demo_part1.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Demo: Part 1\n",
    "\n",
    "This notebook is part 1 of a 3-part demonstration of the model, tested on a Windows 11 machine.\n",
    "\n",
    "### Instructions:\n",
    "1. **Setup**: Follow the instructions in [README.md](README.md) (also given [below](#setup-details)) to:\n",
    "   - Set up the required [Conda environments](README.md#setup).\n",
    "   - Download all necessary [data and models](README.md#download-models-and-data).\n",
    "   - Prepare codebase of the [submodule](README.md#prepare-codebase)  \n",
    "\n",
    "2. **Environment**: Use the `retrosub` conda virtual environment for this notebook. Refer to the [README.md](README.md) file for detailed instructions on setting up the required environments.  \n",
    "\n",
    "3. **Next Step**: After completing this notebook, proceed to [./RetrievalModel/demo_part2.ipynb](RetrievalModel/demo_part2.ipynb).  "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Setup Details\n",
    "\n",
    "1. **Setup Conda environments**\n",
    "\n",
    "- clone repo. If you are in windows, best if you clone in the system drive (C:)\n",
    "```bash\n",
    "        git clone https://github.com/MDFahimAnjum/RetroSub_Research\n",
    "```\n",
    "\n",
    "- fix typo, and change some codes to run the submodules with recent pytorch version. In windows, open git bash and run following. If you try the following in cmd or powershell, you will get errors due to path notations\n",
    "```bash\n",
    "        bash scripts/step0_fix_submodule.sh \n",
    "```\n",
    "- conda environment for reaction retrieval. In windows, open anaconda prompt in the root directory\n",
    "```bash\n",
    "        conda create -n retrieval python=3.6\n",
    "        conda activate retrieval\n",
    "        #conda run -n retrieval pip install -r RetrievalModel/requirements.txt -f https://download.pytorch.org/whl/torch_stable.html\n",
    "        pip install torch sacrebleu transformers==2.11.0 jsonlines regex scikit-learn scipy\n",
    "        conda install -c pytorch faiss-cpu\n",
    "```\n",
    "- conda environment for substructure extraction, seq2seq model inference, ranking model training. for windows, open anaconda prompt and cd to the root folder\n",
    "```bash\n",
    "        conda create -n retrosub python=3.7 -y\n",
    "        conda activate retrosub\n",
    "        conda install -c pytorch pytorch torchvision -y #Alternatively: pip install torch torchvision\n",
    "        pip install rdkit-pypi tqdm func-timeout future six pandas gputil notebook\n",
    "        cd MolecularTransformer\n",
    "        conda run -n retrosub pip install torchtext==0.3.1 \n",
    "        conda run -n retrosub pip install -e .\n",
    "```\n",
    "- conda environment for model training (requires python 3.5)\n",
    "```bash\n",
    "        cd MolecularTransformer\n",
    "        conda create -n mol_transformer python=3.5 -y\n",
    "        conda activate mol_transformer\n",
    "        conda install -c pytorch pytorch torchvision -y\n",
    "        pip install future six tqdm pandas\n",
    "        conda run -n mol_transformer pip install torchtext==0.3.1\n",
    "        conda run -n mol_transformer pip install -e . \n",
    "```\n",
    "\n",
    "2. **Download models and data** \n",
    "- Download processed data, models and results [here](https://figshare.com/ndownloader/files/41144306).\n",
    "- Extract in root folder\n",
    "\n",
    "3. **Prepare Codebase**\n",
    "\n",
    "- Change code of submodule (reaction retrieval) to run on CPU in the code folder.\n",
    "```bash\n",
    "        bash scripts/step0_fix_demo.sh \n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "import operator\n",
    "import pickle\n",
    "import random\n",
    "from collections import Counter\n",
    "\n",
    "from rdkit import Chem\n",
    "from rdkit.Chem import Draw\n",
    "\n",
    "from demo import (extract_all, load_retrieved_candidates,\n",
    "                  load_sub_seq2seq_output, merge,\n",
    "                  merge_predicted_frag_with_substructures,\n",
    "                  prepaire_sub_seq2seq_input, rerank_candidates,\n",
    "                  save_input_for_dual_encoder, save_input_for_subseq2seq,\n",
    "                  view_sub_extractions)\n",
    "from utils.smiles_utils import canonicalize_smiles, smi_tokenizer\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# this file is a dictionary, key: target (reactants), value: source (all possible products)\n",
    "# it is generated on the train and val data by 'data_utils/collect_reaction.py', and is used\n",
    "# to filter substructures extracted from candidate reactants (target side) but not exists in\n",
    "# the product molecules (source side). It means that the substructures shall also remain\n",
    "# unchanged during the reaction from source (product of retrieved candidate reactants) to \n",
    "# target (retrieved candidate reactants). \n",
    "# load only for the first time\n",
    "try:\n",
    "    all_reaction_t2s\n",
    "except NameError:\n",
    "    all_reaction_t2s = pickle.load(open('data/uspto_full/reaction.pkl', 'rb'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "input_product_smiles: COC(=O)Cc1cc2ccc(F)cc2c(-c2ccc(S(C)(=O)=O)cn2)c1C\n",
