTC1KlTgyMjI2MnTpwYWl9fL1u0aJH/gAEDYiIiIuKmTZumE0UR69at88zIyFCnpKSERkdHxzY0NAhpaWnqgQMHRkdFRcXGx8fHVFdXywCgtLTUYdSoURE6nW5Aampq0PX6nzBhQqNOpzNeWS4IAurq6uQAUF1dLff19TUCgJubm3jHHXc0qFSqy6ZGuLq6ipMnT64HAJVKJSUkJDSdPXtWCQBRUVGGoUOHNstksqv6aGlpEfR6vdDc3CwzmUxCQEDAVbH0BExcEBGRRcqMHwRKMFz2eSLBICszfmDRzIaPP/5YM23atGoAuP/++6s2bNigufXWWxs3b97svWDBgoDDhw87eXp6tjvf0d/f3zBhwoRGAPj++++db7311vqAgACTg4MDHnrooaq0tDQXS2IjIiIi61n5TV5gi0m8bCzRYhJlK7/Js3iWZEFBgSo1NbU8Nzc309XVVXzttdd8Fi9eXJaRkZGVl5d3srm5WfbRRx+5z549u3rAgAFN69ev/zU7OztToVBgxowZYStWrDiTk5OTmZaWluPi4iICQGZmpnrbtm2/ZmVlnfziiy88T5061elZGi+99NK5Tz75RKPVahOmTp0asXLlyjMdvbeiokL+9ddfe9x5551116t32223NY4YMaLe398/MSAgIGHcuHF1gwYN0nc2VnvAxAUREVmkbaZFR8s7orS0VP7jjz+6PfXUU7rAwMD4t99+2++LL77wvOOOOxr27duXExgYaHjkkUdC3n77ba/27ler1RcTGpLEDc2JiIjsWXl9S7tjhmuVd4afn5/h9ttvbwSAWbNmVR48eNBl9+7drgkJCdGRkZGxBw8edM3IyHC68r7jx4+rfH19jWPGjGkCAI1GIzo4tOYnRo4cWefl5WVWq9VSeHi4Pj8/37Gzca1bt04zbdq0yvPnzx///PPP8x555JGQjuzLZTQaMXXq1NAnnnjifGxsrOF6dTMyMhxzc3NVRUVFx4uKio7v37/fdffu3T3y4Q0TF0REZBGF4NXuh+a1yjtiw4YNnlOnTq08d+7cieLi4hOlpaXHg4KCDLt373YJDAw0Lly4sGLmzJkVP//8sxoAFAqF1NLSIrTX1ujRoxsPHTrkWlJSojCZTPjkk080Y8eObbjZ2IiIiMi6fFwd2x0zXKu8MwRBuOr1woULdZ9//nl+bm5u5syZMyv0ev1VvxdLkgRBENp9+qFUKi+Wy+VyyWg0tjsGuZ6NGzd6z5o1qwponRnR0tIiKy0tVdzovunTpweHhobqly5dWnajulu2bPEYMmRIo7u7u+ju7i7edttttQcOHHDubKz2gIkLIiKyiK/DI8UClJct2RCgFH0dHim+2TY/+eQTr6lTp1ZfWnbvvfdWP/HEEyGxsbFxMTExsdu3b/d85plnzgPAjBkzymNiYmLvueeekCvb0ul0xqVLlxaPGTMmMiYmJi4hIaFp5syZNTcbGxEREVnXvAkRxY4K2WVjCUeFTJw3IeKmxxJtSkpKlHv37nUGgE2bNmmGDx/eAAB+fn6m2tpa2Y4dOy6ewuHi4mKura2VA0BiYqL+/PnzyrS0NDUAVFdXy4xG620PERAQYNi1a5cbAPz8888qg8Eg+Pv7m653z7x58wLq6urk77///tmO9NG/f3/DgQMHXI1GI1paWoQDBw64xsbG9silIjfM6BBR7yCJIiAIV2WdiSzlrZxSBbTudWGSKpUKwcvg6/BIcVv5zTh8+HDOlWVLliwpW7JkSbtPF/7+978XA7g4uMnLyzt56fXU1NSq1NTUm46HiIgAmI2A3KoHLhABAGbeqqsCWve6KK9vUfq4OhrmTYgobiu3RGhoqH7t2rVec+bM0YWEhLQsWrSovLq6Wh4bGxsXFBRkSExMbGyrm5KSUjF37lzd4sWLxfT09KwPP/wwf968ef31er1MpVKJ+/bty+1s/6mpqUFbt27V6PV6mVarTZgxY0bFm2++ee6tt946+/jjjwe/8847WkEQsGbNmoK2zTUDAwPjGxoa5EajUfjyyy89du3alevh4WFetWqVf0hIiD4uLi4WAJ544omyBQsWVKSlpal/85vfhNfV1cm/+eYbjxdffDHg1KlTJ2fPnl397bffukVFRcUJgoBx48bVTp8+vdbSP1NbEOxx7e/gwYOl9PR0W4dB1CuIDQ0w/ZwOmEyAgwMQEAiFfwBkarWtQ6MuIgjCEUmSBlvSxrFjxwoSExMrrBVTT3Ls2DHvxMTEYFvHQTeP4wgiK8v9Cn/b+zR2OYq4XR2CIbG/QULUFHioeEBTb8WxRN/SE8Y+nHFB1EtJzc0wnsoFysv/W2g0AoUFMJ0phDw+AXJvH9sFSERERPbtzI/ISfsLUgz5aFLLAcjxob4Q//7pFXif+AdW3/spvF0DbB0lEfUBTFwQ9VLGvByg4hpJbkmCOfMkhPhECCoVBKerNlImIiKiPm7r3ufwub4QTWrVxTJ3owl/raqGUqrA11veRvTkaQj16A93R3cbRkpkewkJCdEGw+XHw69fv/50cnJys61i6k2YuCDqrSorr3/dZIIp4zgEtTMUAwdBkHGvXiIiImpVUPMr1oqlKLiQtPAwm1Erl6PSQYF17m6Y2NCIrYZvUb//GJIDBuNPw/9k44iJbOv48ePZto6hN2Pigqi3UjoCLTfYNNhoBFxdAG7YSURERJfo56bD4GYZClwlhBiMgAAk6g1Ic3bCQScVDqocAaEeXs1mxGhibB0uEfVyfMRK1Fs5d2zzTamqCva4SS8RERHZjlwmx3E3XwBAs0yADICbKOL31bXwNZkuPvSoNDdDceYnG0ZKRH0BExdEvZTQ0Y03m5pg+uXnrg2G6CbI5fKk6Ojo2LavP/7xj36duT8wMDC+pKSEMwuJiG5SkE/roRKlCgXyHRzgI5oxrrkZdzY2XVJLwp/OfYkTGVtsEyQR9Qkc0BH1UnIXF5g6Wrm+ritDIbopjo6OYnZ2dqat4yAi6qtClL9DaO1h/OreekLZVhdnVMlk+FZ99abe32fvR/yAh7o7RCLqIzjjgqiXknl4QtB4dayyKHK5CFnkxzVrNH8NDIj/g1yW9NfAgPgf16zRdFVfgYGB8U8//XRAbGxsTGRkZOwvv/yiAoDS0lL5iBEjImJiYmKnT5+uu/Tv9LJly7QRERFxERERccuXL/ftqtiIiHqT1DFhaPJIhCC1LguplsuxzdUFtXL5ZfXcTSLyTkfbIkSi68rJyVFGRETEdbT+ypUrvQoKChxuVCclJaV/R9ucO3duoJ+fX4JarR54aXlubq5y2LBhkZGRkbHJyclR+fn5F/tdtWqVl06nG6DT6QasWrXq4oB+1KhREVFRUbHh4eFx06dP728ytT6mPH/+vHz48OEROp1uwPDhwyPKy8vlALBz505XV1fXW9pmry5atMi/o3HbGyYuiHopSZIAhaJjG2/6+ELgBp10k35cs0bz7wVP6+pLSpSQJNSXlCj/veBpnaXJi5aWFtmlS0Xee+89z7Zr3t7epszMzKxHH320/OWXX9YCwLPPPhswbNiwhqysrMx77rmnpqSkRAkA+/fvV2/atMnryJEjWenp6Vnr16/3OXDgAM8AJiK6gRaxAT6aajiYfeDXcu1xwrwKX0QMmdiNkVGv9NP7GrweGY9lHkl4PTIeP73fZQ9BrmXjxo3eZ86cuW7iorOmTJlSc+jQoawry+fPnx80ffr0ytzc3MwlS5acW7hwYRDQmoR45ZVXAg4fPpyVnp6e9corrwS0JSK2b9+en5OTk5mbm3uysrLSYe3atZ4A8MILL/iPHTu2vrCwMGPs2LH1S5cuvbi8dvDgwQ3Z2dmZ2dnZma+//nqJNd9bd2Ligqi3EkVIjQ2tiQvFdVaFqVRQxid0X1zU6/znz8sDTXr9ZZ8nJr1e9p8/Lw+0pN22pSJtX48//nh127Xp06dXA0BycnLT2bNnHQHgxx9/dH300UcrAeDhhx+udXNzMwPAd99953LXXXfVuLm5ie7u7uLdd99d/e2337paEhsRUV9Q2liKiqZ6QPQDqucDuDp5IUDAqaFD8OTY8O4PkHqPn97X4Ms/6NBwXglIQMN5Jb78g84ayQuTyYSpU6cGR0ZGxk6cODG0vr5etmjRIv8BAwbERERExE2bNk0niiLWrVvnmZGRoU5JSQmNjo6ObWhoENLS0tQDBw6MjoqKio2Pj4+prq6WAUBpaanDqFGjInQ63YDU1NSg6/U/YcKERp1OZ7yyPC8vz+muu+6qA4BJkybV79271wMAtm3b5j569Og6rVZr9vHxMY8ePbru888/dwcAjUYjAoDRaBSMRqPQ9uBxz549Hr///e8rAeD3v/995e7duz2v7K+nY+KCqJcS5HIILi6AJAEyOeBwjeSx0hGSKHZvcNSr1JeWKjtTbg0qlUoCAIVCIZlMposjaZns6o81LoMiIro5UZooKKrvR33xZHi4x8JBcsd/kxcCIAiQC4CTU7c/GKfeJu2VQJhaLv8QN7XIkPaKRQ9BAKCgoECVmppanpubm+nq6iq+9tprPosXLy7LyMjIysvLO9nc3Cz76KOP3GfPnl09YMCApvXr1/+anZ2dqVAoMGPGjLAVK1acycnJyUxLS8txcXERASAzM1O9bdu2X7Oysk5+8cUXnqdOner0LI2YmJimTZs2eQLAhg0bPBobG2WlpaXy4uJih6CgIENbvcDAQENxcfHF9keOHBnh4+OT6OzsbJ49e3Y1AFRWVirakiM6nc5YVVV18anlL7/84hIVFRU7evToiPT0dNXN/0naFhMXRL2YPDyydbaFTAB82jllxMEBaG6C8WQGf7mjm+bq52foTHlXufXWW+vXrl3rBQAff/yxW11dnRwAxo8f37Br1y6P+vp6WV1dnWzXrl2e48aNq+/O2IiIeqr5I8fDSeaF+hYT4j0mXCi9kLyQ5NAo/fFj3naUH1hhsxipF2goa/9hx7XKO8HPz89w++23NwLArFmzKg8ePOiye/du14SEhOjIyMjYgwcPumZkZFy1hPT48eMqX19f45gxY5qA1tkODhceBI4cObLOy8vLrFarpfDwcH1+fr5jZ+NatWpV0f79+11jYmJiv/vuO1dfX1+jg4NDu2PyS5d0f//993mlpaXHDAaDbMeOHW7X62P48OGNhYWFx3NycjKfeuqpsvvvv7/HTo1i4oKoF5M5OkIWFASYzEBl5dUVjMbWr4pyiEVnuz9A6hXGP7+0WKFSXTZtR6FSieOfX1psSbtX7nExZ86c6z51efnll88dOHDAJTY2NubLL7909/f3NwDAyJEjm6ZPn145aNCgmKSkpJhZs2aVjxgxotmS2IiI+opx0VqMjvBGaW0zztQX4mLSwuwM0eSMMv156PWNmJu/GUfLjto0VurBXHzbf9hxrfJOuHIfN0EQsHDhQt3nn3+en5ubmzlz5swK/RVLXoHWGZuCILT7ZE+pVF4sl8vlktFo7PRmccHBwcavvvoqPysrK3PFihXFAODl5WUOCgoyFhUVXUzYFBcXKwMCAi5baqJWq6VJkybVbN261ePCfabCwkIHACgsLHTQaDQmoDXZ4u7uLgLAQw89VGsymYSeelQ8ExdEvZ3BBJiMQEvLtesIAiCXw9wioSZHhL6Ssy+o425NTa26+823Cl39/Q0QBLj6+xvufvOtwltTU6ssaddsNh+5dI+L1atXFwNAcXHxCX9/fxMAjB49uunw4cM5AODn52c+cOBAXmZmZtb7779/9ty5cxfrLVu27HxeXt7JvLy8k0uXLi2z9D0TEfUVShnweP1qPGdajfOlXoAogwBAiSbIZM2QBBElkgdKBE84yByQUZGBv/z4F+TX5Ns6dOpJxvxfMRSOl69dVjiKGPN/Fj0EAYCSkhLl3r17nQFg06ZNmuHDhzcAgJ+fn6m2tla2Y8eOi/tBuLi4mGtra+UAkJiYqD9//rwyLS1NDQDV1dUyo/GqrSosiUthNpsBAEuWLPGfNm1aBQBMmTKlNi0tza28vFxeXl4uT0tLc5syZUptbW2trC05YTQasWfPHvfo6OhmALjjjjtq3n33XS8AePfdd70mTpxYAwBnzpxRiBeWhH/77bdqURSh1WpNVnsT3ahHZluIqGMkgwFi6blrVxCE1j0wRBFSQwOMCsBQKwESoPLiKSPUcbemplZZmqggIiL7cyi/FA4lJxAumDBaXoMfBRNEAG4QMaAyBvs0lTArauGhb4S2thSfNuThRMUJhJWGIcwjzNbhU08x5HetY4i0VwLRUKaEi68BY/6v+GK5BUJDQ/Vr1671mjNnji4kJKRl0aJF5dXV1fLY2Ni4oKAgQ2JiYmNb3ZSUlIq5c+fqFi9eLKanp2d9+OGH+fPmzeuv1+tlKpVK3LdvX25n+09NTQ3aunWrRq/Xy7RabcKMGTMq3nzzzXN79uxxXbZsWaAgCBg6dGj9Bx98cAYAtFqtefHixeeSkpJiAOCZZ545p9VqzWfPnlXcfffd4QaDQRBFURgxYkTd4sWLywHgT3/6U8l9990XptPpvAMCAgzbtm3LB4CNGzd6rl271lcul0sqlUpcv379r+3tB9YTCPa4rn3w4MFSenq6rcPoUifPl+F4SSne+P4gREnEH8eNwQMDOnzEMNENSZIEqbISphPHWpMTV3JwAEQRuJDpBQDJwwvmwEQo3QUoVExc9FSCIByRJGmwJW0cO3asIDExscJaMfUkx44d805MTAy2dRx08/rCOAIADuZXoLzyHH79+a+IVPTDoLvmwj+gv63Dol6kxWTGP/b9is/2HoQoNCLU730UOjVBkhR4vBqAegTedfsFRnMjolpaYJQpERuxAEnBgRgVNArODs62fgt0kziW6Ft6wtiHMy5sQBQlzPr0M5w8X36xbMbHn+G+2BjIe2gGjOyLZDbDdOwopKZGwM0d4rnzEFQKQCb77zq/dqa6CTWVcIpqgqBy6eaIiYioM0pqmrHw3++jxXEb6ionQa7+FWH/vg+7Hj9i69ColyiuaUbK+4dgMIlw9gnGr6atqHYxAoIcKlFEk6IOjoadcGpxhkkmR6Jejx/VMlRmfII7xu6+al8BIiJLMHHRzX44cxbzvtiJk2XlV10ra2iEv5urDaKiXsdohNTcDJhMgKMjDFln4RivAxxk7R3BfvmtOTlwSLwFglzePbESEVGnfH36S7zz3dtocC2Aqbk/RKM30GzCWc/Dtg6NepGckjrUNhkBAYjyUaBY+BYQJAAC9IKEdzxcLyYnTIKAr5zVmNjYjOm16RB2PQPc/Zpt3wBRN0tISIg2GAyXPYVev3796eTkZG4IbgVMXHSzp7bvwMny9mdMMWlB1iLW1QGSCPj4QoAARYAGkiihQw8/aqphPJIOhyHJfFpCRGRnmk3N+PO+Z1ELA5QSAHUhBO02CA5VkEtXneZHdNPSC6shCAIeGa7D7TnP4CGVGRfnagoCzFeMEZoEAVvcXJCucsTvMjZgjCRBmPR6t8dNZCvHjx/PtnUMvRnXJXSjFpPpmkkLNZ9ukxWJ5WWtS0HKyiCdL4EiQAOZk7LjiYiGekiieON6RETUrc7WnkW1ZMRL5VX4a0UlQgxGyJ3OwsXUhADTY7YOj3qR/2SXobbZgDWH9+AFsRBDGpva3zPrgiqFAvUyGU44KvF3Tzdk5HzWjdESUW/HxEU3ej/953bLQzw98Nz4Md0cDfVmsqAgwNGxddbFTZKam6wYEVHnqdXqgVeWLVu2TBsWFhYXGRkZO2zYsMjc3FzlpddHjRoVcfr0aYff/OY3uqioqNjIyMjYiRMnhtbW1vLzjnqFZ//zEgCgn8mIje6uaJbJEF0dAtfGVVgw4l4bR0e9ye/GaOHdbyec3T5HhrMBB13U6MjUTbMgQGMyI6SpufU4diIiK+BArhs5OyivKpvYvx9Ozv8fLBo10gYRUW8lUzlddlrIzTCfPGmlaIisJykpqeno0aNZubm5mVOmTKl++umng9quNTQ0CDU1NYqQkBDjmjVrzubk5GTm5uZmBgUFGV555RVfW8ZNZC06Vetf+fm+3vAwiRANHogM/jP2LhyL22K1No6OepN+ftVwUx6DRlEO4ZozLdpPZBxzdMRb7iqc3PeXrguQiPoUJi660ayBiRjg6wMvlQoPR4QhxlkNV6OBJ4mQfWpsgNlgsHUU1EM0bNytKU1+JP5cyL1JpcmPxDds3K3pin4mT55c7+rqKgLAyJEjG0pKSi5mhHft2uU6YsSIegDQaDQiAIiiiObmZhn3a6He4o0Rj+LVFjVmwxViy//h17NLcDC/2tZhUS/UVBsMr+qBmFxthBkCDDVDYKyPu6JW+wmNerkM21yd8dGprV0fKBH1CVb7jVkQhLWCIJQJgpBxSdkyQRCKBUE4euHrLmv11xPJZAL+/dtZOPDk41g37SH8eWgSnrttvK3Dol7IlJvTeqKIpRobLW+Der2Gjbs1dX9+XyeWVSshAWJZtbLuz+/ruip50ebdd9/1ue2222rbXu/atcv9rrvuuvj6gQceCPbx8Uk8deqU6tlnny3ryljIOjiWuDGZJgR3PvgRZk3bjf+dMQ2jwr3x2gOJtg6Lepl6vRF/23sKxwUDVvs6QDJ5wFhzKwyV4yBJHUsECwAi9E2AgUtPyXZycnKUERERV2bcrmnlypVeBQUFDjeqk5KS0r+jbc6dOzfQz88v4colsLm5ucphw4ZFRkZGxiYnJ0fl5+df7HfVqlVeOp1ugE6nG7Bq1SqvtvJRo0ZFREVFxYaHh8dNnz69v+nCeH/t2rWe4eHhcTKZLGnfvn3qtvo5OTlKlUo1KDo6OjY6Ojp2+vTpHY7b3ljzUf8HACa2U/6WJEm3XPjaZcX+eiQ/VxeEeHpCplBg8thxiImIsHVI1AsJLi7X3UCrQ1zdIHN3t05A1Ks1rNwSiBbj5Z8nLUZZw8otgV3V5+rVqzXHjh1T/+lPfyptK/vpp59cbr/99oa2159++mnB+fPnj0VEROjXrl3r2VWxkFV9AI4lbkwTCjh7I0LrivW/G4rBwV2aI6Q+yMlBjuHe+VB6/gAIAgRFDZQeB6H0+g8E4cbjC0dRgrfZDK9bZgJK9Q3rEwHAlpwtmnEfj4tP+FdC0riPx8VvydnS7f9z27hxo/eZM2eum7jorClTptQcOnQo68ry+fPnB02fPr0yNzc3c8mSJecWLlwYBADnz5+Xv/LKKwGHDx/OSk9Pz3rllVcCysvL5QCwffv2/AtLYU9WVlY6tI1vbrnllubPPvvs1ODBgxuu7Kdfv34t2dnZmdnZ2ZmbNm06Y8331p2slriQJGkfgCprtUdEN0+m8QIsPKlGHhsHgcuYqAPE8uqrN/C5Trmltm3b5vr666/779q165STk5MEAJmZmUp/f3+DSqW6bEStUCgwbdq0qm3btjFx0QNwLEFkHxRyGZ4Ol5Bgah0HCIIAB48jcHDN7ND924rOYV1ZLe4eubQrw6ReZEvOFs2rP72qq2iuUEqQUNFcoaHi1jgAACAASURBVHz1p1d11khemEwmTJ06Nbhtw+76+nrZokWL/AcMGBATERERN23aNJ0oili3bp1nRkaGOiUlJTQ6Ojq2oaFBSEtLUw8cODA6KioqNj4+Pqa6uloGAKWlpQ6jRo2K0Ol0A1JTU4Ou1/+ECRMadTrdVTvV5uXlOd111111ADBp0qT6vXv3egDAtm3b3EePHl2n1WrNPj4+5tGjR9d9/vnn7sB/l8IajUbBaDQKbUthBw0apE9MTGyx9M/KnnXHbyX/IwjC8QvTPzlwJOoGgloNCBb88/YPhNzZ2XoBUa8m8/FsdzOUa5Vb4sCBA05z587Vbd++/VRgYODF9VDbt293v/322+uA1n0tMjIyHNt+3r59u0dERITe2rFQt+JYgqibHfO4AyX69mYGC7jWppxtVz9yd4X24S0AH4BQB605tibQYDZc9hfGYDbI1hxbY/HszYKCAlVqamp5bm5upqurq/jaa6/5LF68uCwjIyMrLy/vZHNzs+yjjz5ynz17dvWAAQOa1q9f/2t2dnamQqHAjBkzwlasWHEmJycnMy0tLcfFxUUEgMzMTPW2bdt+zcrKOvnFF194njp1qtOzNGJiYpo2bdrkCQAbNmzwaGxslJWWlsqLi4sdgoKCLo6hAgMDDcXFxRfbHzlyZISPj0+is7Ozefbs2Tfc5KioqEgZExMTO2TIkKg9e/a4dDZOe9HV/zf5O4AwALcAKAHwxrUqCoLwhCAI6YIgpJeXl3dxWES9nEIBON78w25Fv35WDIZ6O5d5DxXD0eHys3cdHUSXeQ8VW9KuXq+XabXahLavZcuWaRcvXtyvqalJ/uCDD4ZFR0fHjh8/PhwAvv76a/d77723FgAkSUJKSkpIZGRkbFRUVFxpaanDyy+/fM6SWMimOjSW4DiCyLoiA71R79z2APfK5SHXXi6iNpvh5uAOWb8hXRYb9T6VzZXtDlyvVd4Zfn5+httvv70RAGbNmlV58OBBl927d7smJCRER0ZGxh48eNA1IyPD6cr7jh8/rvL19TWOGTOmCWid7eDg0Jo/GDlyZJ2Xl5dZrVZL4eHh+vz8fMfOxrVq1aqi/fv3u8bExMR+9913rr6+vkYHBwdI7Sz3vnST8e+//z6vtLT0mMFgkO3YscPten3079/fePr06eNZWVmZb7755tlHHnkktKqqqkdmFBVd2bgkSefbfhYE4T0AO69T9x8A/gEAgwcPtnBxPlHfJggC5AMHwfz9/s7fGxYOmUuPTcaSDbjMvLMKaN3rQiyvVsp8PA0u8x4qbiu/WaIoHrmybNmyZeevLGtubhbKysocoqKiDAAgl8vx888/Z1vSN9mPjo4lOI4gsi53Jwe8POFpzPluPgDhkr2zrv3Py9MsYkNJOfqnHgIcrvo9kOiavJy8DBXNFVclKbycvCyevXnlyWKCIGDhwoW6Q4cOZYaHhxsXLFgQoNfrr/plXpIkCNfY1EWpVF4sl8vlktFo7PTxZcHBwcavvvoqHwBqa2tlu3bt8vTy8jIHBQUZ09LSXNvqFRcXK8eMGVN/6b1qtVqaNGlSzdatWz3uu+++umv14eTkJDk5OZkBYNSoUU39+/dvycjIUI0ePbrH7ZrbpdkWQRD8L3l5H4CMa9UlIuuSKx0hHzS449M0fbWQBYdA0a/HbjZMNuQy884qv8MfnAg4vf2I3+EPTliatOgMJycnKSMj46pNr6h34FiCyHZG6cbjj8l/hEMHfmVIEBX4m/8E6O56E4J3WDdER71JamJqsVKuvGz2plKuFFMTUy2avQkAJSUlyr179zoDwKZNmzTDhw9vAAA/Pz9TbW2tbMeOHReXILq4uJhra2vlAJCYmKg/f/68Mi0tTQ0A1dXVMqPxqq0qLIlLYTabAQBLlizxnzZtWgUATJkypTYtLc2tvLxcXl5eLk9LS3ObMmVKbW1traywsNABAIxGI/bs2eMeHR3dfL0+zp07p2g7eSQzM1NZUFDgGBUV1SP3wrDmcaibAfwAIEoQhCJBEH4H4FVBEE4IgnAcwDgAT1urPyK6MbmHBxQjRgE3Oh1E6wflgHgoQsO4IScR2QzHEkT2Z1rMNOybvBV/Ed3hKoqQSxIEAN4mMxwkARAEyAUZHhr+Bwy8axUwKMXWIVMP9FDUQ1XPDHmm0NvJ2yBAgLeTt+GZIc8UPhT1kMUPQkJDQ/Vr1671ioyMjK2urlYsWrSofMaMGeWxsbFxd955Z3hiYmJjW92UlJSKuXPn6qKjo2NNJhM+/PDD/Hnz5vWPioqKHTt2bGRTU1OnB8qpqalBWq02oW0J7IIFCwIAYM+ePa6hoaEDgoODB5SVlSleeumlEgDQarXmxYsXn0tKSopJSkqKeeaZZ85ptVpzXV2d7O677w6PjIyMjYmJifP29jYuXry4HADWr1/vodVqE44ePep83333RYwcOTICAL766iuX6OjouKioqNgHHnggbMWKFYVardZs6Z+pLQjtraGxtcGDB0vp6em2DoOoV5Gam2GuqYZUVw/JQQG5pwaSUgmhuRkyd3cIDlY9+YlsSBCEI5IkDbakjWPHjhUkJiZWWCumnuTYsWPeiYmJwbaOg24exxFEXUMs/hmn6wqRVXIYt7lGILv/QHg4eqKosQhD/YbCQc6xRG/BsUTf0hPGPl26xwUR2Q/ByQkKJyfA/4oLPD2EiIiIOkAWOAhhgYMQFnMfgNYdcwEg2CPYZjERUd/AxAURERERERGRBRISEqINhsuPdF2/fv3p5OTk6+5DQR3DxAURERERERFZmyiKoiCTyexvb4IucPz48R55opkoigIA8YYVbYy78BERkV1Sq9UDr3Vt3bp1noIgJO3bt0/dnTERERFRh2WUl5e7X/jFmOyQKIpCeXm5O3rAiV2ccUFERD1KdXW17J133vFNSEhovHFtIiIisgWTyfRYaWnpP0tLSweAD8ztlQggw2QyPWbrQG6EiQsiIrLYztO7NBuzNwdWtVQrNY6ehpnR04onhdxl8RFm7Vm4cGHgwoULS9966y2/rmifiIiILJeUlFQG4B5bx0G9AzNfRERkkZ2nd2nWnHhPV9VSrQSAqpZq5ZoT7+l2nt6lsXZfBw4ccCouLlZOmzat1tptExEREZF9YuKCiIgssjF7c6BBNF72eWIQjbKN2ZsDrdmP2WzG008/3X/lypVnrdkuEREREdk3Ji6IiMgibTMtOlp+s2pqauR5eXmq8ePHRwUGBsYfO3bM+YEHHgjnBp1EREREvRv3uCAiIotoHD0N7SUpNI6eBmv24+XlZa6urj7W9jo5OTnq9ddfPzt69Ogma/ZDRERERPaFMy6IiMgiM6OnFStlDped/62UOYgzo6cVW9KuXq+XabXahLavZcuWaS2LlIiIiIh6Is64ICIii7SdHmLtU0VEUTxyveuHDx/OsaR9IiIiIuoZmLggIiKLTQq5q6qrjj8lIiIior6NS0WIiIiIiIiIyG4xcUFEREREREREdouJCyIiIiIiIiKyW0xcEBEREREREZHdYuKCiIiIiIiIiOwWExdERGSX1Gr1wCvLVq5c6eXp6ZkYHR0dGx0dHfvmm2962yI2IiIiIuo+PA6ViIh6lMmTJ1evX7/+jK3jICIiIqLuwRkXRERkubq1GhTFxKPQKwlFMfGoW6uxdUhERERE1DswcUFERJapW6tBzXM6mM8rAQkwn1ei5jldVyUvdu/e7REZGRk7ceLE0FOnTjl0RR9EREREZD+YuCAiIsvUvRYIqeXyzxOpRYa61wKt3dVvfvObmjNnzpzIzc3NHD9+fP3MmTNDrN0HEREREdkXJi6IiMgy5jJlp8ot4OfnZ3ZycpIAYMGCBeUnT55UW7sPIiIiIrIvTFwQEZFl5L6GTpVboLCw8OLSkE2bNnmEhobqrd0HEREREdkXnipCRESWcVtcjJrndJctFxEcRbgtLrakWb1eL9NqtQltr5988snz5eXlii+//NJDLpdLHh4epg8++KDAkj6IiIiIyP4xcUFERJZxe7QKQOteF+YyJeS+BrgtLr5YfpNEUTxyjUsWJUSIiIiIqGdh4oKIiCzn9miVpYkKIiIiIqL2cI8LIiIiIiIiIrJbTFwQERERERERkd1i4oKIiIiIiIiI7BYTF0RERERERERkt5i4ICIiIiIiIiK7xcQFERHZJbVaPbC98n/+85+eYWFhceHh4XGTJ08O6e64iIiIiKh78ThUIiLqMU6cOOH4xhtv+P/444/ZPj4+5uLiYn6OEREREfVyVptxIQjCWkEQygRByLikTCMIwteCIORd+O5prf6IiMh+rEkv0wS8eSxetvxIUsCbx+LXpJdpuqKfd955x+fxxx8v8/HxMQNAYGCgqSv6IdvgWIKIiIjaY82lIh8AmHhF2bMAvpEkKQLANxdeExFRL7ImvUzz9FdFupIGk1ICUNJgUj79VZGuK5IXp06dcszNzVUNGjQoOjExMfrTTz91s3YfZFMfgGMJIiIiuoLVEheSJO0DUHVF8b0A/nXh538BmGKt/oiIyD4s31cSqDdJl32e6E2SbPm+kkBr92U2m4X8/HzHH374IWfLli2/PvXUU8EVFRVya/dDtsGxBBEREbWnqzfn1EqSVAIAF777dnF/RETUzUobTMrOlFvC39/fMHny5BpHR0cpOjraEBoaqj958qSjtfshu8KxBBERUR9nN6eKCILwhCAI6YIgpJeXl9s6HCIi6iA/F4WhM+WWmDp1as13333nCgAlJSWK06dPq6Kiolqs3Q/1PBxHEBER9V5dnbg4LwiCPwBc+F52rYqSJP1DkqTBkiQN9vHx6eKwiIjIWpaO9i9WKQTx0jKVQhCXjvYvtqRdvV4v02q1CW1fy5Yt006dOrVOo9GYwsLC4saMGRO5fPnys35+fmbL3gHZuQ6NJTiOICIi6r26+hi5LwD8FsDLF75v7+L+iIiom6UO9q0CWve6KG0wKf1cFIalo/2L28pvliiKR9or/+c//1kEoMiStqlH4ViCiIioj7Na4kIQhM0AxgLwFgShCMALaB1kfCwIwu8AnAHwoLX6IyIi+5E62LfK0kQFEccSRERE1B6rJS4kSZp2jUsTrNUHERER9V4cSxAREVF77GZzTiIiIiIiIiKiKzFxQURERERERER2i4kLIiIiIiIiIrJbTFwQERERERERkd1i4oKIiOySIAhJU6ZMCWl7bTQa4enpmThu3LjwtrJPP/3ULT4+PiYkJCQuOjo69u677w7Ny8tT2iZiIiIiIuoKVjtVhIiIyJqcnJzEnJwcp4aGBsHFxUXaunWrm1arNbZd/+mnn1QLFy7sv3Xr1lODBg3SA8CHH37ofurUKWVERITBdpETERERkTVxxgUREVmsoUjUlOw3xxd/Y04q2W+ObygSNdZod8KECbWffPKJBwBs3rxZc//991e1XXvxxRf9FyxYUNKWtACAGTNm1N55550N1uibiIiIiOwDExdERGSRhiJRU5sn6UQDlAAgGqCszZN01khezJo1q2rLli2eTU1NQlZWlnrYsGGNbddyc3NVycnJTZb2QURERET2jYkLIiKySP1pKRDiFZ8nImT1p6VAS9seOnRoc1FRkeN7772nue2222qvVa+0tFQeHR0dGxwcPGDp0qVaS/slIiIiIvvBxAUREVmkbaZFR8s7a+LEiTUvvPBCv5SUlKpLyyMjI/WHDx9WA4Cfn585Ozs7MyUlpbyhoUFujX6JiIiIyD4wcUFERBaRKdHuRpjXKu+sJ598smLhwoXnkpOTmy8t/+Mf/1j6xhtv+P/888+qtrKmpiZ+rhERERH1MjxVhIiILOIaIhTX5km6y5aLyCC6hgjF1mg/LCzM+Pzzz5ddWZ6cnNz86quvnk1JSQlpbGyUeXp6mgMDA1tefPHFc9bol4iIiIjsAxMXRERkEZcgWRUgov60FCgaoJQpYXANEYpby29eU1PTL1eWTZo0qX7SpEn1ba8ffvjh2ocffviae18QERERUc/HxAUREVnMJUhW5RIEixIVRERERETt4VpgIiIiIiIiIrJbTFwQERERERERkd1i4oKIiIiIiIiI7BYTF0RERERERERkt5i4ICIiIiIiIiK7xcQFERHZJUEQkqZMmRLS9tpoNMLT0zNx3Lhx4QCwcuVKL09Pz8To6OjYsLCwuDfeeMPbdtESERERUVdh4oKIiOySk5OTmJOT49TQ0CAAwNatW920Wq3x0jqTJ0+uzs7Ozty3b1/OX/7yl8CzZ8/ymG8iIiKiXoaJCyIislj+/o2anc8Nif90bnDSzueGxOfv36ixRrsTJkyo/eSTTzwAYPPmzZr777+/qr16gYGBpv79+7ecOnVKaY1+iYiIiMh+MHFBREQWyd+/UXPs8+U6fV2ZEpCgrytTHvt8uc4ayYtZs2ZVbdmyxbOpqUnIyspSDxs2rLG9epmZmcqzZ886xsbGtljaJxERERHZF06pJSIii2Tt+VugaGq5LBEumlpkWXv+Fhg2ama7MyQ6aujQoc1FRUWO7733nua2226rvfL6jh07PKOjo12USqW4YsWKQq1Wa7akPyIiIiKyP0xcEBGRRfR15e0uz7hWeWdNnDix5oUXXuj31Vdf5ZSVlV32uTV58uTq9evXn7FGP0RERERkn5i4ICIii6jcfAyty0SuLrdG+08++WSFu7u7OTk5uXnnzp2u1miTiIiIiHoO7nFBREQWiZk4v1imcBQvLZMpHMWYifOLrdF+WFiY8fnnny+zRltERERE1PNwxoUdkEQJzVUGqL0dbR0KEVGnte1jkbXnb4H6unKlys3HEDNxfrGl+1s0NTX9cmXZpEmT6idNmlQPAPPmzasEUGlJH0S9hskAGBsBJ09bR0JERGR1TFzYga//9xhObS8FAAxbGomBvwu1cURERJ0TNmpmlaWJCiK6eT+9dztWohiOohpLHtiMYG2srUMiIiKyGi4VsTFJki4mLQDgh+W5SH/7lA0jIiIiop6krlGP5xSlOOrkiEPOZvz546k4UVBi67CIiIishokLGyr6vgJrB35zVfnh104h6+OzNoiIiIiIepL3vs/Abzfcj1K5/GLZYVcFHtv7/1BalmnDyIiIiKyHiQsbqsisR0utqd1r3y4+CbPB3M0REdmX5przEM38d0BEdC3Hqw6jwLEYkky4WGZqiERN3XA8snOWDSMjsg91dTW2DoGIrICJCxuoLWzCNwuPQ6VxgHDxv4B0xRdwYsMZ2wRIZGNN1SX4dK4O/34+GZ//L/d8ISK60sFTFViw5ShSfDQY26iHXJIuXjNUjYKxLglnWrxQXn/OhlES2c53+d8gfl0cRnw2AnvWPWfrcIjIQtycs7uYjIBMDvywCec+KkBFxhBom3fAURUPfYNDu7cY6tqfjUHU25XmfA9JlKCvNkKlaf/fBxFRX6M3muGokOGZT4/jYH4F5DIBTWdroHRygBkSRJMzBJkBSq9vIRq94e1QBM/mesDV1pETdb83v18GCK0zkT5v2I2JeNGm8RCRZTjjoqtlfgu8dQ/w4mhg4/8C9eUoyNaipaYFQlUxAn1z0TbD4lIqbwcMepJPmqlvCoy/DWXH61HwTRWay3i0X18lCELS448/HtT2eunSpdoFCxYEtFdXrVYPvFF7e/bscQkPD4+Ljo6ObWhoEG5Un8gumE0oOPgB7vrXaExdNwZv7D2B4ppmVDcZUV7fgq8bWvCtOAiiwQvNRY9AX3o/FM6/Qul+CA96DoDCN8rW74DIJn47cP7Fn8eMesGGkRCRNXTLjAtBEAoA1AMwAzBJkjS4O/q1uR8/Afa89t/Xp38CKgswbkgtZIIZ3/4wBWfORv73ukxA1AMBcHCUY/D/hkOhkl/dJlEfcL78HE54R8KrMBtJ0/9g63DIRpRKpbRr1y7PkpKSUn9/f4unoK1fv14zd+7c0vnz51daIz7qXn11LLHm8//Durrd+FNlFYJMZrwlpKLOzQPy4HKIRk84tvhCUlZAaXJCi2CEl0KNJdr7EOzmjuhh82/cAVEv5VqjxcBaH+SqnDBSl2TrcIjIQt25VGScJEkV3dif7TRUQ8r8FhV7voG3BEAABACQzJBqS+F4Yeb7LePLYMwbC4WLDG791Eh+OgIKRyYrqG8TRQkzd3+PIwMmIjHmFjw3dqytQ6KO+OkzDdLeD0RDhRIu3gaM+V0xhtxfZUmTcrlcSklJKf/rX/+qXbVqVfGl17Kzs5UPP/xwqMlkEiZMmFDbVr5z507X5cuXB2g0GmNOTo5TfHx807Zt206vWLHC+9///rcmLS3N/ZtvvnH74osvTlsSG9lMnxlL1JQV4WReHj6qOYImBxmKHBRwESXkC9WobK6DJBMgKMsxyutRGORn4G4ahsQhjngkOdnWoRPZXEVzBdacWYJKlyqMbIxCkFugrUMiIgtxjwtryz0AbHoazadVUFY6IM99AJSqFgT3ywPQmsAQBcDoGg7t7HmYpAm6fntEfYxRNKPOZIYKZoTIGtFQfhpqDz9bh0XX89NnGnz5lg4mQ+vyw4YKJb58SwcAliYvFi9eXBYfHx+3bNmy0kvL58yZ0/+xxx4r/5//+Z/Kl156yefSa1lZWU5Hjx79NTg42JiUlBT99ddfuyxYsKDiwIEDLpMmTaqdPXt2tSUxEXW19/+zCu/mrcbUsjDcJnngo6Aq/M3T47I6gtEFA91n4ZW774e7E/cCIrpUXUsd6hzkqDEDv2grUN1SCW8nb1uHRUQW6K49LiQAXwmCcEQQhCfaqyAIwhOCIKQLgpBeXl7eTWF1gR0vAQCaf3VCbZEnjp4YiZ9+GQ+z+b8zKWS+4XBcuAlg0oLoKo4KBXb9dhZebdyJe2v3I/urd2wdEt1I2vuBF5MWbUwGGdLet/gRl0ajER988MHKl19+2ffS8p9//tnl8ccfrwKA3//+95ct/YiPj28MCwszyuVyxMXFNeXn5ystjYPswnXHEr1mHAEg9+gGKOQy3G3OR7BjLuRX7MjiIIr4bfDfsf43c5i0IGpHqEcoVt72NiQIKGuuxJ7sn20dEhFZqLsSFyMkSRoE4E4ATwmCMPrKCpIk/UOSpMGSJA328fG5uoWewq11bO02qB7et5QhJuYIEuJ+gFypBDyCgLjbgN/yFzGi6wlwc4W7sRoyAGXZ+9FUVXzDe8iGGiraTwxcq7yT/vCHP5zftGmTd2Nj42WfWTKZ7OqdjQE4OjpeLJfL5TCZTNyIs3e47lii14wjAAyr06BeJuFNbzn2OMuhNUpIqtYgoPK3uM84CM8ErcC8cQm2DpPIrmnK8yFJIgAJB77+m63DISILdctSEUmSzl34XiYIwlYAyQD2dUff3W7WSuDd38JBXgEHTz3iXU4D014D/GMAGQ9xIeqI+rLLtx+QOzrbKBLqEBdvQ7tJChdvgzWa12q15smTJ1dv2rTJe9q0aZUAMGjQoIb33ntPM2fOnKr33nvPyxr9kH3rS2OJmEc/wFMbX8RBSYWjLQmYpy3GI7+dAaWmn61DI+oxfij470OPQ17c1oiop+vyxIUgCM4AZJIk1V/4+XYAy7u6X5tRuQDzP2v92WwCZPKLZ0gTUce0NFx+6IODk5uNIqEOGfO74sv2uAAAhVLEmN9ZbarMc889V/qvf/3r4mP01atXn3n44YdDV69erb3nnnu4Z0Uv19fGElH9/RD1x1VIBWA0i3CQ88EHUWeVOg4F8DYAwCjjWJyop+uOGRdaAFuF1l/eFQA2SZK0pxv6tT059z4luhneoZefcigw+Wff2jbgtPKpIk1NTb+0/dyvXz9Tc3PzxdfR0dGGo0ePZre9/utf/1oKAJMmTaqfNGlSfVv5+vXrz7T9/NlnnxVYEg/ZVJ8dSzBpQXRznhgWgfdOizDKZFB02+p4IuoqXf6btSRJvwJI7Op+iKj3uGxPC0EGs1EPhdLJdgHRjQ25v8rSRAXRtXAsQUSd9f2JYxCE1s3xXZ08bRwNEVmK6UcisjtKZ09odLdA4eSGhKlLmbQgIiKiTgkOvQUx0ij4Kf3xxug3bB0OEVmIaxmIyO4oHNUYv2i7rcMgIiKiHkrn5YyNj662dRhEZCWccUFEdqc0ax8KD30GSWr3tEsiIiKia5IkCZ8dKcK+3HJbh0JEVsIZF0TUpZoMRqiVDp26J2PHq2isOIOy3AwMuPN5OHkzx0pERNQniSIgGgGFY4dvOV3RiH/8eAA1qm24vTwc06KnIdYrtguDJKKuxt8GiKjLLP9sP8as+ADbP8nq1H3BQx+AwtEDdWcKUJ0pQTRx5gUREVFfdPSNgTj0YhIW/WNah+8J9nLGkAgRCscz2HFqO5YdeKELIySi7sDEBd0cQyZgqgQA1LbUoqK5EgaTwcZBkb1pqGmBBMAsOOLsN3qIJvGG90iihJChv0XU0H8g9JaXIJkBmYLHofZF+fn5DhMmTAjT6XQD+vXrN2D27Nn99Hq9sHPnTldXV9dbYmJiYoODgwcMHjw4avPmze6X3vv22297RURExIWHh8eFhYXFLV26VGur90FE11CUDhj1rT/XFgMNZYDZaNuYyO6UtoiQS3IcqlNi1meTO3SPTDRiZtJACOYGmCUJTib+vSLq6bhUhDqu6TugfOp/X8uDUevzLX7/n7moaqmCn7IFnk6hWJD0HHRu/W0WJtmP5yePwfSfBsHb0Q2QgJI0CT7JZihd5RfrGOok1J8WodIKUKiAiuMtkAxyqFThEAQBEAHJLEGQ943khclggEyhgEzWt/PKoihiypQp4Y899ljZ/Pnz800mE6ZPn66bP39+4OTJk2sHDx7c8O23354CgIMHDzo9+OCD4Wq1uuDee++t//jjj91Wr17t+/XXX+cGBwcbm5qahL///e9etn5PRAT8eedx7CieA/cWJ/xfVSMmROnQMHQhNu+ejRw0YWy1H9L6P4HlDz0EtaPK1uGSHciPHYrnC25wWAAAIABJREFUi8rR7HIcNfVGPPavYXjzwV1wU//3iNPtR4ux41gx5owPwKDKfXjrh9fxgcoIUWidsXle37cerrWYW+Ao7/jSGqKegIkL6rja9y5/bS6AuWggTMZQBKvMWNj/FM60FOLxvcUYHTgSS4Y+a5s4yW64+jogeowrKg9JaJvgVX4YkLuZYTD8Ao9wDcRaHfQVgL6ibTnI1fth1J4S4RElv6q8NxHNZjzn0Pq/ZP+BgzDvyBEbR9Q5pqIijVjwayAMBiWUSoMsOLRYERRUdbPt7dixw9XR0VGcP39+JQAoFAqsWbPmbGhoaML48ePrL607fPjw5sWLF597++23fe+99976V1991f/ll18uCg4ONgKAWq2WFi5cWGHZOyQia/jPyQ/R7FuLitLpeEKS447zH+OXXa+jyVmCoX4otprGQFF2GOn/+hv+MmAVRg4fYeuQycaemvpP6HK3448HlkACcEiqx8iPR+EVzXDINCFISnocezP2Qln1IR77ugr9jHL86nR5oqLYeBb7zu7D6H6jbfMmukllcyXGfjwWAJASk4LFyYttGxCRFfXtR3rUOT5/v6rIWVaLB7SlaDTL0SgqUN6iAgRgX/H3KKw7Y4Mgyd6oXBzgN/rypENTaQn+P3t3Hh5Vefd//HPPTDJZCWRhCyEBEpYAohJRKyoCtrYiihZQcanV+lSrYov4+OivrVq19lFr5VGqtra11dYiuGvVKuJaF1D2HWTfsu/JLOf+/ZGAAQKEbGcS3q/rykXmPmfO+cxcQ/LN95xznxVv/lKfPnG7Krcf+AwjYyTT4ASLym1SsCbc5lndsnvFin1NC0na+dWXLqY5eqFt25Kd9WszFQhES5ICgWhn/drM0LZtyc3d5rJly2JHjBhR1XAsOTnZ6dWrV2DdunUHHUYaNWpU1YYNG2Ikad26dbGnnXZa1YHrAHDfzLMvU1o4KF/CCnnj1+vDpHxVxq+UNUbGVy6ZkDxRpSqMDuj1j36lmkDI7ciIABMGnq+/f+eFfcWBlXRr0Se6Zf2z+vac8VoaeFyVMdskE9DmqEZ+/FurV976uRSqbd/g7ej9re9rzHNjJGsla7Vg47/djgS0KhoXOAr7HeSUleQJGI2tKtcVPbfozg2D9Nye3jLGkTyOXnlnqmz5F+5ERUTxRnnU87Rvftz4/MlK6n66unY/Q2p02gtT//WN4mVtGtEVodpaLXv+ef3hO9/ebzzzrLEuJWoeZ9PGdDnO/r9PHMfjbNqY3txtWmtljDloVtb68UbXBxD5/HHdlVGeJn/K+4rp/i8ZhZUUlnqGgvLFbVJcxh/lS6z7gf9J1z064Z4XVVDRef/YRNMN652jP4x9/KDxhFCtfrP7a91UVKLji3spfMCvg+6hkAYGArp090p9/cm8dkrbjirypeUv6IZ/3yA1+LU56+zHXAwFtD4aF2iybRVJmrnoZt3yxQ+1vTJZslLZfQPlfbKPRpdW67cDl+neAavUzReURx6tiomWU7jS7diIEN4Yo24j6r/3+pWZO1PpOf91FBtom1xuWvSXv2juNVercseO/cZ//O67LiVqpr1nWjR1vAmGDx9evXjx4viGY0VFRZ5du3ZFZ2dnH/RXzBdffBGXnZ1dI0nZ2dnVH3/8cVxz9w2g7Zio7Vqc6JGcugZ1V8fqgT179P8Ki+UL+VW97SpV75imqu1TtGPPFMUkvKeSKiZWRJ1TMk7T5d2/td9YF8dRV8dR0GNVmLBDxiM1PPhxZ2Gx7i4o0kdxMQr7Yts3cDsom/9bvfbmTMnzTdPibP9xGpCS42IqoPXRuECTVNSG1f/RNXpw+ff0f6un6JFVV8gYKfGyrUqavlF7kqSiYN3Z23WzGTgqC0fJSctwNzgiSlRM8yfY7GxTc+5YvFgf/vYhBcr3P5PpVzUd8MhidHTjs54darwJJk6cWF5TU+N59NFHUyQpFArp+uuvz5g8eXJBfHz8fufpfPbZZ7EPPPBA75/85Cd7JOnWW2/ddfvtt/fZsmWLT5Kqq6vNPffc0725WQC0jhXbt+tn718tx1ekKF+sUp1oRTtGv07rqvtSkhXc93eXlcJJCocTNTxmobL9JW7GRoTpnjpYDauCzVHR+ln3ZN2fnKQzq8qUHgyqrhqts9QfrY1RUXojIU4q29X+gduItVZvb/xYr63eoHW+kEz9mYfdQj799pK/u5wOaH1Mzokm+eErXytY/6dC/4RtuiZnriTJm10pYyTflmj9omCIYr0hxXtD8nvCOj25UJ6KP0txZ0iGjxqkqASPpLq5Khqe2t/Yqf8H6pbbOVoXK95+W8+c8x1JUlxKiqITE+Xz+3X8FVcqOsYvG+54c3l4svpvd9avzdzvchGPx/Fk9T9oBpMmb9Pj0UsvvbT+2muvzXzggQd6OY6jsWPHls6aNWv7u+++m7Bw4cKEIUOG5FZXV3tSUlKCDzzwwJbzzz+/XJKmTp1aumvXLt+4ceMG7b20ZNq0aUzOCbjIWqu//f1PMqn1c1aEqjWttEznlVXqlh4pqlaUhu85Qcv6/FkyjmzYL8mrQKBGz36xQ9PG93E1PyLHD076qR5a+ecGI1Zbo3zqHQrrp8UluqK0XN/L6KVqT13d8HjXurtlZzix6j/+GhcSt77b5v6XXi//SFHyKDN1sO7c1V0x0dma0G+EeidzEy10Tvw1iSY5vme8XlpdqiiPo1hvQF2iKiUZvbd7hPL3JGnM3wr0c4XU544V8sQ4ejO/u76XukteJ1myVZLp4vZLQKQwanggpO6B30hByRcrhSoPfkrSUMkX0/FPEKsoLNzXtJCk+LTu+umKFTLG6M/nfk+Ve/Yod+L5yhg1ysWUR8/Xp09RSPVzXbTSXUUkKTs7Ozh//vz1B45PmDChvLy8fPHhnjt9+vTCvXckAeA+Y4x83U/S93d9pNfTytUjVKhTq6tU4omTDaUpozhT71adrd7la7UzqVDGVyVjrYqLxumjXV5Nc/sFIHIYU1dHNDie4bFSZuhsbY/5WDFdUjSlYrueblB6ZgbD+vslr8rjO/jOZR3NK5/P1esVH0vGKCir7x53vkb88DKNqCyQnrlI2mqkEy6VouOPvDGgA6FxgSPaVFKrykBY389N0qhe8Vq3c4WsSVBFsFa3L7lIheUZen7FP+WkVSjh9WRF1Xo0pdYo6cZcKWW25KFpgW8kDZFK66c+CYUq5YmqVp9Te8jjratAglWOCpdYOQHJEyuljpB8/s4xwcXG997b7/GFf/zjvrNNxt95lwrXrVN6Xp4b0VrM16dPkVrYqADQeS3eWiJfSoKqe1+sybGpemXr7/VsKEE7HOlLbzd5exSpfGO8KvKvVWb8fbLG0Q8KHX3e70r98ryhbsdHhPlB8on6S3HdbcMH1gY0utcp+unEWfuWn7r1fX3yn3uVUV2taVEn66TL7pPxNXvapYiy+ct/SjH1DxxHkwZOrvs+PlU663bJ46NpgU6JxgWO6KLn1urLPXWXqv9jRbEGJiTp57nlWlLuV6/u76lPcoqyLoxR7auOap/KVGhguUyPrrK9X5DxdPyj5GhdCb28ssGwKrYZdemVoC79kvZbHhXnUfc8K+OVjKdzXB6y17BJk+SNjVW4ulr/9cEHyvrWNxOMZZx0kjJOOsnFdADQdn77/FR9v2arnNru+qU/Tza5RK/F1Sqcf4ZqK0bohMRFSupRpZsqHlW3rdVaEJ2k9b1/qdnTRrodHRFoxvlPK2X+bYre9oUuPX26lDtxv+WnZZypIWnD1dXfVR7TuWrRKy5/Rub3QxXwSj+97kuZmAZ3CB/4nUM/EejgaFzgiPY2LeoYpScUKjWmVJ8XD1FpTYr8oV7yxeerpqJQ1vEotDpJtihZTkGJvN2TXcuNyJXY16vEvode7onqXA2LvTxer+6pPPj+8mvfflsb3n1XZ8ycqfjUVBeSAUAbCtZofeIW9aoMyKpcwepBClVUyx/zvmzsbvmry9Urarei0paotqRGO0yCrKTl1WluJ0cE+8HY+w+7PDmmc9agSfF+3XDLQVdR6vm1z6ukpkRXDbtKPg9/4qHz4VONI7pocBfNW10mSYr2BDU1631truih7LgSVRedqZM3b1TJm0Xy+axM2MqRUcxZJ9K0AJpo6XPPaefSJUrJztaoH/3I7TgA0Lq8UbownKKZPQsVru6hcOU2JcVtlBM2OiFQoMXx67W4/ESVx3+m97v3kePfI09tb90ykgk5gaZwrKOX1r2k2nBQJ6WO0/G9+7sdCWh1nevcKbSJuVNyFP75ibrrzF7603cdVdgcjXz9SY149WkNXVGut4PH6dox35fjj5Vk5JEU/PBLBb5a43Z0oEM487bbFJ2YqFd/erPm33uv23EAoHV5vLrlxx/p1cuX6uzB0zXj9Ayd49+jPo70qXOKyspPVXHqF4pxMhWKK1LQW6PSslw9/cnXKqsJup0eiHge49EtJ92ioq1natrja7VwE3NTo/OhcYEm8RijX5zZW9NGnqYBmdMVE5WoKFld/NUSxQWDCng8ClbU7lvfllfJl9XLxcRAx5E2cKDKtmxVqKpK79x9l0q3bXM7EgC0usSYKP3i29/RtaOv0cThP1Bx8BQFSkdJJiRrwgo7UijsSMaRN3abEhNqlBDNycFAU5zQ/QRt39lfCeG1+uW/fqlAKHDkJwEdCI0LHLULBnfV8O6xskaanTdKf+71B/00baWuG3ORvuieXrdS2JHpmuhuUKADGXfXXZIkJxjU/X0zlL92rcuJ3Ldhw4aocePGDcjMzByWkZEx7Kqrrsqoqakxr732WmJiYuLxQ4YMye3fv//QGTNm7OuSvvfee3F5eXmDsrKyhvXr12/o1KlTM8vLy/ldB0SYYcdfL6f8EoXDsQrV9FTUnh+pa3QPGW+FjKTYpC/lTfmXPJ1skmagLX0vN0Vd+s3W9pj3dM6fh6s2FHY7EtBqKObQLPecla4TeyVoY/YwfXTm5QqeP0U7U1K1NaFb3QqxMQqt3uRqRqAjGf7978vWf28l/XbwIDfjuM5xHF1wwQXZEydOLNm8efPyr7/+enllZaVn+vTp6ZKUl5dXsWrVqpWLFy9eNXfu3JQPP/wwbuvWrb5p06YNuP/++7dt2rRp+YYNG1acc845ZSUlJfyuAyJMbLRX/+/cIcpIjlXvpBhdMGKQxucmKyqUrBMqpJ4Br6rLclRSxVFjoKl+eVYX+axV2BjlR0Xp//35FrcjAa2G8+/QLHkxQf1faqnezumjCccNV2q8T8fNv0dxhSV1KxSWquKJF9Xtdz9zNyjQQQSrqmTi4mSrDr7rSEfw5OcLk+9d8EH67oqK6B4JCYE7xpyx/dpReUXN3d6rr76a6Pf7nenTpxdKks/n0+OPP761f//+x40dO7Z873pdunRxhg8fXrVmzRr/vHnzuk6ZMqVw/PjxlZLk8Xh01VVXFbf81QFoCzk9EnVJXl9FR3l0yai+2lHSV6+/UStP6vP6wZ6wnq8O6/WlOzXtlEy3owIdQtdgqUaUVWpPcpICxqgylHTkJwEdBI0LNEvF4y8o47Pl+snZJ6vq5/O1u6JaMYWlciR93Kuf+pUVqs87nyuppFweLhkBjiguOVkzly5TKBTSx7Nm6axbZ7odqcme/Hxh8sw3386sCYU8krSroiJ65ptvZ0pSc5sXy5Ytix0xYsR+XZzk5GSnV69egXXr1u27af2uXbu8X331Vfydd965Y86cOclXXHEFM5IBHcRdr65UcWVA44d011kPLlBlbVA9es3VotgqfZlo9dTW/9WTX2Tq4lF95eWSEeDI+p6s+6e9qWtrfHpt5Xb9aPyZbicCWs0xcfpsftjR7jDXeLUm/+nHK2pofxWUViu8I1+2sFRBSZ/16Kv7Thqn+0eOlVNRrbLHnnc7KtBhJPfvr+4DB2rSo4+qa9+Oc4Tx3gUfpO9tWuxVEwp57l3wQXpzt2mtlTHGHmJcCxcuTBgyZEjuuHHjBk6fPn1XXl5eTXP3BTTF2t3lKqrksoXWdHZuD53Qt6s+/7pQu8tqVFYTVkH9MTVrjK7N6KquFY/r040FLicFOpCew9Q/a7Bu+t44xTK5LTqRTt+4qHEczSwu1c+Ly1TlHFQDo5lizjxRcyZP1aTE43Tb2EkqjYpWyBj1Ky3UiD3bdcaOjTKSbNF82YoX3Y4LoA3trqiIPprxphg+fHj14sWL4xuOFRUVeXbt2hWdnZ1du3eOixUrVqy69dZb8yVpyJAh1QsXLoxr7j6BQ1m+eZd++uyvdePz892O0qlcfkqmPMZoY0GVYqM8Ot6s18UVZfuWh4xRTeJ6lb0/W9r0sYtJAQBu6/SNi9+UlKtE0h5JdxSVuB2nU/EYowpvlCSjyqhofd01VUmBGv36039p8vqlkiRnx3ap8CEpxO0dgc6qR0JCo4ehDzXeFBMnTiyvqanxPProoymSFAqFdP3112dMnjy5ID4+3mnsObfccsueOXPmpMyfP39fw2P27NnJW7Zs4ZATWmTO6+drS/IrWqqZuv+z+92O06mY+itAQlaqtH69meTdb/kun9HJW/+kkn/d5UI6AECk6PSNi2r7TX37taT7SkrdC9PJXHNiquZfMUhXlG9Wr6oKDSnOl9fZ/++JwMYEOZ4LJW9vl1ICaGt3jDlje4zPt99//hifz7ljzBnbm7tNj8ejl156af0LL7zQLTMzc1i/fv2G+f1+Z9asWYfcZkZGRuivf/3rxpkzZ/bJysoa1r9//6EfffRRYrdu3RptdABN9bUntO/755e+oCc++MrFNJ3L/zs3V09dmadoj9E6T4q2lXxb0jfzWWzw+/Vk3EBVjLjavZAAANd1+qNQg6Ojtbo2IFmrrLVBpX8W1IZL/RrQM8btaJ1C36RopZxzgko/+VxeSV7Zfbd0NJLsbo9qFgxR/OWdvkcGHLP2TsDZmncVkaTs7Ozg/Pnz1x84PmHChPIJEyaUN/ac8ePHVy5atGhNS/YLHKhnlwwpsF6SVSCqSnNWXKcLjn9HPbokuB2tw/N6jHJ7J+n4vl314cZSGe/eOXnrmhdBYzW3e6HG9x+gPu7FBAC4rNM3LlIbzO2WVOxo6OKAivpXasAFNC5aS8y5p6n0F0/se2wkyVh5koJySqIV3t2iv10AdADXjsorammjAohUQ8vi9S+/UbD4FBlvtUoSluvz7St0XpeT3Y7WaVx2cqY+XLdH/pT3dXxNrVLCYb0bHytJCiuo6lC1ywkBAG7q9I2Lv9cE932/bKRf0bXSGafHH+YZOFpOZZUU5ZWC39y5JeHibYoZU6DyP2QpsPygA6YAAHQY79lFkpMob/zXqi04S76YdcrqMtDtWJ1KSXVAXl+5PNZoRnGxoqzVZ7F+VXg8CknaVsFcWQBwLOvU5++/XFmlKkmydWddOD5p8ZkxOi3Z72quzqb88Rf2a1pIkq31SFayAY8CX61zKRkAAM3nWEe3/Pt/9GWMV8ZbLeMrU0aXBfqxRmh4eje343Uqz3y6ViZ+vbpXdNNz/ky9ERir8mB39QyFdWNxsT5e97nbEQEALurUjYu5FbX7mhYX/alCv7yjVA/HxstjzBGeiaaqfvcL1fz97YPGK19IV8FNxymwNEkqLVf+Zb9wIR0AAM1XGazUe9tf33fri1jV6PjEZF165e9cTta5/OLl5Vpd+aGiu/1H3W21Xo7qp99rnALlx+mu/EKdVl0ju/lNffXW025HBQC4pNM2LlZsrlJtjSNZyRNwdMrHAWWuD6lX2ZGfi6ZzdhUeemHgm1uaBb9YKWvtodcFACDCJL59l84vLZXXWvnDRgHbXUu9fvlj4tyO1qkUVQYUDnRTuKaHanwB+RJXKbb33xWdskD3pnbTPSnd9GG8R2b9v92OCgBwSbs0Lowx5xhj1hhj1htjbmvr/YWLQ/rgo2LF1Fp5g1b33VehHn2i5RseK09qVFvv/pgSNSJb8nmPvGJcjGx51ZHXAwCgEe1dS+zML9S8Nc9pS3SUfpNfqAnhW5QVvk5nJP+krXd9zDm+nxTT/U3ZmnStKD9b1hp5ogtljKMtUVFa7vfLGqMnevfiIAgAHKPafHJOY4xX0mOSzpa0TdIXxphXrLUr22J/1a+XqPpXe5R9crRWZkcpe21Yg2Zlyni5PKQt2JDTpPWM16Pw1t3yDO3fxokAdBZer3dkTk5OdTgcNtnZ2dVz5szZlJiY2LQfOg3ExcWdUFVV9dVrr72W+NBDD/V47733mDG4g2nXWiIU0NNvPq3f7XpcNjlJRkYm6gT94ZorZa2V4XLTVudTnJzaNDk1feXt/kb9pTl1DYrsQFCFXq+KPV5tCW9SbbhWMT7uDAcAx5r2uKvIKEnrrbUbJckY85yk8yW1erFR/Uaxqn+VL0k64fOAEiqt0jeFZE93ZJKacFYAjp7jSKHwEVfr+usbFEXTAsBR8Pv9zurVq1dK0sSJE/s99NBDaXfeeedut3PBFe1WS7z2xvV6sOjT+grJyEjabUpUv9/W3h0kVQVr5YndpOiYnTJRpTKq6092CYf12O49WhQTo43R3TT+vF/RtACAY1R7XCqSLmlrg8fb6sdaXfWfCmRV16P3WGnQ8qASKqzCWwNtsTtIqnzmrSOuY/r2VOzZo9ohDQC3vFldk3xVYcnwCwqKR15VWDL8zeqa5Nbc/ujRoyvWr1/vv+6669Lvv//+tL3jP/vZz3r/8pe/7FFaWuo59dRTB+bm5g4ZOHBg7jPPPNO1se2Ul5d7zz777AEDBgwYeumll/YNh4/ceEVEaLda4ueFn+332FqrPkUxKqmilmgrL678j7zRpfL6d8vjqdk3XuXxaFNUlPoGQ0oe+j0NTR3sYkoAgJvao3HR2OGJgy5QNMZca4xZaIxZmJ+f37w99fnmNqcNdxAuCDZvezgi/8m5h1/B61HClPHtEwaAK96srkl+qrI6s9jaaEkqtjb6qcrqzNZqXgSDQb311ltdhg8fXn3ZZZcVzZs3b992X3755W6XXXZZcVxcnPP666+vX7ly5ar3339/7e23397HcQ6+qmTZsmXxjzzyyNY1a9as2LRpk/+vf/0r97TsGI5YS7RKHSEpWbGSJCcULyeYKBmjr7ypCoWZW6GtDOpyohIblGr9A0HN2p2vb1dW6b96pOmmvlkaMORC9wICAFzXHo2LbZIyGjzuI2nHgStZa5+01uZZa/PS0tIOXNwkXR/os6+KaVjhBP7NrUTaStyU8fLkZjW+MLOXeqyeo8QbJrdrJgDt659VNenBA36fBCXPP6tqWnREvLa21jN48ODc4cOH5/bp0ycwffr0gtNOO626sLDQt2nTpqj//Oc/sUlJSeGcnJyA4zjm5ptv7jNw4MDcs846a+CePXuit23bdtDlkMOHD6/Mzc0N+Hw+TZkypejDDz9MaElGtJsj1hKtUUdI0pxJc2Qdo5odl6p6x+Wy4ThVdFmrV5YeVLqglfzmohN15e5vLifNDIXUPRxWXk2tnkk7Wwumfa5vpX/LxYQAALe1xxwXX0jKMcb0k7Rd0sWSLm2LHdmq8EGdmFDwE2mDX9IlbbHLY54xRt3n/ka7cqfuN+7NyVCPfz/qUioA7WnvmRZNHW+qhnNcNHTeeecVP/PMM9127doVddFFFxVJ0hNPPJFcWFjoW7Zs2Sq/32/T09OHV1dXH9ScP3COAuYs6DDarZZYufYDyViZqGKZcIwSnYCCJl1ryj6V1K8tdnnMi/f7dP4PH9ETb52roDFaEONXZXI3nTXgAo04+wG34wEAIkCbn3FhrQ1JukHSW5JWSZpjrV3RFvvy+Osm4Nw7z4X3TCOlLVG4/As5ZZVtsUtI8sTFKOmRn0kJsVJMtJIem0nTAjiGdDOm0Yv/DzXeUpdffnnRvHnzkl977bVul112WbEklZaWelNTU4N+v9+++uqriTt27Gi0abJs2bL41atXR4fDYc2dOzf59NNPL2+LjGhd7VlL2ITRipZRbK+5Sk5/VrMzz1ZmWpSqo//TFrtDvZ69szR74JU6p8bROYrTvZfM12U0LQAA9drjjAtZa9+Q9EZb78fEeOS9oqt2/LtMjkfqttujLuddJF9urDxd4tt698e0+PPPVPz5Z7odA4ALpsbFbH+qsjqz4eUiUZIzNS5me1vsLy8vr6aystLTo0ePQGZmZlCSrrnmmqLvfve72cOGDRsydOjQqn79+tU09tzjjz++YsaMGX1Wr14de/LJJ5dffvnlJW2REa2vvWqJk7IzNWbJFM2vmaNox6MhmxfpByddqgEZXKrQ1k457VadctqtbscAAESgdmlctCezO6SosOQEraJqraJzRyj6rES3YwFAp3VObEyRVDfXRbG10d2MCUyNi9m+d7y5qqqqvjrUsrVr1+53CUmvXr1CixcvXn247UyYMKF8woQJnGGBwzKqlY1bIafWp6Cx+o+p1Xm9R0sp3NECAAC3dKrGRcmPNslZFlCyJHmkuNtSaFoAQDs4JzamqKWNCsBtYSesqc9+R2vK+8jj6y1vUqmyxs2WutO0AADATe1xV5F24yxvcDm1I1U/Qw0NAACapiYU0vYarwJFY1Sbf46qAqWat3CB27EAADjmdZrGhbX2gDu6S3Z3SLbacScQAADoUMJhIxtVIH+Pl+Xv9YJqq/opWPSR5ITdjgYAwDGt8zQu8kMHD9ZK5Xdw33UAAHBkH28oVNewlS92i4xxFCg6Uy+WVirwyeNuRwMA4JjWaRoXnu5R8oyOPXg8tVNN4wEAANrIWYO6a2pNtiQpXNNDHl+ZhtjdCsb3dDkZAADHtk71V33XBzNU9c8i1bxYLE/PKMVf311RA2PcjgUAADqAmCivrr3xRXkWPKQP8heqT06Gbj7tI8UnUksAAOCmTtW4kKS4qcmKm5rsdgwAQCv47//+757z5s1L8Xg81uPxaPbs2ZvHjh1beeB6f/vb37rm5ubWjBx0mTWEAAAgAElEQVQ5ssaNnOhEPB5dM3amrnE7BwAA2KfTNS4AAJ3DO++8E//WW291XbZs2crY2Fi7c+dOX21trWls3ZdeeqlrKBQqpXEBAADQ+dC4AAC0WM0LJcnVTxWm28JwtEnxBmKvTtkec2HXFt2Tevv27VHJycmh2NhYK0m9evUKSdL111+f/tZbb3X1er12zJgxZZMnTy5+5513un766aeJv/nNb3rNmzdvw9ChQ2tb43UBAADAfTQuAAAtUvNCSXLV7/IzFbAeSbKF4eiq3+VnSlJLmhcXXHBB2a9//eveWVlZw0aPHl12ySWXFOXl5VW/8cYb3TZu3Ljc4/GooKDAm5qaGh4/fnzJhAkTSq+66qri1npdAAAAiAyd5q4iAAB3VD9VmL63abFPwHqqnypMb8l2k5KSnOXLl6989NFHN6elpYWuvPLKAc8++2w3v9/vXHzxxZlPP/1014SEBKdF4QEAABDxaFwAAFrEFoajj2b8aPh8Pk2YMKH84Ycf3vHAAw9see2117ouXrx41UUXXVTy0ksvdR0zZkxOS/cBAACAyMalIgCAFjEp3kBjTQqT4g20ZLtLlizxezweDR8+vFaSvvrqq9i0tLRgUVGRd+rUqaVjxoypGDhw4HBJSkhICJeVldGMBwAA6IRoXAAAWiT26pTtDee4kCRFGyf26pTtLdluWVmZ96abbupbVlbm9Xq9Nisrq3b27NlbzjnnnJy9dxe55557tkrStGnTiq677rqsxx9/vMfcuXOZnBMAAKAToXEBAGiRvRNwtvZdRU4//fSqr776avWB48uWLVt14Ni3v/3tyg0bNqxoyf4AAAAQmWhcAABaLObCrkUtbVQAAAAAjeF6YAAAAAAAELFoXAAAAAAAgIhF4wIAAAAAAEQsGhcAAAAAACBi0bgAAAAAAAARi8YFACDijBo1atC8efO6NBy7++67u0+ePDnrnHPO6e9WLgAAALQ/GhcAgIgzefLkwn/84x/JDcfmzZuXfPXVVxe8+eabG93KBQAAgPZH4wIA0GLLn9mS/JeT3hs+O+vNkX856b3hy5/ZknzkZx3a5ZdfXvzuu+8mVVdXG0las2ZN9J49e6KysrICOTk5QyVp1qxZKePGjRtw+umn52RlZQ2bMWNGr9Z4LQAAAIgsNC4AAC2y/JktyR/fvTqzak9ttKxUtac2+uO7V2e2pHnRs2fP8IgRIyrnzZuXJElPP/108sSJE4uNMfutt3Tp0vjnn39+4/Lly1e88soryR988EFcC18OAAAAIgyNCwBAiyx8ZEN6uNbZ7/dJuNbxLHxkQ3pLtjtlypSif/7zn90k6YUXXki+/PLLiw5cZ/To0WU9e/YMJyQk2HPPPbd4wYIFCS3ZJwAAACIPjQsAQItU5ddGH814U02bNq3k448/7vLRRx/F1dTUeEaPHl114DoHnoFx4GMAAAB0fDQuAAAtEpfmDxzNeFMlJSU5p5xySvk111yTdeGFFx50toUkffTRR112797traioMG+88UbXM888s6Il+wQAAEDkoXEBAGiRvOkDtnv9HqfhmNfvcfKmD9je0m1ffPHFRWvWrIlt7DIRScrLy6uYOnVqv2HDhg0977zzis8444yDzsoAAABAx+ZzOwAAoGMbdlnfIqlurouq/NrouDR/IG/6gO17x1viiiuuKLniiisW7X08aNCgwLp161bsfZyamhr661//uqWl+wEAAEDkonEBAGixYZf1LWqNRgUAAABwIBoXAIAO6aabbiqUVOh2DgAAALQt5rgAAAAAAAARi8YFAKAxjuM4x9y9Retfs3PEFQEAANBu2rRxYYy50xiz3RizuP7re225PwBAq1men5+fdCw1LxzHMfn5+UmSlrudBd+glgAAAO0xx8XD1toH22E/AIBWEgqFrtm1a9cfd+3aNUzHztl5jqTloVDoGreD4CDUEgAAHMOYnBMAcJCRI0fukTTR7RwAAABAexxFu8EYs9QY8ydjTLdDrWSMudYYs9AYszA/P78dYgEAgA7iiLUEdQQAAJ2Xsda2bAPGvCOpZyOL7pD0qaQCSVbSryT1stb+8EjbzMvLswsXLmxRLgA4VhljFllr89zOATRVa9cS1BEA0DLUEog0Lb5UxFo7vinrGWP+IOm1lu4PAAB0LtQSAADgcNr6riK9GjycJGZqBwAAR4FaAgAAtPXknP9rjDledad3bpL0X228PwAA0LlQSwAAcIxr08aFtfbyttw+AADo3KglAABAe9xVBAAAAAAAoFloXAAAAAAAgIhF4wIAAAAAAEQsGhcAAAAAACBi0bgAAAAAAAARi8YFAAAAAACIWDQuAAAAAABAxKJxAQAAAAAAIhaNCwAAAAAAELFoXAAAAAAAgIhF4wIAAAAAAEQsGhcAAAAAACBi0bgAAAAAAAARi8YFAAAAAACIWDQuAAAAAABAxKJxAQAAAAAAIhaNCwAAAAAAELFoXAAAAAAAgIhF4wIAAAAAAEQsGhcAAAAAACBi0bgAAAAAAAARi8YFAAAAAACIWDQuAAAAAABAxKJxAQAAAAAAIhaNCwAAAAAAELFoXAAAAAAAgIhF4wIAAAAAAEQsGhcAAAAAACBi0bgAAAAAAAARi8YFAAAAAACIWDQuAAAAAABAxKJxAQAAAAAAIhaNCwAAAAAAELFoXAAAAAAAgIhF4wIAAAAAAESsVmlcGGMmG2NWGGMcY0zeAcv+xxiz3hizxhjzndbYHwAA6DyoIwAAwOH4Wmk7yyVdKOmJhoPGmFxJF0saKqm3pHeMMQOtteFW2i8AAOj4qCMAAMAhtcoZF9baVdbaNY0sOl/Sc9baWmvt15LWSxrVGvsEAACdA3UEAAA4nLae4yJd0tYGj7fVjx3EGHOtMWahMWZhfn5+G8cCAAAdAHUEAABo+qUixph3JPVsZNEd1tqXD/W0RsZsYytaa5+U9KQk5eXlNboOAADomKgjAABAczW5cWGtHd+M7W+TlNHgcR9JO5qxHQAA0IFRRwAAgOZq60tFXpF0sTHGb4zpJylH0udtvE8AANA5UEcAAIBWux3qJGPMNkmnSnrdGPOWJFlrV0iaI2mlpDcl/YSZwAEAQEPUEQAA4HBa5Xao1toXJb14iGX3Srq3NfYDAAA6H+oIAABwOG19qQgAAAAAAECz0bgAAAAAAAARi8YFAAAAAACIWDQuAAAAAABAxKJxAQAAAAAAIhaNCwAAAAAAELFoXAAAAAAAgIhF4wIAAAAAAEQsGhcAAAAAACBi0bgAAAAAAAARi8YFAAAAAACIWDQuAAAAAABAxKJxAQAAAAAAIhaNCwAAAAAAELFoXAAAAAAAgIhF4wIAAAAAAEQsGhcAAAAAACBi0bgAAAAAAAARi8YFAAAAAACIWDQugAhWG9qqrdX3qSK0xO0oAACgA9q16QMt/dfPZEu2uR0FAJqNxgUQocpDX2hNzaUqDv9LG2tmytqQ25EAAEAH8tiiR3T2gus1bffbeuztG9yOAwDNRuMCiECODWpzzd0NRqq1vHKiasM7XcsEAAA6jk2lm/Tk8j/ue/xE1Xot+vNYKVjrYioAaB4aF0CEqQlt0/LKiXJUst+4VbnWV1/tUioAANBRvLV8p877x0/l2P3Hr7G79MncS9wJBQAtQOMCiDBban4hqaLRZWFVt28YAADQ4dz6xt8V7/ta0v6di5DHo3+YcndCAUAL+NwOAKCOY2tkrVSjjYdZK6RguFBR3pR2ywUAADqImjKVBCvUI+mvCjsBVcp70CqLgjQuAHQ8NC4Al1nrKGj3aFP1f6vGbpLkHHZ9RzXtkgsAAHQMjmPl+XqBSl/5H01KLFdBtCTjlWR04FkX1U6lGxEBoEVoXAAusTak3YE/qTj4vkKqkdWeJj3PY/xtnAwAAHQEJVUB3f3qSn25pUhDvc9rZ2yh4qxXfcLSNq9XMkaDq4zWREvWV9fA6GZiXE4NAEePxgXgkurwRhUH31VQOw6zVhdJNZJCkjyK8xwnj4ltn4AAACCifbJmuz7Z9rmCCf/QB7H5SgxHaWp5hWqN0d+7JMprrR6uTVTtnnJ9GO/o5RS/8gZNdDs2ABw1GheACyrDK7S55m6FDtu0kDyKUrxnmIIqVIwnU338t8ljotopJQAAiFhfPaPwx79RbbIUMnVnU5R7PZqTGK9YxyO/Y3RDcYXC8YPULSVO36ks0xX9TpXn1NtcDg4AR4/GBeCCosBbR2xaSFHyGp8SfCOVFj2lXXIBAIAO4j+/1/2xIfmsUch8c6PAMq9XjoyiSvPUr2abevh3KOZb90jZ41wMCwAtw+1QgXZmrVVx+JUjrudTisK2TGHL7N8AAKCBgnUqKFipEbUBTS2vUFdn/4m9K7xWpstynRq7XTFVO6UQE3sD6Ng44wJoZ8YYGfllVXXY9eK9uXJstdKiprVTMgAA0CEk9dGvU7opZIzOqaxShfFoXmK8oq1VwFN3XDLPsfKmDJDSBkmDz3U5MAC0DI0LwAVJ3rNUEn79kMuNYpQZe1c7JgIAAB1GVKzSkvrp7dBufR7TQ2FjdH1xib5VXas7U7tpfXS0+qdmSxc+53ZSAGgVXCoCuKBPzHRFa8AhlyfojHZMAwAAOpr/nvAX9bcehY2RJAWNkUd1xb2xVtcMu8rVfADQmlqlcWGMmWyMWWGMcYwxeQ3Gs4wx1caYxfVfj7fG/oCOzmNilRo16RBLu6lfws/bNQ8AuIk6Ajh6pmuGzu55qmTr7ijyVFIX/bhHmtZGR+uRuBGKHfgdlxMCQOtprTMulku6UNIHjSzbYK09vv7rx620P6DDS4k+Tz19N8ijxH1jfg3X8PiXXUwFAK6gjgCa4YazHtD8jMnqLZ8caxUw0vTcq3XW1L+7HQ0AWlWrzHFhrV0l1U06CKBpjPGoe8xUdddUhZ1qhWyhoj29+H8E4JhDHQE0U0wXpY2/U2/pTqmyQKXhWiV1SXc7FQC0uvaY46KfMeYrY8z7xpjT22F/QIfj9cTK7+0jY7xuRwGASEMdATRFfCpNCwCdVpPPuDDGvCOpZyOL7rDWHurc9p2S+lprC40xIyW9ZIwZaq0ta2T710q6VpL69u3b1FgAAKADoI4AAADN1eTGhbV2/NFu3FpbK6m2/vtFxpgNkgZKWtjIuk9KelKS8vLy7NHuCwAARC7qCAAA0FxteqmIMSbN1J/7bozpLylH0sa23CcAAOgcqCMAAIDUerdDnWSM2SbpVEmvG2Peql90hqSlxpglkuZK+rG1tqg19gkAADoH6ggAAHA4rXVXkRclvdjI+DxJ81pjHwAAoHOijgAAAIfTHncVAQAAAAAAaBYaFwAAAAAAIGLRuAAAAAAAABGLxgUAAAAAAIhYNC4AAAAAAEDEonEBAAAAAAAilrHWup3hIMaYfEmb23g3qZIK2ngfbYHc7Yvc7acjZpYiM3emtTbN7RCAW9qpjpAi8/9/U5C7fXXE3B0xs0Tu1kQtgYgSkY2L9mCMWWitzXM7x9Eid/sid/vpiJmljpsbQMt11P//5G5fHTF3R8wskRvozLhUBAAAAAAARCwaFwAAAAAAIGIdy42LJ90O0Ezkbl/kbj8dMbPUcXMDaLmO+v+f3O2rI+buiJklcgOd1jE7xwUAAAAAAIh8x/IZFwAAAAAAIMLRuAAAAAAAABHrmGtcGGMmG2NWGGMcY0zeAcv+xxiz3hizxhjzHbcyHokx5k5jzHZjzOL6r++5nelwjDHn1L+n640xt7mdpymMMZuMMcvq39+Fbuc5FGPMn4wxe4wxyxuMJRtj/m2MWVf/bzc3MzbmELkj/nNtjMkwxrxnjFlV/3Nkev14xL/nAFoPtUT76oh1hEQt0dY6Yi1BHQE03zHXuJC0XNKFkj5oOGiMyZV0saShks6RNNsY423/eE32sLX2+PqvN9wOcyj17+Fjkr4rKVfSJfXvdUdwVv37G8n31f6L6j6vDd0m6V1rbY6kd+sfR5q/6ODcUuR/rkOSZlhrh0g6RdJP6j/PHeE9B9B6qCXaSQevIyRqibb0F3W8WoI6AmimY65xYa1dZa1d08ii8yU9Z62ttdZ+LWm9pFHtm65TGiVpvbV2o7U2IOk51b3XaAXW2g8kFR0wfL6kp+u/f1rSBe0aqgkOkTviWWt3Wmu/rP++XNIqSenqAO85gNZDLdGuqCPaGLVE+6GOAJrvmGtcHEa6pK0NHm+rH4tUNxhjltafJhfJp5N1tPd1LyvpbWPMImPMtW6HOUo9rLU7pbpfkJK6u5znaHSUz7WMMVmSTpD0mTr2ew6g9XS033kd4WduR3tPG6KWcEdH+FxTRwBHqVM2Lowx7xhjljfydbgOvWlkzLV7xR7hNfxe0gBJx0vaKekht3I2QUS9r0fhNGvtiao7NfUnxpgz3A50DOgwn2tjTIKkeZJuttaWuZ0HQOujlogYEfWeHiVqifbXIT7X1BHA0fO5HaAtWGvHN+Np2yRlNHjcR9KO1kl09Jr6Gowxf5D0WhvHaYmIel+bylq7o/7fPcaYF1V3quoHh39WxNhtjOllrd1pjOklaY/bgZrCWrt77/eR/Lk2xkSprth41lr7Qv1wh3zPARwatUTEiKj39GhQS7S/jlBLUEcAzdMpz7hoplckXWyM8Rtj+knKkfS5y5kaVf8Dba9JqpskLFJ9ISnHGNPPGBOtuknLXnE502EZY+KNMYl7v5f0bUX2e3ygVyRdWf/9lZJedjFLk3WEz7Uxxkh6StIqa+1vGyzqkO85gFZHLdH6OlwdIVFLuCXSP9fUEUDzGWs7ytl2rcMYM0nS/0lKk1QiabG19jv1y+6Q9EPVzfh7s7X2X64FPQxjzN9UdwqclbRJ0n/tvS4uEtXfiup3kryS/mStvdflSIdljOkv6cX6hz5Jf4/UzMaYf0gaIylV0m5Jv5T0kqQ5kvpK2iJpsrU2oiavOkTuMYrwz7UxZrSkDyUtk+TUD9+uuutTI/o9B9B6qCXaV0erIyRqifbQEWsJ6gig+Y65xgUAAAAAAOg4uFQEAAAAAABELBoXAAAAAAAgYtG4AAAAAAAAEYvGBQAAAAAAiFg0LgAAAAAAQMSicQEAAAAAACIWjQsAAAAAABCxaFwAAAAAAICIReMCAAAAAABELBoXAAAAAAAgYtG4AAAAAAAAEcvndoDGpKam2qysLLdjAECHtGjRogJrbVoLt9Hd5/P9UdIw0eTGkTmSlodCoWtGjhy5x+0w1BEA0DLUEmhnR6wjIrJxkZWVpYULF7odAwA6JGPM5pZuw+fz/bFnz55D0tLSij0ej22NXOi8HMcx+fn5ubt27fqjpIlu56GOAICWoZZAe2pKHUHnCwDQmGFpaWllFBpoCo/HY9PS0kpVd1QNAACJWgJN1JQ6gsYFAKAxHgoNHI36zwt1BQBgL2oJNNmR6ggKDAAAAAAAELFoXAAAAAAAgIhF4wIAEJG8Xu/IwYMH5+79uv3223u2xnZnzZqVcsUVV/Q9cPzGG29M79mz53FxcXEnHO75p59+es6gQYNys7Ozh1566aV9Q6HQYff3i1/8oke/fv2G5uTkDB00aFDuo48+miJJ6enpw3fu3LlvkuzXXnst8ayzzsqWpDvvvLPHlClTMvcu+/3vf588ZsyYfcsGDBgwdODAgbmnnnrqwLVr10Yf1RsAAMAxglqi89QSNC4AAC32zKebk0fd+87wfre9PnLUve8Mf+bTzckt3abf73dWr169cu/Xfffdt6s1sh7KBRdcUPLZZ5+tOtJ6L7/88oY1a9asXLt27YrCwsKoP/3pT90Ote7//u//ps2fP7/LokWLVq1bt27FJ598ssbaI1/ue8cdd+xeunRp/Ntvvx1fUFDgveeee9Jnz569RZJGjhxZtXjx4lVr165decEFFxT/9Kc/7XNULxQAgAhELdE4aok6NC4AAC3yzKebk3/12srMPeW10VbSnvLa6F+9tjKzNQqOAxUWFnqzsrKGLVmyxC9J5513Xr+HHnooVZLmzp3bJTc3d8igQYNyTz311IFHu+1x48ZVZmZmBo+0XnJysiNJwWDQBINBY4w55LoPP/xwzyeeeGLL3uekpKSEb7zxxsIj7SMqKkr/93//t/mmm27KvPHGG/tceumlBbm5uQFJOu+888oTExMdSRo9enTFzp07I/4oCQAAh0MtQS1xJL4jrwIAwKHNenddem3I2a8RXhtyPLPeXZd+2SmZRc3dbm1trWfw4MG5ex/PmDFj549+9KPihx9+eMuVV17Z7/rrr99dUlLimzFjRsGOHTt8N9xwQ9aCBQtWDx48OLB7925vS17TkYwePTpn6dKl8WeeeWbpVVddVdzYOsXFxZ7Kykrv0KFDa5uzj7PPPrsyJyen+sMPP+yybt265Y2t88QTT6SNHz++tDnbBwAgUlBLUEscCY0LAECL5JfXNtqlP9R4U+09vfPA8UmTJpXNmTOn26233pq5aNGiFZK0YMGC+FGjRpUPHjw4IEk9evQIt2TfR/LRRx+tq6qqMpMmTer/6quvdpk0aVLZgetYa3W4IyiNabh+aWmpZ+nSpfGhUMjs2LHDN2DAgP2O4MyePTt5yZIlcU888cSa5r4OAAAiAbUEtcSRcKkIAKBF0hL9gaMZb6lwOKy1a9fG+P1+p6CgwCc17xd7U4RCIe2d0Ovmm2/u3XBZXFycnTBhQsmLL77YtbHnJicnO7Gxsc7KlSsbLbq6desWKigo2Hc0p7Cw0JucnLxvdq5bbrml9/e///3Cn/3sZztvuOGGjIbPfemllxIffPDBXm+88cb62NjYI1/oCgBABKOWoJY4EhoXAIAWuWlczna/z+M0HPP7PM5N43K2t8X+7r777h4DBw6sefrppzdeffXVWbW1teass86q/OyzzxJXr14dLUmtdXqnz+fT3gm9fve73+0oLS31bN68OUqSgsGg3nzzzaTBgwdXH+r5N998884f//jHmUVFRR5JKioq8jz44IOpkvStb32r/KmnnkqR6oqaZ599NmXMmDHlkvT555/HvvPOO0m/+tWvds2YMSN/27Zt/hdffLGLJH388cexN954Y+bLL7+8Pj09/fDTkAMA0AFQS1BLHAmXigBABKmoDemFL7fp1P4pyumR6HacJtl77emsd9el55fXRqcl+gM3jcvZ3pJrUqWDr0sdO3Zs6Y9//OOCv/3tb6mLFi1a1a1bN2fu3Lnlt912W6+HH354x6xZszZNmjQp23EcpaSkBD/55JN1h9r23LlzU9566619Rzc++eSTVQ888ECPF198MbmmpsbTo0eP46ZNm1bw29/+dkfD55WVlXnOPffc7EAgYBzHMaeddlrZzJkz8w+1n1tvvTW/oqLCc+KJJ+ZGRUVZn89nb7zxxl2S9Otf/3rnD37wg76DBg3KtdZq7NixZdddd12h4zi67rrr+v7mN7/ZGhcXZyXpscce23zVVVf1++53v7ty5syZGVVVVd7JkycPkKTevXsH5s+fv7757zQAdAwbv/pC4VBIOSed6nYUtDJqCWqJIzFNuZVKe8vLy7MLFy50OwYAtLtXl+zQkx9s1IiMJN1zwfBmbcMYs8ham9eSHEuWLNk0YsSIgpZsA8eeJUuWpI4YMSLL7RzUEUDnE6yt0T/vvE2SdOFtd8pxHH32wj818JTTlDH0OJfTdT7UEnDD4eoIzrgAgAgyOjtVmwoqdfrANLejAAAQMaL8MRpx9vcUDgYV2yVJr/z2Pm1a/KUKt29VxtDj5DhheTxtehMIAC6icQEAEaRbfLRuHJfjdoxO4ZFHHkn5/e9/36Ph2EknnVTxt7/9bUtr7+vyyy/v+8UXXyQ0HLvuuut2T58+/Yj3WQcANM3wsd/e933fYSNUtG2rjv/2uXrnj4+paPs2fe+mmUroluxiQnQ21BKRg0tFAKCT4fROuIVLRQC0BScc1n/mPSd/bKzyzrtQpXt265Xf3qeygnz1yh6oXevXKRCoUXS0X0k9euqSXz3A2RctRC0BN3CpCAAAAIAO58t/vaqSXTu0fc1KebxenfDdiVq/8FOV7tmt2uoqbVmxVOFA3R0za4IhOTt2KBwMyuOncQF0JjQuAAAAAEQUxwmreMd2rfp4gWw4rLzzLpQ/Nk5v/v532rp8iWorKyVZyVoZj0e+6GjFJXWVLypaHi9NC6CzoXEBAAAAIGJYx9GrD/1aO9atVlL3Hjrhe+dryGlnatOSL1W8Y5t8UdGSrGSM+gweqpMnTdGiN15WdGy84pK6KBwKyeuLcvtlAGhFHrcDAADQGK/XO3Lw4MG5e79uv/32nq2x3VmzZqVcccUVfRuOlZeXe8aMGZPdr1+/odnZ2UOvv/769CNtZ+zYsdk5OTlDD7fORRddlJWenj588ODBuYMGDcp9+eWXE/cuq6mpMT/84Q8zMjIyhmVmZg4bN27cgA0bNuyrtPe+/pycnKFjx47NLigo8ErSmjVromNiYk7cu80TTjhh8JIlS/xH/04AQOQp3rVDFcVFyt+6SVVlZdr99QYte/ctWcdR5vDj9a0p0zTo9DFKTO2u4WO/rYxhI7Rs/tvavHSxtq1cqi3Ll2rB039QsKbG7ZeCCEAt0XlqCRoXAIAW27irKvmNRfnDX/x098g3FuUP37irqsXTuvv9fmf16tUr937dd999u1oj66HMmDFj99dff71i+fLlKz/77LOEOXPmdDnUuk8//XTX+Pj4cFO2e88992xbvXr1ygcffHDrTTfdlLl3/KabbkqvqKjwfP3118s3b968fOLEiSUXXHBBtuM4kr55/evWrVvRtWvX0AMPPLDvHrkZGRm1q1evXrlmzZqVl156acFdd93VqwUvHQAiwqqPFuj1R/5X6z7/RMm90uXxemQdR3s2bdSif72itZ9/on//4TF9/sI/VV6Qr6+/Wqi1//lQHl+UUjL66rjx31WotlZbVy7T/2fv3uOirvL/gfbXx+cAACAASURBVL/OzDAzwIDAgAOMwqBchgHFFCjXshX7mlZkpH3L8rJl9lXXlpJCM7f159fvVpZL0jdtLTeJbStjzW5uVt803TAVqkHUAW8IDNdhuMPcz+8PnULl6qBcfD8fj3ngnM/nc8778Nj28+Z8zjmfY/u+GujukD6iXKJrlEvQwAUhhBAXna1q8zt2vjnUbHWIAcBsdYiPnW8O7Y+E43J1dXVClUoV63wqkJycHLZp0yZ/AMjJyfHWaDTRUVFRmsmTJ0f2pV4vLy9HcnJyMwBIpVI+fvz4trKyMnFn5zY2NgoyMzMV69atq+xLG9OnT2+pqalxAy48ldm5c6f/m2++WSYSXVi1mZqaWicWix2fffaZ1+XX3nLLLa16vb7TeJqamoQ+Pj69SnwIIWQw8/D2gd1uR0NVJcp1J+Cw2SHzk0MkFuPHLz7Bt397E631xotnc5jb2+Dh4wuxuxRx/zEL8tEhcNhsaDHWwagvH9C+kL6hXKJ3buRcgva4IIQQ4hKdvlXp4JcOhDs4BDp9q3JMoIexq+t6YjabBWq1WuP8npaWVrlkyZL6jIyM0kWLFoUtX768uqGhQZSWlmaoqKgQrVixQrV//36dWq22VFdXX/XObAaDQfj111/7PPvss9WdHV+5cqUyNTW1WiaTOfpS7z//+c8Rd9xxRwMAnDhxQhIUFGTx8/O7pI4JEya0HTt2zH327NnNzjKbzYZ9+/Z5LV68+JdXypWVlUnUarWmtbVVYDKZBLm5ubq+9ZIQQgaf0PETYLda8PNXe8AdF/6GEokl4JwDAgaZrx8CwyNgrKhAc10tHLYLG3g2G2pQc+4sJt09G+7e3mACAYRutMfFUEK5RO/cyLkEDVwQQghxifPpSG/Le8s5vfHy8pSUlKadO3f6pqenh+bn5x8HgP3793smJiY2q9VqCwAoFIqrempgtVpx//33j3niiSeqNRqN5fLjubm57ufOnZNs3769rKioqFf9W7t27ag//vGPo4xGo+i77747CQAOhwOMMX75uZxzMMYA/Jps6fV6cWxsbNt9993X5DzPOb0TAN566y3fxx57LPTgwYOnrqbPhBAy0CpPFcFiaofE3QPm1tZfBi0AoKFSD4GbGyRSdzS1tcFmtcLLzw92iwVhN02Cl9wfjQYD4LDj3x/+HW0N9Ui4by4mzbp3AHtE+opyie5RLkFLRQghhLhI4ia44qbcXbmr7HY7iouLpRKJxGEwGETApTdpVzz88MOqMWPGmF544YUa4MITCueGXk899VTwwYMHZYWFhR5KpXLc1KlT1SUlJZLExMSo7urcsGFD+fnz54+tXr1a/7vf/S4MAGJiYswVFRWS+vr6S+7DBQUFHrGxse3Ar8lWSUnJMYvFwl566aWRndU/b968hry8PJnLnSeEkAHgsNvx7Y6/4uA/duDgB+9CJBFD7OEJL/+L/5fHGNzEEshHh8I3eDREYjEm3JkMZXQMEu6dC98gJU4d/h7Fh7+Hu5cXhG5uaK6tgaH0PPRX/r1KBinKJSiX6AkNXBBCCHGJWumpFzBcMk1RwOBQKz3116K99evXKyIjI01ZWVlnFy9erDKbzWzatGmthw8f9tLpdGIAuJrpnX/4wx+Cm5qahNu3by9zlolEIjg39HrttdcqVq1aVVtTU1Og1+uPHThwQKdSqcxHjhwp6qluoVCItWvX1jgcDvbPf/7T29vb2zF37lzDsmXLRttsNgDA//7v/8pNJpPAuT7WSS6X2zMzM0vfeOMNhdlsviKj+vrrr71Gjx5t7mt/CSFkMBAIhYiecjuCI6PRYqwDHIDPSAVGjFRAJJEgIESFEQEKtDc3XVgeYrfDUHYe9RXlKP7h3wgIHQM3sQQCgRDjp8/EjCdWIDgiGl+/9b/Yl7UNbY0NA91F0guUS1Au0ZMel4owxv4G4B4ANZzz2Itl6wAsAVB78bQ1nPM9nVw7E8BmAEIAb3POX+qnuAkhhAwSzrWnOn2r0mx1iCVuAota6al3ZU0qcOW61KSkpMalS5casrOz/fPz80/6+vo6cnJymlevXh2UkZFRkZmZWZKSkhLucDggl8utubm5XU53zMnJke/du9fH+T03N/fk66+/HhQWFmaKiYnRAMATTzxRs3LlSkNXdfSVQCDAqlWrKl599dXAOXPmNL3++uv6pUuXjgoLC4sVCAQYO3asaffu3acFgiufKUyZMqU9Ojq6/e233/a94447WpzrUjnncHNz42+++eb5/oqTEEKul5qSs6ivqkBgRBTaW5phaW+DMioGblIpCvd/DXAOP+UoaKZOR+kxLU78ex/EUilibpsGoUgE36BguEkluH3hYpzX5iN03ASc/ekotF/tgcRTBg/vESg7UYioybcOdFdJDyiX6J0bOZdgnF+xLObSExibCqAFwLuXDVy0cM5f7eY6IYBiAP8BoBzAUQDzOOc9ztmKj4/neXl5ve0DIYSQDhhj+ZzzeFfq0Gq1JXFxcf12oyU3Bq1W6x8XF6e62uu7eFjiB+BDACoAJQD+k3Ne3109lEcQMjS8/0I6as+fg8Nug5vUHTJfX9zz4F1oLjmGf370b4AxJNyTgqnzH0OLsQ4/f/UFvAMU0H79L3j7B6C5rhYjRgbCYbejxWjA3amrYDW148gnOdBMTcL3H/4dAMe9ac/D2z+gx3jIryiXIAOhuzyixxkXnPMDjLFOL+5BIoDTnPOzAMAY+wDAbAC02IwQQgghndkB4H8BvNuhbDWA/+Ocv8QYW33x+6oBiI0Q0s+iJt8Gc1sL6iv0gIMjZtoMyE/9HfVnquGvCIV8jBpT5z8GAKgtLYEu9yBMzU0wt7Wirvw8oibfhsCxEWisqYLdaoHY3R0+ikDc/YdnAQCtDfUwt7bCy08+kN0khPQDV94qsoIxthBAHoC0Tp5+KAGUdfheDuBmF9ojhBBCem3z5s3yrVu3KjqWJSQktGRnZ5f2d1sLFiwIOXr06CWbWi1btqw6NTW1rr/bGs66eFgyG8BvL/47C8B+0MAFIcPCxFnJuGnmPfjH2mdQfe40Du18D+I7JyO/7BAaWxoQECWCw2GHQCBE5SkdrKZ2cABMIIDY3ROquInI/+IT3DL3Idz60CL89OVncPf2QdTkKXCTSBH72zsGuotkiKNcYvC42oGLrQD+GwC/+HMTgMcuO6ezLVm7XJfCGHsCwBMAEBIScpVhEUIIIRekpqbWXa+b/bVIYMgvFJzzSgDgnFcyxjrdEZ0QMjS11hthamkGHA5Y2tvw3VeHYGlrAwAU5R5EfUUFJs99CJZ2EwDA08cXbpJAtDU24PDuj9BYXYWCr7+E1FOGwn1fo7WhHlVninHH4mUD2S0yTFAuMXhc1VtFOOfVnHM759wB4C1cWBZyuXIAozt8HwWgops6t3HO4znn8QEBtAaNEEIIIb3HGHuCMZbHGMurra3t+QJCyKAg85PjlvsfRNyd90Dg5gZLWxsEIhFkfv5wl3nBWF4K7Vf/QtmJAlja29De3ARzWxscdjvCJyUiQDUGMbcnISg8CtG3/RbeIxW0nwUhw9BVDVwwxoI6fE0BUNjJaUcBRDDGwhhjYgAPAfj0atojhBBCyA2r2pl3XPxZ09lJ9ACEkKEr4ubfoLbkLNjFCdvc4YDHCB+Mnz4TIokEtSVnME5px623xSI8/maMS5qB0HETUFLwE+Lvno2oybeBCRhqz5eAAWisqUJ7S3P3jRJChpQeBy4YY+8DOAQgijFWzhhbDGAjY+wYY6wAwDQAT188N5gxtgcAOOc2ACsA7AVwEsBOzvnxa9QPQgghhAxPnwJYdPHfiwB8MoCxEEL62bmf87Hz/62B2N0d0bf9FiPHREIgFMFQeg4/f/U5gqM0sLc1wHH+B0zyKcHkOfNw+kgujBXlcJNIIbu48WZLcwtaGxvQXGdA5eliGMvLemiZEDKU9OatIvM6Kd7exbkVAO7q8H0PgD1XHR0hhBBCbhgXH5b8FoA/Y6wcwJ8AvARg58UHJ6UAHhi4CAkh/c1mMQPgUMVNgsxPjsrikxCJxbDbrBAIRQgMj0S9/izEPnYg4XGIJBK4e3lDJg+Aw26DXncC/qqxOF9aiqi770egrw/ajXUINlsAgwHw9x/oLhJC+sFVLRUhhBBCrjWhUDhJrVZrnJ81a9YE9ke9mZmZ8oULF3a5C3RSUlJ4RERETFfHlUrluMjISI1ardbExsZGd9fWnDlzVEqlcpyzDxs2bBgJAHV1dcKUlBTV6NGjY0ePHh2bkpKiqqurEwLAgQMHPCIiImJMJhMDgOPHj0tGjRo1zmg0Cj7//HMvLy+vCdHR0ZoxY8bEpKWlBQHAxx9/7B0TExMdGRmpiYmJif7000+9ruZ3M9A45/M450GcczfO+SjO+XbOeR3nfDrnPOLiT+NAx0kI6T8Rib9Byqo/wUcRiNyP3kNbYxOEIhGUURrEzbgLh3d9CFOrGcfM0fj34RJ4eI/A/c/9P0yclYzKM6dwruAniEQiyGQy+AcGISg8EmPkI8FOnwZO0GTvGx3lEsMnl3DldaiEEEIIAKDqnNGvrMigtJptYjeJyDI6yl8fGObn0h+YEonEodPpTvRXjL2RlZXl4+npae/pvO+++644KCjI1ps6N2zYUP7oo49e8srwRx55JFSj0Zg+/vjjQgB4+umng+fPnx/6r3/96+zUqVPbbr755uZ169YpXnrppaqlS5eG/PGPf9T7+fk5ACA+Pr5l3759p5uamgTjxo3T3HfffY0jR460fvHFF6dVKpX16NGj0rvvvjuypqam4Gp+B4QQcr15+vjC2z8AMl8/xN0xC6aWZshHh0DqKcNx72/g7jUCDdUVsLS3IT55DsytLXAIhPALVyNIMw4CgQBhYWG/VjhqFNDcBChHDVynSJ9RLtE1yiVoxgUhhBAXVZ0z+p0rrA61mm1iALCabeJzhdWhVeeMfv3dVl1dnVClUsVqtVoJACQnJ4dt2rTJHwBycnK8NRpNdFRUlGby5MmRfa27sbFRkJmZqVi3bl1lf8fdUWFhoeTYsWOeGzdu/OVNW6+88kpFQUGB5/HjxyUA8Nprr+n//ve/B6xdu1Zhs9nYf/3Xf12RuHl7ezvGjRvXVlRUJJkyZUq7SqWyAsCkSZNMFotF0N7e3tlryQkhZFDyD1Fh7vP/jZtm3oPJc+ch8uYpCImNw+Ovb8ek5BQIhG4Qurnh622v47OMl5D38YcwntZBZDFdWZlEAky4Cbi4Ua/ZbAbn/Dr3iPQF5RJ9cyPmEjRwQQghxCVlRQYld/BL7ifcwQVlRQalK/WazWZBx+mdb731lq9cLrdnZGSULlq0KGzbtm2+DQ0NorS0NENFRYVoxYoVql27dp0pKio6sXv37jN9bW/lypXK1NTUaplM5ujp3OnTp0fExMREv/rqqz0unl67du0oZx+OHDnirtVqpRqNpk0k+nXSo0gkgkajafv555+lAODv729PTU2tfPHFF0e9+eab5zurt6qqSvjTTz95Tpgwob1jeVZWlq9Go2lzd3enLJ0QMiwc2f0R2hrr0VBdhfaWJkhlMoxSa2Bpa0V18UkAgMNux/c7/44jn+Rccm1tbS2Ki4tRU9PpC4nIIEG5RPcol6ClIoQQQlzkfDrS2/Le6mp6Z0pKStPOnTt909PTQ/Pz848DwP79+z0TExOb1Wq1BQAUCkWPUzQ7ys3NdT937pxk+/btZUVFRd3G/f333+tUKpVVr9eLkpKSImNiYkyzZs1q6er8y6d3FhcXSxi78gEG5xwdy/fu3TtCLpdbtVqte1xcnNlZnpeXJ4uOjtYIBAKemppaFR8fb+pwTPrCCy8ov/zyy1O97z0hhAxu4fG3oF5fDnNrC8oKC6CKm4gmQy1kvn6wtLUBANqbm3DupzwwgQCT7p4NocgNACAUCgFc+KOODF6US1Au0ROacUEIIcQlbhKRpS/lrrLb7SguLpZKJBKHwWAQAVfeqPvq4MGDssLCQg+lUjlu6tSp6pKSEkliYmLU6dOn3ZxPODZu3BgAAM5plEql0nb33Xc3HDp0yLMvbU2YMKH9+PHjHnb7r/mQ3W7HyZMnPcaPH28CgPfff39Ec3Oz8PPPPz+1du3aUc3Nzb/cr+Pj41tOnjx54vjx4yfT09NrneVnzpxxmzt3bvj27dvPxcTEmEEIIcPEbx54GLOffR5uUneI3T3Q2lAPY0U5Zqc9j5nLnwZwYZ+M3y54HNMfW/bLoAUA+Pn5ISYmBnK5fKDCJ71AuQTlEj2hgQtCCCEuGR3lr2cCdsmUSCZgjtFR/vpr0d769esVkZGRpqysrLOLFy9Wmc1mNm3atNbDhw976XQ6MQBUV1cL+1LnqlWramtqagr0ev2xAwcO6FQqlfnIkSNF4eHhVp1Od0Kn051IT0+vbWpqEtTX1wsAoKmpSbBv3z7v8ePHt/dUf0exsbHmmJiYtlWrVgV1aD8oNja2LTY21tzS0sKee+650Vu2bClNTExsv/POOxvWrFkT1F2dBoNBeNddd0WsW7eufMaMGa19iYcQQoYCVdwk/CHrIzy5YyfuXJqKpEeXQuYnh8TD45dzRmliETg24oprBQL6k2ewo1yCcome0H/FhBBCXBIY5mcMi1Wcdz4VcZOILGGxivOu7gR++brU5cuXKwsKCiTZ2dn+W7ZsKZs5c2bLLbfc0rx69eqg4OBgW2ZmZklKSkp4VFSUJiUlZUx3defk5MgVCsV45+fMmTNu3Z3vVF5eLrrlllvUUVFRmokTJ0bPmDGjYe7cuU197dt7771XcurUKWlISEjs6NGjY0+dOiV97733SgBg9erVwbNmzaqfNGmSCQA2btxYsXv3bt9jx45Juqpv48aNI0tLSyUvvfRSsPP3pdfraV40IWRYUowJh9VsgsXUp7/1yCBGuQTlEj1hg3GH3fj4eJ6XlzfQYRBCyJDEGMvnnMe7UodWqy2Ji4sz9FdM5Mag1Wr94+LiVAMdB+URhAxvVaeL8c32LRgZNhYznnhyoMMZliiXIAOhuzyCZlwQQgghhBBChg7G0NbYAJt5UC3BJ4RcQ4Nm6gchhBDSnzZv3izfunWromNZQkJCS3Z2dml/t7VgwYKQo0ePyjqWLVu2rDo1NbWuv9sihJAbXeG+r8E5MPjmjZPhhnKJwYMGLgghhAxLqampddfrZn8tEhhCCCFXsrS3oepMMYRuIkyade8lx37+ag8s7W1ISL4frIcNOV19gwS5MVAuMXjQwAUhhBBCCCFkSBC7e2DqI4/CYbcjMDzyl3K7zYrj+78G5xya26ZB5idHk6EGIjcxPEb4XFKHwWBAZWUlRo8eDR8fn8ubIIQMQjRwQQghhBBCCBkyQmLjAADH9n0Fc2srJt01G0KRG26dtwg2sxkyPznaGhvw+Wsvw00iwdznN1wyA8NqtQIAbDbbgMRPCOk7GrgghBBCCCGEDCkOux0FX/8LnHNE3jIFEg8Zjn76T7hJpBgzMQFuEim8/PwhkcmuWDYSGBgIX19fSKXSAYqeENJXNHBBCCGEEEIIGVIEQiFue/h3MLe2wtt/JMxtrXDYbHCILsyicJNKkbzyuU6vZYzRoAUhQwy9DpUQQsigJBQKJ6nVao3zs2bNmsD+qDczM1O+cOHCkK6OJyUlhUdERMR0dVypVI6LjIzUqNVqTWxsbHRX5y1YsCBErVZrxo4dGyOVSic6+/HOO+/4OhwOpKenB4WGhsaqVKrYm2++OTIvL08KAPX19YLRo0fHHjt2TAIAZrOZRUZGar799ltPV/pNCCHDTUhsHMbE34yysjK0msyI+e0dCIpQw+GwD3RoZJCgXGL45BI044IQQojryn/ww9lvlbA0iyH2smBMkh6jbjG6UqVEInHodLoT/RVib2RlZfl4enr2mPF+9913xUFBQd0ujnbuDl5UVCS+5557Ijr25c9//nPA4cOHPQsLC094eXk5du3a5Z2SkhJeVFR03NfX1/GnP/1Jv3Tp0pDvv//+1Lp16xTx8fGtSUlJra73kBBChpe2tjYYjUYYDAacyz+MttoajI2/GQEhqoEOjfQV5RJXoFziVzTjghBCiGvKf/BD8RehsDSLAQCWZjGKvwhF+Q9+/d1UXV2dUKVSxWq1WgkAJCcnh23atMkfAHJycrw1Gk10VFSUZvLkyZHd13SlxsZGQWZmpmLdunWV/R335TIzM4O2bNlS5uXl5QCA+++/v2nSpEmtf/3rX+UA8Pjjj9cLBAKsXbtWkZWVFZCRkVF+rWMihJChSCaTwdPTEwKBAOqkmZh4932Qj+ryQTgZrCiX6LMbLZegGReEEEJcc/ZbJRy2SwfCHTYBzn6rdOVJidlsFqjVao3ze1paWuWSJUvqMzIyShctWhS2fPny6oaGBlFaWpqhoqJCtGLFCtX+/ft1arXaUl1dLexreytXrlSmpqZWy2QyR0/nTp8+PYIxhkcffbT2mWeeMfSlHaPRKGhvbxfExMSYO5ZPmjSp9fjx478sus7MzCybOHFizKuvvnpeoVDQvGdCCOkEYwwqlQoNDQ0QiUQoKytDRUUFRo0aNdChkb6gXIJyiR7QwAUhhBDXOJ+O9La8l7qa3pmSktK0c+dO3/T09ND8/PzjALB//37PxMTEZrVabQGAvt6cc3Nz3c+dOyfZvn17WVFRUbdxf//99zqVSmXV6/WipKSkyJiYGNOsWbNa+tJeZzjnYIz98v2zzz7zDggIsBYWFrq7WjchhAwX3OGA/WwtmLsYQqUvAEAkEsHf3x/V1dVob2+H3W5HcHAwBAKaXD5kUC5BuUQP6L9mQgghrhF7WfpU7iK73Y7i4mKpRCJxGAwGEXDljbqvDh48KCssLPRQKpXjpk6dqi4pKZEkJiZGnT592s25EdbGjRsDAEClUlkBQKlU2u6+++6GQ4cO9WmjKz8/P4e7u7vjxIkTlyQ1P/30k4dGozEBQElJidu2bdsUhw4dOvntt9+OOHz48JBPOAghpD/wNgvs1c2wldVfUl5XV4fy8nI4HA6YTCa0tHT9N6DZbMbZs2dRX/9rHQ0NDSgpKYHZbO7yOnINUS5BuUQPaOCCEEKIa8Yk6SEQXTolUiByYEyS/lo0t379ekVkZKQpKyvr7OLFi1Vms5lNmzat9fDhw146nU4MAH2d3rlq1arampqaAr1ef+zAgQM6lUplPnLkSFF4eLhVp9Od0Ol0J9LT02ubmpoE9fX1AgBoamoS7Nu3z3v8+PHtfe3DihUrqn7/+9+HtLS0MADYvXu319GjR72WLFlSBwC///3vR69cubJy7Nix1hdffLFs+fLlIQ5Hj7NOCSFk2BPIpBCNDYCb+tKXQ7i5uV0yw6KsrAwAYDKZYDKZLjm3tbUVra2taGxshMPhgNFoRHV1NQwGA86fPw+L5Zr8rUy6Q7kE5RI9oKUihBBCXONce9rPO4Ffvi41KSmpcenSpYbs7Gz//Pz8k76+vo6cnJzm1atXB2VkZFRkZmaWpKSkhDscDsjlcmtubu6prurOycmR792718f5PTc39+TYsWOtPcVUXl4uSklJCQcAu93O5syZUzd37tymvvZtzZo1NfX19UKNRhMjEAgQEBBg3bVr12mZTMY//vhjb71eL37qqacMAPDwww83vv322wFvvPGG/Mknn6zra1uEEDLcCANHXFHm7e0NhUKB0tJSAIDNZoPJZMKpU6fgcDgQHR0NsfjCw2lfX18wxuDp6Qm9Xo+amhqIRCJwztHU1ISamhraI+N6o1yCcokeMM75QMdwhfj4eJ6XlzfQYRBCyJDEGMvnnMe7UodWqy2Ji4vr00ZRhGi1Wv+4uDjVQMdBeQQhNxaHw4GGhgZIpVKcPHkSVqsVQuGFh+V2ux0CgQACgQBjx46Fr6/vJdcWFRWhubkZcrkcRqMRVqsVAQEBGDt27EB0ZdCgXIIMhO7yCJpxQQghhBBCCBmyGhoaoNfrIRaL4eHhgZaWFojFYlgsFgiFQri5uYFzjs6myY8aNQpGoxFVVVXgnNOGnoQMUjRwQQghZFjavHmzfOvWrYqOZQkJCS3Z2dml/d3WggULQo4ePSrrWLZs2bLq1NTUITkdkxBChhKZTAaZTAabzYaWlhYIBAJwzqEYORLBSiXMZjOampqumG0BAJ6enrDZbNDrL2ylIBKJ0N7ejubmZnh5eV3vrpBBhnKJwYMGLgghhAxLqampddfrZn8tEhhCCCG9IxaLERYWBovFgh9//BEAYDWbICw9D5PRiCovL7S3t0MqlXY6eOHt7Q2pVAqTyfTL4EdFRQWioqKud1fIIEO5xOBBAxeEEEIIIYSQIU/IORgADgBMgBpvb1S0tUIiEMDH1xceHh6dXufcB8NJJBLRbAtCBhlaxEUIIYQQQggZ8oSNDYiqrYW71Qp3iQTggIMJYHc4IBQKUVxcjObm5iuuE4lEGDt2LHx8fCAVu0Hh74+goKAB6AEhpCs0cEEIIYQQQggZ+gJGwicqEnGREfAwtcMsFkMADgVjEJhMAC68gaSztyp6enrCy12KhppqnPopD4yx6x09IaQbNHBBCCGEEEIIGfoYAyKjAA8P8JZWCACMsNoQeLoYgSVnER4ejtpTp9C2KwfQnbzicj+5HDajAUKr+frHTgjpFg1cEEIIGZSEQuEktVqtcX7WrFkT2B/1ZmZmyhcuXBjS1fGkpKTwiIiImK6OGwwG4cyZM8eEhYXFjBkzJuabb77x7OrcOXPmqN55551LdoIrKioSS6XSiWq1WjN27NiYlJQUldlsZgDw+eefe3l5eU2Ijo7WhIWFxTzxxBOjOovbbrfj/vvvVz3wwAMqh8OBJ598UhkYGDjew8Pjpr78LgghZFiSeSE0wB/B7lKExsVB4CcHC1JCKBRC1N4OodkM1F2536K7zAt3PLwQt//nIwMQNLkWKJcYPrlEj5tzMsb+BuAeADWc89iLZa8ASAZgs+AxvgAAIABJREFUAXAGwKOc84ZOri0B0AzADsDGOY/vv9AJIYQMFh8Wfej3pvZNZV17nVjuLrcsjVuqfzDqQaMrdUokEodOpzvRXzH2RlZWlo+np6e9u3OeeOKJ0TNmzGj68ssvz5pMJtbS0tLnhwCjR48263S6EzabDbfeemvk3/72N99ly5YZASA+Pr5l3759p1taWti4ceM0X331Vf2MGTNandc6HA7Mnz8/1Gq1so8++uicQCDAfffd1/DMM8/UREdHx/a914QQMswwBslNE/HLX2tJ0wEAYs4x2k0EgacMiIlFS0sL3NzcIJFIBixU8ivKJfrmRsslevML2gFg5mVlXwOI5ZyPB1AM4Llurp/GOZ9AgxaEEDI8fVj0od/GoxtDDe0GMQeHod0g3nh0Y+iHRR/69XdbdXV1QpVKFavVaiUAkJycHLZp0yZ/AMjJyfHWaDTRUVFRmsmTJ0f2te7GxkZBZmamYt26dZVdnWM0GgWHDx/2euqppwwAIJVKub+/f7fJSXdEIhEmTpzYqtfr3S4/JpPJeExMTHtpaam4Y/ljjz022mg0inbt2nVOKBQCAKZPn94aGhpqvdo4hgLG2NOMseOMsULG2PuMMelAx0QIGWKOHIawQAtmrEP7D7lo+79vYDh6ZKCjIqBcgnKJnvU4cME5PwDAeFnZV5xz28WvPwC/DmgSQgi5sbypfVNpsVsuuZ9Y7BbBm9o3la7UazabBR2nd7711lu+crncnpGRUbpo0aKwbdu2+TY0NIjS0tIMFRUVohUrVqh27dp1pqio6MTu3bvP9LW9lStXKlNTU6tlMpmjq3N0Op3Ez8/P9sADD6iio6M1Dz74YGhTU9NVL7tsa2tj+fn5nsnJyU2XH6utrRWeO3dOMmPGjF+2wP/kk0/8CgoKPD/99NOzbm5X5CfDFmNMCeAPAOIvzv4UAnhoYKMihAw5NivAOWCxQFxZCe/GBvh1smSEXH+US1Au0ZP+2OPiMQD/6uIYB/AVYyyfMfZEd5Uwxp5gjOUxxvJqa2v7ISxCCCHXQ117nbgv5b3lnN7p/CxZsqQeAFJSUpqio6Pb09PTQ3fs2FECAPv37/dMTExsVqvVFgBQKBR9enKRm5vrfu7cOcnChQuvWPbYkc1mYydPnvT4/e9/X3vy5MkTHh4ejj/+8Y99Xi9bVlYmUavVGrlcPkGpVFpuvvnmduexvLw8WWRkpEapVMbNmDGjMSQkxPmgADExMW16vV783XffefS1zWFABMCdMSYC4AGgYoDjIYQMNTdNAsQXbk0C7xFwxI6D2+TfDHBQBKBcgnKJnrk0cMEYex6ADcB7XZwyhXM+EcAsAL9njE3tqi7O+TbOeTznPD4gIMCVsAghhFxHcne5pS/lrrLb7SguLpZKJBKHwWAQAQDn3KVX1x08eFBWWFjooVQqx02dOlVdUlIiSUxMjDp9+rSb8ynNxo0bA1QqlUWhUFiSkpJaAeDBBx+s12q1fb7xO9el6nS6wh9//FH23nvvjXAei4+PbykuLj6Rl5d3PCsrKyA3N9fdeSw8PNz097///cyCBQvG5uXl3TBLJTjnegCvAigFUAmgkXP+Vcdz6AEIIaRHzU2AhycwYgSMIxU4I3KDvr0dpWdO46fPP0NrU4cH1m1tgOWa3MZIJyiXoFyiJ1c9cMEYW4QLm3Y+wjt7GTIAznnFxZ81AD4GkHi17RFCCBmclsYt1YuF4kumRIqFYsfSuKX6a9He+vXrFZGRkaasrKyzixcvVpnNZjZt2rTWw4cPe+l0OjEAVFdXC/tS56pVq2pramoK9Hr9sQMHDuhUKpX5yJEjReHh4VbnU5r09PTakJAQW2BgoMW5Lvarr77yjoqKMl1tX0JDQ63r168vf+WVV4IuPzZ+/Hhzampq5YsvvnjJU5j/+I//aH3ttdfOz549O+LUqVMuPYkaKhhjvgBmAwgDEAzAkzE2v+M59ACEENKT3D2f4LMfvkO7sQ4e+nJIhUJ4eXmh7fBhhOjLYdr3f4DJBGi1wJ7PgW+/GeiQbxiUS1Au0ZOrGrhgjM0EsArAvZzzti7O8WSMeTn/DWAGgMKrDZQQQsjg9GDUg8b0hPTz/u7+FgYGf3d/S3pC+nlXdwK/fF3q8uXLlQUFBZLs7Gz/LVu2lM2cObPllltuaV69enVQcHCwLTMzsyQlJSU8KipKk5KSMqa7unNycuQKhWK883PmzJleL/J8/fXXSx955JExkZGRmoKCAvcNGzZ0uQEXADz99NOhznYmTJigvvz4/PnzG9rb2wVffvml7PJjaWlptR2TKKd58+Y1pqenV8ycOTOiqqpKuHTp0lEKhWK8yWQSKBSK8StXrgzubX+GiDsAnOOc13LOrQB2AaD53YSQPqkx1KBRwNDu7QX38HBExMTAz88PwZoYjBAI4dvQAOQdBbQ/AfX1QFs7cKoYKDkH1BkGOvxhjXIJyiV6wrqYLPHrCYy9D+C3APwBVAP4Ey68RUQCwLmbzQ+c86WMsWAAb3PO72KMjcGFWRbAhXWp/+Cc/09vgoqPj+d5eXl97QshhBAAjLF8V9/kpNVqS+Li4ihLI32i1Wr94+LiVP1dL2PsZgB/A5AAoB0X3niWxzl/vbPzKY8ghHSmtaEe7c1N8B8deqHAYAAO7AeCglGq/RltZjNGhoggKPoUMq/bIfIJBdzEgEAAuEuBu5MHNP7riXIJMhC6yyNEPV3MOZ/XSfH2Ls6tAHDXxX+fBRDX+zAJIYQQQq7EOT/MGMsB8CMu7K31E4BtAxsVIWSo8fTxhaeP768FJ09cmFnBOaSMQSkWw1L8L7Saz0MvOoPQ38wCfP2A6mrAx2fgAieE9DxwQQghhAxFmzdvlm/dulXRsSwhIaElOzu7tL/bWrBgQcjRo0cvmZ65bNmy6tTUVHrPXj/hnP8JF2Z9EkJI/xgfd2EDzohIyPd/C0FbG4Set0IsCoJUGg0UFQEeHkBCIhA0qGbNk+uEconBgwYuCCGEDEupqal11+tmfy0SGEIIIddYSzMwahRQpIOwvR1gDG4ibzDhBDTa7fi5sgoTPNwh+/FH4G4auLgRUS4xeNDABSGEEEIIIeTGk5cH2GyAQgEbE0DAHeAAjptMONjShga7HcXt7ZgZFAwatiBkYF3161AJIYQQQgghZMgaPx6IjAR+MwWvNTbh06Zm2AD4itzQAsAOoMJmR6FYMsCBEkJoxgUhhBBCCCHkxhM2BrBagbIyJN0+FaUVlWATb0LooUO4VeyGQ43N8PHxQcLNNw90pITc8GjgghBCCCGEEHLjqa0FtD8DjY2YGBmJiffdB9jtgFKJ20JCEBIUDD+5HF5eXgMdKSE3PFoqQgghZFASCoWT1Gq1xvlZs2ZNYH/Um5mZKV+4cGFIV8eTkpLCIyIiYrqrY86cOSqlUjlOrVZrNBpN9DfffOPpLH/nnXd8Lz+/oKBAcvvtt4eHhITEjhkzJuauu+4aU1ZWJvr888+9pk2bFn553c46/vznPweEhITEMsYmVVZW0sMGQgjpT3lHAIMBcHcHAi/eYoRC4NbbgMm/QahKRYMWQxzlEsMnlxiygRNCCBk8jO9/4Fe3ZYvSZjCIRf7+Fvny5Xq/eQ8ZXalTIpE4dDrdif6KsTeysrJ8PD097b05d8OGDeWPPvpo/a5du7yXL18eWlxc3GmsbW1tLDk5OeLFF18se/jhhxsB4LPPPvOqqqrq8R58++23t8yZM6cxKSkpqm89IYQQ0qOoaKDeeOG1qG5uAx3NDY9yCcolukMzLgghhLjE+P4HfjUvvRRqq60Vg3PYamvFNS+9FGp8/wO//m6rrq5OqFKpYrVarQQAkpOTwzZt2uQPADk5Od4ajSY6KipKM3ny5Mi+1t3Y2CjIzMxUrFu3rrIv182cObO5rKysy53btm3b5jdx4sQWZ6JxMe7mhIQEU091T5kypT0qKsrSl3gIIYT00pgxwKR4GrQYBCiXoFyiJzTjghBCiEvqtmxRcrP5koFwbjYL6rZsUbrypMRsNgvUarXG+T0tLa1yyZIl9RkZGaWLFi0KW758eXVDQ4MoLS3NUFFRIVqxYoVq//79OrVabamurhb2tb2VK1cqU1NTq2UymaMv133wwQc+ERER7V0dLywsdJ84cWJbV8fz8vJkHftZWVkpvueeexq7Op8QQggZbiiXoFyiJzRwQQghxCU2g0Hcl/Le6mp6Z0pKStPOnTt909PTQ/Pz848DwP79+z0TExOb1Wq1BQAUCkWvpmg65ebmup87d06yffv2sqKiol7FvXbt2lEvv/xykJ+fn3X79u0lfWmvo/j4+JZ9+/addn6fM2eO6mrrIoQQQoYiyiUol+gJDVwQQghxicjf32Krrb3iBi3y978m0xLtdjuKi4ulEonEYTAYRGPHjrVyzsEYu+o6Dx48KCssLPRQKpXjbDYbMxqNosTExKgjR44UdXWNc11qT3XHxMSYDhw4ILvq4AghhJBhjnKJ7lEuQXtcEEIIcZF8+XI9k0gumRLJJBKHfPly/bVob/369YrIyEhTVlbW2cWLF6vMZjObNm1a6+HDh710Op0YAPo6vXPVqlW1NTU1BXq9/tiBAwd0KpXK3F2i0RdLliypy8/Pl33wwQcjnGU5OTneR44cce+P+gkhhJChjnKJ7lEuQQMXhBBCXOQ37yHjyNWrz4sCAixgDKKAAMvI1avPu7oTuHNdqvOzfPlyZUFBgSQ7O9t/y5YtZTNnzmy55ZZbmlevXh0UHBxsy8zMLElJSQmPiorSpKSkjOmu7pycHLlCoRjv/Jw5c6bfdmZ7+umnQ531TpgwQS2Tyfgnn3xy+o033hgZGhoaO3bs2JgdO3b4BwUFWXuqa8OGDSMVCsX46upqcVxcnObBBx8M7a84CSGEkMGCcolLUS5xJcY5H+gYrhAfH8/z8vIGOgxCCBmSGGP5nPN4V+rQarUlcXFxhv6KidwYtFqtf1xcnGqg46A8ghBCXEO5BBkI3eURNOOCEEIIIYQQQgghgxZtzkkIIWRY2rx5s3zr1q2KjmUJCQkt2dnZpb2tY8GCBSFHjx69ZDOsZcuWVaemptb1V5yEEEIIGZwolxg8aOCCEELIsJSamlrnalLQl8SEEEIIIcML5RKDBy0VIYQQQgghhBBCyKBFAxeEEEIIIYQQQggZtGjgghBCCCGEEEIIIYMWDVwQQgghhBBCCCFk0KKBC0IIIYOSUCicpFarNc7PmjVrAvuj3szMTPnChQtDLi9PTEyMUqlUsc729Hp9pxtYm0wmNm/evFCVShUbFhYWs2PHDp+u2lq5cmXwyJEjx6vVak1ERETMe++9N+Ly8tDQ0NgZM2aMzc/PlwLA6dOn3Tr2W61Wa2Qy2U3Lli1TAoBSqRwXGRmpiYyM1CQkJEQVFxeLne0xxiYtWbJklPP7Cy+8oFi5cmVwx5jeeecdX8bYpAMHDnj09XdHCCGEDCWUSwyfXILeKkIIIcRlttI6P9vpGiXMNjEkIosofKReFCI3ulKnRCJx6HS6E/0VY2+8++67Z6dOndrW3TnPPfdcUEBAgLWkpKTQbrejpqam23vp0qVLq9evX1/9448/SqdPnx710EMPaTuWA8Bbb73le+edd0YVFBQcDw8Pt3bs95EjR9zvvffeiNWrV9c4y7777rvioKAg29NPPx38wgsvBH3wwQfnAUAsFvM9e/b4VlZWVgUFBdkuj6W+vl7wxhtvjBw/fnxr334zhBBCyLVFuUTXKJegGReEEEJcZCut87OdrAyF2XZhtN5sE9tOVobaSuv8+ruturo6oUqlitVqtRIASE5ODtu0aZM/AOTk5HhrNJroqKgozeTJkyP7u22n999/33/Dhg1VACAUCtHZTb0zEydONAmFQlRVVV2RnCxZsqT+tttua9y+ffslv7O2tjY2f/78sE2bNp0PDQ21Xn7dlClTWiorK92c34VCIV+4cGHtn//8Z8Xl5wJAWlqaMi0trUoikfDexEwIIYRcD5RLUC7RExq4IIQQ4hLb6RolHPzS+4mDC2yna5Su1Gs2mwUdpzi+9dZbvnK53J6RkVG6aNGisG3btvk2NDSI0tLSDBUVFaIVK1aodu3adaaoqOjE7t27z1xNm48//rhKrVZrnn322SCHw3HFcYPBIAQuTM/UaDTRs2bNGlNWVtar2Yvffvutp0Ag4F0lJzfddFObTqeTdixbvnz5qISEhJZHHnmksbNr9uzZMyI5ObmhY9mzzz5bs2vXLr+6ujphx/Lvv//eXa/Xi+fNm9dpXYQQQshAoVyCcome0FIRQgghrnE+HelteS91Nb0zJSWlaefOnb7p6emh+fn5xwFg//79nomJic1qtdoCAAqFwt7X9j788MOzYWFh1vr6esE999wzdsuWLfIVK1bUdTzHarWy6upqt1tvvbXl7bffLl+3bp3iySefHL179+5zXdX75ptvKnbu3Cn39PS0v/vuu2cFgs6fGXB+6YOLnTt3eh88eND7559/Pnn5ubfffnukwWBwk8vltoyMDH3HY35+fo4HHnig7qWXXhrp7u7uAAC73Y6nn346JDs7u8s4CSGEkAFDuQTlEj2gGReEEEJcIxFZ+lTuIrvdjuLiYqlEInEYDAYRcOFGzRhzqd6wsDArAPj6+joefPBB45EjRzxtNhucT2meeuqpYIVCYZNKpY4FCxY0AMD8+fONhYWF3W5MtXTp0mqdTnciPz+/aObMmS1dnffzzz97REdHmwCgoqJClJqaGvruu++e9fLyuuJxzXfffVdcWlpaEBkZ2Z6WlhZ8+fHnnnuu+h//+Id/a2urAAAaGhqEp06dkiYlJUUplcpxWq3Wc+7cueG0QSchhJBBgXIJyiV6QAMXhBBCXCIKH6mHgF16QxQwhyh8pL6LS1yyfv16RWRkpCkrK+vs4sWLVWazmU2bNq318OHDXjqdTgwA1dXVwp7q6chqtaKyslIEAGazme3Zs2dEbGxsu0gkgk6nO6HT6U689tprFQKBANOnT2/84osvvABgz5493hEREe2u9mnHjh0+Bw8eHPHYY48ZAWD+/Pmqxx9/vGbKlCld1i2TyfiWLVvK/vnPf8ov769CobAnJyfX/+Mf//AHALlcbq+vr9fq9fpjer3+WFxcXGtOTs7pnjYPI4QQQq4HyiUol+gJLRUhhBDiEueO3/29E7hzXarze1JSUuPSpUsN2dnZ/vn5+Sd9fX0dOTk5zatXrw7KyMioyMzMLElJSQl3OByQy+XW3NzcU13VnZOTI9+7d+8vrx47ePCgLiUlZazVamUOh4PddtttTStXrqzt7Nq//OUv5Q8//HDYM888I5TL5bZ333235Gr655z22d7eLoiMjGzfu3dvUXBwsO2bb77x3Ldv3wi9Xi/+6KOP5M7zb7/99qa//vWv5R3rCA0Ntd57773GV199deQrr7xS2fHY888/X5WVlRVwNbERQggh1xPlEpRL9IRdvg5mMIiPj+d5eXkDHQYhhAxJjLF8znm8K3VotdqSuLg4Q3/FRG4MWq3WPy4uTjXQcVAeQQghrqFcggyE7vIIWipCCCGEEEIIIYSQQatXS0UYY38DcA+AGs557MUyPwAfAlABKAHwn5zz+k6uXQRg7cWvGzjnWa6HTQghhHRv8+bN8q1bt17yDvKEhISW7Ozs0v5ua9WqVYGffPLJJe9Nnz17tvHll1+u6u+2CCGEEHJ9UC4xePRqqQhjbCqAFgDvdhi42AjAyDl/iTG2GoAv53zVZdf5AcgDEA+AA8gHMKmzAY6OaIonIYRcPZreSQbKtVwqwhjzAfA2gFhcyCke45wf6uxcyiMIIcQ1lEuQgeDyUhHO+QEAl2+MMhuAc/ZEFoD7Orn0TgBfc86NFwcrvgYwszdtEkIIIYR0sBnAl5xzNYA4AFe8k54QQgghw5MrbxVRcM4rAYBzXskYG9nJOUoAZR2+l18sI4QQQgjpFcaYN4CpAH4HAJxzCwDLQMZECCGEkOvnWm/OyTop63RtCmPsCcZYHmMsr7a207fGEEIIIeTGNAZALYB3GGM/McbeZox5DnRQhBBCCLk+XBm4qGaMBQHAxZ81nZxTDmB0h++jAFR0VhnnfBvnPJ5zHh8QMCheFUsIIYSQwUEEYCKArZzzmwC0Aljd8QR6AEIIIYQMX64MXHwKYNHFfy8C8Ekn5+wFMIMx5ssY8wUw42IZIYQQ0i2hUDhJrVZrnJ81a9YE9ke9mZmZ8oULF4ZcXm4ymdi8efNCVSpVbFhYWMyOHTt8Ort+1apVgc6YOsa4YcOGkStXrgx+4YUXFJdf01NfFi1aNNrDw+Om/ujfMFUOoJxzfvji9xxcGMj4BT0AIYQQcjnKJYaP3r4O9X0AvwXgzxgrB/AnAC8B2MkYWwygFMADF8+NB7CUc/4459zIGPtvAEcvVrWec375Jp+EEEKGuONNP/n92HhI2WZvFXsIPS0TR0zWx3jf5NL/30skEodOpzvRXzH25LnnngsKCAiwlpSUFNrtdtTU1HR6j3z55ZernK8m8/DwuKljjCtXrgzu7Jru+nLgwAGPxsZGV/acGvY451WMsTLGWBTnvAjAdADX7X8b/cHmsOHDog8hFUoxK2wWPNw8BjokQggZVCiXuIByic719q0i8zjnQZxzN875KM75ds55Hed8Ouc84uJP48Vz8zjnj3e49m+c8/CLn3euVUcIIYQMjONNP/kdqt8X2mZvFQNAm71VfKh+X+jxpp/8erq2r+rq6oQqlSpWq9VKACA5OTls06ZN/gCQk5PjrdFooqOiojSTJ0+O7Gvd77//vv+GDRuqAEAoFCIoKMjWv9FfyWaz4dlnnx21efPm8mvd1jDwJID3GGMFACYA+PMAx9MnZxrO4APdB1h/aD3u230f2qxtAICq1ir05tX0hBAynFEucfVulFziWm/OSQghZJj7sfGQ0s7tl9xP7Nwu+LHxkEtvkTKbzYKOUyLfeustX7lcbs/IyChdtGhR2LZt23wbGhpEaWlphoqKCtGKFStUu3btOlNUVHRi9+7dZ/rSlsFgEAIXnnJoNJroWbNmjSkrK+u3Jxed9QUAXnzxxZF33XVXQ2hoqLW/2hquOOc/X1wKMp5zft/F16wPenaHHQfKD0AkECFBkQAOjsq2Svzn5/+JV46+gtR9qfiw6MOBDpMQQgYU5RJX1xfgxsklhvV0EkIIIdee8+lIb8t7q6spkSkpKU07d+70TU9PD83Pzz8OAPv37/dMTExsVqvVFgBQKBT2vrRltVpZdXW126233try9ttvl69bt07x5JNPjt69e/c5V/rQXV9KSkrcdu/e7fvDDz8U9UcbZHD6vuJ7vPHzG2gxt6DOVAd+8eVq+mY9/q3/N8pbynGw/CAeiHwAQoEQAPDxqY+RcyoHaZPSMFExsbvqCSFkWKBc4ur6ciPlEjTjghBCiEs8hJ6WvpS7ym63o7i4WCqRSBwGg0EEAJxzMNbZG7h7R6FQ2KRSqWPBggUNADB//nxjYWHhNd2E4IcffvA4f/68VKVSjVMqleNMJpMgJCQk9lq2Sa6/SN9IRPtFw8ZtaLW1/lLOGIO70B3gwNGqo1j2zTK0WFoAAHXtBtjsVtSbh8SkEkIIcRnlElfnRsolaOCCEEKISyaOmKwXMqGjY5mQCR0TR0zWX4v21q9fr4iMjDRlZWWdXbx4scpsNrNp06a1Hj582Eun04kBoLq6WtiXOgUCAaZPn974xRdfeAHAnj17vCMiItqvRfxODz30UKPBYNDq9fpjer3+mFQqdZSWlhZeyzbJ9RfoGYh1v1mHh9QPgeHXhNjqsKK0sRTuQnc44EBFSwWaLc0AgN9Jg/Cq702Y7hczUGETQsh1RbnE1bmRcglaKkIIIcQlzh2/+3sncOdaTuf3pKSkxqVLlxqys7P98/PzT/r6+jpycnKaV69eHZSRkVGRmZlZkpKSEu5wOCCXy625ubmnuqo7JydHvnfv3l9eUZabm3vyL3/5S/nDDz8c9swzzwjlcrnt3XffLbmauDMyMoL++te//vIas+rq6oLO+rJly5ZrkoyRwemk8STkUjnqzfW/bMYpk8gwPmA8DuoPotHSiL0le/G72N9BxIQYLfYGOmzayTlHzqkceLl5YWbYzF/KjVXNaKprgzLCH27iPuXYhBAyaFAucSnKJa7EBuNO1vHx8TwvL2+gwyCEkCGJMZbPOY93pQ6tVlsSFxdn6K+YyI1Bq9X6x8XFqQY6jsGYRzy17ykcrTwKDzcPBHkGQciE8JH6wN3NHccNxyERSpCekI4Y/xi4CyWA3QKIpL9cX9VahdR9qQAHtt7yFkb4yiAUCXAqX4+2ZjNCNSPhM1I2gD0khAwnlEuQgdBdHkFLRQghhBBCrrExI8ZglPco/M+t/wMviRdKmktQaChEfnU+EgITsGPmDpS3lON3X/4O/yrZi7K2Wqz4vxX4UHfhjSOBnoFYoFmA5aOfQtlxA/SnLvwtMDJkBIQiASyma/7GPUIIIWTA0FIRQgghw9LmzZvlW7duVXQsS0hIaMnOzi7tbR2rVq0K/OSTTy55h/zs2bONL7/8clV/xUluDA9GPYi4kXFIDErE8/9+HvXt9ZgXNQ/B3sH4TdBv8OyBZ1HaVAqJUIJWayvePfEuzjedR6BnIFoa2mFus+LusLvR1mRG6ckaWEw2VJUYUXHaAG53QH+6DlKZGN5+HuCcg3OgtdEEsVQEibvbQHefEEKGJMolBg9aKkIIIcMMTe8kA4WWinTtuYPP4aTxJKYET8GuU7vQYm3BpJGTsGPWDpyoO4HXf3wddm7H87c8j6rWKvwl/y/wcvNCxrQMlOUbYTXbMXZCEGQ+7jBUNOH88WrYrTZEiA/ATWDDGZ4Eb4UvQtQjce5YJVoaTOAODncvCdSJowe6+4SQIYZyCTIQussjaMYFIYQQQsg1Fusfi2OGYziXDr7PAAAgAElEQVRUcQh2hx0MDCKBCF+VfIVnDzwLsUCM3bN3Q+mlRKBnIJLHJCNaHg0vzqGS/IRm9xC4y1TgnKPqnBHmdivcxEK4uzNIxCK4mRjqq1pgt9hRV9UMoVAAT28pvOXX9E18hBBCyHVBAxeEEEIIIdfYI9GPIF4Rj7/k/QVGkxFeYi+M8hr1yyCGm8ANb/z8BurN9Xh56suYr5l/4cLqAnhYyyH1BpqMkXA4OJqN7QAH/AK9II5ZCGNlMxp+qgVggbndAu4ARGIB5KO8IRQI4HBwCASs2/gIIYSQwYwGLgghhBBCroMovyjE+MfgmOEYODi+K/8Op4w6ZIfNhTpwEv5w/lM0WZpgtpkB8cWL/KMBhw2GJh9UnqiBr0IGv0AviMQCSNzdkP9/5/8/e3ceH1V1/4//de6dPTOTyWSykXWyTiZACISgoCJqW2hVVFopFOlP2yLVfpQPKGAXa/u1Cn6KW0VcatVg1VKsimilVFFbUSCAYckCSchC9nUy+9zt90dMSCBhSyQs72cfPJqce+455w609z1nhaIAvIqDJMqALEPFQjCH6VG9vxmyIiMpEI0xaZGj+uyEEELIcFDHBSGEEELIN0iQBRQeLMQY4xjcnnM7vmz8EpIsQa/So6qjHMyQCnXIjd9M+TUUVw2ieO2xm4PdQNANjSkBQk0XWuq6kJYbBzCG2pJmiIIEgIFX9RwUx/EMmXHt6NZmQ2zpBgBodHxPpwbQl48QQgi5kNDbixBCyHmJ5/lJDofD2fvnl7/8ZexIlPv0009HLly4MOn49EAgwObNm5eckpIy1m6357zyyiuWwe5fsWJFbG+b+rfx4Ycfji4uLtYWFBRkORwOZ2pqas68efOSAWDz5s2mGTNmpPcvZ86cOSkvv/xyREVFhbr/czocDqfRaMz7+c9/Hu92u7mrr7463W6356Snp+fcdddd8SPxGZBzq7a7Fv+q+RfeLHsTJq0JT814CmuvWwsVp0K36Me7TEBz4uVYsW0Jntr1f0DFlmM3N+4FmveD6zoCgEESRPhbGlBb2gJZVmCNMyEq0QwA4NUcdEYdWpkTLQ0+gAF6oxquNj8OfF6N8l1HIcvn36bshBDyTaFY4uKJJWjGBSGEkGFrLN1vrdtbFC/4fRq13hBKzMuvj8se1zGcMrVarVxWVlYyUm08lQceeCAuKipKqK6uPiBJElpaWgZ9R65evbqp9wgzg8GQ17+NV1xxRcY999zTvGDBgi4A2Llzp/5U9aanpwv9y9i5c6f+xhtvzFi5cmULACxbtqz5hhtucAcCATZt2rTMDRs2mG+99dbu4T4v+WZ5O9rRcLAYMZnZSI1Oxe05tyMmrOdEvShDFADAI3jA8zzM4Ung1XrwKh00HNBmiMRzXz6My8dcjmvj8gBeDVOsA6nqJqjqP4fJ7cWuoB6egA0ZSjbiEqKg0avgd4fgc4fg94bAcQw8zxD0i+ho7Aav4qDW8qja24Awix5xadaTNZ8QQs45iiV6UCwxOJpxQQghZFgaS/dbj+z4PFnw+zQAIPh9miM7Pk9uLN0/4t+M2tvb+ZSUlLHFxcVaALjhhhvsa9assQHAxo0bzU6nMzsrK8t5+eWXZ55p2W+88Ybt4YcfbgIAnucRFxcnnmkZLS0t6uTk5FDv7wUFBf4zud/n87EFCxbY16xZU5OcnCyYTCb5hhtucAOATqdTxo8f76urq9Ocqhwy+rqbG+HtaIOroR6MMcy0z0RedF7f9VZfK+7IuQPLJy/H4gmLYdPb8MJ3XsLvvvsKStVq7G/bj0/rPgUMkUDKdDBdOKzuz2Hm28G0YSj3f47PAh9A1AWh1ashCTIUWQFjgCUqDLyKh6IAHMdBpeYRFq6DKTIMLUddqD7YBFmWR/HTIYSQgSiWOIZiicHRjAtCCCHDUre3KF6RpAEd4YokcXV7i+KHM1ISDAY5h8Ph7P192bJljT/72c86n3jiidof//jH9rvuuqu5q6tLtWzZsraGhgbVL37xi5RPPvmkzOFwhJqbm/kzqautrY0HgKVLl47Zvn27KTk5OfjCCy/UJiYmnlHAcffddzd/97vfzczLy/Nee+21rrvvvrvdZrNJAFBUVGTs/zyNjY2a66+/3tX//rvuuith8uTJnh/96Eeu48tua2vjt27darn//vubz6RN5Nxyi91QMzVkUYBGZ0B0Rtag+X79+a/hCrqw5uo1UHNqAICG74kjp46ZClEW4TAlAs37AVsWwGuAuDzA2wIkTcPNwgRUurxIixsDyeuCIAp9ZQtBCSG/AJVWBZWaQ8ATgqIoiEu1okHNQa3hwRidMkIIOX9QLHEMxRKDo44LQgghw9I7OnK66adrqOmdN998c/eGDRsili9fnrx79+6DAPDJJ5+EFRQUuB0ORwgAYmJipDOpSxAE1tzcrL7iiis8f/7zn48+9NBDMf/zP/+T+M477xw5k3Luvffe9tmzZ3e/88475vfee8/yyiuvRJWUlJQAQH5+vmfbtm0VvXnnzJmT0v/eDRs2mP/zn/+Yv/rqq9JB2odbbrklddGiRc1OpzN0/HVyfnCL3fh3yyboeAOSjqoghAKQBAG+zg7IkgSjLaovrzPSiXpPPcI14X1pXYEuHGjbjymGOMzQxwIdR4C2ckD0A/EFAKdGkJnhahKgt6YgwyCio7IaktSJDwLbUSBOh00Kg6hXQ1EAMShCDPaUbY0zwRwZhvzvZEKWFBwqqodGp4J93Igs9yaEkGGhWOIYiiUGR0tFCCGEDItabxj05TdU+nBJkoRDhw7ptFqt3NbWpgIARVGGNYIcExMj6nQ6+bbbbusCgAULFnQcOHDAcDZlpaSkCEuWLGn/6KOPKlUqFYqKik65NrWhoUF17733JhcWFlaZTKYT5vDPnz8/JTU1NfDggw+2nE2byLmhZmpoeT0kRYRfE0Kg2wVfVycq/rsNlZ9/AsF/bLbvvRPvxWNXPQajxggAkOUQSo78BfWN7+GFLx4Gqj4CwmJ7loq46oDOI0Dlh6g+7EX1/ga01LpgsYVBgA6yYkJQ2w2IIWjc7UjIjIRKw4FxDGqtCnE2Duq//gneL7+ELMlQJBlBXwi+7iAUhTbrJISMPoolBqJY4kTUcUEIIWRYEvPy6xnPD3hBMp6XE/Py67+J+n7/+9/HZGZmBl599dWqn/zkJynBYJDNmDHDu2PHDlNZWZkGAM50eifHcbj22mtd77//vgkAPvjgA3NGRsYZrSkFetbGBoNBBgC1tbWqrq4uvv861aEsWLAg5ac//WnLtGnTTqjznnvuGdPd3c2/9NJLdWfaHnJu6Xg9JpgnQ+j2wl/bBCkUQk3Rlwiz2mCOiQOvHXrgUJQE6FUaaDiGz0PtgGkMoAkDzAmAuxGo2gpZCMHGVYLxKlhijAiPCkPWFU5kXT0Zv576KBJiMqBxZEMUFQAMCgCNXo3wxhIEy8tRU9qEg5/XoLPFi/SJ8ciYFE9LRggh5wWKJY6hWGJwtFSEEELIsPSuPR3pncCPX5d6zTXXuBYvXty2fv162+7du0sjIiLkjRs3uleuXBn3xBNPNDz99NPVN998c7osy4iMjBS2b99+eKiyN27cGLlly5a+I8q2b99e+vjjjx+dP3++/b777uMjIyPFwsLC6jNt84cffmi+7777krRarQwAv/vd744mJSWJ+/btG/Kef//732Hbtm0Lr6+v1/z973+P7E2fPn169/Lly5v/9Kc/xdnt9kBOTo4TABYtWtSydOnStjNtGzk3orSxiNcmQVRXgokKeLUKlvgERKakDXnPk7ufxN6WvfjN5JWorf8Q30+aBKmrHuiqB587D1Dp0NUlw9j5IQzoQIbqY5jDc8FUMaja1wR3uxe8moNGp0YwIPQclSrK4HgOfncQ3VmXQ7gjDPW6DBh8IXQ0daO5phNpuXHQ6CgUJISMPooljqFYYnDsfJwimJ+frxQVFY12Mwgh5ILEGNutKEr+cMooLi6uzs3NvaBeaGT0FRcX23Jzc1NGux3nQxwR9HnR3dQAxhgiEpLBq9VD5v1/X/w/lLSX4NfjH4JYpQbPM0QKe8AYEH3lbHCHN+NIYzh42QON7IIJjZA1ERCTv42yfW4ojCHMrEVyTgwOFdVDDEmALMEcaIUQk4ToMWHgu/6Lr8R8RHA8os0adLf7kDouFibrWc1iJoRc5CiWIKPhZHEEdbMTQgghhIwwb1sr6vfvBcfxqN+/F2lTp8Noi+673hnoxJ/2/gkToidgRcEKdIe6EamLRCtc0BrUcDVOglZqQ0NlJyzNjYjXNyNgvwUyp0fDwQOIV3bCd/BDQJoCju+ZbVFX1gJZlsE4QNVYA7b1DSjX/QCazKthM6XhO+iEEjcB7n9/BFt8InVaEEIIuWBQxwUhhJCL0lNPPRW5bt26mP5pkydP9qxfv772dMtYsWJF7LvvvjvgDPnZs2d3rF69ummk2kkuTnpLBPRmC2RRRMjnheIOQgp1gYsLB2MMR1xHcLD9IDwhD25MuxE2vQ0AYBqjhZpTI/zop4CvA/V+H45IV0EHBquiRWS0EabwXChFu8ACDIosArwWoYAIXs0BCqA1aCDFxsP1w/sAjkNttQsloQgYOQUZrQfR8fzzQOwYsNt+BK3dDnVM9CmehhBCLk0US5w/aKkIIYRcZGh6JxkttFTkRIosQwgGgNJWKH4BamccuIgwKIqCLxq/QIo5BRG6CHxQ9QGSzEl4Zu8zyDOMwRKVFZBCEMb/DK3tPFrrXFBreZisBoyJFcCX/QNtQgKq/TkQAgLCbWHwugJQZCAq0YyuVi+CXgGyrECKMEDwiVDJMnImRKP91VfgDmqg//QdWAomYszqVaP9MRFCzjMUS5DRQEtFCCGEEEJGAeM4aPQGSPEWyK4AmEnXk84Ypo6ZCgD4qOYjbDi0AXazHQoUtDMFiB4L6COhtkQjLlyB3qhBfUU7Ohq6EWGNgT75GoQxM1JkM+oPtcEYoYfPHYJWz0MISZAEGdHJFnS3+8ArIhrFIBB0Q2NJQVPeDGgPlEDWhcEwZcpofjyEEELIaaGOC0IIIYSQbxgfEw4+JnxgYrAbOPxPTAqLwpXxV+LyMZfDYXVAw2kAXg20lQFt5WC2LOhNWpgi9DAHQlAfasRhRQ+3xwddhAK3V0CovhtjUiMQEW1C6bZShKCGAsAxJRG8EIDp/hVQ6bUIu7kAMVkp8L//d/BiEOwkR7QSQggh5wvquCCEEEIIGQ3+DsDbCgsYfpH3i4HXBB9Q+e+eny3JaKrqQlerF7KaIRhSAA0ghQSIPg7gOciiAo7n4C2vAkTAEOqGNSYJOoMGgAZjnngCsgL4i4pg2r4LQkIKWNUB8GbzOX9sQggh5ExRxwUhhBBCyClIigSf6IFJHX7qzP0pMtCwG9BbAWvawGuWFCDrhp5rx1PpgTGT0HNEiA4hkwFCQIar2w8X48H5/OAba2D5z3+R+svfoLvFjc4mNxBSg3ECNAYDfO4Q1Fo/XG1edLV4IQkSDEV74EY4YsfFI/a+t8B4/qw/E0IIIeRc4Ua7AYQQQshgDAZD3qnyxMfHj2tsbDztTvgNGzaYx44dm52amppjt9tzFi1alNB77dlnn7VmZmY609PTc7Kyspxz585Nbmtr4wGgoKAgKyUlZazD4XA6HA7nzJkzU4dqz6RJk7L6pzkcDmdGRkZO7+/btm0zFBQUZCUnJ491Op3ZV199dfrOnTv1ALB06dIx0dHR43vrcTgczt42DGbbtm2G/Pz8rJSUlLF2uz1n7ty5yW63m3v66acjFy5cmNQ/b0FBQdZnn31mAIAnn3wyMjMz05mZmenMyMjIee211yy33XZbksPhcKalpeXodLqJvfW//PLLEV988YV+woQJjszMTOc111yT3tHRcUnFD6Is4LO2LdjS8g5qfVUnXJcUaeibPc3A0R3AkY8Hv25JBrSmE9MZAxIvBxKmQJYV1HYE0SYyGKPCEJdqBVRq8Fo1zI50cAC6230IeAUoACy734NYtQ8qNYfm6k401bjg6nBDcndDfeW1UNKckOzZENva4d+//6w+E0IIuRBQLHHxxBI044IQQsiw7dq1y/rpp5/GezwejdFoDE2fPr1+8uTJHaPdrv527dqlW7ZsWdKmTZsq8vLyAoIgYM2aNVEAsHHjRvPatWtjtmzZcthutwuiKOKZZ56JrK+vV9lsNgkACgsLq6666irfqerxer18RUWFOj09XdizZ4+u/7W6ujrVggUL0l555ZWqb33rW14A2LJli7G8vFxbUFDgB4DFixc3//73v28+VT11dXWqH/3oR2mFhYVV1113nVeWZbz66qsRXV1dJw0EKisr1WvWrIn76quvSiMjIyWXy8U1NjaqFixY0AUA5eXlmuuvvz6jrKyspPeesWPHZq9evbrue9/7nufJJ5+M/N3vfhf71FNPNZyqjReLEncxGgK1EBQRJe6vsL97N4wqMwQlBIsqEoc8B6Dltbgq8juI0sYOvDksGojLG3xWxWniOIYxOoZOdwhhUUbEJEcgItoI7opUqDQ8Wmq7IIZkWGOMsHRUoLN0LyIzvYhJvg2BDa9BdCvgD2wH31AHVjAZSf+zAtI/N6L+y+1QgiGYvv0tBCsrEb1kCdRjxgzz0yKEkLNDscQxFEuc6JIaMSGEEDLydu3aZd2yZUuyx+PRAIDH49Fs2bIledeuXWf/Ta2fzZs3mwoKCrJmzpyZarfbc2688Ua7LMt91x977LFop9OZnZmZ6dy7d69uqHIeeeSR2GXLljXm5eUFAECtVmPlypWtAPDoo4/GrVq16qjdbhcAQKVSYcmSJe25ubnBM23vTTfd1FFYWGgFgMLCQuucOXP6gq4//vGP0bfeemt7b6ABAN/5znc8t912W9eZ1rNmzZroW2+9tf26667zAgDHcbj99ts7ExMTxZPd19jYqA4LC5PDw8MlAAgPD5cdDkfoZPdUV1frZs2a5QGA66+/vnvz5s0RZ9reC1mUNhYRahtCcgA1/ko0BGpxyHMAld5y7O/ehYDsg1/y4ZDnIL7o2IaQ3O+fDccDSdOAqOzTrq/eU49/1/wbonzsr9KRFYkMVSeEPz4Iz3/+A41eDZWGR32XH8/sOQqXRQejVY96bgwMt92B6F/cDUWWEdz0d5je/jNC3hZwvm4E9u6BtOl1dBW+itCRamjSUhE8Uo1QZRWClZUj+bERQshpo1hiIIolTkQdF4QQQobl008/jRdFccD7RBRF7tNPP40fqTpKS0v1a9eurauoqDhYW1ur3bp1q7H3ms1mE0tKSkrvuOOO1lWrVsUMVUZ5ebl+ypQpg45yVFRU6KdOnXrSEZCFCxem9k55vPPOOxOGyjd//vzO9957LwIAtmzZYrnlllv6AonS0lL9pEmTTlrPc889F9Nbz5QpUzKHyldSUqLPz88fsqz33nsvov800QMHDhgA4LLLLvPZbDYhMTFx3Pe///2U119//ZSbNmRkZPhff/11CwC89tpr1qampkvqKIo4XQKyTOOgQAEAyF//B1AgKhI4qCApElqCjWgI1KFTaD+tcpWAAOFQM2SXf0D688XP48X9L2J7w3YAQLcrgKbaLvBHD0NqaUbw0KG+vHtrO3Go2YMvG9wI+UWIEoC8y6GOj0fnm29C6uoCUxREq/WwfGcmTN/+NvwHSyAHg2A6HeJ+9zvE3H8fou+/H2HTpo3I50UIIWeKYomBKJY40Vl3XDDGshhjX/X7080YW3JcnqsZY65+eR482/oIIYScn3pHR043/WyMGzfOm5aWJvA8j5ycHF9lZWVf2fPnz+8EgIKCAl9dXZ12uHXt3LlT73A4nImJiWNffPHFvtGAwsLCqrKyspKysrKS559//uhQ90dFRUnh4eHiCy+8EJGenu43Go3yUHnHjx/vSE1Nzbn99tsTe9MWL17c3FvPjh07Dg1176nccMMNnb3llJWVlYwdO9YH9IwAffbZZ4dff/31yoyMjMDKlSsTly5detL1AX/5y1+q161bF5WTk5Ptdrs5tVqtnG27LlRskDQOHDRMgwi1FTpOjwR9CiZZpqI12IQvOj6BKAsDb/C2AO5js2KlVjfkVjekxoGDZNcmXYu86Dw4I52QFQX79zahqqwVwdxp0C34MSqm3IBDDT0Dbd92xmLexBRMGBMJfawJaRPiEJ1kgdDQgM6//Q2QJKhiYsA8Pvh374HximngNGowgwFMp0Pnm3+DKjIShoLJENvaRvpjI4SQ00KxxEAUS5zorDsuFEUpVxRlgqIoEwBMAuAD8PYgWf/Tm09RlN+fbX2EEELOT0ajcdCpgUOlnw2tVtv3cuN5HqIo9n2P1Ol0CgCoVCqlf/rxMjMzAzt27DAMdi09Pd2/fft2AwAUFBT4y8rKSmbMmNHt9/uHfE+KoojeEYglS5YMeFl///vf71y+fHnyvHnzBqzNzc7O9u/evbuvDfv27Sv7zW9+09Dd3X3GRztkZ2f7i4qKBn2eU+E4DjNmzPA9+uijTa+99lrV5s2bLSfLn5eXF/j8888PHzx4sPTHP/5xR2Ji4hlPex0JjDGeMbaXMbb5XNedGpYFqzqq73cDM0ABIEJEnC4R4eoIuMVuJOrtOOI7jIZALTyS+1gBsgSUvAWUvg0Ee9L52HBwsWZwCQNny05PnI6VBSth09vAAKgtenBhWgh/fw37JSuaPBIa2nv+CjQqDmMTzLCZ1AjTqaBXSWh7dh1qf34XxJZWMKMRUBRILS0Qm5rQ+c478BXthuLxQKitReuzz8K7qwgdrxbi6M/vgvvjbd/0R0kIISegWIJiiVPWdzaNHMS1ACoVRakZofIIIYRcIKZPn16vUqkGjASoVCp5+vTp9aPVpsE88MADTY8//njcvn37tAAgSRIeeuihGABYvnx508qVKxMqKyvVvfkDgcCQgQvQM9rQOwLx5JNPDthc6kc/+lHn3Xff3XTLLbd0909ftmxZy9/+9rfIrVu3hvWmeb3es3oX33fffS0bNmyI/Pjjj/vKevbZZ621tbUn3Xi7urpa/d///rcvSCkqKjLEx8efNDCsr69XAT2f2W9/+9u4n/zkJy1n0+YRcC+A0nNdqSiL+E/bVqiZGir0/BPxKX4okCEqIqyaSATlALrFTkiKCLshA1pOhwpvKQ607YIsiD17XUSkAeGJgFoPRZAgB0KQm7shHxl6pgNjDAUTYjD1ymSYEmMQ3nQENh0wIbVvhjPS4gxwik3gAz40/PJXaH/pJUhdXTBeeSUi5twCTqcF0+uhc2bDcv314CMiAFn++o8ELswAzmAAGANn0H/jnychhByPYgmKJU5lpE4V+SGAN4a4djljrBhAA4D7FEU5OFgmxtgiAIsAICkpabAshBBCzkO9O36f7zuBT5kyxb969eq6efPmpfr9fo4xhuuuu84FAHPnznW1tLSoZs2alSFJEjObzZLD4fDPnj27L1hYuHBhqk6nkwHAarWK27dvH3LqZUREhPyHP/yh6fj0pKQkcf369VUrV65M+MlPfqKOjIwUIyIixIceeqgvWHnuuediNmzYENn7+7vvvluRlZV1QjCQmJgoFhYWVt1///0J7e3tao7jlMsuu+yUm3OFQiF23333JTQ3N6u1Wq1itVqFF198sfZk9/zlL3+xvvTSS9EA8N3vfrfznnvuOb1NHEYQYywBwPcA/AHA0nNZd0D2oTFYh27BBR48THw4vJIHMiRYeCvM6ghcY/sexMpWsI42uGO6EZT9qHKV4oqaHAR11dDkJkNoSAbjAlC7/gEhOA7MGg5wDFAPPkimfL1xHON64lHrj3+MKcfl8Yck/O3Fd5D7fiFip0zsyStJ4NRq2BYvRs0HH8HP6aBJzQCkIDwffQzT1dPR+ebfoB4zBkl/fhGaxETonU6E33wTOM0ltX0JIeQ8QbHEiSiWGIgpyvCWqTLGNOjplMhRFKX5uGtmALKiKB7G2HcBPKUoSsapyszPz1eKioqG1S5CCLlUMcZ2K4qSP5wyiouLq3Nzc2nBOzkjxcXFttzc3JRvomzG2EYAjwIwoWcg5Pqh8n4TcURLoBH1gRp0Cu3oFNrhl7xg4KBiKoiKiOujfgDd3k5AUaBMjEWr0gomKDAd8EEb4AFZRkdAgUsBksIqoOLTwI2Jhio9GoydOCCniDKEvTUAx6DOSwbjBh+023/UhT++8hF+8PnfMMERD8ucH8D9739Ddrtgu/suVN7xMygd7VDiEqD1d4O32aCKjkZg/wFY5t6K6HvuGdHPiRBycaBYgoyGk8URI7FUZBaAPcd3WgCAoijdiqJ4vv75AwBqxphtBOokhBBCyCWCMXY9gBZFUXafJM8ixlgRY6yotbV1xNsQrYtDnuUyZBidMKrMmGiZimnWaxGmMgEAqgKHoR6fAPW4BOh0RiTq7Ugwp8LoSOmZUaEo8ChAUGHwW3KgcqaAj7cM2mnRQ+npBJFlCA0nHnPf7RPhC0oYG2/G3NmXI+XnP4VQW4vmh/8A/87dsNw8F9rUVEQvXgTjTTfDFB8NzmyG6bpvQWptgy4rC8Zp01D387vg+fzzEf+8CCGEkJE0EktF5mGIZSKMsVgAzYqiKIyxAvR0lJzz6aWEEEIuHU899VTkunXrBhxlNnnyZM/69etPOoXxfPXWW2+Zf/WrXw04Mi0xMTG4devWytFq0yiYBuDGr2dv6gCYGWOvKYqyoDeDoigvAHgB6Jlx8U01JFFvR6Le3vd7lDYWVb5ypBqywKlO3IietxrBxicCHJAclOBtccMcZ4JY0gSmVUGTn3LCPWXu/ajylePycVcB73yMrlUbEXHbAlhuugkA4AtK2FPpAs8zXOm04rvj4iCnhaOjuRqBvQcQqj4CRemZEWz74VxY/X5U3XQTxNY2iJ2d4MLDAbUKgZJSiC0tCB46DCMdhUoIIX0oljj/DGupCGPMAKAOQKqiKK6v02GDD6AAACAASURBVBYDgKIozzHGfgHg5wBEAH4ASxVF2X6qcmmpCCGEnD2a3klGyze5VKQXY+xqjMJSkZGkCBLE0kYwoxYqmwAc+RgYMxGIcgIAvujYhoZAHSZbroDl03K0v/QX2O66C6ZrZgAAREnBV0e6oddwyEky9ZXrC0rQQYTU3g71mIGn0vlLSuH+8EMYr7oS9fcvh9TVBdvPF8OQPxm6sTm0twUhZACKJchoOFkcMawZF4qi+ABEHpf2XL+fnwHwzHDqIIQQQgi5mDA1D/X4rwe+GvdC9neC627o67jIt0xDl9CJKG0MMCsVpm9/G0EJaO3yIMJkgCApyE8PH1BmXZsfFY0+JEfpkXpcpwUA6J3Z0DuzAQBR99wD13vvQT9+PAwT877ZhyWEEEJGwEidKkIIIYQQ8o1TFOUTAJ+McjNOICkSvnLtgI7TI8d8+p0BH3fFYsuBZEybmIobv05Tc5qeTouvMZ5HVfVHUKR6VDVNg0+wIC/VDEtY34l7UPM925ZpVKfevsxy802w3HzTabeREEIIGW3UcUEIIYQQMkw+0YNqXwU4xsFpmnCSTTePu09QUOK3Ymxo8CNRe+k1DMEAoNdwCEgMquNOGYmN0CLGojntegkhhJALCXVcEEIIIYQMQ0uwEYc8B5FtnACbNmrw402VnlNCGDdwRsT148dgcooV0aYTN/bs8go4UOtBok2HlIQZ8AcD0Gt1ADBoHdRpQQgh5GI1EsehEkIIISPOYDCccr59fHz8uMbGxtPuhN+wYYN57Nix2ampqTl2uz1n0aJFfTtsP/vss9bMzExnenp6TlZWlnPu3LnJbW1tPAAUFBRkpaSkjHU4HE6Hw+GcOXNm6lB1LF26dMyDDz4YM9i1wsJCC2Ns0t69e3W9aeXl5RrG2KR77723b2OCxsZGlUqlmrhw4cKk3jKjo6PHOxwOZ0ZGRs5f//rXcAB4+umnIyMiInJ72/X444/TkeOjoNZfheZgAxRIiNbGDZpHPFCP0K5qSG1uiDXtUES571qMWTdop4PLI8DtF9HpFtDYGcSOwz5UNfupg4IQQk4TxRIXTyxBMy4IIYQM38EDVuzZHQ+fTwODIYSJk+qRM7ZjtJvV365du3TLli1L2rRpU0VeXl5AEASsWbMmCgA2btxoXrt2bcyWLVsO2+12QRRFPPPMM5H19fUqm80mAUBhYWHVVVdd5RtOG958803rxIkTPevXr7fm5eU19KYnJCQE//Wvf1kANHxdV0R6enqg/72LFy9u/v3vf9+8Z88e3bXXXpv1wx/+sBgAbrjhhs7CwsIL8ni2i0WyLg1uoRtJ+rQh8yiSDMgKpLpOKL4QmE4FPiZ8yPwAEJIUqHkGnYYD//XSEJ6jTgtCyEWKYonTcqnGEjTjghBCyPAcPGDFF9uT4fP1nKfo82nwxfZkHDxgHYniN2/ebCooKMiaOXNmqt1uz7nxxhvtsnxstPqxxx6Ldjqd2ZmZmc7+ow/He+SRR2KXLVvWmJeXFwAAtVqNlStXtgLAo48+Grdq1aqjdrtdAACVSoUlS5a05+bmBkfiGQDA5XJxRUVFxpdffrn67bffjuh/TafTKenp6f7PPvvMAABvvfWW9aabbho0WJs4cWKA53k0NTXR4MN5QJFl1B0tQZ27Eh+3vT9kPvW4BGgmJUOVEgk+1gzOajxl2fGROiTa9Ei06RFj0eKqHCtSovUj2XxCCDk/UCxxWi7lWII6LgghhAzPnt3xkKSB7xNJ4rBnd/xIVVFaWqpfu3ZtXUVFxcHa2lrt1q1b+7712Ww2saSkpPSOO+5oXbVq1aDTKgGgvLxcP2XKlEFHOSoqKvRTp0496QjIwoULU3unUd55550JJ8s7mL/+9a+Wq6++2jV+/PigxWKR/vvf/xr6X//hD3/Y8dprr1krKyvVPM8rY8aMEQYr5+OPPw7jOE6Ji4sTAeCf//ynJTMz0zlz5szUiooK9WD3kG+O3OhCemM0xnXYEcYP3RnBeA5MowIXEQZVWjSY+uSbcQLAvo5d0BpbEabryUuzLQghFy2KJU7LpRxLUMcFIYSQ4ekdHTnd9LMwbtw4b1pamsDzPHJycnyVlZV9Zc+fP78TAAoKCnx1dXUn7nB4hnbu3Kl3OBzOxMTEsS+++GLfaEZhYWFVWVlZSVlZWcnzzz9/9EzL3bBhg3XevHmdADBnzpyO9evXDxhFmjNnTvenn35qfuWVV6xz5sw5YYTkueeei3E4HM77778/obCwsIrjONx6661dtbW1+w8dOlRyzTXXuBcsWGA/m2cmZ89nkODR+hEf58A1Ud8bsXLLOsrwxO4n8H9F/3fCtQ6PgFC/PTIIIeSCR7HEabmUYwnquCCEEDI8BkPojNLPglarVXp/5nkeoij2