{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Contextual Word Embeddings and Text Embeddings\n",
    "* Last update: 11.11.2024\n",
    "* Author: Johannes Maucher"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Word Embeddings has been described in [section Representations](05representations). These Word Embeddings (), e.g. CBOW, Skipgram, Glove or the FastText subword embeddings are not contextual. This means that a word is mapped to a unique vector, independend of the context in which the word appears. This is a drawback because the meaning (semantic) of a word depends on the context (the surrounding words), in which the word appears. That's where **contextual Word Embeddings** come into play. In a  **contextual Word Embeddings** the vector representation of a single word or token, varies with the word's context. For example the word `play` will be mapped to another vector in the context of the sentence\n",
    "\n",
    "* `the members of the house of parliament play a crucial role in this debate`\n",
    "\n",
    "than in the context of the sentence\n",
    "\n",
    "* `girls like to play with dolls`\n",
    "\n",
    "Contextual word embeddings can be learned by transformers, as will be described in [section Transformers](../07neuralnetworks/attention). A common approach is to apply the pre-trained encoder-only transformer *BERT*. This approach will be demonstrated below.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load BERT Model and corresponding Tokenizer\n",
    "\n",
    "In this section we will load a pre-trained *BERT* model and the corresponding tokenizer from [HuggingFace](https://huggingface.co/google-bert/bert-base-uncased). *BERT itself* will be described in [section Transformers](../07neuralnetworks/attention). \n",
    "\n",
    "**Import required Python modules:**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import random\n",
    "import torch\n",
    "from transformers import BertTokenizer, BertModel\n",
    "from sklearn.metrics.pairwise import cosine_similarity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "random_seed = 1234\n",
    "random.seed(random_seed)\n",
    "\n",
    "# Set a random seed for PyTorch (for GPU as well)\n",
    "torch.manual_seed(random_seed)\n",
    "if torch.cuda.is_available():\n",
    "    torch.cuda.manual_seed_all(random_seed)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Load pre-trained BERT Model and associated Tokenizer:**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')\n",
    "model = BertModel.from_pretrained('bert-base-uncased')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Define Input Input Sentence and Tokenize it\n",
    "\n",
    "We start from an arbitrary sentence, which is assigned to the variable `text`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "text = \"The lesson on artificial intelligence will be held on Monday\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we tokenize the example sentence. For this we apply the `batch_encode_plus()`-method, which can also be applied for a bunch of texts, each represented as a variable of type `string`. The method automatically integrates the `CLS`- and the `SEP`-special token.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "encoding = tokenizer.batch_encode_plus( [text],# List of input texts\n",
    "    padding=True,              # Pad to the maximum sequence length\n",
    "    truncation=True,           # Truncate to the maximum sequence length if necessary\n",
    "    return_tensors='pt',      # Return PyTorch tensors\n",
    "    add_special_tokens=True    # Add special tokens CLS and SEP\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The tokenizer-method returns a dictionary (`encoding` in the cell below), which keeps the keys `input_ids` and `attention_mask`.\n",
    "The values of these keys are printed below:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input ID: tensor([[  101,  1996, 10800,  2006,  7976,  4454,  2097,  2022,  2218,  2006,\n",
      "          6928,   102]])\n",
      "Attention mask: tensor([[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]])\n"
     ]
    }
   ],
   "source": [
    "input_ids = encoding['input_ids']  # Token IDs\n",
    "# print input IDs\n",
    "print(f\"Input ID: {input_ids}\")\n",
    "attention_mask = encoding['attention_mask']  # Attention mask\n",
    "# print attention mask\n",
    "print(f\"Attention mask: {attention_mask}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The attention mask defines the attention, which is drawn to each single token. Since in our example all elements of the `attention_mask` are the same, each token is considered equally in the calculation of the contextual embedding.\n",
    "\n",
    "## Generate Word-Embeddings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of Word Embeddings: torch.Size([1, 12, 768])\n"
     ]
    }
   ],
   "source": [
    "# Generate embeddings using BERT model\n",
    "with torch.no_grad():\n",
    "    outputs = model(input_ids, attention_mask=attention_mask)\n",
    "    word_embeddings = outputs.last_hidden_state  # This contains the embeddings\n",
    "\n",
    "# Output the shape of word embeddings\n",
    "print(f\"Shape of Word Embeddings: {word_embeddings.shape}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As can be seen in the output above, `BERT` outputs for each of the 12 words in the input a contextual word embedding of length `768`.\n",
    "\n",
    "Below we just demonstrate the tokenizer's `encode()` and `decode()`-method:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Decoded Text: the lesson on artificial intelligence will be held on monday\n",
      "tokenized Text: ['the', 'lesson', 'on', 'artificial', 'intelligence', 'will', 'be', 'held', 'on', 'monday']\n",
      "Encoded Text: tensor([[  101,  1996, 10800,  2006,  7976,  4454,  2097,  2022,  2218,  2006,\n",
      "          6928,   102]])\n"
     ]
    }
   ],
   "source": [
    "# Decode the token IDs back to text\n",
    "decoded_text = tokenizer.decode(input_ids[0], skip_special_tokens=True)\n",
    "#print decoded text\n",
    "print(f\"Decoded Text: {decoded_text}\")\n",
    "# Tokenize the text again for reference\n",
    "tokenized_text = tokenizer.tokenize(decoded_text)\n",
    "#print tokenized text\n",
    "print(f\"tokenized Text: {tokenized_text}\")\n",
    "# Encode the text\n",
    "encoded_text = tokenizer.encode(text, return_tensors='pt')  # Returns a tensor\n",
    "# Print encoded text\n",
    "print(f\"Encoded Text: {encoded_text}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Run the following code-cell only if you like to see the 12 embeddings, each of length 768. We don't execute this cell, because of the very long output:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "#for token, embedding in zip(tokenized_text, word_embeddings[0]):\n",
    "#    #print(f\"Token: {token}\")\n",
    "#    print(f\"Embedding: {embedding}\")\n",
    "#    print(\"\\n\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we calculate the text-embedding from the contextual word embeddings at the BERT-encoder's output. Note that are different methods to calculate a text embedding from the list of word embeddings. The most popular options are:\n",
    "\n",
    "1. The encoder's output at the position of the `CLS`-token (first position in the input sequence) is a representation of the entire text and therefore can be applied as text-embedding.\n",
    "2. Just calculate the mean of all contextual word embeddings at the output of the transformer-encoder.\n",
    "\n",
    "Below we implement the second option:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of Sentence Embedding: torch.Size([1, 768])\n",
      "Sentence Embedding:\n",
      "tensor([[-3.5001e-01, -3.3019e-01,  2.6465e-01,  7.5814e-02,  2.6252e-01,\n",
      "         -4.2781e-01,  3.2828e-01,  6.8058e-01, -1.1294e-01, -1.3759e-01,\n",
      "         -2.1079e-01, -2.7575e-01,  1.1764e-01,  1.5474e-01, -1.3306e-01,\n",
      "         -6.5089e-03,  1.0366e-01, -2.5773e-02,  6.5012e-02,  4.5614e-02,\n",
      "         -8.1534e-02, -9.1516e-02, -9.6652e-02,  6.5698e-01,  3.4971e-01,\n",
      "          1.5745e-01, -2.1174e-02,  4.4196e-01, -1.7211e-01, -1.0172e-01,\n",
      "         -2.7239e-02,  8.8936e-02, -7.1396e-02,  1.2314e-01,  1.8995e-01,\n",
      "          1.0141e-01,  1.3849e-01, -2.6853e-01, -3.7551e-01,  5.1383e-01,\n",
      "         -4.5088e-01, -8.5676e-02,  3.5888e-01, -8.5598e-02, -3.4887e-01,\n",
      "         -3.2465e-01,  2.8293e-01, -2.6047e-01,  9.4687e-02, -3.1299e-01,\n",
      "         -6.5420e-01,  3.6622e-01,  2.0525e-01,  1.6598e-01, -1.8717e-02,\n",
      "          3.6724e-01,  4.8422e-02, -1.5004e-01, -2.3705e-01, -1.4570e-01,\n",
      "          3.3358e-02,  1.8378e-01, -3.5432e-01, -3.2449e-01,  4.1561e-01,\n",
      "          2.9656e-01,  2.2482e-02,  6.1011e-02, -5.8971e-01, -5.3985e-02,\n",
      "         -7.3305e-01,  2.3496e-02,  3.8414e-02,  1.3239e-01, -2.0427e-01,\n",
      "         -1.1042e-01,  1.4433e-01,  2.5816e-01,  1.5394e-01, -2.2436e-01,\n",
      "         -2.3106e-02,  4.3833e-01,  3.8140e-01,  2.4284e-01,  3.2330e-01,\n",
      "          4.8820e-01, -3.6262e-01, -4.5515e-02, -4.6194e-01,  5.3348e-01,\n",
      "         -1.6715e-01, -1.2979e-01, -1.6823e-02,  2.2146e-01,  2.5272e-01,\n",
      "         -4.0833e-03, -8.4667e-02,  2.3374e-02, -1.8031e-01,  5.2358e-01,\n",
      "          3.6706e-01, -3.6619e-01, -1.0291e-01,  3.6938e-01,  3.5335e-01,\n",
      "         -1.5034e-01, -3.5996e-01, -6.9851e-01,  1.5057e-03,  8.2513e-02,\n",
      "         -1.0837e-01, -1.4382e-01,  2.8829e-02, -4.0238e-01, -3.9316e-01,\n",
      "          1.7526e-01, -2.8134e-02, -7.4224e-02, -2.6436e-01,  8.5258e-02,\n",
      "         -2.2381e-01, -3.5146e-02,  2.9077e-01,  8.7016e-01, -3.0988e-01,\n",
      "          2.5399e-01,  6.1589e-02,  3.6281e-01,  2.3399e-01, -4.6509e-01,\n",
      "          3.7167e-01,  2.6715e-01,  6.8925e-01, -4.8475e-01, -1.4224e-02,\n",
      "          2.1432e-01,  1.4330e-01, -2.2511e-01, -3.2580e-01,  4.6141e-01,\n",
      "         -2.1279e-01, -1.2618e-01,  2.7136e-01, -5.1448e-01,  3.4573e-01,\n",
      "          1.5227e-03, -2.3293e-01, -3.1338e-01, -1.7379e-01,  7.6039e-02,\n",
      "         -5.6773e-05,  2.7782e-01, -6.1375e-01, -2.3112e-01, -2.6738e-01,\n",
      "         -1.6650e-01, -1.7668e-01,  3.2612e-01,  1.8533e-02,  2.2091e-01,\n",
      "          2.4410e-01, -3.2489e-01, -4.4623e-02,  2.8956e-01, -2.4651e-01,\n",
      "          4.0018e-01,  1.2866e-01,  8.7871e-02,  2.9744e-01,  1.9692e-01,\n",
      "         -3.2523e-01, -1.4092e-01,  5.1488e-01, -8.8644e-02, -1.4185e-01,\n",
      "         -1.8024e-01, -3.3525e-01, -1.0060e-01,  5.8080e-02,  4.4633e-02,\n",
      "         -9.4789e-01,  3.4185e-02,  1.8970e-01, -3.8252e-01,  7.1670e-02,\n",
      "         -4.9928e-01,  1.6674e-01, -4.2020e-01, -2.9474e-01,  4.1578e-01,\n",
      "         -3.0630e-01,  1.6033e-01, -4.2800e-01, -3.7009e-01,  5.4866e-01,\n",
      "         -3.3470e-01,  1.0239e-01,  1.5431e-01, -4.5977e-01,  3.8354e-01,\n",
      "          3.6915e-01, -7.4490e-02,  5.9416e-02,  3.2153e-01, -3.1410e-01,\n",
      "          6.1120e-02, -8.8887e-02, -4.9073e-01, -1.9812e-01,  6.4406e-02,\n",
      "         -2.9210e-01,  4.5945e-01,  7.2102e-02, -2.6303e-02,  2.4272e-01,\n",
      "         -2.4792e-01,  4.9993e-02,  2.1967e-01, -3.7672e-01, -1.7933e-01,\n",
      "          2.7616e-01,  2.6674e-01, -4.0560e-01,  4.4782e-01,  3.0446e-02,\n",
      "          3.3901e-01, -1.0932e-02,  9.4912e-02,  2.4630e-01,  5.3611e-02,\n",
      "         -2.0549e-01, -2.9059e-01,  2.4120e-01,  3.3106e-02, -1.3610e-02,\n",
      "         -2.5216e-01, -7.8993e-01, -2.2043e-01,  1.0860e-01, -4.3867e-01,\n",
      "         -1.3193e-01,  1.8721e-01,  1.1624e-01,  1.2369e-01,  3.4737e-01,\n",
      "         -1.7266e-01,  2.1957e-01,  4.3804e-01, -9.4424e-02, -3.5881e-01,\n",
      "         -4.2728e-01, -9.7560e-01, -5.6313e-02, -4.3410e-01, -6.1033e-01,\n",
      "          4.9483e-02, -4.7547e-01, -3.3615e-02,  2.7971e-01,  4.4012e-01,\n",
      "          2.6933e-01,  2.4647e-01,  3.2145e-01,  1.9147e-01, -1.8743e-01,\n",
      "         -5.1153e-01,  4.6308e-02, -1.8047e-01,  2.4499e-01,  4.6406e-02,\n",
      "          2.6046e-02, -6.1334e-01,  4.6881e-01,  1.5311e-01, -4.2407e-01,\n",
      "         -5.3723e-01,  3.9752e-01,  3.2234e-02, -2.2177e-01,  3.1721e-03,\n",
      "          1.5712e-01,  4.1237e-01, -3.2868e-01,  4.7417e-01,  6.7273e-02,\n",
      "         -3.8105e-01,  1.6562e-01, -2.2771e-01, -2.1024e-01, -2.7373e-01,\n",
      "         -7.0715e-02,  5.3727e-01, -1.6749e-01, -2.6956e-01,  9.9304e-02,\n",
      "          1.2754e-01,  1.5457e-01, -1.9934e-01,  2.5971e-01, -2.5183e-01,\n",
      "          3.2693e-02,  4.1218e-02, -3.1145e-02,  1.1443e-01, -1.2888e-01,\n",
      "         -3.2929e-03, -1.3154e-01, -2.1232e-01, -4.0269e+00, -6.0261e-02,\n",
      "         -1.2884e-01,  1.3271e-01,  2.1156e-01,  6.5302e-02,  7.5073e-03,\n",
      "         -8.8058e-02, -5.0960e-01, -2.6623e-01, -4.8730e-01, -2.1142e-01,\n",
      "          3.3794e-01,  1.7242e-01,  2.9090e-01,  1.3112e-01,  5.8265e-01,\n",
      "         -2.3833e-01,  1.0395e-01,  5.4858e-01, -2.4176e-01, -2.4934e-01,\n",
      "         -2.0685e-02, -7.9416e-02,  4.1731e-01,  1.0205e-01, -5.3768e-01,\n",
      "          1.9558e-01, -3.5325e-01, -7.3941e-03,  1.4926e-01, -5.9997e-02,\n",
      "          3.4697e-01,  2.0086e-01,  1.8412e-01, -1.5491e-02, -1.0581e-01,\n",
      "         -1.1161e-03,  1.9717e-01,  1.1820e-02, -4.6035e-02,  9.1656e-02,\n",
      "         -2.7942e-01, -3.2009e-03,  9.4838e-01, -1.4958e-01,  4.1843e-02,\n",
      "         -2.7940e-01, -6.7275e-02,  6.5053e-02, -1.0996e-01,  1.7743e-01,\n",
      "         -1.5813e-01, -7.0556e-02, -2.1256e-01, -1.4250e-01,  4.0637e-01,\n",
      "          1.9027e-01, -1.8110e-01, -3.8689e-01,  4.0752e-01, -1.9994e-01,\n",
      "         -3.3271e-01,  2.1929e-02, -1.0427e-01, -2.4982e-01, -6.1833e-01,\n",
      "         -3.3414e-01,  9.3490e-02,  1.8034e-01,  2.6657e-01, -3.7683e-02,\n",
      "         -5.9592e-01, -9.9460e-01, -3.1050e-01,  4.9618e-02, -1.8509e-02,\n",
      "         -1.9436e-01,  3.6091e-01,  1.1406e-01,  9.9842e-02, -6.3950e-01,\n",
      "         -5.0970e-04,  1.7271e-01,  6.2532e-02, -2.7805e-01,  9.7467e-02,\n",
      "         -1.0300e-01, -1.7447e-01, -4.6657e-01,  6.9973e-02,  4.7557e-01,\n",
      "          2.8477e-02, -5.7240e-02,  1.1543e-02,  9.7917e-02,  4.2295e-01,\n",
      "         -2.7138e-01,  1.5050e-01,  2.2147e-01, -1.7118e-01, -6.0872e-01,\n",
      "          2.8294e-01,  9.7952e-02,  3.4805e-01,  5.8905e-02, -3.4898e-01,\n",
      "          2.9768e-01,  3.2487e-02,  9.4686e-02,  5.3735e-01, -4.1341e-01,\n",
      "          3.9576e-01, -1.5443e-01,  5.1577e-03,  1.7772e-01, -2.7315e-02,\n",
      "          5.1828e-01, -7.3625e-02, -8.1733e-02, -9.8084e-02,  2.2278e-01,\n",
      "         -8.4016e-03,  1.1859e-01, -2.8934e-01, -1.2390e-01, -6.1686e-02,\n",
      "         -2.9135e-01,  1.3129e-01, -1.8670e-01,  3.4517e-02,  8.4539e-02,\n",
      "         -1.2491e-01,  3.3284e-01,  2.3134e-01, -1.2112e-01, -9.6196e-02,\n",
      "         -2.2862e-01,  1.7466e-01,  1.1974e-01,  1.5807e-01,  3.9687e-01,\n",
      "          2.7971e-01, -2.4582e-02,  8.7744e-02,  2.1067e-01, -1.4803e-01,\n",
      "         -1.3783e-02, -2.0617e-01,  1.8067e-01, -2.8827e-01, -2.3446e-01,\n",
      "         -2.0790e-01, -3.0983e-01, -1.0412e-01,  1.5453e-01,  2.1305e-01,\n",
      "         -1.9187e-01,  5.0736e-02, -1.3197e-01, -1.9838e-01, -1.4006e-01,\n",
      "         -1.2290e-01,  3.0371e-01, -4.0152e-02,  7.6067e-01, -2.7431e-01,\n",
      "         -6.8675e-02, -3.8117e-01,  7.5752e-02, -1.5208e-01, -1.4443e-01,\n",
      "         -2.9547e-02, -3.5503e-01,  1.8192e-01, -2.5108e-02,  1.9621e-02,\n",
      "         -5.7128e-02, -1.7207e-01,  3.7231e-01, -6.8502e-02, -2.5235e-01,\n",
      "         -4.1659e-01,  1.0478e-01,  8.4711e-02,  3.4261e-01,  8.1984e-02,\n",
      "          6.0531e-02,  1.1377e-01,  2.9045e-02, -7.8550e-02,  2.8527e-01,\n",
      "         -3.6355e-01, -1.4366e-01, -5.3405e-01, -3.4555e-01,  5.4226e-01,\n",
      "          6.1720e-02,  2.8459e-01,  6.7661e-02,  2.7893e-01,  6.7947e-02,\n",
      "         -2.9389e-01,  6.1849e-01,  3.4890e-01, -4.4523e-02,  1.2058e-02,\n",
      "          1.8805e-01, -1.7503e-01,  1.1556e-01, -3.1747e-01, -1.3132e-01,\n",
      "         -3.1923e-01, -2.9955e-01,  1.1621e-01, -4.7510e-01, -2.4903e-01,\n",
      "          5.0876e-02, -1.7723e-01,  4.4682e-01, -3.9996e-01,  1.3958e-01,\n",
      "          3.1837e-01,  4.3305e-02, -2.4848e-01, -2.6041e-01,  1.5642e-02,\n",
      "         -1.1489e-01, -1.9163e-01, -4.1618e-02,  1.1694e-02, -1.2724e-01,\n",
      "          2.3057e-01,  2.1819e-01, -1.8967e-01, -1.9907e-02, -5.0891e-01,\n",
      "         -3.5877e-01,  7.2083e-02, -1.8683e-01, -1.2663e-01, -2.9630e-01,\n",
      "         -3.1224e-02, -2.2234e-01, -2.2673e-01, -6.4638e-02, -1.9064e-01,\n",
      "          1.9712e-01,  9.9471e-04, -2.8300e-02, -1.4606e-01,  1.4804e-01,\n",
      "         -2.5294e-01, -6.8988e-02, -4.0995e-01, -6.0582e-01,  2.8313e-02,\n",
      "         -3.8813e-02, -7.6255e-02, -3.3953e-02, -5.8762e-02, -1.8847e-01,\n",
      "         -2.8234e-01,  3.4244e-01,  2.2865e-01,  4.3288e-01,  5.5052e-02,\n",
      "         -1.6130e-01,  2.2415e-01,  4.4876e-01, -1.9497e-01, -2.1223e-01,\n",
      "         -1.0491e-01, -1.6519e-01, -1.6083e-01,  1.3944e-01,  6.6475e-02,\n",
      "         -5.8687e-01, -2.9872e-01, -1.7971e-01,  1.1832e-01,  2.1972e-01,\n",
      "         -2.9425e-01, -1.0049e-01,  3.6542e-01, -4.7722e-01, -3.1582e-01,\n",
      "          5.3376e-01, -1.1965e-01, -1.2218e-01, -2.5652e-02,  1.4879e-01,\n",
      "         -6.7233e-02,  9.2874e-02,  1.2487e-01,  4.9830e-01,  2.5855e-01,\n",
      "         -1.4293e-01, -1.1309e-01,  1.8774e-01, -1.4081e-01,  4.3225e-01,\n",
      "          7.1667e-02,  1.8105e-01, -4.7223e-01,  3.2808e-01,  1.0556e-01,\n",
      "         -2.2698e-01,  3.4445e-01,  3.5342e-02, -1.3785e-01,  3.2490e-01,\n",
      "          5.4810e-01,  7.1389e-01, -4.8402e-01, -3.2673e-02,  6.6809e-02,\n",
      "         -5.0849e-01, -3.4669e-01,  4.0453e-01, -8.7078e-02, -6.8191e-02,\n",
      "          4.1731e-01, -5.4170e-02, -4.6859e-02,  4.7115e-01, -1.7083e-01,\n",
      "          9.9105e-02, -1.2825e-01,  2.7454e-01, -2.0900e-01,  7.4443e-01,\n",
      "         -1.7870e-01,  5.1283e-01, -2.5385e-01, -5.6341e-01,  7.2013e-02,\n",
      "         -1.8784e-01,  2.7014e-01,  6.1444e-02,  4.0886e-01,  4.9157e-01,\n",
      "          3.8884e-01,  2.7868e-01,  1.3078e-01,  3.0113e-02,  9.5825e-02,\n",
      "          3.1781e-01,  4.1155e-01,  5.6720e-01,  2.3901e-01,  5.9801e-02,\n",
      "          1.2964e-01, -2.9834e-01,  5.2350e-01,  1.1932e-01, -1.5596e-01,\n",
      "          6.4928e-01,  3.5083e-01,  4.2034e-01,  5.1993e-01,  3.2745e-01,\n",
      "          5.0950e-01, -5.3826e-01, -2.9767e-01,  4.2256e-01,  7.3183e-02,\n",
      "          1.7045e-01, -3.5584e-01,  2.6905e-01,  2.8558e-01,  2.5299e-01,\n",
      "          4.5645e-02, -3.0377e-01,  6.7047e-02,  2.7107e-01, -1.6785e-01,\n",
      "         -3.8109e-01, -3.7910e-02, -1.0643e-01,  4.1419e-01,  9.4466e-02,\n",
      "         -4.4282e-01, -4.5154e-02, -2.8278e-02, -4.7644e-01, -5.1879e-01,\n",
      "          2.0594e-01,  7.6130e-03, -1.2532e-01, -5.3660e-02,  2.2753e-01,\n",
      "          4.1270e-01, -5.2486e-01, -2.3312e-02, -1.2330e-01,  1.8143e-01,\n",
      "         -4.6706e-02,  4.2303e-01,  2.0783e-01,  3.4429e-01, -3.5113e-01,\n",
      "         -1.3195e-01, -2.3233e-02,  6.8567e-02, -1.8425e-01, -1.0333e-01,\n",
      "          1.8997e-01, -4.5926e-01, -2.3890e-01,  3.7245e-02,  2.4184e-01,\n",
      "         -7.7843e-01, -6.9345e-02,  2.4815e-01,  2.4974e-01, -1.9161e-01,\n",
      "         -1.1773e-01, -1.5767e-01, -2.4952e-01, -2.2291e-01, -2.2567e-01,\n",
      "          1.7225e-01,  3.7873e-01, -2.4091e-01,  3.8792e-01, -9.1646e-03,\n",
      "          7.4300e-02,  8.9568e-02, -3.7235e-01,  2.3798e-01,  7.5409e-02,\n",
      "          3.5144e-01, -1.3033e-01,  1.0079e-01, -6.0263e-02, -3.6118e-01,\n",
      "          3.2302e-01, -8.2100e-02, -1.8806e-01,  4.9064e-02, -1.2287e-01,\n",
      "          2.8884e-01, -9.0585e-02, -2.6741e-01, -3.8880e-01, -1.4117e-01,\n",
      "         -9.0293e-02, -2.0516e-01, -4.1247e-01,  1.4714e-01,  6.7941e-02,\n",
      "          4.5147e-01, -3.0173e-01, -1.4309e-01,  1.5541e-02, -3.4537e-02,\n",
      "         -9.3550e-02,  3.1032e-02,  2.7587e-02]])\n"
     ]
    }
   ],
   "source": [
    "sentence_embedding = word_embeddings.mean(dim=1)\n",
    "print(f\"Shape of Sentence Embedding: {sentence_embedding.shape}\")\n",