      "golden_reactants_smiles: COC(=O)Cc1cc2ccc(F)cc2c(B2OC(C)(C)C(C)(C)O2)c1C.CS(=O)(=O)c1ccc(Br)nc1\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlgAAAEsCAIAAACQX1rBAAAABmJLR0QA/wD/AP+gvaeTAAAgAElEQVR4nOzdZ1yT19sH8F8SNgjIkuXCxagTq1XcYquirbYqiEXUWtTaP07E0RZb24qtVay1fXDvgXXhbN1blOFgCyqCspEdQsZ5XhyMSBUUktzBnO+nL5qY3NeVVrhyzn3OdXiEEDAMwzCMpuJznQDDMAzDcIkVQoZhGEajsULIMAzDaDRWCBmGYRiNxgohwzAMo9FYIWQYhmE0GiuEDMMwjEZjhZBhGIbRaKwQMgzDMBqNFUKGYRhGo7FCyDAMw2g0La4TYBiGYZSsoAAbNiAmBhIJBg3C1KnQ0eE6JzXCY023GYZh3mVFRejRA336wMcHQiFWrICREY4eBY/HdWbqgo0IGYZh3mmhobC1xaZNVQ/79YODA86dw+DBnKalRtg9QoZhmHdadDSGD3/x0NAQAwciOpq7hNQOK4QMwzAveHvjk08gv2U0YQL+/ZfThBouPx+mpi89Y2aG3FyOslFHrBAyDMO8kJmJK1ewdeuLh2VlXOajAC1bIiPjpWceP0arVtwko5ZYIWQYhnnJwoVYvBj5+VznoSiDBmHnTpSXVz188ADnzmHQIE5zUi9ssQzDMMxLXFwwdiwWLHixvmTMGFy7Vse7unWbHx29u5YXyGQyACtWrPD19VVMom/IywtHj+L99+HpifJybN2KpUvh6KjSHNQbK4QMwzA1LVsGJydcvVr1MC8PmZl1vKWoqDCzzhcBU6dO/eyzz4yMjBqc4xvj87FnDy5dwq1bMDXF2bNwcVFd9MaAFUKGYZiaTEzwyy+YPRuGhgBw4ABEojrewuf/JpP9UMsLxGJx9+7d8/LyVq1a9d133yku2bo8eoRff0Xr1pg/HwBmzYJEgt9/h0CguhzUG9tQzzAM88LAgQgIwPDhIAQDB+LGDezZg9GjFXPxa9eu9enTR09PLyEhoWXLloq5aJ0iIvDBB+jRAxERAKCjA7EYIhFrLiPHRoQMwzCvwOPhzz/RpQsA/PAD4uJqe7Gp6fzCwvRaXrBhwwZjY+PevXuPHTs2LCxs8eLFu3btUmi+TP2xESHDMEyVmzdhbg4bGxgYVD3z5AlMTeHhgYsXa3tjmzYuqanxtbwgOzvbysoKQHp6uqOjo1AovHjxYt++fRWWei3YiLAubETIMAwDAM+ewcMD+vq4ceNFIbSzA4CgIOTk1PZebe3fxOKiWl5gYmJC/6V58+bz58//4YcfZs2aFRkZyeezPWzcYyNChmEYAJg1C7//joEDce6ccgMJhUInJ6e0tLRNmzZNmTJFucHARoR1Y19GGIZhkJCAv/6CQICQEKXH0tfX//nnnwEsWrSoqKi2cSSjGqwQMgzDYO5ciMWYNg2dOqki3Pjx4/v27ZuTk7N8+XJVxGNqxQohwzCaLjwcp06haVN8/72KIvJ4vJCQED6fv3r16uTkZBVFZV6DFUKGYTRaZSUCAgBg6VJYWKgubrdu3Xx9fSsrKxcsWKDUQBIrrbKv+5eNqeqpVjazX9nX/SFgp/K+wBbLMAyj0X79FQsWwMkJd+5AW1ulobOzszt06FBUVHTq1KmPPvpISVHKyiISEz8wNOzh6BgBIDpahxBxt24iHo8tlqnCRoQMw2iunBz89BMArFql6ioIoFmzZosWLQIwZ84csVis6vDMc6wQMgyjuRYvRlERRo7E0KHcJDBnzpx27dolJCSEhoZykwHDCiHDMBorOvp2cnKKri5WruQsBx0dnV9++QXAd999l5eXx1kemo0VQoZhNBEhZPZs/8uX2y1cuL59ey4zGTVq1EcfffTs2bPvVbZolXkZK4QMw2iiffv2Xb582crKas4cT65zwapVq7S1tf/666979+5xnYsmUpdCePny5T179qSn19a+nWEYRiGEQuHChQsB/PTTT/IuoBxydnaeNm2aVCqdPXs217loInUphJs2bfL29j558iTXiTAM8+775Zdf0tLSunbtOnnyZK5zqfLDDz+Ym5ufO3fuyJEjXOeicdSlEMbHxwNwdnbmOhGGYd5xGRkZv/76K4CQkBCB2pzS3rRp06CgIABz584ViUQKvLKWloWZmbexcdW6WDOz8WZm3oC6fHB1oBYb6gkhpqamxcXFubm5Fqps7cAwDNeEJYXR/+wCoGdkYtXSsblTd2VH9Pb23rNnj6en5969e5Ud661IpdKuXbveu3cvODg4MDBQ0RcvLi7+VyYr0dN7z9CwO8A6y7ygFiPCjIyM4uJiS0tLVgUZRtOUFxdc3LcaQElBdvjvc89s/VGp4a5fv7537159ff3g4GClBqoHgUAQEhIC4KeffsrMzFTglcXip3FxToWFRyoqkp4+XSyR5Cvw4u8AtSiECQkJYPOiDKOptHX03MbMHOA9b+T/Vt45fwDA0/t3KivKHt29mpeRosBAMpls1qxZhJAFCxa0atVKgVdWlEGDBn3yySclJSVLlixR4GWfPTtgYNC5desddnbB7dqd1tJiQ46XqMUJ9bQQOjk5cZ0IwzCcqRSWJkX826yVE4B/N39PZLImZtYufUeKyksyEqPqfPuTMq30nDrO9ouIiLh165adnV0AbbOtllatWnXq1Klt27Y1bdq0RYsWtb/Y05PP58tqf02TJoMEAuOysqiysluGhu/TJ0tKzgmFdWzVKCvTDw8X1pmwr6+vqalpnS9TZ6wQMgzDMWFp0e9f9pJUVlSUl4wJ+Is+2cXdq+sQLwAX9666sKvu1i/p2h02HbpQ58vMzc0///xzQ0NDACkpKQEBAd9//30n1RxC+Hp79+69fPkyXTjq4ODQp0+fxMTEVatW1fnGnj11tbXrWFnTosVfFhZfVlQk3L//oa6ug7X14qZNPyso2JeXt772N8pkLrNnx9WZw9ChQ1khVAC6ZJQVQobRTPpGJv4brgPIeZSwKeDjGevOATAwNqN/at++W8+RX9R5EZtyA8PmddSz6OjoK1euHD16dNmyZdra2uvWrTt8+HBRUdG5c+ca/CHqr6ysLCAgICMjo0ePHr6+vllZWbdu3SouLh49enTz5s1rf6+1tUAgkNb+Gn39TjyewM4u2MZmaWHh32lpk7W1LY2NB/P5enUlZuLvP7jO/Js2bVrna9QdUQN0jUxGRgbXiTAMo2p5Gam/eLvQf5fJZCt9OqcnRG5ZODrxxj8KjyUSidq3bw9gzZo1hJCCggL6y+fvv/9WeKw398033wDo1q2bVColhNCtjZ988omSwt2/Pzw7e42SLt5Icb9YJj8/Py8vz9jY2NbWlutcGIbhgERcGX/laOylw0fXztMzNLZ2eE9JgXR0dFauXAlg6dKleXl5TZs2/eGHHwAEBARUVFQoKWjt0tPTV61axePx1qxZw+fzY2Jitm3bJu/ErShFRcfy83cIhfeePTtQWnrNyMhNgRd/B3BfCBPj4wE4OjryeGxfC8NoHP0mpu97+D5NvZv1MM6mbaepK49p6eh27D/a3La1MsKNHDly6NChz549o7vX/fz8OnXq9PDhwze5IacM8+bNKy8v9/b27tOnDyFk1qxZMpls9uzZ7RXaCFxb27qs7EZGRsCzZ/scHPYaGLgq8OLvADXYUL9+PVm0KN/X14Kjv4gMw2iUhISEzp07y2Sy6OjoTp06nT9/ftCgQUZGRklJSSqel7p69Wrfvn319PQSExNbtGixe/fuCRMmWFlZJScnK7ADqlRaWFp6RSBoSgeCRUUnAJmJiQfbUy/H/YgQ8fG8ggILa2uu82AYRiM4OTnNmDFD3uF64MCBn376aWlpKT0sXmXkmxoXLVrUokULoVC4ePFiAMuXL1dsH/CKiqSUlJEZGXPpw9TUUSkpIwkRKzBEY6cGhTAhAQDYklGGYVRl6dKlFhYW58+fP3ToEICVK1fq6ent2LHjypUrKsth48aNUVFRzZs3nzdvHoDg4GDaB3zSpEkqy4GhWCFkGEbjNG3alJ6CO3/+/IqKitatW8+dO5cQMnv2bJmsjv3pClFcXExvUv72228GBgYZGRkrV66UL5lRQQJMdVz/Fy8tRUYG9PTQWik3xhmGYV5p2rRpnTp1evDgwerVqwEsWrTI1tY2Kipq165dKoj+/fffZ2Vlubm5jRkzBsD8+fPLy8u9vLz69u2rguhMDVwXwoQEEIL27aE2h6EwDKMJBAIBLYE///zz06dPjYyMli9fDmDBggXFxcVKDZ2SkrJu3To+nx8SEsLj8a5duxYWFqavr08TYFRPDQoh2LwowzAcGDRo0OjRo0tLS+kqFR8fn549e2ZlZSl2D99/zZo1SyQSTZ06tXv37vIlM4GBgS1btlRqXOZ1WCFkGEZz0WUy27dvv3nzJo/Ho0O0lStXpqQo8tSL6s6cOXPixIkmTZrQm5SbNm2KjIy0t7efP3++kiIydWKFkGEYzeXg4DB79my6k50Q8sEHH3z++ecikWjhwoXKCCeRSOiejaCgIGtr65KSku+++w7AypUraR9whhNcF8L4eIAVQoZhOLNkyRJbW9sbN27QZTK//PKLsbHxgQMHTp8+rfBYa9eujYuLa9u27ddffw3ghx9+yMrK6t2797hx4xQeS04gFFgmdzRJq+rfbXH/PcvkjiBsN301HPY5JSIR0dIiAgERCrlMg2EYzbZlyxYAdnZ2paWlhJAff/wRgIuLi1gsVmCU/Px8MzMzAMeOHSOEpKSk6Orq8vn8mzdvKjDKK9y4QQDSo0fVQ21tAhCRSLlBGxVOR4QiERYtwvTp0KvjNBCGYRjl8fX17dGjx5MnT1asWAFg/vz5bdu2jYuL27hxowKjLF68uKCgwN3d3cPDA8Ds2bNFItGUKVPef/99BUZh6oG7XqM3byIqClIpunaFG2uFzjAMl65fv+7m5qarq5uQkNCqVau///577NixZmZmycnJ5ubmDb9+XFxcly5dANy+fdvFxeXMmTNDhgxp0qRJUlKSjY1Nw69fm4gIfPABevRARAQA6OhALIZIBB0d5cZtPDgaEc6cifHjkZuLwkJMmQJfX3De+5thGA3Wq1cvb2/vioqKBQsWABgzZsyQIUMKCgroNGnDzZkzRyKRfP311y4uLhKJZM6cOQC+++47pVdB5k1wMB177hwxMyP5+VUPi4qItTUJD+cgE4ZhmOcyMjLo0s0LFy4QQmJjY7W0tLS0tO7du9fAK+/fvx+AmZlZXl4eISQkJARAmzZtKioqFJB3ndg9wrpwMSK8dAmjRsHMrOqhsTHGjcPFixxkwjAM85ydnV1gYCCA2bNnS6VSFxeXL774wsbGJjc3t4FXLikpMTEx+fHHH83NzQsKCpYtWwZg9erVurq6CsibaTAuCmF6Omoc+mVnh/R0DjJhGIapZv78+a1atbp9+/amTZsArFixIjExceDAgQ287OTJk5OTk/38/AB88803+fn5gwcPHjlypAIyZhSBi0JobY28vJeeyc0FmyhnapBKsXEjfH0xahS++QYZGVwnxLz79PX16cLRb775prCw0MTExMDAQCFXtrKyEggEcXFxGzZs0NLSorOjjJrgohB2744TJyB+fiykVIqjR8EWEDM1TJ6MjRvh7Y1FiyAS4f33WS1kVGDcuHH9+/fPzc319PRU+MVHjx4tkUhmzpz53nvvKfzir2VgACenFyf8ODvDyQk8tqH+BS62T0ilGDoUPB5mzQKfjz//RGEhzp8Hj4cbN9hWCgYA7t1D7954+BAWFlXPjB8PGxusWsVpWoxGOH369IcffgiAz+dra2sr6rKVlZWEED6fn5WVZWlpqajLvimJBCkp0NVFq1asCtbAxYhQIMCxYxg1Ctu3Y/NmDBmCf/+FlhYWLULfvli4kJutFEeP4n//w7hx+OYb3L/PQQJMdbdv4/33X1RBAMOGISaGu4QYDTJkyJDPP/+cx+PJZDKR4hBCtLW1v/32Ww6q4Nq1sLHBV1/Bywv29jh1StUJqDctbsLq6uKrr/DVVy89aW4OHg8rVuDhQ2zZAgVNzb+R4GCsX48ff0SLFjhzBh98gPPn0amT6hJgasjPh4nJS8+YmdW8tcwwSrNjx47NmzdLpVLFXpbux1DsNet28SKCgnD9Ojp0AICjRzFuHJKTYW2t6kzUFUeF8JUCA9G5M8aPR1gYkpJw+DBatVJF3KIi/PADrl1Dly4A0KcPKiqwdCkOHlRFdOaVWrbE48cvPZOWpqK/DwwDANDW1lbgvCiXwsIwaVJVFQQwciS6dMHx4/jiC07TUiNcnz5Rw9ChuHkTTk64cwfvv4/z55UVqLAQ//d/8PcHgPh4WFhUVUFqxAhERSkrNPMm3NyQmIgbN6oeikTYuBEffcRpTgzTOKWmvqiClJMTUlM5ykYdqVkhBNCuHW7cwMcfIy8PH36I339X5MVlMly5gmnTYG+PGTPwxx9ITUVeHpo2fellFhZo8BZapkGsrPB//wcPD/j7IygIH3wAe3tMn851WgzTCBkYQCh86RmhUKX3ntSe+hVCAMbGOHgQgYGQSDBrFqZNQ2VlQ6/58CGWLoWDA/r2xfr1EAoxZAh27oSdHVq0QEbGSyt0Hj9Gy5YNjcg0kI8PYmLQqROaNsXq1Th6FBkZWLoUDx5wnRnDNCpOTlXttimZDBERcHHhLiH1w3GLt9rt2EH09AjwePTonJyc+lxBKCRhYWTECCIQEIAAxN6eBAaS1NQXrxGJiLU12b+/6qFMRj77jHz9tQLyZxSivJzQloy+vgQg337LdUIM06g8fEhMTcmWLaSykpSWkvnziaMjUU2b00aCu2OY3lBMjNDLq69AkF1ScujQoe7du7/h+27cuOH899/GGzeiqAgADAzw2WeYPBkDBlTtoSEEFy9i82ZMmYKKCnh7w9sbLVvizBlkZeHcOSji7BWmoX77DcuWYd06TJiAixcxYADs7fHoEQQCrjNjmMYjMhILFyImBlpaGDQIK1fCzo7rnNSIWk6NVte1a8mFC9omJhkZGf379