DT3rdDoFAFQqldI//XiZmZmBHTt2GAa7lp6e7t++fbsBAAoKCvxlZWUlM2bM6Pb7/SPynmxqauK//PJL8913350cHx8/7plnnondtGlTRP9pqjqdThk/frxv3bp1sQsWLOg8vozFixc3l5WVlezevbt85syZHgCIjY2V9Hq9AgBLly5tPXjw4KDPR7459apGfJFQiiOGenDs9P+5CKIMf0ga8nqCMQEToiZgesL0AemtrhCKj3TjYK3nrNtMCCHnHYolTulSjyWo44IQQsjwTJxUD54fOPzL8zImTqofpRYN6oEHHmh6/PHH4/bt26cFAEmS8NBDD8UAwPLly5tWrlyZUFlZ2Tc9MhAIjNi8/PXr10fccsst7Q0NDfvr6+v3NzU17UtISAj961//GrC2YMWKFU2//e1vj8bGxg79jbafmpqavva+/vrrltTU1MDJ8pORlxaWhfSwbKiZGgHJf8r8ij8ERZKxu7IbOw654AsO/ldt1BjxwJQHcGvWrQPSDVoeBi0PS9gFsyyZEEJOjWKJU7rUYwl66xFCCBme3h2/z/OdwKdMmeJfvXp13bx581L9fj/HGMN1113nAoC5c+e6WlpaVLNmzcqQJImZzWbJ4XD4Z8+e3d17/8KFC1N1Op0MAFarVdy+ffuhoep64okn4p5//vm+NbLJycnB5cuXN/bPM3v27M7169dbH3zwwabetPz8/EB+fv5pBwyPPfZY9JYtWyw8zysWi0V85ZVXqk/3XjIy1JwGQTmAOv8RaDkdcsxDn7wnd/kgHGwAF2GATm+EJCsD9q1o7gqiuSuEzDEG6DSD74ERpuMxJdMy4s9BCCGjimKJE1AsMRBTFOXUuc6x/Px8paioaLSbQQghFyTG2G5FUfKHU0ZxcXF1bm5u20i1iVwaiouLbbm5uSmj3Y5zHUe0h1pR5S1Htmk8jCrzkPlkTxDCgaPgbSao0qMBAF1eAY0dQdhj9Cg96kWXV4AjPgxx1iE3tSeEkG8cxRJkNJwsjqAZF4QQQgghwxCpiUKkJqrv90MNXnT7RIxPMYGra4fc5YfaGQfOqIX2srQB99a2BtDWHUKXV0CcVQtLmAoxlmHvC0cIIYRcVKjjghBCyEXlqaeeily3bt2Ao8wmT57sWb9+fe1I1rNixYrYd999d8Bu3rNnz+5YvXp101D3kEtDuzuEQEhGICRD5/JD8YWgBEQw7YknzKXG6iHLCto9IZTUeqBScbAa1QgPOy9PoyOEkEsCxRLnH1oqQgghFxma3klGy6W6VOR4/qCEgCAjwqiGEhShBARw4fq+67KioK07BEuYGhoVB1lWcKjBi8MNXsgKMGOcFToNj0BIhtmggiDJ8AUk6swghJwzFEuQ0XCyOIJOFSGEEEIIGUF6LY8IY08nA9OqBnRaAEB9ewAHaz043OADAHAcQ1Z8GDQq9vX1IIqPuLG70gWXV0BpnQd7qrrR6hqxUwEJIYSQCwp1XFxChOB5ebINIYQQcskQJQVH2wLwBCSo+WMnijDGkGgzQKPiEBR7ZlroNTy0ag4mvQpaNQedhsI2Qgghlyba4+ISse+jLSja/Da0YWFIm1iAy26ZO9pNIoQQQi457e4QurwiGACDduCRp9mJYYi1amHUcmj3iEiI1EGn4WGPMcAeYxidBhNCCCHnAeq4uER8tfV9dDYcBQA0HT6EiDEJyLps2ii3ihBCCLnw7a7pwIZdR3HHFXZkxZpOmjfSpEFGlAaWbh+sejbgGmMMVqMaDe0B7Knqhl7D49rcyG+y6YQQQsgFgeYcXiJajxzp95uCLc89ifryklFrDyGEnIrBYMg7VZ74+PhxjY2Np90Jv2HDBvPYsWOzU1NTc+x2e86iRYsSeq89++yz1szMTGd6enpOVlaWc+7cucltbW08ABQUFGSlpKSMdTgcTofD4Zw5c2bqUHUsXbp0jF6vz6uvr+9rV/9n4Xl+ksPh6KvnoYceipEkCQCwefNmE2Ns0uuvvx7em3/GjBnpmzdvNg3WjpdffjnidJ+dfHM+r2hHSWM3dlZ3nDKvimdIYRJMLi+kug6IkoKgIA/II/dunM5OvL/rrbfQ9PAfILndI9F0Qgi5qFEs0eNiiCVoxsVFThIFBDweqHQ6iAF/X7rg9+OtRx7ET//0Egzm8JOUQAghp9bU1GStr6+PFwRBo1arQ/Hx8fWxsbGn/hZ3Du3atUu3bNmypE2bNlXk5eUFBEHAmjVrogBg48aN5rVr18Zs2bLlsN1uF0RRxDPPPBNZX1+vstlsEgAUFhZWXXXVVb7TqctisYgPP/xwzLp16+qPv6bVauWysrISAKivr1f94Ac/SHW5XPwTTzzRAAAxMTHC6tWr4+bPn+8arOwzaQc5N358eQoyY0yYnhl1Wvn5aBMQFMHFmLGr0gV/SEZBRjgMWh5BQUZFow8GLY8pmSe+n90ffQyxuRmh6hrox40d6UchhJBRQ7HEQBRLDEQzLi5iR0sP4K+/XIq3Hv3toNeFQAB7tmw+x60ihFxsmpqarDU1NcmCIGgAQBAETU1NTXJTU5P1VPeejs2bN5sKCgqyZs6cmWq323NuvPFGuywfG6F+7LHHop1OZ3ZmZqZz7969uqHKeeSRR2KXLVvWmJeXFwAAtVqNlStXtgLAo48+Grdq1aqjdrtdAACVSoUlS5a05+bmBs+mzfPmzWvftGmTtbm5mT9Zvvj4ePHPf/5z9csvvxzd+0zZ2dk+k8kkvf322+azqZuce+EGNWaOjYVec9K/7j5Mq4YqPRqcSQeNioOKY+C5nukVKp7BbFAh0qw+YQ8MAIhZfj+i/vd/qdOCEHJRoVjiRBRLDEQdFxexoN+HzqZGtNVWD5ht0d+OjW/g/T/93zluGSHkYlJfXx+vKMqA94miKFx9fX38SNVRWlqqX7t2bV1FRcXB2tpa7datW42912w2m1hSUlJ6xx13tK5atSpmqDLKy8v1U6ZMGXR0oaKiQj916tSTjjwsXLgwtXda5Z133plwsrxGo1GaN29e28na08vpdIZkWUb/6aC//vWvGx955JG4U7Wjqanp9L4pk3PKH5LQ3BWELyCiviMAWVaGzJuXakZqrB5t3UEIogyeY5iYFo7cFDMYO3GtiCYlBcYraI8qQsjFhWKJE1EsMRB1XFzEUvMmI23iZJhsUVBpNEPmq9xTdA5bRQi52PSOjpxu+tkYN26cNy0tTeB5Hjk5Ob7Kysq+sufPn98JAAUFBb66ujrtcOvauXOn3uFwOBMTE8e++OKLfes+CwsLq8rKykrKyspKnn/++aOnKmflypUtGzZsiOzo6Djlu1ZRBn6xnTlzpgcAPvzwQ+Pxefu3IzY2VjqthyLnVEmdByV1HhRXu3Go3ovGrqEH20RJRnm9F8XVHvynpANdXuEctpQQQs4PFEsMjmKJY6jj4iLGGEPm5VeC51WQJXnIfILPi5qDX53DlhFCLiZqtTp0JulnQ6vV9r2NeZ6HKIp9Q9E6nU4BAJVKpfRPP15mZmZgx44dg54pmZ6e7t++fbsBAAoKCvxlZWUlM2bM6Pb7/Wf9nrTZbNLNN9/c8cc//jH6ZPlKSko0PM8jPj5e7J/+wAMPNP7hD38YdKSEnN+iwzUIN6gQH6lDhFGNiDD1kHlVPAebWQ2jjgc3yAwLSVYgiEO/wwkh5GJAscTgKJY4hjouLjJ+vx+HDh1Ce3s7AOA/r78CV0szGMdBZzQBkAEo/f70aKk6MlhxhBBySvHx8fWMsQHfrBhjcnx8/AmbSY2mBx54oOnxxx+P27dvnxYAJEnCQw89FAMAy5cvb1q5cmVCZWVl3zfMQCAwZOByun71q181v/rqq1GSJA1aVkNDg+pnP/tZ8u23397CcQNfybfccku3y+XiS0tLBw2QyPmjyyug23csVtTwDCqeITZCiwl284C9KvwhCV+Wd6GktudUEJdXQFu3AJ4B05wRsBzXybHzsAtflHedcPIIIYRcTCiWGBrFEj3oVJGLhKIocLvdCIVCCAaD8Hg8iIyMxGW33IrtG9+Ep6Mdai2HkEeBfML5agwBT/eotJsQcuHr3fH7fN8JfMqUKf7Vq1fXzZs3L9Xv93OMMVx33XUuAJg7d66rpaVFNWvWrAxJkpjZbJYcDod/9uzZff/nuHDhwlSdTicDgNVqFbdv337oVHXGxcWJs2bN6nzppZf61qcGg0HO4XA4RVFkPM8rc+fObf/tb3/bPNj9K1asaFywYEH68J+efFOCgoy9Vd1QlJ6ZFmE6HtV1bghBEYZACOkO24D8gZCMTo+Atu4QIs0aRJrUiDRpYNLzUPMnjifxDJAYwyCTMQgh5KJBscTQKJbowY5fC3M+yM/PV4qKaN+FM9Ha2oqmpiZYLBaEh4dD3dkOYe8ehE29An97ajUaK3r/N6GgZ9ZFb3DEYIkbgx8+tBphlgviCF9CyCkwxnYripI/nDKKi4urc3Nz20aqTeTSUFxcbMvNzU0Z7XacizhCkGRIkgKNisOBWg86PSF0+0TwPEMsk6Dt9mNMjAER2bHo9AjYU+VCQqQOWfFGHKx1o7kriAl2M6ymky/fVhQFigJwHPVcEELOHYolyGg4WRxBMy4uUD6fD11dXQgLC0N4eDgMBgN4Buz8619gtFiQodHD7PbAqygouPlWvPv4KkASATAAPMAYOJ6H1hCGqXPmUacFIYQQcppkWcG2fR2QZAVX5URgfIoJB2vccHlFKDKQPT4Kan8IvFkPAHjvqwaEgoCiAFnxRjgTjdCoOLS5BUQY1YOeHtKL0WwLQgghhDouLjQulwuBQAA1NTVQFAUGgwHjxo1DWFgYUpKS8JXXDVkIYq8oQ9Ptgq+9EWG1VVCpVRAlEWAMudfNgmPa1ags+gJJ4ybAPmHSaD8WIYSMmKeeeipy3bp1A44Omzx5smf9+vW1I1nPihUrYt99990B58vPnj27Y/Xq1U0jWQ85/wiSAkGSIcsAxxi6fSIaOv198xm/OuLG5IxwMI7hQL0LG/bUQq9W4bnJPe9bSQaOtgcAACnRemhU1DNBCCHnE4olzj+0VOQC4na7cfDgwQFpiYmJiI/vOd5YURSU7t6FsHALDm37F2oOfAVvZyc4noeiKJBEEVlTr8S3fnoX1FrdaDwCIeQcoOmdZLRcCktFmruCKDvqBUPPcpHLsiLAoOCzkk5Ics/SkaAgg+cYLs+ygFcBj289hOxYM26dnNhXTlt3CIoC6DQ9SzdNehpLIoScPyiWIKOBlopcBBRFQVVV1YA0xhgEQcCePXug0+kQ9HkRCInoam1F7k234rJb5mLnprcQERcPS0wMIsYkICJ2zCg9ASGEEHLh84dkyIoCUVKgAAiKEg7UuBH6+sjStFg9So96ERJl7K9xI0yvwtX2OIxNNkJ2B8DUPJhODZtZg5AoY3tpJxhjmJYdARVPMy8IIYSQwQy744IxVg3ADUACIB7fM8d6Fm4+BeC7AHwA/j9FUfYMt95LjSxJkMVjR60xxpAeE4PKlhbIsoxQqOeIYwUAB4ZAIIDopCRce8fiUWoxIYQQcvFJjtIhTMthe1kXAKDsqAd1HX6oOQ7WMDUaOoII0/IID1OhyyviaJsfigLUVLug6eiEyDGYp6WD5xhUHIPFqAbHGAY5UIQQQgghXxup1+QMRVEmDDGdaBaAjK//LAKwboTqvCSI7W1AaSn2ffAujrz/FlJraxDp6oIuGEBzeTlkSRqQ32AwID0zs2/5CCGEEEJGTnN3EL975wAONXWCY0BzlwBBVMDA4A5K+Ki8GZIiw5lkhCPBiOz4MKh4hm5RRpfCoZ2p4A30vLs5jmGC3QyLicf7+xshSPIoPx0hhBByfjoX/fuzARQqPb4EYGGMxZ2Dei9skoSad97G54WF6PzofThrS5EYZobZ40Z0SzPsNdUQ1F9PmPl6nxKdTocJEybAZrOB5/lRbDwhhAyfwWDIO1We+Pj4cY2Njac9e3DDhg3msWPHZqempubY7facRYsWJfRee/bZZ62ZmZnO9PT0nKysLOfcuXOT29raeAAoKCjISklJGetwOJwOh8M5c+bM1MHK37x5s2nChAmO/mmCICAyMjK3pqZGPWfOnJSXX345on+ZWVlZzokTJzqKi4u1p/sclxrGWCJjbBtjrJQxdpAxdu9otcUbEHBvtA7z9ADknvdvbLgGFqMa75cexd/21uDVXVXQqnjERWiREW/E+GQjspUQwqMMEJNtaOoMQhBlBAUZIVHGnz4+jOc/rcJ/DreO1mMRQshFiWKJi8dI7HGhAPgXY0wB8LyiKC8cdz0eQF2/349+ndbYPxNjbBF6ZmQgKSlpBJp1YfMcqYLh6FHU+/24MvgvMLEVOcY5+Ki4Gka9AVfaMxDV2YlavaHnBllGVFTU6DaaEHLJ+mrrB9YvN74R7+3q1IRZIkKXfX9e/YRvfbdjtNvV365du3TLli1L2rRpU0VeXl5AEASsWbMmCgA2btxoXrt2bcyWLVsO2+12QRRFPPPMM5H19fUqm80mAUBhYWHVVVdd5TtZHbNmzXL/9Kc/1ZSXl2uysrJCAPDuu++aMzMz/cnJycLx+XvL/OMf/2j73//938SPP/644pt49ouACGCZoih7GGMmALsZY1sVRSk51w1J8gXQquYggfX0WzDAZtYhzqrFtd5YVLR4YDFoBtyTYNEidDgIBACPSocOn4SDjV2o7wxiQkIEZo2NhVmvxqFmD460+XD71BRwHO13QQi5tFAs0YNiicGNxIyLaYqiTETPkpC7GWNXHXd9sDfvCUeZKIrygqIo+Yqi5NMXcKCmohKtQghhjAPjIwDOAEFWodPrRrOrEwyAPiQguqUZFlcX1BoNzGbzaDebEHIJ+mrrB9ZPXn0x2dvVqQEAb1en5pNXX0z+ausH1lPdezo2b95sKigoyJo5c2aq3W7PufHGG+2yfGxK/WOPPRbtdDqzMzMznXv37h3yyKRHHnkkdtmyZY15eXkBAFCr1Vi5cmUrADz66KNxq1atOmq32wUAUKlUWLJkSXtubm7wTNrK8zyuv/76jsLCwr5nf+ONN6w/+MEPThp4XXvttZ6ampqLdpRkuBRFaezdH0tRFDeAUvQMgpwTsqzgYK0Hhxu9cDV7UGowoyTcAoUxcKxn34va1gDMGg3um5GN+RPtqGj0YuehTnxZ3oUWr4C9ehM6Em2IMGsRECQ880kF3vqqFn5BxJUZUfj+hES8V9yAd/bWo8MXOlePRggh5wWKJY6hWGJww+64UBSl4ev/bgHwNoCC47IcBZDY7/cEAA3Drfdil5qTAyPPI1GrRjDsOijWOxBujMFM5wTMmnIlmMGAiOxspM6+CY6Zs5A3cSJMJtNoN5sQcgn6cuMb8ZIgDHifSILAfbnxjRH7YllaWqpfu3ZtXUVFxcHa2lrt1q1bjb3XbDabWFJSUnrHHXe0rlq1KmaoMsrLy/VTpkwZdJSjoqJCP3Xq1JOOgCxcuDC1d3rnnXfemTBUvttuu63jH//4hxUA/H4/27ZtW/htt93WebKy//GPf4Q7HA7/yfKQHoyxFAB5AHYcl76IMVbEGCtqbT27JRdyMIj2l1+Bd8fOAekBQUaLK4i6tgA4uw0eRYbByCMgCtjf2IWqZh+cCWFIjTGgzRvEC/+tQOH2GtS2+uH2i6hpDaBFAEo6en42G9S4YfwYfD8/AdeOtaG61Y8jzX7cPC4RS7+VCZvxoo07CSFkUBRLDESxxImGtVSEMRYGgFMUxf31z98G8Pvjsm0C8AvG2JsApgBwKYrSCHJSXHQ0wrRamFRq8BwDxxjAGGxJKcC4cUBqGqBWH8vP0XbkhJDR0Ts6crrpZ2PcuHHetLQ0AQBycnJ8lZWVfWXPnz+/EwAKCgp8mzZtihhuXTt37tQvXLjQ7vV6uQcffLD+Zz/7WSdwetM7AWD69Ok+n8/HFRcXa/ft26efMGGCNyoqShos78KFC1N1Op2ckJAQfO6552qH2/aLHWPMCOAtAEsURenuf+3rpaovAEB+fv4JMztPh7+4GN2bN8O/ZzfCphwbhzFoeTjijSg96sF/a30obvVC1+XDwQYXqtq90KgYJqVZ4EhQ4c3dtdjX0IWAIEGv5jHHnoAxVi38QQl+QUJIUJAeF4YlYzP6yhfFnjUn+SlWTLDT7ElCyKWHYomBKJY40XD3uIgB8HbPiadQAXhdUZQPGWOLAUBRlOcAfICeo1Ar0HMc6u3DrPOSoNVq0X3L9yErCtRtrUDJQSAQAK66CogeshOQEELOuTBLRGiwwCLMEjFi8921Wm3fF1Ge5yGKYt8yRJ1OpwCASqVS+qcfLzMzM7Bjxw7D5ZdffsJoRHp6un/79u2GG264wV1QUOAvKysrWbhwYZLf7z+rXuGbbrqpo7Cw0FpeXq6fO3fukFM7TzeAIQBjTI2eTou/Koryj2+iDsOECQi/5WbosrNPuBYVrsGRFg6KwqO52w+rXovs2HBwPEO+PQI8B/z/7N15fJTVvfjxz/PMPpNJMpN9JwlJhoQQIewqiCtqi1VQUStcF7y1ekv1Vm29/d1r8epVLZyvIAAAIABJREFU20tLq61L1RYxV8UNRRTjUpBFdgOSfd/3STKT2Wee3x8xUQqIrEE479crr1fmPM8855wHkvnmPOd8T32nm2mpVoKhEAlhJrLiwshKMCJJEnqtiq4BH/0eH9FemeLSQc4bG41BqyIj3ohBJ5NgOezsZEEQhDOaiCUOJmKJAx3XY3pFUWoVRSn46itPUZRHvip/+qtBC77aTeQuRVEyFUXJVxRl54lo+NkgJiaGmNhYyM2DBdfBjxeJQQtBEE470xfc0KLSaA7Yx1Gl0YSmL7ihZbTadCi/+tWv2pcvX56wd+9eHUAwGOShhx6KA7j//vvbf/nLXybX1NSMTGXzeDzHnB1x0aJFva+//nrUli1bzDfccEPf8bf+7CYNPSF5HihTFGX5SatHq8V6000YJ0066JhaJTEjJ5K0eB0VXQNsru9iRnoMD1w6DpNazctb6njo3f1srunBYtJy3bQkpmdGUvtFKy989gK7B97BHQjQ3O9ia00vf/iokqLtDQDoNDJjYo3oNGL2pCAIZycRSxxMxBIHOhG7igiCIAhnseGM36d7JvBp06a5H3/88aYbbrghw+12y5IkcfHFF/cDXH/99f2dnZ3qyy+/PCsYDErh4eFBm83mvuqqq0aWIwxPxQSwWq2BLVu2VB6ursLCQo9erw/l5+e7wsPDQ4c7T/jOzgVuBvZJkvTFV2UPKoqy7lQ2QpIkxsaZ+fklWejVKhIidNR2uPC6fLy5vZnKPhfVBge3nZ9OTISWzno74T39jPWY+Uv/mzSXJoEi8YvLsnl9t493Slq5fkoqYbqhcExRFOq6BxkTZRK7igiCcFYRscTBRCxxIElRjmkZ6Ek1efJkZedOMTFDEAThWEiStEtRlMnHc42SkpL6goKC7hPVJuHsUFJSEl1QUDBmtNtxMuOIimYnTd1uMhOMlDe76Pf4CLV2UdIxgCfcRHqimRkZUUzLiMLh9LJ7Uw2K1YkUZeSXr3Zgd/lY9oPxfFjRjtsf5JmbJ48MXLy+q5m/b6nnusnJ3DxjDMWlHdR0Obn13HS0ajEbQxCEU0fEEsJo+LY4QnwKCoIgCIIgfAc763vZ3+zkpR0NLHlpJ30eL31uL8/U9tGq1nHZhCQ+Ku3g9x9V4fEH2d7Ux69L2vnDdj8hVwIXjI3DrNfQ2dDLjy06ZieGo1UNhWLBkEJ0mBa1SiLGPLSryMqt9by3t42Kdsco9loQBEEQRp9YKiIIgiCcUVasWBH1l7/85YCEQFOmTHG+9NJLJzTT9gMPPBC/Zs2aA/aXv+qqq3off/zx9hNZj3D6ePLTahJNRso7+vEEg0SGaYgN03HL9AzSoow43UFSLWE02J38+K/beGx+Phq1jMcfwB8Kcc3EZHKTzVwcqeWXH1TQo0BhXgLRBj3VbS4y4sN466fnjtT3s4uyqO8eJC9R7DQiCIJwKolY4vQjlooIgiCcYcT0TmG0nOlLRT4t76SsrZ9XdzTTM+jl2R9PpqXXw77WPpzeAFaDgS11nbT2u4kP17H04mye+6yWZIuRFQvP4atd2FAUhRc+qaJuwMMvL89l4347W+u6cQV9PDp/PDq16oS3XRAE4WiIWEIYDd8WR4gZF4IgCIIgCN/BHFssSRYDb+1pwWrS4vIHeXV3I9WdTsINGsYnhZhfmExLnxurScODb+3D6fFjNer4vLKPKLOW7EQT/qDCO2WdhBSwO70oikJypIF15T30OH0kRhpGu6uCIAiCcFoROS4EQRAEQRAOwVtbS/vD/03nph1sLu8gEAyxva4XrVrF7OxYLs6Nw+MPYtSqWDwjjR9PS8UbDLJkVjo1XYM4PAE0KpkJSRFUtTup6RgEQKuWWVCYzKTUSFRqiexEExNSw/nPH+aKQQtBEARBOAQx40IQBEEQBOEQBj//HNfuPXxRP8gL2Rez5DIb8yclEx2mZXpGFAD/cWUur+1sYlxiODtqe/movBOtSubH09P4vLaHeLOBHqeP7d29TBlj5XJiAShMs/LazmaWvVvG0zcXkoVpNLsqCML3VG3dn+jv38U426Po9Ymj3RxBOGnEwIUgCIIgCMIhRMybh2wwEeaLJq1PIjXSRFe/j8vy4kfyVQRCIao7nfzvhxU09riwmLTsa+lnT2Mf09KtlLc6yY4OZ0qGhXGJZm5+fhvnZ0VzbWEKmTEmxidFjHIvBUE4HfR3uVBrVJgidd/pfKfdi96kxuNuxO+309e3h0HXa+i0McTGXoZWG32SWywIp5ZYKiIIgiCcloxG48QjnZOUlJTf1tb2nQbh//jHP0ZJklS4Zs0a83DZypUrIyVJKnzxxRctAH6/n7vvvjspLS1tvM1my7XZbLkPPPBA/PD5KpWqcLjcZrPlPvjgg/GHqgtg6tSpOWPGjBlvs9lyMzIy8n73u9+NRJHnn39+Vk5OTu7YsWPzbrzxxtRAIDDyvoceeiguPT09Lzs7OzcnJyf39ttvT/Z6vdLw8c2bNxskSSp84403wgHa29tVw+2Jjo4uiI2NnTD8urKyUjtt2rTsjIyMvLFjx+Y9/PDDsd/lXglDVGFhRF59FZdefy5P3jGD9h4/+xoc9LsCOD0B9tY7yI4J54apqczIiMLpDTLoDdBid9PS5yaoKEiygo8g675s47mNdeyst/O3zfVYTFp+MCGRT8o72VrTM9pdFQRhFLkGfHz4fCkfPr8fJaTQ0+LknRVfsGNd3UHnOp0VVJasZ93Te3nrD59Su1OPyZBNVfXD1Nf/marqR9jzxW14PG2j0JPTj4glzpxYQsy4EARBEI6b8/NW68DHTUkhh08rm7W+8ItSWsKmJ/aOdrv+WVZWlruoqMh61VVXOQBeffVVa05Ojnv4+NKlS5M6Ojo0ZWVl+41Go2K32+WHH354JKDQ6XSh8vLy0u9a38qVK2tnzZrl6ujoUGVlZeXffffdPXq9XlmzZk2N1WoNhUIhLr/88swXXnjBcscdd9ifeOKJmI8//jh8x44d5dHR0UGPxyMtW7YsbnBwUNLpdArASy+9FDVp0iRnUVGRdf78+QPx8fHB4Tbde++9iWFhYcFly5Z1ADQ0NGj+93//t/m8885z2e12eeLEiblXXHHFQGFhoedE3dOzhaKASobqTifaeoUNFV30Ov389IKx3DgtFYBJaRb2NffT7/aTGRvGleMTaHd4+Li0g5rOQa6cEM9leXHMzo4BoL5nKA9GU6+LGZlRo9k9QRBGkUavIiJGz0C3hw9f2M+Y/Gg66vtpKu+lcnsHl92eR8DvIKD5lJqq51ECFpB+hrO/Fym8l67u7QRDvYCMTpeMVhOBLH+3mRunExFLHJqIJYaIgQtBEAThuDg/b7X2ra1LIxCSAUIOn7ZvbV0awIkIONauXWtetmxZotVq9VdUVBjy8/Ndb7/9dp0sD00afOKJJ2LXr18fEQgEpFdffbV24sSJh/0gnTZtmnPbtm1hXq9X8ng8Un19vS4vL88F4HA45KKiopi6urq9RqNRAbBYLKHly5e3Hm8fBgYGVAaDIaRWqxUAq9UaAvD7/ZLf75eGlx0sX7484R//+Ed5dHR0EECv1yuPPvroyF7uoVCItWvXWtavX185Z86cHJfLJQ239VDS0tL8aWlp/uG+ZGZmuhsbG7Vi4OLoybKE1aLm/Y0tbKpX4/YHcXj8GPUjD7Bo7HHxyo4mBtx+LCYt4xMjyIk382lFFyFFYeGUVGLD9SPnL545hukZUeQmhI9GlwRBOE1UbW+npbIP76Cf3tZBfO4AkXEGOuudOO0eNhRVEtBswBi/C7UxiBoLV/97Ift2/56BwQp0eisuVy8gkZKyiKTEG5Dl79e2yiKWOLKzPZYQS0UEQRCE4zLwcVPScKAxIhCSBz5uSjpRdZSVlRmeeuqppurq6v2NjY264uLisOFj0dHRgdLS0rJbb72167HHHov7tutIksSsWbMG3nzzzfCioqLIuXPn9g0fKy0t1SUkJPgsFkvocO/3er3yN6d3Pvfcc5Zvq2/RokUZ2dnZufn5+eN/8YtftKrVXz8vOO+887JiYmIKTCZT8JZbbrHb7XbZ5XLJNpvNd7jrFRcXh6WkpHjz8vK806ZNc6xevfo7J0ioqKjQlpaWGmfPnu38ru8RDjQ2Jozrp6Rw15yxLJ45BpUs89ymWgC8gSBv7WlBp5ZJiNDj8Qf5uKyD+1aXoFPLuHxBnt9Uh8cfHLmeRiUzPikCWZYOV6UgCGeBupJuvC4/skpGrZXp73SjVqtQaWRUKhm1TqanejwdJfPwtF9BYuy/o9GEMXbcVSSlXEhExCS02jj0+hTq6/9EVdV/j3aXjpqIJQ5PxBJDxMCFIAiCcFxCDp/2aMqPRX5+/mBmZqZfpVKRl5fnqqmpGbn2jTfeaAeYOnWqq6mp6YhzY2+66abeV155xbp69Wrr4sWLD/sUZ8WKFVE2my03Pj5+QnV1tQa+nt45/LVkyRL7t9W1cuXK2srKytLa2tq9Tz75ZHxlZeVIuzdt2lTV3t5e4vP55HfffTdcUZSRhI8Ab7zxRrjNZstNSkrKLy4uNgGsWrXKumDBgl6AhQsX9r7yyivWI/UXoL+/X77mmmsyH3vssabhJzTC0WnscrO5vJeUCBOv72rG7QuSFGkgPXpoNxCdWsWDV4zjoXl5XDQuDlt8ON5AiMoOJ619btr63fxtSz3vfHHcD90EQTiDVO/qxO30M+HCZGbOz0SlkXE7fHQ2OlBrZBRFobNhACWoxtsXw0D9eRAYmqUVHj6B6JhLCAScGAxJRERMJBTy4hwsG+VeHT0RSxyeiCWGiIELQRAE4bjIZu0hR/UPV34shtdkAqhUKgKBwMinsl6vVwDUarXyzfLDmTNnjquiosLQ29urnjBhgne4PDc319vW1qa12+0ywNKlS3vKy8tLzWZzMBgMHtcj8cTExMD48eNdGzduPGDPS6PRqPzgBz/oe+uttyKtVmvIYDCEysvLtQDz588fKC8vL83OznZ7vV45EAjw/vvvW377298mJiUl5d93332pGzZsiBhu7+F4vV7pyiuvzLz22mt7Fy9e3Pdt5wqH99quRn73SRmba7qo6x6kptPJc4smc97YGN7e08zyDyvIiTczJsrElpoeXL4AE1Mj8QaCRBi16NUyOrXMrgY7wZBCKHToWblfNPVR3ek4xb0TBGG0VO3ooK/DRZhFj86gwtXvw+cJoiggyRI+d5CAPwhffQoN9nvZ9UEDva2DKEqItrY3GRwsx+ftJhh0kJf7e3Kyl41up46BiCWO7GyPJcTAhSAIgnBcwi9KaUEtHzjyrpZD4ReltIxSk45o2bJlzQ8//PAB7TObzaGFCxd233bbbakul0sCCAQC+P3+457H73A45P379xtzcnK8/f39ckNDgwaGMo9/8MEHETabzQ3w85//vG3JkiVp3d3dKhhah+r1emWANWvWhNtsNld7e/velpaWfa2trfvmzp1rLyoqijxcvaFQiIULF6ZlZ2d7HnrooY7j7cfZzKRXYzaomZgWiVmvZmttD4te2Mb9r5fwyLpyPq3oYke9nWSLgcvz47lpWhqVHQ7c/iCKohAXYSAx0sDYWBOzf/spS1buPKiOtn43/+/tL/nlG/tGBjaqOhyUNInxJkE4UyiBEIGekTyOqLUyxggtg3Yv299rQGHoZ19WfJh0Q+eF/IACkgRagwq1Rqbmi042fbiM1kovCfHXEhU9h9iYuURHzyEsLHs0unZcRCxxZGd7LCGScwqCIAjHZThp1vchE/iw6667buBQ5StWrGi55557Em02W57JZArp9frQ9ddf3z2clGp4Xerw+RdeeGH/n//858MGVYsWLcrQ6/Uhn88nLVy4sPv88893NTU1qa+88sqxPp9PCoVC0rnnnjtw3333dQHcf//9XS6XS548efI4rVYbMplMoalTpzpnzJjhuuOOO1LmzZt3wF+w8+fPtz/zzDOxd9111yHvdXFxcdjbb78dlZWV5R5u929+85uW66+/vv/o79rZ7SezM/jx9FR2NtgZcPvpd/uxGLXY4s1MTI0kPymS87OiUatkfnrBWACWrNyJUatiXIKZ8nYH/oDC9jo7vYM+Gntd/POUXqtJS2GahSiTFlmWCIYUHnhjL4GQwl8XTT4gsacgCN9PAx814qmyE3H5GPRjLZy7IIvW6j52vleHP1BDyvStdOybjeSUsfeEAQrD0y0UBZSggjFCS2SCgQ1FE1FpPUSHTWKgWyHtB+mj2rfjIWIJEUsciaQoh00gOmomT56s7Nx58JMIQThRgqEgf9v/NybFTmJi3ND2zsFgiL4OJ3qjFkktoTdokFViUpLw/SNJ0i5FUSYfzzVKSkrqCwoKuk9Um4SzQ0lJSXRBQcGY0W7H8cYRn5R30Ov0kRkbxvjEcDTqr7Pzf9ncz/tftnH9lBRizHoMWhU7G3p5dG0ZN01P49K8OExaNbIsUdLUx91Fu5EkmJ0dQyCkcOu56Wyo7OLy8fEkWYwj1/UHQ7z/ZTvjE8PJiBnJF8eKj6oY8Ph5YK4NrXroMykYUmjtc5Ni/fr9giB8Pzg3t+Da203kvEy0SUM/67vXN7DtnVpkfR+Ryc24eqLArscdNKMgAxKyGgxhWvQmDQPdbmS1TMAXJCJGT2RcGI4eD+dfl0XC2MM+OD8qIpYQRsO3xRFixoVw1lhTtYb/t+X/oaAwLW4aJd0lxBpjee+a9wDobu6nubIbvy8ACmj1asKjTGRMiBcDGIIgCGeJYEjhDx9V0TngxRsIYtZrePn2aSODBCs+qaK938McWyypUUPLjF/+vJF9Lf08v6mW5zfVMj4pgseumcD2ul4UFPpcAd7a3YJOI5OfFMHt52ccVO/m6m6e/KSKSKOWV5ZMH9lpZOnFWQed++zGWtbta+PfLhzLpXnxJ/FuCIJwooWdm0TYuQdulJEwNgJTuBavJ4q+OgvRqWZUVgmlzYnbEUCSJWJSzUTGGAn4grgcflRqCb87SNAPM67OpK/DRXzmd94cQhC+d8TAhXDWeGbvMyPrBrd1bENGZmLM0GwLRVEY6HHhdflRUDBIDizeZuwdWTRVqUmzxY5m0wVBOAorVqyI+stf/nLAVmZTpkxxvvTSS40no75LLrkk858zkD/yyCPN8+fPP+QUUuH0ppIlpqVb2d1gZ29LP75AiEFfYOT4eWOjqOkc5JH3yugd9HLu2Bg+q+rCH1Jo7/egAD2DPj4s7WDyGAtlbQOY9Co+KeukY8DLPyq6uHFaGq19boq2NZJiNXB5fgIFyZEEQgodAx621fUyIzPqsG0cXkoSYdCcgjsiCMLJlpAZyTX3F7LtnVraqvvQmzQ0lfYS9H+V8kFSyJwYQ9WOTrQGFT+6ZyK1ezqp2tFJZmEMkbFGImPFDKwTScQSpx8xcCGcNW4ZfwvLPv86y3KIEI2ORj6qXsve6nfResdjC03FIIURq6kiUtVOMKChpUJDS0U3E2anExZpGMUeCILwXSxdurRn6dKlPaeqvuLi4ppTVZdw8imKwva6XnyBEHkJ4Zj1aga9QwMXH5d3sHJrA1mxYXzZ2o8/qNC5p4VIo4Y4s44rJyQwLjGcjRVddDk8uP0Bll9/DgBqeT/FpR3MHR9PVYeDX725j5LmPgwaFS12N/demsP9l+Wwp7GP3MTwA9rzzTwYANdNSWFBYfLIrAxBEL7f3A4faq2K2Tfm8MnKcuxtg5gitPg8waGHakHY9m4d0UkmLrx5HNvfraO3bZAZ12SSlnf4QU7h2IlY4vQjBi6Es8a1OdeyumI1Zfav97be07WHpoEG+rx9zNUrTDWcg1/x0RnIIIgWeyCZ4SwwzVXd2KakjE7jBUEQhFNCkiSWXpxFj9PHFfnx3PDcNn7zbikv/MsUKtudtPd7sBg1I7t+5MSHcf9cG5EGDb9YXUJDj4v75uZw6992YB/0gwJXT0omJ87Mlupu9rX08+LmeqwmDXHhehIj9RSOsQIwd3wCc8cnAEMDFj9/9QscngB/umEiJt2BIZsYtBCEM8Ngn5e1T5XQ1+kiKjGMqOQwZJWE3x9EVklY4kzYOwcJ+kIM9HhwO/0M9nuRZNj9QQMqjUyKzTra3RCEk04MXAhnlYyIjAMGLgC0oQBWtR61yoIKLQFUeJRIWvyRKMNZnIH+LhduhxeDWXeIKwuCIAhnigttX88OvsgWy64GOw09Ls4bG822uh4utsXS2Osm0qhhyhgrj79fTs+gj2a7m4p2B4VpFhIjDXgDIf5vRxNdTi86tQqtWoVBo0ItS4DEnRdkcv3kFNSHyKOkKNA76MPtC+IPhg46LgjCmUFWSwR8QUIBBWePk2B3A+6gGVlrIvf8BArnjmFjUTmVOzvR6FR8/nYNXleA8BgDA11uBrrcYBvtXgjCyScGLoSzRpOjiQ3NGw4ok5HRhhR+ZEzhsvB4wv07afGPx6nEACAFgyhBP7JWT9AfpHFnDem5UWjj4g5VhSAIgnCGOSfVwkdlnfx9Sz0XjYtlXHw4l+bF0zPowz7oR6OSQZIwaFRIEljDdKRYDXj9IbJiwhj0BSlp6udPN0zk4tw4kiIN7G608+u3vuSR98r4fXEl4+LDuefSbCalWkbqlWWJH09L4x+VXXgDYuBCEM5UhjAt1/x7IeWft5FqqaXtw81gsOLLu5lxheGovH3MuTmX9ImxfP5WDV6Xn/Gzkkk/Jxpnr5fo5LAjVyIIZwAxcCGcNUwaE1aDFZfDhSzJBJQAKklFT9DHW4NNXBPWilWtxq/o8fnD8HsUCPiRNNqhC3hc9H/2EaXvOCn47UNIavHjIwiCcKYrTI0kFFKo7HCwubobh8fPC5vrCDeo8QUU/rqokFtmpmPUqeh3+4kx6/D4g1R2OEmxGogO05GfFIksSyR9lSdpYkokP7toLPe9vpcBT4C2AQ/VnU5+u76Ctj43r/3rDGLD9Wyq7ubLln52Ndi5Ij+B8vYBosN0RIeJmX+CcCYxhGuZeGkakEZUlAzWdBzBGD544nUitV1En38lWedmMGthNhq9mpgU89D7wrSj23BBOIXEHo/CWcOqt/Jvk/6NjMgMEk2JaGQNASWAV1a4LLGAzoQI3HoNMeoGxmk/JGzb22hLd4JWD45+GBxAMYShVkugUo12dwThjGc0Gice6ZykpKT8tra27zSK+Mc//jFKkqTCNWvWmIfLVq5cGSlJUuGLL75oAfD7/dx9991JaWlp4202W67NZst94IEHRvabVKlUhcPlNpst98EHHzzsXpRer1f66U9/mpSWljY+KysrLz8/f9xrr702knVx8+bNBkmSCt94443wb77vUP2+9957E2NjYyfYbLbc9PT0vJtuuik1GAwC8MILL1jGjh2bJ8ty4caNG0Va+RNEURSe3lDDi5vriTBqcHj8aNUS/qCCLxAiK9rMFeMS2VnTx5cNToJBiDXreOz9cp78tJqfzMpk9c5mfvNuKY+9X4Z90DdybUmSSLYYmZAUwZ2zM3hiwQQsBg1fNvfTMeClxe4G4CcXZHLHrAwuGhdLZYeD+1bv5T/e2jdat0QQhFMhdTp88giOv/8r3YNW2pwplO2wU72rk8QsCzH2D2DVAmjfx2C/l9bqvtFu8WlNxBJf+77HEmLgQjhrDPbZybVH8RvzbTyT/d9cPfZqVLIKnaxlpi4cnSTjCqnwoae9OxapZA+6bR9CMAg6PUREEcyfQe5vfnVQhndBOOvteN7K77LzeSiykN9l57Pj+dMyU1hWVpa7qKhopG2vvvqqNScnxz38eunSpUltbW2asrKy/eXl5aVbt24t9/v9I5+VOp0uVF5eXjr89eijj7Yfrq577rknsb29XVNeXr6/qqpq/7p166oGBgZGRj1feumlqEmTJjm/2Z5v85Of/KSjvLy8tLq6en95eblh3bp1ZoBzzjnH/cYbb1RPnjzZebT3Qzg0XyBESb2DkoZ+iss6+PlF2UQatVhMOixGNVaThulpMSRFmNhQ1UVl1wBatcSgN8jO+l621/USCIawGrXYB328t6+N/3n/6/xKH3zZxms7mxjwBJAlGbUs8cs396HXyvz6ynEYdEP/TWRFJi8uEq1KJsqkJdVqJD8pYrRuiyAIp4JvEHprCR/cxY+i/5vzLw8jJS8GWSXhdvigqwwGu6C3jq1v1rDptSpaq86QwQsRSxxExBJfE3PdhbPCQEcb+z94h1AwiEqjRfYE+FnhLYx5+TN685LIDk2kuS4JlUqiVwngMYFvzjUohjBAQVVbSjBjHFJfNx1/fJKEX94/2l0ShNPHjuetrP9VGgHv0Ieys0PL+l+lATDltt7jvfzatWvNy5YtS7Rarf6KigpDfn6+6+23366T5aHqnnjiidj169dHBAIB6dVXX62dOHGi53DXmjZtmnPbtm1hXq9X8ng8Un19vS4vL88F4HA45KKiopi6urq9RqNRAbBYLKHly5e3Hm2bh69VW1u712AwKAApKSmB22+/3Q4QCoVYu3atZf369ZVz5szJcblc0nCdR+L1eiWv1ytHRUUFACZNmnTY/grH5i+f1lC8v4M52XEkR+l564tmwnRqki0GgkEFu8vHmyWNJEQY2dtmZzDgxxqu5trJKTw2fwK7Guxc++xWmu1uxiWYaexxMS5h6GGY0xvgqU9rUBSFey7JZkZmFF0OL5IEBo2akuY+XtnRxEPzchl0QjCkEB2uJSpMx1M3TRrlOyMIwklntML8v6Lb8gz6+s9QpypUbfdR+lkLXxQ3ojddTsaYS8hLKySxtxuAiFjDKDf6BBCxxEFELHEgMeNCOOMpikJN6XaCgQAoChGJyWTMnI1cWsW5beG4Gup4z96GghoDDSQObkOzZR3+8dMJZI6HUJBgdDxSbwe6vVtwf751tLskCKeXDY8njQQawwJemQ2PJ52oKsrKygxPPfVUU3V19f7GxkZdcXHxSDay6OjoQGlpadlZjMkGAAAgAElEQVStt97a9dhjj31r5lxJkpg1a9bAm2++GV5UVBQ5d+7ckcdUpaWluoSEBJ/FYjlsJkSv1yt/c3rnc889ZznUecPXslqth7xWcXFxWEpKijcvL887bdo0x+rVq4/4GP3pp5+Os9lsufHx8QXp6ememTNnuo/0HuHoKYrC6l1NNPW5aHY42d86wK4GOw9ePo6MmKH/dgoKPW4fY2L1nJMcSXXnIE9+UoWiKOTEmfngy3ba+jx4/EFquwaJjzCgVcsMegI8+XEV45PCWTRzDBeNi8OoVZMWZeKFf5nC84sno1PL6NQysWY9OUkm0uMMmA1ieaIgnFUsaWguewj19S+C7QoG+70oQMAfpKPeyZ7PvFT89UvGzUzg4ltyMUWcAXlvRCxxEBFLHEgMXAhnvJa9eyg31+LXhUCrgjAdP/3kLh4NvYfv4lTez3GwyvE+SfIm4tiP2x2kPhgNLgdSbweq+nIkYxjqzhYUnQHzxRfh3r9/tLslCKcPZ+ehs4MdrvwY5OfnD2ZmZvpVKhV5eXmumpqakWvfeOONdoCpU6e6mpqajhi93XTTTb2vvPKKdfXq1dbFixcf9inOihUror76cJ9QXV2tgYOndy5ZssR+LP1ZtWqVdcGCBb0ACxcu7H3llVeOOMVzeHpnV1dXicvlkp999tlDBjrC8ZEkiR9PSSc/MYLSVgdOT4D67kGWvLSTTys6UaskNCoZvUbFPRfnsOxH47k8P54H5tqQJImKDgfdTi/RYVoyYsJIsRq4emISc/PiWfTidp7fXMvbe1oOyHkBMCE5EpUssb3ejj+kkBChJy5Sx5hYo1ieKAhnOMUfwrG5BW99/9eFai1EjwVgypXpXHZ7HlOuTMcUocOglYm26OBM+tUgYomjdrbFEmKpiHDGU5QQBsVA03Q/0zWFVP5jA6ZQiNmqPFIyx/JvIRWD1vG4nPHIe9cx2KZQf+GthHW1kF/0ezQuJ+7ZP0JTvRdNzZcMlMcwuHkLY/6vaLS7Jginh7BYH86OgwOLsFjfIc4+JjqdbmTqo0qlIhAIjIRrer1eAVCr1co3yw9nzpw5rjvvvNOg1+tDEyZM8A6X5+bmetva2rR2u122WCyhpUuX9ixdurQnKysrLxgMHlV4+M/X+uaxQCDA+++/bykuLo5cvnx5gqIo9PX1qQ917qHodDrl0ksvHdi4caP5jjvuOKZgR/h2Vxcm0uvy8I/qTryBIIGQQo/Ty0W2WHyhEC12N+F6Nat3NQ/NpPAG+LSiixmZ0dR2OZlXkEhWnJmcODM6jUxcuJ5QSKGhZ5BgCEDh/S/b+dfZmQAEgiHueGkXXzTa6XP7MWpVvL2rlGsmZyPLYtcAQTjTeRsGcO3uxFc/gG7MgQ/NOxsG+OzVSiLjjKh1KiZclEzO5Dj0Zu2ZNagpYomDiFjiQGLgQjjjJRcUEu8bj1qro6+lkR59FNMCmdiMMahNsVyT/x/4DXr0aj1uSwxyVTVqKUR/TAqeDBty6S70H6xC5R5EArRpUzDNnDna3RKE08fsB1oOWJcKoNaFmP1Ayyi26lstW7aseXi96DCz2RxauHBh92233Za6atWqBqPRqAQCAfx+/1FHhsPXWrJkSeqqVasa9Hq90tDQoHnvvffMcXFxAZvN5tq0aVPV8PnXXHPNmKKiosi77rrriOt4Q6EQW7duDSsoKHAdbbuE70hW2FrfgyxBW58bRQG9VqbH6aW130Oy1Uh6lIm/flaLPxjCH1Qw69TIskRp6wAX2mK5JHdopnHHgIdQSEGWJe65OJutNT2oVRL/MnPMSHWf1/awrbaHQDCEAqRGDBKl2sS+mjoKsn44OvdAEIRTRpdmxjgxFm1y2EHHtM56VFIAvzeIvd2FIUyDIUJHS4Wd/ZtamXhJKjGp5kNc9XtGxBIHEbHEgY554EKSpBRgJRAPhIBnFUVZ8U/nXACsAeq+KnpTUZRlx1qnIBwLSZLQ6PQAWJLTyL/oB0ypfwc5UE8gLIPfbluB0aviuoR5+J97AWdLK+Nu+Xdic7Jwzr4Yec8mVO5BlLQMzJljiHvwV2iTk0e5V4JwGhlOmrXh8SScnVrCYn3MfqDlRCTTOlmuu+66gUOVr1ixouWee+5JtNlseSaTKaTX60PXX399d1pamh++Xpc6fP6FF17Y/+c///mQQdUf/vCHlp///OdJ2dnZeTqdTjEYDMH/+q//ai0qKrLOmzfvgBTw8+fPtz/zzDOxd911V6/H45Hj4uImDB+78847O2BoXeprr70WFQgEpHHjxrnuu+++Thjahu2+++5Ltdvt6quvvjpr3LhxBwQywtGLNeu4tjCZTdVdlLcPJVj3+ENsqu5Gr1WjkiW+aLSjkiX0ag2D3gBBRWF7XQ9jokycOzYagE/LO1leXMkV+QnceUEmN01P46bpaQfVFxeux5YQznljo0iPNlHd1sqgr5eSSjcTxipn1lNVQRAOImlUmM87RCqHnhoiN97JVeFqOi98DacDkrKHZva31fbT1+Gis2HgzBi4ELGEiCWOQFKU75R49OA3SlICkKAoym5JkszALuBHiqKUfuOcC4BfKIryg6O59uTJk5WdO3ceU7sE4Ttx9aI47PiqFf6rZwWXyzPJNqZjDnjYt3k33ZcsICkxAsv/3I/ji72o/F7M9/yC1MU3jnbLBeGIJEnapSjK5OO5RklJSX1BQUH3iWqTcHYoKSmJLigoGDPa7TjeOKLX6ef5DfV8XNlGZacDRQEFkCVIiNCTlxjBlppuchMj+NfzM1i1rYGmXhcdDi9jrEZ+v/AcTDo1jT0uHl1XxoLCZG6eMeageprtLv5zzX5mZkZx+/kZI+WflHew4qMqfjAhgSWzMo+5H4IgfM95+uH5y9jbNoFyribvsgnknT80wOHzBGiv6Scp24JKc+LTFopYQhgN3xZHHPOMC0VR2oC2r753SJJUBiQBpd/6RkE4RfwhP592r6PfZycQDJBgSMSoNjMQ6KMgYgrRUSlo2tbyS/9CWt06vLE6EsZPJ2v6TCL6fWTGG3FMn0pHegHBS3/IuEliloUgCMJokSRpLrACUAF/VRTlsZNVl0qCL9v66Bn0MTHVwjkpEaz5ohWNSibZYsQdCKJVqZiYEolBq6J9wEOvy0eYTk2EUcudq3bj9AZ4aF4ur94xg2c21vL0hhr+dVbGAbMnuhxeuhxevmjs492SVi7Li0eWYP3+dkIKxIbrT1YXBUH4PtBHwA//QMSHH6HtDUdWyex4rw7bjATMVj2peVGj3UJBOGVOSI4LSZLGABOBbYc4PEOSpBKglaHZF2I7BuGUCCoBWlyN+BjalnjAZUeDliABepxtzG6ZQKQrgcGAG18oAkNHD5g+JTp9DtHhWraW27EX/BBvQIEAVHU4yEkIH+VeCYJwJCtWrIj6y1/+csBWZlOmTHG+9NJLjSejvksuuSTznzOQP/LII83z588/5BRS4ehJkqQCngIuAZqBHZIkvfPNWZ4nUoRJw2MLxvOzV/bg9gX5YUES56RYWF5cSbfTi8MTQCXDwikpKECyxciV+QnMK0hi9a4mVm6px+UL8MbuFp7ZUEt5u4M0q5Fok5Zel4+bp4/BoFUxMdXCf/0wl//9sIJnNtagliU+2N/O9rperipIxOsP8dA7+7nn4mwijJqT0VVBEE53qdNJu306KSGFtX/6gr4ONxqdinMuTh3tlp3RRCxx+jnugQtJksKAN4CfK4ryzzd2N5CmKIpTkqQrgLeBrMNc5w7gDoDUVPGDKBw/laQaGbQYpgmqSHPEIockzIN6FEXGLDnxqXxYgyFwtgPgCwTp7PcRUoBQiJoeB6v31VG0ZPoo9EQQhKMxnMH7VNVXXFxcc6rqOotNBaoVRakFkCTpFeAqTuIsz4RIA3+6YRLNdhfnpERyTkokMzOjWL2rmT9/Wk1IAX9I4b7XS6jsGNpJJDXKyA1TUyku7cDh8ZMYruej0g4koMnu5ldv7UOvUZEVG8Yc21A8/PgH5ZQ09ZMVG0ZtVx9bazpxehUmpETybkkrzXY3z26sweUPcu8l2Zj1YgBDEM5Gjh4P/V1uXANeJFnkvTnZRCxx+jmuBVGSJGkYGrR4WVGUN//5uKIoA4qiOL/6fh2gkSQp+lDXUhTlWUVRJiuKMjkmJuZ4miUIALR6mg4q80hewn1GzH4jPimAAmiN0SRGhGGMTcGd8QNqOrro79+HV+nBHxokOuTD4XExJyf21HdCEARBgKGlqN/8pd78VdlJFR+hZ/IY68hrvWZoechv5uXxwwkJODx+fIEQwWAIrz8IgMWoxWLUkhRpoLZ7EKNWzdjYMLRqmWBIwe0L0tjr4sP97bT1u0mPNqFWSeTEh/HW7jpSwt3cONnMPyo6MWlVXFuYTFnbADvr7dR3f2+SvwuCcIJFxBjImBhLmFVPZNOr8M7PwHXa5q0UhBPueHYVkYDngTJFUZYf5px4oENRFEWSpKkMDZScspEr4eyWoE+hMOJc6gYr6Qv0EiJIhjeV3MFMtGmxSIlqKrQ1DGicTLQnEnJ4+d3HDRikMq6wOelTqjAqN9OjtnLDjLHYDrFFlSAIgnBKHOrx4gHZxU/FzM1nNtbywZdtZMaE0dTrYtAXJCvWTLfDxx8+riI6TMv60g7OSY3ApFXz9y31qGQIN6g5PyuGjgEPA54AH5d1sqPeTrLFQE68mdgwLRsqu5CRubZA4YXtDgZ9ThRFYf3+dq4tTEavUdHS5yI/OeKk9E0QhNNYZzk4Wjl3/oXkFkhEfroeWr3QWwdG65HfLwhngONZKnIucDOwT5KkL74qexBIBVAU5WlgAXCnJEkBwA0sVI51GxNBOEpaWctky0wmW2ZS0r+DBlc1U/Xno/Z4qNU3EmFNoLK3gmAgyDkNseBXSNSo2NUVg0YXzYTYGXT2hiMpoFGBEgohySc+a7MgCIJwRM1AyjdeJzOUO2uEoijPAs/C0K4iJ6MRFqOGtn43vkCQhAgjY+PCCARDVHQ4aOge5JnP6ugc8ODw+lHLMvERBhIj9DT0uthU3Y0/GOKS3Dj2twwQVBR6Bn2Utg5gd/kZ9AXQqlV80ZGCSeck0ijh9AZoH/Dy2q5mwvUaFGDu+IST0TVBEE5nH/wS3HYIi8dS8QL4nZD7I0guHO2WCcIpczy7imzi0E9AvnnOk8CTx1qHIJwoBRFTKIiYAkBjeB0l9n14u3YgI5FiyEAVG0mob5DFadHcGpeFpFUjtTnp73cD0PPRP1B/8gZJv/sd6iiRwVkQBOEU2wFkSZKUDrQAC4FTvj/1sxtrcXmDJIar6Hf7eWlrA/Hhen5xSTYvbqlndnY0b+xqYdAdwK/A+MQInr65kI/LOmm0u/jjR1VsqOwizWokM9pEitXI5LRIdjX0saGqC28gxNaaHjJiTHgDCjFmPalWIxUdTiRJ4ryxh1xtKwjCmW7cD6GnBixpkHUpBDxgr4N/PA4XPDDarROEU0I8PhbOOgn6ZGJ08chIBJUgkRoLX5gMrHcEGKzponl3Cw63n6Z1nxD98FLYvY0Bk5WQy4Xi9Y528wXhrGE0Gice6ZykpKT8tra27zQI/8c//jFKkqTCNWvWmIfLVq5cGSlJUuGLL75oAfD7/dx9991JaWlp4202W67NZst94IEH4ofPV6lUhcPlNpst98EHH4w/VF0AU6dOzRkzZsz4nJyc3EmTJtlKSkp09957b+Jdd911QG6GLVu2GDIyMvKGX2/evNkgSVLhG2+8ccA2Roe6H/fee29ibGzsBJvNlpuenp530003pQaDwe9yO75XFEUJAHcD64Ey4LXR2KXMrFMjyxLnZUVx8bg45k9KYmZmFDPHRjMhOZK/flZHVZcTRYKZGVZmZlppH/BwcW4cC6ekMHd8HC12N59VdWM1adjVYGd9aSdT061cPC4OW7wZtz/E/lYHk9MsLDk/HYtRy4JJSVySG8cFOSIHmCCclabcBnMfBa0Jsi+FC34JXeVQ+ymIyezfSsQSZ04sIQYuhLOORtYwK+pSZlov4qrIa8kLjuONXc2sbbDTGZLoVGkYcAXQ97QjuwfRdrUxmJhJ0lNPoUlMHO3mC8Jpad+GZuuLD2zKf+onnxS++MCm/H0bmk/LRbdZWVnuoqKikba9+uqr1pycHPfw66VLlya1tbVpysrK9peXl5du3bq13O/3j3xW6nS6UHl5eenw16OPPtr+bfWtXLmytqKiovTGG2/svueee1IWL17cs2bNmgPuzapVq6zz588fybD20ksvRU2aNMn5zXZ+m5/85Ccd5eXlpdXV1fvLy8sN69atMx/5Xd8/iqKsUxQlW1GUTEVRHhmNNvzmqvHMn5SMVq3i04pOZmXF8KsrxmHSqekY8ODxB1FJEtmxZlKsRp78tIZH3isDwKhV89C88cSah3a7q+4cxBMIUtY2wFt7Wrh/bg7LrytgUqqFSIOGAY+f9/a18VFZJ6s+b8SoVfHTl3ezcmv9aHRdEITR0rIb+puHvq/dAC9dDX1NcOVymPckSGfODiMiljg0EUsMEQMXwllJJakIU5txlzbh3NfEedHhTI0LJyrdSkp2FL5AiPbzfkDbrffTN/sHhAAiT8vfnYIw6vZtaLZuXl2d5ur3aQFc/T7t5tXVaScq4Fi7dq156tSpOXPnzs1IT0/PmzdvXnooFBo5/sQTT8Tm5uaOy87Ozt2zZ4/+2641bdo05549e0xer1fq7++X6+vrdXl5eS4Ah8MhFxUVxfz1r39tNBqNCoDFYgktX7689duu+V1cdNFFzoaGBl1BQYE3PDw88Mknn5iGj73zzjvWRYsW9QKEQiHWrl1rWblyZf1nn30W7nK5vnNE6vV6Ja/XK0dFRQWOt73CoT33WS1lbQPIsjT0t8JX/zqKojBljIWUKBOJkXo8gSAffNlOMKRgizfjC4TwBULoNSre+9n5rFg4kRizjjFWE+dmRuMLhrjlxR3c+red2F0+os06SlsHSLMY0aplkMA+OLRFt8t3+j0FEwThJOksh3d/Dv9349BgxfbnUBq24in+PSRNguixo93CE0bEEkd2tscSYuBCOGvVDlbQbOzAZfSRGqnnhigdprY+YsK1RHb3k+7z4k3JhK8Scla2DI5yiwXh9LRzXX1SMBA64PMkGAjJO9fVn7DtKsvKygxPPfVUU3V19f7GxkZdcXHxyDY/0dHRgdLS0rJbb72167HHHov7tutIksSsWbMG3nzzzfCioqLIuXPn9g0fKy0t1SUkJPgsFkvocO/3er3yN6d3Pvfcc5bv0v4333wzwmazuQHmz5/f+/LLL1sBPv74Y1NkZGQgPz/fC1BcXByWkpLizcvL806bNs2xevXqI24h8fTTT8fZbLbc+Pj4gvT0dM/MmTPdR3qPcGwyY0zER+hZcl4GL98+jZw4M3/+RzXrS9v5v+1NtPW50alVeAMhrpmUxEPz8siKDeOC337KBb/7lJZeF7f9fSfr97dz95yx1HYPsq2uh6gwHW39HjodHuq7ndR2OQmEFOaOTyA6TIuiKESFaXnhX6Zwx/kZo30bBEE4FQbaYPtzoATB2Q4vLwBLKnv5F96pv5WG/WfWRo0iljiysz2WEAMXwlkrN3wiYenJ9OdIeNOq0GRYUGfHQSCEsd+FTRMgUjf0IyIBRp1qdBssCKep4acj37X8WOTn5w9mZmb6VSoVeXl5rpqampFr33jjjXaAqVOnupqamnRHutZNN93U+8orr1hXr15tXbx4ce/hzluxYkXUVx/iE6qrqzVw8PTOJUuW2L+trkWLFmXYbLbcrVu3hq1YsaIJYPHixb3vvfeeJRgM8vLLL1sXLFgw0oZVq1aNvF64cGHvK6+8csQnTcPTO7u6ukpcLpf87LPPfqcASDh6/3FlLk/eOIlla0v5zzX7KS7r4P197Wyr7eWqcxJ5+KrxSBK093sZcAfYUNnF8uJKOh1eepw+nL4ALl8Au8vPjLFRZMaYCDdouPqcoVwZalnCF1CQgMvHx/OL10voHfRhNWmJCtMRY9Yhy2fOtHBBEA6hYz94ndD0ObTshPh8SDsPHO1Q8QESASSfA1n6KreFogwl6Xz3Hgh9f2dkiVji8EQsMeR4tkMVhO81szocm3kC6zpexx10kR6TRYxuaDmXKjuO5h4PIY9EmB4K0s1Em0/Y701BOKMYI7S+QwUWxgit70TVodPpRrKPqVQqAoHAyF9ver1eAVCr1co3yw9nzpw5rjvvvNOg1+tDEyZMGMm4m5ub621ra9Pa7XbZYrGEli5d2rN06dKerKysvGAweEx/La5cubJ21qxZrm+WjR071p+UlORdt26ded26dZbNmzeXAQQCAd5//31LcXFx5PLlyxMURaGvr0893J4j1aXT6ZRLL710YOPGjeY77rjjW4Mg4dgFggrNdjcKCrfEj+HS3Dguz49nbOzQ50eT3UXRtkZyEyPQa2R2N9jRqSSSo4zsrO8lKkyLXi1x7dNbMes19PW4+F1xBRF6DQbN0GwNo06NUaum2e7GGwgyZYyVW85NH+WeC4Jw0tVugOL/hNTpcPFvIBiAlKkw0AqV74O9gQnx27FZv0C7/iGouRicXUODHH4XJJ4DhYtHuxfHRMQShydiiSFi4EI4602zzKLf30e09utZYS6jgaomD05vEEmC8iYHLp9CboqJ1BjjKLZWEE4/k68Y07J5dXXaN6d4qtRyaPIVY1pGs13fZtmyZc0Gg+GAVOxmszm0cOHC7ttuuy111apVDUajUQkEAvj9/hP+iPvaa6/tve+++1JSU1O9mZmZfoA1a9aE22w216ZNm6qGz7vmmmvGFBUVRd51112HfZozLBQKsXXr1rCCggLXkc4Vjp1Bq+JPN07EPujjwbf2YdKpCTeoefjdUvKSIrgiP4E7Lxhad/5peSd9bj+SLDM3L4GH15bhC4aG8lYAUSYtGlnCrFUzJsqI1aTly5Z+pqVbuXpiIn/bUo8hJHPdlJTR7LIgCKdKeCIYoyBqLGj0EJk6NKMiZRrowkAZgEAArasVPP2wpwhCvqHdRiJSwdUDvbVg/f4tKROxxNE722IJMXAhnPWitLFEaWMPKDMbVPjlIA6vH6NGTZ8LUKDN7hMDF4LwT/JnJ/fC0PpUV79Pa4zQ+iZfMaZluPx0dN111w0cqnzFihUt99xzT6LNZsszmUwhvV4fuv7667vT0tL88PW61OHzL7zwwv4///nPRx1ULVq0yP7rX/865dFHH20aLisqKrLOmzev75vnzZ8/3/7MM8/E3nXXXb0ej0eOi4ubMHzszjvv7IChdamvvfZaVCAQkMaNG+e67777Oo+2PcLRiQvXY9KpsZq0RBi0VHY4qetx0Wh3Y3f5+J9rhv6ZJo+xsKAwGbNOzaIZaXxU2k7bgJerzknAoFGzoDAZpzfAxspu3tzTTOeAB4tRy4TkSCKNOh69Oh+H189leYfdKU8QhDNJdBbc/ObQ9w1b4W9XAgqMXwAzfw6WVNCYoHnH0CDF7r9DCAhPgkv/Gz59BDrLYP5zo9mLYyJiCRFLHImknIZ7/06ePFnZuXPnaDdDOMP1uXy8vquZObZYMmPCDji2v7Wf17Y1kxMdiQzoNGpizBrOyQgnzCDG+4TTmyRJuxRFmXw81ygpKakvKCjoPlFtEs4OJSUl0QUFBWNGux2nOo4Y8Pip/P/t3Xl8XHW9//HXd9ZksqfpkqZpkrZp06RtgK7Qy2YpVqAFKSjCFS4gotBr3X6C4r1XEa5eUa5FFBFQkP0KyI5QFkFUCsUutEm6b+matdln/f7+mHRP9zZnkryfj0cfmTln5sx7Tk/mfPOZ7/l+tzazuLqRKcP6kZbkwe9xM6BzTIpwNMbv3l/HvW+vIhKzDMpI5jNjBvGdGSXMr9jOxrpWHv7HesKRGNkBL43tEa6dWsgNZw3vtvcgIgkk2Aw1K+Gh6fHBOf1ZkF0AN74LT38Rdm6CSx+Ev/8Stn0C5VdAxUvQ0QD/MhfGXn7cEdSWECccqh2hv8Ckz3pj+XZeWLyF7U0d3HZhKRWbWmjtiJCd5uXuN1aS6fPjc7uIWdiysw0IsHjdTiaNzNrdzVdERCQ9ycuEomwmFGVT1xLk+kcW0hqMkOx1MedTxfRP8/PLt1fREowPnJfsdVGUk8LmxnbueWsVLgOv/fuZ/PzNlTy/aDM728L8ZcUOUvxeLirPJT3J6/A7FJFu9cz1UFMJU78BS5+M965wdf7Zlj0MIkHwpcK0/4B178Fffgz168Hljg/UuW0ZTP4KpOc6+jZETiQVLqTP+tToAdS0BDlv9ECstexoDNIajFLXHKYoI40Xl1czfkg/0mMRhjQ30eoxbI36+HtVA+eM6ed0fBE5iHnz5vW777779pnKbOLEiS2PPvroxpPxetOnTx++/wjkd955Z/Xs2bO77EIqvVuK38OwnBS2NXXQ1B6msS3MOaMGMC4vgw/W1RPwuXn0+sn0S/UTica4cFwuGcleUpO9fLiunrZQFLfbsGxLE3WtYdwu+PzEoU6/LRHpLi07oH4NNFbDh/fHl0XDMGRi/Pb0H8Z/vv3fsOgP8fEwQh2ABbcPQi1Q+RJ4/HDu9xx5C72B2hKJR4UL6bNyUv3cfG58ADVrLcaA123we10U908nN9mL3+NiUHOQNC8sa2qB5H5sbmgDVLgQSVS7RvDurtebP3/+mu56LUl8SV43d3/+FKyNzz4yJCsZYwyPfWkKbeEoAa+butYQT3+0iT8tqmZAWhIpfjeluekMyUqmvjWIx+UiI9nLxMJszizu7/RbEpHutPiJeAHCGIh0QFouRIPgS9v3cS4D7Q2w6UPAxC8pSc6CwjOh4oX4JSSREHg0K96xUFsi8fS5wkVzSxVNOxcxaNBncbuTnI4jCcBGY9hghPSAl1AkRsxaPG4X5xQPwsZi/HJFDWV+Q+7wwaQYQ2ayj+raDobk6PgREZGuGfJUbQMAACAASURBVGPIz94zmLPLZUj1x5tdtzy7hMqtTXjdbjY3tIOBodkpfPv8Udzy7FI6wlG+fl4x55VqUE6RPmf0zHhBoq0+3nti8Gmw8lVY8SpMuw0qX4b374acUfCZu6DyxfjsI+0N4E+DZc/Gn7fmbdj8TyiY4vQ7Ejkh+lzhYuPGB2ltXYXPP4D+OdOcjiMJIFK5ldjOdsaVDMIVSGJlfYhXl27hrVXbmZifzYaOMMt3Rrggu40hFvKykvF6DOHt2/EMGIAxJ3x2IxER6aUaWkOsq2kjGLZ8/4KR/L9nlhIDppX0p3hgGj+cNYaFG+qZOkI9LUT6pH7DYdp/gcsH/7gHti2OX/aR0tnb98P7Ycsi2LIYckbAvz4TX75tGTx3Q7xoAWCB3HFdvoRIT9SnCheRSDO1te8SjbaSnKQ50SXOJHkxrUEia2pojcRIzUzFF/uEEel5nD0wja31LQxOS+K5ZZuwMbjjwnLc782n+pGHyfrCFWRedpnTb0FERHqIrBQfcz41Ao/bcHbJQAZnJePCUJSTAsDYIRmMHZLhcEoRcdQr34xfMmJjYDsHhI90xH9OuhG2V8