    "# Print the sentence embedding\n",
    "print(\"Sentence Embedding:\")\n",
    "print(sentence_embedding)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Calculate Similarity between texts\n",
    "Up to now, we generated a text-embedding of our sample text `text`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'The lesson on artificial intelligence will be held on Monday'"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "text"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now, we like to compare this text with other texts. For this, we first calculate the text-embeddings of the new texts. Then for each of these new texts the text-embedding is determined and the **cosine similarity** between our original `text` and all new texts is calculated. We expect, that we obtain the largest cosine similarity score for the text, whose semantics is closest to the sematics of the original `text`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Example sentence for similarity comparison\n",
    "example_sentence0 = \"Cuttlery and plates are piled up on the table\"\n",
    "example_sentence1 = \"European cities suffer from severe air pollution\"\n",
    "example_sentence2 = \"The machine learning module is in the master's program\"\n",
    "example_sentence3 = \"Artificial intelligence is the future of technology\"\n",
    "example_sentence4 = \"Waste of energy is a major concern for the environment\"\n",
    "example_sentences=[example_sentence0,example_sentence1,example_sentence2,example_sentence3,example_sentence4]\n",
    "# Tokenize and encode the example sentence\n",
    "example_encodings = tokenizer.batch_encode_plus(\n",
    "    example_sentences,\n",
    "    padding=True,\n",
    "    truncation=True,\n",
    "    return_tensors='pt',\n",
    "    add_special_tokens=True\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[  101,  3013, 25091,  2100,  1998,  7766,  2024, 17835,  2039,  2006,\n",
       "          1996,  2795,   102],\n",
       "        [  101,  2647,  3655,  9015,  2013,  5729,  2250, 10796,   102,     0,\n",
       "             0,     0,     0],\n",
       "        [  101,  1996,  3698,  4083, 11336,  2003,  1999,  1996,  3040,  1005,\n",
       "          1055,  2565,   102],\n",
       "        [  101,  7976,  4454,  2003,  1996,  2925,  1997,  2974,   102,     0,\n",
       "             0,     0,     0],\n",
       "        [  101,  5949,  1997,  2943,  2003,  1037,  2350,  5142,  2005,  1996,\n",
       "          4044,   102,     0]])"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "example_input_ids = example_encodings['input_ids']\n",
    "#example_input_ids = example_encodings['input_ids'][1]\n",
    "example_input_ids"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "example_attention_mask = example_encodings['attention_mask']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Generate embeddings for the example sentence\n",
    "with torch.no_grad():\n",
    "    example_outputs = model(example_input_ids, attention_mask=example_attention_mask)\n",
    "    example_sentence_embedding = example_outputs.last_hidden_state.mean(dim=1)\n",
    "    #example_sentence_embedding = example_outputs.last_hidden_state[0,0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Similarities of 5 example sentences and the original text \n",
      "\t The lesson on artificial intelligence will be held on Monday\n",
      "\t\t0 : Cuttlery and plates are piled up on the table : Similarity Score = 0.560899555683136\n",
      "\t\t1 : European cities suffer from severe air pollution : Similarity Score = 0.552908182144165\n",
      "\t\t2 : The machine learning module is in the master's program : Similarity Score = 0.724198043346405\n",
      "\t\t3 : Artificial intelligence is the future of technology : Similarity Score = 0.6622300744056702\n",
      "\t\t4 : Waste of energy is a major concern for the environment : Similarity Score = 0.5618582367897034\n"
     ]
    }
   ],
   "source": [
    "\n",
    "# Compute cosine similarity between the original sentence embedding and the example sentence embedding\n",
    "similarity_score = cosine_similarity(sentence_embedding.reshape(1, -1), example_sentence_embedding)\n",
    "\n",
    "# Print the similarity score\n",
    "print(f\"Similarities of {len(example_sentences)} example sentences and the original text \\n\\t\",text)\n",
    "for idx,sample in enumerate(example_sentences):\n",
    "    print(f\"\\t\\t{idx} : {sample} : Similarity Score = {similarity_score[0][idx]}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As can be seen in the output of the previous cell, we actually obtain the largest similarity-scores for the sentences, which are semantically closest to the original `text`."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Calculate Similarity between different contextual embeddings of the same word\n",
    "BERT outputs contextual word embeddings, i.e. a given word is not mapped to a unique vector. Instead the vector representations of a single word vary with the context of the word.\n",
    "\n",
    "In the experiment below we generate 6 different sentences. In each of these sentences the word `show` is at the 2nd position. We like to investigate and compare the 6 different vectors generated at the BERT output for the word `show`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "s0 = \"I show you the easiest way to solve this problem.\"\n",
    "s1 = \"They show remarkable resilience in difficult situations.\"\n",
    "s2 = \"We show our appreciation by sending a thank-you note.\"\n",
    "s3 = \"You show kindness to everyone you meet.\"\n",
    "s4 = \"Teachers show their students how to find reliable information.\"\n",
    "s5 = \"Friends show support when times get tough.\"\n",
    "sents=[s0,s1,s2,s3,s4,s5]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Encode the 6 sentences:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Tokenize and encode the example sentence\n",
    "example_encodings = tokenizer.batch_encode_plus(\n",
    "    sents,\n",
    "    padding=True,\n",
    "    truncation=True,\n",
    "    return_tensors='pt',\n",
    "    add_special_tokens=True\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "By checking the output of the following cell, we can see, that in each of the 6 sentences the ID at position 3 (index = 2) is the same. This is because the second word in each sentence is the word `show`. Note that the second word is at the third position, because the special token `CLS` has been attached in the tokenisation method above. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[  101,  1045,  2265,  2017,  1996, 25551,  2126,  2000,  9611,  2023,\n",
       "          3291,  1012,   102,     0],\n",
       "        [  101,  2027,  2265,  9487, 24501, 18622, 10127,  1999,  3697,  8146,\n",
       "          1012,   102,     0,     0],\n",
       "        [  101,  2057,  2265,  2256, 12284,  2011,  6016,  1037,  4067,  1011,\n",
       "          2017,  3602,  1012,   102],\n",
       "        [  101,  2017,  2265, 16056,  2000,  3071,  2017,  3113,  1012,   102,\n",
       "             0,     0,     0,     0],\n",
       "        [  101,  5089,  2265,  2037,  2493,  2129,  2000,  2424, 10539,  2592,\n",
       "          1012,   102,     0,     0],\n",
       "        [  101,  2814,  2265,  2490,  2043,  2335,  2131,  7823,  1012,   102,\n",
       "             0,     0,     0,     0]])"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "example_input_ids = example_encodings['input_ids']\n",
    "#example_input_ids = example_encodings['input_ids'][1]\n",
    "example_input_ids"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "example_attention_mask = example_encodings['attention_mask']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we generate for each of the 6 sentences the corresponding BERT output. As before, for each sentence the i.th output vector of is the contextual embedding of the i.th input word. Again, we have to take into account, that the special token `CLS` has been added at the first position of each sentence.\n",
    "\n",
    "In contrast to the previous experiments, we now do not calculate the mean over all contextual word embeddings of a sentence. Instead we just pick out the contextual word embedding at postion with index 2. Because at this position we have in each of the 6 sentences the same word `show`. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Generate embeddings for the second word (IDX=2) of the example sentence\n",
    "IDX=2\n",
    "with torch.no_grad():\n",
    "    example_outputs = model(example_input_ids, attention_mask=example_attention_mask)\n",
    "    #example_sentence_embedding = example_outputs.last_hidden_state.mean(dim=1)\n",
    "    example_sentence_embedding = example_outputs.last_hidden_state[:,2]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The output `example_sentence_embedding` is a tensor of shape (6, 768), where 6 is the number of example sentences and 768 is the dimension of the embeddings."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([6, 768])"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "example_sentence_embedding.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we calculate the pairwise cosine similarity between all 6 sentences:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "similarity_score = cosine_similarity(example_sentence_embedding)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We obtain a $6 \\times 6$ array, whose value at row $i$, column $j$ is the cosine-similarity between sentence $i$ and sentence $j$."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6, 6)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "similarity_score.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1.   , 0.39 , 0.487, 0.508, 0.685, 0.472],\n",
       "       [0.39 , 1.   , 0.495, 0.655, 0.483, 0.6  ],\n",
       "       [0.487, 0.495, 1.   , 0.538, 0.543, 0.533],\n",
       "       [0.508, 0.655, 0.538, 1.   , 0.536, 0.709],\n",
       "       [0.685, 0.483, 0.543, 0.536, 1.   , 0.59 ],\n",
       "       [0.472, 0.6  , 0.533, 0.709, 0.59 , 1.   ]], dtype=float32)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np\n",
    "np.set_printoptions(precision=3)\n",
    "similarity_score"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Finally we visualise the pairwise similarities in a heatmap:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: >"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA6MAAANCCAYAAACTfho7AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8pXeV/AAAACXBIWXMAAA9hAAAPYQGoP6dpAAEAAElEQVR4nOzdd1RURxvA4d/Se1fAig17x4IFsGKNXSxRsXfErp+xG41RLNFo1CjWWBJLEnsFe0OxEiuIBcQCFlCU8v1BWF1ZWgKyhvc5Z89h5869d2Z2dtn3ztxZRWJiYiJCCCGEEEIIIcRnpJXTBRBCCCGEEEIIkftIMCqEEEIIIYQQ4rOTYFQIIYQQQgghxGcnwagQQgghhBBCiM9OglEhhBBCCCGEEJ+dBKNCCCGEEEIIIT47CUaFEEIIIYQQQnx2EowKIYQQQgghhPjsJBgVQgghhBBCCPHZSTAqhBBCCCGEEOKzk2BUCCGEEEIIIXKxo0eP0rJlS/Lly4dCoWDHjh3p7uPv70/VqlUxMDCgaNGi/PTTT5k+rwSjQgghhBBCCJGLRUdHU7FiRRYvXpyh/MHBwTRr1oy6dety8eJF/ve//+Hl5cXWrVszdV5FYmJi4j8psBBCCCGEEEKI/xaFQsH27dtp3bp1qnnGjh3LH3/8QVBQkDJtwIABXLp0iVOnTmX4XDIyKoQQQgghhBD/MbGxsbx8+VLlERsbmyXHPnXqFI0bN1ZJc3d35/z587x//z7Dx9HJktIIIYQQX6iEcMecLoJGcM9XMaeLoBEU1cvndBE0QryRbk4XQSOENNfP6SJohBLronK6CBphb+C0bD9HVv5PmvVTF6ZOnaqSNnnyZKZMmfKvjx0eHo6tra1Kmq2tLXFxcTx9+hR7e/sMHUeCUSGEEEIIIYTQAAkkZNmxxo8fz4gRI1TS9PWz7gKLQqFQeZ589+en6WmRYFQIIYQQQggh/mP09fWzNPj8mJ2dHeHh4SppERER6OjoYG1tneHjSDAqhBBCCCGEEBogPjHrRkazM9Bzdnbmzz//VEnbv38/Tk5O6OpmfJq/LGAkhBBCCCGEEBoggcQse2TG69evCQwMJDAwEEj66ZbAwEBCQ0OBpCm/3bt3V+YfMGAA9+7dY8SIEQQFBbFq1SpWrlzJqFGjMnVeGRkVQgghhBBCiFzs/Pnz1KtXT/k8+V7THj16sHr1asLCwpSBKUCRIkXYvXs3w4cP58cffyRfvnz88MMPtGvXLlPnlWBUCCGEEEIIITRAVi5glBlubm7KBYjUWb16dYo0V1dXLly48K/OK8GoEEIIIYQQQmiA+DQCwv8iCUaFEEIIIYQQQgNk9l7PL50sYCSEEEIIIYQQ4rOTkVEhhBBCCCGE0ADxuWxkVIJRIYQQQgghhNAAMk1XCCGEEEIIIYTIZjIyKoQQQgghhBAaQFbTFUIIIYQQQgjx2eXMr4zmHJmmK4QQQgghhBDis5ORUSGEEEIIIYTQALKarhBCCCGEEEKIzy4+d8WiMk1XCCGEEEIIIcTnJyOjQgghhBBCCKEBctsCRhKMCiGEEEIIIYQGiEeR00X4rCQYFUIIIYQQQggNkCD3jAohhBBCCCGEENlLRkaFEEIIIYQQQgPktmm6MjL6H+Hn54dCoSAqKiqni6IxHBwcWLBgQU4XI9MUCgU7duzI6WLkWjnV/lOmTKFSpUpp5nFzc8Pb2/uzlCczMlJ2T09PWrdu/VnKI4QQQnyp4lFk2eNLICOjGsTT05OoqCgJRDJp9erVeHt7/2cC8bCwMCwtLbPkWFOmTGHHjh0EBgZmyfGyk0KhYPv27bkiYPmndd22bRu6urrZUyjx2Z27BKs2wrWb8OSZgkUzEmlYN6dLlXVaDmxMh1GtsLa3IOTaA5YO9+Xq8b/U5i1buxR9v+tKwVL50TfS5/G9J+xafoBtC3Yp82jraNN5fBsadXfFJr8V92884udxGzi/L/Az1eifadnOiQ5dnbGyNuVecARL5+/n6qXQdPcrU6EgPkt6EHI3goHdlyvTtbW16NSjDo2aVcAmjxn3Q5+y8sdDnD99Jzur8a999VUVOnrUwNrahJCQJyz58SBXrjxQm7dixULMm981Rbpnj2Xcv/8cgMIONnh61sXR0Q47Owt+/PEg27aey9Y6ZJWuFSvSt5oTeY2NufXsGdOP+HH+4cNU8+tpazO0Zk1alSmNjZER4a9fs+TMGX67eg0AHS0tBlSvTtuyZbAzMeHu80i+P3aMoyEhn6lG/0yLjtVo36MOVjYm3LvzhJ/m7OHaxXvp7lemUiHm/NyTkDsRDPZYqkz//ueeVHAqkiL/2WM3mTR0fZaWXWQdCUaF0DB2dnY5XQShoaysrD7r+d6/fy/BbzZ68wZKFoc2zWDYxJwuTdZy7ViLgfN7smjwCq6duEHz/o2YuXsCvcsO58n9pynyv41+y+8/7uXu5Xu8jY6lXJ1SDPupH2+jY9m94iAAPWd0okFXF+b3+4nQvx7i5F6JKdtGM6z2BO4EhnzmGmaMa8MyDPB2Z9Gc3Vy7fJ/mravw7fwu9Om8hCePX6a6n5GxPmMmteLi+WAsrYxVtnkOqEcD9/LMn7WT+/ee4lSzGJO/64h3P1/u3AzP7ir9I25upRk0uCE/LNzH1asPaNGyMrO+86BXzxVERKTeDj26LyM6Olb5/MWLGOXfBvq6hIVFcdT/LwYOapit5c9KzUs68k09NyYfOkTAw0d0rlCBVW3b4L56DWGvXqnd54cWzbExMmbcvv3ci4rC2sgIHa0Po14jatemVenSTDhwgDvPn1PXoTBLv/qKDps2cj3iyWeqWea4NC5H/9FN+XHmTq4FhtKsfTVm/Pg1/dou5kn4i1T3MzLRZ9T0tgSeDcbCWvW9MW3EJnR1tZXPzSwMWbJ5EMcOXM22emSHhMQvY0Qzq8g03S/IvXv3aNmyJZaWlhgbG1O2bFl2796tkicgIAAnJyeMjIyoVasWN27cUNm+dOlSihUrhp6eHiVLlmTdunXKbSNHjqRly5bK5wsWLEChULBr14cr0yVLlmTZsmUpypaYmEjx4sWZO3euSvrVq1fR0tLizp2kK7ahoaG0atUKExMTzMzM6NixI48fP1bmVzeVz9vbGzc3N7Vt4ufnR8+ePXnx4gUKhQKFQsGUKVOU22NiYujVqxempqYUKlSI5cuXq+z/8OFDPDw8sLS0xNramlatWhGSzpXE69ev06xZM0xMTLC1taVbt248ffrhy9XevXupU6cOFhYWWFtb06JFC2X9Ad69e8eQIUOwt7fHwMAABwcHZs2apdz+6TTR9Mro5+dH9erVMTY2xsLCgtq1a3Pv3j1Wr17N1KlTuXTpkrJtVq9enaI+V65cQUtLS1mHyMhItLS06NChgzLPrFmzcHZ2BiA+Pp7evXtTpEgRDA0NKVmyJAsXLlTmPXr0KLq6uoSHq34pGjlyJC4uLmrb1MHBAYA2bdqgUCiUzyHtPqtOau3xT47n7OzMuHHjVNKePHmCrq4uR44cAZJezzFjxpA/f36MjY2pUaMGfn5+qR4zrboCrFu3DgcHB8zNzenUqROvPvpy8uk03SVLllCiRAkMDAywtbWlffv2qZ539erVWFhYsGPHDhwdHTEwMKBRo0bcv39fmSd5uu2qVasoWrQo+vr6JCYmpvu+TbZs2TIKFiyIkZERHTp0SHO2QmJiIt9//z1FixbF0NCQihUr8ttvvym3J996sG/fPipXroyhoSH169cnIiKCPXv2ULp0aczMzOjcuTMxMTGpnkeTudQE7z7QWP3b4ovWbngL9q46zJ6Vhwn96yFLh6/myf2ntBzYWG3+O4EhHNl0gnvXH/D43hMObThGwL5LlK9TWpmn4dcubJy1jbN7LhIeHMHOn/Zzfl8g7Ue0VHtMTdCuszN7/7zI3j8ucj/kKT8t2M+TiBe0bOuU5n7e45pzZP9Vgq6mHDls2KQCG9cc59yp24Q/imLntgACztyhfZea2VWNf619h+rs2XOJ3bsvERr6jCU/HiQi4iUtv6qc5n6RkdEqj4SPlhm9cSOM5cuOcORIEO/fx2V3FbJMr6pV+fXKVbZcucqd58+Z4edH2KtXdK1YUW1+FwcHahQoQO/t2zkZGsrDly+5HB7OhUdhyjyty5Rm6dkz+AUHc//FC365dJlj90LoXTXtfpaT2narxb7tF9i7/QL3g5+ybM4enoS/pEWHamnu5/XNV/jtuUzQ5fsptr1++YbIZ6+Vj8o1i/P27XuO7r+WXdXIFrltmq4Eo1+QwYMHExsby9GjR7ly5QqzZ8/GxMREJc+ECRPw8fHh/Pnz6Ojo0KtXL+W27du3M2zYMEaOHMnVq1fp378/PXv2VH6pdnNz49ixYyQkJP3crr+/PzY2Nvj7+wMQHh7OzZs3cXV1TVE2hUJBr1698PX1VUlftWoVdevWpVixYiQmJtK6dWueP3+Ov78/Bw4c4M6dO3h4ePzjNqlVqxYLFizAzMyMsLAwwsLCGDVqlHK7j48PTk5OXLx4kUGDBjFw4ED++itpmlhMTAz16tXDxMSEo0ePcvz4cUxMTGjSpAnv3r1Te76wsDBcXV2pVKkS58+fZ+/evTx+/JiOHTsq80RHRzNixAjOnTvHoUOH0NLSok2bNsp2/eGHH/jjjz/YsmULN27cYP369SkCkmTplTEuLo7WrVvj6urK5cuXOXXqFP369UOhUODh4cHIkSMpW7assm3UtXW5cuWwtrZWvs5Hjx7F2tqao0ePKvP4+fkpX/eEhAQKFCjAli1buH79OpMmTeJ///sfW7ZsAcDFxYWiRYuqBHlxcXGsX7+enj17qq3nuXNJU6t8fX0JCwtTPk+vz34qrfb4J8fr2rUrGzduJDHxwxegzZs3Y2trq2yPnj17cuLECTZt2sTly5fp0KEDTZo04datW5mqK8CdO3fYsWMHO3fuZOfOnfj7+/Pdd9+pPc758+fx8vJi2rRp3Lhxg71796Ya7CeLiYnh22+/Zc2aNZw4cYKXL1/SqVMnlTy3b99my5YtbN26VTm9OyPv2+T9/vzzT/bu3UtgYCCDBw9OtSzffPMNvr6+LF26lGvXrjF8+HC+/vprZT9MNmXKFBYvXszJkye5f/8+HTt2ZMGCBfzyyy/s2rWLAwcOsGjRojTrLT4vHV0dHKsWJWD/JZX0gAOXKetcMkPHKFbJgTK1SnL56Icvkbr6urx7+14l37s37yhXp9S/L3Q20NHRokRJey6cUZ0+G3DmLmXKF0x1v8bNK2Kf35J1K/3VbtfV0+b9O9XgKzY2jrIVC/37QmcDHR0tHB3tOH8+WCU94HwwZcsWSHPfZct7seXXocyZ25lKlTSzfpmhq6VFOVtbjt9TnYp6/N49quTLp3afBsWKcuXxY/pVc+JEv34c7NmT8a4u6Ot8mNyop61NbFy8yn5v4+Jwyq/+mDlNR0ebEqXtuXBK9b1x4fRtSqfRjxu1qky+glasX+aXofO4t66C/76rxH7yuSE0i0zT/YKEhobSrl07ypcvD0DRokVT5Pn222+VX5LHjRtH8+bNefv2LQYGBsydOxdPT08GDRoEwIgRIzh9+jRz586lXr16uLi48OrVKy5evEiVKlU4duwYo0aNYtu2bQAcOXIEW1tbSpVS/4+/Z8+eTJo0ibNnz1K9enXev3/P+vXrmTNnDgAHDx7k8uXLBAcHU7Bg0j/idevWUbZsWc6dO0e1amlfDVNHT08Pc3NzFAqF2umtzZo1U9Z37NixzJ8/Hz8/P0qVKsWmTZvQ0tLi559/VgYrvr6+WFhY4OfnR+PGKa/gL126lCpVqjBz5kxl2qpVqyhYsCA3b97E0dGRdu3aqeyzcuVK8ubNy/Xr1ylXrhyhoaGUKFGCOnXqoFAoKFy4cKr1S6+MTk5OvHjxghYtWlCsWDEASpf+MJJgYmKCjo5OmlN/FQoFLi4u+Pn50a5dO/z8/OjRowdr1qzh+vXrODo6cvLkSYYPHw6Arq4uU6dOVe5fpEgRTp48yZYtW5RBee/evfH19WX06NEA7Nq1i5iYGJWg/WN58uQBwMLCQqWs6fXZT718+TLN9sjs8Tw8PBg+fDjHjx+nbt2km/l++eUXunTpohzx37hxIw8ePCDf318kRo0axd69e/H19VXpJ+nVFZIC/dWrV2NqagpAt27dOHToEN9++22K44SGhmJsbEyLFi0wNTWlcOHCVK6c9ijD+/fvWbx4MTVq1ABgzZo1lC5dWvmehaSR3nXr1inLeeDAgQy9b9++fcuaNWsoUCDpy+WiRYto3rw5Pj4+KeoZHR3NvHnzOHz4sHLEvWjRohw/fpxly5apXPCaMWMGtWvXBpL61fjx47lz547y8699+/YcOXKEsWPHpll38fmY25iiraNN5OMolfTIx1FY2lmkue8voT9hnscMbR1t1k3dwp6Vh5Xbzu+7RLvhLbhy9DqP7jymcoPyOLeqhpa2Zl5XN7MwQltHi8jn0Srpkc+jsfxkemGyfAWt6D24ASP6ryYhXv2PDZ4/fYe2nWtyOTCUsAfPqVytKM4uJdHS0sxREHNzI7S1tYiM/KQdIqOxslLfDs+ev8bHZze3boajq6tDw0blmDO3CyNGbOCKmhGxL4WloSE6Wlo8jVFti6fRMeRxMFK7TyFzC5zy5yc2Lp6Bf/yBlaEhUxvUx9zAgHH79gNwLOQevapW4dyDB9yLiqJW4UI0LFYMLYVm9gkzS6Okz4jnr1XSI59FY2VjonaffIWs6OXViFE9V5IQn5DuORzL5adICVvmT92RFUX+rOJz2Vhh7qrtF87Ly0v5xWzy5Mlcvnw5RZ4KFSoo/7a3twcgIiICgKCgIOWXumS1a9cmKCgIAHNzcypVqoSfn59y6mb//v25dOkSr169UhkdU8fe3p7mzZuzatUqAHbu3Mnbt2+V0z2DgoIoWLCg8gstQJkyZbCwsFCWIat93B7JAWtyewQEBHD79m1MTU0xMTHBxMQEKysr3r59qzKt9mMBAQEcOXJEmd/ExEQZnCfvc+fOHbp06ULRokUxMzOjSJGkm+lDQ5MWrPD09CQwMJCSJUvi5eXF/v37Uy1/emW0srLC09MTd3d3WrZsycKFCwkLC0v1eKlxc3NTTi319/dXXpzw9/fn3LlzvHnzRqXv/PTTTzg5OZEnTx5MTExYsWKFsn7Jdbx9+zanT58GkgL2jh07Ymys/otHatLrs59Krz0ye7w8efLQqFEjNmzYAEBwcDCnTp2ia9ekhTUuXLhAYmIijo6OKn3C398/1T6UFgcHB2UgCknvqeT++qlGjRpRuHBhihYtSrdu3diwYUO601V1dHRwcvowbatUqVIp3n+FCxdWBqKQ8fdtoUKFlIEoJE1xTkhISHGrACRNdX/79i2NGjVSabe1a9emaLeP38O2trYYGRmpXIiztbVNtY3UiY2N5eXLlyqP2Nj0v9iIzEv8JJZSKBQqswzUGeEyicHVxrFw4HLaDmtOvU4f3q9LvH15eCuMlUEL2RO7kSGLerN/9ZEMfTHNSSnbAVDTDFpaCsZPbcPaFf48/HuRHnWWzt/Ho/vPWblpELuPfcPgkU3YvzNQZQqrRvq0eApFirZJ9uD+c3bvusStW4+5fv0hPyzcx5kzt+nYsUa2F/NzUNcnUnv1FIqk/MN37+ZyeDh+wcF86+dPu7JllaOj048c4V5UFPt7evLXcG+m1K/Pb9eukZDO+y3HqWsHNWXW0lIwblYH1i09zMPQZxk6dJPWVQi+9ZibV1NfGEpTJSQqsuzxJZCR0S9Inz59cHd3Z9euXezfv59Zs2bh4+PD0KFDlXk+XmwkeSQteXrox2nJEhMTVdKSgxI9PT1cXV2xtLSkbNmynDhxAj8/v3R/VqJPnz5069aN+fPn4+vri4eHB0ZGRmrPpa4MWlpaKT6I3r//59MrPl18RaFQKNsjISGBqlWrKgONj338ZfxjCQkJtGzZktmzZ6fYlhz8t2zZkoIFC7JixQry5ctHQkIC5cqVU079rVKlCsHBwezZs4eDBw/SsWNHGjZsqHK/3MfnS6+Mvr6+eHl5sXfvXjZv3sw333zDgQMHqFkz4/cPubm5MWzYMG7fvs3Vq1epW7cud+7cwd/fn6ioKKpWraoMkrZs2cLw4cPx8fHB2dkZU1NT5syZw5kzZ5THy5s3Ly1btsTX15eiRYuye/fuNO+jTEt6ffZT6bVHZo/XtWtXhg0bxqJFi/jll18oW7YsFf++tychIQFtbW0CAgLQ1tZW2e/TKfQZkVZ//ZSpqSkXLlzAz8+P/fv3M2nSJKZMmcK5c+ewsLBI9Rzq6vpx2qcXDDLyvk3rPOryJNdp165d5M+fX2Wbvr6+yvNPP9My00bqzJo1S2VkH2DSSCsmj7LO8DFE2l48fUV8XDxWn4yCWuQ1J+px6guTAISHJF1YCLkaiqWtOd0md+TIphN/H/clU9rOQVdfFzNrU549ek6f77oSHpzxixGf08uoGOLjErD6ZBTUwtI4xWgpgKGRHiXL5Ke4oz1DRjYFQKGlQEtLwZ7j3zB+2HoCA0J4ERXDlLFb0NXTxszciGdPXtF7cAPCH0V+lnpl1osXMcTHJ6RYiMnSwijFaGlarl9/RMOGZbO6eJ9V5Js3xCUkkOeTz1lrIyOeRqu/mPgkOprHr1/z+qPbh+48f46WQoG9iQkhUVE8f/OGAb//gZ62NpaGhjx+/Zoxdety/0Xa77ec8jIyhvi4eCytVf9PWlgZE/lMzXvDWB/HsvkpVtKOweOaA8nvDS12nZ/M/wau5dK5D9PA9Q10cXUvz9qlh1McS2geCUa/MAULFmTAgAEMGDCA8ePHs2LFCpVgNC2lS5fm+PHjdO/eXZl28uRJlWmMbm5urFy5Eh0dHRo2TFqdztXVlU2bNqV6v+jHmjVrhrGxMUuXLmXPnj0q9x2WKVOG0NBQ7t+/rxxluX79Oi9evFCWIU+ePFy9qrrqWWBgYJoreurp6REfH5/q9tRUqVKFzZs3kzdvXszMzDK8z9atW3FwcEBHJ+Xb59mzZwQFBbFs2TLltM7jx4+nyGdmZoaHhwceHh60b9+eJk2a8Pz58xSrpWa0jJUrV6Zy5cqMHz8eZ2dnfvnlF2rWrJnhtkm+b3TGjBlUrFgRMzMzXF1dmTVrFpGRkSqv+7Fjx6hVq5ZyqiugdhSwT58+dOrUiQIFClCsWLEUI5Kf0tXVTVHWjPRZdVJrj39yvNatW9O/f3/27t3LL7/8Qrdu3VTOEx8fT0REhPL1zgh1df0nkt+nDRs2ZPLkyVhYWHD48GHatm2rNn9cXBznz59XTsm9ceMGUVFRqU69h4y9byFp5P/Ro0fK6cqnTp1CS0sLR0dHtcfU19cnNDQ03c+UrDZ+/HhGjBihkqYbWeWzluG/Lu59HDcD7lKlUQVO7DirTK/SsAIn/8j4T28oFAp09VN+zr6Pfc+zR8/R1tGmTtuaHP31ZJaUO6vFxSVw60YYVaoX5YT/hxkCVaoX5dTRlDMGYqJj6ddlqUpay3ZOVKpahOn/+5XwR1Eq296/i+fZk1doa2tRx600Rw9dz5Z6/FtxcQncvBlO1apFOHH8pjK9atUinDh5M409VZUobsvzT6Z1fmneJyRw9fFjahcuxP7bt5XptQsX5uDtVGZkPXxEU0dHjHR1ifn74nwRS0viExIIe63aHu/i43n8+jU6Wlo0KVGCXTdT9jNNEBcXz62gMCo7F+PkkQ8zbCrXKMZpv5Q//xTzOpb+7RarpLXwqE6lakWYMWoz4Q9VL8S4NC6Lrp42h3ep3rf+pfhSFh7KKhKMfkG8vb1p2rQpjo6OREZGcvjw4XS/lH9s9OjRdOzYkSpVqtCgQQP+/PNPtm3bxsGDB5V5ku8b/fPPP5kxYwaQFKC2a9eOPHnyUKZMmTTPoa2tjaenJ+PHj6d48eLK+8EAGjZsSIUKFejatSsLFiwgLi6OQYMG4erqqpw6WL9+febMmcPatWtxdnZm/fr1XL16Nc174RwcHHj9+jWHDh2iYsWKGBkZKUdj09K1a1fmzJlDq1atmDZtGgUKFCA0NJRt27YxevRolSmHyQYPHsyKFSvo3Lkzo0ePxsbGhtu3b7Np0yZWrFihXPF2+fLl2NvbExoammI11vnz52Nvb0+lSpXQ0tLi119/xc7OTu1oVnplfP/+PcuXL+err74iX7583Lhxg5s3byqDLQcHB4KDgwkMDKRAgQKYmpqmGHmCD/eNrl+/XnlvaIUKFXj37h2HDh1i2LBhyrzFixdn7dq17Nu3jyJFirBu3TrOnTunnI6czN3dHXNzc2bMmMG0adPSfT0cHBw4dOgQtWvXRl9fH0tLywz12Y8FBwen2R6ZPR4kjRS2atWKiRMnEhQURJcuXZTbHB0d6dq1K927d8fHx4fKlSvz9OlTDh8+TPny5WnWrFmG65pZO3fu5O7du7i4uGBpacnu3btJSEigZMnUF4jR1dVl6NCh/PDDD+jq6jJkyBBq1qypDE7Vycj7FsDAwIAePXowd+5cXr58iZeXFx07dlR7v7KpqSmjRo1i+PDhJCQkUKdOHV6+fMnJkycxMTGhR48emW6PZIsXL2b79u0cOnRI7XZ9ff0U74GEmJy5YyU6BkI/mkH2IAyCboG5GeSzzZEiZZmt83cydu1Qbp6/Q9CpmzTr15C8hWzY+VPSbQm9ZnbBJp8V33smfcH8apA7EaFPuf9XUoOUq1OKDiO/YsfiPcpjlqpeHJv8VtwODMEmvxXdJ3dES0vB5u9///wVzKCtG08xZnIbbgaFcf3qA5q3qkJeW3N2bg8AoNfA+ljnMWXOtN9JTISQu6o/wxEVGcO7d3Eq6aXK5sc6jyl3boZjk8eMbn1c0dJSsGX9ic9at8z47dezjBvfkps3wrh+/SHNW1Qir60Zf/55EYDefVyxsTFl9nc7AWjbrhqPw6MICXmKjq42DRuWxcW1FJMnb1UeU0dHi8KFbf7+WxsbGxOKFcvLmzfveaSho8QAqwICmNu0KVceP+biozA6VShPPlNTfrmUFDiNqlMHOxMTRu3dC8Aff/3FEOeazHZ3Z+HJk1gaGjLOxYXfrl4jNi5pIauKdnbYmpgQ9OQJtiYmDHN2RqGA5efO51g907Nt3UlGf9uWW9ceEnT5Pk3bOZHX3pxdvyVdsOo5tCHWec2YO3EbiYmJ3LujOgPixfNo3r2LS5EO4N66KieP/MWrF28+S12yWnxi7rqLUoLRL0h8fDyDBw/mwYMHmJmZ0aRJE+bPn5/h/Vu3bs3ChQuZM2cOXl5eFClSBF9fX5WfTTE3N6dy5cqEhoYqA8+6deuSkJCQ4RGM3r17M3PmTJWVfOHDT5YMHToUFxcXtLS0aNKkicpKmO7u7kycOJExY8bw9u1bevXqRffu3bly5Uqq56tVqxYDBgzAw8ODZ8+eMXnyZJWfd0mNkZERR48eZezYsbRt25ZXr16RP39+GjRokOooZL58+Thx4gRjx47F3d2d2NhYChcuTJMmTdDS0kKhULBp0ya8vLwoV64cJUuW5IcfflBpYxMTE2bPns2tW7fQ1tamWrVq7N69Gy2tlB8+6ZXxzZs3/PXXX6xZs4Znz55hb2/PkCFD6N+/PwDt2rVj27Zt1KtXj6ioKHx9ffH09FRbt3r16rFt2zZlWRUKBXXr1mXnzp3UqVNHmW/AgAEEBgbi4eGBQqGgc+fODBo0iD179qgcT0tLC09PT2bOnKkyEpkaHx8fRowYwYoVK8ifPz8hISEZ6rOftlda7ZHZ4yXr2rUrzZs3x8XFhUKFVFf68/X1ZcaMGYwcOZKHDx9ibW2Ns7NzqoFoanXNLAsLC7Zt28aUKVN4+/YtJUqUYOPGjZQtm/o0NiMjI8aOHUuXLl148OABderUUd7jnZqMvG8h6SJF27ZtadasGc+fP6dZs2YsWbIk1eNOnz6dvHnzMmvWLO7evYuFhQVVqlThf//7X+Ya4hNPnz79R/fr5oRrN6CH94cr4LN/TPq7dZNEZo3PqVJlDf8tJzGzNuHrie2xsrck5Op9JjSfSURo0k9IWdtZkreQjTK/QktBr5ldsCuSl4S4BB7dCefn8RvYteyAMo+egR6e0ztjXzQvb16/5ezui8zuvojoF5r70z7+B69jZm5E194uWFmbcO9uBN+M+IWIv39H0crGhLx25pk6pq6eDp7962Gfz5I3b95x9uQtZk/dTvTr2PR3ziF+fkGYmRnSrXttrKxMCAl5wvjxW4j4+7dWra1MyJv3w/9dXR0t+g9ogI2NCbGxcdwLecr48Vs4+9HKxNbWpixf0Vv53MOjJh4eNQkMvMfIEb98vspl0q4bN7EwMGRozZrkMTbm1rNn9N62nUd//4xXXmNj7M0+rB0Q8/493X/7jcn167Pj665EvX3Lrhs3mHfiw4wAfR0dRtSpTSFzc6Lfv8f/bjAj9+zhVazm9omj+69iZmFI1/5uWNqYcu92BBOHrCci7O/3Rh5T8tpn7r0BkL+QNeWqFGb8gDVZXeTPJiGXLemjSExvNQEhMunEiRO4ubnx4MEDbG2/8Mv74l/p27cvjx8/5o8//sjpouR6q1evxtvbO83f/sytEsJTTiXOjdzzqf+dw9xGUb18ThdBI8QbpX57TG4S0jzlbKLcqMS6qJwugkbYG5j+TK9/yz8k6/4nuTpkfCp8TpGRUZFlYmNjuX//PhMnTqRjx44SiOZiL1684Ny5c2zYsIHff9fcKXRCCCGEEJokt90zmrvGgUW22rhxIyVLluTFixd8//33OV0ckYNatWrFV199Rf/+/WnUqFFOF0cIIYQQ4osQn6iVZY8vgUzTFUIIkavJNN0kMk03iUzTTSLTdJPINN0kMk03yeeYpnsgOOOLk6anURH1v6OuSWSarhBCCCGEEEJogIRcNk1XglEhhBBCCCGE0ADxuewuytxVWyGEEEIIIYQQGkFGRoUQQgghhBBCA3wpCw9lFQlGhRBCCCGEEEIDJOSyiau5q7ZCCCGEEEIIITSCjIwKIYQQQgghhAaIT5TVdIUQQgghhBBCfGa5bTVdCUaFEEIIIYQQQgMk5LIFjHJXbYUQQgghhBBCaAQZGRVCCCGEEEIIDSDTdIUQQgghhBBCfHa5bQGj3BV6CyGEEEIIIYTQCDIyKoQQQgghhBAaICGXjRVKMCqEEEIIIYQQGiBeVtMVQgghhBBCCCGyl4yMCiGEEEIIIYQGSCB3LWAkwagQQgghhBBCaACZpiuEEEIIIYQQIldZsmQJRYoUwcDAgKpVq3Ls2LE08//444+ULl0aQ0NDSpYsydq1azN9ThkZFUIIIYQQQggNEJ9DY4WbN2/G29ubJUuWULt2bZYtW0bTpk25fv06hQoVSpF/6dKljB8/nhUrVlCtWjXOnj1L3759sbS0pGXLlhk+r4yMCiGEEEIIIYQGSEhUZNkjNjaWly9fqjxiY2PVnnfevHn07t2bPn36ULp0aRYsWEDBggVZunSp2vzr1q2jf//+eHh4ULRoUTp16kTv3r2ZPXt2puorwagQQgghhBBCaIB4tLLsMWvWLMzNzVUes2bNSnHOd+/eERAQQOPGjVXSGzduzMmTJ9WWMzY2FgMDA5U0Q0NDzp49y/v37zNcX5mmK4QQIldzz1cxp4ugEfY9upTTRdAI7u2kPwA8cDNIP1MuUHzTy5wugkYIr2OZ00UQ/8D48eMZMWKESpq+vn6KfE+fPiU+Ph5bW1uVdFtbW8LDw9Ue293dnZ9//pnWrVtTpUoVAgICWLVqFe/fv+fp06fY29tnqIwSjAohhBBCCCGEBkjIwtV09fX11QafqVEoVH9WJjExMUVasokTJxIeHk7NmjVJTEzE1tYWT09Pvv/+e7S1tTN8TpmmK4QQQgghhBAaIB5Flj0yysbGBm1t7RSjoBERESlGS5MZGhqyatUqYmJiCAkJITQ0FAcHB0xNTbGxscnwuSUYFUIIIYQQQohcSk9Pj6pVq3LgwAGV9AMHDlCrVq0099XV1aVAgQJoa2uzadMmWrRogZZWxkNMmaYrhBBCCCGEEBogK6fpZsaIESPo1q0bTk5OODs7s3z5ckJDQxkwYACQdP/pw4cPlb8levPmTc6ePUuNGjWIjIxk3rx5XL16lTVr1mTqvBKMCiGEEEIIIYQGyMz02qzk4eHBs2fPmDZtGmFhYZQrV47du3dTuHBhAMLCwggNDf1Qzvh4fHx8uHHjBrq6utSrV4+TJ0/i4OCQqfNKMCqEEEIIIYQQudygQYMYNGiQ2m2rV69WeV66dGkuXrz4r88pwagQQgghhBBCaICcmqabUyQYFUIIIYQQQggNEJ/LgtHcVVshhBBCCCGEEBpBRkaFEEIIIYQQQgMk5NACRjlFglEhhBBCCCGE0AC5bZquBKNCCCGEEEIIoQESEnPXyGjuCr2FEEIIIYQQQmgEGRkVQgghhBBCCA0Qn8vGCiUYFUIIIYQQQggNINN0hRBCCCGEEEKIbCYjo0IIIYQQQgihARJy2VihBKNCCCGEEEIIoQHiZZquEEIIIYQQQgiRvWRkVAghhBBCCCE0QG5bwEiCUSGEEEIIIYTQAAmJuWviau6qrRBCCCGEEEIIjSAjo0IIIYQQQgihAeKRabpCCCGEEEIIIT6z3HbPqEzTTUVISAgKhYLAwMCcLoqSJpZJU/n5+aFQKIiKivrHx3Bzc8Pb2zvNPA4ODixYsOAfnyO7KBQKduzYAaTsN1nRNp+Dp6cnrVu3zvbjfvo6x8TE0K5dO8zMzJTtpC4tq1/77KhvRvpwdpoyZQqVKlXKsfMLIYQQX5qERK0se3wJcuXIqEKR9hWHHj16MGXKlM9TGCGyQVhYGJaWlmq31apVi7CwMMzNzT9zqTJn4cKFJCYmZvt5tm3bhq6urvL5mjVrOHbsGCdPnsTGxgZzc3N++umnFGnnzp3D2Ng428rl5uZGpUqV/lXA+2ndHBwc8Pb2zpYAVaFQsH37dpWAetSoUQwdOjTLz6XpWg5sTIdRrbC2tyDk2gOWDvfl6vG/1OYtW7sUfb/rSsFS+dE30ufxvSfsWn6AbQt2KfNo62jTeXwbGnV3xSa/FfdvPOLncRs4vy/wM9Uoe527BKs2wrWb8OSZgkUzEmlYN6dLlXVatqlKhy7OWFubEBL8hKU/7Ofqpfvp7le2fAF8FncnJDiCAZ4/K9O1tbXo3L02jZpWwMbGlPuhz/h56SHOn7mbndX417o4VaC3sxN5TI25FfGMmfv9CQh9mGp+XW1tBrvU4KvypcljYkT4y9f8dPwsWwOvAdCmYhm+a+WeYr/y3/7Au/j4bKtHVmjR3okOX9fCysaUe3cj+GnePq4Ghqa7X5kKBZm7zJOQuxEM6rpMZVubzjVo3s6JvLbmvHwRw7FDQaz68SDv32l2W3ysY90K9GjghI25MXfCnjFnqz8X76TRR3S06d+0Bs2qlcbG1IjHUa/5ed9Zfj997TOWWvxbuTIYDQsLU/69efNmJk2axI0bN5RphoaGREZG5kTRvkjv379X+cKb096/f5/TRfhXsqI97ezsUt2mp6eX5nZN8bmCZSsrK5Xnd+7coXTp0pQrVy7NtDx58nyW8v0bn9btczMxMcHExCRHy/C5uXasxcD5PVk0eAXXTtygef9GzNw9gd5lh/Pk/tMU+d9Gv+X3H/dy9/I93kbHUq5OKYb91I+30bHsXnEQgJ4zOtGgqwvz+/1E6F8PcXKvxJRtoxlWewJ3AkM+cw2z3ps3ULI4tGkGwybmdGmylmuDMgwc1phFPnu4dvk+zVtXYebczvT++ieePH6Z6n5GxvqMmdiKiwHBWFqpXvTq2c+NBu7lmD97F6H3nuFUvShTZnVgWP/V3Ln1OLur9I80LePIeHc3pu4+zIX7j+hUpTwrurSm+ZK1hL18pXafhe2bY21sxIQ/DxD6PAorY0N0tFRHel69jaXJj6tV0jQ9EHVtVJYBI5qwePYurl26T/O2VZmxsCt9O/6Ybp8YPbU1F8/dxdJa9XO1XpPy9BrckHnTf+f65fvkL2TNqMmtAVg2f192VifLNK7iyOh2bszcfJjAu49oX6c8Pw5qTdsZawmPVN9Hvu/VHGtTI6ZuOMD9J1FYmRqirfVljAamJSGX3TP65b9i/4CdnZ3yYW5ujkKhSJGW7O7du9SrVw8jIyMqVqzIqVOnVI518uRJXFxcMDQ0pGDBgnh5eREdHQ3AtGnTKF++fIrzV61alUmTJqktW2RkJF27diVPnjwYGhpSokQJfH19VfKkV6atW7dStmxZ9PX1cXBwwMfHR7lt0aJFKmXasWMHCoWCH3/8UZnm7u7O+PHj1ZYvecrnli1bcHNzw8DAgPXr1wPg6+tL6dKlMTAwoFSpUixZskTtfnXr1sXQ0JBq1apx8+ZNzp07h5OTEyYmJjRp0oQnT54o9zt37hyNGjVSjka5urpy4cIFlTIpFAp++uknWrVqhbGxMTNmzEhR7jdv3tC8eXNq1qzJ8+fPefbsGZ07d6ZAgQIYGRlRvnx5Nm7cmGK/uLg4hgwZgoWFBdbW1nzzzTdpjta9ePGCfv36kTdvXszMzKhfvz6XLl1KNf8/bc93794xZMgQ7O3tMTAwwMHBgVmzZqm0SfI03U+pm6abVj+GpBG1mTNn0qtXL0xNTSlUqBDLly9XOe6DBw/o1KkTVlZWGBsb4+TkxJkzZ5Tb//zzT6pWrYqBgQFFixZl6tSpxMXFpdo26qbTenl5MWbMGKysrLCzs0t3BkN8fDwjRoxQvn5jxoxJ8fp9PJXVzc0NHx8fjh49ikKhwM3NTW1acpt8PGoZFRVFv379sLW1xcDAgHLlyrFz505A/XTVBQsW4ODgkGrd/f39WbhwIQqFAoVCQUhIiNq8S5YsoUSJEhgYGGBra0v79u1Trdu9e/cYPny48pgZLVt678PkvG3atEGhUCiff3rshIQEpk2bRoECBdDX16dSpUrs3btXuT35/bBt27ZUP+Pu3btHy5YtsbS0xNjYmLJly7J79261bZMT2g1vwd5Vh9mz8jChfz1k6fDVPLn/lJYDG6vNfycwhCObTnDv+gMe33vCoQ3HCNh3ifJ1SivzNPzahY2ztnF2z0XCgyPY+dN+zu8LpP2Ilp+rWtnKpSZ494HGLjldkqzXzqMGe3cGsufPQELvPWPpwgM8iXhJyzZV09zPe0wzDh+4yvWrKUeFGjYpz8a1Jzh76g7hj6LYueMC58/cpX3nmtlVjX+tp3MVtl68ym8Xr3L36XNm7vcn/MUrOjtVUJu/brHCVCucn36/bOdUcCgPX7zkyqPHXHwQppIvkUSeRseoPDRd2y412ff7Rfb+fpH7IU/5ad4+njx+QYv21dLcb9j/WnBk31WCrjxIsa1M+QJcuxzKkX1XeRz2ggtn7uK3/yqOpe2zqxpZrlv9Kmw/dZXtp64S/Pg5c7b6Ex75ig511feRWqUL41Q8P0OWbufMjVAePX/J1XuPuRQcpjb/lyQ+UZFljy9BrgxGM2PChAmMGjWKwMBAHB0d6dy5s/IL9JUrV3B3d6dt27ZcvnyZzZs3c/z4cYYMGQJAr169uH79OufOnVMe7/Lly1y8eBFPT0+155s4cSLXr19nz549BAUFsXTpUmxsbDJcpoCAADp27EinTp24cuUKU6ZMYeLEiaxevRpI+kJ67do1nj5NukLv7++PjY0N/v7+QFLwdfLkSVxdXdNsl7Fjx+Ll5UVQUBDu7u6sWLGCCRMm8O233xIUFMTMmTOZOHEia9asUdlv8uTJfPPNN1y4cAEdHR06d+7MmDFjWLhwIceOHePOnTsqgfqrV6/o0aMHx44d4/Tp05QoUYJmzZrx6tWrFMdt1aoVV65coVevXirbXrx4QePGjXn37h2HDh3CysqKt2/fUrVqVXbu3MnVq1fp168f3bp1UwmeIGnKpo6ODmfOnOGHH35g/vz5/Pzzz6iTmJhI8+bNCQ8PZ/fu3QQEBFClShUaNGjA8+fPs7Q9f/jhB/744w+2bNnCjRs3WL9+faqBTXrS68fJfHx8cHJy4uLFiwwaNIiBAwfy119JUw9fv36Nq6srjx494o8//uDSpUuMGTOGhIQEAPbt28fXX3+Nl5cX169fZ9myZaxevZpvv/02U2Vds2YNxsbGnDlzhu+//55p06Zx4MCBVPP7+PiwatUqVq5cyfHjx3n+/Dnbt29PNf+2bdvo27cvzs7OhIWFsW3bNrVpn0pISKBp06acPHmS9evXc/36db777ju0tbUzVb9kCxcuxNnZmb59+xIWFkZYWBgFCxZMke/8+fN4eXkxbdo0bty4wd69e3FxUf+Nftu2bRQoUIBp06Ypj5lR6b0Pkz/jfH19CQsLU/nM+7RePj4+zJ07l8uXL+Pu7s5XX33FrVu3VPKl9Rk3ePBgYmNjOXr0KFeuXGH27NkaM/qqo6uDY9WiBOxXvQAVcOAyZZ1LZugYxSo5UKZWSS4f/TDNTFdfl3dvVWd8vHvzjnJ1Sv37Qotso6OjhWNJewLOqk6fDTh7l7LlCqS6n3uziuTLb8m6VUfVbtfV1ebdJ1Mv38XGUa5Cys8ITaCrpUVZe1uO37mnkn7ibiiVC+ZTu099x2JcfRRBn9rVOOrdl72DPRnTqC76OqqfqUZ6ehz26o2/dx9+6tSK0naaPWNFR0eLEqXyEXDmjkp6wJm7lKmQep9o3LIS9gUsWb/CT+32q4GhlCiVj5JlktrTLr8F1WoV58yJW2rzaxodbS1KF7TlVJBqHzkdFErFIur7iFv5YlwLjcCzYTX2z+jL75M8Gd6mLvq6/+z/rsg5uXKabmaMGjWK5s2bAzB16lTKli3L7du3KVWqFHPmzKFLly7KkYcSJUrwww8/4OrqytKlSylQoADu7u74+vpSrVrSFS9fX19cXV0pWrSo2vOFhoZSuXJlnJycANQGGGmVad68eTRo0ICJE5PmOjk6OnL9+nXmzJmDp6cn5cqVw9raGn9/f9q1a4efnx8jR45k/vz5QNKXyrdv31KnTp0028Xb25u2bdsqn0+fPh0fHx9lWpEiRZRBR48ePVTK7u6edI/HsGHD6Ny5M4cOHaJ27doA9O7dWxk4A9SvX1/lvMuWLcPS0hJ/f39atGihTO/SpYtKEBocHAzA48eP8fDwoFixYmzcuBE9PT0A8ufPz6hRo5T5hw4dyt69e/n111+pUaOGMr1gwYLMnz8fhUJByZIluXLlCvPnz6dv374p2uTIkSNcuXKFiIgI9PX1AZg7dy47duzgt99+o1+/flnWnqGhoZQoUYI6deqgUCgoXLhwqsdOT3r92MDAAIBmzZoxaNAgICl4nj9/Pn5+fpQqVYpffvmFJ0+ecO7cOeXU0OLFiyvP8e233zJu3DhlXyhatCjTp09nzJgxTJ48OcNlrVChgjJ/iRIlWLx4MYcOHaJRo0Zq8y9YsIDx48fTrl07AH766Sf27Ut9ypKVlRVGRkYppjKrS/vYwYMHOXv2LEFBQTg6Oirr+E+Zm5ujp6eHkZFRmlOqQ0NDMTY2pkWLFpiamlK4cGEqV66cat20tbUxNTXN9DTt9N6HyVOWLSws0jz23LlzGTt2LJ06dQJg9uzZHDlyhAULFqjMzkjrMy40NJR27dopZ3hktp1jY2OJjY1VSUtIjEdL8e+/wJjbmKKto03k4yiV9MjHUVjaWaS57y+hP2GexwxtHW3WTd3CnpWHldvO77tEu+EtuHL0Oo/uPKZyg/I4t6qGlrZcT9Zk5hZGaOtoEfk8WiU9MjI6xTTLZPkLWNJ7YD2GD1pLQrz6WTjnz9ylXacaXAm8x6OHkVR2KoJzXUe0tDRzFMTSKGl67bNPRi2fRkeTx1j9/66CluZULZSP2Lg4Bm/5A0sjQyY3q4+FgQH/+zPpAuTdp88Z//s+bkQ8xURfn+7VK7Oxpwetlq3n3vOo7K7WP2L2d5+Iev5aJT3q2WssrYup3SdfQSt6DW7AyH6+qfYJ/wPXMLc0xufnXigUoKOjzZ+/nWPLmhNZXofsYGliiI62Fs9fqfaRZ6+isTFT30fy25hTuVg+3sXFMWLFH1gYG/I/j/qYGxkwZUPqF6m/BF/KwkNZJXfV9h+oUOHD9AB7+6TpDhEREUDSKOTq1auV90WZmJjg7u5OQkKCMhjq27cvGzdu5O3bt7x//54NGzakGLn72MCBA9m0aROVKlVizJgxnDx5MlNlCgoKUgZ2yWrXrs2tW7eIj49HoVDg4uKCn58fUVFRXLt2jQEDBhAfH09QUBB+fn5UqVIl3ZGG5GAZ4MmTJ9y/f5/evXurtMWMGTO4c0f16t/HZbe1tQVQmTZsa2urrEtyvQYMGICjoyPm5uaYm5vz+vVrQkNVb/T/uDwfa9iwIUWLFmXLli3KQBSSpm9+++23VKhQAWtra0xMTNi/f3+K49asWVNlwStnZ2dlW34qICCA169fK4+X/AgODk7RDp/KbHt6enoSGBhIyZIl8fLyYv/+/WkePy0Z6ceg+tolT21Pfq0CAwOpXLlyqvcoBgQEMG3aNJVzJI/6xcRkfFrVx2WApP7/cX/52IsXLwgLC8PZ2VmZpqOjk2pf+TcCAwMpUKCAMhD9XBo1akThwoUpWrQo3bp1Y8OGDZlqz4zK6PswLS9fvuTRo0dqP5+CgoJU0tL6jPPy8mLGjBnUrl2byZMnc/ny5UzVZdasWco6JD+CUb+40D/16Ux+hUKR7mJcI1wmMbjaOBYOXE7bYc2p1+lDOy3x9uXhrTBWBi1kT+xGhizqzf7VR0iIT8jScovs8elrr1CTBqClpWD8lDasXXmUh/dTn02zZOF+Ht5/zspfBrLH738MGdGE/bsukZCQ/Qu+/Ruflk6BIkWacptCQWIijNq+hyuPHnP0dgjf7T9Km0pllaOjlx6G88eVv7jx+CkBoQ/x/m0nIc8i+bpapeysRpZQ9xmhrjG0tBSMm9GWdcv9eBiaep+oUKUwnXvVZfHsXQz+ejlTR2+mRh1HuvT+sua+p+gjf/cDdbT+3va/1Xu4eu8xx6+HMHfbUb6qUfaLHx1NSFRk2eNLICOj6fh4IZnkoCR56mFCQgL9+/fHy8srxX6FChUCoGXLlujr67N9+3b09fWJjY1VjtKo07RpU+7du8euXbs4ePAgDRo0YPDgwcydOzdDZUpMTEyxWrC6e+SWL1/OsWPHqFixIhYWFri4uODv74+fn5/ynri0fLyKaPK5V6xYoTKqCKSYpqiu7J+mJR8PkoKuJ0+esGDBAgoXLoy+vj7Ozs68e/cu1fJ8rHnz5mzdupXr16+rBL0+Pj7Mnz+fBQsWUL58eYyNjfH29k5x3MxISEjA3t4ePz+/FNssLCzS3Dez7VmlShWCg4PZs2cPBw8epGPHjjRs2JDffvvtH5U7vX4MpFhU6ePXytDQMN1zTJ06VWX0N1nyyGtGpFWGnJRe/bW0tFK8D7NioS1TU1MuXLiAn58f+/fvZ9KkSUyZMoVz586l2+cyU7aMvg8zQt3n06dpaX3G9enTB3d3d3bt2sX+/fuZNWsWPj4+GV61d/z48YwYMUIlrY25Z2arodaLp6+Ij4vH6pNRUIu85kQ9fpHmvuEhScF2yNVQLG3N6Ta5I0c2nfj7uC+Z0nYOuvq6mFmb8uzRc/p815XwYPUXYoRmeBEVQ3xcAlafjIJaWBoT9cloKYChkR4lS+ejeAk7hgxvAoBCS4GWloK9/v9j3PBfCLwQwouoGKaM/xVdPW3MzIx49vQVfQbWJzws6nNUK9MiY94Ql5CAjbGRSrq1sVGq93g+eR3N41eveR374TPmztPnaCkU2JmZqh35TASuPHqMg7VFFpY+a738u098OjJubmVM5CejpfB3nyiTn+KO9gwe3Qz40Cd2n5rI+KHruHQ+hB4D6nFo92X2/n4RgJA7ERgY6jLsfy3ZuOpoqgGdpoh8/Ya4+ASsTVX7iJWJEc9eqe8jT19GE/HiNa/ffugjweHP0dJSYGthSuiTqOwssshCEoz+C1WqVOHatWsq0xE/paOjQ48ePfD19UVfX59OnTphZGSUan5IWqXT09MTT09P6taty+jRo1WC0bSUKVOG48ePq6SdPHkSR0dHZSDj5ubGsGHD+O2335SBp6urKwcPHuTkyZMMGzYsQ+dKZmtrS/78+bl79y5du3bN1L7pOXbsGEuWLKFZs6QP4fv37yvvd82I7777DhMTExo0aICfnx9lypRRHrdVq1Z8/fXXQNIX3Vu3blG6dGmV/U+fPp3ieYkSJdTeC1ilShXCw8PR0dH5x/dvQsbb08zMDA8PDzw8PGjfvj1NmjTh+fPnmV5BNSP9OD0VKlTg559/TvX8VapU4caNG//qHJllbm6Ovb09p0+fVt5HGRcXp7yXNytVqFCBBw8ecPPmTbWjo3ny5CE8PFwl8Erv94L19PTUjsB/SkdHh4YNG9KwYUMmT56MhYUFhw8fVhv4qztmRsqWkfehrq5umuU1MzMjX758HD9+XOW+1pMnT1K9evV06/mxggULMmDAAAYMGMD48eNZsWJFhoNRfX195TT6ZFkxRRcg7n0cNwPuUqVRBU7sOKtMr9KwAif/UH8frToKhQJd/ZT/nt/HvufZo+do62hTp21Njv6acuaM0BxxcQncvBFGlWpFOHH0w4r9VaoV4eTxmynyx0TH0vdr1Z/raNm2KpWqOjB9wtYUweb7d/E8e/oKbW0t6riV4uhh1RkGmuJ9QgLXwh5Tu2hhDt74MEuoVtFCHLqhftbQhfuPaFKmBEa6usT8fXGsiJUF8QkJhKey+i5Aads83IzI+HeEzy0uLoFbfz2iSo2inPT7MCOjSvWinPqojySLiY6lX6clKmkt21ejklMRpo/bQvjDKAD0DXRJ/GRkPCE+EQUZm5mR0+LiEwi6/xjnUoU5cvlDn6hRqhB+V9T3kcC7j2hYuQSGerq8eZfURwrnTeojj6NS7yNfgty2mq4Eo//C2LFjqVmzJoMHD6Zv374YGxsTFBTEgQMHWLRokTJfnz59lEHOiRNpz9+fNGkSVatWpWzZssTGxrJz584UAVJaRo4cSbVq1Zg+fToeHh6cOnWKxYsXq6zEmnzf6IYNG/j999+BpAB15MiRAOneL6rOlClT8PLywszMjKZNmxIbG8v58+eJjIxMMQqRGcWLF2fdunU4OTnx8uVLRo8ene4o1Kfmzp1LfHw89evXV97jWLx4cbZu3crJkyextLRk3rx5hIeHp2jr+/fvM2LECPr378+FCxdYtGiRyurEH2vYsCHOzs60bt2a2bNnU7JkSR49esTu3btp3bp1pqaHptee8+fPx97enkqVKqGlpcWvv/6KnZ1dhkfDPpbRfpyWzp07M3PmTFq3bs2sWbOwt7fn4sWL5MuXD2dnZyZNmkSLFi0oWLAgHTp0QEtLi8uXL3PlyhW1qx9nlWHDhvHdd99RokQJSpcuzbx581RWEc4qrq6uuLi40K5dO+bNm0fx4sX566+/UCgUNGnSBDc3N548ecL3339P+/bt2bt3L3v27MHMzCzVYzo4OHDmzBlCQkIwMTHBysoKrU+WrN+5cyd3797FxcUFS0tLdu/eTUJCAiVLql8sx8HBgaNHj9KpUyf09fWxsbHJUNky8j50cHBQ3v+tr6+v9nduR48ezeTJkylWrBiVKlXC19eXwMBANmzYkOG29vb2pmnTpjg6OhIZGcnhw4dV3relSpVi1qxZtGnTJsPHzEpb5+9k7Nqh3Dx/h6BTN2nWryF5C9mw86ekqfS9ZnbBJp8V33suBuCrQe5EhD7l/l9Jq6aWq1OKDiO/YsfiPcpjlqpeHJv8VtwODMEmvxXdJ3dES0vB5u9///wVzAbRMfDxz00+CIOgW2BuBvlsc65cWWHr5jOMndiKm3+FEXT1Ac1aVSGvrTk7tyetRt1rQD1sbEz5fsYfJCZCSPATlf2jImN4/y5OJb1UmXzY5DHl9q3H2OQxpXsvF7QUCjZv0NyLE76nLvB9myZcDUtaEdejSnnszU3ZFJA0zX5E/drYmpow9veke/p3XvmLQXVrMKtVY37wO4WlkSGjG7mwNfAasXFJF70Gu9Tk0sMwQp5FYaKvR/fqlShll4epew6nWg5NsO2X04ye2oab1x8RdOUBzdpUJa+dObu2ngeg5+AG2OQxZc6UHSQmwr07n/SJ59G8exenkn762E3adnHm9o0w/rr2kPwFrOgxoB6nj93Q+OnbydYdvsC33ZtwLfQxl4PDaFe7PPZWpvx2LKmPDP2qNnnNTZi4LqmP7D73F32b1GDa141ZuvsUFsaGDG/jwu+nrhH7XrN/3ic9X8r02qwiwei/UKFCBfz9/ZkwYQJ169YlMTGRYsWK4eHhoZKvRIkS1KpVi2fPnqWYdvkpPT09xo8fT0hICIaGhtStW5dNmzZluExVqlRhy5YtTJo0ienTp2Nvb8+0adNUVu9VKBS4urqyY8cO6tatq6yLubk5RYsWTfMLcmr69OmDkZERc+bMYcyYMRgbG1O+fHnlojj/1KpVq+jXrx+VK1emUKFCzJw5U2XhoYyaP3++SkA6ceJEgoODcXd3x8jIiH79+tG6dWtevFCdSte9e3fevHlD9erV0dbWZujQoakuRKRQKNi9ezcTJkygV69ePHnyBDs7O1xcXJT3x2ZUeu1pYmLC7NmzuXXrFtra2lSrVo3du3enCFYyIqP9OC16enrs37+fkSNH0qxZM+Li4ihTpoxyURp3d3d27tzJtGnT+P7779HV1aVUqVL06dMn0+XNjJEjRxIWFoanpydaWlr06tWLNm3apHids8LWrVsZNWoUnTt3Jjo6muLFi/Pdd98BULp0aZYsWcLMmTOZPn067dq1Y9SoUSl+Hudjo0aNokePHpQpU4Y3b94QHBycYsTdwsKCbdu2MWXKFN6+fUuJEiXYuHEjZcuWVXvMadOm0b9/f4oVK0ZsbCyJiYkZKltG3oc+Pj6MGDGCFStWkD9/frU/RePl5cXLly8ZOXIkERERlClThj/++IMSJUpksJWT7vcePHgwDx48wMzMjCZNmigXYAO4ceNGtry+GeW/5SRm1iZ8PbE9VvaWhFy9z4TmM4kITRqtsbazJG+hDyukK7QU9JrZBbsieUmIS+DRnXB+Hr+BXcs+LMChZ6CH5/TO2BfNy5vXbzm7+yKzuy8i+oXm/4xFRly7AT28P3z5mv1j0t+tmyQyS/2vjH0x/A9dx8zMkK971sXK2oSQu0+YMGoTEX9P27a2NiGvbeZ+U1lPTwfPvm7Y57PkzZt3nD11m9nTfyf6dWz6O+eQPddvYmlkwCCXGuQ1MeZmxDP6/bKDRy+SRrDymBhjb26qzB/z/j291m/lm6b12Nq3C1Exb9lz/SYLjny4oG9moM+05g3JY2LEq9h3XA+P4Os1v3LlkWb+1moy/wPXMDU3pGsfV6xsTLh3J4JvvDcQEZ7UJ6xsTMhjl7k+8cvfU3E9B9bHOo8pL6JiOH3sJquXHMqOKmSL/RduYmFsQP+mNbAxM+Z22DOGLNlB2N+/MZrHzBh7qw995M279wxYvJVxHeqxYUwXXkS/Zf+Fm/y488tYtEl8oEjU9LH7/4DExERKlSpF//79/9UooRBCiKzXSKtDThdBI+x7lPpvIucm7u2653QRNEKou/q1GHIbhz9f5nQRNMLjGpkfqPgvClw8PNvP4XFqQJYda7PzT1l2rOwiI6PZLCIignXr1vHw4UN69uyZ08URQgghhBBCaCiZpiuylK2tLTY2NixfvlztPVRCCCGEEEIIAbKAkchiMgtaCCGEEEIIIVKSYFQIIYQQQgghNIBM0xVCCCGEEEII8dnltmA0878DIYQQQgghhBBC/EsyMiqEEEIIIYQQGiC3jYxKMCqEEEIIIYQQGiC3BaMyTVcIIYQQQgghxGcnI6NCCCGEEEIIoQHkd0aFEEIIIYQQQnx2Mk1XCCGEEEIIIYTIZjIyKoQQQgghhBAaILeNjEowKoQQQgghhBAaQIJRIYQQQgghhBCfXW4LRuWeUSGEEEIIIYQQn52MjAohhBBCCCGEBkjMZSOjEowKIYQQQgghhAbIbb8zKtN0hRBCCCGEEEJ8dhKMCiGEEEIIIYQGSEhUZNkjs5YsWUKRIkUwMDCgatWqHDt2LM38GzZsoGLFihgZGWFvb0/Pnj159uxZps4pwagQQgghhBBCaIDEREWWPTJj8+bNeHt7M2HCBC5evEjdunVp2rQpoaGhavMfP36c7t2707t3b65du8avv/7KuXPn6NOnT6bOK8GoEEIIIYQQQuRi8+bNo3fv3vTp04fSpUuzYMECChYsyNKlS9XmP336NA4ODnh5eVGkSBHq1KlD//79OX/+fKbOK8GoEEIIIYQQQmiArJymGxsby8uXL1UesbGxKc757t07AgICaNy4sUp648aNOXnypNpy1qpViwcPHrB7924SExN5/Pgxv/32G82bN89UfSUYFUIIIYQQQggNkJXTdGfNmoW5ubnKY9asWSnO+fTpU+Lj47G1tVVJt7W1JTw8XG05a9WqxYYNG/Dw8EBPTw87OzssLCxYtGhRpuorwagQQgghhBBC/MeMHz+eFy9eqDzGjx+fan6FQvU+08TExBRpya5fv46XlxeTJk0iICCAvXv3EhwczIABAzJVRvmdUSGEELmaonr5nC6CRnBvVzGni6AR9m1dm9NF0AiuA/vldBE0QnQh45wugkYwux+X00XINf7JKrip0dfXR19fP918NjY2aGtrpxgFjYiISDFammzWrFnUrl2b0aNHA1ChQgWMjY2pW7cuM2bMwN7ePkNllJFRIYQQQgghhNAAiYlZ98goPT09qlatyoEDB1TSDxw4QK1atdTuExMTg5aWaiipra39dx0yfnIZGRVCCCGEEEIIDZBA1o2MZsaIESPo1q0bTk5OODs7s3z5ckJDQ5XTbsePH8/Dhw9ZuzZp9kjLli3p27cvS5cuxd3dnbCwMLy9valevTr58uXL8HklGBVCCCGEEEKIXMzDw4Nnz54xbdo0wsLCKFeuHLt376Zw4cIAhIWFqfzmqKenJ69evWLx4sWMHDkSCwsL6tevz+zZszN1XkViZsZRhRBCiP+YxjWn5XQRNEKitty5A3LPaDK5ZzRJorwtANCKk3AB4Ojvo7P9HFV2f5Nlx7rQbEaWHSu7yMioEEIIIYQQQmiArFzA6Esg13uEEEIIIYQQQnx2MjIqhBBCCCGEEBogt91AKcGoEEIIIYQQQmiARJmmK4QQQgghhBBCZC8ZGRVCCCGEEEIIDZDbRkYlGBVCCCGEEEIIDSCr6QohhBBCCCGEENlMRkaFEEIIIYQQQgPIarpCCCGEEEIIIT47uWdUCCGEEEIIIcRnl9uCUblnVAghhBBCCCHEZycjo0IIIYQQQgihAXLZLaMSjAohhBBCCCGEJpBpukIIIYQQQgghRDaTkVEhhBBCCCGE0AS5bJ6uBKNCCCGEEEIIoQFkmq4QQgghhBBCCJHNZGRUCCGEEEIIITRAYi6bpisjo9nEwcGBBQsW5HQxRAZNmTKFSpUqacxx0qNQKNixY0e2nye7hYSEoFAoCAwMBMDPzw+FQkFUVFSOliszPtd73c3NDW9v72w/jxBCCCFyTmKiIsseX4L/3MjoTz/9xOjRo4mMjERHJ6l6r1+/xtLSkpo1a3Ls2DFl3mPHjuHi4sKNGzdwdHTMqSILDTBq1CiGDh2aqX0UCgXbt2+ndevW/+o44oNatWoRFhaGubl5ThclhdWrV+Pt7f1FBcpfMgcHB7y9vb/IALxlOyc6dHXGytqUe8ERLJ2/n6uXQtPdr0yFgvgs6UHI3QgGdl+uTNfW1qJTjzo0alYBmzxm3A99ysofD3H+9J3srMa/1rJNVTp0ccba2oSQ4Ccs/WE/Vy/dT3e/suUL4LO4OyHBEQzw/FmZrq2tRefutWnUtAI2NqbcD33Gz0sPcf7M3eysxmdz7hKs2gjXbsKTZwoWzUikYd2cLlXWad2kEp1bV8Pa0piQ+0/5YeURLgc9VJu3UtmCLJrhkSK965BVhD58niK9QZ2STBnZkmNnbvG/737P8rJntTbulejc6kNbLPRNvS0qly3Iomkp26KLVyptUbskU0e05OjZW/xvtma3ReumlejcphpWliaEhD5l0crDXL6eSp8oV5Afvu2UIv3rQSuV7eBSswRfd6hJfjsLdHS0ePAois2/n2O/3/VsrYf4d/5zwWi9evV4/fo158+fp2bNmkBS0GlnZ8e5c+eIiYnByMgISBqFyZcvnwSimfDu3Tv09PRyuhgq3r9/j66u7r86homJCSYmJv+6LFl1nNxKT08POzu7nC6GEP+Ya8MyDPB2Z9Gc3Vy7fJ/mravw7fwu9Om8hCePX6a6n5GxPmMmteLi+WAsrYxVtnkOqEcD9/LMn7WT+/ee4lSzGJO/64h3P1/u3AzP7ir9I64NyjBwWGMW+exRtsPMuZ3p/fVP6bfDxFZcDEjZDj37udHAvRzzZ+8i9N4znKoXZcqsDgzrv5o7tx5nd5Wy3Zs3ULI4tGkGwybmdGmyVv3aJfHqVY95yw9y5a+HfNW4InMmtqObly8RT1+lul+XwSuJjolVPo96+SZFHts8Zgzq4UbgtfQvdGiC+rVK4tWzHj4rktqiVeOKzJ3Qjm7evjxOoy06D1lJ9Jv022JwDzcCr2t+W9SvU5Khveszb9kBrgY95Cv3inw/qT3dh6xKu08M/JmYmHfK51EvY5R/v3z9lnW/nib0wTPexyVQy6ko47yaEvkihnMXQ7KzOlnrCxnRzCr/uWm6JUuWJF++fPj5+SnT/Pz8aNWqFcWKFePkyZMq6fXq1QOSgqwxY8aQP39+jI2NqVGjhsox1JkyZQqFChVCX1+ffPny4eXlpbI9JiaGXr16YWpqSqFChVi+fLnK9itXrlC/fn0MDQ2xtramX79+vH79WrlNS0uLp0+fAhAZGYmWlhYdOnRQ7j9r1iycnZ1TLV9kZCTdu3fH0tISIyMjmjZtyq1bt1TK/+mU0gULFuDg4KB87unpSevWrZk1a1aagfudO3do1aoVtra2mJiYUK1aNQ4ePKiSx8HBgenTp9OlSxdMTEzIly8fixYtUsmjUChYunQpTZs2xdDQkCJFivDrr78qtydP69yyZQtubm4YGBiwfv16AHx9fSldujQGBgaUKlWKJUuWqBz7wYMHdOrUCSsrK4yNjXFycuLMmTNq2+LcuXM0atQIGxsbzM3NcXV15cKFCyp1AWjTpg0KhUL5/NPjJCQkMG3aNAoUKIC+vj6VKlVi7969Keqzbds26tWrh5GRERUrVuTUqVNq2/ljYWFhqbZT/fr1GTJkiEr+Z8+eoa+vz+HDh9Ue79KlS9SrVw9TU1PMzMyoWrUq58+fV24/efIkLi4uGBoaUrBgQby8vIiOjlZpk5kzZ6bZ58+ePUvlypUxMDDAycmJixcvqmz/dJru6tWrsbCwYN++fZQuXRoTExOaNGlCWFiYcp+4uDi8vLywsLDA2tqasWPH0qNHD5UR6089e/aMzp07U6BAAYyMjChfvjwbN25MNb+fnx89e/bkxYsXKBQKFAoFU6ZMUW5P770+duxYHB0dMTIyomjRokycOJH3798rtyf3m3Xr1uHg4IC5uTmdOnXi1avU/yHv3bsXc3Nz1q5dq3Z7r169aNGihUpaXFwcdnZ2rFq1CoDY2Fi8vLzImzcvBgYG1KlTh3PnzinzJ7f/x3bs2IFCkfo/yoz26a1bt1K2bFn09fVxcHDAx8dHuc3NzY179+4xfPhwZXsnS68f5rR2nZ3Z++dF9v5xkfshT/lpwX6eRLygZVunNPfzHtecI/uvEnT1QYptDZtUYOOa45w7dZvwR1Hs3BZAwJk7tO9SM7uq8a+186jB3p2B7PkzkNB7z1i68ABPIl7Ssk3VNPfzHtOMwweucv1qytGRhk3Ks3HtCc6eupPUDjsucP7MXdp31tx2yAyXmuDdBxq75HRJsp7HV07sOnSFnQevcO/BcxatOkLEs1e0aVIpzf0io2J4/tEjIUH1ZjotLQWThjdj1aYThD1+kY01yDqdWjqx8/AVdh66wr2Hz/nBN6ktWrtXSnO/yBfpt8XkYc1YufkEj76AtujYyoldB6+w68DffWLlEZ48fUXrppXS3C/qRQzPo6KVj4/bIfDqfY6dvsW9B895FB7FbzsvcDfkCRVK58/m2mStxMSse3wJ/nPBKCR9kTly5Ijy+ZEjR3Bzc8PV1VWZ/u7dO06dOqUMRnv27MmJEyfYtGkTly9fpkOHDjRp0kQlePvYb7/9xvz581m2bBm3bt1ix44dlC9fXiWPj4+P8gv3oEGDGDhwIH/99ReQ9OW1SZMmWFpacu7cOX799VcOHjyoDCDKlSuHtbU1/v7+ABw9ehRra2uOHj2qPL6fnx+urq6ptoOnpyfnz5/njz/+4NSpUyQmJtKsWTOVL8EZcejQIYKCgjhw4AA7d+5Um+f169c0a9aMgwcPcvHiRdzd3WnZsiWhoapT0+bMmUOFChW4cOEC48ePZ/jw4Rw4cEAlz8SJE2nXrh2XLl3i66+/pnPnzgQFBankGTt2LF5eXgQFBeHu7s6KFSuYMGEC3377LUFBQcycOZOJEyeyZs0aZflcXV159OgRf/zxB5cuXWLMmDEkJCSorc+rV6/o0aMHx44d4/Tp05QoUYJmzZopg4PkL+y+vr6EhYWpfIH/2MKFC/Hx8WHu3LlcvnwZd3d3vvrqqxT9asKECYwaNYrAwEAcHR3p3LkzcXFxao+ZkXbq06cPv/zyC7GxH66ibtiwgXz58in7/Ke6du1KgQIFOHfuHAEBAYwbN0454nzlyhXc3d1p27Ytly9fZvPmzRw/fjxFwJtWn4+OjqZFixaULFmSgIAApkyZwqhRo9KsIyS9V+bOncu6des4evQooaGhKvvNnj2bDRs24Ovry4kTJ3j58mW699O+ffuWqlWrsnPnTq5evUq/fv3o1q2b8uLEp2rVqsWCBQswMzMjLCyMsLAwlTKkVW8AU1NTVq9ezfXr11m4cCErVqxg/vz5Kue4c+cOO3bsYOfOnezcuRN/f3++++47teXZtGkTHTt2ZO3atXTv3l1tnj59+rB3716VwH337t28fv2ajh07AjBmzBi2bt3KmjVruHDhAsWLF8fd3Z3nz1NO/cqstPp0QEAAHTt2pFOnTly5coUpU6YwceJEVq9eDcC2bdsoUKAA06ZNU7Y3ZLwf5hQdHS1KlLTnwhnV6bMBZ+5SpnzBVPdr3Lwi9vktWbfSX+12XT1t3r9T/TyIjY2jbMVC/77Q2UBHRwvHkvYEnFWdPhtw9i5lyxVIdT/3ZhXJl9+SdauOqt2uq6vNu3fxKmnvYuMoVyH1thU5T0dHC8ditpwNDFFJPxcYQrlS+dLcd+W87uxYOYAFUztQuVzK19mzozNRL96w69DVrCxytklui3OftsWlEMqVTLstVs3tzo6fB7Bgcipt0cGZqJdfRlsktYNdynYIDKFcqbQDx5Xzu7PddyDzp3WkchqfqwBVKhSiYH5LLl1LeZFPoyVm4eML8J8NRk+cOEFcXByvXr3i4sWLuLi44OrqqhztPH36NG/evKFevXrcuXOHjRs38uuvv1K3bl2KFSvGqFGjqFOnDr6+vmrPERoaip2dHQ0bNqRQoUJUr16dvn37quRp1qwZgwYNonjx4owdOxYbGxvl+Tds2MCbN29Yu3Yt5cqVo379+ixevJh169bx+PFjFAoFLi4uyvx+fn706NGDhIQErl+/TlxcHCdPnsTNzU1t+W7dusUff/zBzz//TN26dalYsSIbNmzg4cOHmV74xtjYmJ9//pmyZctSrlw5tXkqVqxI//79KV++PCVKlGDGjBkULVqUP/74QyVf7dq1GTduHI6OjgwdOpT27dun+FLeoUMH+vTpg6OjI9OnT8fJySnFCKq3tzdt27alSJEi5MuXj+nTp+Pj46NMa9u2LcOHD2fZsmUA/PLLLzx58oQdO3ZQp04dihcvTseOHVMdWa5fvz5ff/01pUuXpnTp0ixbtoyYmBjlxYE8efIAYGFhgZ2dnfL5p+bOncvYsWPp1KkTJUuWZPbs2VSqVCnFgjejRo2iefPmODo6MnXqVO7du8ft27fVHjMj7dSuXTsUCgW///7hfhFfX188PT1THdUKDQ2lYcOGlCpVihIlStChQwcqVqwIJF1E6NKlC97e3pQoUYJatWrxww8/sHbtWt6+fas8Rnp9Pj4+nlWrVlG2bFlatGjB6NGj06wjJE3D/umnn3BycqJKlSoMGTKEQ4cOKbcvWrSI8ePH06ZNG0qVKsXixYtTjOZ9Kn/+/IwaNYpKlSpRtGhRhg4diru7u8ro8sf09PQwNzdHoVBgZ2eHnZ2dynTstOoN8M0331CrVi0cHBxo2bIlI0eOZMuWLSrnSEhIYPXq1ZQrV466devSrVs3lXomW7JkCQMGDOD333+nVatWqdaxVq1alCxZknXr1inTfH196dChAyYmJkRHR7N06VLmzJlD06ZNKVOmDCtWrMDQ0JCVK1em2X4ZkVafnjdvHg0aNGDixIk4Ojri6enJkCFDmDNnDgBWVlZoa2tjamqqbG/IeD/MKWYWRmjraBH5XHWkNvJ5NJbWxmr3yVfQit6DG/Dd5O0kxKv/5nD+9B3adq5JvoJWKBRQpXpRnF1KYmWtmbcEmKfWDpHRWKZS5vwFLOk9sB6zpu5IvR3O3KVdpxrkL2CZ1A7ViuBc11Fj20EkMTc1REdbi8ioGJX0yKgYrCzUvy+eRb7m+yX7mPj970z4/ndCH0ayYGpHKpb5cDGjfKl8NG9Qnu+X7M/W8mel5LZ4/kK1LZ5HxWCdSls8jXzN7KX7+GZOUlvcfxTJwsmftEXJfLRoUJ7ZS7+MtjA3S+4Tqp8Rz6OisbJMo08s3sfE737nm+9+J/Thc+ZP81BpBwBjIz32bhrG4a0jmD2xHQuXH+L8pXvZVhfx7/3n7hmFpPtGo6OjOXfuHJGRkTg6OpI3b15cXV3p1q0b0dHR+Pn5UahQIYoWLcqvv/5KYmJiiimosbGxWFtbqz1Hhw4dWLBgAUWLFqVJkyY0a9aMli1bKhdNAqhQoYLy7+QvsREREQAEBQVRsWJFjI0/vOlq165NQkICN27cwNbWFjc3N+V0P39/f6ZPn05wcDD+/v68ePGCN2/eULt2bbXlCwoKQkdHhxo1aijTrK2tKVmyZIpRxvSUL18+3ftEo6OjmTp1Kjt37uTRo0fExcXx5s2bFCOjnwZ/zs7OKQIzdXmSV1tN5uT0YcrbkydPuH//Pr1791a5IBAXF6dcCCcwMJDKlStjZWWVZj2SRUREMGnSJA4fPszjx4+Jj48nJiYmRX3S8vLlSx49epTiNapduzaXLl1SSfu4r9jb2yvLUKpUqVSPn1Y76evr8/XXX7Nq1So6duxIYGAgly5dSvNCxIgRI+jTpw/r1q2jYcOGdOjQgWLFigFJI1m3b99mw4YNyvyJiYkkJCQQHBxM6dKlU9QjtT6ffM+2ujqoY2RkpCwHJLVP8jFfvHjB48ePqV69unK7trY2VatWTXXUGyA+Pp7vvvuOzZs38/DhQ2JjY4mNjVV5P2ZGWvWGpJkUCxYs4Pbt27x+/Zq4uDjMzMxUjuHg4ICpqanaeibbunUrjx8/5vjx4yp1PnbsGE2bNlU+X7ZsGV27dqVPnz4sX76cMWPGEBERwa5du5QB7p07d3j//r1K/9TV1aV69eqZ/oxQJ60+HRQUlCKQrl27NgsWLCA+Ph5tbW21x8xoP0xL8mv9sYSEOLS0su7f4adToxQK1F6h1tJSMH5qG9au8Ofh/dRHo5fO38fw8S1YuWkQJMKjh8/ZvzOQxi0qZVmZs0PiJw2hUJMGf7fDlDasXXk0zXZYsnA/w8c2Z+UvA5Pa4VEk+3ddonHzillddJENEj99EyjU9weA+48iuf8oUvn82o0w8tqY0qmVE5euP8DQQJdvvJvz/dL9vHiV8t5JTZfivaFQ0z5/S9EWN5PaovNXH9pi4rAvsy1SflYqUu8TDyO5//DjPvEoqU+0qcal6x9GPmPevKO39xoMDfWoWqEQg3vV49HjFwRe1fz7aJN9KavgZpX/ZDBavHhxChQowJEjR4iMjFROZbWzs6NIkSKcOHGCI0eOUL9+fSBpREJbW5uAgIAUX4JSW4ymYMGC3LhxgwMHDnDw4EEGDRrEnDlz8Pf3V05t/HRRHYVCofyCnJiYmOoIVXK6m5sbw4YN4/bt21y9epW6dety584d/P39iYqKomrVqipfXj+W2pv54/NqaWmlyKduCm9GvqCPHj2affv2MXfuXIoXL46hoSHt27fn3bt36e6b1v1nqeX5uEzJbbpixQqV4BtQvp6GhobpnuNjnp6ePHnyhAULFlC4cGH09fVxdnbOUH3SK7u61/7jvpK8La1gKiPn6tOnD5UqVeLBgwesWrWKBg0aULhw4VT3nTJlCl26dGHXrl3s2bOHyZMns2nTJtq0aUNCQgL9+/dPcV80QKFCH6YKptfn/wl1x0z5jzxlG6fFx8eH+fPns2DBAsqXL4+xsTHe3t7/6PVNrYzJ9T59+jSdOnVi6tSpuLu7Y25uzqZNm1TukUzvGMkqVarEhQsX8PX1pVq1asp6Ozk5qVywsbW1BaB79+6MGzeOU6dOcerUKRwcHKhbN2l5zuQ2Sqt/ZvQzIr02+bRPq3sPZKR/ZLQfpmXWrFlMnTpVJa1ofjeKFVA/fT0zXkbFEB+XgNUno6AWlsYpRgkBDI30KFkmP8Ud7RkyMuligkJLgZaWgj3Hv2H8sPUEBoTwIiqGKWO3oKunjZm5Ec+evKL34AaEf/QFVZO8ULaD6v9PC0tjolJrh9L5KF7CjiHDmwAf2mGv//8YN/wXAi/83Q7jf01qBzMjnj19RZ+B9QkPi/oc1RL/0ItXb4iLT0gxCmppbkTkJyOEabl28xGNXcsAkN/Ogny25nz3vzbK7Vp/f6Yc+W0EXYes5FG45t03mdwWn46CWpob8Twqk23h8klbjE/ZFn5bRtBl6EqNu4f0xcu/+4Slmj6RiXa4fjNM2SeSJSbCw/AoAG4HR1C4oDVft6/xRQWjX8r02qzynwxGIWl01M/Pj8jISJWpgK6uruzbt4/Tp0/Ts2dPACpXrkx8fDwRERHKL2oZYWhoyFdffcVXX33F4MGDKVWqFFeuXKFKlSrp7lumTBnWrFlDdHS0MrA6ceIEWlpayhHa5PtGZ8yYQcWKFTEzM8PV1ZVZs2apBNmpHT8uLo4zZ85Qq1YtIGnRlps3bypHD/LkyUN4eLjKF8NPRyAz6tixY3h6etKmTdKH4evXrwkJCUmR7/Tp0ymefzr6d/r0aZX74E6fPk3lypVTPbetrS358+fn7t27dO3aVW2eChUq8PPPP/P8+fMMjY4eO3aMJUuW0KxZMwDu37+vXEwqma6uLvHx8ep2B8DMzIx8+fJx/PhxXFw+rEhx8uRJlVGtfyq9dipfvjxOTk6sWLGCX375JcVUZ3UcHR1xdHRk+PDhdO7cGV9fX9q0aUOVKlW4du0axYsX/8flLVOmDOvWrePNmzfKiwOf9ofMMjc3x9bWlrNnzyrfu/Hx8Vy8eDHN33s9duwYrVq14uuvvwaSgpxbt26lObKmp6eX5uudmhMnTlC4cGEmTJigTLt3759NGSpWrBg+Pj64ubmhra3N4sWLgaTPInWvjbW1Na1bt8bX15dTp04pP/Mg6aKdnp4ex48fp0uXLkBSoHn+/Hnlz6nkyZOHV69eqXxO/dPPiI+VKVOG48ePq6SdPHkSR0dH5QUkde2dFf1w/PjxjBgxQiWtbcO5//h4H4uLS+DWjTCqVC/KCf8byvQq1Yty6uiNFPljomPp12WpSlrLdk5UqlqE6f/7lfBHUSrb3r+L59mTV2hra1HHrTRHD2nmzxXExSVw80YYVaoV4cRH9a5SrQgnj99MkT8mOpa+Xy9TSWvZtiqVqjowfcLWFMHm+3fxPHua3A6lOHr434/ki+wTF5fAzTuPqVbRgWNnPtx+Uq2iA8fPpn07yscci9jyLDLpYkbow+d0H7ZaZXvfLrUxMtRj4cojaa7GmpM+boujH9XdqYIDx89lvC1KfNIW3bxXq2zv26U2RgZ6LPx7oShNk9QO4ThVLMyx0x/W0HCqVJjjZzLRDkXz8izydZp5FICujvrZNkIz/KeD0cGDB/P+/XuVoM3V1ZWBAwfy9u1b5UIujo6OdO3ale7du+Pj40PlypV5+vQphw8fpnz58sqA5GOrV68mPj6eGjVqYGRkxLp16zA0NExz5OljXbt2ZfLkyfTo0YMpU6bw5MkThg4dSrdu3ZSjGsn3ja5fv57hw4cDSUHVu3fvOHToEMOGDUv1+CVKlKBVq1b07duXZcuWYWpqyrhx48ifP79yepybmxtPnjzh+++/p3379uzdu5c9e/akmD6YEcWLF2fbtm20bNkShULBxIkT1Y7snThxgu+//57WrVtz4MABfv31V3bt2qWS59dff8XJyYk6deqwYcMGzp49m+49bFOmTMHLywszMzOaNm1KbGws58+fJzIykhEjRtC5c2dmzpypXBnY3t6eixcvki9fPrVTRYsXL866detwcnLi5cuXjB49OsXoqoODA4cOHaJ27dro6+tjaWmZ4jijR49m8uTJFCtWjEqVKuHr60tgYKDKNMN/KiPt1KdPH4YMGYKRkZHyQoE6b968YfTo0bRv354iRYrw4MEDzp07R7t27YCkBaNq1qzJ4MGD6du3L8bGxspFrTIS5AJ06dKFCRMm0Lt3b7755htCQkKYO/ffBwFDhw5l1qxZFC9enFKlSrFo0SIiIyPTHHEvXrw4W7du5eTJk1haWjJv3jzCw8PTDEYdHBx4/fo1hw4dUk43/njKcVrnCg0NZdOmTVSrVo1du3axffv2f1RXSPq8Sl6UTUdHJ8U090/16dOHFi1aEB8fT48ePZTpxsbGDBw4kNGjR2NlZUWhQoX4/vvviYmJoXfv3gDKz7f//e9/DB06lLNnzyoXGfo3Ro4cSbVq1Zg+fToeHh6cOnWKxYsXq6yA7eDgwNGjR+nUqRP6+vrY2NhkqB+OHz+ehw8fprrKsL6+Pvr6+ippWTlFd+vGU4yZ3IabQWFcv/qA5q2qkNfWnJ3bAwDoNbA+1nlMmTPtdxITIeTuE5X9oyJjePcuTiW9VNn8WOcx5c7NcGzymNGtjytaWgq2rD+RZeXOals3n2HsxFbc/CuMoKsPaKZsh6RVyXsNqIeNjSnfz/gjqR2CU7bD+3dxKumlyuTDJo8pt289xiaPKd17uaClULB5w0n+C6JjIPSjRYQfhEHQLTA3g3y2OVeurLD5j/N8M6wZf90J59qNR3zVqAJ5bUzZsS/plpX+X9fFxsqEb3/YA0CHFlUIj3hJ8P2n6Oho4+5aBrdajkz4+3cz372PJzhU9QLx6+ik6fefpmuaTX+eZ6JXUltc/bstbG1M2bH/77boWpc8VibMWPR3WzSvQviTpLbQ1dGmsUsZ6jk78r/vP2qL+6m0xX3NbYstv59ngndzbtxO6hMt3SuS18aM3/cmtUO/bnWxsTZl5oLdAHRoWZWwiBeEhCb1icZuZXCrVZJvZu1QHrNruxrcuB3Ow/AodHW0qVm1KO71yuLz0wF1RdBYMk33P6JevXq8efOGUqVKKYM7SApGX716RbFixShY8MMqXL6+vsyYMYORI0fy8OFDrK2tcXZ2VhuIQtLCNd999x0jRowgPj6e8uXL8+eff6Z6j+mnjIyM2LdvH8OGDaNatWoYGRnRrl075s2bl6Ie27ZtUy5UpFAoqFu3Ljt37qROnTppnsPX15dhw4bRokUL3r17h4uLC7t371ZOnytdujRLlixh5syZTJ8+nXbt2jFq1KgUP0uREfPnz6dXr17UqlVL+aXx5cuUvyU3cuRIAgICmDp1Kqampvj4+ODu7q6SZ+rUqWzatIlBgwZhZ2fHhg0bKFOmTIpjfaxPnz4YGRkxZ84cxowZg7GxMeXLl1eO8Ojp6bF//35GjhxJs2bNiIuLo0yZMvz4449qj7dq1Sr69etH5cqVKVSoEDNnzkyx8quPjw8jRoxgxYoV5M+fX+1IsJeXFy9fvmTkyJFERERQpkwZ/vjjD0qUKJFmfTIiI+3UuXNnvL296dKlCwYGBqkeS1tbm2fPntG9e3ceP36MjY0Nbdu2VU5nrFChAv7+/kyYMIG6deuSmJhIsWLF8PBI+UPcqTExMeHPP/9kwIABVK5cmTJlyjB79mxlwPtPjR07lvDwcLp37462tjb9+vXD3d091fsOIWkl4uDgYNzd3TEyMqJfv360bt2aFy9Sn8pUq1YtBgwYgIeHB8+ePWPy5MkqP++SmlatWjF8+HCGDBlCbGwszZs3Z+LEiRnaNzUlS5bk8OHDyhHST6f8fqxhw4bY29tTtmxZ8uVTXa3xu+++IyEhgW7duvHq1SucnJzYt2+f8sKKlZUV69evZ/To0SxfvpyGDRsyZcoU+vXr94/LDkkjnFu2bGHSpElMnz4de3t7pk2bhqenpzLPtGnT6N+/P8WKFSM2NpbExMQM9cOwsLBM3dud1fwPXsfM3IiuvV2wsjbh3t0IvhnxCxF/Txm0sjEhr515po6pq6eDZ/962Oez5M2bd5w9eYvZU7cT/To2/Z1ziP+h65iZGfJ1z7pYWZsQcvcJE0ZtIuLv6YLW1ibktc1cO+jp6eDZ1+1DO5y6zezpv2t0O2TGtRvQw/vDl9DZPyb93bpJIrPG51SpssbhEzcwMzXEs6Mz1pbGBIc+ZcyMbTx+kvQ9wdrSGNs8Hy6E6+poM8jTlTxWJsS+iyP4/jNGT9/K6QvBOVWFLHP45A3MTQ3x7PChLUbP/KQtbFTbYnB31bYY9e2X3xaHjyf1iR4etbC2Mib43lPGTtv6UTuYYGvz4VY0HR1tBvV0+9AOoc8YM+03Tgd8aAdDA11GDGhEHuukPKEPnzNj/i4OH085M0Wj5bJpuorEf3ojlxCZ5ODggLe3tzJAVEehULB9+/Y0fyNSZNz9+/dxcHDg3LlzGZo+/l+QkJBA6dKl6dixI9OnT8/p4uS4mJgY8uXLx6pVq2jbtm1OF0cjNa45LaeLoBEStf+TC+xn2r6t6kfVcxvXgf/uotN/RaK8LQDQipNwAeDo7+n/CsC/5bBW/c+6/RMh3cdl2bGyy392ZFSI3Oz9+/eEhYUxbtw4atas+Z8ORO/du8f+/ftxdXUlNjaWxYsXExwcrLwPMrdKSEggPDwcHx8fzM3N+eqrr3K6SEIIIYRIl0zTFUJ84U6cOEG9evVwdHTkt99+y+niZCstLS1Wr17NqFGjSExMpFy5chw8eDBDP/PxXxYaGkqRIkUoUKAAq1evVvnZKSGEEEJoqFw2CC3fTsRno+6eyk/JrPGs4ebmlmvasmDBgpw4obkLueQUBweHXNMHhBBCCPFlkmBUCCGEEEIIITRBLruOLMGoEEIIIYQQQmgC+WkXIYQQQgghhBCfW267w0YWrBZCCCGEEEII8dnJyKgQQgghhBBCaIJcNjIqwagQQgghhBBCaIJcds+oTNMVQgghhBBCCPHZycioEEIIIYQQQmgAhUzTFUIIIYQQQgjx2eWyYFSm6QohhBBCCCGE+OxkZFQIIYQQQgghNEEuW8BIglEhhBBCCCGE0AS5bJquBKNCCCGEEEIIoQlyWTAq94wKIYQQQgghhPjsZGRUCCGEEEIIITRBLhsZlWBUCCGEEEIIITRBLlvASKbpCiGEEEIIIUQut2TJEooUKYKBgQFVq1bl2LFjqeb19PREoVCkeJQtWzZT55RgVAghhBBCCCE0gCIx6x6ZsXnzZry9vZkwYQIXL16kbt26NG3alNDQULX5Fy5cSFhYmPJx//59rKys6NChQ6bOK8GoEEIIIYQQQmiCxCx8ZMK8efPo3bs3ffr0oXTp0ixYsICCBQuydOlStfnNzc2xs7NTPs6fP09kZCQ9e/bM1HklGBVCCCGEEEKI/5jY2Fhevnyp8oiNjU2R7927dwQEBNC4cWOV9MaNG3Py5MkMnWvlypU0bNiQwoULZ6qMEowKIYQQQgghxH/MrFmzMDc3V3nMmjUrRb6nT58SHx+Pra2tSrqtrS3h4eHpnicsLIw9e/bQp0+fTJdRVtMVQgghhBBCCA2Q2Xs90zJ+/HhGjBihkqavr5/6uRWqK/kmJiamSFNn9erVWFhY0Lp160yXUYJRIYQQQgghhPiP0dfXTzP4TGZjY4O2tnaKUdCIiIgUo6WfSkxMZNWqVXTr1g09Pb1Ml1GCUSGEELlavJFuThdBIzxwM8jpImgE14H9croIGsF/6fKcLoJGcB3QN6eLoBFC3XPXb1/mqBz4nVE9PT2qVq3KgQMHaNOmjTL9wIEDtGrVKs19/f39uX37Nr179/5H55ZgVAghhBBCCCE0QRZO082MESNG0K1bN5ycnHB2dmb58uWEhoYyYMAAIGnK78OHD1m7dq3KfitXrqRGjRqUK1fuH51XglEhhBBCCCGE0AQ5FIx6eHjw7Nkzpk2bRlhYGOXKlWP37t3K1XHDwsJS/Oboixcv2Lp1KwsXLvzH55VgVAghhBBCCCFyuUGDBjFo0CC121avXp0izdzcnJiYmH91TglGhRBCCCGEEEIDZOVqul8CCUaFEEIIIYQQQhPksmBUK6cLIIQQQgghhBAi95GRUSGEEEIIIYTQBLlsZFSCUSGEEEIIIYTQALntnlGZpiuEEEIIIYQQ4rOTkVEhhBBCCCGE0ASJipwuwWclwagQQgghhBBCaAKZpiuEEEIIIYQQQmQvGRkVQgghhBBCCA2Q2xYwkmBUCCGEEEIIITSBBKNCCCGEEEIIIT633DYyKveMCiGEEEIIIYT47GRkVAghhBBCCCE0QS4bGZVgVAghhBBCCCE0QS4LRmWarhBCCCGEEEKIz05GRoUQQgghhBBCA8gCRkIIIYQQQgghRDaTYFQIIYQQQgghxGcn03SFEEIIIYQQQhPINF0h/jumTJlCpUqVNLoMq1evxsLC4l+fx83NDW9v7399HCGEEEIIkTMUiVn3+BLIyKjIUomJiTRq1AhtbW327dunsm3JkiWMHz+eK1euUKhQoRwqoebx8PCgWbNmOV2MbKdQKNi+fTutW7fO6aKI/wAHBwe8vb018gLMV19VoaNHDaytTQgJecKSHw9y5coDtXkrVizEvPldU6R79ljG/fvPASjsYIOnZ10cHe2ws7Pgxx8Psm3ruWytQ1bo4lSB3s5O5DE15lbEM2bu9ycg9GGq+XW1tRnsUoOvypcmj4kR4S9f89Pxs2wNvAZAm4pl+K6Ve4r9yn/7A+/i47OtHv9W6yaV6Ny6GtaWxoTcf8oPK49wOUh9O1QqW5BFMzxSpHcdsorQh89TpDeoU5IpI1ty7Mwt/vfd71le9pxw7hKs2gjXbsKTZwoWzUikYd2cLlXWat00uU+Y/N0nDnP5eip9olxBFs3olCK96+CVqfSJUkwZ9XefmLUjq4uepb4uW4n+lauR18iYm8+fMu3EEc6FqW+HufWb0L5UuRTpN58/pfGm1QCUsLRmRPXalM9jSwEzc6YdP8yqyxeyswoiC0gwKrKUQqHA19eX8uXLs2zZMvr37w9AcHAwY8eOZdGiRRKIfsLQ0BBDQ8OcLoZIQ3x8PAqFAi0tmUwi0ubmVppBgxvyw8J9XL36gBYtKzPrOw969VxBRMTLVPfr0X0Z0dGxyucvXsQo/zbQ1yUsLIqj/n8xcFDDbC1/VmlaxpHx7m5M3X2YC/cf0alKeVZ0aU3zJWsJe/lK7T4L2zfH2tiICX8eIPR5FFbGhuh88p579TaWJj+uVknT5EC0fu2SePWqx7zlB7ny10O+alyRORPb0c3Ll4in6tsBoMvglUTHfOgPUS/fpMhjm8eMQT3cCLx2P1vKnlPevIGSxaFNMxg2MadLk/WS+kR95i07kNQn3CsyZ2J7ug1dlXafGPQz0THvlM+jXsakyGObx4xBnl9Gn2hRvCST6tRj4tGDnA9/SNcyFVndoh2NNvry6HXKdph6/DCzTx1VPtfW0mKPRw9237mpTDPU1SX05Qt237nBxNr1Pks9ssUXMqKZVeSblchyBQsWZOHChYwaNYrg4GASExPp3bs3DRo0oEiRIlSvXh19fX3s7e0ZN24ccXFxyn0dHBxYsGCByvEqVarElClTUj2fn58f1atXx9jYGAsLC2rXrs29e/dU8qxbtw4HBwfMzc3p1KkTr159+KCLjY3Fy8uLvHnzYmBgQJ06dTh37sOoQ9WqVfHx8VE+b926NTo6Orx8mfTFMjw8HIVCwY0bNzLUPsHBwRQvXpyBAweSkJCQYppu8rTetMocHR1N9+7dMTExwd7eXqV8yRwcHJg5cya9evXC1NSUQoUKsXz5cpU8Dx8+xMPDA0tLS6ytrWnVqhUhISEZattLly5Rr149TE1NMTMzo2rVqpw/f15tnR0cHABo06YNCoVC+Rxg6dKlFCtWDD09PUqWLMm6devSbUNfX19Kly6NgYEBpUqVYsmSJcptzs7OjBs3TiX/kydP0NXV5ciRIwC8e/eOMWPGkD9/foyNjalRowZ+fn7K/Mmvyc6dOylTpgz6+vocO3YMXV1dwsPDVY49cuRIXFxclM+3bt1K2bJl0dfXx8HBIcVrkxWvy8cSExMpXrw4c+fOVUm/evUqWlpa3LlzB4DQ0FBatWqFiYkJZmZmdOzYkcePHyvze3p6phi19vb2xs3NTe15P22nkiVLYmRkRPv27YmOjmbNmjU4ODhgaWnJ0KFDif8oYEiv/QFOnjyJi4sLhoaGFCxYEC8vL6Kjo4GkKen37t1j+PDhKBQKFApFqmX83Np3qM6ePZfYvfsSoaHPWPLjQSIiXtLyq8pp7hcZGa3ySEj48G3kxo0wli87wpEjQbx/H5fGUTRHT+cqbL14ld8uXuXu0+fM3O9P+ItXdHaqoDZ/3WKFqVY4P/1+2c6p4FAevnjJlUePufggTCVfIok8jY5ReWgyj6+c2HXoCjsPXuHeg+csWnWEiGevaNOkUpr7RUbF8Pyjx8f9AUBLS8Gk4c1YtekEYY9fZGMNPj+XmuDdBxq7pJ/3S+TRyoldBz/qEyuPEPE0A33iRQzPo6KVD/V9ovkX0yf6VHRiS9AVNgdd4U7kc6adOELY61d8Xa6S2vyv3r3jyZsY5aNCXjvM9Q34NeiqMs/liHBmnfLnz9s3NPoiVboSs/DxBZBgVGSLHj160KBBA3r27MnixYu5evUqCxcupFmzZlSrVo1Lly6xdOlSVq5cyYwZM/7xeeLi4mjdujWurq5cvnyZU6dO0a9fP5Uvp3fu3GHHjh3s3LmTnTt34u/vz3fffafcPmbMGLZu3cqaNWu4cOECxYsXx93dnefPk6a/uLm5Kb8oJyYmcuzYMSwtLTl+/DgAR44cwc7OjpIlS6Zb3qtXr1K7dm06dOjA0qVLUx1pS6/Mo0eP5siRI2zfvp39+/fj5+dHQEBAiuP4+Pjg5OTExYsXGTRoEAMHDuSvv/4CICYmhnr16mFiYsLRo0c5fvw4JiYmNGnShHfv3qXbtl27dqVAgQKcO3eOgIAAxo0bh66urtr6JAf3vr6+hIWFKZ9v376dYcOGMXLkSK5evUr//v3p2bOnMmhUZ8WKFUyYMIFvv/2WoKAgZs6cycSJE1mzZo2yXBs3biQx8cOn8ObNm7G1tcXV1RWAnj17cuLECTZt2sTly5fp0KEDTZo04datW8p9YmJimDVrFj///DPXrl3DycmJokWLqgTLcXFxrF+/np49ewIQEBBAx44d6dSpE1euXGHKlClMnDiR1atXZ9nr8imFQkGvXr3w9fVVSV+1ahV169alWLFiJCYm0rp1a54/f46/vz8HDhzgzp07eHiknA6YWTExMfzwww9s2rSJvXv34ufnR9u2bdm9eze7d+9m3bp1LF++nN9++025T3rtf+XKFdzd3Wnbti2XL19m8+bNHD9+nCFDhgCwbds2ChQowLRp0wgLCyMsLExt2T43HR0tHB3tOH8+WCU94HwwZcsWSHPfZct7seXXocyZ25lKlb7s2SO6WlqUtbfl+B3Vi4In7oZSuWA+tfvUdyzG1UcR9KldjaPefdk72JMxjeqir6Otks9IT4/DXr3x9+7DT51aUdouT7bV49/S0dHCsZgtZwNDVNLPBYZQrpT6dki2cl53dqwcwIKpHahcrmCK7Z4dnYl68YZdh66q2VtoqqQ+YZdKn8if5r4r53Vnx6qBLJjWMZU+UYuol2/YdfBKVhY5W+hqaVEujy3H7oeopB+7H0JV27TfG8k6li7P8Qf3ePg69RknXyq5Z1SILLJ8+XLKlSvHsWPH+O2331i+fDkFCxZk8eLFKBQKSpUqxaNHjxg7diyTJk36R1MgX758yYsXL2jRogXFihUDoHTp0ip5kkcfTU1NAejWrRuHDh3i22+/JTo6mqVLl7J69WqaNm0KJAU7Bw4cYOXKlYwePRo3NzdWrlxJQkICV65cQVtbm6+//ho/Pz+aNWuGn5+fMshJy6lTp2jRogXjx49n1KhRaeZNq8yvX79m5cqVrF27lkaNGgGwZs0aChRI+WW3WbNmDBo0CICxY8cyf/58/Pz8KFWqFJs2bUJLS4uff/5ZGWD6+vpiYWGBn58fTk5OabZtaGgoo0ePplSpUgCUKFEi1frkyZP0hdHCwgI7Oztl+ty5c/H09FSWccSIEZw+fZq5c+dSr576KTbTp0/Hx8eHtm3bAlCkSBGuX7/OsmXL6NGjBx4eHgwfPpzjx49Tt27SjUa//PILXbp0UY4Ubty4kQcPHpAvX9I/vVGjRrF37158fX2ZOXMmAO/fv2fJkiVUrFhRee7evXvj6+vL6NGjAdi1axcxMTF07NgRgHnz5tGgQQMmTkyaW+bo6Mj169eZM2cOnp6eWfK6NG7cOEWb9OzZk0mTJnH27FmqV6/O+/fvWb9+PXPmzAHg4MGDXL58meDgYAoWTPoSs27dOsqWLcu5c+eoVq1aqq9det6/f68c3QZo374969at4/Hjx5iYmFCmTBnq1avHkSNH8PDwyFD7z5kzhy5duijvBy1RogQ//PADrq6uLF26FCsrK7S1tTE1NVXpTxkRGxtLbGysSlpCQhxaWv/+36G5uRHa2lpERkarpEdGRmNlZax2n2fPX+Pjs5tbN8PR1dWhYaNyzJnbhREjNnDlsuZPtVPH0ihpeu2zT0Ytn0ZHk8e4sNp9ClqaU7VQPmLj4hi85Q8sjQyZ3Kw+FgYG/O/PAwDcffqc8b/v40bEU0z09elevTIbe3rQatl67j2Pyu5qZZq5qSE62lpERqm2Q2RUDFYWqfSHyNd8v2QfN+48RldXG3fXsiyY2hGviZu5dD3pvuPypfLRvEF5eo1Ym+11EFnrQ5/45DPiRTRWlql/Rnz/4z5u3En6jHB3K8OCaR54fbPpoz6Rn+YNy9Nr+Jpsr0NWsDRI+ox48kb1vfEkJgabgurb4WN5jIxxK1SEYQd2ZVcRxWckI6Mi2+TNm5d+/fpRunRp2rRpQ1BQEM7OziqjlrVr1+b169c8eKB+cY/0WFlZ4enpibu7Oy1btmThwoX/Z+/O42LO/ziAv6b7vg9JOiSJROWMFCH3tdjVUop1XxH6OYtdy7qtc1FhF7vkWFeIinJVQkr3QYcklXSoZn5/zDaMmZLdmu8w7+fjMY/V9/udmff3M9/57ry/n/fn8xXoJTExMeEldQBgYGCAgoICANweyOrqajg4OPDWy8rKolu3bkhMTAQAODo64s2bN3jw4AHCw8PRt29fODs7Izw8HAAalYxmZ2fDxcUFK1as+GQi2piY3717h549e/K1g7Ce2U6d3pfEsVgstGjRgvc6MTExSE1NhaqqKlRUVKCiogItLS1UVlYiLS3tk23r7e2NqVOnwsXFBT///DOvHPRzJCYm8rU9wD0m6tr+Yy9fvsSzZ8/g5eXFi1lFRQXr1q3jvb+uri4GDBiA33//HQC3LPr27dtwc+NOEhMbGwsOhwMLCwu+1wgPD+fbBzk5Ob72A7ilrKmpqbhz5w4Abu/j+PHjoays3OD+pKSk8JWp/pfPRRgDAwMMHToUhw4dAgCcP38elZWVGDduHC8uIyMjXiIKAFZWVtDQ0Ki3rRtLSUmJl4gCgL6+PkxMTKCiosK3rG7/GtP+MTExCAwM5Fs/aNAgsNlsZGTw9zp+rvXr10NdXZ3vkZkV9p9eU8DHV6NZLHDquUL9/FkRLl54iJSUF0hIyMGO7SG4ezcV48d3b9qYGCDQDGDVWzXG+qeNFp++hMe5LxCRmomfr0RgdOcOvN7Rhzn5OPf4KZJeFCImOwcLTp5H5qvX+L5r5+bcjf+M8/Fes8BXufGhZ7mv8ffVx0hOL8CTpDxs2X8Nt2PS8e1IewCAooIsViwYio17rqDkjeA4UvJlEPz0WZ84Jh79c0zkYsu+a7gdk4ZvR3EvIioqyGLFwiHYuDvkyzsmPtpn7s/DT3fnjbPsgNKqSlzJSPnktl8kCSvTpZ5R0qxkZGQgI8M9zDgcjsDYrrqTb91yKSkpgRNydXV1g+8REBCAefPm4fLlyzhx4gRWrFiBq1evokePHgAgUDrKYrHAZrOFvv+HcdUtU1dXR+fOnREWFoaoqCj069cPffr0QVxcHFJSUpCcnNzguDqAmyC1bNkSx48fh5eXF9TU1BrcvjExN0ZDr8Nms2FnZ8dL2j6OF2i4bdesWYOJEyfiwoULuHTpElavXo3jx49j9OjRjY6vLqYPCTtO6tTF/ttvv6F7d/4f69LS78v53NzcMH/+fOzcuRN//PEHOnTowOvhZLPZkJaWRkxMDN9zAPAlUIqKigJx6OnpYfjw4QgICICZmRkuXrzIN9axoWP8Q//1cxFm6tSpmDRpErZu3YqAgABMmDABSkpK9cb18fJ/892rb18+tX+fan82m43p06dj3rx5Au/3XydA8/X1hbe3N9+ykSO2/6fXrFNSUo7aWjY0P+oF1dRQEugtbUhCQi5cXDo0SUxMeF1egRo2GzrKSnzLtZWV6h3j+bLsLV68KUNZ1ftS9LTCIkixWGihpiq055MD4HHuC5hoazRh9E2n5E0FamrZAr2gmupKeF3S+LGuT5JzMbCvFQDAsIUGWuqr4+f/vT/PSv3zHb5x0htucw4iN1/8xwtKqgaPieLPOCaS8t4fEwaaaKmvgZ+Xj+Gt5x0TpxbBbfZB5OYX//fgm9DrSu45QleJvx10FJVQWP7pdhhnaY3TyQmo/uf/K1+bL6W8tqlQMkpExsrKCqdOneL7ARwVFQVVVVUYGnLHSujq6vL1vpWWljaqJ6RLly7o0qULfH190bNnT/zxxx+8ZLQh5ubmkJOTw61btzBx4kQA3B/g0dHRfLeMcHJywo0bN3D37l34+/tDQ0MDVlZWWLduHfT09ARKgz+mqKiI8+fPY8iQIRg0aBCuXLnC1/P5OczNzSErK4s7d+7wfpi/fv0aycnJjSoXrmNra4sTJ05AT0+vweS4oba1sLCAhYUFFi5ciO+++w4BAQH1JqOysrJ8vYMAt+z31q1bmDx5Mm9ZVFRUve2pr68PQ0NDpKen83o6hRk1ahSmT5+Oy5cv448//sCkSZP49qe2thYFBQW8Mt7PMXXqVHz77bdo1aoV2rRpw9cTamVlxRtL/OH+WFhYCCRe9Wns5/KxIUOGQFlZGXv27MGlS5cQEfF+1kErKytkZ2fj2bNnvN7RhIQElJSU8NpaV1cX8fH848/i4uLqHQf8bzWm/W1tbfHkyROYm5vX+zpycnICx1NjyMvLQ15enm9ZU5ToAkBNDRvJyfmwszNF5K33Mzza2ZkiMiq5gWfya2uuj6KisiaJiQnVbDae5L2Ag5kxriW9783vZdYaoUnCe/djn+XC1aotlGRlUf7PRRBTLQ3UstnIr2f2XQBor6+L5ILCpt2BJlJTw0Zy2gt0tTHBzbupvOVdbUxw615qA8/kZ2Gqj1f/XMzIzinC5PmBfOunTXSAkqIctv8zEQ4RX9xjIh9dOxvj5t33vXpdOxvj1t3POCbM9PDqNfcckf38FSbP458zYJpbb+4xceA6CgrFb0xlNZuN+Jcv0NvIBCEZ7/e7dysTXM1suB16tDSCqYYmTlym8dJfCyrTJSIza9YsPHv2DHPnzsXTp09x9uxZrF69Gt7e3rzxov369cORI0dw8+ZNxMfHw93dvcEf8RkZGfD19cXt27eRlZWFK1euIDk5+ZPJYR1lZWXMnDkTPj4+uHz5MhISEjBt2jSUl5fDy8uLt52TkxMuX74MFosFKysr3rLff/+90QmgsrIyLly4ABkZGQwePBhlZf/ux6aKigq8vLzg4+OD0NBQxMfHw8PD47PH3Lq5uUFHRwcjR47EzZs3kZGRgfDwcMyfPx/Pnz9vsG0rKiowZ84chIWFISsrC5GRkbh//36D7W5iYoLQ0FDk5+fj9evXALgTMQUGBmLv3r1ISUnBli1bEBwc3GAp85o1a7B+/Xps374dycnJePz4MQICArBlyxbeNsrKyhg5ciRWrlyJxMRE3oUGgJtAu7m5YfLkyQgODkZGRgbu37+PDRs24OLFi59st0GDBkFdXR3r1q3jTVxUZ9GiRQgNDcXatWuRnJyMoKAg/Prrr40qza7zqc+lPtLS0vDw8ICvry/Mzc35yrhdXFzQqVMnuLm5ITY2Fvfu3cPkyZPRt29f2Ntzy//69euH6OhoHD58GCkpKVi9erVActoUGtP+S5cuxe3btzF79mxeBcK5c+cwd+5c3uuYmJggIiICOTk5KCzkJiM5OTmwtLTEvXv3mjzuxjr51z0MGWIDV9dOaN1aGzNn9Yeevhr+/vsBAMBral8sXTaMt/2YsV3h4NAWhoaaMDbRgdfUvnDsa4kzZ95PSCYjI4U2bfTQpo0eZGSkoaOjgjZt9NCypabI96+xAm7H4hvbjhjbuQPMdLTgO7AvDNRVcTzmEQDAu58DNnxwz9Dzj5+iuLwS60cORBsdLdi3NoTPAEecinuCqhruRYfZjj3Qu40xWmmow1JfFz8NHwDLFro49s9riqMT56IxzMUaQ/p3hHErLcyd4gQ9HVWcCXkIAJj+fR8snzeYt/24Ybbo080crQw0YGKkjenf94FTLwsEX+QeP++qa5GRXcj3KHtbhfKKd8jILkRNzZffU/S2HEhM4T4A4Hke99+5Lxp+3pfixNloDHPp9P6Y8HSGno4a/zEx//29x8cNt0Of7h8fE+2++GPiwMNoTGhvjXGWHdFGUwsrHZzQUlUVv8dz22FJjz7Y3H+wwPPGt++IB/m5SC4SvAglKyUFK21dWGnrQlZaGvrKqrDS1oWxmkZz707TojJdQpqHoaEhLl68CB8fH9jY2EBLSwteXl5YsWIFbxtfX1+kp6dj2LBhUFdXx9q1axvsGVVSUsLTp08RFBSEV69ewcDAAHPmzOHd37Qxfv75Z7DZbEyaNAlv3ryBvb09QkJCoKn5/ode3a07+vbty+vV7du3L7Zt2/ZZvZEqKiq4dOkSBg0ahCFDhuDSpUuNfu6HfvnlF5SVlWHEiBFQVVXFokWLUFLyeaVZSkpKiIiIwNKlSzFmzBi8efMGhoaG6N+/P9TU1FBRUVFv29bU1ODVq1eYPHkyXrx4AR0dHYwZMwZ+fn71vt/mzZvh7e2N3377DYaGhsjMzMSoUaOwfft2/PLLL5g3bx5MTU0REBDQYNnz1KlToaSkhF9++QVLliyBsrIyrK2t+XqyAW5SN3ToUDg6OgqUdgYEBGDdunVYtGgRcnJyoK2tjZ49e2LIkCH4FCkpKXh4eOCnn37i69EFuD16f/75J1atWoW1a9fCwMAA/v7+fJMXfcqnPpeGeHl58W4b8yEWi4UzZ85g7ty5cHR0hJSUFFxdXbFz507eNoMGDcLKlSuxZMkSVFZWwtPTE5MnT8bjx00/M+On2r9Tp04IDw/H8uXL0adPH3A4HLRp04Zv9l9/f39Mnz4dbdq0QVVVFTgcDqqrq5GUlITyRpR5NZewsESoqSli0mQHaGmpIDPzJXx9/0TBC27vhLaWCvT03n+OsjJSmD6jP3R0VFBVVYOszEL4+v6Je3ff9yBqa6ti/2/vL45NmNADEyb0QFxcFhZ5/yG6nfsMlxKSoamkgFmO3aGnoozkglf44Y8zyC3h9tzpqijDQP19dUh5dTU8j57CisHOODVtIorLK3EpIRnbbkTytlFTkIf/UBfoqijhTdU7JOQX4Pugv/BYjLOU65FJUFNVhMf4ntDWVEZGdiGWrAvGi5f/HA+aytDX/fB4kMYsj77Q1VJB1bsaZDx7BZ+1p3An9r+Nlf6SPEkC3Be8H1awYRf336NcOVjvy1RUTed6ZBLU1BThMaHX+2Ni7an3x4SWCvR13383uMeE00fHxEncifmyj4nzqUnQkFfEfPue0FVWRvKrQkw5H8ybHVdPSRmGKvz/z1OVk8NgMwv43bou9DX1lVVwcYI77+/pXbpiepeuuJPzDN+ePdF8O9PUvpAksqmwOJ8zAI0QQgimTZuGFy9e4Ny5c0yHwicyMhJOTk54/vw59PX1mQ7ni9G/33qmQxALz50UmA5BLOg9+ILvT9iEwvfs//RGEqDvjGlMhyAWng0Un/s5MylzVuMrnf4tq/9tbbLXSvhpYZO9VnOhnlFCCGmkkpIS3L9/H7///jvOnj3LdDg8VVVVePbsGVauXInx48dTIkoIIYR8oSRtAiMaM0oIIY00cuRIjBgxAtOnT+fd41UcHDt2DO3atUNJSQk2btzIdDiEEEII+bdozCghhBBhPryNizjx8PD4rHGphBBCCCHigJJRQgghhBBCCBEHX0iPZlOhZJQQQgghhBBCxICkjRmlZJQQQgghhBBCxIGEJaM0gREhhBBCCCGEEJGjnlFCCCGEEEIIEQNUpksIIYQQQgghRPQkLBmlMl1CCCGEEEIIISJHPaOEEEIIIYQQIg4krGeUklFCCCGEEEIIEQMspgMQMSrTJYQQQgghhBAictQzSgghhBBCCCHigMp0CSGEEEIIIYSImqTd2oXKdAkhhBBCCCGEiBz1jBJCCCGEEEKIOKCeUUIIIYQQQgghIsdpwsdn2r17N0xNTaGgoAA7OzvcvHmzwe2rqqqwfPlyGBsbQ15eHm3atMGhQ4c+6z2pZ5QQQgghhBBCxABTY0ZPnDiBBQsWYPfu3XBwcMC+ffswePBgJCQkoHXr1kKfM378eLx48QIHDx6Eubk5CgoKUFNT81nvS8koIYQQQgghhEiwLVu2wMvLC1OnTgUAbNu2DSEhIdizZw/Wr18vsP3ly5cRHh6O9PR0aGlpAQBMTEw++32pTJcQQgghhBBCxEETlulWVVWhtLSU71FVVSXwlu/evUNMTAwGDhzIt3zgwIGIiooSGua5c+dgb2+PjRs3wtDQEBYWFli8eDEqKio+a3cpGSWEEEIIIYQQMcDiNN1j/fr1UFdX53sI6+UsLCxEbW0t9PX1+Zbr6+sjPz9faJzp6em4desW4uPjcfr0aWzbtg0nT57E7NmzP2t/qUyXEEIIIYQQQr4yvr6+8Pb25lsmLy9f7/YsFovvbw6HI7CsDpvNBovFwu+//w51dXUA3FLfb775Brt27YKiomKjYqRklBBCCCGEEELEQRNOYCQvL99g8llHR0cH0tLSAr2gBQUFAr2ldQwMDGBoaMhLRAGgffv24HA4eP78Odq2bduoGKlMlxBCCCGEEELEQFOW6TaWnJwc7OzscPXqVb7lV69eRa9evYQ+x8HBAbm5uSgrK+MtS05OhpSUFFq1atXo96aeUUIIIRItc+inrxpLAvPjpUyHIBbetlZmOgSx0HfGNKZDEAvhe39jOgSxYL9mJtMhkGbm7e2NSZMmwd7eHj179sT+/fuRnZ2NGTNmAOCW/Obk5ODw4cMAgIkTJ2Lt2rWYMmUK/Pz8UFhYCB8fH3h6eja6RBegZJQQQgghhBBCxAND9xmdMGECXr16BX9/f+Tl5aFjx464ePEijI2NAQB5eXnIzs7mba+iooKrV69i7ty5sLe3h7a2NsaPH49169Z91vtSMkoIIYQQQggh4oChZBQAZs2ahVmzZgldFxgYKLDM0tJSoLT3c9GYUUIIIYQQQgghIkc9o4QQQgghhBAiBj5n4qGvASWjhBBCCCGEECIOKBklhBBCCCGEECJqLI5kZaM0ZpQQQgghhBBCiMhRzyghhBBCCCGEiAPJ6hilZJQQQgghhBBCxIGkTWBEZbqEEEIIIYQQQkSOekYJIYQQQgghRBxIWM8oJaOEEEIIIYQQIgaoTJcQQgghhBBCCGlm1DNKCCGEEEIIIeJAwnpGKRklhBBCCCGEEDFAZbqEEEIIIYQQQkgzo55RQgghhBBCCBEHEtYzSskoIYQQQgghhIgBSSvTpWSUEEIIIYQQQsQBR7KyURozSgghhBBCCCFE5KhnlBBCCCGEEELEAJXpEkIIIYQQQggRPQlLRqlMlxBCCCGEEEKIyFHPKCGEEEIIIYSIARab6QhE64vpGXVycsKCBQuYDqNBmZmZYLFYiIuLYzoUAR4eHhg1atR/eo3AwEBoaGg0STzNhcVi4cyZM0yHIcDExATbtm1jOowms3//fhgZGUFKSgrbtm3DmjVr0Llz5//8us3VTh8f/597PgkLCwOLxUJxcXG92zTF9+PffE/z8/MxYMAAKCsri/3380NfwvmEEEIIETlOEz6+AI3uGWWxWA2ud3d3R2Bg4H+Nh/xHmZmZMDU1xYMHD5okOfjQhAkTMGTIkCZ9zbCwMDg7O+P169di+8PUxMQECxYsEPuLIZ/SVPtRWlqKOXPmYMuWLRg7dizU1dXBZrMxd+7cpglUBIKDgyErK8t0GAK2b98OzmdO6b5161bk5eUhLi4O6urqzRTZfyPs2GuO84m4cLOxwbSu9tBTVkbKq1dYeyMM0Tk59W4vJy2NuT16YKRVe+goKSG/rAy7797FyfgnAAAZKSnM6NYNYzpYoYWKCtKLXmPjzZuIyMwU0R79O8O+sce473tBS0cVWekF2LslBPFx2Z98nlUnI2za54HM9ALMctvHt270d90xdKw99PTVUVpSjpuhiTi06xqq39U21278Z6MHdcZ3I7tCW1MZmc8KsT3gBh4lCj8eunQwwk7/CQLLJ847hOycIoHl/R3awc97OCLupeB/G842eexNadTgzvhuVFdoa6og81khdhy8jkcJwtuhc0cj7Fz3rcByt9kHhbdDb0usWTwcN++m4H/rzzR16Iy4/xA4dAx4kgy8fMXCznUcuPRhOqqmNa5vJ0weYA8ddWWk577Cpr/C8SBV+DFhZ9EKv3mPE1g+ZnUgMl+8BsA9V05x7YphPa2gp6GCrBevsSP4JqISspp1P8h/0+hkNC8vj/fvEydOYNWqVUhKSuItU1RUbNrIRKC6ulosf5CKK0VFxQY/Z2pPyZCdnY3q6moMHToUBgYGvOUqKioij+Xdu3eQk5P77OdpaWk1QzT/3b9JJtPS0mBnZ4e2bdv+6/dl4rv7qfPJl2poOwuscHbC6tBQxOTk4rtOnXBozGgMCgxC3ps3Qp+zY9hQ6CgpY1nIFWQVF0NbSQkyUu8vAHs7OGBk+/ZYfvUq0oqK0MfEGHtGjMC448eQUPBSRHv2efoO6IAZ3q74dcMFPHn4DEPH2GHddjdMG78LL1+U1vs8JWV5+PiNwoP76dDU5j+nOLtaw3O2C7asPYuER89g2Fobi1ePAgDs2xrSnLvzr/Xr1Q7zpjhj82/X8PhpDkYOtMGm5WMxaUEAXhQKPx4A4Ls5B/G2oor3d3FphcA2+rpqmO3uhLiEZ80Se1Pq59AO8zz7Ycu+q3j8NAcjBtngl5XfYNLcQyhooB0mzjqAt+XveH8Xl5YLbKOvq4ZZHk6IeyL+7fA5KiqAdubA6CHA/JVMR9P0BtpZYPE4J6w/dh0P03Ixto81ds4ZhW/8DiP/df3HxKhVAXhb+f6YeP3m/Xdj1sheGNK9PdYevYrM/NfoaWWMTTNGYMovx5H0TDzPlcJI2my6jS7TbdGiBe+hrq4OFovFtywiIgJ2dnZQUFCAmZkZ/Pz8UFNTw3v+li1bYG1tDWVlZRgZGWHWrFkoKyvje4/IyEj07dsXSkpK0NTUxKBBg/D69WveejabjSVLlkBLSwstWrTAmjVr+J5fUlKCH374AXp6elBTU0O/fv3w8OFD3vq6UsJDhw7BzMwM8vLy4HA4OHnyJKytraGoqAhtbW24uLjg7du3Qtvh9evXcHNzg66uLhQVFdG2bVsEBATwbZOeng5nZ2coKSnBxsYGt2/f5lt/6tQpdOjQAfLy8jAxMcHmzZt563bu3Alra2ve32fOnAGLxcKuXbt4ywYNGgRfX1+h8ZmamgIAunTpAhaLBScnJ771mzZtgoGBAbS1tTF79mxUV1fz1r179w5LliyBoaEhlJWV0b17d4SFhfHWf1xWV197fiwrKwvDhw+HpqYmlJWV0aFDB1y8eBGZmZlwdnYGAGhqaoLFYsHDwwOA8HLNzp07833mKSkpcHR0hIKCAqysrHD16lWB987JycGECROgqakJbW1tjBw5Epkf9CbUlUXW1y5OTk7IysrCwoULwWKxeBUC9e1TQ8rLy+Hp6QlVVVW0bt0a+/fv51v/+PFj9OvXj3cc/vDDD7zvyOPHjyElJYXCwkIA3ONQSkoK48a9v0q4fv169OzZU+h717cfQMPH48cCAwN5x6eZmRlYLBYyMzMFynQ/1a4AUFBQgOHDh0NRURGmpqb4/fffG2y/D193/fr1aNmyJSwsLAB8+nMW1h4f9tIdPXoU9vb2UFVVRYsWLTBx4kQUFBQIPC8yMhI2NjZQUFBA9+7d8fjx4wbj/fvvvxs8L9a3fx/GOW/evHrPeyYmJjh16hQOHz7M9/3Jzs7GyJEjoaKiAjU1NYwfPx4vXrzgPa++7y6LxcK+ffswbNgwKCkpoX379rh9+zZSU1Ph5OQEZWVl9OzZE2lpabzXSktLw8iRI6Gvrw8VFRV07doV165d49sHYceesDLdPXv2oE2bNpCTk0O7du1w5MgRvvUsFgsHDhzA6NGjoaSkhLZt2+LcuXMNfgai5mlnh78ex+PPx/FIKyrCurAw5L15AzcbG6HbO5qYoHurVvA6fRpR2dnIKS3Fo/x8xOa+vwA8yqo99ty7i7CMDDwrKcEfDx/hZlYmvOzsRbVbn23MxB4IOfsAl88+wLPMQuzdEoKXL0ow7JuuDT5v/v+G4UZIPBIfPxdYZ2XdCk8eZeNGSDxe5JUg9m46wq7Ew6K9gZBXEg/fDrfH+euPcT70MbJyirAj4AYKXr3BqEGdG3ze65JyFBW/f7DZ/P9vlZJiYfX8ITh4IhK5L0qacQ+axoSR9rhw7THOX3uMrOdF2HnwBgoK32C0a+cGn8dth7e8h7B2WLVwKA4dj0TeF9AOn8OxB7BgKjDQkelImoebiy3ORMbjTGQ8MvKLsOmvcLx4/Qbf9O3U4POK3lTgVWk578H+4Hfn0O7tcejSPUTGZyKnsAQnIx7hdkImJrnYNffuNC0Op+keX4AmGTMaEhKC77//HvPmzUNCQgL27duHwMBA/Pjjj+/fSEoKO3bsQHx8PIKCgnD9+nUsWbKEtz4uLg79+/dHhw4dcPv2bdy6dQvDhw9Hbe370pugoCAoKyvj7t272LhxI/z9/XkJCIfDwdChQ5Gfn4+LFy8iJiYGtra26N+/P4qK3pd0pKam4s8//8SpU6cQFxeH/Px8fPfdd/D09ERiYiLCwsIwZsyYekvlVq5ciYSEBFy6dAmJiYnYs2cPdHR0+LZZvnw5Fi9ejLi4OFhYWOC7777j/QCNiYnB+PHj8e233+Lx48dYs2YNVq5cyStxdnJywpMnT3hJR3h4OHR0dBAeHg4AqKmpQVRUFPr27Ss0vnv37gEArl27hry8PAQHB/PW3bhxA2lpabhx4waCgoIQGBjIV1o9ZcoUREZG4vjx43j06BHGjRsHV1dXpKSkCH0vYe0pzOzZs1FVVYWIiAg8fvwYGzZsgIqKCoyMjHDq1CkAQFJSEvLy8rB9+/Z63+tDbDYbY8aMgbS0NO7cuYO9e/di6dKlfNuUl5fD2dkZKioqiIiIwK1bt6CiogJXV1e8e/f+qlpD7RIcHIxWrVrB398feXl5vAqB+vapIZs3b4a9vT0ePHiAWbNmYebMmXj69CkvVldXV2hqauL+/fv466+/cO3aNcyZMwcA0LFjR2hra/OOg4iICGhrayMiIoL3+mFhYfUeF/Xtx6eOx49NmDCBl2jcu3cPeXl5MDIyErrtp443Dw8PZGZm4vr16zh58iR2794tNAH8WGhoKBITE3H16lWcP3++0Z9zQ969e4e1a9fi4cOHOHPmDDIyMniJ3Yd8fHywadMm3L9/H3p6ehgxYgRfgv2hxpwXG6Oh8979+/fh6uqK8ePH874/HA4Ho0aNQlFREcLDw3H16lWkpaVhwgT+0r/6vrtr167F5MmTERcXB0tLS0ycOBHTp0+Hr68voqOjAYB3XAJAWVkZhgwZgmvXruHBgwcYNGgQhg8fjuxsbjlmfcfex06fPo358+dj0aJFiI+Px/Tp0zFlyhTcuHGDbzs/Pz+MHz8ejx49wpAhQ+Dm5sZ3jmeSrJQUOurr41YWf0nYraws2LZsKfQ5/duY4fGLF/ihqz0if/gB16ZMgW9fR8jLvC9ckpOWRlUNfxlqZU0N7A2FvybTZGSk0NayJWLupvEtj7mbDqtOrep93sDhnWHQShNHfwsTuj4+LhttLVuinRV3v1sYaqBrL3Pcjaz//1FMkpGRgkUbfdyPy+Rbfv9hJjq2a/izO7RpMs4cmIFtq8ehS0fBc6zHuJ4oLq3AhdD4pgy5WXDboQXufdwOcZnoaGnY4HMPbpmMM4dmYpv/eOHtML4Xtx2uNXxhkIgXGWkptG+tjzuJ/OfK24nZsDFr+LtxbLkbQjb8gL0LxsLegv98Iisjjapq/gu+VdU16GwunudKwtUks+n++OOPWLZsGdzd3QFwe0zWrl2LJUuWYPXq1QDA1wthamqKtWvXYubMmdi9ezcAYOPGjbC3t+f9DQAdOnTge59OnTrxXq9t27b49ddfERoaigEDBuDGjRt4/PgxCgoKIC8vD4DbC3jmzBmcPHkSP/zwAwDuj84jR45AV1cXABAbG4uamhqMGTMGxsbGAMDXM/mx7OxsdOnSBfb23CvSJiYmAtssXrwYQ4cOBcD94dShQwekpqbC0tISW7ZsQf/+/bFyJbfmwsLCAgkJCfjll1/g4eHBl3SMHTsWYWFhWLRoEbZu3QqA++OzsrISvXv3Fhpf3X5pa2ujRYsWfOs0NTXx66+/QlpaGpaWlhg6dChCQ0Mxbdo0pKWl4dixY3j+/Dla/vOjafHixbh8+TICAgLw008/CX2/j9uzvjYbO3YsX49anbpyST09vc8aM3rt2jUkJiYiMzMTrVpxT0Y//fQTBg8ezNvm+PHjkJKSwoEDB3i9MQEBAdDQ0EBYWBgGDhz4yXbR0tKCtLQ0r8esMftUnyFDhmDWrFkAgKVLl2Lr1q0ICwuDpaUlfv/9d1RUVODw4cNQVlYGAPz6668YPnw4NmzYAH19fTg6OiIsLIx3XLi7uyMoKAgJCQmwsLBAVFQUFi5cKPS969uPTx2PH6vrtQW4x9rHx9iHGmrX5ORkXLp0CXfu3EH37t0BAAcPHkT79u0/2Y7Kyso4cOAArzz30KFDjfqcG+Lp6cn7t5mZGXbs2IFu3bqhrKyM7yLD6tWrMWDAAADcJLFVq1Y4ffo0xo8fL/CajTkvNkZD5z1dXV3Iy8tDUVGR91lcvXoVjx49QkZGBu9CwZEjR9ChQwfcv38fXbtye6bq++5OmTKFtz9Lly5Fz549sXLlSgwaNAgAMH/+fEyZMoW3vY2NDWw+6PVbt24dTp8+jXPnzmHOnDn1Hnsf27RpEzw8PHjfEW9vb9y5cwebNm3iVVAA3IsY3333HQDud37nzp24d+8eXF1dG9WeVVVVqKqq4lvGqakBS+a//+9QU1ERMlJSKCznr6wpfFsOXRMloc9pra4Be0NDVNXUYua5c9BSVIRf/35QV1DAspArAICbmVnwtLPF/efPkVVcjF7GreHSpg2kPjGXA1PUNJQgLSOF4iL+6qfiV2XQ1G4j9DktjbTgObs/Fv0QAHat8IvB4VefQF1TGZsPeILFAmRkpPH3yfv4MyiyyfehKairKkJGWgpFJfylpUXF5dDWUBb6nMLXZdiwJwRJaS8gKysN174dsH31eMxdfQIPE7i9xdbtWmJYf2tMWXS42fehKdS1w+ti/u/F65K30NIU3g6visqwcVcIktLyISsrg0FOVtjmPwHzVhx/3w6WhhjqYg3PhUHNvg+kaWmocI+JVx+VXReVvoW2mrHQ5xSWvMXao1eRmFUAORlpDOnRHnsXfIMftvyF2H/Gmd5OyML3LnaITc3B85fF6GbZGn1t2kBaTM+V9ZG0Mt0mSUZjYmJw//59viv+tbW1qKysRHl5OZSUlHDjxg389NNPSEhIQGlpKWpqalBZWYm3b99CWVkZcXFxfCWHwnTqxN91b2BgwOtJiYmJQVlZGe+Hcp2Kigq+kjJjY2O+H182Njbo378/rK2tMWjQIAwcOBDffPMNNDU1hcYwc+ZMjB07FrGxsRg4cCBGjRqFXr161Rtn3Zi6goICWFpaIjExESNHjuTb3sHBAdu2bUNtbS2kpaV5SUf//v3x5MkTzJgxA5s2beL13Nra2v6r8XkdOnSAtLQ0X2x1ZYaxsbHgcDi8ssc6VVVVAm36oY/bU5h58+Zh5syZuHLlClxcXDB27FiBz/JzJSYmonXr1rxEFIBAiWpMTAxSU1OhqqrKt7yyspLvmGioXerzb/bpw/V1Ze51x29iYiJsbGx4iSjAPS7YbDaSkpKgr68PJycnXmlveHg41q5di4yMDISHh6OkpAQVFRVwcHBoMIaPNeZ4/LcaatfExETIyMjwLuoAgKWlZaMuSFhbW/ONE23s59yQBw8eYM2aNYiLi0NRURHYbO686tnZ2bCysuJt9+ExpqWlhXbt2iExMVHoazbmvNgYDZ33hElMTISRkRFfj7WVlRU0NDSQmJjIS0br++5++H76+voA+C/Q6evro7KyEqWlpVBTU8Pbt2/h5+eH8+fPIzc3FzU1NaioqOD1jDZWYmIi76JhHQcHB4FqiQ/jU1ZWhqqqaqN61OusX78efn5+fMs0BgyE1j/JdlP4uLCGxap/UkMWi7v9wosXUfZPT/6PYeHYNWI4VodeR1VNDdbeuIGfBg7AlSke4ADILi7GySdP8M1HF2zFjWA7CG8IKSkWlq0bgyP7w5CTXX8vdydbY3zn2Qe/briAp/E5aGmkhZmLXFFUWIY/DkbU+zymfVxpxW0G4UfEs9zXeJb7fnjSk+Q86Omo4rsR9niY8ByKCrJYOX8oNu65gpI3guNIxZngHrPqrUITaIekXOjpqOLbUV157bBi4RBs3B3yxbUD+YCQc0R958qsF6+R9eL9MfEoIw8tNFUxaYAdLxn95c8wrPzeBcFr3MHhAM9fFuPvqCcY3ku8z5UCKBn9fGw2G35+fhgzZozAOgUFBWRlZWHIkCGYMWMG1q5dCy0tLdy6dQteXl68ErfGTGTx8QQbLBaL96ORzWbDwMCAb4xjnQ9/4H74Yx8ApKWlcfXqVURFReHKlSvYuXMnli9fjrt37/LGX35o8ODByMrKwoULF3Dt2jX0798fs2fPxqZNm4TGWddTUxdn3bisD318Mq5LOm7evAkbGxtoaGjA0dER4eHhCAsLExgH2lifaj9paWnExMQIJCANJb4ft6cwU6dOxaBBg3DhwgVcuXIF69evx+bNmxucfVVKSkqgXT4shxT2P7CP25XNZsPOzk7oWMQPf4Q31C5NuU8NvY+w4+Lj/XJycsL8+fORmpqK+Ph49OnTB2lpaQgPD0dxcTHs7OwEErJPaczx+G99an/rln2uj4+5xn7O9Xn79i0GDhyIgQMH4ujRo9DV1UV2djYGDRrUqDLf+vbhU+fFxvrc47O+Y+nj5fV9d4Wdvxo6p/n4+CAkJASbNm2Cubk5FBUV8c033zS6RPpDwo7Fj5f9m+/rh3x9feHt7c23rPOevZ8ZqXCvKypQw2ZD96O21VZSQuFbwYlXAODl27d4UVbGS0QBIK2oCFIsFgxUVJBZXIyiigrMOHsOctLS0FRUxIuyMizp0wfPSsRzjFxpcTlqa9gCExCpaynj9Ue9pQCgqCSHdlaGMLcwwGwf7gzLLCkWpKRYuHh7JXznHsHD6Ey4z3BG6MVHuHz2AQAgM60ACoqymP+/4Th2KELshkeVvKlATS1boBdUU10JRcXCjwdhniTnYqAj96KYYQsNtNRXx8++o3nr63rIw/70xsS5B8VuDGldO2gJaYfXn9MOSXkY2PefdjDQREt9Dfy8/P35ta4dbpxaBLfZB5GbX/zfgyfNorjsn++GOv9FWU1VJRQJmaSqPo8z8jCkmyXf6y7a+zfkZKShrqKAl8VvMW90b+QW1j9pmjiintF/wdbWFklJSTA3Nxe6Pjo6GjU1Ndi8eTOkpLjDVP/880++bTp16oTQ0FCBK9afE0N+fj5kZGSEls42hMViwcHBAQ4ODli1ahWMjY1x+vRpgR8sdXR1deHh4QEPDw/06dOHN46sMaysrHDr1i2+ZVFRUbCwsOAlgXVJx8mTJ3mJZ9++fXHt2jVERUVh/vz59b5+XY/Rh2NtG6NLly6ora1FQUEB+vRp+rnDjYyMMGPGDMyYMQO+vr747bffMHfu3Hrj1dXV5RtbVlpaioyMDN7fVlZWyM7ORm5uLq+s+OOJomxtbXHixAnehFb/lpycnND2rG+f/g0rKysEBQXxKgUA7mQ5UlJSvN7quhLudevWwcbGBmpqaujbty/Wr1+P169f1ztetKH9aMzx2Bzat2+PmpoaREdHo1u3bgC444Ybuo9nff7r5/z06VMUFhbi559/5vUm1o2N/NidO3fQunVrANxJpJKTk2FpaSl020+dF5tL3Xfj2bNnvP1JSEhASUlJo8qgP9fNmzfh4eGB0aO5P47LysoEJo+q7zv0ofbt2+PWrVuYPHkyb1lUVFSTxywvL88bylGnKUp0AaCazUb8ixdwMG6NK6mpvOUOxsa4liq8lz4mJxeDLSygJCuL8n8uuJlqaqKWzUbeR5P8vautxYuyMshIScG1bVtcSE4S9pKMq6lhI+VpLmy7myEq7ClvuW03M9yOEIy5/G0Vfvh2N9+y4d90RWd7U6xd9ifyc4oBAPIKsuB8NIENu5YDFv7pURGzbLSmho3ktBfoamOCiHvvjwf7Tia4dT+1gWfya2uqj1evuSWu2TlFmLQgkG/9tIkOUFKQw/ZD3MmRxA23HfLRtbMxbt59P763a2dj3Lrb+HawMNPDq9fc70T281eYPI9/8shpbr2hpCiH7Qeuo+ALSz4kTU0tG4nZL9C9vTFuxL0/N/Zo3xphDxtX0QQA7Yz0UFgiOOHou5pavCx+CxkpKfTv0hZXY5KbJG7SPJpkAqNVq1bh8OHDWLNmDZ48eYLExEScOHECK1asAAC0adMGNTU12LlzJ9LT03HkyBHs3ct/JdrX1xf379/HrFmz8OjRIzx9+hR79uzhTeTzKS4uLujZsydGjRqFkJAQZGZmIioqCitWrKj3hyUA3L17Fz/99BOio6ORnZ2N4OBgvHz5st4fQKtWrcLZs2eRmpqKJ0+e4Pz585/1Y2nRokUIDQ3F2rVrkZycjKCgIPz6669YvHgxb5u6pOP333/nJaNOTk44c+YMKioq6h0vCnDHXioqKuLy5ct48eIFShp55dzCwgJubm6YPHkygoODkZGRgfv372PDhg2fnCX2UxYsWICQkBBkZGQgNjYW169f57WZsbExWCwWzp8/j5cvX/Jmj+3Xrx+OHDmCmzdvIj4+Hu7u7nzJkYuLC9q1a4fJkyfj4cOHuHnzJpYvX873vm5ubtDR0cHIkSNx8+ZNXknr/Pnz8fy54EyN9TExMUFERARycnJ4x2ND+/RvuLm5QUFBAe7u7oiPj8eNGzcwd+5cTJo0iVcqyWKx4OjoiKNHj/KOi06dOuHdu3cIDQ39ZI+5sP1ozPHYHNq1awdXV1dMmzYNd+/eRUxMDKZOnfqvbvXxXz/n1q1bQ05Ojnd+OnfuHNauXSt0W39/f4SGhiI+Ph4eHh7Q0dHhm/32Q586LzYXFxcXdOrUCW5uboiNjcW9e/cwefJk9O3bl68suqmYm5sjODgYcXFxePjwISZOnCjQUyns2PuYj48PAgMDsXfvXqSkpGDLli0IDg7+7GPR0tISp0+f/tf7818dionBeGtrfNOxA9poaWG5U1+0VFXFH//M7L64d29s+mB867mnT1FcWYkNgwbBXEsLXQ0NsczRESfjn6Dqn4nvbFq0wEBzcxipq8Pe0BABY8aAxQL236///21MC/7jDlxH2mLg8M4wMtHB9IWDoNdCHRdOcWOeMrs/fNaMAsAt581Ke8n3KC56i3fvapCV9hJVldwk/c7NZAwda4++AzpAv6UGbLuZwX2GM+7cTBKYZVVcHP87GsP6W2Nov44wNtTCXA8n6Ouo4swV7vEw3a0PVsx9P9fBuKG26NPNHK0MNGBqpI3pbn3g3NMCpy5xe4PfVdci41kh36PsbRXKK98h41khamoaXyUgSifORmOYSycM6d8Rxq20MNfTGXo6ajgT8k87fN8Hy+e/v+/wuOF26NOd2w4mRtqY/n0fOPVqh+CLH7RDdiHfo+xtFcor3iEjW3zb4XO8LQcSU7gPAHiex/137ouGn/el+P1aLEY7dMTIXh1g2kILi8b1RQtNVZyKeAQAmDPKAf4e74dPTOzXBU42bWCkpwEzA23MGeUAF9u2OBH2/q4ZHU1aoF9ncxjqqKOLuSF+nTcaLBYLgVfE91wplITNptskl4MHDRqE8+fPw9/fHxs3boSsrCwsLS0xdepUANxbcmzZsgUbNmyAr68vHB0dsX79er4r4BYWFrhy5Qr+97//oVu3blBUVET37t15E1V8CovFwsWLF7F8+XJ4enri5cuXaNGiBRwdHXk/5oVRU1NDREQEtm3bhtLSUhgbG2Pz5s18E+F8SE5ODr6+vsjMzISioiL69OmD48ePN7qtbG1t8eeff2LVqlVYu3YtDAwM4O/vzzdZDIvFQt++fXHmzBleL2WnTp2grq4OMzOzBnt/ZGRksGPHDvj7+2PVqlXo06eP0NJlYQICArBu3TosWrQIOTk50NbWRs+ePf/zjelra2sxe/ZsPH/+HGpqanB1deVNyGRoaAg/Pz8sW7YMU6ZMweTJkxEYGAhfX1+kp6dj2LBhUFdX542PrCMlJYXTp0/Dy8sL3bp1g4mJCXbs2ME3iYmSkhIiIiKwdOlSjBkzBm/evIGhoSH69+//WT1o/v7+mD59Otq0aYOqqipwOJwG9+nfUFJSQkhICObPn4+uXbtCSUkJY8eOxZYtW/i2c3Z2RnBwMC/xZLFY6NOnD86fP9/gRYr69qMxx2NzCQgIwNSpU9G3b1/o6+tj3bp1vImUPsd//Zx1dXURGBiI//3vf9ixYwdsbW2xadMmjBgxQmDbn3/+GfPnz0dKSgpsbGxw7ty5eu9z+qnzYnNhsVg4c+YM5s6dC0dHR0hJScHV1RU7d+5slvfbunUrPD090atXL+jo6GDp0qUoLeXvlRB27H1s1KhR2L59O3755RfMmzcPpqamCAgI+OxhCUlJSY2+CNccLiQlQ0NBEXN79ICusjJSXr2CV/Bp5P5zj1E9ZWUYqL0vpy+vrsbkkyexul8/nPneDcWVlbiQlIQtkVG8beRlZODd2wGt1dXxtroa4ekZWHTpEt58NBGTOAm/+gSq6opwm9oXWjoqyEorwIoFv6Mgn/vZaOmoQLfF591X949/SnE9ZvaDtq4qSorLcedmMgJ3hzbHLjSJ61FJUFdVhMe4ntDWVEZGdiF8fgrGi5fc74i2pjL0dd6fp2RlpDF7cl/oaqmg6l0NMp69wuIfT+FObEZ9b/FFuB6ZBDU1RXhM6MVrhyVrT71vBy0V6Ou+/17IykhjlocTXzv4rD2JOzFfdjt8jidJgPuC98MUNuzi/nuUKwfrhd/d74tyJSYZ6ioKmDa0O3TUlJGW+wrzfj2DvCLuuVJHXRkttPiPiYVjHaGroYKq6hqk577C3F9PIzI+k7eNnKw0Zo3sBUMddZRXVSMyPgMrAi6jrEJ8z5XCSFqZLosjbnUthBBCiAi12bzl0xtJAPPj4lfiyYS3rT89F4IkYNUzq7GkCd/7G9MhiAX7NTOZDkEsxO4VfteCpuQ48pcme62Isz5N9lrNpWkGyhBCCCGEEEII+W8k7DoQJaOEEEIIIYQQIgYkrUy3SSYwIoQQQgghhBBCPgf1jBJCCCGEEEKIOBDT2cGbCyWjhBBCCCGEECIOJCsXpTJdQgghhBBCCCGiRz2jhBBCCCGEECIGJG0CI0pGCSGEEEIIIUQccCQrG6VklBBCCCGEEELEgKT1jNKYUUIIIYQQQgghIkc9o4QQQgghhBAiDiSsZ5SSUUIIIYQQQggRAywJGzNKZbqEEEIIIYQQQkSOekYJIYQQQgghRBywmQ5AtCgZJYQQQgghhBAxQGW6hBBCCCGEEEJIM6NklBBCCCGEEELEAacJH59p9+7dMDU1hYKCAuzs7HDz5s16tw0LCwOLxRJ4PH369LPek8p0CSGEEEIIIUQcMFSme+LECSxYsAC7d++Gg4MD9u3bh8GDByMhIQGtW7eu93lJSUlQU1Pj/a2rq/tZ70s9o4QQQgghhBAiwbZs2QIvLy9MnToV7du3x7Zt22BkZIQ9e/Y0+Dw9PT20aNGC95CWlv6s96VklBBCCCGEEELEAIvTdI+qqiqUlpbyPaqqqgTe8927d4iJicHAgQP5lg8cOBBRUVENxtulSxcYGBigf//+uHHjxmfvLyWjhBBCCCGEECIOOJwme6xfvx7q6up8j/Xr1wu8ZWFhIWpra6Gvr8+3XF9fH/n5+ULDNDAwwP79+3Hq1CkEBwejXbt26N+/PyIiIj5rd2nMKCGEEEIIIYSIAVYT3mfU19cX3t7efMvk5eXrf28Wi+9vDocjsKxOu3bt0K5dO97fPXv2xLNnz7Bp0yY4Ojo2OkZKRgkhhBBCCCHkKyMvL99g8llHR0cH0tLSAr2gBQUFAr2lDenRoweOHj36WTFSmS4hhBBCCCGEiIMmLNNtLDk5OdjZ2eHq1at8y69evYpevXo1+nUePHgAAwODRm8PUM8oIYQQQgghhIgHZu7sAm9vb0yaNAn29vbo2bMn9u/fj+zsbMyYMQMAt+Q3JycHhw8fBgBs27YNJiYm6NChA969e4ejR4/i1KlTOHXq1Ge9LyWjhBBCJFrbI8VMhyAW8ntrMh2CWFB7VsN0CGIhe5DwcWKSxn7NTKZDEAvRaxq+vYfkWMh0AM1mwoQJePXqFfz9/ZGXl4eOHTvi4sWLMDY2BgDk5eUhOzubt/27d++wePFi5OTkQFFRER06dMCFCxcwZMiQz3pfFofD0J1VCSGEEDHg2nkV0yGIBUpGuSgZ5aJklEvr0efdM/FrRckol1SL5GZ/j4E91zbZa125vbLJXqu5UM8oIYQQQgghhIgDCesnpAmMCCGEEEIIIYSIHPWMEkIIIYQQQog4aML7jH4JKBklhBBCCCGEEDHAojJdQgghhBBCCCGkeVHPKCGEEEIIIYSIAwnrGaVklBBCCCGEEELEASWjhBBCCCGEEEJETsImMKIxo4QQQgghhBBCRI56RgkhhBBCCCFEDEjabLqUjBJCCCGEEEKIOJCwZJTKdAkhhBBCCCGEiBz1jBJCCCGEEEKIOJCwnlFKRgkhhBBCCCFEHEhYMkpluoQQQgghhBBCRI56RgkhhBBCCCFEHEjYfUYpGSWEEEIIIYQQMSBpt3ahMl1CCCGEEEIIISJHPaOEEEIIIYQQIg4krGeUklFCCCGEEEIIEQdsSkYJIYQQQgghhIiahPWM0phRQgghhBBCCCEiR8ko+c9YLBbOnDnT7O9jYmKCbdu2NfnrBgYGQkNDo8lfl/w3YWFhYLFYKC4uZjoUAHScMC0zMxMsFgtxcXFMh0IIIYQ0Hw6n6R5fACrTJXw8PDwQFBQksDwlJQXm5uZCn5OXlwdNTc3mDo2IsTVr1uDMmTNfTaJgYmKCBQsWYMGCBbxlEyZMwJAhQ5gLqhl5eHiguLj4kxeVnJyc0Llz52a5KPQ1GTa+K75x7w0tHRVkpb3E3l8u4cmDrE8+z6pza/xyYAoy0wowe8Ie3vKNB6agk72pwPb3biZj1dyjTRp7cxvfpxPc+9tDR10ZaXmv8MupcDxIy6l3e1kZaUwf3B1DuraHjqoSXhSX4UDIPZy980SEUf83owZ3xneju0JLUwWZ2YXYefA6HiUI3+fOHY2w48dvBZZ/P+sgsnOKAACOPdri+3E9YNhCAzIyUnieW4wTZ+/jSlhCs+7Hf/V9h86Y3qUr9JSUkVxUCP/IG7ifJ7wdNvVzxTeWHQWWJxcVYuDxQABAW01teHdzgLWuPlqpqcP/1nUcehTbnLvQZMb17YTJA7jfg/TcV9j0VzgepApvCzuLVvjNe5zA8jGrA5H54jUAQEZKClNcu2JYTyvoaagg68Vr7Ai+iaiET593xN39h8ChY8CTZODlKxZ2ruPApQ/TUTWzLySJbCqUjBIBrq6uCAgI4Fumq6srsN27d+8gJyeHFi1aiCo0ImY4HA5qa2uZDkMkFBUVoaioyHQYRMw5DuyI6T6Dseun83gSl40h33TFul3f44cxv+Jlfkm9z1NSkcfitWMQdy8DGtrKfOv8vY9DVlaa97eahiJ2n5iFm1fjm20/msNAWwv4jHXCTyeuIy49F9/0tsauWaMwZt1h5L9+I/Q5Gz2HQltVCX6/X8Wzl8XQUlWEtNSXU9TVr3c7zPXqhy37riI+MQcjBtlg46pvMHnOIRQUCt9nAJg48wDKy9/x/i4uLef9u7SsEkf+uoPs569QXcNGL3szLJs3GK9LynH/QWZz7s6/Nsy8HVb1dsbKiGuIzs+Bm5UNAoeNxYBjAcgtE2wHv1vXseF2BO9vaSkpXJrgjotpybxlirKyyC4twcW0JKx0cBbJfjSFgXYWWDzOCeuPXcfDtFyM7WONnXNG4Ru/+r8HADBqVQDeVr4/Jl6/qeD9e9bIXhjSvT3WHr2KzPzX6GlljE0zRmDKL8eR9Oxls+5Pc6uoANqZA6OHAPNXMh0NaQ5fzhmdiIy8vDxatGjB95CWloaTkxPmzJkDb29v6OjoYMCAAQAEy3RzcnIwYcIEaGpqQltbGyNHjkRmZiZvvYeHB0aNGoVNmzbBwMAA2tramD17Nqqrq3nbFBQUYPjw4VBUVISpqSl+//13gTjXrFmD1q1bQ15eHi1btsS8efPq3aeHDx/C2dkZqqqqUFNTg52dHaKjo/m2CQkJQfv27aGiogJXV1fk5eXx1rHZbPj7+6NVq1aQl5dH586dcfnyZd76sWPHYu7cuby/FyxYABaLhSdPuFfva2pqoKqqipCQEKHxZWVlYfjw4dDU1ISysjI6dOiAixcvAhBeHnrmzBmwWCy+tujcuTP27dsHIyMjKCkpYdy4cXwlrnXt7ufnBz09PaipqWH69Ol49+79/9yqqqowb9486OnpQUFBAb1798b9+/d56+tKZ0NCQmBvbw95eXkcOXIEfn5+ePjwIVgsFlgsFgIDAwX28fHjx5CSkkJhYSEA4PXr15CSksK4ce+v+K5fvx49e/bke15MTAzs7e2hpKSEXr16ISkpiW/933//DTs7OygoKMDMzAx+fn6oqanhrWexWDhw4ABGjx4NJSUltG3bFufOnRP6OQDc3r+srCwsXLiQtz/CPoe6Nj906BBat24NFRUVzJw5E7W1tdi4cSNatGgBPT09/Pjjj3yvX1JSgh9++IH3GfTr1w8PHz7krW/Msfqhp0+fonfv3lBQUICVlRWuXbv2Wd/JNWvWICgoCGfPnuXtb1hYmMD7eHh4IDw8HNu3b+dtV/ca4eHh6NatG+Tl5WFgYIBly5bxfQbCSuw7d+6MNWvWfNZ+AEB6ejqcnZ2hpKQEGxsb3L59u962YcKYSb0QcjoWl0/H4llGIfb9cgkv80sxbFzXBp83b8UIhF16hMRHzwTWlZVW4PWrMt6jSw9zVFZWI+LKl9M7CACT+tni9O14nL4dj4wXRfjlVDjyX7/BuD6dhG7fq70x7M0NMWfPadxNykZuUSnis17gYUae0O3F0fiR9rhw7TEuXH2MrOdF2HnwBl4WvsGowZ0bfF5xSTmKit/yHuwPZteMi3+Gm3dSkPW8CLn5xTh5PhbpmS/Rqb1hM+/NvzfVxh5/Jj7GicTHSHtdBP/IG8gre4PvO3YWuv2bd+/wsqKc9+ik1wLq8gr4K/H9BZhHBflYfzscf6cm4d0XdFHUzcUWZyLjcSYyHhn5Rdj0VzhevH6Db/oK/x7UKXpTgVel5bwH+4Pes6Hd2+PQpXuIjM9ETmEJTkY8wu2ETExysWvu3Wl2jj2ABVOBgY5MRyJCbE7TPb4AlIySzxIUFAQZGRlERkZi3759AuvLy8vh7OwMFRUVRERE4NatW7zk7sOk58aNG0hLS8ONGzcQFBSEwMBAvgTGw8MDmZmZuH79Ok6ePIndu3ejoKCAt/7kyZPYunUr9u3bh5SUFJw5cwbW1tb1xu3m5oZWrVrh/v37iImJwbJlyyArK8sX96ZNm3DkyBFEREQgOzsbixcv5q3fvn07Nm/ejE2bNuHRo0cYNGgQRowYgZSUFADcBObDH/Dh4eHQ0dFBeHg4AOD+/fuorKyEg4OD0Phmz56NqqoqRERE4PHjx9iwYQNUVFTq3R9hUlNT8eeff+Lvv//G5cuXERcXh9mzZ/NtExoaisTERNy4cQPHjh3D6dOn4efnx1u/ZMkSnDp1CkFBQYiNjYW5uTkGDRqEoqIivtdZsmQJ1q9fj8TERAwcOBCLFi1Chw4dkJeXh7y8PEyYMEEgvo4dO0JbW5vXJhEREdDW1kZExPur32FhYejbty/f85YvX47NmzcjOjoaMjIy8PT05K0LCQnB999/j3nz5iEhIQH79u1DYGCgQALo5+eH8ePH49GjRxgyZAjc3NwE9qlOcHAwWrVqBX9/f97+1CctLQ2XLl3C5cuXcezYMRw6dAhDhw7F8+fPER4ejg0bNmDFihW4c+cOAG5P8tChQ5Gfn4+LFy8iJiYGtra26N+/Py+eTx2rH2Kz2Rg1ahSUlJRw9+5d7N+/H8uXL+fb5lPfycWLF2P8+PG8CzB5eXno1auXwHtt374dPXv2xLRp03jbGRkZIScnB0OGDEHXrl3x8OFD7NmzBwcPHsS6devqbbd/sx91li9fjsWLFyMuLg4WFhb47rvv+BJfJsnISKNtewPE3k7jWx57JxXtbVrX+7wBI7ugpZEWju4La9T7DBpli/CQeFRVVn96YzEhIy2F9kb6uJ3IXzZ4JzEbNqYthT7HyboNnmQXwMOlK66sm4azqzywcHQfyH/QSyzOZGSkYNGmBe7HZfItvx+XiY6WDSeOB7dOxumAmdjqPx5drI0a3Na2U2sYGWri4ZPn/zXkZiErJYWOuvq4+SyTb/nNZ5mw0xf+2X9sfHtr3HqehZyy0maIUHRkpKXQvrU+7nz0PbidmA0bs4bb4thyN4Rs+AF7F4yFvUUrvnWyMtKoquY/D1ZV16CzeePal4gZDrvpHl8AKtMlAs6fP8+XCA0ePBh//fUXAMDc3BwbN26s97nHjx+HlJQUDhw4wOtRCggIgIaGBsLCwjBw4EAAgKamJn799VdIS0vD0tISQ4cORWhoKKZNm4bk5GRcunQJd+7cQffu3QEABw8eRPv27Xnvk52djRYtWsDFxQWysrJo3bo1unXrVm9c2dnZ8PHxgaWlJQCgbdu2fOurq6uxd+9etGnTBgAwZ84c+Pv789Zv2rQJS5cuxbffcsfybNiwATdu3MC2bduwa9cuODk5Yf78+SgsLIS0tDSePHmC1atXIywsDLNmzUJYWBjs7OzqTTCzs7MxduxYXkJtZmZW777Up7KyEkFBQWjVivs/qZ07d2Lo0KHYvHkzr5RaTk4Ohw4dgpKSEjp06AB/f3/4+Phg7dq1qKiowJ49exAYGIjBgwcDAH777TdcvXoVBw8ehI+PD++9/P39eT3jAKCiogIZGZkGS7ZZLBYcHR0RFhaGsWPHIiwsDO7u7ggKCkJCQgIsLCwQFRWFhQsX8j3vxx9/5CWoy5Ytw9ChQ1FZWQkFBQX8+OOPWLZsGdzd3XnttnbtWixZsgSrV6/mvYaHhwe+++47AMBPP/2EnTt34t69e3B1dRWIU0tLC9LS0lBVVf1kCTqbzcahQ4egqqoKKysrODs7IykpCRcvXoSUlBTatWuHDRs2ICwsDD169MCNGzfw+PFjFBQUQF5eHgD32Dpz5gxOnjyJH3744ZPH6oeuXLmCtLQ0hIWF8WL98ccf+T6bxnwnFRUVUVVV1eD+qqurQ05ODkpKSnzb7d69G0ZGRvj111/BYrFgaWmJ3NxcLF26FKtWrYJUI0oqG7MfdRYvXoyhQ4cC4F5k6NChA1JTU3ntxSQ1TSVIy0jjdVEZ3/LXr95CS0f4d79lay14zhuAxVMOgl376R8OFh0NYdpWH1v9zjRFyCKjqaIIGWkpFL0p51v+6s1b6KgZC32OoY46urRpiXc1NfD+7Rw0lBXxvwn9oK6kgDW/XxVF2P+Juhp3n18Xv+VbXlT8FlqaykKf8+p1GTb+GoLktHzIyspgoJMVtvpPwPzlx/Ew4X2yqawkh1OHZkJOVhq1bA627r2K6IfiOT5QU0ERMlJSeFnB/9m/LC+HjpHwdviQrpIynFqbYv7VC80Vosho/PM9eFXK3xZFpW+hXc/3oLDkLdYevYrErALIyUhjSI/22LvgG/yw5S/E/jPO9HZCFr53sUNsag6evyxGN8vW6GvTBtIfVFARIq4oGSUCnJ2dsWfP+8kzlJXf/8/C3t6+wefGxMQgNTUVqqqqfMsrKyuRlva+t6BDhw6Qln5/ddvAwACPHz8GACQmJkJGRobvvSwtLflKJMeNG4dt27bBzMwMrq6uGDJkCIYPHw4ZGeGHtLe3N6ZOnYojR47AxcUF48aN4yWeAKCkpMT3t4GBAa8ntrS0FLm5uQK9mg4ODrzyyg97/WRlZWFjY4MRI0Zgx44dAIT3+H1o3rx5mDlzJq5cuQIXFxeMHTsWnTo1XLLzsdatW/MSUQDo2bMn2Gw2kpKSeD/wbWxsoKSkxLdNWVkZnj17hpKSElRXV/Ptp6ysLLp164bExES+9/rUcVAfJycn7N+/HwC393jt2rXIyMhAeHg4SkpKUFFRIdDOH7aDgYEBAG4Zd+vWrRETE4P79+/z9YTW1taisrIS5eXlvH398DWUlZWhqqrK19P+b5mYmPAd6/r6+pCWluZLwPT19XnvFRMTg7KyMmhra/O9TkVFBe/78alj9UNJSUkwMjLiSw4/vijT2O/kv5WYmIiePXvylY07ODigrKwMz58/R+vW9fcI1mnMftSp73hobDJaVVWFqqoqvmVsdg2kpJrwf4cfVUaxWNxe8Y9JSbGwbP04HNlzHTnZrxr10q6jbJGR8gLJ8fVP+iPOPm4FFotV71wdUv+s+1/gJZT9M1ZuU3AENnkNw/o/r6Oq+ssozfx4/7j7LHynn+W8xrOc17y/nyTlQk9HFd+O7sqXjJZXvIPXgiAoKsrBrlNrzPZ0Ru6LEsTFC5Z5i42P9pl7yvh0GeE4yw4orarElYyU5omLCcKOiXo2zXrxGlkv3h8TjzLy0EJTFZMG2PGS0V/+DMPK710QvMYdHA7w/GUx/o56guG9OjTTDpBmRRMYEUmnrKxc78y5HyamwrDZbNjZ2Qkd4/nhJEgflx2yWCyw2dxegbr/SbMauKJnZGSEpKQkXL16FdeuXcOsWbPwyy+/8JLBj61ZswYTJ07EhQsXcOnSJaxevRrHjx/H6NGj643n4x8LH8fD4XB4yz7s9ZOTk4OTkxM6duyI2tpaPH78GFFRUXwzs35s6tSpGDRoEC5cuIArV65g/fr12Lx5M+bOnQspKSmBWD4cX1ufD2NrzLb1tfuH+1nnU8dBfep6kFNTUxEfH48+ffogLS0N4eHhKC4uhp2dnUDS9OFnUxdH3bHCZrPh5+eHMWPGCLyXgoKC0Neoe5261/gvhL1uQ+/FZrNhYGAgdExm3cWWTx2rHxL22Xyssd/Jf0tYDB8fS586hhuzH3UaOh4aY/369Xyl6QDQRt8R5i3qv1jUWKWvy1FbUwtNbf5eUA0tZbx+9VZge0VleVh0MESbdi0wexm3t5clxYKUlBQuRK/G/2YexsP7Gbzt5RVk0XeQNQ7vuf6fYxW112UVqKllQ1tViW+5looSXn3UW1qnsPQtCkrKeIkoAGTkF0FKigV9DVVkvyxuzpD/s5JS7j5/3Auqqa6E18XC91mYhOQ8DOxrxbeMwwFy8osBAKkZBTA20sb333QXy2T0dWUFaths6Crxt4OOohIKyz/dDuMsrXE6OQHVTXDOZlpx3fdAnf97oKmqhKLSxh8TjzPyMKTb+wtwxWUVWLT3b8jJSENdRQEvi99i3ujeyC38ssuaJdYXMtazqdCYUdKkbG1tkZKSAj09PZibm/M91NXVG/Ua7du3R01NDd+kLUlJSQL3m1RUVOT1PoaFheH27du83lVhLCwssHDhQly5cgVjxowRmDG4PmpqamjZsiVu3brFtzwqKoqvdLhu3GhYWBicnJzAYrHQp08fbNq0SWiP38eMjIwwY8YMBAcHY9GiRfjtt98AcBOGN2/e4O3b9z9mhd1CJTs7G7m5uby/b9++DSkpKVhYWPCWPXz4EBUV72fgu3PnDlRUVNCqVSuYm5tDTk6Obz+rq6sRHR3Nt5/CyMnJNWpW3boe5HXr1sHGxgZqamro27cvwsPDP9l7LIytrS2SkpIEjjVzc/NGlYfWp7H787lsbW2Rn58PGRkZgXh1dHR42zX2WLW0tER2djZevHjBW/bhhFN17/mp72Rj91fYdlZWVoiKiuJLNqOioqCqqgpDQ+64OF1dXb6xt6WlpcjIeJ9kNWY/moqvry9KSkr4HmZ6DX83G6umphYpiXno0pO/J7tL9zZIfJgtsH15WRWmj/0Vsybs4T0unIzGs4yXmDVhD54+5h8D6DiwA2TlpHH9wkOB1xJ3NbVsJD57gZ6W/KWI3S1b42FGrtDnxKXnQlddGYpy7y9AGOtpoJbNxovi+mcdFRc1NWwkp+XD3oZ/n+07GyP+aeN7ttua6eHV67IGt2GBO25QHFWz2Yh/+QK9jUz4lvduZYKYF8I/+zo9WhrBVEMTJxK/rJmj61NTy0Zi9gt0b89/TPRo3xoP0xtuiw+1M9JDYYngBa53NbV4WfwWMlJS6N+lLcIf/vfqF0KaGyWjpEm5ublBR0cHI0eOxM2bN3klmPPnz8fz542bXKFdu3ZwdXXFtGnTcPfuXcTExGDq1Kl8t9UIDAzEwYMHER8fj/T0dBw5cgSKioowNhYcc1FRUYE5c+YgLCwMWVlZiIyMxP379z+ZYH3Ix8cHGzZswIkTJ5CUlIRly5YhLi4O8+fP523j5OSEJ0+e4PHjx+jTpw9v2e+//w5bW1uoqanV+/oLFixASEgIMjIyEBsbi+vXr/Pi6969O5SUlPC///0Pqamp+OOPP4TOVqugoAB3d3c8fPgQN2/exLx58zB+/Hi+0sd3797By8sLCQkJvF63OXPmQEpKCsrKypg5cyZ8fHxw+fJlJCQkYNq0aSgvL4eXl1eD7WNiYoKMjAzExcWhsLBQoAyyTl0P8tGjR+Hk5ASAW3b57t07hIaG8pY11qpVq3D48GGsWbMGT548QWJiIk6cOIEVK1Z81usI25+IiAjk5OTwZv9tCi4uLujZsydGjRqFkJAQZGZmIioqCitWrEB0dPRnH6sDBgxAmzZt4O7ujkePHiEyMpI38U9dr2FjvpMmJiZ49OgRkpKSUFhYWG/Pu4mJCe7evYvMzEwUFhaCzWZj1qxZePbsGebOnYunT5/i7NmzWL16Nby9vXkXBPr164cjR47g5s2biI+Ph7u7O1+ZfmP2ozFycnJgaWmJe/fu1buNvLw81NTU+B5NWaIbfCQKrqNtMXBkFxiZ6uCHxa7QM1DHhZPc5HrKXBcsXsvtyedwOMhKK+B7lBS9xbt3NchKKxCYoGjQKDtE3XiKNyUVAu/7JThyPRaje3XEyB4dYKqvhcVj+sJASxUnbz4CAMwd4YC1kwbxtr94/ylK3lbC//uBMGuhBds2hlg42hFnbz/5Ykp0/zwbjWEDOmFI/44wbqWFOV7O0NNRw9nL3AsKP0zqg/8teH//4nHD7dC7uzlaGWjAxEgbP0zqA6de7RB84QFvG7ex3WFvYwwDfXW0NtTC+BH2GOTcAVfCxfc+owceRmNCe2uMs+yINppaWOnghJaqqvg9ntsOS3r0web+gwWeN759RzzIz0VykeB5WFZKClbaurDS1oWstDT0lVVhpa0LYzWN5t6d/+T3a7EY7dARI3t1gGkLLSwa1xctNFVxKoL7PZgzygH+Hu+/BxP7dYGTTRsY6WnAzEAbc0Y5wMW2LU6Evb8o1dGkBfp1NueOszY3xK/zRnNntr9S/0zsX4q35UBiCvcBAM/zuP/OfdHw875oHE7TPb4AVKZLmpSSkhIiIiKwdOlSjBkzBm/evIGhoSH69+/fYDL2sYCAAEydOhV9+/aFvr4+1q1bh5Ur399gSkNDAz///DO8vb1RW1sLa2tr/P333wJj8QBAWloar169wuTJk/HixQvo6OhgzJgxAqV6DZk3bx5KS0uxaNEiFBQUwMrKCufOneObXKZjx47Q0dGBsbExb1/79u2L2traT/b41dbWYvbs2Xj+/DnU1NTg6uqKrVu3AuBOqHP06FH4+Phg//79cHFxwZo1a/DDDz/wvYa5uTnGjBmDIUOGoKioCEOGDMHu3bv5tunfvz/atm0LR0dHVFVV4dtvv+W7vcbPP/8MNpuNSZMm4c2bN7C3t0dISAg0NTUbjH/s2LEIDg6Gs7MziouLERAQAA8PD6HbOjs7Izg4mJd41vUgnz9/Hr17927wfT42aNAgnD9/Hv7+/ti4cSNkZWVhaWmJqVOnftbrfMzf3x/Tp09HmzZtUFVVVe/4rs/FYrFw8eJFLF++HJ6ennj58iVatGgBR0dH3njTzzlWpaWlcebMGUydOhVdu3aFmZkZfvnlFwwfPpxXptyY7+S0adMQFhYGe3t7lJWV4caNG0IvDCxevBju7u6wsrJCRUUFMjIyYGJigosXL8LHxwc2NjbQ0tKCl5cX3wUBX19fpKenY9iwYVBXV+eNFf6c/WiM6upqJCUlobwRpX/NJeJKPNQ0FOE23QmaOqrISi3AyjlHUZDHvceolq4q9AwaVyXyIcPW2uhoawzfGUFNHbLIXIlNhoayAqYP7g4dNWWk5r3CnN1nkPfPvRV11ZRhoPW+TL/iXTVm/HoKy8Y54/clE1HythJXYpOx63wkU7vw2a7fSoKaqiLcJ/SCtpYyMrIKsdT/FF685JZPamuqQF/n/T7LyEhj1hQn6GqpoOpdDTKyX2GJ/0nciXn/fVFUkIX3jAHQ1eZuk51ThHVbL+D6raSP315snE9Ngoa8Iubb94SusjKSXxViyvlg3uy4ekrKMFTh/42gKieHwWYW8LslvCxdX1kFFye48/6e3qUrpnfpijs5z/Dt2RPNtzP/0ZWYZKirKGDaUO73IC33Feb9egZ5RdzvgY66Mlp88D2QlZHGwrGO0NVQQVV1DdJzX2Hur6cRGZ/J20ZOVhqzRvaCoY46yquqERmfgRUBl1FWIfzC8JfkSRLgvuD9RckNu7j/HuXKwXpfpqJqZl9IEtlUWJym+pVFCGHMmjVrcObMGaHlu3U8PDxQXFwscN9G8nWJjIxE7969kZqaWu/ER18CUe6Ha+dVzfr6X4r83g1fdJIUas/E41ZBTMseRDOxAoDWI/Esfxa16DV7Pr2RBJBqkdzs7zHYcO6nN2qkSzk7m+y1mgv1jBJCyBfs9OnTUFFRQdu2bZGamor58+fDwcHhi0tEv5b9IIQQQkjjUTJKCCFfsDdv3mDJkiV49uwZdHR04OLigs2bNzMd1mf7WvaDEEII+U++gpmjPweV6RJCCJFoVKbLRWW6XFSmy0VlulxUpstFZbpcIinTbTGryV7rUv7uT2/EMJpNlxBCCCGEEEKIyFGZLiGEEEIIIYSIAwkrWqVklBBCCCGEEELEAVuyklEq0yWEEEIIIYQQInLUM0oIIYQQQgghYoDDkazZdCkZJYQQQgghhBBxQGW6hBBCCCGEEEJI86KeUUIIIYQQQggRBzSbLiGEEEIIIYQQkWPTmFFCCCGEEEIIIaImYT2jNGaUEEIIIYQQQojIUc8oIYQQQgghhIgBjoSV6VLPKCGEEEIIIYSIAw6n6R6faffu3TA1NYWCggLs7Oxw8+bNRj0vMjISMjIy6Ny582e/JyWjhBBCCCGEECLBTpw4gQULFmD58uV48OAB+vTpg8GDByM7O7vB55WUlGDy5Mno37//v3pfSkYJIYQQQgghRBywOU33+AxbtmyBl5cXpk6divbt22Pbtm0wMjLCnj17Gnze9OnTMXHiRPTs2fNf7S4lo4QQQgghhBAiDjjsJntUVVWhtLSU71FVVSXwlu/evUNMTAwGDhzIt3zgwIGIioqqN9SAgACkpaVh9erV/3p3KRklhBBCCCGEkK/M+vXroa6uzvdYv369wHaFhYWora2Fvr4+33J9fX3k5+cLfe2UlBQsW7YMv//+O2Rk/v2cuDSbLiGEEEIIIYSIAc5nltc2xNfXF97e3nzL5OXl692exWLxx8LhCCwDgNraWkycOBF+fn6wsLD4TzFSMkoIIYQQQggh4oDTdLd2kZeXbzD5rKOjowNpaWmBXtCCggKB3lIAePPmDaKjo/HgwQPMmTMHAMBms8HhcCAjI4MrV66gX79+jYqRynQJIYQQQgghRELJycnBzs4OV69e5Vt+9epV9OrVS2B7NTU1PH78GHFxcbzHjBkz0K5dO8TFxaF79+6Nfm/qGSWEEEIIIYQQMdCUZbqfw9vbG5MmTYK9vT169uyJ/fv3Izs7GzNmzADALfnNycnB4cOHISUlhY4dO/I9X09PDwoKCgLLP4WSUUIIIYQQQggRB01Ypvs5JkyYgFevXsHf3x95eXno2LEjLl68CGNjYwBAXl7eJ+85+m+wOBwOM+k3IYQQQlBVVYX169fD19e3UWN7vlbUDlzUDlzUDlzUDlzUDl8vSkYJIYQQBpWWlkJdXR0lJSVQU1NjOhzGUDtwUTtwUTtwUTtwUTt8vWgCI0IIIYQQQgghIkfJKCGEEEIIIYQQkaNklBBCCCGEEEKIyFEySgghhDBIXl4eq1evlvhJOagduKgduKgduKgduKgdvl40gREhhBBCCCGEEJGjnlFCCCGEEEIIISJHySghhBBCCCGEEJGjZJQQQgghhBBCiMhRMkoIIYQQQgghROQoGSWEEEIIIYQQInKUjBJCCCGEEEIIETlKRgkhhBAx4OLiAjMzM6bDYNzZs2dx+PBhpsNgnLu7O/r168d0GIyLjo5GREQE02Ewjr4XXHQ8fH3oPqOEEEKIGNi1axcKCwuxevVqpkNhlKWlJVJSUlBbW8t0KIz63//+h7y8PAQEBDAdCqPat2+P5ORkiT8e6HvBRcfD14eSUUIIIYQQIpZyc3NRXV0NY2NjpkMhYoCOh68PJaOEEEIIIQx79+4dMjIy0KZNG8jIyDAdDiGEiASd7QghhBARqqysxM6dO3Hjxg0UFBSAzWbzrY+NjWUoMtG6fPkyVFRU0Lt3bwDcMuXffvsNVlZW2LVrFzQ1NRmOUDTKy8sxd+5cBAUFAQCSk5NhZmaGefPmoWXLlli2bBnDEYqGtLQ08vLyoKenx7f81atX0NPTk6iyzOLiYty7d0/o+WHy5MkMRSV6bDYbqampQtvB0dGRoahIU6NklBBCCBEhT09PXL16Fd988w26desGFovFdEiM8PHxwYYNGwAAjx8/xqJFi+Dt7Y3r16/D29tbYsZK+vr64uHDhwgLC4OrqytvuYuLC1avXi0xyWh9hXpVVVWQk5MTcTTM+fvvv+Hm5oa3b99CVVWV7/zAYrEkJhm9c+cOJk6ciKysLIFjg8ViSdTFia8dJaOEEEKICF24cAEXL16Eg4MD06EwKiMjA1ZWVgCAU6dOYdiwYfjpp58QGxuLIUOGMByd6Jw5cwYnTpxAjx49+BIPKysrpKWlMRiZaOzYsQMAN8E4cOAAVFRUeOtqa2sREREBS0tLpsITuUWLFsHT0xM//fQTlJSUmA6HMTNmzIC9vT0uXLgAAwMDib1oJwkoGSWEEEJEyNDQEKqqqkyHwTg5OTmUl5cDAK5du8br8dHS0kJpaSmToYnUy5cvBUpTAeDt27cS8QN869atALg9o3v37oW0tDRvnZycHExMTLB3716mwhO5nJwczJs3T6ITUQBISUnByZMnYW5uznQopJlRMkoIIYSI0ObNm7F06VLs3btXomeE7N27N7y9veHg4IB79+7hxIkTALhjJlu1asVwdKLTtWtXXLhwAXPnzgUAXgL622+/oWfPnkyGJhIZGRkAAGdnZwQHB0vMWOH6DBo0CNHR0RJ/z+Hu3bsjNTWVklEJQMkoIYQQIkL29vaorKyEmZkZlJSUICsry7e+qKiIochE69dff8WsWbNw8uRJ7NmzB4aGhgCAS5cu8Y2d/NqtX78erq6uSEhIQE1NDbZv344nT57g9u3bCA8PZzo8kblx4wYAmlV46NCh8PHxQUJCAqytrQXODyNGjGAosub36NEj3r/nzp2LRYsWIT8/X2g7dOrUSdThkWZCt3YhhBBCRMjFxQXZ2dnw8vKCvr6+QCmmu7s7Q5ERpjx+/BibNm1CTEwM2Gw2bG1tsXTpUlhbWzMdmshUVFRgzpw5Ej+rsJSUVL3rvvaJe6SkpMBiseqdzKpu3dfeDpKGklFCCCFEhJSUlHD79m3Y2NgwHQrj6NYNpM78+fMRGRmJbdu2wdXVFY8ePYKZmRnOnTuH1atX48GDB0yHSJpZVlZWo7eV5CEOXxvJq38ghBBCGGRpaYmKigqmw2Ac3bqBi+6vySXpswoTSjAlFSWjhBBCiAj9/PPPWLRoEX788UehY6HU1NQYiky06NYNXHR/TS5Jn1X4Q+Hh4di0aRMSExPBYrHQvn17+Pj4oE+fPkyHJjLnzp0TupzFYkFBQQHm5uYwNTUVcVSkOVAySgghhIhQ3eQ8/fv351suaWOhJP3WDXR/TX6SPqtwnaNHj2LKlCkYM2YM5s2bBw6Hg6ioKPTv3x+BgYGYOHEi0yGKxKhRo4SOH/1w3Gjv3r1x5swZiZ+B+UtHY0YJIYQQEfrUDKl9+/YVUSTM6tevH5YsWSJRM+d+qK5XJysrC61atRJ6f01/f390796dqRBFKioqCq6urnBzc0NgYCCmT5/ON6uwnZ0d0yGKRPv27fHDDz9g4cKFfMu3bNmC3377DYmJiQxFJlqhoaFYvnw5fvzxR3Tr1g0AcO/ePaxYsQIrV66Euro6pk+fju7du+PgwYMMR0v+C0pGCSGEECJyp0+fxooVK+Dj4yPRt26g+2u+R7MKA/Ly8njy5IlAxUBqaio6duyIyspKhiITrY4dO2L//v3o1asX3/LIyEj88MMPePLkCa5duwZPT09kZ2czFCVpClSmSwghhIjYzZs3sW/fPqSnp+Ovv/6CoaEhjhw5AlNTU/Tu3Zvp8ERi7NixAABPT0/eMkm8dUPd/TWB9+NHJW2MZB1ra2verV0klZGREUJDQwWS0dDQUBgZGTEUleilpaUJHT+vpqaG9PR0AEDbtm1RWFgo6tBIE6v/ZkaEEEIIaXKnTp3CoEGDoKioiNjYWFRVVQEA3rx5g59++onh6EQnIyND4JGens77ryQ5fPgwrK2toaioCEVFRXTq1AlHjhxhOiyRS0tLw4oVKzBx4kQUFBQAAC5fvownT54wHJnoLFq0CPPmzcPMmTNx5MgRHD16FDNmzMD8+fOxePFipsMTGTs7O/j4+ODly5e8ZS9fvsSSJUvQtWtXANxx561atWIqRNJEqEyXEEIIEaEuXbpg4cKFmDx5MlRVVfHw4UOYmZkhLi4Orq6uyM/PZzpEIkJbtmzBypUrMWfOHDg4OIDD4SAyMhK7du3CunXrBMYOfq3Cw8MxePBgODg4ICIiAomJiTAzM8PGjRtx7949nDx5kukQReb06dPYvHkzb3xo3Wy6I0eOZDgy0UlKSsLIkSORkZEBIyMjsFgsZGdnw8zMDGfPnoWFhQXOnDmDN2/eYNKkSUyHS/4DSkYJIYQQEVJSUkJCQgJMTEz4ktH09HRYWVlJzJgwgNsTtm3bNr5bWMyfPx9t2rRhOjSRMTU1hZ+fHyZPnsy3PCgoCGvWrEFGRgZDkYlWz549MW7cOHh7e/N9L+7fv49Ro0YhJyeH6RCJiHE4HISEhCA5ORkcDgeWlpYYMGAApKSosPNrQmNGCSGEEBEyMDBAamoqTExM+JbfunULZmZmzATFgJCQEIwYMQKdO3fm9QhGRUWhQ4cO+PvvvzFgwACmQxSJvLw8gUlaAKBXr17Iy8tjICJmPH78GH/88YfAcl1dXbx69YqBiAjTWCwWXF1dJXbGbUlBySghhBAiQtOnT8f8+fNx6NAhsFgs5Obm4vbt21i8eDFWrVrFdHgis2zZMixcuBA///yzwPKlS5dKTDJqbm6OP//8E//73//4lp84cQJt27ZlKCrR09DQQF5eHu+WN3UePHgAQ0NDhqISDS0tLSQnJ0NHRweampoNTmBVVFQkwsiY4+/v3+B6STpXfu0oGSWEEEJEaMmSJSgpKYGzszMqKyvh6OgIeXl5LF68GHPmzGE6PJFJTEzEn3/+KbDc09MT27ZtE31ADPHz88OECRMQEREBBwcHsFgs3Lp1C6GhoULb52s1ceJELF26FH/99RdYLBbYbDYiIyOxePFigRLmr83WrVuhqqrK+7ekzqb8odOnT/P9XV1djYyMDMjIyKBNmzaUjH5FaMwoIYQQwoDy8nIkJCSAzWbDysoKKioqTIckUkZGRtiyZQvGjRvHt/zPP//E4sWLJeregTExMdi6dSsSExPB4XBgZWWFRYsWoUuXLkyHJjLV1dXw8PDA8ePHweFwICMjg9raWkycOBGBgYGQlpZmOkTCsNLSUnh4eGD06NE0adFXhJJRQgghhIicv78/tm7dimXLlqFXr168HsENGzZg0aJFWLFiBdMhEgakpaXhwYMHYLPZ6NKli0SVKgOAtLQ08vLyoKenx7f81atX0NPTk5j779YnPj4ew4YNQ2ZmJtOhkCZCZbqEEEJIMxszZkyjtw0ODm7GSMTHypUroaqqis2bN8PX1xcA0LJlS6xZswbz5s1jODrRYrPZSE1NRUFBAdhsNt86R0dHhqJiRps2bSRqNuWP1ddHVFVVBTk5ORFHI36Ki4tRUlLCdBikCVEySgghhDQzdXV1pkMQOywWCwsXLsTChQvx5s0bAOCNm5Mkd+7cwcSJE5GVlSWQiLBYLInpCeNwODh58iRu3LghNCn/2i/S7NixAwD3Mz9w4ABf2X5tbS0iIiJgaWnJVHgiV9cedTgcDvLy8nDkyBGaXfcrQ2W6hBBCCCEM6dy5MywsLODn5wcDAwOByWsk5ULGvHnzsH//fjg7O0NfX1+gHQICAhiKTDTqZhHOyspCq1at+MbIysnJwcTEBP7+/ujevTtTIYrUx7MqS0lJQVdXF/369YOvr69EXrj6WlEySgghhDCgoKAASUlJYLFYsLCwEBgj9jWytbVFaGgoNDU10aVLlwZnDY2NjRVhZMxRVlbGw4cPYW5uznQojNLS0sLRo0cxZMgQpkNhlLOzM4KDg6Gpqcl0KISIBJXpEkIIISJUWlqK2bNn4/jx47wSTGlpaUyYMAG7du36qnvCRo4cCXl5ed6/6RYWQPfu3ZGamirxyai6ujrMzMyYDoNxN27cYDoEsfP8+XOwWKyv/n6zkop6RgkhhBARGj9+POLi4rBz50707NkTLBYLUVFRmD9/Pjp16iRR95Yk3PsprlixAj4+PrC2toasrCzf+k6dOjEUmWgFBQXh8uXLOHToEBQVFZkOh1HPnz/HuXPnkJ2djXfv3vGt27JlC0NRiRabzca6deuwefNmlJWVAeCOKV+0aBGWL18OKSkphiMkTYWSUUIIIUSElJWVERISgt69e/Mtv3nzJlxdXfH27VuGIhMtMzMz3L9/H9ra2nzLi4uLYWtri/T0dIYiEy1hP6pZLBY4HI5ETWBUXl6OMWPGIDIyEiYmJgJJuaSUbYeGhmLEiBEwNTVFUlISOnbsiMzMTHA4HNja2uL69etMhygSvr6+OHjwIPz8/ODg4AAOh4PIyEisWbMG06ZNw48//sh0iKSJUJkuIYQQIkLa2tpCS3HV1dUlapxYZmam0ESrqqoKz58/ZyAiZmRkZDAdgljw8PBATEwMvv/+e6ETGEkKX19fLFq0CP7+/lBVVcWpU6egp6cHNzc3iZpFNigoCAcOHMCIESN4y2xsbGBoaIhZs2ZRMvoVoWSUEEIIEaEVK1bA29sbhw8fhoGBAQAgPz8fPj4+WLlyJcPRNb9z587x/h0SEsKXmNfW1iI0NFRgJs2vmbGxMdMhiIULFy4IrRiQNImJiTh27BgAQEZGBhUVFVBRUYG/vz9GjhyJmTNnMhyhaBQVFQm9lY2lpSWKiooYiIg0F0pGCSGEkGb28cyxKSkpMDY2RuvWrQEA2dnZkJeXx8uXLzF9+nSmwhSJUaNGAeCWorq7u/Otk5WVhYmJCTZv3sxAZIRJRkZGUFNTYzoMxikrK6OqqgoA0LJlS6SlpaFDhw4AgMLCQiZDEykbGxv8+uuvAvcb/fXXX2FjY8NQVKQ5UDJKCCGENLO6BIxwJyYBuPcRvH//PnR0dBiOiIiDzZs3Y8mSJdi7dy9MTEyYDocxPXr0QGRkJKysrDB06FAsWrQIjx8/RnBwMHr06MF0eCKzceNGDB06FNeuXeOb6O3Zs2e4ePEi0+GRJkQTGBFCCCGEEEZpamqivLwcNTU1UFJSEpjASFJKM9PT01FWVoZOnTqhvLwcixcvxq1bt2Bubo6tW7dKVFl3bm4udu3ahadPn4LD4cDKygqzZs1Cy5YtmQ6NNCFKRgkhhBAGxMTEIDExESwWC1ZWVujSpQvTIYnc27dvER4eLvQWFvPmzWMoKsKEoKCgBtd/XNJNvm7Z2dkwMjISOpFVdnY2b4gD+fJRMkoIIYSIUEFBAb799luEhYVBQ0MDHA4HJSUlcHZ2xvHjx6Grq8t0iCLx4MEDDBkyBOXl5Xj79i20tLRQWFgIJSUl6OnpScytXQipT1lZGa+svY6kjKuVlpZGXl4e9PT0+Ja/evUKenp6EnPLI0lAd4wlhBBCRGju3LkoLS3FkydPUFRUhNevXyM+Ph6lpaUS1Ru4cOFCDB8+HEVFRVBUVMSdO3eQlZUFOzs7bNq0ienwREZKSgrS0tL1PohkycjIwNChQ6GsrMy73ZOmpiY0NDQk6tZPdffZ/VhZWRkUFBQYiIg0F5rAiBBCCBGhy5cv49q1a2jfvj1vmZWVFXbt2oWBAwcyGJloxcXFYd++fbykq6qqCmZmZti4cSPc3d0xZswYpkMUidOnT/P9XV1djQcPHiAoKAh+fn4MRUWY4ubmBgA4dOiQRN5v1dvbGwB3tu2VK1dCSUmJt662thZ3795F586dGYqONAdKRgkhhBARYrPZApOzANzbmnxckvc1k5WV5f3Q1tfXR3Z2Ntq3bw91dXVkZ2czHJ3ojBw5UmDZN998gw4dOuDEiRPw8vJiICrClEePHiEmJgbt2rVjOhRGPHjwAAC3Z/Tx48eQk5PjrZOTk4ONjQ0WL17MVHikGVAySgghhIhQv379MH/+fBw7dow3K2ROTg4WLlyI/v37Mxyd6HTp0gXR0dGwsLCAs7MzVq1ahcLCQhw5cgTW1tZMh8e47t27Y9q0aUyHQUSsa9euePbsmcQmozdu3AAATJkyBdu3b5eYMbKSjCYwIoQQQkTo2bNnGDlyJOLj43mzRWZnZ8Pa2hpnz55Fq1atmA5RJKKjo/HmzRs4Ozvj5cuXcHd3593CIiAgQKJvbF9RUQFfX19cunQJSUlJTIdDRCgtLQ0zZszA999/j44dOwpUUXTq1ImhyAhpHpSMEkIIIQy4evUq3/3zXFxcmA6JMEBTU5NvXCCHw8GbN2+gpKSEo0ePYsSIEQxGJzrOzs4Njo+8fv26CKNhzp07dzBx4kRkZmbylrFYLN6EPjSLLPnaUJkuIYQQIiI1NTVQUFBAXFwcBgwYgAEDBjAdEmHYtm3b+P6WkpKCrq4uunfvLlGzp348KU11dTXi4uIQHx8vUfcY9fT0RJcuXXDs2DGJnMCISB7qGSWEEEJEqE2bNggODpboMlQAMDU1bfCHNt1nlADAmjVrUFZWJjG3+1FWVsbDhw9hbm7OdCiEiAT1jBJCCCEitGLFCvj6+uLo0aPQ0tJiOhzGLFiwgO/vuluaXL58GT4+PswExZDi4mIcPHgQiYmJYLFYsLKygqenJ9TV1ZkOjXHff/89unXrJjHJaL9+/SgZBRAREYFevXpBRoY/VampqUFUVBQcHR0Ziow0NeoZJYQQQkSoS5cuSE1NRXV1NYyNjaGsrMy3PjY2lqHIxMOuXbsQHR2NgIAApkMRiejoaAwaNAiKioro1q0bOBwOoqOjUVFRgStXrsDW1pbpEBl15MgRLF26FLm5uUyHIhL79+/HunXr4OnpCWtra4EJjCRlDLG0tDTy8vKgp6fHt/zVq1fQ09OjsbNfEUpGCSGEEBHy8/NrcP3q1atFFIl4Sk9PR+fOnVFaWsp0KCLRp08fmJub47fffuP1AtXU1GDq1KlIT09HREQEwxGKxpgxY/j+5nA4yMvLQ3R0NFauXCkx3wspKal610nSBEZSUlJ48eIFdHV1+ZYnJyfD3t5eYs4PkoDKdAkhhBARkpQf1f/WyZMnJap8OTo6mi8RBQAZGRksWbIE9vb2DEYmWh+XJEtJSaFdu3bw9/fHwIEDGYpK9NhsNtMhMKruogSLxYKHhwfk5eV562pra/Ho0SP06tWLqfBIM6BklBBCCGFAdHQ0b4xg+/btYWdnx3RIItWlSxeBW5rk5+fj5cuX2L17N4ORiZaamhqys7NhaWnJt/zZs2dQVVVlKCrRk5Sy7IZ8ONt2x44dmQ6HEXUXJTgcDlRVVaGoqMhbJycnhx49emDatGlMhUeaASWjhBBCiAg9f/4c3333HSIjI6GhoQGAO4FNr169cOzYMRgZGTEboIiMHDmSLxmtu6WJk5OTQGL2NZswYQK8vLywadMm9OrVCywWC7du3YKPjw++++47psMTuZiYGL6JnLp06cJ0SCIjIyMDY2NjiSnFFabuooSJiQkWL14sMKaefH1ozCghhBAiQgMHDkRpaSmCgoLQrl07AEBSUhI8PT2hrKyMK1euMBwhEaV3797Bx8cHe/fuRU1NDQBAVlYWM2fOxM8//8xXpvg1KygowLfffouwsDBoaGiAw+GgpKQEzs7OOH78uMDYwa9VQEAA/vrrL4mfbRvg9hSHhYUhLS0NEydOhKqqKnJzc6GmpgYVFRWmwyNNhJJRQgghRIQUFRURFRUl0OMTGxsLBwcHVFRUMBSZaNFsmfzKy8uRlpYGDocDc3NzKCkpMR2SSE2YMAFpaWk4cuQI2rdvDwBISEiAu7s7zM3NcezYMYYjFA2abZsrKysLrq6uyM7ORlVVFZKTk2FmZoYFCxagsrISe/fuZTpE0kSoTJcQQggRodatW6O6ulpgeU1NDQwNDRmIiBn1XQuvqqqCnJyciKNhTmBgICZMmAAlJSVYW1szHQ5jLl++jGvXrvESUQCwsrLCrl27JGoCo1GjRjEdgliYP38+7O3t8fDhQ2hra/OWjx49GlOnTmUwMtLUKBklhBBCRGjjxo2YO3cudu3aBTs7O7BYLERHR2P+/PnYtGkT0+E1ux07dgDgzpZ54MABvnK72tpaRERESNSYUV9fX8ybNw/jxo2Dl5eXxM4UymazBe6pCXBLliVphlmabZvr1q1biIyMFLgwZWxsjJycHIaiIs2BynQJIYQQEdLU1ER5eTlqamr47ispIyMjUJJXVFTERIjNytTUFAC3DK9Vq1aQlpbmrZOTk4OJiQn8/f3RvXt3pkIUqdraWly4cAGBgYG4cOECTE1NMWXKFLi7u6NFixZMhycyI0eORHFxMY4dO4aWLVsCAHJycuDm5gZNTU2cPn2a4QhFS5IncgIALS0t3Lp1C1ZWVlBVVcXDhw9hZmaGW7duYezYsXjx4gXTIZImQskoIYQQIkJBQUGN3tbd3b0ZI2GWs7MzgoODoampyXQoYqOgoABHjx5FYGAgnj59CldXV3h5eWH48OGQkpJiOrxm9ezZM4wcORLx8fEwMjICi8VCdnY2rK2tcfbsWbRq1YrpEEWCJnLimjBhAtTV1bF//36oqqri0aNH0NXVxciRI9G6dWu6FdBXhJJRQgghhBAxcffuXRw6dAhBQUEwMDBAcXExNDQ0EBAQACcnJ6bDa3ZXr17F06dPweFwYGVlBRcXF6ZDEimayIkrNzcXzs7OkJaWRkpKCuzt7ZGSkgIdHR1EREQITHxGvlyUjBJCCCFEJLy9vbF27VooKyvD29u7wW23bNkioqiY9+LFCxw5cgQBAQFIT0/HqFGj4OXlBRcXF1RUVGDFihU4efIksrKymA612WRkZPBKuCWZuro6rl27hq5du/Itv3fvHgYOHIji4mJmAmNARUUFjh07htjYWLDZbNja2sLNzQ2KiopMh0aaEE1gRAghhBCRePDgAW8m4QcPHtS7HYvFElVIjBs+fDhCQkJgYWGBadOmYfLkyXz3l1RUVMSiRYuwdetWBqNsfubm5nB0dISXlxe++eYbKCgoMB0SI2gip/cUFRXh6ekJT09PpkMhzYh6RgkhhBBCGOLl5YWpU6eiZ8+e9W7D4XCQnZ0NY2NjEUYmWvHx8Th06BB+//13VFVVYcKECfD09JSYiazq0EROXOfOnRO6nMViQUFBAebm5tST/pWgZJQQQgghjCstLcX169dhaWkpUbd2+dCxY8cwYsQIgVmVJUlNTQ3+/vtvBAYG4tKl86F94QAAjDNJREFUS2jbti28vLwwadIkiZi8hyZy4pKSkgKLxRK4H3HdMhaLhd69e+PMmTM0CdoXjpJRQgghhEF1SVi7du14E5ZIgvHjx8PR0RFz5sxBRUUFbGxskJmZCQ6Hg+PHj2Ps2LFMhyhyampqiIuLg5mZGdOhMK6qqgq7d++Gr68v3r17B1lZWUyYMAEbNmyAgYEB0+E1O0mfyCk0NBTLly/Hjz/+iG7dugHgjptdsWIFVq5cCXV1dUyfPh3du3fHwYMHGY6W/BeUjBJCCCEiREkYV4sWLRASEgIbGxv88ccfWL16NR4+fIigoCDs37+/wTGlX6sP76coqaKjo3Ho0CEcP34cysrKcHd3h5eXF3Jzc7Fq1Sq8efMG9+7dYzrMJqWlpYXk5GTo6OjA09MT27dvh6qqKtNhMapjx47Yv38/evXqxbc8MjISP/zwA548eYJr167B09MT2dnZDEVJmsLXfdMqQgghRMxERESgT58+AIDTp0+Dw+GguLgYO3bswLp16xiOTnRKSkp4E/VcvnwZY8eOhZKSEoYOHYqUlBSGoyOitmXLFlhbW6NXr17Izc3F4cOHkZWVhXXr1sHU1BQODg7Yt28fYmNjmQ61yb179w6lpaUAuPchrqysZDgi5qWlpUFNTU1guZqaGtLT0wEAbdu2RWFhoahDI02MZtMlhBBCRKihJMzHx4fh6ETHyMgIt2/fhpaWFi5fvozjx48DAF6/fi2xM6leunQJhoaGTIfBiD179sDT0xNTpkxBixYthG7TunXrr7Iks2fPnhg1ahTs7OzA4XAwb968em9fcujQIRFHxww7Ozv4+Pjg8OHDvLHCL1++xJIlS3i3vUlJSZGYMbRfM+oZJYQQQkSoLgl7+/YtLl++jIEDBwKQvCRswYIFcHNzQ6tWrdCyZUs4OTkB4PYcW1tbMxucCB09epT37969e0NeXp73tyRdnEhJSYGvr2+9iSgAyMnJwd3dXYRRicbRo0cxZMgQlJWVgcVioaSkBK9fvxb6kBQHDx5ERkYGWrVqBXNzc7Rt2xatWrVCZmYmDhw4AAAoKyvDypUrGY6U/Fc0ZpQQQggRod27d2P+/PlQUVGBsbExYmNjISUlhZ07dyI4OBg3btxgOkSRiYmJQXZ2NgYMGAAVFRUAwIULF6ChoQEHBweGoxMNDQ0NHD16FMOGDeNbvnDhQhw/fhx5eXkMRUaYYGpqiujoaGhrazMdCuM4HA5CQkKQnJwMDocDS0tLDBgwAFJS1Jf2NaFklBBCCBGx6OhoPHv2TKKTMMJ1+fJlfPvttzh37hwcHR0BAHPnzkVwcDBCQ0Ml9jY3hBDJQMkoIYQQQgiDjh8/jlmzZuHKlSs4dOgQzp49ixs3bsDCwoLp0AgDQkNDERoaioKCArDZbL51kjJmlEgOmsCIEEIIaWbe3t5Yu3YtlJWV4e3t3eC2W7ZsEVFURFx8++23eP36NXr37g1dXV2Eh4fD3Nyc6bAIA/z8/ODv7w97e3sYGBiAxWIxHRIhzYqSUUIIIaSZPXjwANXV1bx/14d+eEqG+i5I6OnpoUuXLti9ezdvmSRdnKipqUFYWBjS0tIwceJEqKqqIjc3F2pqarxy9q/d3r17ERgYiEmTJjEdCiEiQWW6hBBCCBG57OxsGBkZCSTgHA4Hz549Q+vWrRmKrPk5Ozs3ajsWi4Xr1683czTiISsrC66ursjOzkZVVRWSk5NhZmaGBQsWoLKyEnv37mU6RJHQ1tbGvXv30KZNG6ZDIUQkKBklhBBCiMhJS0sjLy8Penp6fMtfvXoFPT091NbWMhQZYcKoUaOgqqqKgwcPQltbGw8fPoSZmRnCw8MxdepUpKSkMB2iSCxduhQqKioSf8uS2NhYyMrK8m7zdPbsWQQEBMDKygpr1qyBnJwcwxGSpkJluoQQQogIvX37Fj///HO9E5Skp6czFJlocTgcoWXJZWVlEnW/VcJ169YtREZGCiQZxsbGyMnJYSgq0ausrMT+/ftx7do1dOrUCbKysnzrJaVse/r06Vi2bBmsra2Rnp6Ob7/9FqNHj8Zff/2F8vJybNu2jekQSROhZJQQQggRoalTpyI8PByTJk2SyAlK6sZLslgsrFy5EkpKSrx1tbW1uHv3Ljp37sxQdKJHFye42Gy20N7w58+fQ1VVlYGImPHo0SPe8R8fH8+3TpLOFcnJybx2+Ouvv+Do6Ig//vgDkZGR+PbbbykZ/YpQMkoIIYSI0KVLl3DhwgWJvZ9o3QROHA4Hjx8/5usJk5OTg42NDRYvXsxUeCIn6Rcn6gwYMADbtm3D/v37AXATr7KyMqxevRpDhgxhODrRuXHjBtMhiAUOh8O7MHPt2jUMGzYMAGBkZITCwkImQyNNjMaMEkIIISJkamqKixcvon379kyHwqgpU6Zg+/btUFNTYzoURmloaEj0xYk6ubm5cHZ2hrS0NFJSUmBvb4+UlBTo6OggIiJCYGwx+br169cPRkZGcHFxgZeXFxISEmBubo7w8HC4u7sjMzOT6RBJE6FklBBCCBGho0eP4uzZswgKCuIrUSWSiS5OvFdRUYFjx44hNjYWbDYbtra2cHNzg6KiItOhNasxY8YgMDAQampqGDNmTIPbBgcHiygqZj169Ahubm7Izs6Gt7c3Vq9eDQCYO3cuXr16hT/++IPhCElToWSUEEIIaWZdunThK79MTU0Fh8OBiYmJwAQlsbGxog5PZD71Q/tDkvKjmy5OkClTpmDHjh1QVVXFlClTGtw2ICBARFGJp8rKSkhLSwucN8mXi8aMEkIIIc1s1KhRTIcgFtTV1ZkOQexs3rwZaWlp0NfXl7iLEx9LTk5GWFiY0ImcVq1axVBUze/DBFPSk80PFRcX4+TJk0hLS4OPjw+0tLSQkJAAfX19GBoaMh0eaSLUM0oIIYQQwhA/P78G19eVJ37tfvvtN8ycORM6Ojpo0aIFXyUBi8WSqKSccMt0+/fvDw0NDWRmZiIpKQlmZmZYuXIlsrKycPjwYaZDJE2EklFCCCFEhMzMzHD//n1oa2vzLS8uLoatra3E3MqDkA8ZGxtj1qxZWLp0KdOhEDHg4uICW1tbbNy4Eaqqqnj48CHMzMwQFRWFiRMn0gRGXxEq0yWEEEJEKDMzU+j9FKuqqvD8+XMGImKGqalpg7cxoaRcsrx+/Rrjxo1jOgwiJu7fv499+/YJLDc0NER+fj4DEZHmQskoIYQQIgLnzp3j/TskJIRv/GRtbS1CQ0NhamrKRGiMWLBgAd/f1dXVePDgAS5fvgwfHx9mgmJAbW0ttm7dij///BPZ2dl49+4d3/qioiKGIhOtcePG4cqVK5gxYwbToRAxoKCggNLSUoHlSUlJ0NXVZSAi0lwoGSWEEEJEoG4SIxaLBXd3d751srKyMDExwebNmxmIjBnz588XunzXrl2Ijo4WcTTM8fPzw4EDB+Dt7Y2VK1di+fLlyMzMxJkzZ77qSXs+Zm5ujpUrV+LOnTuwtrYWmMhp3rx5DEXGnMrKSigoKDAdBiNGjhwJf39//PnnnwC4583s7GwsW7YMY8eOZTg60pRozCghhBAiQqamprh//z50dHSYDkUspaeno3PnzkJ7Rb5Gbdq0wY4dOzB06FCoqqoiLi6Ot+zOnTsScz/FhqoCWCyWxJRts9ls/Pjjj9i7dy9evHiB5ORk3sQ9JiYm8PLyYjpEkSgtLcWQIUPw5MkTvHnzBi1btkR+fj569uyJixcvQllZmekQSROhnlFCCCFEhDIyMpgOQaydPHkSWlpaTIchMvn5+bC2tgYAqKiooKSkBAAwbNgwrFy5ksnQRIq+F1zr1q1DUFAQNm7ciGnTpvGWW1tbY+vWrRKTjKqpqeHWrVu4fv06YmNjwWazYWtrCxcXF6ZDI02MklFCCCGEiFyXLl34JjDicDjIz8/Hy5cvsXv3bgYjE61WrVohLy8PrVu3hrm5Oa5cuQJbW1vcv38f8vLyTIfHiLqivYYmuPpaHT58GPv370f//v35xs926tQJT58+ZTAyZvTr1w/9+vVjOgzSjCgZJYQQQojI1Y2hrSMlJQVdXV04OTnB0tKSmaAYMHr0aISGhqJ79+6YP38+vvvuOxw8eBDZ2dlYuHAh0+GJ1OHDh/HLL78gJSUFAGBhYQEfHx9MmjSJ4chEJycnB+bm5gLL2Ww2qqurGYiIOffu3UNYWBgKCgrAZrP51m3ZsoWhqEhTo2SUEEIIISK3evVqpkMQCz///DPv39988w1atWqFqKgomJubY8SIEQxGJlpbtmzBypUrMWfOHDg4OIDD4SAyMhIzZsxAYWGhxCTmHTp0wM2bN2FsbMy3/K+//kKXLl0Yikr0fvrpJ6xYsQLt2rWDvr4+Xy+5JPaYf81oAiNCCCGEiNzFixchLS2NQYMG8S0PCQkBm83G4MGDGYqMMMHU1BR+fn6YPHky3/KgoCCsWbNGYsaU/v3335g0aRJ8fX3h7+8PPz8/JCUl4fDhwzh//jwGDBjAdIgioa+vjw0bNsDDw4PpUEgzo55RQgghRMTS0tIQEBCAtLQ0bN++HXp6erh8+TKMjIzQoUMHpsMTiWXLlvH1CtbhcDhYtmyZRCWjycnJ9ZYjSsrtXfLy8tCrVy+B5b169UJeXh4DETFj+PDhOHHiBH766SewWCysWrUKtra2+PvvvyUmEQW4ZfsODg5Mh0FEgHpGCSGEEBEKDw/H4MGD4eDggIiICCQmJsLMzAwbN27EvXv3cPLkSaZDFAlFRUUkJibCxMSEb3lmZiY6dOiAt2/fMhOYiP3222+YOXMmdHR08P/27jys5vz///jjlJK0qyxJiyylRdmGtNllhpgxTIYQxhiyZJ+xM4OxxDAYY1rGzthmYVCKIlspEaVSlrLFNFTUOef3h8v5zvlk5jO/6+q8X306j9t1zXXxPueP+8d3zNfT6/V+vRo1alRlO2JKSorAOum4uroiODgYc+fOVXu+dOlS7NmzB1evXhVURiKsXLkS9+/fR0REhOgU0jCujBIREUlo9uzZWLp0KaZNmwZjY2PV84CAAKxbt05gmbRMTU2Rm5tbZRi9deuWVt0huHTpUixbtgyzZs0SnSLUokWLMGTIEJw+fRre3t6QyWRITExEbGws9u7dKzqPJDZ9+nT069cPzZs3h4uLC/T09NQ+P3DggKAyqm4cRomIiCR09epV7Ny5s8pzKysrPHnyRECRGP3798eUKVNw8OBBNG/eHMDrQTQ8PFyrDu55+vQpBg8eLDpDuPfffx/nz5/H2rVrcejQISiVSri4uODChQu1/uAec3Pzf30oT3FxsYZraoZJkybh1KlTCAgIQIMGDXhoUS3GYZSIiEhCZmZmKCwshIODg9rz1NRU2NjYCKqS3tdff40+ffqgdevWaNq0KQDg7t278PHxwapVqwTXSWfw4ME4fvy42p2S2qpdu3bYvn276AzJcStqVTExMfjpp5/Qr18/0SmkYRxGiYiIJBQcHIxZs2Zh3759kMlkUCgUSEpKwvTp06ucJFqbmZqa4uzZszhx4gTS0tJQr149uLu7w9fXV3Saxq1fv171YycnJ8ybNw/Jyclwc3Orsh0xLCxM6jwh/P39MXr0aAwePBj16tUTnSOpkJAQ0Qk1joWFhWrHBNVuPMCIiIhIQhUVFRg5ciR2794NpVKJOnXqQC6XIzg4GFFRUdDV1RWdSBr2n6vif0cmkyE3N1fDNTVDeHg4duzYgbKyMnz44YcIDQ3FO++8IzpLCLlcjoMHDyIzMxMymQzOzs4YMGAA6tTRnjWkyMhIHDt2DJGRkTA0NBSdQxrEYZSIiEiAnJwcpKamQqFQwNPTEy1atBCdpHHr16/HuHHjYGBgoLY6+DbasiJI/0cul+OXX35BZGQkfvvtNzg5OWH06NEYPnw4GjZsKDpPEhkZGRgwYACKiorQqlUrAK+v/rGyssKRI0fg5uYmuFAanp6eyMnJgVKphL29fZUdA9pyyrQ24DBKREQkoYSEBPj5+YnOEMLBwQGXLl1CgwYN/nF1UJtWBBcvXozp06dXWf0pKyvD119/rTX3jP6nR48eYcuWLVi2bBnkcjkCAwMRFhaGbt26iU7TqHfeeQfW1taIjo6Gubk5gNeHXI0cORIPHz7EuXPnBBdKY9GiRf/4+YIFCyQqIU3jMEpERCQhfX19NGrUCMHBwfj444/h6uoqOkkyf/zxB0xNTUVn1Ci6urooLCyEtbW12vMnT57A2toacrlcUJk4Fy5cQGRkJHbt2gVTU1OMHDkShYWF2LFjBz799NNafcBVvXr1cOnSJbRp00bteUZGBjp06ICysjJBZUSaoSM6gIiISJvcv38fM2fOxJkzZ+Du7g53d3esXLkSd+/eFZ2mcRYWFnj48CEAoFu3bnj27JnYoBpAqVS+9dqKtLQ0WFhYCCgS4+HDh1i9ejVcXV3h4+ODR48eYffu3bh9+zYWLVqE7777DocPH8bmzZtFp2pUq1at8ODBgyrPHz58CCcnJwFFRJrFlVEiIiJB8vLysHPnTuzatQs3btyAr68v4uLiRGdpjKmpKZKTk+Hs7AwdHR08ePAAVlZWorOEeHO35B9//AETExO1gVQul+P58+cYP348Nm7cKLBSOvr6+mjevDlGjx6NkSNHvvXfi5KSEgwYMACnTp0SUKg5JSUlqh8nJiZi5syZWLhwoeoAp+TkZCxevBjLly9HYGCgqEyNs7CwQFZWFiwtLf/r3avact+qNuAwSkREJJBcLsfRo0cxb948pKen1+ptme+//z6SkpLg7OyMhIQEdOnSBfr6+m/9bm0eygEgOjoaSqUSo0ePRkREhNr2ZX19fdjb26Nz584CC6WjVCpx5swZtG/fXitPTtXR0VEbvN780fzNs7/+vDb/9yE6OhpDhw5F3bp1ERUV9Y/DKK/DqT04jBIREQmQlJSEHTt2YP/+/SgvL0f//v0xbNgw9O3bV3SaxpSVlSE6Oho5OTlYvXo1xo4d+7fDx9q1ayWuEyMhIQHe3t5adW3Hf1IoFDAwMMC1a9e04lTp/5SQkPCvv6uth59R7cVhlIiISEJz587Frl27cP/+ffTo0QPDhg1DUFCQVqwIlZSUwMTEBAAQEBCAgwcPwszMTGwU1Qht2rTBtm3btPZuUVLHg720h/b+NRwREZEA8fHxmD59OoYMGQJLS0vROZIyNzdX/QHzn7bgkfZZuXIlZsyYgU2bNmnVCdN/p7S0FAUFBXj16pXac3d3d0FF0vq7tbKXL1/+7dZ++t/EYZSIiEhCZ8+eFZ0gjJGRkWplIyEhARUVFaKTqIb4+OOPUVpaCg8PD+jr66NevXpqn2vLgTWPHj3CqFGjcPTo0bd+XttXBNevXw/g9fux33//PYyMjFSfyeVynD59Gq1btxaVRxrAYZSIiEjDjhw58q+/279/fw2WiNWjRw8EBATA2dkZSqUSAwcO1NoDjEhdRESE6IQaYcqUKXj69CmSk5NVW9kfPHiApUuXYvXq1aLzNO7Nu+JKpRKbN2+Grq6u6rM3B3vV9ut9tA3fGSUiItIwHZ1/d613bT8tkwcYEf2zxo0b4/Dhw+jYsSNMTExw6dIltGzZEkeOHMHKlSuRmJgoOlESAQEBOHDgAMzNzUWnkIZxGCUiIiLJ8QCj/3Px4kXs27fvre8IHjhwQFCV9HJychAZGYmcnBysW7cO1tbWOHbsGGxtbdGmTRvReZIwMTFBeno67O3tYW9vjx07dsDb2xt5eXlo06YNSktLRScSVat/91e1RERERNXo1KlTHEQB7N69G97e3rh+/ToOHjyIiooKXL9+HXFxcWp3j9Z2CQkJcHNzw/nz53HgwAE8f/4cAJCeno4FCxYIrpNOq1atcPPmTQBA27ZtsWXLFty7dw+bN29G48aNBdcRVT+ujBIREUksISEBq1atQmZmJmQyGZydnTFjxgz4+PiITtOoadOmYcmSJahfvz6mTZv2j99ds2aNRFViubu745NPPsFnn30GY2NjpKWlwcHBAZ988gkaN26MRYsWiU6UROfOnTF48GBMmzZN9evg6OiIixcvIigoCPfu3ROdKIkdO3agoqICI0eORGpqKnr37o0nT55AX18fUVFRGDJkiOhEomrFYZSIiEhC27dvx6hRozBo0CB4e3tDqVTi7NmzOHjwIKKiohAcHCw6UWP+ujU3ICDgb78nk8m05gCj+vXr49q1a7C3t4elpSVOnToFNzc3ZGZmolu3bigsLBSdKAkjIyNcvXoVDg4OasPo7du30bp1a5SXl4tOFKK0tBQ3btxAs2bNtO4qKNIOPE2XiIhIQsuWLcPKlSsxdepU1bPJkydjzZo1WLJkSa0eRk+dOvXWH2szCwsL/PnnnwAAGxsbZGRkwM3NDc+ePdOq9wPNzMxQWFgIBwcHteepqamwsbERVCWeoaEhvLy8RGcQaQyHUSIiIgnl5ubivffeq/K8f//+mDt3roAiEsnHxwcnTpyAm5sbPvzwQ0yePBlxcXE4ceIEunfvLjpPMsHBwZg1axb27dsHmUwGhUKBpKQkTJ8+HSNGjBCdp1Hcvv52z549w4ULF/Dw4UMoFAq1z2r7vxPahMMoERGRhGxtbREbGwsnJye157GxsbC1tRVUJY1Bgwb96+9qyymyGzZsUG1BnTNnDvT09JCYmIhBgwZh3rx5guuks2zZMowcORI2NjZQKpVwcXGBXC5HcHAwvvjiC9F5GpWamoqKigrVj/+OTCaTKkm4n3/+GcOGDcOLFy9gbGys9r9dJpNxGK1F+M4oERGRhDZt2oQpU6Zg9OjR6NKlC2QyGRITExEVFYV169bhk08+EZ2oMaNGjVL9WKlU4uDBgzA1NUX79u0BAJcvX8azZ88waNAgREZGisokgXJycpCamgqFQgFPT0+0aNFCdBIJ0LJlSwQGBuLLL7/827uIqXbgMEpERCSxgwcPYvXq1cjMzAQA1Wm6AwYMEFwmnVmzZqG4uBibN2+Grq4uAEAul2PChAkwMTHB119/LbiQpJSQkAA/Pz/RGVRD1K9fH1evXoWjo6PoFNIwDqNEREQkOSsrKyQmJqJVq1Zqz2/evIkuXbrgyZMngspIBH19fTRq1AjBwcH4+OOP4erqKjpJMty+XtWgQYMwdOhQfPjhh6JTSMP4zigREZGE7ty5A5lMhqZNmwIALly4gJ07d8LFxQXjxo0TXCedyspKZGZmVhlGMzMzqxxWQrXf/fv3sXv3buzatQsrV66Eq6srPv74YwQHB6t+r9RWpqamohNqnH79+mHGjBm4fv063NzcoKenp/Z5//79BZVRdePKKBERkYR8fHwwbtw4DB8+HEVFRWjZsiVcXV2RlZWFsLAwzJ8/X3SiJKZNm4aoqCjMnTsX77zzDgAgOTkZy5cvx4gRI7Tq1FBSl5eXh507d2LXrl24ceMGfH19tebeWXpNR0fnbz+TyWSQy+US1pAmcRglIiKSkLm5OZKTk9GqVSusX78ee/bsQVJSEo4fP47x48cjNzdXdKIkFAoFVq1ahXXr1qGwsBAA0LhxY0yePBnh4eGq90hJO8nlchw9ehTz5s1Denq6Vg0flZWViI+PR05ODoKDg2FsbIz79+/DxMQERkZGovOIqhWHUSIiIgkZGRkhIyMD9vb26N+/P7y9vTFr1iwUFBSgVatWKCsrE50ouZKSEgCAiYmJ4BLpDRw48K1XdshkMhgYGMDJyQnBwcFVtjPXVklJSdixYwf279+P8vJy9O/fH8OGDUPfvn1Fp0kiPz8fffr0QUFBAV6+fImsrCw4OjpiypQpKC8vx+bNm0UnSq68vBwGBgaiM0hD/n4NnIiIiKpdmzZtsHnzZpw5cwYnTpxAnz59ALx+Z65BgwaC68QwMTHRykEUeP2+YFxcHFJSUlRDaWpqKuLi4lBZWYk9e/bAw8MDSUlJgks1a+7cuXBwcEC3bt2Qn5+PiIgIFBUVYfv27VoziALA5MmT0b59ezx9+hT16tVTPR84cCBiY2MFlklLLpdjyZIlsLGxgZGRkWrHyLx587Bt2zbBdVSdOIwSERFJaMWKFdiyZQv8/f3x0UcfwcPDAwBw5MgRdOzYUXAdSe3NCbK5ubn46aefcODAAeTk5ODjjz9G8+bNkZmZiZCQEMyaNUt0qkbFx8dj+vTpuHfvHn799VcEBwdr5f2SiYmJ+OKLL6Cvr6/23M7ODvfu3RNUJb1ly5YhKioKK1euVPu1cHNzw/fffy+wjKobt+kSERFJTC6Xo6SkBObm5qpnt2/fhqGhIaytrQWWkdSsrKyQlJSEli1bqj3PyspCly5d8PjxY1y9ehU+Pj549uyZmEiSjIWFBRITE+Hi4gJjY2OkpaXB0dERiYmJeP/99/HgwQPRiZJwcnLCli1b0L17d7Vfhxs3bqBz5854+vSp6ESqJlwZJSIikpiurq7aIAoA9vb2HES1UGVlJW7cuFHl+Y0bN1SH9hgYGLz1vdLa5scff4S3tzeaNGmC/Px8AEBERAQOHz4suEw6PXv2REREhOrnMpkMz58/x4IFCxAYGCguTGL37t2Dk5NTlecKhQIVFRUCikhTOIwSERERCTJ8+HCEhoZi7dq1SExMRFJSEtauXYvQ0FCMGDECAJCQkIA2bdoILtWsTZs2Ydq0aQgMDMSzZ89Ug7iZmZnacFbbrVmzBgkJCXBxcUF5eTmCg4Nhb2+Pe/fuYcWKFaLzJNOmTRucOXOmyvN9+/bB09NTQBFpCrfpEhERkRCxsbGIjY3Fw4cPoVAo1D774YcfBFVJSy6XY/ny5diwYYNqC2bDhg0xadIkzJo1C7q6uigoKICOjg6aNm0quFZzXFxc8OWXXyIoKEhtW2ZGRgb8/f3x+PFj0YmSKSsrw+7du3H58mUoFAp4eXlh2LBhagca1XY///wzhg8fjjlz5mDx4sVYtGgRbt68iZiYGPzyyy/o2bOn6ESqJhxGiYiISHKLFi3C4sWL0b59ezRu3LjKNtSDBw8KKhNHm6+4qVevHm7cuAE7Ozu1YTQ7Oxvu7u5aceVRRUUFWrVqhV9++QUuLi6ic4T7/fff8eWXX6oN5fPnz0evXr1Ep1E1qiM6gIiIiLTP5s2bERUVheHDh4tOqTG0cQh9w8HBAVeuXIGdnZ3a86NHj2rNYKanp4eXL19qxfvB/0bv3r3Ru3dv0RmkYXxnlIiISGIJCQl477334OTkhBYtWqB///5vfT+qNnv16hW6dOkiOkO4Bw8eYPjw4WjSpAnq1KkDXV1dtX+0xYwZM/DZZ59hz549UCqVuHDhApYtW4a5c+dixowZovMkM2nSJKxYsQKVlZWiU2qEV69e4e7duygoKFD7h2oPbtMlIiKS0Pbt2zFq1CgMGjQI3t7eUCqVOHv2LA4ePIioqCgEBweLTpTErFmzYGRkhHnz5olOEapv374oKCjAxIkT37pdecCAAYLKpLd161YsXboUd+7cAQDY2Nhg4cKFCA0NFVwmnYEDByI2NhZGRkZwc3ND/fr11T4/cOCAoDJpZWdnY/To0Th79qzac6VSCZlMpjrgiv73cRglIiKSkLOzM8aNG4epU6eqPV+zZg22bt2KzMxMQWXSmjx5MmJiYuDu7g53d3fo6empfb5mzRpBZdIyNjbGmTNn0LZtW9EpNcbjx4+hUCi08qqjUaNG/ePnkZGREpWI5e3tjTp16mD27Nlv/UsaDw8PQWVU3TiMEhERSahu3bq4du1alTv0bt26BVdXV5SXlwsqk1ZAQMDffiaTyRAXFydhjTguLi7YsWMHr6sg+ov69evj8uXLaN26tegU0jAeYERERCQhW1tbxMbGVhlGY2NjYWtrK6hKeqdOnRKdUCNERERg9uzZ2LJlC+zt7UXnENUILi4uWnWdjzbjyigREZGENm3ahClTpmD06NHo0qULZDIZEhMTERUVhXXr1uGTTz4RnUgSMjc3R2lpKSorK2FoaFhlu3JxcbGgMiJpvbnaCAAuXbqEL774Al9++SXc3Nyq/L7Q5pOnaxsOo0RERBI7ePAgVq9erXo/1NnZGTNmzKj1h9UMGjQIUVFRMDExwaBBg/7xu9pyUEt0dPQ/fh4SEiJRCZFYOjo6au+Gvjms6K94gFHtw226REREEhs4cCAGDhwoOkNypqamqj9cmpqaCq6pGThsEr3GrfvaiSujREREEnJ0dMTFixfRoEEDtefPnj2Dl5cXcnNzBZWRVEpKSlTbDP+6NfFttGk7YkJCAlatWoXMzEzIZDLVjgEfHx/RaUSkIRxGiYiIJKSjo4OioqIq11Y8ePAAzZo1w8uXLwWVkVR0dXVRWFgIa2vrKlsT39C27YjafP/u+vXr//V3w8LCNFhSc3h7e8PPzw/+/v7w9vauct8q1R4cRomIiCRw5MgRAEBQUBCio6PVtqnK5XLExsbixIkTuHnzpqhEkkhCQoLqHsWEhIR//K6fn59EVWJp8/27Dg4Oaj9/9OgRSktLYWZmBuD1rglDQ0NYW1trzc6Jr776CgkJCTh79izKy8vRrl071XDatWtXGBkZiU6kasJhlIiISAI6OjoAXt+h+Z//r1dPTw/29vZYvXo13n33XRF5RELx/t3Xdu7ciW+//Rbbtm1Dq1atAAA3b97E2LFj8cknn2DYsGGCC6Ull8tx8eJFxMfHIz4+HnFxcZDJZNxBUovwACMiIiIJKBQKAK9XQS5evAhLS0vBRSRKenr6v/6uu7u7BktqDt6/+9q8efOwf/9+1SAKAK1atcLatWvxwQcfaN0wmp2djbS0NKSlpSE9PR0mJiZ8h7iW4TBKREQkoby8PNEJNU55eTkMDAxEZ0imbdu2qhXyt70v+lfa8s5oeHg4wsLCcOXKlbfev6stCgsLUVFRUeW5XC7HgwcPBBSJMWTIEJw+fRoKhQK+vr7w9fXFnDlztOYvZ7QJt+kSERGR5BQKBZYtW4bNmzfjwYMHyMrKgqOjI+bNmwd7e3uEhoaKTtSY/Px81Y9TU1Mxffp0zJgxA507dwYAnDt3DqtXr8bKlSsRFBQkqFJ62nr/7l+99957KCgowLZt29CuXTvIZDJcunQJY8eOha2trerd89pOR0cHlpaWGDlyJAICAuDj48P3RGspDqNEREQkucWLFyM6OhqLFy/G2LFjkZGRAUdHR+zduxdr167FuXPnRCdKomPHjli4cCECAwPVnv/222+YN28eLl++LKiMRHj06BFCQkJw7Ngx6OnpAQAqKyvRu3dvREVFVTmFu7Z69uwZTp8+jfj4eCQkJODatWvw8PCAv78//P390bdvX9GJVE04jBIREZHknJycsGXLFnTv3h3GxsZIS0uDo6Mjbty4gc6dO+Pp06eiEyVRr149pKSkwNnZWe15ZmYmvLy8UFZWJqhMWrx/V11WVhZu3LgBpVIJZ2dntGzZUnSSUDk5OVi6dCm2b98OhUKhNdvXtQHfGSUiIiLJ3bt3r8phNcDr7btve2eutnJ2dsbSpUuxbds21XuzL1++xNKlS6sMqLXZ7du33zpgvHz5Evfu3RNQJFbLli21egAtLi5GQkKC6hTda9euwcLCAgMGDEBAQIDoPKpGHEaJiIgk5O/vj9GjR2Pw4MGoV6+e6Bxh2rRpgzNnzsDOzk7t+b59++Dp6SmoSnqbN2/Ge++9B1tbW3h4eAAA0tLSIJPJ8Msvvwiu07y/vgP5+++/v/X+XXt7ewFlYsjlckRFRSE2NhYPHz5UncL9RlxcnKAyaVlZWcHS0hI+Pj4YO3Ys/P394erqKjqLNIDDKBERkYTatWuHmTNnYtKkSfjwww8RGhqKd955R3SW5BYsWIDhw4fj3r17UCgUOHDgAG7evImYmBitGMLe6NixI/Ly8rB9+3bVtswhQ4YgODgY9evXF52ncW8OaJLJZAgJCVH77K/372qLyZMnIyoqCv369YOrq+t/PW25tkpLS+PwqSX4zigREZHE5HI5fvnlF0RGRuK3336Dk5MTRo8ejeHDh6Nhw4ai8yTz+++/48svv8Tly5ehUCjg5eWF+fPno1evXqLTSGK8f/c1S0tLxMTEVDnQiqi24jBKREQk0KNHj7BlyxYsW7YMcrkcgYGBCAsLQ7du3USnkUSysrIQHx//1m2Z8+fPF1RFIjRp0gTx8fFa/b4oaRcOo0RERIJcuHABkZGR2LVrF0xNTTFy5EgUFhZix44d+PTTT7Fq1SrRiRpz8eJFKBQKdOrUSe35+fPnoauri/bt2wsqk9bWrVvx6aefwtLSEo0aNVLblimTyZCSkiKwjqS2evVq5ObmYsOGDVq7RZe0C4dRIiIiCT18+BA//vgjIiMjkZ2djffeew9jxoxB7969VX/4PHnyJIKCgvD8+XPBtZrTsWNHzJw5Ex988IHa8wMHDmDFihU4f/68oDJp2dnZYcKECZg1a5boFKoBBg4ciFOnTsHCwgJt2rRR3TX6xoEDBwSVEWkGDzAiIiKSUNOmTdG8eXOMHj0aI0eOhJWVVZXvdOzYER06dBBQJ53r16/Dy8urynNPT09cv35dQJEYT58+xeDBg0VnUA1hZmaGgQMHis4QLiYmBkOGDEHdunXVnr969Qq7d+/GiBEjBJVRdePKKBERkYTOnDkDHx8f0RnCNWjQAL/88gs6d+6s9vzs2bPo168fnj59KqhMWqGhoejQoQPGjx8vOoWoxtDV1UVhYSGsra3Vnj958gTW1tZvvZOW/jdxZZSIiEhCHERf69mzJ+bMmYPDhw+r7pZ89uwZ5s6di549ewquk46TkxPmzZuH5ORkuLm5VdmWGRYWJqhMWikpKdDT04ObmxsA4PDhw4iMjISLiwsWLlwIfX19wYUkJaVS+dZ3Zu/evat2Fy397+PKKBERkcT279+PvXv3oqCgAK9evVL7TFsOrLl37x58fX3x5MkTeHp6AgCuXLmChg0b4sSJE7C1tRVcKA0HB4e//UwmkyE3N1fCGnE6dOiA2bNn4/3330dubi7atGmDgQMH4uLFi+jXrx8iIiJEJ2qMl5cXYmNjYW5uDk9Pz388uKi2//fhzf/+tLQ0tGnTBnXq/N+6mVwuR15eHvr06YO9e/cKrKTqxJVRIiIiCa1fvx6ff/45QkJCcPjwYYwaNQo5OTm4ePEiPvvsM9F5krGxsUF6ejp27NiBtLQ01KtXD6NGjcJHH31UZXWwNsvLyxOdUCNkZWWhbdu2AIB9+/bB19cXO3fuRFJSEoYOHVqrh9EBAwao3o0MCgoSGyPYm//9V65cQe/evWFkZKT6TF9fH/b29nj//fcF1ZEmcBglIiKS0LfffovvvvsOH330EaKjozFz5kw4Ojpi/vz5KC4uFp0nqfr162PcuHGiM2qEV69eIS8vD82bN1dbDdIWSqVSdcfqyZMn8e677wIAbG1t8fjxY5FpGmdubg4dHR0AwKhRo9C0aVPVz7XNggULIJfLYWdnh969e6Nx48aik0jDuE2XiIhIQoaGhsjMzISdnR2sra1x4sQJeHh4IDs7G++88w6ePHkiOlEyWVlZiI+Px8OHD1WDyBvz588XVCWt0tJSTJo0CdHR0QBe/5o4OjoiLCwMTZo0wezZswUXSqNbt26wtbVFjx49EBoaiuvXr8PJyQkJCQkICQnB7du3RSdqTJ06dXD//n1YW1v/7cE92sbAwACZmZn/uI2dagft+6s3IiIigRo1aoQnT57Azs4OdnZ2SE5OhoeHB/Ly8qBNfz+8detWfPrpp7C0tESjRo3U3pOTyWRaM4zOmTMHaWlpiI+PR58+fVTPe/TogQULFmjNMBoREYFhw4bh0KFD+Pzzz+Hk5ATg9fvVXbp0EVynWU2aNMFPP/2EwMBAKJVK3L17F+Xl5W/9brNmzSSuE8PNzQ25ubkcRrUAV0aJiIgkNGbMGNja2mLBggXYvHkzpk2bBm9vb1y6dAmDBg3Ctm3bRCdKws7ODhMmTMCsWbNEpwhlZ2eHPXv24J133oGxsTHS0tLg6OiIW7duwcvLCyUlJaIThSovL4eurm6tfo/4u+++w6RJk1BZWfm333lzuqy2XGly/PhxzJo1C0uWLEG7du1Qv359tc9NTEwElVF14zBKREQkIYVCAYVCoXovcO/evUhMTISTkxPGjx+vNVdYmJiY4MqVK3B0dBSdIpShoSEyMjLg6OioNoympaXB19cXf/zxh+hESdy5cwcymQxNmzYFAFy4cAE7d+6Ei4uLVrxX/OeffyI/Px/u7u44efIkGjRo8NbveXh4SFwmxl/fmf3rrgltG8q1AYdRIiIiklxoaCg6dOiA8ePHi04Rys/PDx988AEmTZoEY2NjpKenw8HBARMnTsStW7dw7Ngx0YmS8PHxwbhx4zB8+HAUFRWhVatWaNOmDbKyshAWFqY127ajo6MxdOhQ1em62iohIeEfP/fz85OohDSN74wSERFJyNvbG35+fvD394e3t3eV7WfawsnJCfPmzUNycjLc3NyqbMMMCwsTVCatr776Cn369MH169dRWVmJdevW4dq1azh37tx//QN5bZKRkYGOHTsCeL1bwNXVFUlJSTh+/DjGjx+vNcNoSEiI6IQagcOm9uDKKBERkYS++uorJCQk4OzZsygvL0e7du1Uw2nXrl3V7tWrzf7pYBKZTIbc3FwJa8S6evUqVq1ahcuXL0OhUMDLywuzZs2Cm5ub6DTJGBkZISMjA/b29ujfvz+8vb0xa9YsFBQUoFWrVigrKxOdSBJ79uwZtm3bhszMTMhkMri4uGD06NEwNTUVnUbViMMoERGRAHK5HBcvXkR8fDzi4+MRFxcHmUyGly9fik4jklynTp0QEBCAfv36oVevXqpTppOTk/HBBx/g7t27ohNJQpcuXULv3r1Rr149dOzYEUqlEpcuXUJZWRmOHz8OLy8v0YlUTbTzRl0iIiLBsrOzkZaWhrS0NKSnp8PExASBgYGisyT36tUr3Lx58x9PEq3NUlJScPXqVdXPDx8+jKCgIMydOxevXr0SWCatFStWYMuWLfD398dHH32kOqjnyJEjqu27pD2mTp2K/v374/bt2zhw4AAOHjyIvLw8vPvuu5gyZYroPKpGXBklIiKS0JAhQ3D69GkoFAr4+vrC19cXfn5+cHd3F50mqdLSUkyaNAnR0dEAgKysLDg6OiIsLAxNmjTRmvs1O3TogNmzZ+P9999Hbm4uXFxcMGjQIFy8eBH9+vVDRESE6ETJyOVylJSUwNzcXPXs9u3bMDQ0hLW1tcAyMd78Ef2vp8lqi3r16iE1NRWtW7dWe379+nW0b98epaWlgsqounFllIiISEL79u2DXC5HSEgIRo8ejVGjRmndIAoAc+bMQVpaGuLj42FgYKB63qNHD+zZs0dgmbSysrLQtm1bAK//3fDz88POnTsRFRWFn376SWycxJRKJS5fvowtW7bgzz//BADo6+vD0NBQcJm0YmJi4Obmhnr16qFevXpwd3fHjz/+KDpLUiYmJigoKKjy/M6dOzA2NhZQRJrC03SJiIgkVFxcjNOnTyM+Ph5ffPEFrl27Bg8PD/j7+8Pf3x99+/YVnSiJQ4cOYc+ePXjnnXfUVn5cXFyQk5MjsExaSqUSCoUCAHDy5Em8++67AABbW1s8fvxYZJqk8vPz0adPHxQUFODly5fo2bMnjI2NsXLlSpSXl2Pz5s2iEyWxZs0azJs3DxMnToS3tzeUSiWSkpIwfvx4PH78GFOnThWdKIkhQ4YgNDQUq1atQpcuXSCTyZCYmIgZM2bgo48+Ep1H1YjbdImIiATKycnB0qVLsX37digUCq25zN3Q0BAZGRlwdHSEsbEx0tLS4OjoiLS0NPj6+uKPP/4QnSiJbt26wdbWFj169EBoaCiuX78OJycnJCQkICQkBLdv3xadKImgoCAYGxtj27ZtaNCggerfh4SEBIwZMwbZ2dmiEyXh4OCARYsWYcSIEWrPo6OjsXDhQuTl5Qkqk9arV68wY8YMbN68WfU+uZ6eHj799FMsX75c6+9hrU24MkpERCSh4uJiJCQkqE7RvXbtGiwsLDBgwAAEBASIzpNMhw4d8Ouvv2LSpEkA/u+9uK1bt6Jz584i0yQVERGBYcOG4dChQ/j888/h5OQEANi/fz+6dOkiuE46iYmJSEpKgr6+vtpzOzs73Lt3T1CV9AoLC9/6f/cuXbqgsLBQQJEY+vr6WLduHb766ivk5ORAqVTCyclJ67ZsawMOo0RERBKysrKCpaUlfHx8MHbsWPj7+8PV1VV0luS++uor9OnTB9evX0dlZSXWrVuHa9eu4dy5c0hISBCdJxl3d3e103Tf+Prrr6GrqyugSIy/2xVw9+5drXpH0MnJCXv37sXcuXPVnu/ZswctWrQQVCWOoaEhzMzMIJPJOIjWUtymS0REJKGMjAytHD7f5urVq1i1ahUuX74MhUIBLy8vzJo1C25ubqLTSGJDhgyBqakpvvvuOxgbGyM9PR1WVlYYMGAAmjVrhsjISNGJkvjpp58wZMgQ9OjRA97e3qp3JWNjY7F3714MHDhQdKIkKisrsWjRIqxfvx7Pnz8HABgZGWHSpElYsGAB9PT0BBdSdeEwSkRERERC3b9/HwEBAdDV1UV2djbat2+P7OxsWFpa4vTp01p1tcvly5exdu1aZGZmQqlUwsXFBeHh4fD09BSdJpnx48fj4MGDWLx4sWrb/rlz57Bw4UIMGDBAaw600gYcRomIiEgSJSUlMDExUf34n7z5HmmPsrIy7Nq1CykpKaqV8mHDhqFevXqi00hipqam2L17d5XTxY8ePYqhQ4dqzQFn2oDDKBEREUlCV1cXhYWFsLa2ho6OjtqVLm8olUrIZDKtOVWY6K/++nvkr548eQJra2ut+X3RsGFDxMfHw9nZWe15ZmYmfH198ejRI0FlVN14gBERERFJIi4uDhYWFgCAU6dOCa6pGeLj4+Hv7y86o0bIyspCfHw8Hj58qLp79Y358+cLqpLW360RvXz5sspJw7XZZ599hiVLliAyMlJ1jcvLly+xbNkyTJw4UXAdVSeujBIREUkoJiYGQ4YMqXJP3qtXr7B79+4q9wtS7WZgYAAbGxuMGjUKISEhsLW1FZ0kxNatW/Hpp5/C0tISjRo1Uls1l8lkSElJEVineevXrwcATJ06FUuWLIGRkZHqM7lcjtOnT+P27dtITU0VlSipgQMHIjY2FnXr1oWHhwcAIC0tDa9evUL37t3VvnvgwAERiVRNOIwSERFJSJu34aWnp//r77q7u2uwpOYoLi7G9u3bERUVhfT0dHTv3h2hoaEICgrSqpUwOzs7TJgwAbNmzRKdIoSDgwMAID8/H02bNlW71kdfXx/29vZYvHgxOnXqJCpRUqNGjfrX39WWk5ZrKw6jREREEtLR0cGDBw9gZWWl9jwtLQ0BAQEoLi4WVKZ5b94T/W9/9NDWd0avXLmCH374Abt27YJCocCwYcMQGhqqWhmqzUxMTHDlyhU4OjqKThEqICAABw4cgLm5uegUIklwGCUiIpKAp6cnZDIZ0tLS0KZNG9Sp83/HNsjlcuTl5aFPnz7Yu3evwErNys/P/9fftbOz02BJzXX//n189913WL58OerUqYPy8nJ07twZmzdvRps2bUTnaUxoaCg6dOiA8ePHi06hGuThw4e4efMmZDIZWrZsqVVX/GgLHmBEREQkgaCgIACvV7969+6t9k7Ym21477//vqA6aWjrgPnfVFRU4PDhw/jhhx9w4sQJtG/fHhs2bMBHH32E4uJizJo1C4MHD8b169dFp2qMk5MT5s2bh+TkZLi5uUFPT0/t87CwMEFlJEJJSQk+++wz7N69W7VLQldXF0OGDMHGjRthamoquJCqC1dGiYiIJBQdHY2hQ4dWOcBIGxw5cgR9+/aFnp4ejhw58o/f7d+/v0RVYk2aNAm7du0CAHz88ccYM2YMXF1d1b5TUFAAe3v7KifM1iZv3pl8G5lMhtzcXAlrSLQPP/wQV65cwTfffIPOnTtDJpPh7NmzmDx5Mtzd3Wv1DhJtw2GUiIhIQnfu3IFMJkPTpk0BABcuXMDOnTvh4uKCcePGCa7TLB0dHRQVFanuGf072vTOaPfu3TFmzBi8//77f3tgUWVlJZKSkuDn5ydxHZEY9evXx++//46uXbuqPT9z5gz69OmDFy9eCCqj6sZtukRERBIKDg7GuHHjMHz4cBQVFaFHjx5wdXXF9u3bUVRUVKvvU/zryl5tXuX7/xEbG/tfv1OnTh0OoqRVGjRo8NatuKampjzcqZbhyigREZGEzM3NkZycjFatWmH9+vXYs2cPkpKScPz4cYwfP16rtyM+e/YMZmZmojMkd/PmTXzzzTfIzMyETCZD69atMWnSJLRq1Up0mkZNmzbtX393zZo1GiypOby9veHn5wd/f394e3ujfv36opOE+O6777Bv3z7ExMSgcePGAICioiKEhIRg0KBB+OSTTwQXUnXhyigREZGEKioqVO+Lnjx5UvVuZOvWrVFYWCgyTVIrVqyAvb09hgwZAgAYPHgwfvrpJzRu3Bi//fabVlxnAgD79+/HRx99hPbt26Nz584AgOTkZLi6umLnzp0YPHiw4ELNSU1NVfv55cuXIZfLVUN4VlYWdHV10a5dOxF5Qrz77rtISEjAhg0bUF5ejnbt2qmG065du6odfFabbdq0Cbdu3YKdnR2aNWsG4PW703Xr1sWjR4+wZcsW1XdTUlJEZVI14MooERGRhDp16oSAgAD069cPvXr1QnJyMjw8PJCcnIwPPvgAd+/eFZ0oCUdHR2zfvh1dunTBiRMn8OGHH2LPnj3Yu3cvCgoKcPz4cdGJknB0dMTHH3+MxYsXqz1fsGABfvzxR61ZKV+zZg3i4+MRHR2t2ob59OlTjBo1Cj4+PggPDxdcKC25XI6LFy8iPj4e8fHxiIuLg0wmw8uXL0WnSWLRokX/+rsLFizQYAlpGodRIiIiCcXHx2PgwIEoKSlBSEgIfvjhBwDA3LlzcePGDRw4cEBwoTTq1auHrKws2NraYvLkySgvL8eWLVuQlZWFTp064enTp6ITJWFoaIj09HQ4OTmpPc/OzoaHhwdKS0sFlUnLxsYGx48fr3KXakZGBnr16oX79+8LKhPjxo0bSEhIQHx8PBISEvDq1Sv4+Pjg4MGDotOIqhW36RIREUnI398fjx8/RklJidpBHOPGjYOhoaHAMmmZm5vjzp07sLW1xbFjx7B06VIAgFKp1JqTdIHX/z6cOXOmyjCamJgIHx8fQVXSKykpwYMHD6oMow8fPsSff/4pqEp6Q4YMwenTp6FQKODr6wtfX1/MmTMH7u7uotOINILDKBERkcSUSiUuX76MnJwcBAcHw9jYGPr6+lo1jA4aNAjBwcFo0aIFnjx5gr59+wIArly5UmUwq23+esdq//79MWvWLFy+fBnvvPMOgNfvjO7bt+//a6vi/7qBAwdi1KhRWL16tdqvw4wZMzBo0CDBddLZt28fLC0tMXLkSAQEBMDHx0dr3hP9Kx0dHchksr/9XJv+wqq24zZdIiIiCeXn56NPnz4oKCjAy5cvkZWVBUdHR0yZMgXl5eXYvHmz6ERJVFRUYN26dbhz5w5GjhwJT09PAEBERASMjIwwZswYwYWa8093rP6VNt23WlpaiunTp+OHH35ARUUFgNdX2oSGhuLrr7/WmlNlnz17htOnT6u25167dg0eHh7w9/eHv7+/6i9tarvDhw+r/byiogKpqamIjo7GokWLEBoaKqiMqhuHUSIiIgkFBQXB2NgY27ZtQ4MGDZCWlgZHR0ckJCRgzJgxyM7OFp1IJMyLFy+Qk5MDpVIJJycnrRlC/05OTg6WLl2K7du3Q6FQaM1fTvydnTt3Ys+ePVWGVfrfxW26REREEkpMTERSUhL09fXVntvZ2eHevXuCqsT48ccfsWXLFuTm5uLcuXOws7NDREQEHBwcMGDAANF5JED9+vW1+v3I4uJi1cFF8fHxuHbtGiwsLDBgwAAEBASIzhOuU6dOGDt2rOgMqkYcRomIiCT0d6sbd+/ehbGxsYAiMTZt2oT58+djypQpWLZsmerXxMzMDBERERxGtcyLFy+wfPlyxMbG4uHDh1AoFGqfa8sVN1ZWVrC0tISPjw/Gjh0Lf39/uLq6is6qEcrKyvDNN9+gadOmolOoGnEYJSIiklDPnj0RERGB7777DsDr9wKfP3+OBQsWIDAwUHCddL755hts3boVQUFBWL58uep5+/btMX36dIFlJMKYMWOQkJCA4cOHo3Hjxv94eE1tlpaWxuETr0/b/uu/A0qlEn/++ScMDQ2xfft2gWVU3fjOKBERkYTu37+PgIAA6OrqIjs7G+3bt0d2djYsLS1x+vRpWFtbi06URL169XDjxg3Y2dnB2NhY9e5sdnY23N3dUVZWJjqRJGRmZoZff/0V3t7eolNqhEePHuHmzZuQyWRo2bIlrKysRCdJKioqSm0Y1dHRgZWVFTp16qR2JRb97+PKKBERkYSaNGmCK1euYNeuXUhJSYFCoUBoaCiGDRuGevXqic6TjIODA65cuQI7Ozu150ePHoWLi4ugKhLF3NwcFhYWojOEe/HiBSZNmoSYmBjVVmVdXV2MGDEC33zzjdZc/zRy5EjRCSQRrowSERGR5CIjIzFv3jysXr0aoaGh+P7775GTk4OvvvoK33//PYYOHSo6UTIKhQK3bt1667uSvr6+gqqktX37dhw+fBjR0dFaM3C9zSeffIKTJ09iw4YNqlXixMREhIWFoWfPnti0aZPgQmkcO3YMRkZG6Nq1KwBg48aN2Lp1K1xcXLBx40aujtYiHEaJiIg07MiRI+jbty/09PRw5MiRf/xu//79JaoSb+vWrVi6dCnu3LkDALCxscHChQu16g7B5ORkBAcHIz8/H//5RzJtumfU09NTdaWLvb099PT01D5PSUkRVCYtS0tL7N+/H/7+/mrPT506hQ8//BCPHj0SEyYxNzc3rFixAoGBgbh69Srat2+P8PBwxMXFwdnZGZGRkaITqZpwmy4REZGGBQUFoaioCNbW1ggKCvrb72nL8FFZWYkdO3bgvffew9ixY/H48WMoFAqteV/2r8aPH4/27dvj119/1eqDe/7p94U2KS0tRcOGDas8t7a2RmlpqYAiMfLy8lTb9X/66Se89957+PLLL5GSkqJVB71pA66MEhERkeQMDQ2RmZlZ5Z1RbVO/fn2kpaXByclJdArVAN27d0eDBg0QExMDAwMDAK+vNAkJCUFxcTFOnjwpuFAaFhYWSExMhIuLC7p27YoRI0Zg3LhxuH37NlxcXLRqMK/tuDJKREQk2LNnz2BmZiY6Q1KdOnVCamqq1g+jnTp1wq1btziMEgBg3bp16NOnD5o2bQoPDw/IZDJcuXIFBgYG+P3330XnSaZr166YNm0avL29ceHCBezZswcAkJWVxXtGaxkOo0RERBJasWIF7O3tMWTIEADA4MGD8dNPP6Fx48b47bff4OHhIbhQGhMmTEB4eDju3r2Ldu3aoX79+mqfu7u7CyqT1qRJkxAeHo6ioiK4ublVeVeyNv86WFhYICsrC5aWllXulfxPxcXFEpaJ4+rqiuzsbGzfvh03btyAUqnE0KFDte607Q0bNmDChAnYv38/Nm3aBBsbGwCvT9vu06eP4DqqTtymS0REJCFHR0ds374dXbp0wYkTJ/Dhhx9iz5492Lt3LwoKCnD8+HHRiZLQ0dGp8kwmk0GpVGrNu7OAdv86REdHY+jQoahbt26VeyX/U0hIiIRlRCQVDqNEREQSqlevHrKysmBra4vJkyejvLwcW7ZsQVZWFjp16oSnT5+KTpREfn7+P36uLdt3+evw35WVlWnVqmBWVhbi4+PfetXP/PnzBVURaQa36RIREUnI3Nwcd+7cga2tLY4dO4alS5cCAJRKZa1eBftPHLJe46/Da5999hk2btxY5fmLFy/Qr18/xMfHSx8lwNatW/Hpp5/C0tISjRo1UlstlslkHEap1uEwSkREJKFBgwYhODgYLVq0wJMnT9C3b18AwJUrV3iIjZbKyclBREQEMjMzIZPJ4OzsjMmTJ6N58+ai0yRz/PhxfPHFF6q/nAFeD6La9n7g0qVLsWzZMsyaNUt0CpEkqr6oQERERBqzdu1aTJw4ES4uLjhx4gSMjIwAAIWFhZgwYYLgOpLa77//DhcXF1y4cAHu7u5wdXXF+fPn0aZNG5w4cUJ0nmSOHz+OyMhIrF27FgDw559/omfPnpDJZDh27JjgOuk8ffoUgwcPFp1BJBm+M0pEREQkiKenJ3r37o3ly5erPZ89ezaOHz+OlJQUQWXSy8jIgL+/P+bNm4fdu3ejbt26+PXXX6uctFybhYaGokOHDhg/frzoFKFOnDgBb29vGBoaik4hDeMwSkREJMD169dRUFCAV69eqT3v37+/oCISwcDAAFevXkWLFi3UnmdlZcHd3R3l5eWCysRITk5Gjx490KlTJ/zyyy9acXDR+vXrVT9+8eIF1qxZg379+r31qp+wsDCp84QwMTHBy5cv0a5dO/j5+cHf3x/e3t6qnSRUe3AYJSIiklBubi4GDhyIq1evqq7wAKA6qESbDjF69uwZ9u/fj5ycHMyYMQMWFhZISUlBw4YNVfcK1na2trZYs2ZNla2Ze/fuxfTp01FQUCCoTPM8PT3fep1Lfn4+rK2t1QbR2rxC7ODg8K++J5PJkJubq+GamkEul+PChQtISEhAfHw8zp49i/Lycnh5ecHf37/KTgL638UDjIiIiCQ0efJkODg44OTJk3B0dMSFCxfw5MkThIeHY9WqVaLzJJOeno4ePXrA1NQUt2/fxtixY2FhYYGDBw8iPz8fMTExohMlMXbsWIwbNw65ubno0qULZDIZEhMTsWLFCoSHh4vO06igoCDRCTVCXl6e6IQaR1dXF507d0bnzp0xe/ZsZGRkYNWqVdixYwcuXrzIYbQW4cooERGRhCwtLREXFwd3d3eYmpriwoULaNWqFeLi4hAeHo7U1FTRiZLo0aMHvLy8sHLlShgbGyMtLQ2Ojo44e/YsgoODcfv2bdGJklAqlYiIiMDq1atx//59AECTJk0wY8YMhIWFvXXlkKi2y8zMVK2KJiQkQC6Xo2vXrvD394efnx88PDxEJ1I14TBKREQkIXNzc1y+fBmOjo5o3rw5vv/+ewQEBCAnJwdubm4oLS0VnSgJU1NTpKSkoHnz5mrDaH5+Plq1aqV170oCr0+QBQBjY2PBJURi6ejowMrKClOmTEH//v3Rpk0b0UmkIdymS0REJCFXV1ekp6fD0dERnTp1wsqVK6Gvr4/vvvsOjo6OovMkY2BggJKSkirPb968CSsrKwFF4nEIJXotLCwMp0+fxsKFC3Ho0CH4+/vD398fPj4+PMSoluHKKBERkYR+//13vHjxAoMGDUJubi7effdd3LhxAw0aNMCePXvQrVs30YmSGDduHB49eoS9e/fCwsIC6enp0NXVRVBQEHx9fRERESE6UWO8vLwQGxsLc3Pzvz3E543afHAP0X/z7NkznDlzBgkJCUhISMDVq1fRtm1bJCcni06jasKVUSIiIgn17t1b9WNHR0dcv34dxcXFMDc316r3A1etWoXAwEBYW1ujrKwMfn5+KCoqQufOnbFs2TLReRo1YMAA1K1bV/Vjbfq/O9H/D4VCgcrKSrx69QovX75ERUWF1rxPri24MkpERCTArVu3kJOTA19fX9SrVw9KpVIrh5K4uDikpKRAoVDAy8sLPXr0EJ1EAmRnZ1e5a1UbHTt2DEZGRujatSsAYOPGjdi6dStcXFywceNGmJubCy6UxuTJkxEfH49r167BwsICvr6+qq26rq6uovOoGnEYJSIiktCTJ0/w4Ycf4tSpU5DJZMjOzoajoyNCQ0NhZmaG1atXi04kCTk6OuLixYto0KCB2vNnz57By8tLa+6V1NHRQePGjeHn5wc/Pz/4+/ujVatWorMk5+bmhhUrViAwMBBXr15Fhw4dMG3aNMTFxcHZ2RmRkZGiEyXxwQcfcPjUEhxGiYiIJDRixAg8fPgQ33//PZydnVWnyB4/fhxTp07FtWvXRCdqzPr16//1d8PCwjRYUnPo6OigqKgI1tbWas8fPHgAW1tbvHr1SlCZtB48eIC4uDjVdR5ZWVlo2LChajAdP3686ERJGBkZISMjA/b29li4cCEyMjKwf/9+pKSkIDAwEEVFRaITiaoVh1EiIiIJNWrUCL///js8PDzUrjTJy8uDm5sbnj9/LjpRYxwcHP7V92QyWa1fETxy5AgAICgoCNHR0TA1NVV9JpfLERsbixMnTuDmzZuiEoW6desWli5dih07dkChUEAul4tOkoSFhQUSExPh4uKCrl27YsSIERg3bhxu374NFxcXrbn6CQBycnIQERGBzMxMyGQyODs7Y/LkyWjevLnoNKpGPMCIiIhIQi9evIChoWGV548fP1YdalNb5eXliU6oMYKCggC8HrxDQkLUPtPT04O9vb1Wbdl+/vw5EhMTER8fj4SEBFy5cgXOzs6YNGkS/Pz8ROdJpmvXrpg2bRq8vb1x4cIF7NmzBwCQlZWFpk2bCq6Tzu+//47+/fujbdu28Pb2hlKpxNmzZ9GmTRv8/PPP6Nmzp+hEqiZcGSUiIpJQv3794OXlhSVLlsDY2Bjp6emws7PD0KFDoVAosH//ftGJknvzRxFtPMDJwcEBFy9ehKWlpegUofT09GBhYYHhw4cjICAAXbt2VVst1hYFBQWYMGEC7ty5g7CwMISGhgIApk6dCrlc/v+11f1/maenJ3r37o3ly5erPZ89ezaOHz/OK49qEQ6jREREErp+/Tr8/f3Rrl07xMXFoX///rh27RqKi4uRlJSkVVvQtm3bhrVr1yI7OxsA0KJFC0yZMgVjxowRXEZSCwoKQmJiInR1dVUH1/j7+8PZ2Vl0GglgYGCAq1evVjlhOSsrC+7u7igvLxdURtWN23SJiIgk5OLigvT0dGzatAm6urp48eIFBg0ahM8++wyNGzcWnSeZefPmYe3atZg0aRI6d+4MADh37hymTp2K27dvY+nSpYILpfPixQskJCSgoKCgyoFF2nKQ06FDhwAA6enpSEhIQGxsLBYuXAiZTAZ/f3/s3r1bbKBEdHV1UVhYWOVAqydPnsDa2lpr3p21srLClStXqgyjV65cqfJrQ//buDJKREREkrO0tMQ333yDjz76SO35rl27MGnSJDx+/FhQmbRSU1MRGBiI0tJSvHjxAhYWFnj8+DEMDQ1hbW1d6w9yepvU1FScOnUKp06dwrFjxyCTybTmVOG/O135/v37aN68OcrKygSVSWvx4sVYu3YtZs+ejS5dukAmkyExMRErVqxAeHg4vvjiC9GJVE24MkpERESSk8vlaN++fZXn7dq1Q2VlpYAiMaZOnYr33nsPmzZtgpmZGZKTk6Gnp4ePP/4YkydPFp0nmbVr1yI+Ph5nzpzBn3/+ibZt28LPzw+ffPIJfH19Redp3Jt3QWUyGb7//nsYGRmpPpPL5Th9+jRat24tKk9y8+bNg7GxMVavXo05c+YAAJo0aYKFCxdqzW4BbcGVUSIiIpLcpEmToKenhzVr1qg9nz59OsrKyrBx40ZBZdIyMzPD+fPn0apVK5iZmeHcuXNwdnbG+fPnERISghs3bohOlET79u1V74n6+vrCxMREdJKk3lx7lJ+fj6ZNm0JXV1f1mb6+Puzt7bF48WJ06tRJVKIwf/75JwDA2NhYcAlpAldGiYiISBLTpk1T/fjNCtDx48fxzjvvAACSk5Nx584djBgxQlSi5PT09FSnCDds2BAFBQVwdnaGqakpCgoKBNdJ59KlS6IThHpz7VFAQAAOHDgAc3NzwUU1B4fQ2o3DKBEREUkiNTVV7eft2rUD8Ppye+D1oSVWVla4du2a5G2ieHp64tKlS2jZsiUCAgIwf/58PH78