w8LC6v95VlZWWvWrOnSpUuvXr1uRUWhqAiurggJQXo6tm/HwIHg8fDkCVasQLt2GDgQO3Zg40YMHYrYWLi4QCbDl1/i1i1WBdWFkRGKirBlCwD064e2bZGRgXPnuE6LYRqV7t1x5gzy85GdjT17WBWsiesh6RupqKiYMmUKAB6PFxgYKJVKa7ygsrLy0KFDI0eOlO/RsbGxWbd0KYmPr36VmtOkdnYkMJCkpKj0wzBvpaiIGBgQHq9qNnvZMgKQ8eO5TothmHdH4yiEVGhoKN3WM2zYsGfPntEn4+PjAwMDmzVrRuufQCBwd3cPCwuj3YyqxMaSwEBiYVFV/3R1ydixJDycSCTcfBLmrXz+OQHI0qWEEJKeTgQCoqdHCgq4TothmHeE2t8jfNmlS5fGjBmTm5vbtm1bb2/vkydP3rp1i/5Rx44dp0yZMmHCBHn7opycnLhDhwauXYu4uKr39+iByZPh5QVTU07yZ+rj3DkMHoxWrZCaCj4fH32Ef//FunU1OxMxDMPUSyMrhABSU1NHjRoVGxtLH5qYmHh6evr4+PTp04c+I5VKz58/v379+sOHD+vyeMVGRjwAY8ZgxoyXds0zjQUhaNcOqak4exaDBmHvXowfj+7d8fw7EMMwTEOo/WKZ/2jTps21a9dGjRoFgM/nT548OTQ0lFbB+Pj4gIAAe3v7IUOG7N+/H8DgYcOe7NyJrCyEhrIq2FjxePDxAVC1ZGbUKFhYSJrIRAWx3ObFMMy7ofGNCClCyJw5c9auXWtmZpaamnr48OEdO3acPXuWfhxHR8dJkyZNmjRJfu+QadzS0uDgAF1dPH0KU9OnSfMyS1c1azbH3p6dysQwTEOpU9Ptt8Hj8bp06SKTyUpKSqytrYVCIQBTU1MvL6/Jkyf36NGD6wQZhWrZUhQ0JdchUk+8zwLTTJt7Zyasys/fbmcXzOPpcJ0cwzCNW+ObGpUrKSkBQA/6cnNzCw0NzcjI+Ouvv1gVfCeVfT0g2+l2Xv5WAAYGrvr6nSWS/KKi44oMUVaWy3rMMozmacSFMCEhAcDQoUMfPHhw5coVPz8/Q0NDrpNilMXU9DOBwLSs7EZFRTwAC4tJAPLytijk4lFRUbNmzWrRokVQUJBCLsgwTCPSiAthfHw8gNmzZ7dkDbI1AJ+vZ2bmBSAvbysAMzMfHk+3uPikWJxZ72tmZmauWLHCycmpe/fuv//+e0FBweMahyAyDKMBGnEhpCNCZ2dnrhPRCFFRUdOmTTM2NrawsAgPD+ckB3PzyQAKCrYTItbSMjcx8SBEUlCw822vU1lZefTo0XHjxrVs2XLhwoWJiYk2Njb+/v537949duyYEhJnGEatNdZVowUFBebm5kZGRsXFxTzaRJtRgidPnmzfvn3r1q3JycnyJ/X19ffu3fvxxx+rPp/4+M5C4d02bY6Ymn5cVHQsJWWknl4HF5cE4I3+DsTHx2/fvn3Lli05OTkAdHR0Pvzww4kTJ44ePVrenI9hGE3TWEeEdF7U0dGRVUFlkI+ZWrVqtXjx4uTkZFtbW39//8uXLw8aNEgoFI4aNWrhwoUymUzFiZmbTwSQn78FgLHxUG1tG7E4p7KyjvlMiaQgJ+ePO3f69OzZY8WKFTk5OZ07d16zZs2TJ0+OHj06duxYVgUZRpM11p9/9ZwX7dWrl1QqPXDgQPPmzbnOpZ7i4uJ27NixefNmun5SV1d36NCh1cdMZ8+eXbNmzbx581asWPHgwYMtW7aoco2SubnPkyeLioqOicVZ2trW7dqd0tVtz+frveblsuLi03l5W4qKjshkFQAmThwqELSbPHly165dVZYzwyiWWCwODAzs2rWrD200wTRY4y6ETk5OXCfygkQiiYmJEYvFZmZmXOfy1vLzsXs3btw4uXv3cPpM165dJ0+ePGHCBPnHKSoqMjExATBr1qwOHTqMHz9+//79SUlJhw8fbt26tWry1NKysrScrqVlRbcP8vlNysquaWvb6Om99DehsjK9oGB3Xl6oSPQQAMA3NnY3M/NZu3YMn2+gmlQZRkmOHTu2evVqQ0NDV1dXdRsMNFZcdvxugKFDhwI4fPgw14m8QGtz69atuU7kLUil5PRp4uNDDAwIQFq1EjVtaubn5xcZGSl/jUQiOX369NixY/X19TMzM+XPJycn0y8i5ubmtKePiqWnB9y92/LBA6/ExN65uaGEEKlUWFAQlpzsHhnJi4xEZCTu3m2RkREoEj1UfXoMozy+vr4AXFxcSktLuc7lXdBYCyHdMpGUlMR1Ii8cPHgQwPDhw7lOhBBChEJSXPzSw7Kyl16QlEQWLSJ2dlUnUwkEZPhwsn8/EYnE8tfcvXt39uzZ8tM8dHR0anzzKC4u/uSTTwBoaWkFBwcr9yO9TCzOj4zkicXZ8mfy83fExDSh9S862ujhw0klJZcIkakyK4ZRjfLy8s6dOwMYr/yzOSMjI1etWqXsKNxqlIWwtLSUx+Pp6OiIxeK6X60qP/74I4B58+ZxnQghhCxdSnR1SWJi1cMffiD/+x8hhJSXk7Aw4u5OeLyqEti+PQkKIo8evXhvYWFhaGiom5ubfNrA2dk5ODg4Ozv7v4FkMllQUBBdsuTn5ycSiVTw6QghYnFuZCT/2bOD8mdKS29GRiI+3jU3N1QiKa7lvQzzDkhKSmrSpAmAzZs3Ky/KhQsXjI2NARw5ckR5UTjXKAthZGQkgPfee4/rRF4yYcIEABs3buQ6EUIIWbqUNG9OPvqo6iEthN9+S4yMqupfkybkiy/IlSsv3iKVkrNn8728vPT0qtaeNG3adObMmdWnSV9n7969BgYGANzc3LKyspTzmWrKyVkXE2MSH98tNzeUjvwqKh6oJjTDqIPdu3cD0NPTi4mJUcb1jx07pq+vD8DLy+ulo87fOY1y+wTdO6FWK2WgfgtZx4xBTg727XvxjJ4eSkvh6orQUDx9io0bQUd96elYsQJt22LYMLPTp29XVla6u7tv27YtIyPjjz/+cHV1rTOWp6fn1atXW7ZsefXq1e7du99S2kmBQuGd9PRZmZk/AbC0/KpTpyxr64VZWSuePv0egK6uitbsMIw6GD9+/JQpUyoqKsaNG1dcXKzYi+/Zs2f06NFCoXD69Om7du3S1tZW7PXVC9eVuD4WLVoEICgoiOtEXpDJZEZGRgAKCgq4zoUQQpYuJfPnkytXiK0tKSysGhHm5ZGUlBevKS0lW7eS/v1fTJO2bUvWro1IT0+vX9Dc3NwBAwYA0NPT2759u2I+CSGEELE4Pzv79/j4rvQW4J07zWSyF7PieXmbExN7KzAcwzQWQqGQ7gXy9PRU4GX/+usvPp8PIDAwUCZ792+0N8oRoRrunUhLSystLbW2tm7atCnXubzg5oYPP8R331U9NDdHmzYAEBWFadNgY4NJk3DxInR1MXYsTp9GcjK+/rqHvb19/cJZWFj8888/U6dOraio8PX1VcSOe1lx8ZlHjybeu9c8Pd2/vDxGIDC1sPBr2/a4TFZaULBLIskVi7MLCw/r63duWCCGaZT09PT27dtnbGy8b9++DRs2KOSaK1asmDFjBiFkxYoVwcHBGtG0hOtKXB8dOnQAcOfOHa4TeeHEiRMABg0axHUiVeiIkBCSk0OsrIi3d9VimaQk0qFD1fgPIH36kE2bXlpfqhChoaF0ImXYsGHPnj2rxxVEorTMzOC7d1vRIWBkJD852T0vb5tUWk5fUFmZ8ejRlLi492JjndLSZkgkhYQQsThfKmWryRmNs2/fPgB6enrR0dENuY5MJgsICAAgEAjWr1+vqPTUX+MrhCKRSEtLi8/nl5eXc53LC7/99huAmTNncp1IFXkhJIT89Rfh86sKoUhELCyIjQ3x9yd37yoxgUuXLllZWQFo165dfHz8G76rpKRk69YtiYmD5RsBY2PbZ2b+XFmZUed7MzN/jorSzcn5q2GJM0yjNG3aNABt27YtKiqq3xUkEsmXX34JQEdHZ9++fbtDoWEAACAASURBVIpNT801vqnR+/fvSyQSBwcHupxJTajbbK2+PuT/efz80KdP1UMdHVy6hPR0rFmDjh2VmEDfvn0jIyNdXV3v37//wQcfHD16tPbX09MtbG1tJ02anJVVzufrNW06tl270y4uidbWi7S17V75LpEo9cmTJWVl1wHo6LQiRETbkDKMplmzZk23bt1SUlJoMXtblZWV3t7eGzZsMDAwOHLkyLhx4xSeoVrjuhK/tf379wMYOXIk14m8hO6646TByuvk55OkJMLtTsuSkpLPPvsMgEAgeOWO+4yMjJ9//rldu3b0byOPx+vbt++ZM7ul0pI3uf6TJ0GRkXj40IcQIpUKY2KaRkaivFyN5swZRmXu379PmyD++eefb/XGsrIy2qvL1NT08uXLSkpPnTW+Qvj9998DWLBgAdeJvMTc3BxA9Q5knNu2jQDEy4vjNGQyWXBwMF2B5uXlVVZWRggRiUTh4eHVj32wtbUNDAxMTk5+q4uLRA8jI/nR0foSyTNCSFrajMhIpKerRU8DhlG9sLAwALq6um+y/Zd69uwZ/R7frFkzJe1HVH+NrxB6eXkB2LJlC9eJvJCVlUW/THGdyEsWLiQAWbqU6zwIIYQcO3aMfld1dnaeNGkS/d5Af2LHjRt38uRJiURSvysnJQ2KjERu7nryvLnM7dtWMtm7vPmXYWoxc+ZMAG3atCksLKzzxVlZWV26dAHQokULtepYqWKN7x6huu1bh/pu8AcANUnKw8MjIiKibdu2jx492rp1a35+Pm3blp6evm/fvqFDhwoEgvpd2cJiMoC8vC0ADA3f19fvJJHkFBWdUGT2DNN4/Pbbb66urqmpqRMnTiS1nruelpbWr1+/27dvOzo6XrlypX379ipLUt00skIok8noUel0B4WaUMPaDCA+HlCbQgigQ4cOy5YtKy8vd3BwuH37dlxcXGBgoLyjd72Zmn4mEJiWlV2vqEjAyyf3MowG0tXVPXDggJmZWXh4+B9//PG6lyUmJvbt2zc5Oblbt26XLl1qvEeoKkQjK4QPHz4UCoV2dnZ0nk1NqNuSUQCVlXj4EAIBni9D4cxPP/20fPny7OxsABkZGQA8PDxo43yF4PP1mzYdByA/fysAc3MfHk+7qPCkVJitqBAM07i0bNly69atPB5v3rx5169f/+8LoqKi+vXrl56e3q9fv3PnzjX8+2hj18gK4bVr16BmJQdqWQiTkiCRwMEBeq87vF1V1q5du3jxYpFIBKX9h6Kzo/n52wgRa2lZtb83o6unjeDPnYqNwjCNyMiRI/39/cVisZeXV35+fvU/unTp0qBBg3Jzcz08PE6dOqWMQcWff/45ceLE0tJShV9ZSRpZIaQNFNLS0rhO5CVqWAjpDULOJ2ufPXuWnZ1tZGREJ16U9B/K0PADQ60uzS46yK6dBWDUdDAvNQ1b2Owoo9F+/fXX3r17P378eNKkSfKbhcePHx86dGhxcbGXl9ehQ4eUsRs7KSkpICBgx44dZ8+eVfjFlaSRFcK+ffsCSElJ2aI2v+aKi4ufPn1qYGBAzwpWE2qyUoYuI3J0dKTtChMTE6Gcbwwdzkxo5n9d8Nt6ABg+HDY2iIvDzZsKD8QwjYW2tvbevXvNzc2PHTsWEhIClRwoIZFIfH19y8vLJ06cSE/tbhy4Xrb6dkpKShwdHQHweLxFixZJpVKuMyJ0Cr5r165cJ/IST08CkG3bOE6DdgH28fEhhDx9+hRA06ZNlRIpK4toaxNtbZKTQwgh8+cTgEyfrpRYDNN4HDt2jMfjaWtrz58/X36ghPLCffPNNwDs7e3V5ByeN9TIRoRGRkYJCQmhoaFaWlrLly8fMWJEYWEhtymp4bwo1GZEWP0/Dh0dKmttbbNmGDoUYjF27gSAqVMBYM8elJcrJRzDNBIeHh70ZuHKlSsJIb/88ktwcLCSYl2/fn358uV8Pn/Hjh1qdQ5PnRpZIaT8/PzOnj1rZWV18uTJHj160N+2XFHDQiiVIjkZPB4432NSfYel0jeZTJ4MAJs3A0CHDujZE0VFOHRIWeEYpjE4evTo4cOHAejq6hobGw8aNEhJgcrKynx9faVS6YIFC+i5pI1IoyyEePuezsqj3IFOvTx8+KhTp49HjvzF2JjjTKoXP6V/Yxg5EtbWiI1FVBTwvC6qzb1khlGx+/fvDxs27OOPP05LS+vSpYurq2tRUVH//v2V9Nty1qxZ9+/f79q1K+2C2chwPTfbIHX2dFYBBwcHAG9+0pAKHDlyBMCHH37IbRplZWV8Pl9HR0csFhNCBg4cCODkyZNKDDl3LgEIPQyrqIgYGBAej6SmKjEiw6if8vLyoKAgPT09AKampiEhIRKJRCwW06OaBALB2rVrFRtRPui8q9TT3ZSmcRdC8pqeziojFAoFAoG2trZIJFJl3NrRewCzZ8/mNo3IyEgALi4u9KG1tTWAR48eKTFkbCwBiIkJoWdVrl5NDh4klazvKKNBwsPDW7duDYDH4/n4+GRnZ1f/05CQELqE29/fX1GLDbOzs5s1awZgzZo1Crmg6jX6QkjJezp37do1LS1NNUHj4uImT57M5/MtLS0rKipUE/RN+Pr6AggNDeU2jR07dgAYO3YsIeTZs2cADAwMlL7Q9/33iZMTSUxUbhSGUT8pKSkeHh50qq9Lly5Xr1595cv27dtHB4ujR49u+MhBJpONGDECgLu7u0wma+DVuPKOFEJCSGJiIm1AamlpeeHCBeUFKioq2rZtm7u7O/1iZWhoCKB3797qcwZTjx49AFy6dInbNBYvXgzgu+++I4RcvXoVgKurq9Kj5uURQkhZGZk7l3TpQuzsyODBJDxc6XEZhkPl5dKgoHatWwNo2rTpunXraj/O5cqVKxYWFgB69uxZY8j4tv788086Afv48eOGXIdb704hJITk5+e7u7vTqepNmzYp9uJSqfTy5ct+fn608gEwNjb28fEJDQ1t1aoVAFtb24iICMUGrQeZTEYHx7m5udxmMnr0aAB79uwhhGzcuBHA559/rqLYo0aRESNIRgaprCTHjxMzM/LvvyoKzTAqduQIad2aAH/36jV58uQ3LGxxcXH0F5eDg0NCQkL9IqekpBgZGQHYu3dv/a6gJt6pQkgIkUgkgYGBtFD5+flVKuL+UEZGRnBwcJs2behl+Xy+m5tbaGhoaWkpfUFeXh5dCaKrq7t169aGR2yI9PR0OizmNg1CCG19cPv2bULIvHnzAPz000+qCJyaSvT1SfXvAcuXk2HDVBGaYVQpNZWMGEEAApDOncmVK2/17szMzPfffx+AmZnZxYsX3za4WCzu2bMnAHreU6P2rhVCaufOnbSHXr9+/eo98K+oqAgLCxsxYoT8qDx7e/vAwMDUV61CFIvF/v7+8gJM10ly4t9//6UfnKsEqMrKSm1tbT6fX15eTggZNmwYgIMHD6oi9rFjpFOnl565eJG0aqWK0AyjGuXlJCiI6OkRgJiakpAQUq+jrUtLSz/++GP6JX7Xrl1v9d5vv/0WjbCJzCu9m4WQEHLt2jUbGxs68L93795bvTcyMtLf319+irqent7YsWPDw8P/O+0ukUi+/PLLmJgY+jA0NFRHRwfARx99xNVfDtpUcNq0aZxEl4uLiwPQpk0b+pBOwiSqZg3L33+T7t1feiYigjRrporQDKMCp08TR0cCEB6P+PiQrKyGXEwikfzvf/+jq0yDgoLe8F03b96k33TPnz/fkOhq4p0thISQJ0+e0GUjRkZGBw4cqPP1+fn5oaGhXbp0kW+ydHV1DQkJyaPrL15l7dq1dDHkvn376DNXrlyhK4nbtm0bFxensA/zxuhWoZCQEPqwqKjo//7v/1S/muvvv/8GMGLECPJ8Q6G2trZCZqrrFhNDzMxe2jWxdStxc1NFaIZRoPJyEhVFLlwghYVVz6SnEx+fF3Ohly8rKlRISAjdhDZlypQ6f05LS0vbtWsHJbctVaV3uRASQioqKiZNmkS/7AQGBr5y7b5EIjl9+vTYsWPlvdjNzMz8/Pzoza06r//FF1/UuH56enr37t0BNGnS5PDhw4r/VLWiB3T8+3xtCJ33GD9+PJ2iVJlly5YBCAgIIIRERUVV31CodFIpcXQky5dXPczPJx07kj/+UFF0hlGIzZuJpSUZNox4eBALC7JuHSGEeHtX7ZRds4Yo+v7LoUOHDAwMAAwZMqSoqKiWV9Jfel26dFGr/dMN8Y4XQoo26aYDlOr/gxMSEgIDA+kAjjZccHd3DwsLe9uBS2hoKC2iw4YNKywsJIQIhUIfHx/5bIMKBmQymUy+qFVPT08+3S/fYdmlSxflbmZ/mbe3N4DNmzcTQnbu3AlgzJgxKotO4uKIoyNxdSUff0ysrcnMmfW7g8Iw3IiNJQYG5Pk9FxIbSwwNSWQkSU8nU6cSpe3UioiIoL8PO3bs+LrtEI29icwraUQhJIScOnWKdkPv2LHjnTt36EZA+RSok5NTcHBwVgOm2i9dumRlZQWgffv2dC1y9ZY3np6eymt58+jRo6VLl9JeErT00qK+atUq+gL5DksLCwuVTejTGeYbN24QQpYsWQLg22+/VU3oKlIpuXePXLzYwDsoDMOB5cvJuHEvPTNpEvnmGxVEfvDgAV3vbWtrGx0dXeNPc3JyaKWU33x5N2hKISSEJCUl0f/B8o2ApqamM2bMuHnzpkKu//jx427dutH9heHPd3CfOHFCSQOy1y1qvX//vrwAy2dE8/PzhwwZAkBLS+v3339XYBqvJJVK6RwLHR9/+umnAHbv3q3suAzzjvjyS7J48UvP/Pwz8fZWTfCCgoL+/fvT1RXHjx+v/kfvQBOZV9KgQkgIKSwstLW1NTEx6dy5c/WNgIryyibg8gKsqAHZmyxqPXr0qLGxMYBu3brRnnPK2GH5Og8ePKDfKOlD+vHla2sZdVZYWBgQEDCMbbvk1ty5ZNasl55ZvFiVB02LRKIJEybQr85//fUXffKvv/5C428i80qaVQiJ8ls/V58RlQ/IioqKRo4c2cABWUFBwVstar179y6dL7Wxsbl+/Tp9cteuXXSHZZ8+fRoyFVy7Y8eO0a+NpNqGQhX3Q2fqRywW00n+t910xCjSpk2kRj/CPn1UvOBLJpMFBQXRXzX+/v7JycnvRhOZV9KsQqiy1s/yAZm8CXi9B2QNWdSal5c3ePBgemebrlshhERFRTVv3pxOpUZGRtb7M9YiNjY2MDBw/fr1hBB6XqODg4MyAjHKQHfgLFmyhOtENFhJCWndmixcSPLySEEBWbqUNG/+YhOFCm3evJn+5jEzMwPg4+Oj+hxUQLMKYY3Wz05OToMGDSopKVFGrLt379KjCqsPyHbv3v3mAzKFLGoVi8X/LcBPnjyhvZEMDQ33799f78/4JuiGQg8PD6VGYRTo/Pnz9LvLO3YfqJF5/Jj4+JAWLUjz5mT8ePLwIVeJnD592sDAwMDAQCAQcLIpWQU0qxBWb/2cmZlJ57uVFy4/P18+IJM3AY+Ojm7RokUtA7Iap1sAcHR0bOCi1h07dtCDV/r375+Tk0MIqaiomDx5MmrdYakQAwYMAPDxxx8r6fqMwkmlUjs7OwCKWkfGNIhYTCIiyOvbeiibTCZr2bKl/I5Mz549LytuI7+a0KxCWL3187lz5wD07t1bqRFfOSB7+vTpBx98UGNA9t/TLUxMTHx8fE6fPq2QTK5evUrvj7Zp0yY2NpY+Kd9hOXz48EKFTrzQRa20CgKYVePOP6PeZs+eDWDu3LlcJ8IQ4ulJALJlC1fxT5w4AaBFixZbtmyhfSvpCgD5r5F3gGYVwuHDh+N56+d169YB+OKLL1QQVz4g69evn3xANmXKFDogmz59+nfffUe7cQLg8/nu7u67du1SeC+YjIwM2mzeyMhI3v/6n3/+oTssO3ToUO/TWKqLiIiYPn26qamp/OP06tWLwy7kTD3cuHGDrvtV+kHKTJ1WrSIAmTqVq/j01yZdBl9WVhYcHEwXQGhra/v5+SlvzZ0qaVYhpMWG/rr/+uuvAaxcuVI1oa9duyYfkMnX461evVpLS6tJkya0ZtRyuoWiCIXCiRMn1pgRTUlJcXFxoffD/63vuX2vW9RKCz/T6LRt2xZAPU7nUbqiIjJvHnn/fdKhA/nsM6IGh4AqV0QEAYiTEyfBU1JS+Hy+rq5u9WN8cnNz/f396WSSkZFRUFCQijs4KpwGFcIarZ/p3bsa20WVqnoPUnlcOi50dnY+d+6cyu5Ch4SE0G34Y8eOpZspS0pK6Dm61XdAvonXLWpluwYbu0WLFgGYMWMG14m8TCYj/fuTcePIw4ekqIhs2EBMTMg7NEf3CpWVxMCA8HiEi++Uc+bMATBlypT//lFCQsLYsWPlX+JDQ0Mb7/yBBhXC6OhoWnLoQzrZ/VC1a7GEQqGvr6+enh7tPUYIoS1JN2zYoMo0CCEnT56ks5edOnV68OABeXkHpLe3d51f8eiiVjrMrfeiVkZt3blzB4CFhYV6TWtfu0YsLUn1PakzZpAvv+QuIZUYMIAA5Hm/KpUpKyujuyZu3br1utecOXOma9eu8kmgRnoqkwYVwuqtn1W2ofCVqt9kpmPEK295tLRCJCcnOzk5ATA3Nz979ix9MiwszNDQ0Nra+smTJ698lzIWtTLqydnZGcCZf/7hOpFqNmwgH3740jO7dhElL3nj3pIlBCAqP/OItpJxq+sIM6lUGhYWJl/l4O7u3uj6cWtQIaze+vnatWsAunbtym1KMpmM3iCs5chDpSouLv7kk08AaGlpyWdEY2Ji5ANWOWUvamXU0LWQkKJOncirpsU4s2YNqbEV5/Bh0rkzR9moyokTBCB9+qg4bKdOnQDs2bPnTV5M19HQ1spaWlp+fn6ZSjslQ+E0qBBWb/28adMmABMmTOA2pbS0NADNOD08nTZSosM7Pz+//x4wlp6eHhwc3KZNG/kqUDc3N2V0amXUTkoKAYixMREKuU7luaNHay4b+eUX8umnHGWjKoWFRCAgurqq/B9BN5jZ2Ni81aGDeXl58nU0hoaGgYGBSupYolgaVAjpNCBdxDF//nwAP/74I7cpnTp1CsCAAQO4TYMQsnfvXnpehJubG53kFAqFrzzdQqmLWhm1060bAcihQ1zn8VxxMTE3J/J2SDk5pFUr8mZDlkYt38Mjon//xP9M1SjPJC8vAEFBQfV4b2JionwdjZ2dXWhoqES9DwTVlEJYo/Wzh4cHgAMHDnCb1erVq9VnYV5kZKS9vT0AXV1dBwcHulWI3kn18fE5d+5c410SxtTfL78QgHh5cZlDcTGZM4fIBxbnzhEbGzJkCPH0JDY2ZP58LnNTlenTpwP45ZdfVBTv8WOZicmdgQOfPn1a72tcu3atV69e9NeIs7OzKpfovy1NKYQ1Wj/TLqDx8fHcZuXn5wdABQcEvqHc3Fza/q36RkCu7l8yauHxY8LjEQMDwtUEl1hMhg8nwEsH1QqF5PJlcuoUec2SrnfPjh07AHzyyScqirdoEQEafgKiTCYLCwuTHxvu7u5+584dhSSoWJpSCA8cOIDnrZ/Ly8urbyjkUJ8+fQCo1XqT4uLiOXPmDB069NSpU1znwqgHNzcCcDb9OHMmAYiFBUlOVnaosLCw8ePHC9Xnhmg19IxPCwsLVew2rqggVlYEINeuKeR65eXly5cvp+toBALBb7/9ppDLKpCmFMJly5YBmD9/PiEkJiYGgBNHnRqqs7CwAPC6jQoMoxZ+/50ARGVjkeqCgwlA9PTI1auEEFJWRgYNIidOKCkabUDI+R2T16Gd0JOSkpQeacsWAhBFL6rPz88PDAzU1dXlZLdY7fjQDAkJCQDoehk6TUr/nUO5ubl5eXnGxsbyPrYMo448PaGlhZMnUVCg0rj792PxYvD52LkTvXtDKsW4cTh3DgEBkEqVEdDT0xPAvn37lHHxhqP32+hZcsr1558A4O+v2KuamZkFBwc/fPjQzc1NsVduOM0qhHSDcPV/5xCtx87OzvKd6QyjjqysMGAAKitx5Ijqgt68iUmTIJPh11/x2WcAMHs2jh+HuTkOHsTzlcyKNX78eD6fHx4eXlxcrIzrNxCtH0ovhNev49YtWFjAy0sZl1fP7/0aUQhlMllycjKADh064OXRIYfUJA2GqZunJwCobKj04AFGjkR5Ofz8MHcuAPz6K/74A3p6CA9H+/ZKCmtra+vm5lZRUXH06FElhWgIuqRA6YXwjz8AYOpU6OkpN5A60YhCmJaWVlZWZmtrS88bUpMKpCZpMEzdPv0UOjq4dg1lZUqPlZ+PYcOQk4Phw7FuHQD8/TcWLgSPh02b0Lu3UoN7eXlBXWdHu3TpYmRklJSUlJubq6wYOTk4cAACAaZNU1YItaQRhbB6yRGLxfRgkfZK+15Zj6wYRq2ZmeGff/D0KZ432FOSCqFQ9umnSE6GqyvCwqClhVu34OtbNUfq7a3U6ADGjh2rpaV16tSpAhXfEH0DWlpaq1atOnr0qPzgNsX7v/+DSISPP8bzxqEaQuMKYWpqamVlZcuWLQ2V/CNdJzVZs8Mwb8TUFN7ecHCAtTXGjcOjRwqPQAj5YurUBSUlUkdHHD4MQ0M8eIARI1Beji+/xLx5Co/4X5aWloMGDRKLxYcOHVJBuLdy6dKlsLCwkJCQiRMnrl69WiQSKTiARIL16wHg668VfGW1p3GFUE3GYcXFxU+fPtXT02ulYd+8mEYpPR2DBmH4cMTH4/59tG2LgQMVPk0aGBi4e/fuTQ8fJv/9N+ztX8yRDhtWtY5RJejs6N69e1UW8Q1lZ2cXFRUFBwd/9dVXBw8eXLp0qYID5OWhWze89x4GDlTwldWeRhTC6mMvNRmHJSYmEkI6dOggUM76N4ZRpC1bMGAApk+Hnh6aNMFPP8HEBAodM23YsOHXX3/V1tbev3+/k4tLRUVFwVdfITkZ3bpVzZGqyujRo3V1dc+fP5+dna2yoG+oSZMmrq6uAwYM8PHxuXXrlmIuSghSU3HmDEpKcPgwbt2C5q1jV91fLw7dvn0bz5eM+vn59ezZk/MlvGpSjxnmjSQkoGfPFw95PPTsiYQERV3+5MmTX331FY/H27Bhg7u7u0wm8/X1vXD8eNTw4fYbN8LISFGB3oSpqenQoUOPHDmyf//+r9VskjAnJ2f//v1lZWV//PHHwoULARw8eJAeIGppaWlpaamtrf22V8Qnn6CsDJ07IzYWPB7Cw2Fvr5Ts1di7XwiTkpKEQiGA0tJSAJaWlu7u7lwnpS4ztAzzRioroaPz0jO6uqiowLZtWLQIrq5V//TogWbN3vba0dHR48aNk0gkS5cu9fX1BbBo0aKwsDBjY+OC5cvtufjO6unpeeTIkX379qlbISwoKDhz5kxFRYVEIqmoqAAwceLEsmpz1Obm5lZWVpaWls2aNWvWrNlv1tY65uawtoalJSwtYWODGgtt/vc/tGuHrVvB54MQTJ8OPz+cOKHiz8U9rlvbKJ1UKqW7JmxsbNSn3+vIkSMB7JefJsMw6iwggPj5vfSMuzv5808yezYBXvqnRQvyySfk++9JeDjJyKjzwhkZGfTMk/Hjx9Mumhs2bACgra3977//KunT1KmsrMzQ0JDH4z18+JCrHP4rLCxs0KBB9N+TkpJ0dHQKCwsnTJjg7u7esWNHa2vrGvdZdHR0ZDxezf9BenqkeXPy/vvk44+JVEr09UlCwosYjx8TPl+Njp9UlXd/RMjn88+ePTtu3LiUlJTevXtv27btM9qoglNsRMg0JmPHYvBg+PvDxQUAzpxBRAS2bYOVFSZPRnQ0YmIQHY07d/D4MR4/ftGDxspqwfDhWjY23bp169q1q4ODQ/U+SsXFxcOHD8/IyOjfv/+WLVt4PN7JkydnzJjB4/HWr18/ZMgQLj4qABgYGIwYMWLfvn379+8PCAjgKo1aiEQiHo+nr6+/c+dO+ZOEkLy8vNzc3Nzc3MzMzLJnz3iZmcjORnY2cnORk4OsLJSWIj0d6elo1gxZWRAK8fxoCACwt4euLh49gqMjB5+KQ1xXYhWpqKiYNGkSAB6PFxgYyO3RekKhUCAQaGlpVVRUcJgGw7yF0FBiaUnc3UnfvsTGhhw79uqXPXlCwsNJUBAZMYJYWcl4PJPnB1sCMDY2dnNz8/f337Zt240bNwYPHgzA2dm5oKCAEHLv3j16QEH9DoNVLLp9wtXVletECCFELBaLxeKwsDBbW1s/P7/PP//c1tb2119/rc+1ystJWhq5cYOcP09KSghA8vNf/KlIRPh80oAzCBspHiGEmwrMhfXr18+cOVMikXh4eOzatYv+1Kne3bt3O3fu3KFDh8TERE4SYJj6KCvDvXvQ1kbHjjVvGb6GNC0tPDo6JiYmOjo6Jibm6dOnNV6gra3dp0+fPXv2SCSSXr16paene3l57d69m/MGvCKRyNraurCwMCkpifPmGzNnzrx///6ff/756NEjANra2u3bt1fMir/WrbF6NUaNqnp45gy8vZGdrXELR7muxKr2zz//0FuGtA5xkgPdojRq1ChOojMMVwoKCi5fvhwSEuLj4+Ps7Mzn8wEIBAI7O7vOnTsD6Nevn/pMk9A5pB9++IHbNIKDgwHo6+tHRUUp/urr15NWrci5c6S0lFy5Qtq1I6tXKz6K2tO4QkgIuX//vouLCwAzMzNODsUNCgoCsGjRItWHZhj18ezZs+pdPZ2cnOgcqZo4efIkAEdHRw5zCAsL4/P5fD5fiack7tpFevcmrVuTDz4gmzcrK4p608RCSAgpLi4eNWoU/TYaHBys4uiffvopgO3bt6s4LsOooWHDhgEwNjZWqyWahBCxWGxlZQWAq9Xmly9f1tPTA7Bq1SpOEtAcGtFZ5r+aNGly8ODB4OBgmUy2cOHCzz//nO41VCqZTHbmzJmJEyeeOHHC1NT03r17yo7IMOpv4sSJANq2VUC0lQAAIABJREFUbatu7Qa1tLToCnNODqNITU399NNPKyoqpk2bNmfOHNUnoFm4rsQc27dvH+2+3atXr6dKWyt1//79JUuW2D/v10BvjQCYMWNGZWWlkoIyTKNQVlZmZGSkbpv2qAsXLgBwcHCgexxVJi8vr127dgA8PDwkEokqQ2smTS+EhJDbt2/Tr6K2trY3btxQ4JWFQmFYWJi7u7t8CVyLFi0CAwMfPny4Z88eAwMDAG5ubllZWQoMyjCNDu1zvWLFCq4TqUkqldrZ2QGIiIhQWVChUNi7d28Arq6upaWlKouryVghJISQ3NzcgQMHAtDV1d2yZUvDLxgZGenn5yc/NkxfX3/s2LGnT5+u/r0yOjq6RYsWAOzt7W/dutXwoAzTSB0+fBhAt27duE7kFei05Jw5c1QTTiaTeXt7A2jVqlVmZqZqgjKsEFYRi8X+/v60bvn5+YnF4npc5OnTpyEhIR07dpTPPLu6uoaGhhYXF1d/mUQiodfPycnp378/AD09vR07dijmkzBMYyMSiczMzADEx8dznUtNERERAGxsbFQzRTlv3jwAxsbGd+/eVUE4hmKF8CWhoaE6OjoAPvzwwzdfyS0SicLDw+nZ1rT+2draBgYGJicnv/L18+fP79evX3Z2NiGkoqLiiy++gHq0vGEYrkyePBnA999/z3Uir9C2bVsAFy5cUHag9evXA9DW1uZkW5cmY4WwpitXrjRr1gxA27ZtY2Nja39xbGxsYGCgpaUlrX+6urojRowICwurZUCZn59PW0I4ODjcu3ePPhkaGkrPTxk+fHhhYaEiPw/DNAb//PMPuN609zqLFy+mS9uUGuX48eNaWlo8Hm/btm1KDcT8FyuEr5Cent69e3cATZo0OXTo0H9f8OzZs9DQUDc3N/kUqLOzc3BwcE5Ozptc/8mTJz179gRgZGQk3yd78eJFummpffv2CdX7wTOMBpBIJPQL6O3bt7nOpaa7d+8CMDQ0XLhwYVhYWEpKisIXkUZFRRkZGantmPidxwrhqwmFws8//1w+Y0n/3kul0tOnT/v4+NAFnwBMTU39/Pzq0frolU3AU1NT33vvPQBmZmYcnkHDMJz46quvoJYdl9asWcPj8eQ/9Xi5e3hkZKRIJGrI9eVnUXl7e6t4nwZDsUL4WjKZLDg4mO758/DwCAgIoAupAQgEguHDh+/fv7+BPwChoaH0tuKIESOKiooIISUlJbTvDCctbxiGQxcvXgTQsmVLtSoGx44dEwgEPB5v1qxZ3377rYeHh62tbY3d2AYGBr169frqq682bNgQFRX1Vr8WioqKOnXqBGDAgAEN/H3C1BsrhHUIDw9v0qQJnbWg85ZBQUGPHj1S1PVPnTpFm4B37NjxwYMH5OUCPH78+PLyckXFYhh1JpPJ6IYixW7nbYjIyEjacOOnn36q/nyN7uE1zsrQ0tJydnb28fEJCQm5fPlyLXsBKysr6bGLzs7Oz549U/4HYl6NFcK6/fTTT3TtzOXLl5Vx/fv37zs7OwMwNzc/c+YMffLo0aPGxsZ0c1VaWpoy4jKMupk7dy6A2bNnc50IIYQ8fPiQ3rb84osv6DMSiUT4qtPbi4qKqtfFGifF090XI0aMCAoKCg8Pl68kkMlk9P6IjY2NAr9bM/XACmHdFixYACWfxiJvAq6lpSWfEb17927r1q3pz8n169eVF51h1MTNmzdVuWmvFoWFhfSG/cCBA+Uzljdv3pSP9oKDg8PDw/Py8v773tLS0qtXr/7xxx9Tpkzp2rUrXRBenYODw5gxY+i5xEZGRko5X4l5G6wQ1m3EiBEA9u/fTwjJzMx8w6Whb0smkwUFBdE5FtoEnBCSl5dHf1p0dXU3a+oJKYxGoT02z58/z2EOlZWV9OfOxcWl+ozl3r175XuFKT6f3759e09PzxUrVpw+ffqVdVEsFsfGxm7bts3f39/NzY3OtdKldrq6uq9cl86oGCuEdWvTpg2AuLg4QgjtPrNaaWdX7tu3jy5O6927N22wJBaLAwMD6U+On58fa9LNvNuWLFkCYPr06VwlIJPJ6IEYtra2/70rUVlZWb2q6evrv24WNCws7JUbkSUSSWxs7Pbt2y0sLABcvXpVJR+LqQ0rhHUQCoUCgUBLS4senO3u7g7g6NGjyosobwJuZ2d38+ZN+uSOHTvoyWT9+/dX0pCUYdRBfHw8AAsLC66+833zzTcAmjRpEhMTU+eLq4/23N3daaO46kxNTeUbLWJjY6u3jgoICADwv//9T5mfhnkjrBDW4fbt26jW8ILuoEhNTVVqUHkTcD09PXmbiatXr1pbWwNo06ZNnS1vGKbxojfnTpw4ofrQmzdvBiAQCMLDw+vxdqlUmpiYuHv37oCAgMGDB7+yLk6bNo2+OCoqCoC1tTXnN0QZVgjrsGfPHgCjR48mhBQVFfF4PD09PRX8xa3eBNzf359GTE9Pd3V1pT+oS5cuVXYODMOJZcuWAfD19VVx3PPnz9NWw+vWrVPUNZ88eRIeHh4UFDRixAgHBwcAU6dOlf9p+/btAcjXijNcYYWwDt999x2AxYsXE0Ju3LgBoEuXLiqLLm8CPnToUHrTXigU9ujRg07dsA7dzDspJSWFx+MZGxurchNtXFycqakplNzaJjMzs/p9R/rr5csvv1ReROZN8MHUKiEhAYCTk1ONf1cNPz+/f//919LS8tSpUwMGDJBKpXp6eqGhoQDKysrkJ90zzLukTZs2rq6uxcXFp06dUk3EzMzMYcOGFRYWjhs37scff1ReIGtra9o0gBo/fjyAv//+u7KyUnlBmTqx36R1oLfu6YZ31RdCAP3794+Oju7evbu/vz/dqJubmwuAjgsZ5p1Ez6zfu3evCmKVlpZ6eHg8fvy4T58+27ZtU+X3S0dHx06dOj179uz06dMqC8r8FyuEtZFIJHSWhk7lc1IIAdjb21+7dm3KlCn0Ia3NnTt3VnEaDKMynp6efD7/2LFjpaWlSg0klUonTJgQExPTpk2bgwcP0rXZqqTKks+8DiuEtUlNTRWJRC1atKC9RmkFUn0hBFC9OQVX9ZhhVMbe3t7Nza28vPz3339XaqBZs2aFh4dbWFicPHlSfrCoKnl5efF4vCNHjgiFQtVHZyhWCGtDSw6dF62oqHj06JGWlhY9rlpNsmKYdxXdTRsUFOTj47N///6UlBRCiGJDrFixYt26dXp6ekeOHKEdbVSvdevWPXr0KCkpOX78OCcJMGCFsHbVx17JyclSqbRNmza6urrqkxXDvKuWLFliamoqkUh27tw5bty4du3amZiY9OnTZ9asWevXr79y5YpIJGrI9ffv37948WI+n79z587evXsrKu168PT0BJsd5ZRW3S/RYNVLTvVVMxzKz8/Pzc01MjKSH47IMO+kDh063L59m66Rjo2NjY6OfvLkydWrV69evUpfoK+v36lTp65du3br1q1bt27vvffem39JvXnz5qRJk2Qy2W+//fbZZ58p6zO8GS8vr4CAgOPHjxcVFZmYmHCbjGZihbA21W8Kqsk4TF6PaxyBxjDvnpYtW/7888/yh4WFhbGxsVHPJSYmRkRERERE0D/V0tJq3769q6urq6uri4tLt27d/tvYhXrw4MHIkSPLy8v9/PzowU/csrGx6du374ULF8LD/7+9Ow9r4s7/AP4J4Uq4QzkFuSwop4AWFFQUdAsEhVWEFnGtFVp3K2q1tbtrF6tthXbpYtHaqu1WUauoVcFSBUGpIFaBAIb7rJyiHOFIOJLM74/5mSeL2tqaZIB8Xo+Pj5OZzPczPMR35jvf+U