YvJcnv7E3RWgcPXxC/NMTlgVgkfnlJ1Ssw7vhnGBFJBH2qcLF27S+IROJTAS/48DMUFv47w4d93eFU4jTPiAEwYgDhyi1U7rS0NUbIaM+jJDsV4/MyKDuVpo4ws8qG8OKyajoiUdrS4lVvV0qKw+lFRKSnmT1+yO7bL805Mz7GksftYCIRcUxX41BEQ/GihTcAxMDth6yi+LrRF0Hx+dC4IT6LCMBLX4NwR3yMi5nz4ve9Aeg3rFvfisjJ1KfmdMzIGL/P/dbW1Q4lkUTkHT2YgQWZrG/uYGGHj0eWb+Mvq7axcFM9FhiRk0ZhdoDKbTtpGVlG4ZNPkP6ZzzgdW6TXCgQCpx7uMXl5eWO3bt16REX4e+65p58xZvwLL7ywe4SzP/zhD5nGmPG///3vswDC4TBz5szJKygoGFNSUlJaUlJSesstt+weaMDtdo/ftbykpKT0e9/73kEHIdg/28svv5x27rnxEYHvueeefi6Xa/yCBQuSd60vLi4uW7FihQ9g586driuvvLIgPz9/zIgRI8omTJgw6u23307Ztd2RI0eWlpSUlI4ZM2b0rufPnj27MC8vb2xJSUlpUVFR2be+9S3Ng9cDJPvcJHlVtBDpk7Yti/eUePen+y6fdS9cej94k6BpS3wK1At/BmvegUWPx3tVvPPf8Mx1sKMKUvrDkAlw7WswagZ8eyXMXQx547t+3T5EbYne05boUz0ufL4cjEkiFu0A42Hdx4ZVr/+di26ehD+5T+0KOYjKHTuJutxMKhrA0u07eW19LR6XYUujm8cWrmVjYxtr61qZOjqL0rx0p+OKJIzq6sez162/Ny8UqvH5fP1DRYVzNg8ZclW907n2V1xc3P7EE09kX3zxxc0ATz/9dPaoUaPad62fO3du3vbt272VlZXLA4GAbWhocP3oRz/a3aDw+/2xqqqqihORZeDAgaHbb78995VXXlm7/7qrrrqqsKCgILh+/fplbrebiooK39KlS3c3TN59992Vubm5kf2fd8cdd1Rfe+21DW1tbWbkyJFjbrjhhrqSkpLQicgrIiInWKQjfllHsHnf5S4XjJkNS5+Bla/Hp0N1J8G7/wPhdhhYBnWr489LzoKZv4BYLP68HkxtiaPXl9oSfeqv9XCHIdKeQcO6s6lbMR0bSgFa2bKqgaJxGgRLYPSgNP5auR6Py0U4GiMas4SjluZwhAtK8nh04Vo8LsO9b63G63bxqZKBh9+oSC9XXf149qrVdxbEYkEXQCi0w7dq9Z0FACeiwfHyyy+n3X777YOzs7PDK1asSB47dmzb888/v87V2UD76U9/OuD111/PiEQi5umnn1576qmndhxsW5MnT25ZsGBBajAYNB0dHWb9+vX+srKyNoDm5mbXE0880X/dunVLA4GABcjKyordfffdW473PXRl2rRpOxcsWJC2ZMkSf3l5eXDX8uXLl/sXLVqU8vzzz691u+PfxJeWloZKS0uPuNHQ1tbmAkhLS4ud8OAiInJiDJkAV/4RkjP3LGurh+dvAizUVMaLEa074JGZ4E+Fss9CICfe6yItN97bAnpF0UJtiaPXl9oSPfsIP0qRlmJqV5xJ7bJZ2FA64GaE38OGP+mSEYkbMTCNrDQPzy/bhM/jImrB6zZMHppDstdNwOumNRRhQ30bf11V63RckYSwbv29ebsaGrvEYkHXuvX35p2o16isrEz+1a9+tWn16tXLN27c6J8/f37qrnU5OTmRioqKyuuuu67mJz/5ySGricYYzjrrrKbnnnsu/YknnsicMWNG4651FRUV/tzc3FBWVtZBT9DBYNC1d/fOBx54IOtY35PL5WLu3LnbfvjDH+7TDXPx4sVJpaWlbR7Pwb9bmDZtWnFZWdnon/3sZzl7L//+978/pKSkpHTo0KHjLrnkkvq8vLwDvkkREZEEktof3N4994PN0LI9XsBI7geFU6HsEgi3QlIGTL4RAlnxAsan7+zxBYtd1JY4Nn2pLdE7jvQjlDsiA4+/BWz8wyHJQJ7XRb4BG02IQpIkgDU1LYQiMZo74r+joahl/oqtPLhgNa3hKF63ixSfm6tPL3Q2qEiCCIVqfEez/FiMHTu2dfjw4WG3201ZWVnbmjVrdm/7yiuvbACYNGlS26ZNm/yH29ZVV11V/9RTT2X/8Y9/zL7mmmsO+i3OvHnz+pWUlJQOGjRo3OrVq72wp3vnrn833HBDw9G8j/2nT77xxhvr/vnPf6ZWVVUd8b7629/+VlVRUVH5xhtvrHrggQcGvPbaa7sbXnfccUd1VVVVxdatW5e89957afPnz9cIwiIiPUlWAVz+MFzxOGQOha1LoerV+OCbW5fAa7fA0qdh2bPxy0ggXuSoeBFCbY5GPx5qSxy5vtqW6FOFi7rqFmqWXwxEAUOHNVS2R6mKWEIhFS4E6ltDLNzQQDRm91neFo4SjMSIxCyhSIz6tjC/eXeNQylFEovP17/LbocHW34s/H7/7l9Kt9tNJBLZfdZOSkqyAB6Px+69/GDOPffcthUrViTX19d7xo0bt7tbZWlpaXDr1q2+hoYGF8DcuXPrqqqqKtLS0qLRaPSw291fVlZWpLa2dveoi3V1de7s7Ox9vrXwer3MmTNn2+2337772tdTTjmlo7KyMhCNRrvcbmFhYRggLy8vcuGFFzb+4x//OKBBkZGREZs6dWrzu+++m3rgFkREJKFlFYAvFZq3QSAbxl4OxhUfqHPZs+BLiw+8OWIa1KyEF78Gf/0ZLHnS6eTHTG2JrqktsUefKly880QVRA3gBuLH7Y4oJA3PxOvrU7tCDiIj2ct5pQMoz8/A49rz2WKAgIlxbcWrXLrqHVzGMjQr4FxQkQRSVDhns8vl36f663L5Y0WFczY7lelwbr/99uof/ehH++RLS0uLXXHFFbXXX3/90La2NgMQiUQIh8NH3dAAOOOMM5ofeuihfru28/jjj/c755xzmvd/3Jw5c+ref//99Pr6eg9AWVlZcNy4ca3f/OY3B8di8d36ySef+B977LHMpqYm167GUFNTk+udd95JHzduXPv+2wyHw3z88cepI0aMCO6/TkREegBfAM7+Dlzwc/jUbXDpb+JjXLTVw4e/hYvuhvxJ8P7/Qk0VBPrB0NOdTn3M1JbomtoSe/Spv9Yz+icR/xMUIAZYMDD1smJc7j61K+QgXAauPaOIR/5tIl73ns8XCyS3NXPKjpWcsWUZwVCUD9bWORdUJIEMGXJVffGI2zb4fANCYPD5BoSKR9y2IRFHAt/lc5/7XNPMmTMPOPHPmzdv86BBg8IlJSVlo0ePLp04cWLJ5z//+dqCgoIwHHhd6k033XTQa29//OMfb12zZo1/1KhRpaWlpaXDhg0LfvWrXz3ggyMpKcl++ctf3rGrsQHw2GOPrd++fbu3oKBgzMiRI0uvv/76wvz8/FB1dbVnypQpJaNGjSo97bTTRp9//vmNl112WdOu5+26LrUzf9vVV1/duP/riYhID9BWD2/fAe/cEb8EJLMAptwUH4wzc+iex427HEpnweyHYGCpc3mPk9oSXVNbYg9jrT38o7rZhAkT7MKFC0/oNlsbO3jpp29TGHyB7aFimqIDabP9yBjcj8/dNgmX65iKYNLL3PeX1fzy7dXkZyWzdWc7TR37dq8aU7uGiD+ZjoJiLJZ3v3OuQ0lFDs4Y87G1dsLxbGPJkiXry8vLNQKtHJUlS5bklJeXFzqd42S0I0REulU0AvP/E9rr4+Nb1K+OFy4GjYtfSuJPO6kvr7aEOOFQ7Yg+MR3qyo+28e6jnxAK+Rifvp6xOe+QlD+a8KjL8E2ZoaKFAGBtlH7J7QTDUVZub8FlIMVraA3vKe4tyxnOBaPzGD0wnaVb6li+ZSdlgzMcTC0iIiIivUFocwuR2naSx+Zg3J74rCG/+zTUrQJ/Ohg3DBrjdEwRR/SJwoXH6yYUil8Ksil0Cmn+JjaY2ynKydclIrLb3yo+YEvNRlL9GezssEQt+xQtdtmys428jGSWb2vmyb9v5JozCykeeHKr3iJy5ObNm9fvvvvu22cqs4kTJ7Y8+uijG0/G602fPn34/iOQ33nnndWzZ89uOthzRERE9tc0fwPR5hDuLD/+oenQ3gCtNRAOwvm3QvnnnY7YZ6gtkXj6ROGiaFwOKZmG1sYI9aaYlVnTGLYjSNMbG0gamXXAlDLSN931Vjvbm1LwuFwUZPvYurODUPTAwsXiLQ0s3hKftej5pVtYVdvK01+Z0t1xReQg5s6dWzd37txuG4Rm/vz5mmJIRESOW2D8QMLbWvEO6pzgYd178dlEAtmwo8LZcH2M2hKJp9d3N2iqa+eFeYtobQQXlvOHv8KUy8bgsmC8LhUtZLeZpw5j4rB8UpN81DQH8Xv2/Hrsf5RkJHnI8HsAy+nD+3VrThERERHpmZre2Ujji2uw4dgB6wJjc8iYXoDL1zn75aYPwRuA0kvg9Ju7OalIYun1PS4qHn2SHavzMMbLgIwm0nNziARS8WYn4R2cEFPSSoL4t9MLecq7iZeWbtlnuYv4HDR729kRwe2C/il+vnRmUbdlFBEREZGeyVpLx4oGbDhGtDWEJzPp0E8469sw6jNQMBVcvf77ZpFD6t2/AdEwA9vfBRslNdPLJXfMhtkP4MlJJ+eGsWReNMzphJJAvv70Yu54pYLctBDXnFZLXnoIAK/HEPAd+KsSjUHMwkPvryMYiR6wXkRERERkF2MMWZcWk3XJiMMXLSB+icjOanj0Ytiy+OQHFElgvbdwUb0QVr1JrKWRNFcNWe6NuG1o92pdIiL7S/K6CHjdnF3UwqT8FibntwAQjFiCXXTnA2hoDfL68q38c0NCTG8s0qusWbPGO23atOEFBQVj8vPzx1x77bX5HR0dh/zwvvXWWwftfT8QCJwKsH79eu+MGTNUrRYREUd5BwTw5R/FoO71a6CjCZq2HP6xcgC1JXqP3lm4WP8+PDITnrqSoe4PGBt4hamZT0GbphGWg/vBrDGMzk1j/qp03liVwctVmbvXRS343Qd+xo0dksHM8sGcOjTzgHUicuxisRiXXHLJiFmzZjVu2LBh2bp165a1tra65s6dm3eo591zzz25XS0vLCwM//nPf157pK8fiUSONrKIiMgJY2OdA8Sf+S245NdQcoGzgXogtSV6l+MqXBhjZhhjVhhjVhtjbu1ivd8Y83Tn+gXGmMLjeb0jsvSP8Ld7IdwGxPC6OhgTeJ3spO3xUXlFDuK1ZVvZ2thBXbuHFyqyCEX3PV7276XjNnD/1RP46jkjSPK6uzOqSMJpbFySvW7d/WNXr757/Lp1949tbFySfTzbe+mll9L8fn9s14jeHo+H3/zmN5uefvrpnJ/85Cf9r7766qG7HnvuueeOePnll9NuuummvGAw6CopKSmdNWvWPoPPrFixwldcXFwG8YbEjTfeOGTMmDGjR44cWXrXXXflALz88stpkydPHjlz5syiUaNGlR1PfhERkWMVbQ5R++AnNLywGrzJMLDzlLTpQ9hR6Wy4k0htCTmUYx6c0xjjBn4FTAeqgY+MMS9aa/eeq+d6oMFaO8IYcwXwP8DJnYD4tVugvYuZaxo3wM7NkD74pL689FzhaIwdzR0HXd8R2fdykYDPTXuo60tIRPqSxsYl2XV1fymwNl7ti0ZbfXV1fykAyMwsrz+WbX7yySfJ5eXlbXsvy87OjuXm5oYikUiXXTx//etfb3744YcHVFVVHXLOuF/84hc5GRkZ0WXLllW2t7ebiRMnlsycObMJYOnSpSmLFi1aXlJSEjrUNkRERE4WG44R21lLrLEa6t2QXQTbK+DJL8S/iJ1wPfzL18F7BONk9BBqS8jhHM+sIpOA1dbatQDGmKeAi4G9/5MvBn7QefsZ4F5jjLHW2uN43YNr3tZ10QLAuMGXclJeVnq+yq1NzHtrFS1HWIhIT3IzqzyPof0CJzmZSOJraPggb1dDYxdro66Ghg/yjrWxYa3FGHPAuaJz+bFGBeDNN99Mr6qqCrz44otZAM3Nze6Kiookn89nx40b16qGRmIxxtwFzARCwBrgWmutBhYSkV7Lk51ETvHbmC1/h6VboL0R1rwF4XaIBuGT/4OcYhg72+moJ4zaEnI4x3PtRB6waa/71Z3LunyMtTYC7AT6dbUxY8yXjTELjTELa2pqji1RUsah1y9+/Ni2K71e/zT/UX2AtYZiLNzQcBITifQc0Wir72iWH4mxY8e2L168eJ9qc319vWvbtm2+jIyMaCy2p8gYDAaP6lxmrTU///nPN1ZVVVVUVVVVbN68+ZNLL720CSAQCKgbVeKZD4yx1o4DVgLfdTiPiMhJ5554Ga5YE/ztHvjwfqhdBS43pOXGCxgL7oNo2OmYJ4zaEnI4x1O46OqvvP0rWkfymPhCa39rrZ1grZ3Qv3//Y0vkTeagbymrEIaefmzblV4vJ9XPlKLsLg/Y/WUmu5g5Lpcfzio96blEegK3O6XLbxUOtvxIzJo1q7mjo8N177339oP4taQ33XRT/uWXX15bXFwcXL58eSAajbJ69Wrv0qVLdzdKPB6PDQaDh/xVnj59+s777ruv/67HLV261N/U1KRBkBKUtfaNzi8/AD4AhjiZR0TkpFv4e/jTlyHYDNEQ2Bhg4z3IP/V9GPd5GDkD3F6nk54wakvI4RzPzq0G8ve6PwTYf56e3Y8xxniADOCYuvocsYvu7np57Yr4bCMiB3HzuSMY3v/QlxOl+t08ccMZ/OKKU5k8LKebkokktqysKZuNce/z7YIx7lhW1pTNx7pNl8vF888/v/q5557LKigoGFNUVDTG7/fH7rnnns3Tp09vyc/PD44aNaps7ty5+aWlpbuvX73qqqtqRo8efcCAWnv7xje+UVtSUtIxduzY0cXFxWU33HBDQTgc1hzZPcN1wGtdrTghPTdFRBLBpg+hcT0MGAulF4Ovc/rUSDsUnQ3TfwBnfdvJhCec2hJyOOZYh5voLESsBKYBm4GPgCuttcv3eszNwFhr7Vc6B+e81Fr7ucNte8KECXbhwoXHlAuAZ2+IX/u1F+v2037ab/GfcyHulN5TnZQT6/EPNnDb88u6XBfwuVn0H9PxawYRSXDGmI+ttROOZxtLlixZX15efsRzSDc2LsluaPggLxpt9bndKaGsrCmbj/WaVOm5lixZklNeXl54tM8zxrwJDOpi1W3W2hc6H3MbMIF4W+KQjZfjbkeIiDjp9e/DRw9CUhqkDYbWOmiqhuwR8LXu+WxTW0KccKh2xDEPzmmtjRhj5gCvA27gd9ba5caY24GF1toXgYeAR40xq4n3tLjiWF/viNWvh2XP7rvMeIglFdKyOp1o5g7Sph5y6l7pw66cPJT/nb+C2tYDrxn83IR8FS1EDiIzs7xejQs5Vtba8w613hhzDXARMO2kDfAtIpIoxl9DcNVmWtqmkZq6AX/2WqjaAS3b4Y/XwuwH4+Nd9DJqS8ihHM+sIlhrXwVe3W/Zf+51uwO4/Hhe46i98Z9go3stcEH/EmLn3ot/QxbJJcc1HbD0csYYrplayN1vrMISH6RlWkl/Jg3L4ctnDXM6nohIn2OMmQHcApxtrW073ONFRHqkSAiaNsenPs0pJjz+TiJvfkRoQwP+zCqIRSAWhpWvw6LHYfzVTicW6Va9bwCRorP23E7qB6dcCeWfwzswk4zab+NZ8ZBz2aRHuPaMImaV55KZ7MHvdfHFMwpVtBARcc69QBow3xiz2BjzG6cDiYiccH/9Ofzx32DVfABSJgwiY8B7pESfgh3L4l/M2lh8RpGK553NKuKA4+pxkZCSUsETgEgbBBtgzKUwYhpsXACtNVC70umEkuBSk7zM+8Jp1LcEaQqGKeyX6nQkESfEYrGYcblc6pYvRyQWixnghE8BZ60dcaK3KSKScNIGgtsHyfHe4ZH6dlyeEMbjgt2dyU18JhFPEgRbwJ/wbVS1JeSIHa4d0ft6XIw4DybdAOmD4wPY5E+OLx86GT57P5x/h7P5pMfITvWraCF92bKampqMzpOIyCHFYjFTU1OTAXQ9urGIiBzahOvgS/NhyHgAWt6thOr3IdxOxOQRSRkH6XnxwkXDWmjc4HDgI6K2hByRI2lH9L4eFyk5cP7tsKMCqj+i/fVXaW0cQ8aMQrwDSpxOJyLSI0QikS9t27btwW3bto2hNxa55USLAcsikciXnA4iItJThLe3svONDQRO6U9gbH8A2qvqCW3YSmj5OvBcTlrGa9TXXg/kkVO+FFfNIjj1KhhY5mz4I6C2hByFw7Yjel/hYpdgM+3BsbStjRB1B4nUtuMdEHA6lYhIjzB+/PgdwCync4iIiPRW4R3tRBuDhKpbCIztT6wjQtP8DYS3NUI0i2h4HLGYH3e6BxNdh4k0wRWPOR37iKktISdS7ytcxGKw8jWin7qLpmcbwaSQcVER/sIMp5OJiIiIiIgAkFzWD3eaF++gFACM14V/VBaxYIRoXS2xmJ9wx2C8nnoMIahd43BiEef0vsLFRw/A2z/C5UsjI+86IsP+laSiTKdTiYiIiIiI7GZcZveXq+FtrdQ89Amx5jC42vF6GvDFPqbVdSkx9zCMJ8r2mqn4X1hN1sUas1j6nt51rdGHD8Jrt0KwGdO8haS2+aROznU6lYiIiIiIyEHZcAwb7JxQIeYnEh1ENDCKQHl/+n2xFHdGKtG6MOFtrc4GFXFI7ypcVL7ArhlUojE3dTumU/dEJdZqBh4REREREUk84W2ttFfVk3npCPAYwIU1XjqCpbRX1FH/eBU2HMM3NI2sz6q3hfRNvatwMe4LWCBmvXTEzqB95ym0L6sluKHJ6WQiIiIiIiIHaFtcQ0dVPbGWMN6BAfC7we2CkMW2RYjuDBKt7wAL3gEpTscVcUTvGuOiZRtYiMSyaInOxBAET2D3gDciIiIiIiKJJGXSINwZPgLl/UkuyaLmd8uJNnSAAVemn5RTBxBp6MCT4Xc6qohjelePi/V/JxJLoiHyHSJ2EJZkXOk+XD6308lEREREREQO4MlOIvX0wZgkD83vbya6M967AiDWEKT5/c1gwSR5sJGYs2FFHNJ7ChctNbDlY5ojVxK2I7FkAV6itR1s/u8PaF9e63RCERERERGRLgXX7aRjRQO4Xbiz/GA6V4RitC+uoeXjbURq2x3NKOKU3nOpSCAbxl9L5O1+YHf9lnf+bInQ/LctJJflOBZPRERERETkYHy5KSSX5eDJDRAoy6FlwVZCO1rp+LgGAGNceAYGHE4p4oze0+PC5Ybz/pOUgeuB6D6r3INTSJ+W70gsERERERGRw3EFvGReNIzU8YNwJXlIPzuf9DOGYNK94AKXy2CMOfyGRHqh3tPjolPK3LtwrWggtL4Rktwkj8zGl5vqdCwREREREZGj4stLJecLowlvb8M7WBMOSN/V6woXAMmjskgeleV0DBERERERkePiL8rAX5ThdAwRR/WeS0VEREREREREpNdR4UJEREREREREEpYKFyIiIiIiIiKSsFS4EBEREREREZGEpcKFiIiIiIiIiCQsFS5EREREREREJGGpcCEiIiIiIiIiCUuFCxERERERERFJWCpciIiIiIiIiEjCUuFCRERERERERBKWChciIiIiIiIikrBUuBARERERERGRhGWstU5nOIAxpgbY4NDL5wC1Dr12otO+6Zr2S9e0Xw7uZO+bAmtt/5O4fZGE5nA74mj0pM/JnpQVelbenpQVelbenpQVEiuv2hKSUBKycOEkY8xCa+0Ep3MkIu2brmm/dE375eC0b0QEetZnQU/KCj0rb0/KCj0rb0/KCj0vr0h30qUiIiIiIiIiIpKwVLgQERERERERkYSlwsWBfut0gASmfdM17Zeuab8cnPaNiEDP+izoSVmhZ+XtSVmhZ+XtSVmh5+UV6TYa40JEREREREREEpZ6XIiIiIiIiIhIwlLhopMx5nJjzHJjTMwYM2G/dd81xqw2xqwwxnzaqYxOM8b8wBiz2RizuPPfBU5ncpIxZkbnMbHaGHOr03kSiTFmvTHmk87jZKHTeZxkjPmdMWaHMWbZXsuyjTHzjTGrOn9mOZlRRLpXTzp/GGPyjTHvGGMqO9tJc53OdDjGGLcxZpEx5mWnsxyOMSbTGPOMMaaqcx+f7nSmgzHGfKPzGFhmjHnSGJPkdKa99aTz7UGy3tV5HCw1xvzJGJPpZEaRRKPCxR7LgEuB9/ZeaIwpBa4AyoAZwK+NMe7uj5cw/tdae0rnv1edDuOUzmPgV8BngFLgC53Hiuxxbudx0ten9XqY+GfH3m4F3rLWFgNvdd4XkT6gB54/IsC3rLWjgSnAzQmeF2AuUOl0iCM0D/iztbYEKCdBcxtjF5uUBAAABDlJREFU8oCvAROstWMAN/H2cSJ5mJ5zvn2YA7POB8ZYa8cBK4HvdncokUSmwkUna22ltXZFF6suBp6y1gatteuA1cCk7k0nCWgSsNpau9ZaGwKeIn6siOzDWvseUL/f4ouBRzpvPwJc0q2hRMRJPer8Ya3daq39Z+ftZuJ/WOc5m+rgjDFDgAuBB53OcjjGmHTgLOAhAGttyFrb6GyqQ/IAycYYDxAAtjicZx896XzbVVZr7RvW2kjn3Q+AId0eTCSBqXBxeHnApr3uV5PAJ+xuMKezC9vvEqW7nUN0XByaBd4wxnxsjPmy02ES0EBr7VaI/1EADHA4j4h0nx57/jDGFAKnAgucTXJIvwC+A8ScDnIEhgE1wO87L2150BiT4nSorlhrNwM/AzYCW4Gd1to3nE11RHrq+fY64DWnQ4gkkj5VuDDGvNl5Xd7+/w71TYfpYlmvnYrlMPvoPmA4cArxk9bPHQ3rrD51XByDqdba04h3hb7ZGHOW04FERBJEjzx/GGNSgWeBr1trm5zO0xVjzEXADmvtx05nOUIe4DTgPmvtqUAriXMpwz46v6y6GCgCBgMpxph/dTZV72SMuY34JVqPO51FJJF4nA7Qnay15x3D06qB/L3uDyHBusadSEe6j4wxDwAJP+jVSdSnjoujZa3d0vlzhzHmT8S7Rr936Gf1KduNMbnW2q3GmFxgh9OBRKTb9LjzhzHGS7xo8bi19jmn8xzCVGBW5+DhSUC6MeYxa22i/oFdDVRba3f1YHmGBC1cAOcB66y1NQDGmOeAM4DHHE11eD3qfGuMuQa4CJhmrU34gqZId+pTPS6O0YvAFcYYvzGmCCgGPnQ4kyM6P/B3+SzxAU37qo+AYmNMkTHGR3yAqhcdzpQQjDEpxpi0XbeB8+nbx0pXXgSu6bx9DfCCg1lEpHv1qPOHMcYQH4Oh0lp7t9N5DsVa+11r7RBrbSHx/fp2AhctsNZuAzYZY0Z1LpoGVDgY6VA2AlOMMYHOY2IaCTqQ6H56zPnWGDMDuAWYZa1tczqPSKLpUz0uDsUY81ngl0B/4BVjzGJr7aettcuNMf9H/EQSAW621kadzOqgnxpjTiHepXU9cKOzcZxjrY0YY+YArxMfWft31trlDsdKFAOBP8XbNXiAJ6y1f3Y2knOMMU8C5wA5xphq4L+AnwD/Z4y5nnhj8HLnEopId+qB54+pwBeBT4wxizuXfa8vzyx2gv078HhnEWstcK3DebpkrV1gjHkG+Cfx9vAi4LfOptpXTzrfHiTrdwE/ML+zDfWBtfYrjoUUSTBGvZBEREREREREJFHpUhERERERERERSVgqXIiIiIiIiIhIwlLhQkREREREREQSlgoXIiIiIiIiIpKwVLgQERERERERkYSlwoWIiIiIiIiIJCwVLkREREREREQkYalwISIiIiIiIiIJ6/8Dygiklbs3NjgAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1080x720 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "embed = wrapper.fetch_cell_embedding(final_dim=256,\n",
    "\t                                     auto_correct=True,\n",
    "\t                                     restore_order=False)\n",
    "\n",
    "embed = embed['prefer']\n",
    "\n",
    "from umap import UMAP\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "\n",
    "cell_type = model.label_info['cluster label']\n",
    "batch = model.label_info['batch id']\n",
    "neuron_subtype = model.label_info['major']\n",
    "\n",
    "\n",
    "pal1 = {\"L2/3\": \"#e51f4e\", \"L4\": \"#45af4b\", \"L5\": \"#ffe011\", \"L6\": \"#0081cc\",\n",
    "\t\t\t       \"Ndnf\": \"#ff7f35\", \"Vip\": \"#951eb7\", \"Pvalb\": \"#4febee\",\n",
    "\t\t\t       \"Sst\": \"#ed37d9\", \"Astro\": \"#d1f33c\", \"ODC\": \"#f9bdbb\",\n",
    "\t\t\t       \"OPC\": \"#067d81\", \"MG\": \"#e4bcfc\", \"MP\": \"#ab6c1e\",\n",
    "\t\t\t       \"Endo\": \"#780100\"}\n",
    "\n",
    "vec = UMAP(n_components=2, n_neighbors=25, random_state=0).fit_transform(embedding)\n",
    "fig = plt.figure(figsize=(15, 10))\n",
    "ax = plt.subplot(2, 2, 1)\n",
    "sns.scatterplot(x=vec[:, 0], y=vec[:, 1], hue=cell_type, ax=ax, s=5, alpha=0.8, linewidth=0, palette=pal1)\n",
    "handles, labels = ax.get_legend_handles_labels()\n",
    "labels, handles = zip(*sorted(zip(labels, handles), key=lambda t: t[0]))\n",
    "ax.legend(handles=handles, labels=labels, bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0., ncol=1)\n",
    "ax = plt.subplot(2, 2, 2)\n",
    "sns.scatterplot(x=vec[:, 0], y=vec[:, 1], hue=batch, ax=ax, s=5, alpha=0.8, linewidth=0)\n",
    "handles, labels = ax.get_legend_handles_labels()\n",
    "labels, handles = zip(*sorted(zip(labels, handles), key=lambda t: t[0]))\n",
    "ax.legend(handles=handles, labels=labels, bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0., ncol=1)\n",
    "\n",
    "\n",
    "neuron_mask = np.isin(cell_type, np.asarray(['L2/3', 'L4', 'L5', 'L6', 'Ndnf', \"Vip\", 'Pvalb', 'Sst']))\n",
    "pal1 = list(sns.color_palette(\"tab20\"))\n",
    "pal2 = list(sns.color_palette(\"Set2\"))\n",
    "pal3 = list(sns.color_palette(\"husl\", 12))\n",
    "pal = pal1 + pal2 + pal3 + pal1\n",
    "# pal = pal1 + pal3 + pal2\n",
    "pal_all = pal1 + pal2 + pal3 + pal1 + pal2 + pal3\n",
    "pal_all = pal_all[:len(np.unique(neuron_subtype[neuron_mask]))]        \n",
    "\n",
    "\n",
    "ax = plt.subplot(2, 2, 3)\n",
    "sns.scatterplot(x=vec[neuron_mask, 0], y=vec[neuron_mask, 1], \n",
    "                hue=neuron_subtype[neuron_mask], ax=ax, s=5, alpha=0.8, linewidth=0, palette=pal_all)\n",
    "handles, labels = ax.get_legend_handles_labels()\n",
    "labels, handles = zip(*sorted(zip(labels, handles), key=lambda t: t[0]))\n",
    "ax.legend(handles=handles, labels=labels, bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0., ncol=1)\n",
    "\n",
    "ax = plt.subplot(2, 2, 4)\n",
    "vec2 = UMAP(n_components=2, n_neighbors=25, random_state=0).fit_transform(embedding[neuron_mask])\n",
    "sns.scatterplot(x=vec2[:, 0], y=vec2[:, 1], \n",
    "                hue=neuron_subtype[neuron_mask], ax=ax, s=5, alpha=0.8, linewidth=0, palette=pal_all)\n",
    "handles, labels = ax.get_legend_handles_labels()\n",
    "labels, handles = zip(*sorted(zip(labels, handles), key=lambda t: t[0]))\n",
    "ax.legend(handles=handles, labels=labels, bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0., ncol=1)\n",
    "plt.tight_layout()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    " "
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}