GD/++CPc3NxE50nq2bNn2LZtGzIzMyGTyeDs7IzQ0FCtuuLl1KlTohNqhAcPHmD69OmIjY3Fw4cP8Z8bObXlICdtwG26REREEqusrER8fDxycnIQHBwMY2Nj3L9/HyYmJmrvilHtd+nSJfz5558ICAjAo0ePEBISgsTERDg5OSEyMhIeHh6iEyVx6dIl9O7dG/Xq1UPHjh2hVCpx6dIllJWV4fjx4/Dy8hKdKAm5XI6oqCjVEKZQKNQ+j4uLE1Qmrb59+6KgoAATJ05E48aNq9xBPGDAAEFlVN04jBIREUkoPz8fffr0QUFBAV6+fImsrCw4OjpiypQpKC8vx+bNm0UnEknOx8cHTk5O2Lp1K+rUeb1xr7KyEmPGjEFubi5Onz4tuFAaEydORFRUFPr16/fWIWzt2rWCyqRlbGyMM2fOoG3btqJTSMO4TZeIiEhCkydPRvv27ZGWloYGDRqong8cOBBjxowRWCa9ixcvYt++fW+9X/PAgQOCqkiES5cuqQ2iAFCnTh3MnDnzracu11a7d+/G3r17ERgYKDpFKFtb2ypbc6l24jBKREQkocTERCQlJUFfX1/tuZ2dHe7duyeoSnq7d+/GiBEj0KtXL5w4cQK9evVCdnY2ioqKMHDgQNF5GuXl5YXY2FiYm5vD09OzyurXX6WkpEhYJo6JiQkKCgqqXF9y584drTrARl9fH05OTqIzhIuIiMDs2bOxZcsW2Nvbi84hDeIwSkREJCGFQvHWwzfu3r2rVX/o/vLLL7F27Vp89tlnMDY2xrp16+Dg4IBPPvkEjRs3Fp2nUQMGDEDdunUBAEFBQWJjaoghQ4YgNDQUq1atQpcuXSCTyZCYmIgZM2bgo48+Ep0nmfDwcKxbtw4bNmz4x7+kqI3Mzc3V/je/ePECzZs3h6GhIfT09NS+W1xcLHUeaQjfGSUiIpLQkCFDYGpqiu+++w7GxsZIT0+HlZUVBgwYgGbNmiEyMlJ0oiTq16+Pa9euwd7eHpaWljh16hTc3NyQmZmJbt26obCwUHQiSejVq1eYMWMGNm/ejMrKSgCvr7359NNPsXz5ctXwXtsNHDgQp06dgoWFBdq0aVNlCKvN29ejo6P/9XdDQkI0WEJS4sooERGRhNauXYuAgAC4uLigvLwcwcHByM7OhqWlJXbt2iU6TzIWFhaqy+xtbGyQkZEBNzc3PHv2DKWlpYLrpHPx4kUoFAp06tRJ7fn58+ehq6urNe9L6uvrY926dfjqq6+Qk5MDpVIJJycnGBoaik6TlJmZWa3fpv53OGBqJ66MEhERSaysrAy7du1CSkoKFAoFvLy8MGzYMNSrV090mmSCg4PRvn17TJs2DcuWLcO6deswYMAAnDhxAl5eXrV6BeivOnbsiJkzZ+KDDz5Qe37gwAGsWLEC58+fF1RGJE5KSgr09PRUd+0ePnwYkZGRcHFxwcKFC6u8c0//uziMEhERkeSKi4tRXl6OJk2aQKFQYNWqVar7NefNmwdzc3PRiZIwMjJCeno6HB0d1Z7n5eXB3d1dtXpc27148QLLly//2/s1c3NzBZVJj/cQAx06dMDs2bPx/vvvIzc3Fy4uLhg0aBAuXryIfv36ISIiQnQiVRNu0yUiIpJYVlYW4uPj3/qH7vnz5wuqkpaOjg6aNGmi+vHMmTMxc+ZMAMCtW7e0ZhitW7cuHjx4UGUYLSwsVLvmpLYbM2YMEhISMHz48Lfer6kt/vMe4p49e8LY2BgrV67UqnuIs7KyVHeM7tu3D35+fti5cyeSkpIwdOhQDqO1iPb8V46IiKgG2Lp1Kz799FNYWlqiUaNGan/olslkWjOMBgYGIi4uDgYGBmrPb968ie7du+Pu3buCyqTVs2dPzJkzB4cPH4apqSkA4NmzZ5g7dy569uwpuE46R48exa+//gpvb2/RKULxHuLXlEql6i/qTp48iXfffRfA6/tHHz9+LDKNqhmHUSIiIgktXboUy5Ytw6xZs0SnCGVubo6goCD88ssvqhXANyfpfvjhh4LrpLN69Wr4+vrCzs4Onp6eAIArV66gYcOG+PHHHwXXScfc3BwWFhaiM4TjPcSvtW/fHkuXLkWPHj2QkJCATZs2AXi9fb1hw4aC66g66YgOICIi0iZPnz7F4MGDRWcI99NPP+HFixcIDg6GUqlERkYG/P398dFHH2HdunWi8yRjY2OD9PR0rFy5Ei4uLmjXrh3WrVuHq1evwtbWVnSeZJYsWYL58+dr1UnKb8N7iF+LiIhASkoKJk6ciM8//xxOTk4AgP3796NLly6C66g68QAjIiIiCYWGhqJDhw4YP3686BTh/vjjD/j7+6N58+Y4c+YMRowYga+//lp0FknE09NTbZv6rVu3oFQqYW9vX+V+zZSUFKnzhOA9xP+svLwcurq6Vf79oP9dHEaJiIg0bP369aofv3jxAmvWrEG/fv3g5uZW5Q9VYWFhUudJpqSkpMqzoqIi9OjRA++++y6WL1+uem5iYiJlmlA//vgjtmzZgtzcXJw7dw52dnZYu3YtHB0dMWDAANF5GrNo0aJ//d0FCxZosKTmuH//PgICAqCrq4vs7Gy0b99edQ/x6dOnYW1tLTqRqFpxGCUiItIwBweHf/U9mUxWq6+w0NHReespqW/+KCKTyaBUKiGTyd66VbE22rRpE+bPn48pU6Zg6dKluHbtGhwdHREVFYXo6GicOnVKdCJJjPcQkzbhMEpERESSSEhI+Nff9fPz02BJzeHi4oIvv/wSQUFBMDY2RlpaGhwdHVXv0GrLyaF37tyBTCZD06ZNAQAXLlzAzp074eLignHjxgmuIyJN4Wm6REREElq8eDGmT58OQ0NDtedlZWX4+uuva/XVLm8GzMrKSixbtgyjR4/WqkN63iYvL091iu5f1a1bFy9evBBQJEZwcDDGjRuH4cOHq7Zuu7q6Yvv27SgqKqrVvy+OHDnyr7/bv39/DZYQSY8ro0RERBLS1dVFYWFhlXe/njx5Amtra63ZnmpsbIyrV6/C3t5edIpQLi4u+OqrrzBgwAC1ldH169cjOjoaly9fFp0oCXNzcyQnJ6NVq1ZYv3499uzZg6SkJBw/fhzjx4+v9dvX/+rNdvX/fAZAa/778J/kcjmuXr0KOzs7mJubi86hasSrXYiIiCT05p3I/5SWlqZV9yx2794d8fHxojOEmzFjBj777DPs2bMHSqUSFy5cwLJlyzB37lzMmDFDdJ5kKioqULduXQDAyZMnVSuArVu3RmFhocg0jVMoFKp/jh8/jrZt2+Lo0aN49uwZ/vjjDxw9ehReXl44duyY6FTJTJkyBdu2bQPwehD18/ODl5cXbG1t+d+NWobbdImIiCRgbm4OmUwGmUyGli1bqg2kcrkcz58/16rrXvr27Ys5c+YgIyMD7dq1Q/369dU+15btiKNGjUJlZSVmzpyJ0tJSBAcHw8bGBuvWrcPQoUNF50mmTZs22Lx5M/r164cTJ05gyZIlAF6fLtugQQPBddKZMmUKNm/ejK5du6qe9e7dG4aGhhg3bhwyMzMF1kln//79+PjjjwEAP//8M/Ly8nDjxg3ExMTg888/R1JSkuBCqi7cpktERCSB6OhoKJVKjB49GhERETA1NVV9pq+vD3t7e3Tu3FlgobT+c2viX2nLabqVlZXYsWMHevfujUaNGuHx48dQKBRaeX1HfHw8Bg4ciJKSEoSEhOCHH34AAMydOxc3btzAgQMHBBdKo169erhw4QLc3NzUnqenp6NTp04oKysTVCYtAwMD3Lp1C02bNsW4ceNgaGiIiIgI5OXlwcPD463XRNH/Jg6jREREEkpISIC3tzfq1OHmJAIMDQ2RmZkJOzs70SnCyeVylJSUqL0TePv2bRgaGmrNgO7r6ws9PT1s374djRs3BvD6Lt7hw4fj1atX/18nUv8vs7Ozw9atW9G9e3c4ODjg22+/xbvvvotr166ha9euePr0qehEqib8/4REREQS0pYrS+jf6dSpE1JTUzmM4vXhXv95OI22HXD1ww8/YODAgbCzs0OzZs0AAAUFBWjZsiUOHTokNk5Co0aNwocffojGjRtDJpOhZ8+eAIDz58+jdevWguuoOnFllIiIiIR48eIFEhISUFBQgFevXql9FhYWJqhKWvv27cPs2bMxderUt7476+7uLqiMRFEqlThx4gRu3LgBpVIJFxcX9OjR460Hn9Vm+/fvx507dzB48GDV/bPR0dEwMzPDgAEDBNdRdeEwSkRERJJLTU1FYGAgSktL8eLFC1hYWODx48eqLZm1+SqPv3rbu7NvrvbQlndniUh7cZsuERERSW7q1Kl47733sGnTJpiZmSE5ORl6enr4+OOPMXnyZNF5ksnLyxOdQDXI4sWL//Hz+fPnS1QivfXr1//r72rLzgltwJVRIiIikpyZmRnOnz+PVq1awczMDOfOnYOzszPOnz+PkJAQ3LhxQ3QikeQ8PT3Vfl5RUYG8vDzUqVMHzZs3R0pKiqAyzXNwcFD7+aNHj1BaWgozMzMAwLNnz7Ru54Q24MooERGRhAICAv7x3a+4uDgJa8TR09NT/To0bNgQBQUFcHZ2hqmpKQoKCgTXSevmzZv45ptvkJmZCZlMhtatW2PSpElo1aqV6DTJaPOK4F+lpqZWeVZSUoKRI0di4MCBAoqk89ddAjt37sS3336Lbdu2qX4f3Lx5E2PHjsUnn3wiKpE0gCujREREEpo6darazysqKnDlyhVkZGQgJCQE69atE1QmrV69emHkyJEIDg7G+PHjkZqairCwMPz44494+vQpzp8/LzpREvv378dHH32E9u3bq+6ZTU5OxsWLF7Fz504MHjxYcKE0tHlF8N/IyMjAu+++i9u3b4tOkUTz5s2xf//+Kv9eXL58GR988AG3t9ciXBklIiKS0Nq1a9/6fOHChXj+/LnENeJ8+eWX+PPPPwEAS5YsQUhICD799FM4OTkhMjJScJ10Zs6ciTlz5lRZGVywYAFmzZqlNcOoNq8I/hvPnj3DH3/8ITpDMoWFhaioqKjyXC6X48GDBwKKSFO4MkpERFQD3Lp1Cx07dkRxcbHoFJKQoaEh0tPT4eTkpPY8OzsbHh4eKC0tFVRWM2jbiuB/HuKjVCpRWFiIH3/8Eb6+vti1a5egMmm99957KCgowLZt29CuXTvIZDJcunQJY8eOha2tLY4cOSI6kaoJV0aJiIhqgHPnzsHAwEB0BknM398fZ86cqTKMJiYmwsfHR1BVzaFtK4L/uXNCR0cHVlZWCAkJwZw5cwRVSe+HH35ASEgIOnbsCD09PQBAZWUlevfuje+//15wHVUnDqNEREQSGjRokNrP36x8XLp0CfPmzRNUJb0nT55g/vz5OHXqFB4+fAiFQqH2ubasEPfv3x+zZs3C5cuX8c477wB4/c7ovn37sGjRIrUVoP79+4vK1Lh/WhHs06ePoCrp8V3I16ysrPDbb78hKysLN27cgFKphLOzM1q2bCk6jaoZt+kSERFJaNSoUWo/f7Py0a1bN/Tq1UtQlfT69u2LnJwchIaGomHDhlVOGA4JCRFUJi0dHZ1/9T2ZTAa5XK7hGnH+81qPv/6+mDNnDoyNjQWVSWv06NFYt25dlf+9L168wKRJk/DDDz8IKiPSDA6jREREJDljY2MkJibCw8NDdApRjaGrq4vCwkJYW1urPX/8+DEaNWqEyspKQWXSksvliIqKQmxs7Ft3TmjLFVjagNt0iYiIBLh8+bLqXkkXF5cqVxjUdq1bt0ZZWZnoDKqB7t69C5lMBhsbG9EpkikpKYFSqYRSqcSff/6p9v64XC7Hb7/9VmVArc0mT56MqKgo9OvXD66urv94NzP9b+PKKBERkYQePnyIoUOHIj4+HmZmZlAqlfjjjz8QEBCA3bt3w8rKSnSiJC5evIjZs2dj/vz5cHV1VR1S8oaJiYmgMunFxsZi7dq1qr+caN26NaZMmYIePXqITpOMQqHA0qVLsXr1atUVR8bGxggPD8fnn3/+r7cz/6/S0dH5x4FLJpNh0aJF+PzzzyWsEsfS0hIxMTEIDAwUnUIaxpVRIiIiCU2aNAklJSW4du0anJ2dAQDXr19HSEgIwsLCtObqBjMzM/zxxx/o1q2b2nOlUlnr34/8qw0bNmDq1Kn44IMPMHnyZACvDzAKDAzEmjVrMHHiRMGF0vj888+xbds2LF++HN7e3lAqlUhKSsLChQtRXl6OZcuWiU7UqFOnTkGpVKJbt2746aefYGFhofpMX18fdnZ2aNKkicBCaenr61c5YZpqJ66MEhERScjU1BQnT55Ehw4d1J5fuHABvXr1wrNnz8SESaxjx46oU6cOJk+e/NYDjPz8/ASVScvGxgZz5sypMnRu3LgRy5Ytw/379wWVSatJkybYvHlzlRODDx8+jAkTJuDevXuCyqSVn5+PZs2aaf221NWrVyM3NxcbNmzQ+l+L2o4ro0RERBJSKBRVtqQCgJ6eXpVDOmqzjIwMpKamolWrVqJThCopKXnr1SW9evXCrFmzBBSJUVxcjNatW1d53rp1a6255gcAMjMzcefOHXTt2hXA67+U2Lp1K1xcXLBx40aYm5sLLpRGYmIiTp06haNHj6JNmzZV/pt54MABQWVU3Wr3BnwiIqIaplu3bpg8ebLaite9e/cwdepUdO/eXWCZtNq3b487d+6IzhCuf//+OHjwYJXnhw8fxnvvvSegSFp3794FAHh4eGDDhg1VPt+wYYNWnbg8Y8YMlJSUAACuXr2KadOmITAwELm5uZg2bZrgOumYmZlh4MCB8PPzg6WlJUxNTdX+odqD23SJiIgkdOfOHQwYMAAZGRmwtbWFTCZDQUEB3NzccPjwYTRt2lR0oiT27duHhQsXYsaMGXBzc6uy8uHu7i6oTFpLly7FqlWr4O3tjc6dOwN4/c5oUlISwsPD1Q5yCgsLE5WpMWZmZvjmm29gZ2eHwMBANGvWDJ07d4ZMJsPZs2dx584d/Pbbb/Dx8RGdKgkjIyNkZGTA3t4eCxcuREZGBvbv34+UlBQEBgaiqKhIdCJRteIwSkREJMCJEydw48YNKJVKuLi4aNXJqQDeejqqTCbTugOMHBwc/tX3ZDIZcnNzNVwjvW+//RazZ89Gz549ERERgS1btiAzM1P1+2LChAladXCPhYUFEhMT4eLigq5du2LEiBEYN24cbt++DRcXF5SWlopOlExlZSXi4+ORk5OD4OBgGBsb4/79+zAxMYGRkZHoPKomHEaJiIgklJeX968HkNosPz//Hz+3s7OTqIREy8vLQ2hoKK5fv44tW7ZgwIABopOE6d+/P169egVvb28sWbIEeXl5sLGxwfHjxzFx4kRkZWWJTpREfn4++vTpg4KCArx8+RJZWVlwdHTElClTUF5ejs2bN4tOpGrCYZSIiEhCurq68PX1RWhoKD744AO1y+2JtNmba26cnZ1Rp476GZspKSmCqqRVUFCACRMm4M6dOwgLC0NoaCgAYOrUqZDL5Vi/fr3gQmkEBQXB2NgY27ZtQ4MGDZCWlgZHR0ckJCRgzJgxyM7OFp1I1YTDKBERkYQyMjLwww8/YMeOHXj58iWGDBmC0aNHo1OnTqLTJPfjjz9i8+bNyMvLw7lz52BnZ4eIiAg4ODho1erY3bt3ceTIERQUFODVq1dqn61Zs0ZQlbTy8/MxcuRIXL9+HePGjasyjC5YsEBQGYlgaWmJpKQktGrVCsbGxqphVBu3K9d2vNqFiIhIQq6urlizZg1WrlyJn3/+GVFRUfDx8UGLFi0QGhqK4cOHw8rKSnSmxm3atAnz58/HlClTsGzZMtU7omZmZoiIiNCaYTQ2Nhb9+/eHg4MDbt68CVdXV9y+fRtKpRJeXl6i8ySxdetWhIeHo0ePHsjIyNCKf//pnykUire+N3737l0YGxsLKCJN4dUuREREAtSpUwcDBw7E3r17sWLFCuTk5GD69Olo2rQpRowYgcLCQtGJGvXNN99g69at+Pzzz6Grq6t63r59e1y9elVgmbTmzJmD8PBwZGRkwMDAAD/99BPu3LkDPz8/DB48WHSexvXp0wezZs3Chg0bcODAAQ6iBACqA63ekMlkeP78ORYsWIDAwEBxYVTtOIwSEREJcOnSJUyYMAGNGzfGmjVrMH36dOTk5CAuLg737t2r9SuDeXl58PT0rPK8bt26ePHihYAiMTIzMxESEgLg9V9QlJWVwcjICIsXL8aKFSsE12meXC5Heno6RowYITqFapC1a9ciISEBLi4uKC8vR3BwMOzt7XHv3j2t+H2hTbhNl4iISEJr1qxBZGQkbt68icDAQMTExCAwMFB11YmDgwO2bNmC1q1bCy7VLAcHB1y5cqXKqblHjx6Fi4uLoCrp1a9fHy9fvgQANGnSBDk5OWjTpg0A4PHjxyLTJHHixAnRCVQDNWnSBFeuXMGuXbuQkpIChUKB0NBQDBs2DPXq1ROdR9WIwygREZGENm3ahNGjR2PUqFFo1KjRW7/TrFkzbNu2TeIyac2YMQOfffYZysvLoVQqceHCBezatQtfffUVvv/+e9F5knnnnXeQlJQEFxcX9OvXD+Hh4bh69SoOHDiAd955R3QeSaiyshIGBga4cuUKXF1dRecIV69ePYwePRqjR48WnUIaxNN0iYiISIitW7di6dKluHPnDgDAxsYGCxcuVF1noQ1yc3Px/PlzuLu7o7S0FNOnT0diYiKcnJywdu1a3reqZZo3b44DBw7Aw8NDdIrkjhw5gr59+0JPTw9Hjhz5x+/2799foirSNA6jREREJKnKykrs2LEDvXv3RqNGjfD48WMoFApYW1uLTiMSKjIyEvv27cP27dthYWEhOkdSOjo6KCoqgrW1teq1hbeRyWRvPWmX/jdxGCUiIiLJGRoaIjMzkyt/RH/h6emJW7duoaKiAnZ2dqhfv77a5ykpKYLKiDSD74wSERGR5Dp16oTU1FQOo0R/ERQUJDpBuIqKCvTq1QtbtmxBy5YtReeQhnEYJSIiIslNmDAB4eHhuHv3Ltq1a1dlBcjd3V1QGZE4CxYsEJ0gnJ6eHjIyMiCTyUSnkAS4TZeIiEhCW7duhb+/P1q0aCE6Rai3vRMmk8mgVCr5ThiRlgsPD4eenh6WL18uOoU0jMMoERGRhFq3bo3s7Gw0bNgQfn5+8Pf3h5+fX62/V/Q/5efn/+Pn2rB9t6KiAq1atcIvv/yiVXerkjoLCwtkZWXB0tIS5ubm/7giWFxcLGGZOJMmTUJMTAycnJzQvn37Kjsn1qxZI6iMqhu36RIREUnoxo0bKCoqwqlTp5CQkIC1a9diwoQJsLKygr+/P3bv3i06URLaMGz+N3p6enj58iW3I2q5tWvXwtjYGAAQEREhNqaGyMjIgJeXFwAgKytL7TP+fqlduDJKREQkyIsXL5CYmIjdu3dj+/btUCqVqKysFJ1FElq+fDlu3LiB77//HnXqcI2AtFtubi4cHBw4cGoRDqNEREQSOnr0KBISEhAfH4+0tDS0adMGvr6+8Pf3h4+PD8zNzUUnkoQGDhyI2NhYGBkZwc3Nrcp2xAMHDggqI1FycnIQGRmJnJwcrFu3DtbW1jh27BhsbW3Rpk0b0Xkapauri8LCQtWdw0OGDMH69evRsGFDwWWkKRxGiYiIJKSjowMrKyuEh4fjk08+gampqegkEmjUqFH/+HlkZKREJVQTJCQkoG/fvvD29sbp06eRmZkJR0dHrFy5EhcuXMD+/ftFJ2qUjo4OioqKVMOosbEx0tLS4OjoKLiMNIXDKBERkYQiIiJw+vRpnDlzBrq6uqpDjPz9/eHs7Cw6j4gE6ty5MwYPHoxp06apDWIXL15EUFAQ7t27JzpRoziMap+q56oTERGRxkyZMgUHDhzAo0ePcOLECfj4+ODkyZPw8PBA48aNRedJ5vPPP8eJEydQWloqOoWoxrh69SoGDhxY5bmVlRWePHkioEhaMpmsyvuifH+0duOb8kRERAKkpqYiPj4ep06dwpkzZ6BQKNC0aVPRWZK5fPkyvvnmG7x8+RJeXl6qK266du0KIyMj0XmS+W+HteTm5kpYQ6KZmZmhsLAQDg4Oas9TU1NhY2MjqEo6SqUSI0eORN26dQEA5eXlGD9+PN+lrsU4jBIREUmof//+SExMRElJCdq2bQt/f3+MGzcOvr6+MDExEZ0nmWPHjkEul+PChQuqA52+/fZblJWVwcvLC8nJyaITJTFlyhS1n1dUVCA1NRXHjh3DjBkzxESRMMHBwZg1axb27dsHmUwGhUKBpKQkTJ8+HSNGjBCdp3EhISFqP//4448FlZBU+M4oERGRhKZPnw5/f3+tGz7/yc2bNxEfH4+TJ0/i0KFDMDMzw6NHj0RnCbVx40ZcunSJBxhpmYqKCowcORK7d++GUqlEnTp1IJfLERwcjKioKOjq6opOJKpWHEaJiIhIcps2bUJCQgISEhIgl8vh4+OjOszJ3d1ddJ5wubm5aNu2LUpKSkSnkAA5OTlITU2FQqGAp6cnWrRoITqJSCM4jBIREUksISEBq1atQmZmJmQyGZydnTFjxgz4+PiITpPMX6+4GT9+PFeJ/8PKlSvx7bff4vbt26JTiIg0hsMoERGRhLZv345Ro0Zh0KBB8Pb2hlKpxNmzZ3Hw4EFERUUhODhYdKIkDh06hNOnTyM+Ph7Xr1+Hh4eH6oobHx8frTnEyNPTU+0AI6VSiaKiIjx69Ajffvstxo0bJ7COpCaXyxEVFYXY2Fg8fPgQCoVC7fO4uDhBZUSawWGUiIhIQs7Ozhg3bhymTp2q9nzNmjXYunUrMjMzBZWJ88cff+DMmTPYv38/du7cCZlMhpcvX4rOksSiRYvUfv5mxdjf3x+tW7cWVEWiTJw4EVFRUejXrx8aN25c5aTltWvXCioj0gwOo0RERBKqW7curl27BicnJ7Xnt27dgqurK8rLywWVSa+4uFh1km58fDwyMjLQoEED+Pn5Yd++faLziCRnaWmJmJgYBAYGik4hkgSvdiEiIpKQra0tYmNjqwyjsbGxsLW1FVQlPXd3d1y/fh0WFhbw9fXF2LFj4e/vD1dXV9FpkpPL5Th06JDqHWIXFxf079+fJ6dqIX19/Sr/bSCqzTiMEhERSSg8PBxhYWG4cuUKunTpAplMhsTERERFRWHdunWi8yQzbtw4rR0+/+rWrVsIDAzEvXv30KpVKyiVSmRlZcHW1ha//vormjdvLjqRJBQeHo5169Zhw4YNVbboEtVG3KZLREQksYMHD2L16tWq90PfnKY7YMAAwWVivPmjiDb+4TswMBBKpRI7duyAhYUFAODJkyf4+OOPoaOjg19//VVwIWnaoEGD1H4eFxcHCwsLtGnTBnp6emqfHThwQMo0Io3jMEpERERCxMTE4Ouvv0Z2djYAoGXLlpgxYwaGDx8uuEw69evXR3JyMtzc3NSep6WlwdvbG8+fPxdURlIZNWrUv/5uZGSkBkuIpMdtukRERAK8evXqrVc3NGvWTFCRtNasWYN58+Zh4sSJqitukpKSMH78eDx+/LjKacO1Vd26dfHnn39Wef78+XPo6+sLKCKpccAkbcaVUSIiIgllZ2dj9OjROHv2rNpzpVIJmUwGuVwuqExaDg4OWLRoEUaMGKH2PDo6GgsXLkReXp6gMmmNGDECKSkp2LZtGzp27AgAOH/+PMaOHYt27dohKipKbCBJqlu3bjhw4ADMzMzUnpeUlCAoKIj3jFKtw2GUiIhIQt7e3qhTpw5mz5791nsEPTw8BJVJy8DAABkZGVVODs3Ozoabm5vWXHHz7NkzhISE4Oeff1a9H1hZWYn+/fsjKioKpqamggtJSjo6OigqKoK1tbXa84cPH8LGxgYVFRWCyog0g9t0iYiIJHTlyhVcvnwZrVu3Fp0ilJOTE/bu3Yu5c+eqPd+zZw9atGghqEp6ZmZmOHz4MG7duoXMzEwolUq4uLjweg8tk56ervrx9evXUVRUpPq5XC7HsWPHYGNjIyKNSKM4jBIREUnIxcUFjx8/Fp0h3KJFizBkyBCcPn0a3t7eqituYmNjsXfvXtF5knNycuIAqsXatm0LmUwGmUyGbt26Vfm8Xr16+OabbwSUEWkWt+kSERFpWElJierHly5dwhdffIEvv/wSbm5uVa5uMDExkTpPmMuXL2Pt2rVqK4Lh4eHw9PQUnUYkqfz8fCiVSjg6OuLChQuwsrJSfaavrw9ra2vo6uoKLCTSDA6jREREGqajo6P2buibw4r+StsOMCIiIuI2XSIiIg07deqU6IQaSaFQ4NatW2+94sbX11dQFRERSYUro0RERCS55ORkBAcHq7Yn/hVXiImItIOO6AAiIiJtcuzYMSQmJqp+vnHjRrRt2xbBwcF4+vSpwDJpjR8/Hu3bt0dGRgaKi4vx9OlT1T/FxcWi8yTj7e2NuXPn4vjx43jx4oXoHCIiSXFllIiISEJubm5YsWIFAgMDcfXqVbRv3x7h4eGIi4uDs7MzIiMjRSdKon79+khLS9P6E2S/+uorJCQk4OzZsygvL0e7du3g5+cHf39/dO3aFUZGRqITiYg0hsMoERGRhIyMjJCRkQF7e3ssXLgQGRkZ2L9/P1JSUhAYGKh2v2Bt1q1bN8ycORN9+vQRnVIjyOVyXLx4EfHx8YiPj0dcXBxkMhlevnwpOo0k9Pnnn8Pf3x/e3t4wNDQUnUOkcTzAiIiISEL6+vooLS0FAJw8eRIjRowAAFhYWKhdAVMbpaenq348adIkhIeHo6io6K1X3Li7u0udJ1R2djbS0tKQlpaG9PR0mJiYwMfHR3QWSezy5cv45ptv8PLlS3h5ecHf3x9+fn5cJadaiyujREREEurfvz9evXoFb29vLFmyBHl5ebCxscHx48cxceJEZGVliU7UmDdX3PzdHz3efKZNBxgNGTIEp0+fhkKhgK+vL3x9feHn56d1wzj9H7lcjgsXLiAhIQHx8fE4d+4cysrK4OXlheTkZNF5RNWKK6NEREQS2rBhAyZMmID9+/dj06ZNsLGxAQAcPXq01m9ZzcvLE51Q4+zbtw+WlpYYOXIkAgIC4OPjwxUwLaerq4vOnTvDwsIC5ubmMDY2xqFDh5CTkyM6jajacWWUiIiISJBnz57h9OnTiI+PR0JCAq5duwYPDw/4+/vD398fffv2FZ1IEtq0aRMSEhKQkJAAuVwOHx8f1YFWXC2n2ojDKBEREUkuOjoalpaW6NevHwBg5syZ+O677+Di4oJdu3bBzs5OcKEYOTk5WLp0KbZv3w6FQqE125XpNR0dHVhZWSE8PBzjx4+HiYmJ6CQijeIwSkRERJJr1aoVNm3ahG7duuHcuXPo3r07IiIi8Msvv6BOnTo4cOCA6ERJFBcXq94NjI+Px7Vr12BhYQFfX18EBATgs88+E51IEjp06JBqpfz69etqq+Tcwk21EYdRIiIikpyhoSFu3LiBZs2aYdasWSgsLERMTAyuXbsGf39/PHr0SHSiJHR1dWFpaQkfHx/V0OHq6io6i2qAP/74A2fOnMH+/fuxc+dOXvVDtRIPMCIiIiLJGRkZ4cmTJ2jWrBmOHz+OqVOnAgAMDAxQVlYmuE46aWlpHD5JzX+ulmdkZKBBgwbw8/MTnUZU7TiMEhERkeR69uyJMWPGwNPTE1lZWap3R69duwZ7e3uxcRJ6M4g+evQIN2/ehEwmQ8uWLWFlZSW4jERwd3fH9evXVVu1x44dy9VyqtU4jBIREUlg0KBB/+p72vKu5MaNG/HFF1/gzp07+Omnn9CgQQMAwOXLl/HRRx8JrpPOixcvMGnSJMTExEChUAB4vXV3xIgR+Oabb2BoaCi4kKQ0btw4Dp+kVfjOKBERkQRGjRr1r74XGRmp4RKqST755BOcPHkSGzZsgLe3NwAgMTERYWFh6NmzJzZt2iS4kER580d0mUwmuIRIcziMEhEREQliaWmJ/fv3w9/fX+35qVOn8OGHH2rNQU70f2JiYvD1118jOzsbANCyZUvMmDEDw4cPF1xGVP24TZeIiIhIkNLSUjRs2LDKc2tra5SWlgooIpHWrFmDefPmYeLEifD29oZSqURSUhLGjx+Px48fqw76IqotuDJKREREJEj37t3RoEEDxMTEwMDAAABQVlaGkJAQFBcX4+TJk4ILSUoODg5YtGgRRowYofY8OjoaCxcuRF5enqAyIs3gMEpEREQkSEZGBvr06YPy8nJ4eHhAJpPhypUrMDAwwO+//442bdqITiQJGRgYICMjA05OTmrPs7Oz4ebmhvLyckFlRJqhIzqAiIiItM/ChQuRn58vOkM4V1dXZGdn46uvvkLbtm3h7u6O5cuXIzs7m4OoFnJycsLevXurPN+zZw9atGghoIhIs7gySkRERJJr164d0tLS4Ofnh9DQUAwaNEi1TZVIW/30008YMmQIevToAW9vb8hkMiQmJiI2NhZ79+7FwIEDRScSVSsOo0RERCREeno6IiMjsXPnTrx69QpDhw7F6NGj0aFDB9FpksrKykJ8fDwePnyoumv0jfnz5wuqIlEuX76MtWvXIjMzE0qlEi4uLggPD4enp6foNKJqx2GUiIiIhKqsrMTPP/+MyMhIHDt2DK1atcKYMWMwcuRImJqais7TqK1bt+LTTz+FpaUlGjVqpHanpEwmQ0pKisA6IiLN4jBKREREQr169QoHDx7EDz/8gLi4OHTp0gUPHjzA/fv3sXXrVgwZMkR0osbY2dlhwoQJmDVrlugUqiEUCgVu3br11pVyX19fQVVEmsFhlIiIiIS4fPkyIiMjsWvXLtStWxcjRozAmDFjVCeJrl69GitXrsSDBw8El2qOiYkJrly5AkdHR9EpVAMkJycjODgY+fn5+M8/ostkMsjlckFlRJrBYZSIiIgk5+7ujszMTPTq1Qtjx47Fe++9B11dXbXvPHr0CA0bNqyyOlSbhIaGokOHDhg/frzoFKoB2rZti5YtW2LRokVo3Lix2rZtALV+2zppHw6jREREJLklS5Zg9OjRsLGxEZ0iufXr16t+/OLFC6xZswb9+vWDm5sb9PT01L4bFhYmdR4JVL9+faSlpVW5Z5SotuIwSkRERMK8evUKeXl5aN68OerUqSM6RxIODg7/6nsymQy5ubkarqGapFu3bpg5cyb69OkjOoVIEhxGiYiISHJlZWWYOHEioqOjAby+3sTR0RFhYWFo0qQJZs+eLbiQSBrp6emqH+fk5OCLL77AjBkz3rpS7u7uLnUekUZxGCUiIiLJTZ48GUlJSYiIiECfPn2Qnp4OR0dHHDlyBAsWLEBqaqroRCHkcjmuXr0KOzs7mJubi84hCejo6EAmk1U5sOiNN5/xACOqjbRjPwwRERHVKIcOHcKePXvwzjvvqB3S4uLigpycHIFl0poyZQrc3NwQGhoKuVwOX19fnDt3DoaGhvjll1/g7+8vOpE0LC8vT3QCkTAcRomIiEhyjx49grW1dZXnL168qHKCaG22f/9+fPzxxwCAn3/+Gbdv38aNGzcQExODzz//HElJSYILSdPs7OxEJxAJoyM6gIiIiLRPhw4d8Ouvv6p+/mYA3bp1Kzp37iwqS3KPHz9Go0aNAAC//fYbBg8ejJYtWyI0NBRXr14VXEdSi46OVvt9MXPmTJiZmaFLly7Iz88XWEakGRxGiYiISHJfffUVPv/8c3z66aeorKzEunXr0LNnT0RFRWHZsmWi8yTTsGFDXL9+HXK5HMeOHUOPHj0AAKWlpVXuXaXa78svv0S9evUAAOfOncOGDRuwcuVKWFpaYurUqYLriKofh1EiIiKSXJcuXZCUlITS0lI0b94cx48fR8OGDXHu3Dm0a9dOdJ5kRo0ahQ8//BCurq6QyWTo2bMnAOD8+fNo3bq14DqS2p07d1R3jB46dAgffPABxo0bh6+++gpnzpwRXEdU/fjOKBEREQnh5uamutpFWy1cuBCurq64c+cOBg8ejLp16wIAdHV1eb2NFjIyMsKTJ0/QrFkzHD9+XLUaamBggLKyMsF1RNWPV7sQERGREAqFArdu3cLDhw+hUCjUPvP19RVURSTOsGHDcOPGDXh6emLXrl0oKChAgwYNcOTIEcydOxcZGRmiE4mqFVdGiYiISHLJyckIDg5Gfn5+lfsVeZ8iaauNGzfiiy++wJ07d/DTTz+hQYMGAIDLly/jo48+ElxHVP24MkpERESSa9u2LVq2bIlFixahcePGVa5zMTU1FVRGRERS4TBKREREkqtfvz7S0tJUh7UQEZH24Wm6REREJLlOnTrh1q1bojOIiEggvjNKREREkkhPT1f9eNKkSQgPD0dRURHc3Nygp6en9l13d3ep84TJyclBZGQkcnJysG7dOlhbW+PYsWOwtbVFmzZtROcREWkMt+kSERGRJHR0dCCTyaocWPTGm8+06QCjhIQE9O3bF97e3jh9+jQyMzPh6OiIlStX4sKFC9i/f7/oRCIijeEwSkRERJLIz8//19+1s7PTYEnN0blzZwwePBjTpk2DsbEx0tLS4OjoiIsXLyIoKAj37t0TnUgSKisrg1KphKGhIYDXv2cOHjwIFxcX9OrVS3AdUfXjNl0iIiKShJ2dHUaPHo1169bB2NhYdE6NcPXqVezcubPKcysrKzx58kRAEYk0YMAADBo0COPHj8ezZ8/QqVMn6Onp4fHjx1izZg0+/fRT0YlE1YoHGBEREZFkoqOjUVZWJjqjxjAzM0NhYWGV56mpqbCxsRFQRCKlpKTAx8cHALB//340bNgQ+fn5iImJwfr16wXXEVU/DqNEREQkGb4dpC44OBizZs1CUVERZDIZFAoFkpKSMH36dIwYMUJ0HkmstLRUtWvg+PHjGDRoEHR0dPDOO+/8f21zJ/pfwWGUiIiIJCWTyUQn1BjLli1Ds2bNYGNjg+fPn8PFxQW+vr7o0qULvvjiC9F5JDEnJyccOnQId+7cwe+//656T/Thw4cwMTERXEdU/XiAEREREUlGR0cHpqam/3UgLS4ulqhIjFu3bsHJyUn185ycHKSmpkKhUMDT0xMtWrQQWEei7N+/H8HBwZDL5ejevTuOHz8OAPjqq69w+vRpHD16VHAhUfXiMEpERESS0dHRQUREBExNTf/xeyEhIRIViaGjowMbGxsEBASgW7duCAgI0JoThOmfFRUVobCwEB4eHtDReb2J8cKFCzAxMUHr1q0F1xFVLw6jREREJBkdHR0UFRXB2tpadIpQZ86cQUJCAuLj43Hu3DmUl5ejWbNmqsE0ICCABxhpmcrKShgYGODKlStwdXUVnUMkCQ6jREREJBldXV0UFhZq/TD6VxUVFTh37hzi4+MRHx+P5ORkvHz5Ek5OTrh586boPJJQ8+bNceDAAXh4eIhOIZIEh1EiIiKSDFdG/15ZWRkSExPx+++/Y+vWrXj+/DnkcrnoLJJQZGQk9u3bh+3bt8PCwkJ0DpHGcRglIiIiEqC8vBxnz57FqVOnEB8fj4sXL8LBwQF+fn7w9fWFn58ft+pqGU9PT9y6dQsVFRWws7ND/fr11T5PSUkRVEakGXVEBxARERFpGz8/P1y8eBHNmzeHr68vJk2aBD8/PzRs2FB0GgkUFBQkOoFIUlwZJSIiIpKYnp4eGjdujKCgIPj7+8PX1xeWlpais4iIJKUjOoCIiIhI2zx79gzfffcdDA0NsWLFCtjY2MDNzQ0TJ07E/v378ejRI9GJJMizZ8/w/fffY86cOar7dlNSUnDv3j3BZUTVjyujRERERIL9+eefSExMVL0/mpaWhhYtWiAjI0N0GkkoPT0dPXr0gKmpKW7fvo2bN2/C0dER8+bNQ35+PmJiYkQnElUrrowSERERCVa/fn1YWFjAwsIC5ubmqFOnDjIzM0VnkcSmTZuGkSNHIjs7GwYGBqrnffv2xenTpwWWEWkGDzAiIiIikphCocClS5cQHx+PU6dOISkpCS9evICNjQ0CAgKwceNGBAQEiM4kiV28eBFbtmyp8tzGxgZFRUUCiog0i8MoERERkcTMzMzw4sULNG7cGP7+/lizZg0CAgLQvHlz0WkkkIGBAUpKSqo8v3nzJqysrAQUEWkW3xklIiIiktiWLVsQEBCAli1bik6hGmTcuHF49OgR9u7dCwsLC6Snp0NXVxdBQUHw9fVFRESE6ESiasVhlIiIiIioBigpKUFgYCCuXbuGP//8E02aNEFRURE6d+6M3377DfXr1xedSFStOIwSEREREdUgcXFxSElJgUKhgJeXF3r06CE6iUgjOIwSERERERGR5HiAERERERFRDREbG4vY2Fg8fPgQCoVC7bMffvhBUBWRZnAYJSIiIiKqARYtWoTFixejffv2aNy4MWQymegkIo3iNl0iIiIiohqgcePGWLlyJYYPHy46hUgSOqIDiIiIiIgIePXqFbp06SI6g0gyHEaJiIiIiGqAMWPGYOfOnaIziCTDbbpERERERIJMmzZN9WOFQoHo6Gi4u7vD3d0denp6at9ds2aN1HlEGsVhlIiIiIhIkICAgH/1PZlMhri4OA3XEEmLwygRERERERFJju+MEhERERHVAH/88QeKi4urPC8uLkZJSYmAIiLN4jBKRERERFQDDB06FLt3767yfO/evRg6dKiAIiLN4jZdIiIiIqIawMLCAklJSXB2dlZ7fuPGDXh7e+PJkyeCyog0gyujREREREQ1wMuXL1FZWVnleUVFBcrKygQUEWkWh1EiIiIiohqgQ4cO+O6776o837x5M9q1ayegiEiz6ogOICIiIiIiYNmyZejRowfS0tLQvXt3AEBsbCwuXryI48ePC64jqn58Z5SIiIiIqIa4cuUKvv76a1y5cgX16tWDu7s75syZgxYtWohOI6p2HEaJiIiIiIhIctymS0RERERUw5SVlaGiokLtmYmJiaAaIs3gAUZERERERDVAaWkpJk6cCGtraxgZGcHc3FztH6LahsMoEREREVENMGPGDMTFxeHbb79F3bp18f3332PRokVo0qQJYmJiROcRVTu+M0pEREREVAM0a9YMMTEx8Pf3h4mJCVJSUuDk5IQff/wRu3btwm+//SY6kahacWWUiIiIiKgGKC4uhoODA4DX74cWFxcDALp27YrTp0+LTCPSCA6jREREREQ1gKOjI27fvg0AcHFxwd69ewEAP//8M8zMzMSFEWkIt+kSEREREdUAa9euha6uLsLCwnDq1Cn069cPcrkclZWVWLNmDSZPniw6kahacRglIiIiIqqBCgoKcOnSJTRv3hweHh6ic4iqHYdRIiIiIqIapry8HAYGBqIziDSK74wSEREREdUAcrkcS5YsgY2NDYyMjJCbmwsAmDdvHrZt2ya4jqj6cRglIiIiIqoBli1bhqioKKxcuRL6+vqq525ubvj+++8FlhFpBodRIiIiIqIaICYmBt999x2GDRsGXV1d1XN3d3fcuHFDYBmRZnAYJSIiIiKqAe7duwcnJ6cqzxUKBSoqKgQUEWkWh1EiIiIiohqgTZs2OHPmTJXn+/btg6enp4AiIs2qIzqAiIiIiEibjR49GuvWrcOCBQswfPhw3Lt3DwqFAgcOHMDNmzcRExODX375RXQmUbXj1S5ERERERALp6uqisLAQ1tbW+P333/Hll1/i8uXLUCgU8PLywvz589GrVy/RmUTVjsMoEREREZFAOjo6KCoqgrW1tegUIknxnVEiIiIiIsFkMpnoBCLJcWWUiIiIiEggHR0dmJqa/teBtLi4WKIiImnwACMiIiIiIsEWLVoEU1NT0RlEkuLKKBERERGRQHxnlLQV3xklIiIiIhKI74uStuIwSkREREQkEDcqkrbiNl0iIiIiIiKSHFdGiYiIiIiISHIcRomIiIiIiEhyHEaJiIiIiIhIchxGiYiIiIiISHIcRomIiIiIiEhyHEaJiIiIiIhIchxGiYiIiIiISHL/DwCBQO1bhbGZAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.heatmap(similarity_score, annot=True, cmap='viridis', xticklabels=sents, yticklabels=sents)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The heatmap shows, that the cosine-similarity between the embeddings of our word `show` in sentence $i$ and the same word in the context of the same sentence $j=i$ is 1. However, for $j \\neq i$ the cosine similarities are $<1$, which proves that our word `show` has distinct embeddings in the distinct contexts. The embeddings in sentence 0 and sentence 1 show the smallest similarity score. This is plausible, because in sentence 1 the word `show` has a significantly different meaning than in sentence 0. The highest value for different sentences is obtained for sentence 3 and sentence 5. This is also plausible, because in these two sentences the meaning of the word `show` is similar."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "ws2324",
   "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": "3.11.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}