56dHQ01eUoIwzCpyIIoqamBgBmzpwJEyYIJ0gZCCmevr6+r6+vr68vuTgwMFBWVkaGYmVl5d27dysrKysrK1NTU8kNzMzMPB/x8vIyNjYGgO7u7sDAwK6urqCgoAMHDlB2MP8rIiLi+vXrp0+fxiCkBE3m15+njF9++cXa2trExKSzsxMAXFxcuFxuUVEROckZVbZu3ZqcnJyQkCB5JAVCCAD4fH55eXlJSQmHw+FwOHfv3h13l7q1tbWbmxuXy21oaPD09MzLy5M8EYlyDx8+NDc3B4COjg5DQ0Oqy1E6eEb4VNKDM4VCYV1dHY1Gc3BwmAhV4RkhQuMwmUxvb29vb29yUSgU1tTUSPpRS0tLm5ubm5ubbWxs9PT0Lly4MHFSEABeeOEFf3//y5cvnzt3LjY2lupylA6OGn0q6chpbGyUvqGQQhTey4jQJKKqqurk5LR27dp9+/bl5+fzeLzKysrPP/+8ublZKBQaGBgAQF1d3fbt2w8fPkx1sQCP7qw/ffo01YUoIwzCp6J2ltEnGhwcbG1t1dDQsLGxobYShCYXOp0+a9asTZs2eXl5DQ0NZWZmAkB1dXVSUtIXX3xBdXUAAGFhYQwG4/r16+3t7VTXonQwCJ9qAt47UVVVRRCEvb29qir2aSP0R0hPafbyyy+zWKzS0lLyA04tXV3dl19+WSwWnz17lupalA4G4VNJXyOcIGeEOKcMQs8pIiKCTqdnZmbyeDw1NbWwsDAASEtLo7ouALyznjoYhE/W1dXV3d2tq6trZmYGEywIKS8DocnL1NR04cKFw8PDFy9ehAk253VISIi2tvatW7eam5uprkW5YBA+mXRf6LgbCimEQYjQ85M+8Vq8eLGJiUlNTQ2Hw6G6LmAymcuXLycIAofMKBgG4ZNJd0I+ePBATU3NxMSE8vt7MAgRen7h4eHq6urZ2dldXV10Oj08PBwmzEkhGdIYhAqGQfhkJ06cAAAGgwEAxsbG3d3d1dXV1JY0Ojra2NhIp9OpmiYfoamBxWIFBAQIhcLz58/Do97R7777biLMLkKO3+FwOBNh/I7ywCB8soGBAQDIyMgQCoXkK/r6+pRWBLW1tUKh0NbWVvHPDkVoipE+8Zo/f761tXVLS0thYSHVdYG6unpoaChMmPE7SgKD8Mn++te/AsC9e/fYbHZvby/V5QBgvyhCskPetJeXl9fW1kaj0cje0QnSIUmeoR44cCAlJaWgoGBwcJDqiqY+DMIne+ONN06cOGFiYnLlypWXXnqpoqKC6opwThmEZEZHRycwMFBy0x6ZPWlpaSKRiOrSoKGhQVVVtb+/Py4uztfXV0dHx9zcPCQkZNeuXRkZGV1dXVQXOAXhpNu/prW1NSwsrKioSEdH59ixY2SXBVUiIyNPnz797bff/uUvf6GwDISmhjNnzqxevdrb25vsEXVwcKitrb169aq/vz+FVWVmZq5YsUIkEkVHR6urqz9t9nAPDw/yQYzu7u7kLV7ouRDoVwkEAvLBKDQabceOHWKxWGFNc7ncgoICyaKrqysA3L59W2EFIDSFCQQCXV1dAKivrycI4l//+hcAxMTEUFhScXExOZvx7t27JS+OjY1xudyjR4/GxcX5+Pg8Ple4gYGBj49PXFzc0aNHuVyuIv+PmjIwCH+bWCxOSEhQUVEBgPDw8MHBQbk219fXd/To0YCAAADw8vIiXxSJRAwGg0aj8Xg8ubaOkPKIiooCgL179xIEQV6DNzAwGBkZoaSY1tZWCwsLAHjttdfIV54WaW1tbenp6fHx8Ww228jIaFwu6urqSueiUChU4EFMVhiEzyozM5McOOrm5tbU1CTz/YtEouzs7KioKPKeDQDQ19ffuHHj6OgoQRB1dXUAYGlpKfN2EVJa6enpADB79mxykex0ycjIUHwlPB6PbN3Pz0+SxDk5Ofr6+tKpJhKJHn9vQ0PD2bNn//GPfwQGBpqYmIzLRW1tbR8fn5UrV27YsEHeX+InLwzC36GmpoacXMbQ0DA3N1dWu21paUlISLC1tSV/cVVUVHx8fL766ivp31pygvzFixfLqlGE0OjoKIvFAoCKigqCID7++GMAWLNmjeLLIHuAHB0de3t7Ja8nJyePSzU9PT0/P7+tW7empqY+7Wyvp6fnxo0bycnJ0dHRjo6ONBpN8nZ7e3sFHtZkgkH4+/B4vJCQEABQVVXdt2/f8+xKIBCkpaWx2Ww6nU7+mlpYWOzYsaOhoUG6ObKblEajWVtb6+jo5OXlPfdBIIT+3+uvvw4A8fHxBEE0NjbSaDQdHR0+n6+wAsRiMTn8zczMrLm5edxa6V5QyXdlCTU1NUdHx+jo6OTk5Bs3bjyx7O7u7uzsbPL+EAAoKipSyGFNMhiEv5tQKNyxYwf5WxUbG/sHrigUFRXFxcVJJmzT1NQMDw/Pzs6WXBIQi8W5ubnR0dFMJlPyTZB8BqGGhsY333wj62NCSEllZ2cDgIODA7no5eUFAGfOnFFYAfHx8QDAZDJ//vnn39x43NkeOXBBQlVVVToXx3WEstlsAPjoo4/kdiiTGAbhH3Ty5EkypXx8fDo7O5/lLR0dHcnJyW5ubpJfXE9Pz+Tk5O7ubsk2ra2tCQkJdnZ2j3eTjo2NSQcwee0QIfQ8hEIheV2tpKSEIIj//Oc/ALBy5UrFtH7y5EkajUan0y9cuPAH3t7b25ubm/vvf/87Kipq1qxZkr4lEvks4j179pAbX7p0CQBcXV1legRTBAbhH1dSUjJ9+nSyS/POnTtP20woFJJdE2pqauQvqKmpaVxcXFlZmWSb4eHhcd2k06ZN27FjBzmwW1pqaio5xdrChQvv378vr2NDSGm89dZbAPDuu+8SBNHe3k6n0zU1Nfv6+uTd7vXr1zU0NAAgJSVFJjscGRmRvtGC/I9i8+bN5NrR0VGyF4rL5cqkuakEg/C5dHV1LVy4kOzePHbs2Li1LS0tW7duNTY2JrNNXV39z3/+c0ZGxtjYmGQbLpe7Y8eOF154gdxGQ0MjPDw8PT39VwY9FxQUmJqaAoCtre3du3fldWwIKYf8/HwAsLS0JK9N+Pn5AcDjH2fZqqioMDAwAIB33nlHTk2MjIwUFxfX1tZKXomJiQGA999/X04tTl4YhM9reHh4/fr18OiOe+nxzdXV1eSQrVmzZiUkJEifwPX09Hz11VezZ88e10368OHDZ2m0tbX1pZdeAgBtbe1z587J/qgQUhpisdja2hoAyPkrDh48CABBQUHya7Gjo8PKygoAVq1a9cQ7IuQkJycHAOzs7BTW4mSBQSgbX331FdnzGRgYKD0AOjExUXqY1uPdpCwWKzY2lsPh/N4WBQIBOdjs8QBGCP0u77zzDgDExcURBPHgwQN/f/+vv/5aTm3x+XxySM7cuXOHhobk1MoTiUQic3NzHDv6OAxCmcnLyyN7Qe3t7auqqsatraqq2rFjB9mlSV7HDggISEtLe84xL4mJiXQ6naai8mFh4RBmIUJ/SFFREXnxXt7zsAiFQnLKYltbW0qu8cfFxQHA9u3bFd/0RIZBKEsNDQ3Ozs4AoKurSw4Dk74RkIzAmTNnJiQkPONA02eRmZn5l1OnPIuKIisq2oaHZbVbhJSKvb09ABw/flyurWzatAkADA0Na2pq5NrQ09y8eZO8IIp9SNLw6RMyNjAwsHbt2gsXLtBoND09PYFAMDIyAgD6+vqRkZGvvfYaeW1PtlpGRrbV1zcOD+upqu61tX1JR0fmTSA0tQUHB2dmZmpqam7dujUoKMjNzU1H1p+jpKSk7du3a2pqXr161cfHR7Y7f0YEQdja2jY3N9+4ccPX15eSGiYgDELZIwhi3bp1x44dAwAajTZ//vy1a9dGRUU9Pm28DPFFovebm/P6+ug02l/Nzf/yqA8WIfQscnNzAwMDpR94ZGZm5vmIl5eXZPj3H5ORkREWFiYWi1NTU8nJvqny3nvvJSYmvvXWWykpKRSWMaFgEMrLmTNnCgoKVq9ePX/+fMW0SAAcbm8/3NFBAASxWDutrNRV8MHLCD2rwsLClJQUBoNRVlbG5XLJvhwJKysrySMAPTw8yFEnz+jOnTt+fn58Pj8xMfHdd9+VdeG/D4fD8fDwMDIyam9vV1VVpbaYCQKDcKrJ6u3d3dw8LBa7amt/amtr+Gh4KkLo2QmFwpqamuLi4uLi4srKypKSkp6eHukNDAwMHB0dJaeMs2bNUnnK986mpqZ58+bdv39/w4YNhw8fVkj5v8HJyamysjIrK2vp0qVU1zIhYBBOQbV8/raGho7RUWM1tX/b2TnKs0sWISXR3t5e/Mjt27e7urqk1+rq6rq4uEhycebMmeQsUT09PfPnz6+pqQkMDExPT58gZ2C7du364IMPXn/99SNHjlBdy4SAQTg19YyNvdPYWDY4qKGissva+k5//6BYTK4yVVOLs7CgtjyEJrvm5uaSkhIOh0P+3dHRIb1WS0vLzc3N1dX15s2b5eXlzs7O+fn5enp6VFU7DvlEOX19/c7OTnKaNyWHQThliQgiqbX1bFfXf2bM+Fdz81YLC2M1NQBg0OkueI6IkEz19vZWVFRIThnJO4kBYPr06f39/eXl5ZaWllTX+D/c3d1LS0vT09PJ58opOQzCKa6Wz7dnMv3Lyv7r4DBdU5PqchBSCr29vSUlJRcuXNi/f7+pqWlbW5uKikpzc/OJEydcXV0nQvYkJia+9957r7766okTJ6iuhXoYhErBv6xso7m5kZoaALzIYJhjZwhCCvHiiy/W19fn5eUtXLjwyJEjMTExf/rTny5fvkx1XXDv3j1ra2smk3n//n253tk1KeDwemVxu7//el/f9b6+Vqk7pRBCckU+Gv706dMAsHLlSnV19ZycnHEDbSgxffp0b2/voaGhH374gepaqIdBqCzemjYt3to63toa551BSGEiIiIA4MyZM0Kh0MDAYNmyZUKh8Ny5c1TXBfCotlOnTlFdCPUwCBFCSF7c3NwcHR0fPHiQm5sLEyx7IiIi6HR6ZmYmj8ejuhaKYRAihJAcrV69Gh6F34oVK5hMZn5+fmtrK9V1gamp6aJFi0ZGRi5cuEB1LRTDIFQK/7SyegGnmEGICuTMoufOnRseHtbR0QkODhaLxWlpaVTXBfDoDJW8hKnMMAiVwhJ9fSadTnUVCCmjGTNmeHh49Pf3X7lyBSZY9oSHh6urq2dnZ0+E8TsUwiBECCH5kg6/4OBgPT2927dv19XVUVWPQCAg/2FgYBAQECAUCr///nuqipkIMAgRQki+XnnlFRqNdvHixaGhIU1NzeXLlwMAVb2jJSUltra2ly5dIhcjIyMBICkp6dy5c42NjZSURDkMQoQQki9LS8t58+bx+XwyfsjsoaR39N69e2w2u7OzUzJApq+vT1NTs6mpadWqVXZ2dnp6er6+vps3bz527FhFRYVIJFJ8kYqHM8sghJDcpaSkxMXFhYaGnj9/fmxszMzMrLu7m8vlOjk5KayG/v7+BQsWlJeXL1q06MqVKxoaGpmZmStWrBCJRGw2WyQSPW32cHd3d3d3d+8//clh2jRVGu3XW2kZGekTCiWLM5lMtd96C+UwCBFCSO46OzstLCxUVVU7Ozv19fXfeOONQ4cO7dy5c8+ePYopYGxsLDg4ODs729HRMT8/38DAgMvl+vr68ni8+Pj4Xbt2kZs9bfZwAIgoLGxSV7fS0JippTWLyZzFZDowmYzHnsK4s6mpViCYpq5OLsZbW+tPjIdP/QoMQoQQUoSAgICcnJz//ve/69atu3bt2pIlS+zs7Orq6mjyP2EiCGL9+vXffvutmZlZYWGhlZVVW1vbvHnzWlpaIiMjT548+bQayNnDORwOh8Oh/+Mf1cPDYqm1KjSatYbGTCbTgckk/9am03c2NTlraUUaG8v7oGQIgxAhhBTh8OHDsbGx5KTbYrF4+vTpbW1td+7cmTNnjryb/uCDD3bt2sVkMnNzc728vAYGBhYsWFBWVrZw4cKsrKxnfyQhXyyu4/Nr+PxqPr9aIGgUCIRSCUIDeNPcvGl4GIMQIYTQE/T29pqamorF4ra2NmNj4y1btuzbt2/79u2ffvqpXNs9derUq6++qqKicu7cuRUrVoyNjbHZ7KysrFmzZhUUFBgYGPzhPQsJ4t7wcBWfT/6p4fN3Wlnl83g3+/u1VVQAYI6u7r+srGR3KHJDIIQQUgg2mw0AX3zxBUEQN2/eBABLS0uRSCS/FvPy8sgTvpSUFPKVN998EwCMjIzq6+tl25ZQLB4Vi//Z2Hi0s3NQKBwUCofleWgyhLdPIISQgkjfWe/t7W1jY9PS0lJQUCCn5qqqqkJDQ0dGRrZv3/7WW28BwJ49e7788ksGg5Genm5nZyfb5ug0GjlAVJ1G06LTteh0jceG0kxMk6NKhBCaAlasWMFgMG7cuNHW1kaj0eQ63dqDBw9CQkJ6e3tXrVqVmJgIAKdOnYqPj1dRUTl58qS3t7c8Gp2kMAgRQkhBxk26TQZhWlqaUOrGO5kQCATLly9vaGiYO3fu0aNHVVRUfvrpp3Xr1hEEkZycHBoaKtvmpE3X0Jh0U/xjECKEkOKQ08qQT2WaPXu29NMKZUUsFr/66qu3bt0ip1JjMpmSPtK3335706ZNMmzrcbHm5gHPMQCHEhiECCGkOJJJt+vr6+HR0wpl2zu6ZcuWCxcuGBoaZmZmGhsbP3z4cPny5b29vWw2+5NPPpFhQ1MGBiFCCCnOuEm3IyMj3d3dPT09ZbX/zz77LCUlRV1d/cyZMw4ODgKBICQkpL6+fu7cuadOnaLj49ieBO8jRAghhfrhhx/YbLarq2tZWZls95yRkREWFiYWi1NTU6OiosRi8apVq86fP29jY1NYWGhiYiLb5qYMPCNECCGFWrZsmaGhYXl5eX5+vgx3W1RU9Morr4hEor1790ZFRQHA1q1bz58/z2KxfvzxR0zBX4FBiBBCCqWmpubg4AAAK1as+PLLL2XyFMCmpiY2mz00NLRhw4YdO3YAQHJy8ueffy7pI33+JqYw7BpFCCFF++abbzZs2CD571dXV9fFxcXT09PT09PJycnFxUX90dMbnlFMTMyRI0cCAwPT09NVVVUvXboUGhoqFouPHTu2Zs0aORzBlIJBiBBCFDi5/Bf6AAAL5ElEQVR16tTBgwe1tbVLS0vb29ulVzGZTMlTAD08PJydnX8zF0dHRz/66KPt27fr6OgUFRX5+fkNDQ3t3bv3vffek+dBTBEYhAghRLFfeQogAKiqqtrb23s+4u7urqWl9bRdNTU1zZs37/79+6+//vqRI0cUUv6kh0GIEEITS39/f3l5uSQXa2pqRCKR9AZmZmaSXPT29jYyMiJf5/F4vr6+XC735ZdfzsjIUJ3wT8SdIDAIEUJoQhsaGiorK+NwOOQzcrlc7tjYmPQGtra27u7uNBotOzubx+M5Ozvn5+fr6elRVfCkg0GIEEKTiVAorKmpkZwvlpaWDg0NSdaqqqrW1tba2NhQWOGkg0GIEEKTmEgkqq6uvnPnzv79+01NTTdv3rx06VKqi5pkMAgRQggpNbyhHiGEkFLDIEQIIaTUMAgRQggpNQxChBBCSg2DECGEkFLDIEQIIaTUMAgRQggpNQxChBBCSg2DECGEkFLDIEQIIaTUMAgRQggpNQxChBBCSg2DECGEkFLDIEQIIaTUMAgRQggpNQxChBBCSg2DECGEkFLDIEQIIaTUMAgRQggpNQxChBBCSg2DcCqoqamZM2fO9evXqS4EIYQmHwxChVqzZg2NRqPRaGpqara2titWrOBwOM+/Wz6fX1xc3NfX94f3IBAIrl69+vyVIDRB5OTk0B5hMBguLi6ffvopQRDyaIvD4ZSVlT3/fhobG/Py8p5/P+j3UqW6AKVjaGh48OBBAGhvb//ss8+WLl169+5dMzMzaquqrKxcunSpWCym0WjUVoKQDCUkJHh6eg4MDKSlpb377rtqampbtmyReSuJiYkWFhZubm7PuZ+0tLSrV68uWrRIJlWhZ4dnhIrGYDDCw8PDw8M3b96clpbW3d19+vRpqouCtrY2qktASPbc3d0DAgLCwsJOnjw5e/bsr7/+etwGg4ODT3zjyMjI0/bJ5/PHvfK0j8+v7EQsFo+Ojj7/fpBMYBBSaebMmQDQ0dEBADNmzEhLSwsKClJXV9+4cSMAjI6O/v3vf7e0tKTRaObm5lu3bhUIBJL3ZmVlLViwgMFgaGtrr1y5UvL6mjVrIiMjJYu3bt2ys7P7+eefycXy8vLAwEBtbW0NDQ1fX9/bt28DgLu7e0xMDFmDnZ3dZ599poiDR0iBaDSajY1Ne3s7udjf379+/XptbW0dHR1jY+OkpCTy9bq6uoiICENDQ01NTX19/XXr1vF4PHKVWCz+4IMPzMzMtLS0mExmTEzM0NBQamoqi8UqLCw8cOAAi8VisVg8Hu/Bgwfr1683NTXV1NTU1tYODQ1tbW0FgAcPHrBYrMzMzIiICAaDoaWl5e/vT5ZUWlrKYrEOHTqUl5dH7ic/Px8Azp49O2PGDE1NTS0trZCQELFYTMHPThkQSIGioqIsLCwki6mpqQBw9OhRgiBoNJq2tvb27dtTU1Pz8vIIgoiOjtbU1ExKSrp58+b+/fu1tLTCwsLIN168eFFFRSUwMPDy5cv5+fkffvghAJw/f54giCVLlixYsEDSRG5uLgBcu3aNIIjW1lYWi+Xs7Jyenp6TkxMaGrpq1SqCIK5duxYdHQ0AWVlZ2dnZ9fX1ivuJICQf5DXvK1eukIsdHR1GRkZ+fn7k4rJly1gs1nfffVdRUbFt2zYAyM3NJQiis7Nz27Zt165dq66uPnPmDIPB2Lx5M/mWnTt3qqio7Nq1q7S0NC0tzcHB4fr1662trdnZ2QwGY9WqVdnZ2dnZ2WNjYyMjI1u2bLl8+XJ1dfWPP/5oamq6fPlycucAoKen9+GHH3K53MzMTB0dnbVr1xIEwePxsrOznZ2d3d3dyf309PS0tLSoqalFRkbevHkzPT09OTmZgp+jcsAgVKioqCgtLa3Y2NjY2NjAwEAVFRVPT08+n08QBI1G27hxo2TL1tZWGo22c+dOySuffPIJAFRVVREEMXPmTGdnZ6FQSK4qKSl5liCMj49XVVVtaGggV4nFYpFIRP47Pj4eAMRisTyPHiHFIYMwKCgoNjY2NDRUX19fV1f31q1bBEFwuVwAOHDgALmlSCQyNzePiop6fCehoaHu7u4EQQiFQh0dnVdeeUWySvrDoq2tvW3btqdV8vbbb+vp6RGPgnDPnj2SVatXr7a3t5cs+vr6+vv7SxbJceDk5xrJFQ6WUTSxWNzb2wsAVlZWBw8ejI6OZjAY5CojIyPJZpWVlQRBBAQESF5ZunQpAHC5XBaLVV1dvXv3bjqd/rua5nK5dnZ2tra25CI5oO45DwehiWxwcLC3t/fGjRt6enqFhYXkqLSqqioAOHr06Pnz58nN+Hx+Y2Mj+e+2trbjx483Njb29vaWlZUJhUIAaGlpGRgYWLJkiWTPv/7Z6enpOX78eG1tbVdXV1VVFY/HGxsbI1exWCzJZvr6+gMDA0/byZw5c6ZPnx4REREZGRkTE+Pr6/vHfgjoN2EQKpqhoWFaWtpvbtbf3w8AWlpakld0dHQAoK+vb2hoCP43NZ9Rf3+/trb2730XQpPXP//5z2XLlh05ciQmJobD4ZBBSA6QCQ4ONjU1JTcLDw/X09MDgNzc3KCgIDc3t6VLl9ra2jY1Nd2/fx8AyA+d5Dvrr6uoqFi4cKGJiUlISIinpyePxyPPQX8vLS2tO3fu7N+//7vvvjt27NiqVavS0tLwy6s8YBBOUOR5W11d3Zw5c8hXampqAMDOzm7atGl0Ov2XX36RbCw9jI1Op0uPqRm3z+Li4rGxMTU1tXGr8NOFprD169cfPnx48+bN/v7+GhoadnZ2AODh4cFms8dtuWfPHldX18LCQhUVFQCora3Nzs4GABsbGxqNRp5KPhEhdYdiUlISg8EoLi4mg3NsbCwrK+sZSyX+905HY2Pj3bt37969e+fOnR999BGHw/Hw8HjGXaFnh6NGJyhnZ+fp06cnJSWRg9aGhob27t1rbGzs5eWlrq4+b968Y8eOkcNNR0dHP/74Y8kbraysamtryS+wAoHgwIEDklVsNrunp4e81ggAYrGY/LYLAMbGxgBQX1+vqONDSHFUVFQOHDjQ2Ni4b98+AJg7d665ufmePXvIixQAUFNTQ85H0d/fr6WlRaZgXV2d5PZ2JpO5bNmyQ4cOSbLwl19+IXtNAcDMzEw6I/v7+zU1Ncmvm/fv37906dIz1mlqatrQ0CC5raK3t1dyKkmey0p3ESEZwiCcoNTU1E6cONHY2Dhjxgw/Pz87O7uysrLjx48zmUwA+OKLL0QikaOj4+LFi62trSUfSADYuHEjn893cnIKDAy0trYmO1RJbDZ7y5Yt77///syZM5csWWJpablp0yZyVVhYmLGx8fz58z08PKRjFaGpYc6cOdHR0R9++GF7e7umpubp06cbGxutra29vb1ffPFFV1dXcmTK3/72t7y8PA8PDx8fnwULFjg4OEj2cOjQIRMTEzc3N09PTycnJwcHh+LiYnJVXFzc5cuXZ8yYYWho2NPT8+abb967d8/FxcXPz8/JycnCwuIZi9y4cWNnZ6etra25ufn169e///57FxeXadOmOTk5bdq0ac2aNfb29rL+wSAAABohnzmH0BP9/PPPDx48eLxDBgDOnDnj6Ojo5OQk/WJvb+8PP/zQ1tZmamoaFBQkfV2Qx+Pl5OQ0NTXNmjUrMDDw4sWLc+fOnTZtGgA0NjZeuXJFIBAsXrx49uzZZ8+e9fPzk7y3srKyoKBAKBTa29v7+flJRtw8fPgwPT0dAPz9/a2srOT0E0BIMbq6uvLy8shrdeQrDx8+vHbtmpubGxkn/f39WVlZ3d3dJiYmPj4+kg9IcXHxTz/9pKmpGRoaShAEh8MJDg4mVwmFwvz8/MbGRiMjozlz5khPCFVaWlpeXm5tbe3j40On02tray9fvgwAwcHBL7zwQlZW1sqVK0dHRzMyMjw8PMi+WbKt9vb2kJAQyX6ampry8/ONjIwWLVrEYDA4HM7du3cHBwfd3Nx8fHzk/2NTUhiECCGElBp2jSKEEFJqGIQIIYSUGgYhQgghpYZBiBBCSKlhECKEEFJqGIQIIYSU2v8BVYFypW+/1ecAAALjelRYdHJka2l0UEtMIHJka2l0IDIwMjIuMDkuNQAAeJzNk09IFFEcx9+82Z2Z/TP7Z2bdnUkEIwzROlgXSdx9oR46FFKKUsjOsg6M6O4suyP9Mcy6eAjMsEuRlIhGVregk+7QpYKgQ1FdFIJACLLwYEQWzXu7mpndHtSDH7/v+/2+7zO/ecN8WniwBJwlgtKCTlQ6UeXECMMBzckMK4BaJ7OQK2XXrz3uQ8ihUp1DBs4ss0NhW8cDqjF6hxPb898IIULgts+4OduGny/52U1i+YAPMICB1ZCthdAFXG7g5gzI8RovGFDwaB6vAb0+4PVrftGAYkALBKuDIQOGwlpYMqAkAzkC5AoBylEBSjEtphgwoBpQ3QVUtybymhLUIiye0K0GRJ7jBI/X7+RQWIopwVgX48xQvnRQ+e76pH2u7gbCGz4+Y58a6yRa4Ubs5RWFaNix1/ZHm4iOz43ajVNGAuvJ96229axqHusP+zrs4aHHRay98x774GKVjXW3W7WfT58n+nPdZPH21E2iu+oPxMWxy0TX19xLPFm8S7TS/TJ+qbORaEtGxcqBL4TZvAqLM0PLC1ifXUknpLXZZqyH+1kk3uohnvt7etCqa454vvYPoDXfBJlTlqYROvmG6LqVF6jtlUje5eKPq+iY3kb0bM1DdO3OlTjW6vpuZNZwpP52dDwxsf6EnF3qfl38duE78bQ8fVRs+LhO6glX+TalcfydnQg6kUymzUxu0NJ72/NmrsBtmHgnsoOZw3kzQ2wnLD2vm2lDz7SaWT3glFr7ClYqm9aPpqx835nIH5Vke8oyCls4zOYzf4cx+DvLh3JC6ffaaSpYnkrAvZYj7cdT2X7iS1lmpjefOp0zC8n9DVvbEPdp8ioo80KUeVHKPIYyT6DMYynzfJR5kDLPTZkXpsxTKPM4yjw/ZZ6XMi9AmeeizOMp82TKPJUyT6LMi/3nPJEyL0KZ56HMw365yQL/ZsnRn284jLg043R3AAACfnpUWHRNT0wgcmRraXQgMjAyMi4wOS41AAB4nH1VW24bMQz89yl4AQt8SuRnbKdFUcQGmrR3yH/vj5K7SFYBhGotQtLOUhpxSJ+g2q/bz/e/8Nn4djoB4H9+EQF/BBFPL1ADuDx//3GH69vT5WPl+vh9f3sFHsCR3+TzFfv09nj5WCG4wllbsKEoaHMa3Q2w4daOTxkecLZGXZwZpIl0J10AZfeYa0MVuImHOS2AWh65WbiPPGqeQZhXQNs9soQJ5/tAH4QLYC+gNBxdTeCcAHFjXiDHjhQyVoIzNaOQWNHxQnJLYFKvkfbeuS+QsSMlMIzKe+jQGAtkHv5amzIPH1LcPLGyokQVIGw0JGJDSieMJZLhG1DGZeDgCpWgC8sKKZtPpdye6sjuyBQrpNZBsQ1DTiZ55BAOX3KyHWpEhlEjzSujlZSob/t7oJjXe68wrIAVJm5ootLLpXcb6CtkhUlySxPkQoow9iWl2JA9tK6RGmmxXwke4RWsZTDDegK7s+oSWCHqLUPted3cUii4FDxXDmmzzpLMUqeusmSTbh65defcu5TGjjaWSanbBZkQZvjzpqLnna6ABvciy0FaZOoMsQT2XZp9qG1x1hQzrQLOY0eq2MialNIYlcoL5PP99qXm7FXo8rjfjipUDx+1JicgR0XJCehRNyindlQHyt6PGqDZx5Homt2PbNbscaSsZqc5MbUM0ZR/tBme8ky3FZnyScuQTnmjZcim9KDN9CkLtAyNSe5ahnyStZahmORLZXjWKZVhmgTJ2wpPyttXZJKYbn500pKWYZtEUxRSE5M6qAyPSQVaKwfEd+qfHLgumWfmpYc5+jX/+P/K8ekfGkVQcPziCM4AAAFPelRYdFNNSUxFUyByZGtpdCAyMDIyLjA5LjUAAHicJVG7bcUwEFslpQ3Ih/vrDg+pDKR9RUZwnwkyfCgFcCHRJE3S9/s+Pt/n/cjz6PM8x9e5Dsf1f/s+7nO9x/P86PnI/fF7XE6twTacSmb2eF1BklY6jMyyBIgTM08fSlYdBUQpuqqAtJrq5qh12GBqrrlURjzTY1xMbBW6aTCVUB+XUEhbby8gkgMHz0zbkDU3pEbtM4AIqc6aA98pQCAxybRuW5ClMKwEiSeDFGRcpovkAp0s8yreEZhmsPZcGdq0e2MhEvDAyZFGYmmr2VAX+WuHfilxmPtiVcZkTGHgozYQM+U1oFG2w0pIHBO9glClE/cshfiV5D0xsBJqLw+nSMyIxbncZElSoWDS4lj9sbcJCyTcib+FqqKNlLKkvgfK6dHQeKPL3JCv+Gg5Ay3P3z958W+wT3SD/AAAA2B6VFh0cmRraXRQS0wxIHJka2l0IDIwMjIuMDkuNQAAeJzllFtIFFEYgM+c2Z29zl7c1VVX1928khj1ZqnMiGYYCVu+Bbazrptjtc6yrlRg6BIphYZhBEYEGfgQ7ZM9ZOHuRvgUoSJUJPQQdHuIegiMwGjOGWe9JPRyHoQGfs43///Pd/7DDPMtlXwH5IsFygXlKJejUo4higGCvFK0HlTJKw0ZZdVs3KM6hAyv5BleRCtN7ZDYVjEAL1Lv8ISyapUt5K1Rn7wTWjTqQBvrjmn1IdW5rWzDZUY9RfY0/xy6VBlap9zS2boJUICCXkhXQagB8uxaRoSMTtDpRag3CAajCI0mYDQLZlaErEWwWIHVBmx2YM8BdgewO4EzFzjzgNMFLPkizC8AhW7gLtJDd7EcHuApEWGJV/D6ROjbA3ylQmmZ3KUVyjwCqxNcVuDUoANp8y2sjmH0BqOZ1WltVpfTznjKSn3eEtcCJc+9/o5B+Tn/i8ximOXRDfPwaSYxdpRDrB0bzXzqpjAvl53MMNNLmPmli5k393rSiBsm3JnZ0CTmDzd/po/FKzOIr4S+p2bmDmJ21Nytt/PXMD/+keLmuyYwT0gefsh/CvOt5AD/cTKK+Ubbb852ncbsvk9z7VffY//XtYX6RfujFOLTexNc/coRPM/z5WquaNCK57f0/+Jur/KYn60Mc3BhGPP8mSa+oi6BeXzmMB+CEHP13AA/+MCP2fZ6lB9Zm8bOzlovP3W5ArO5fjYtFY7MIZ56uZqudSRx/snnTh4uvOIVhls40U03buSncf7LlIlTezrix1OtDUnMo/veplVucRVn1J4W16Uso56O/aBRfVZlTrP+FnPG0ccnh1WOQCAkRaL98XCXPyZF+xi1SSdHb3+kMSZFcFt7PBwLSyExHGmWesMWOdXc0xcP9obCbcF4rOeC869MwB+Mi32bPFR2z60yCn1fjkPjJuUvstNUcH0qPao1tfpPBHvP4r5gXIp0xYLno1JfoObA5jK0AbK+AsK+HMK+QsI+mrDPRNinJexjCfs0hH0MYZ+DsK+IsE9H2Gch7DMS9gHCvnzCPjthH9zlvv/tvKS/FzNhH0XY5yTsKybsc+9yXy5hn56wz0DYl0fY5yHscxH2obqj7g7Y3Zcj7w8Kn2e0blFIOQAAAxN6VFh0TU9MMSByZGtpdCAyMDIyLjA5LjUAAHichVZbbtwwDPzfU+gCEfjSg5/dJC2KIhugSXuH/ha9PzqUE9kBlF0/hBU9lobkkN5TiuPnw48//9I85OF0SomuXO6efisRnZ5S/Ejnx2/fL+n+9cv53XL//Ovy+pK0JXW8g/Mj9svr89O7hdN9uqtZi5bakmQh13gr0zj2VyU9p7uSe7Vm8bw5i8gCqLGiZXURNjwvVMxXQIsVNRt18544V229rbYusaJmJzCr6Y4ytq5SF8gaSMmm1KukO87sHY4tkC2QnN2EzOKdyvBMF8geSMpWvZAH0kmEVkgHElZ2DSBCAJjwAsjIERw2524ayEqVyhIZ+YFDrZRKI1pNa11tzpK+JgTJOnkJZCuuvkTq2N17bSbhEJVSbBUk5C88Kowo9QhXq955iSyb71LUKcKlhYFdIWs6Y3ciZ4qVWqfaV1nnBn1Qro0FZijThHXpT9/8KbU6JI8EFOaV4NgHkKD1zsmysMChldYjPxpAqBhA0m6NVkDegN2pF3Ak5ZDoAigD6OaESGuWSkZLYCTHUBSMUIKsAr90RqJ8BNWlZqGcWiDk5Yplk7pZYyuRG6rF6yo3Ut+qQgEYuxfU8DJCbfhTDGdJNbdPQ9nTSwTmNtCHP2/A/nlycB2A5QqQN/Xe3FplK3EZa2JJY226KkfVLUDtNtLeWtHY/uruJZ3/jv52G1rT5bB9h9S5U1sh2weXriAfLw8fPgjbJ+L8fHnYPxFxyv4hwCTp3u4xSbY3dca07K2bcde9Pxvutjdhw933Tmu4fW+nhpuPXdNiYD50Rx7DJIdSsWGZ/NDtLAaeFNHVLAaeLDloxjCJokvxGCZXtCMew6SLvsNjmIzRYIZFjp1ks0zOMoKJaw+ovFkmZzSBzTI5y+CMFjM5S3DGIJOzBGcUpkyGMYkKlMkwJjHose54WCbDkUWIWidDTCyqRPesR5xRDToZanCG6nUy1OAclskQUo8p1HlQtIUXOyQ0Ef9cdjp9S9f0SgdjO2QrNHxUbMzf/xDh9+k/S2S9f2HrDjQAAAGGelRYdFNNSUxFUzEgcmRraXQgMjAyMi4wOS41AAB4nG2RO47bUAxFt5LSBmSC/0dikCYC0rrIEtSnSD2Lz6U1QgaDWDYkHPF+Hr0/99v3530/5Dj0OI7bz/s83H4oXux3fD9uT70fstP+a8bnNwrM/bn/PuTb++2RZGGxNiXlNt/eHkGVvmJjWi2qCuRkrSpAweE9yMi5YhNKW1Uv0Az9gwmq1ARScuPK7SEkXTZIqF0ZY0opiAFi8uzgHtY8eUwsfSqRrGxgQt5SPiQ5ObY3uK8IzKDKsoQ7KniNkdGKth5VV65XHAc+4x2CBjWlVnadcRpWnVPeQkBHydwyR17FWT1juURnUa5iL+/ItRnKBgZHgU3W5qSia00dALRwYis/QTWiUcbEz76NoxhpssPCsVTR2aqxp8wZA0t1tMhg93OD7ktiunJGZ/m5VsPzCEPxEtbhuLakdZZB5v9BXXUvEBfg0+OTBpl6DbnYslf04hn6R/DHfJHhoF/IpSpsT0o+W1/o/v4X2IuXJd9gH6MAAAAASUVORK5CYII=",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "# we prepaire 200 samples in demo_data/test_cases.json, in which 's' is the source (product) and 't' is the target (reactant)\n",
    "test_cases = json.load(open('demo_data/test_cases.json'))\n",
    "# some typical cases in test_cases.json, change the value of test_case_idx below to view results\n",
    "#   1 Suzuki-Miyaura Cross Coupling\n",
    "#   2 Amide Condensation\n",
    "#   5 SN2 Substitution\n",
    "#   6 Ester Hydrolysis(Deprotection)\n",
    "#  16 Sonogashira Cross Coupling\n",
    "#  19 Nitro Reduction\n",
    "#  22 Free Radical Halogenation\n",
    "#  43 Nitrile Hydrolysis\n",
    "#  56 Aldol Condensation\n",
    "# 145 Sulphur Oxidation\n",
    "\n",
    "# please change test_case_idx, or change the input below, to try other examples\n",
    "test_case_idx = 1 # Suzuki-Miyaura Cross Coupling\n",
    "\n",
    "input_product_smiles, golden_reactants_smiles = test_cases[test_case_idx]['s'], test_cases[test_case_idx]['t']\n",
    "\n",
    "# the dual encoder was trained on canonicalized SMILES\n",
    "input_product_smiles = canonicalize_smiles(input_product_smiles)\n",
    "golden_reactants_smiles = canonicalize_smiles(golden_reactants_smiles)\n",
    "print(f'input_product_smiles: {input_product_smiles}')\n",
    "print(f'golden_reactants_smiles: {golden_reactants_smiles}')\n",
    "Draw.MolsToGridImage([Chem.MolFromSmiles(input_product_smiles), Chem.MolFromSmiles(golden_reactants_smiles)], \\\n",
    "    legends=['Product', 'Reactants'],\n",
    "    molsPerRow=2, subImgSize=(300, 300))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tokenized_product_smi:  C O C ( = O ) C c 1 c c 2 c c c ( F ) c c 2 c ( - c 2 c c c ( S ( C ) ( = O ) = O ) c n 2 ) c 1 C\n"
     ]
    }
   ],
   "source": [
    "# the input of the dual encoder is tokenized SMILES of product\n",
    "tokenized_product_smi = smi_tokenizer(input_product_smiles)\n",
    "# save the input to file\n",
    "save_input_for_dual_encoder(tokenized_product_smi)\n",
    "print('tokenized_product_smi: ', tokenized_product_smi)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "retrosub",
   "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": "3.7.16"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "f6d49a75f7ba54184f6e117f8690f4c521276805d6e956d52c324f39ee5d316f"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}